Ion pair first and second acidities of some β-diketones and aggregation of their lithium and cesium enediolates in THF

Article abstract of DOI:10.1021/jo0491915

Ion pair pK values were measured for three β-diketones in THF, 1-3, with lithium and cesium counterions. The results showed variations with concentration indicative of aggregation of the metal enolates to dimers. Similarly, ion pair pK values could be det

Full text of DOI:10.1021/jo0491915

Ion Pair First and Second Acidities of Some â-Diketones and
Aggregation of Their Lithium and Cesium Enediolates in THF
Antonio Facchetti† and Andrew Streitwieser*,‡
Department of Chemistry, University of California, Berkeley, California 94720-1460, and Department of
Chemistry and the Materials Research Center, Northwestern University, Evanston, Illinois 60208-3113
astreit@socrates.berkeley.edu
Received May 13, 2004
Ion pair pK values were measured for three â-diketones in THF, 1-3, with lithium and cesium
counterions. The results showed variations with concentration indicative of aggregation of the metal
enolates to dimers. Similarly, ion pair pK values could be determined for some of these metal
enolates going to the corresponding dimetal dienediolates which were also found to form dimers.
These equilibria are more complicated to analyze because aggregation affects both sides of the
proton transfer equilibria. The results show that all of the species measured exist mostly as dimers
at concentrations >0.01 M typical of most organic synthesis reactions and physical measurements.
NMR measurements show that the enols of 1 and 2, which can undergo intramolecular hydrogen
bonding, predominate in both THF and DMSO solutions, whereas 3, whose enols cannot be so
stabilized, is mostly keto in THF but approximately equimolar enol and keto in DMSO. Dimerization
of the monolithium salts is rapid on the NMR time scale but that of the dilithium salts is slow.
Introduction
reagents for â-dicarbonyl dianions.14-17 Aldol addition
reactions2,4,18 of these dianions have been explored to a
lesser degree than alkylation. However, the aldol prod-
ucts can spontaneously cyclize to hemiketals or lactones
depending on the nature of the dianion and this property
has been exploited in the preparation of novel oxygen-
based heterocycles. Yumaguchi and co-workers have
made extensive use of the acylation of â-dicarbonyl
dianions in the synthesis of polyketides, phenols, and aryl
C-glucosides.16 The dianions of simple â-diketones have
been acylated and used in the synthesis of stegobinone19
and xanthones.20
Metal enolates of ketones, amides, esters, and other
carbon derivatives are known to be frequently aggregated
in ethereal solvents and in recent years this group has
determined a number of such aggregation equilibrium
constants and the relative ion pair acidities of the
monomers.21-32 Recently, the general approach was ap-
â-Dicarbonyl functions in aldehydes, esters, or ketones
can generally be deprotonated stepwise between the
carbonyl groups to give monoanions and then at the
external ketone enolate site to give 1,3-dienediolate
dianions. Mono-1 and dianions2,3 arising from the mono-
or bis-deprotonation of â-dicarbonyl compounds have
been widely used in organic synthesis because of their
ready access, predictable reactivity, and expanded reac-
tion repertoire. Several reviews have covered the prepa-
ration and reactivity of these systems; their use is
commonplace in total synthesis and methodology as a
reliable tool when three-, four-, or five-carbon-chain
extension is required. Alkylation reactions of mono-4-8
and dienediolates2,9,10 of â-dicarbonyl compounds with
alkyl halides continue to be attractive pathways to
substituted ketoesters and diketones as well as novel
oxygen-11,12 and nitrogen-containing12,13 heterocycles. Ep-
oxides have also been used extensively as alkylating
(11) Marzinzik, A. L.; Felder, E. R. Molecules 1997, 2, 17.
(12) Nasibov, S. S.; Sadykhov, N. S.; Muradava, F. M. Russ. Chem.
Bull. 1997, 46, 1618.
* Address correspondence to this author.
† Northwestern University.
(13) Osowska-Pacewicka, K.; Zwierzak, A. Synth. Commun. 1988,
28, 1127-1137.
‡ UC Berkeley.
(1) House, H. O. Modern Synthetic Reactions; W. A. Benjamin,
Inc.: Menlo Park, CA, 1972.
(14) Kowalski, C.; Haque, M. S.; Fields, K. W. J. Am. Chem. Soc.
1985, 107, 1429.
(2) Thompson, C. M.; Green, D. L. C. Tetrahedron 1991, 47, 4223-
4285.
(15) Mori, K.; Sasaki, M.; Tamada, T.; Suguro, T.; Masuda, S.
Heterocycles 1978, 10, 111.
(3) Thompson, C. M. Dianion Chemistry in Organic Synthesis; CRC
Press: Boca Raton, FL, 1994.
(16) Yamaguchi, M.; Hirao, I. Chem. Lett. 1985, 337.
(17) Lygo, B. Tetrahedron 1988, 44, 6889.
(18) Peterson, J. R.; Winter, T. J.; Miller, C. P. Synth. Commun.
1988, 28, 949.
(4) Caine, D. In Carbon-Carbon Bond Formation; Marcel Dekker:
New York, 1979; Vol. 1, pp 90-341.
(5) Li, C.-J.; Lu, Y.-Q. Tetrahedron Lett. 1995, 36, 2721.
(6) Ranu, B. C.; Bhar, S. J. Chem. Soc., Perkin Trans. 1 1992, 385.
(7) Antonioletti, R.; Bonadies, F.; Orelli, L. R.; Scettri, A. Gazz.
Chim. Ital. 1992, 237.
(19) Gil, P.; Razkin, J.; Gonzalez, A. Synthesis 1998, 4, 386.
(20) Chandrasekhar, B.; Ramadas, S. R.; Ramana, D. V. Tetrahedron
2000, 56, 5947.
(21) Abbotto, A.; Streitwieser, A. J. Am. Chem. Soc. 1995, 117,
6358-6359.
(8) Shone, T.; Kashimura, S.; Sawamura, M.; Soejima, T. J. Org.
Chem. 1988, 53, 907.
(9) Harris, T. M.; Harris, C. M. Org. React. 1969, 17, 1.
(10) Kaiser, E. M.; Petty, J. D.; Knutson, P. L. A. Synthesis 1979,
509.
(22) Abu-Hasanayn, F.; Stratakis, M.; Streitwieser, A. J. Org. Chem.
1995, 60, 4688-4689.
(23) Gareyev, R.; Ciula, J. C.; Streitwieser, A. J. Org. Chem. 1996,
61, 4589-4593.
10.1021/jo0491915 CCC: $27.50 © 2004 American Chemical Society
Published on Web 10/26/2004
J. Org. Chem. 2004, 69, 8345-8355
8345

Facchetti and Streitwieser
SCHEME 1
plied to the lithium and cesium enediolates of an aryl-
acetic acid,33 which were found to form tight dimers.
The knowledge of the actual reactive species involved
in reactions of such systems is fundamental for mecha-
nism hypotheses and particularly to understand reaction
stereochemistry. Accordingly, in this paper we applied
these approaches to the lithium and cesium enolates of
the dicarbonyl systems 1-3. Since UV-vis spectroscopy
was a principal tool in this study, the aryl groups were
required as chromophores. Moreover, at the low concen-
trations inherent in UV-vis measurements, addition of
the anions to carbonyl groups was slow and did not affect
our study. Compound 1 has a terminal methyl group for
forming the dianion whereas in 2 the dianion has further
conjugation. Both of these compounds form monoanion
ion pairs expected to be of U-shape.34 To compare with a
system constrained to a W-shape we included compound
3.
The resulting ∆pK values are converted for convenience
to absolute numbers36 by setting the pK of fluorene equal
to its ionic pKa of 22.9 (per hydrogen) in DMSO.37 We
have emphasized previously that such “ion pair pK
values” are not pKa values.38
Since all of the indicator anions are known to exist
exclusively as monomeric ion pairs in THF, the indicator
technique can be used to determine the presence of
substrate anion ion pair aggregates. Aggregation of the
enolate ion pairs, eq 3, pulls equilibrium 1 to the right
resulting in an observed pK value lower than that in the
absence of aggregation. In addition, due to the stoichi-
ometry of eq 3, the effect of enolate aggregation is to
increase the observed acidity of the corresponding acidic
partner as the total enolate concentration is increased.
In this paper we employ this principle to determine the
ion pair acidities of lithium and cesium mono- and
dienolates of compounds 1-3 and to study their degrees
of aggregation in THF solution. We compare these results
with other classes of organic compounds previously
studied in this laboratory.
n(RH- M+) h (RH- M+)n
n(R2- M+2) h (R2- M+
(3a)
(3b)
As described previously, the ion pair acidity difference
between a given system and a reference indicator is
defined by the equilibrium in eq 1 with the corresponding
pK difference given by eq 2.35
)
2 n
Finally, 1H and 13C NMR studies in THF-d8 and
DMSO-d6 provide structural details of the neutral and
anionic species derived from 1-3.
In- M+ + RH2 yK1
\
z InH + RH- M+
(1a)
K2
\
In- M+ + RH- M+ y
z InH + R2- 2M+
(1b)
(2)
Results
∆pK ) -logK1 ) pK[RH2] - pK[lnH]
Synthesis. 1-(4-Biphenylyl)butane-1,3-dione (1)39 and
1,4-diphenylbutane-1,3-dione (2)40 were prepared with
slight modifications of known procedures according to
Scheme 1. Phenyl- and 4-biphenylyl methyl ketone were
deprotonated with EtONa in ethanol and reacted with
ethyl acetate or ethyl phenylacetate, respectively, to
afford the corresponding diketones 1 or 2 in high yields.
To the best of our knowledge, cyclic diketone 3 is
unknown in the literature. It was prepared starting from
3-(2-acetylphenyl)propionic acid, which was converted
into the corresponding ethyl ester by reaction with SOCl2
and quenching with ethanol. Cyclization of the ketoester
(Scheme 2) was performed in refluxing t-BuOH with
t-BuOK as base.
In the present case, two protons can be removed to give
first and second equilibrium constants and ∆pK values.
(24) Krom, J. A.; Streitwieser, A. J. Org. Chem. 1996, 61, 6354-
6359.
(25) Streitwieser, A.; Krom, J. A.; Kilway, K. A.; Abbotto, A. J. Am.
Chem. Soc. 1998, 120, 10801-10806.
(26) Streitwieser, A.; Wang, D. Z.; Stratakis, M. J. Org. Chem. 1999,
64, 4860-4864.
(27) Streitwieser, A.; Wang, D. Z.-R. J. Am. Chem. Soc. 1999, 121,
6213-6219.
(28) Wang, D. Z.-R.; Streitwieser, A. Can. J. Chem. 1999, 77, 654-
658.
(29) Wang, D. Z.; Kim, Y.-J.; Streitwieser, A. J. Am. Chem. Soc.
2000, 122, 10754-10760.
(30) Streitwieser, A.; Kim, Y.-J.; Wang, D. Z.-R. Org. Lett. 2001, 3,
2599-2601.
(31) Abbotto, A.; Leung, S. S.-W.; Streitwieser, A.; Kilway, K. V. J.
Am. Chem. Soc. 1998, 120, 10807-10813.
(32) Facchetti, A.; Streitwieser, A. J. Org. Chem. 1999, 64, 2281-
2286.
(36) Kaufman, M. J.; Gronert, S.; Streitwieser, A., Jr. J. Am. Chem.
Soc 1988, 110, 2829-2835.
(37) Bordwell, F. G. Acc. Chem. Res. 1988, 21, 456-463.
(38) Streitwieser, A.; Kim, Y.-J. J. Am. Chem. Soc. 2000, 122,
11783-11786.
(33) Streitwieser, A.; Husemann, M.; Kim, Y.-J. J. Org. Chem. 2003,
68, 7937-7942.
(34) Zaugg, H. E.; Schaefer, A. D. J. Am. Chem. Soc. 1965, 87, 1857-
1866.
(35) Bors, D. A.; Kaufman, M. J.; Streitwieser, A., Jr. J. Am. Chem.
Soc. 1985, 107, 6975-6982.
(39) Spraugue, J. M.; Beckham, L. J.; Adkins, H. J. Am. Chem. Soc.
1934, 56, 2665-2668.
(40) Padwa, A.; Hornbuckle, S. F.; Zhang, Z.; Zhi, L. J. Org. Chem.
1990, 55, 5297.
8346 J. Org. Chem., Vol. 69, No. 24, 2004

Aggregation of Metal Enolates to Dimers
SCHEME 2
whereas the lithium scale42,43 involves solvent-separated
ion pairs (SSIP). In the case of dienediolates (1, 2)-Li2
and (1, 2)-Cs2, both ion pairs involved in eq 1 exhibit
sufficiently distinct visible absorption spectra so that the
concentrations of the species at equilibrium could be
determined by the use of the double-indicator method.
In this technique the concentrations of the ion pairs are
determined by linear least-squares fitting of the spectra
of the separate species (indicator and enolate) to the
spectrum of the equilibrium mixture. In contrast, the
monoenolates (1-3)-Li and (1-3)-Cs and the dienediolate
3-Li2 do not have absorption maxima in the visible region.
Although these enolate ion pairs exhibit distinct absorp-
tion in the near-UV, this region of the spectrum cannot
be analyzed quantitatively due to the large number of
overlapping peaks arising from both indicators and the
neutral dicarbonyl substrates (see Figures S1-S3, Sup-
porting Information). Consequently, we used the single
indicator technique in which the spectrum of the indica-
tor salt of known pK is recorded and the initial anion
concentration is determined from the absorbance at the
appropriate λmax. A measured amount of substrate RH2
is then added and the equilibrium concentration of the
indicator is determined directly from the new absorbance
value. The equilibrium enolate concentration is then
calculated with eq 4.
TABLE 1. Spectroscopic Data for Lithium and Cesium
Enolates of 1-3 in THF at 25 °C
compd
enolates
λmax (nm)
ꢀ (cm-1 M-1
)
1
-H
324
29 600
21 500
12 950
17 500
8 800
18 300
19 500
19 400
15 900
24 700
23 600
-Li
340-342a
379-381a
350-355a
418-421a
312
-Li2
-Cs
-Cs2
-H
2
-Li
333
-Li2
-Cs
-Cs2
400-402a
338-340a
430-438a
444b
3c
-H
<290
-Li
-Li2
-Cs
329
13 300
3 430
337
336
11 000d
a Depending on concentration. b Data are for the isosbestic point.
c Dicesium enolate is not soluble. d Monocesium enolate is poorly
soluble. ꢀ is (20%.
UV-vis Measurements: Ion Pair Absorption Spec-
tra. With the exception of 3 and its completely insoluble
dicesium enolate 3-Cs2, all of the neutral, monometal, and
dimetal enolates (metal ) Li, Cs) of â-dicarbonyl com-
pounds 1-3 were found to absorb in the visible or near-
UV region in THF (Figures S1-S3, Supporting Informa-
tion). The extinction coefficient (ꢀ) of each ion pair was
determined by recording a series of 22-35 absorption
spectra at different metal enolate concentrations (Figures
S4-S14, Supporting Information). A detailed description
is reported in the Experimental Section. Due to its poor
solubility, the extinction coefficient of the cesium enolate
3-Cs was obtained at very low anion concentrations,
which prevented further dilution for each experiment. For
this reason, this value has lower accuracy because only
five absorption spectra were obtained (Figure S15, Sup-
porting Information). The absorption spectra of monoeno-
lates 1-Li, 1-Cs, and 2-Cs were found to be slightly
concentration dependent, with a shift of the maximum
absorption (λmax) ranging from 2 to 5 nm. For the
dienediolates of 1 and 2 the shift of λmax ranges from 2 to
8 nm. No appreciable shift was observed for the other
enolate systems. Extinction coefficient measurements of
2-Cs2 at varying concentrations reveal an isosbestic point
at 444 nm (vide infra). No clear isosbestic point could be
detected for the other enolates, probably because the
concentration dependencies of their spectra are smaller
than the experimental error in the measurements. In
Table 1 are collected the spectroscopic data, λmax, and
extinction coefficient at λmax and/or at the isosbestic point
of neutral systems 1 and 2, and of the soluble lithium
and cesium enolates of 1-3.
{RH- M+}eq ) [In- M+]initial - [In- M+]eq (4)
Details on the single44 and double36 indicator methods
were reported previously.
Initially we ensured that pK measurements would not
be affected by the base employed to deprotonate the
indicator or the mixture indicator-substrate. Reproduc-
ible pK results showed that a nitrogen-containing base
such as diisopropylamine does not influence equilibrium
1 by forming aggregates with lithium enolates. Moreover,
strong nuclophiles such as DPMCs or TMSLi do not react
either with the dicarbonyl or with the monoenolcarbonyl
species. However, it is important in determining the
acidity of dicarbonyl systems to simultaneously maintain
an equilibrium concentration of either mono- or dimeta-
lated enolate and the corresponding conjugate acid. This
represents a potential problem since enolate ions can add
to carbonyl groups. Furthermore, in principle, addition
of the indicator anion to the CdO bond might also occur.
This could be particularly true for those â-dicarbonyl
(41) Streitwieser, A.; Ciula, J. C.; Krom, J. A.; Thiele, G. J. Org.
Chem. 1991, 56, 1074-1076.
(42) Facchetti, A.; Kim, Y.-J.; Streitwieser, A. J. Org. Chem. 2000,
65, 4195-4197.
Ion Pair Acidity. The lithium and cesium ion pair
acidities of the dicarbonyl systems 1-3 were measured
against hydrocarbon indicators (Table 2) with known pK.
The cesium scale41,42 involves contact ion pairs (CIP),36
(43) Streitwieser, A.; Wang, D. Z.; Stratakis, M.; Facchetti, A.;
Gareyev, R.; Abbotto, A.; Krom, J. A.; Kilway, K. V. Can. J. Chem.
1998, 76, 765-769.
(44) Streitwieser, A., Jr.; Reuben, D. M. E. J. Am. Chem. Soc. 1971,
93, 1794-1795.
J. Org. Chem, Vol. 69, No. 24, 2004 8347

Facchetti and Streitwieser
TABLE 2. Experimental Details To Measure Lithium and Cesium Ion Pair Acidities of Enolates 1-3
compd
enolates
basea
time (h)b
indicatorc (pK, ꢀd)
methode
1
-Li
TMSLi, LDA
LDA
DPMCs, CuCs
CumCs
CumCs
TMSLi
4
48
5
f
f
0.5
16
4
0.5
0.25
60
f
DFB (5.61, 79 600)
BnFl (21.36, 15 400)
BBP (12.29, 107 600)
TpTM (33.1, 30 200)
TPM (31.26, 29 900)
DFB
DMAPhFl (19.02, 27 800)
BBP
DP5 (25.62, 117 000)
DFB
S
D
S
D
D
S
D
S
D
S
-Li2
-Cs
-Cs2
2
3
-Li
-Li2
-Cs
-Cs2
-Li
-Li2
-Cs
LDA
DPMCs
CumCs
TMSLi
LDA
DPMCs
2,3-BF (22.95, 25 500)
BBP
S
S
a Base for forming enolate: TMSLi, trimethylsilylmethyllithium; LDA, lithium diisopropylamide; DPMCs, diphenylmethylcesium; CumCs,
cumylcesium. b Approximate time to reach equilibrium. c Indicator for measurement: DFB, 3-(dibenzofulvenyl)-6-(9-fluorenyl)-1,2-
benzofulvene; BnFl, 9-benzylfluorene; BBP, 1,3-bis(biphenylene)propene; TpTM, tri-p-tolylmethane; TPM, triphenylmethane; DMAPhFl,
9-dimethylaminophenylfluorene; DP5, 1,5-diphenyl-1,3-pentadiene; 2,3-BF, 2,3-benzofluorene. d ꢀ in cm-1 M-1 e S, single indicator method.
.
D, double indicator method. f These cesium enolates precipitate before an unchanging absorption spectrum of the equilibrating mixture
was reached.
FIGURE 1. The pK values of the monolithium enolates vary
with concentration. The regression lines are as follows: 1-Li,
1.34 ( 0.27 - (0.265 (0.068)x; 2-Li, 0.952 ( 0.142 - (0.277 (
0.036)x; 3-Li, 2.58 ( 0.095 - (0.376 ( 0.022)x.
FIGURE 2. The pK values of the monocesium enolates vary
with concentration. The regression lines are as follows: 1-Cs,
10.31 ( 0.12 - (0.055 ( 0.027)x; 2-Cs, 9.75 ( 0.18 - (0.115 (
0.044)x; 3-Cs, indeterminate because of solubility limitations.
systems that are mainly present in THF solution as the
keto tautomer. The tautomer equilibria will be discussed
further in a later section. All of these secondary reactions
would affect the position of the acid-base equilibrium
and result in erroneous and irreproducible ion pair
acidity values particularly in cases with long proton
transfer equilibrium times. In practice, such secondary
reactions were not an important problem, probably
because of the high dilutions required for the UV-vis
measurements. They might have contributed to the
observed experimental uncertainties but these are rela-
tively small, especially considering the difficulty of these
measurements. Table 2 also shows the time needed to
establish the proton-transfer equilibria. We assumed that
the species present in eq 1 reached equilibrium when the
UV-vis spectrum of the corresponding mixture did not
change appreciably after 1 h. Previous work from this
laboratory had shown23,45 that proton transfer equilibria
between weak carbon acids are achieved much more
rapidly with cesium compared to lithium counterions.
The current study confirms this general trend. In fact,
reactions involving the mono-/dilithium salts of 1-3
required long equilibration times probably because the
Li-O bond is such a tight ion pair. On the other hand,
the first deprotonations with lithium bases were found
to be reasonably rapid.
Monoenolates. Plots of the observed ion pair acidities
(pK values) vs the logarithm of the metal enolate
concentration are presented in Figures 1 and 2. The
pK values decrease with increasing concentration in-
dicative of aggregation. The average aggregation num-
bers, n, given from the slopes (eq 5)35 are summarized in
1 - n
n
(45) Xie, L.; Streitwieser, A. J. Org. Chem. 1995, 60, 1339-
1346.
pKobs
)
log{enolate} + b
(5)
8348 J. Org. Chem., Vol. 69, No. 24, 2004

Aggregation of Metal Enolates to Dimers
TABLE 3. Summary of Average Aggregation Numbers
(n), Dimerization Constants (K1,2), and Ion Pair pK
Values of Lithium and Cesium Enolates of 1-3 in THF at
25 °C
na
Kd (M-1
)
pKo
b
compd
enolate
monoenolates
-Li
-Cs
-Li
-Cs
-Li
-Cs
dienolates
-Li2
-Li2
1
2
3
1.23 ( 0.02
1.10 ( 0.05
1.36 ( 0.05
1.10 ( 0.02
1.61 ( 0.05 26000 ( 3000
-0.38 ( 0.88
2200 ( 130
1000 ( 200
5100 ( 400
1050 ( 100
2.58
10.98
2.27
10.4
4.6
∼13
1
2
1.43 ( 0.03
1.29 ( 0.03
1.06 ( 0.02c
6000 ( 1000d
760 ( 100d
140 ( 20d
21.9
17.7
24.5
-Cs2
a Determined in the log concentration range of ca. -4.7 to -3.3.
b For monomers; statistical effects in the carbonyl compounds are
not considered. c Determined in the log concentration range of ca.
-5.8 to -4.2 d Error bounds determined by statistics of the Kobs
vs {dienediolate}/Kobs plots. Errors in Kd,1 (for the monoenolates)
are not included.
Table 3. The modest values suggest that the aggregates
are dimers. The scattering of points for 3-Cs is a
consequence of extreme dilution and no useful data could
be extracted for this compound. Note also that throughout
this paper concentrations expressed within curly brack-
ets, {}, are formal concentrations as measured by spec-
troscopy.
FIGURE 3. The circles are experimental pK values for
monoenolate (as acid) and dienediolate (as base) monomers
compared to points calculated from data in Tables 2 and 3 and
from Figures S19-S21 (Supporting Information).
tion of dienediolate, {dienediolate}, together with the
amount of indicator salt could generally be determined
directly from the spectra. The formal monoenolate con-
centration was then determined, in effect, by the single
indicator technique. From Kd,1 the monoenolate concen-
tration was deduced and used to determine the observed
ion pair acidity equilibrium constant, Kobs, with the
indicator of the second acidity constant of the ketone. A
plot of Kobs/{RM2} then gives Ko and the dimerization
constant of the dimetalated ketone, Kd,2. This procedure
could be applied to 1-Li2, 2-Li2, and 2-Cs2, with the
results summarized in Table 3. A few runs gave results
inconsistent with the others and were discarded. The
plots of Kobs vs Kobs/{dienediolate} used are shown in
Figures S19-S21 (Supporting Information). The results
were used to compute the calculated pK values for
comparison with observed values as a function of log-
{dienediolate} in Figure 3.
1H NMR Data and Keto/Enol Tautomerism in
Solution. The pK measurements of the monoenolates
refer to the state of the diketone in solution, generally
some equilibrating mixture of keto and enol tautomers.46,47
To refer the acidities to the individual tautomers meant
determining the keto-enol ratios. For example, aqueous
solutions of acetylacetone have been extensively studied
and were found at equilibrium to be approximately 20%
enol at room temperature.48-50 Base or solvent-promoted
deprotonation of either keto or enol gives the same
enolate anion. In our cases where the base is a metalated
indicator, these equilibria are summarized in Scheme 3.
For a monomer-dimer equilibrium an alternative
expression of the data is by a plot of Kobs vs {enolate-M}/
Kobs. The intercept gives Ko for the monomer and the slope
2 24
is 2KdKo . These plots are shown in Figures S17 and
S18 (Supporting Information) and the calculated dimer-
ization constants for the monoenolate, Kd,1, are sum-
marized in Table 3. The uncertainties are typically of the
order of 10% but some uncertainties are substantially
larger, probably because of the limitations of the single-
indicator method.
Note that 1/Kd is the concentration at which [monomer]
) [dimer] and thus at concentrations >0.001 M all of
these monoenolates are mostly dimer. These results
suggest that most studies in the literature (e.g., NMR)
of monoenolates in THF actually pertain to the dimers.
The dimerizaton constant for the W-enolate 3-Li is much
greater than that of the U-enolates 1-Li and 2-Li,
undoubtedly because the latter monomers are stabilized
by intramolecular coordination of lithium to both enolate
oxygens. The comparable values for 1-Cs and 2-Cs
suggest that the cesium enolate monomers are also
importantly stabilized by such internal coordination.
Dienediolates. Application of these methods to the
dialkali dienediolates by measuring the ion pair second
acidity constants of the â-diketones is not so straight-
forward. The problem is that we must now consider
aggregation in both the acid (the monoenolate) and the
base (the dienediolate). Aggregation of the acid increases
the observed pK and aggregation of the base lowers it.
In principle, one could calculate the amount of monoeno-
late monomer from the now known Kd,1 and formulate
the acidity equilibria in terms of the monomer. In
practice, this means having an indicator together with
the diketone and enough strong base to convert part of
the diketone to the dienediolate. The formal concentra-
(46) Emsley, J.; Freeman, N. J. J. Mol. Struct. 1987, 161, 193.
(47) Bunting, J. W.; Kanter, J. P.; Nelander, R.; Wu, Z. Can. J.
Chem. 1995, 73, 1305.
(48) Eidenhoff, M. L. J. Am. Chem. Soc. 1945, 67, 2073.
(49) Ahrenes, M. L.; Eigen, M.; Kruse, W. Ber. Bunsen-Ges. 1970,
74, 380.
(50) Moriyasu, M.; Kato, A.; Hashimoto, Y. J. Chem. Soc., Perkin
Trans. 2 1986, 515.
J. Org. Chem, Vol. 69, No. 24, 2004 8349

Facchetti and Streitwieser
SCHEME 3. Equilibria among Keto and Enol
Forms and Conjugate Enolate Ion Pairs
SCHEME 4. Keto-Enol Equilibria in DMSO and
THF
TABLE 4. Keq for the Keto-Enol Equilibria in THF and
DMSO and Calculated pKK and pKE in THF at 25 °C
a
b,c
b,c
Keq
pKE
pKK
-Li
compd
THF
DMSO
-Li
-Cs
-Cs
respectively. However, the enol form is largely the
predominant species in both solvents. This result is not
surprising considering that acyclic systems can form
strong intramolecular hydrogen bonds (75 kJ mol-1 for
acetylacetone)57 resulting in a stabilized cyclic enol
structure (Scheme 4). The strength of this intramolecular
bond would limit enol-solvent hydrogen bonding for most
solvents. On the other hand, as discussed above, the
W-shaped dicarbonyl moiety in 3 greatly affects the enol/
keto equilibrium. In THF only 8% of the enol form of 3
is present, which increases to about 50% in DMSO. In
this case, the more polar DMSO shifts the equilibrium
to the enol side rather than to the ketone, probably
because of intermolecular hydrogen bonding to the
solvent of the open enol form(s); DMSO is a better
hydrogen-bonding acceptor than THF.58
1
2
3
11.74 ( 1.4
113.47 ( 7.3
0.082 ( 0.002 1.08 ( 0.1 3.42 ∼12
3.77 ( 0.4 2.54
2.44 ( 0.1 2.27
10.94 1.48
10.4 0.21
9.88
8.3
4.52 ∼13
a [enol]/[keto]. b Probable errors are about (0.1 except for 3-Cs,
which are closer to (1. c For monomeric enolates without consid-
eration of statistical effects.
Important tools for measuring keto-enol equilibria
have generally been H NMR51,52 and IR53 spectroscopy.
1
In most cases, the conversion rates between the two
tautomers are sufficiently slow that the NMR spectra
show separate and distinct proton signals of both the keto
and enol tautomers. Hence, from the relative intensities
of the signals the equilibrium concentrations of each of
1
the two tautomers can be conveniently determined. H
NMR spectra of the neutral systems 1-3 (Figures S22-
S24, Supporting Information) were recorded in dilute
solution (0.01 M). THF and DMSO solutions were made
up and allowed to reach tautomer equilibrium over a
1-day period. In both solvents the enol/keto composition
could be determined by the integration of both keto and
enol proton resonances. Integration of the signals with
longer pulse delays (up to 10 s) gave similar results. The
results of three experiments (Table S2, Supporting
Information) were used to calculate Keq and averaged in
Table 4. From eqs 7 and 8, KK and KE were determined
and combined with the pKo results to give the ion pair
acidity pK values for the keto and enol tautomers also
summarized in Table 4.
The NMR investigation also points out the sensitivity
of the keto/enol equilibrium to solvent polarity. It is
widely recognized that this equilibrium has important
kinetic implications in â-diketone reactivity.54 For acyclic
1,3-diketones, polar solvents favor the more polar keto
side of the equilibrium.55,56In agreement with these
generalizations, we find that going from THF (ꢀ ) 7.6)
to DMSO (ꢀ ) 46.7) the fractions of the enol form of 1
and 2 decrease from 92% and 99% to 79% and 71%,
13C NMR Data and Geometric Isomerism in the
Anions. Carbon NMR spectroscopy can provide further
insights into the structures of 1-3 and of the correspond-
ing enolate anions. Selected 13C NMR resonances for all
of the investigated systems in THF-d8 and DMSO-d6 are
reported in Table 5. The spectra of the neutral and
monolithium species were recorded in 0.1 M solutions
whereas those of the dilithium dienediolates (1, 2)-Li2
were recorded as saturated solutions due to the poor
solubility of these enolates. Unfortunately, 3-Li2 was too
insoluble to be studied by NMR. Note that at 0.1 M
concentration (1-3)-Li and (1, 2)-Li2 are all present
almost wholly as dimers. Figures 4 and 5 (spectra A-C)
show decoupled spectra of 1, 2, and the corresponding
mono- and dilithium enolates in deuterated THF.
The NMR spectra of 1 and 2 exhibit one single set of
resonances. The small amount of the keto form present
was not detected (except for the Câ of 1). The internally
hydrogen-bonded enols of â-diketones are known to be
double-well systems59 but the proton shift has a low
barrier and is fast on the NMR time scale.60 The de-
coupled (A) and coupled (B) 13C NMR spectra of 3 in THF
and DMSO are shown in Figures S25 (Supporting
(51) Reeves, L. W. Can. J. Chem. 1957, 35, 1351.
(52) Reeves, L. W.; Schneider, W. G. Can. J. Chem. 1958, 36, 793.
(53) Isaacs, N. S. Experiments in Physical Organic Chemistry;
Macmillan: New York, 1969; p 132.
(57) Emsley, J.; Freeman, N. J.; Parker, R. J.; Overill, R. E. J. Chem.
Soc., Perkin Trans. 1 1986, 1479.
(58) Matthews, W. S.; Bares, J. E.; Bartmess, J. E.; Bordwell, F.
G.; Cornforth, F. J.; Drucker, G. E.; Margolin, Z.; McCallum, R. J.;
McCollum, G. J.; Vanier, N. R. J. Am. Chem. Soc. 1975, 97, 7006.
(59) Bauer, S. H.; Wilcox, C. F. Chem. Phys. Lett. 1997, 279, 122-
128.
(54) Toullic, J., Ed. Experiments in Physical Organic Chemistry;
Academic Press: London, UK, 1982.
(55) Emsley, J.; Freeman, N. J.; Parker, R. J.; Dawes, H. M.;
Hursthouse, M. B. J. Chem. Soc., Perkin Trans. 1 1986, 471.
(56) Emsley, J.; Freeman, N. J.; Bates, P.; Hursthouse, M. B. J. Mol.
Struct. 1987, 161, 181.
(60) Perrin, C. L.; Kim, Y.-J. J. Am. Chem. Soc. 1998, 120, 12641-
12645.
8350 J. Org. Chem., Vol. 69, No. 24, 2004

Aggregation of Metal Enolates to Dimers
TABLE 5. 13C and 1H NMR Chemical Shifts (δ)a of 1-3 and of the Corresponding Mono-b and Dilithiumc Enolates in
THF-d8 and DMSO-d6 at 25 °C
13C NMR
1H NMR
δ DMSO-d6
CâHx CσHy η,d % CâHx CσHy η,d %
δ THF-d8
Câ Cγ
H-enol 184.07 97.16 194.48 25.61 182.10
194.78
190.83 28.63 179.57
170.16 102.29 160.74 79.36
195.84 46.47 182.31
195.24
191.68 50.31 181.15
161.09 104.60 169.40 98.71
109.19
39.86 163.70g 107.70 198.74g 41.59 5.68
δ DMSO-d6
δ THF-d8
compd enolate
CR
Cσ
CR
Câ
Cγ
Cσ
1
2
3
96.88 194.16
53.67 203.77
94.23 189.40
25.41 6.40
30.65 4.11
29.00 5.91
2.17
2.21
92
99
8
6.63 2.22
4.33 2.26
5.84
79
71
52
H-keto
-Li
-Li2
e
55.00
e
e
182.06 95.11
f
H-enol 184.49 96.80
96.52 195.13
52.44 203.31
93.86 189.62
44.98 6.31
49.34 4.13
48.83 5.87
3.58
3.65
6.63 3.81
4.35 3.95
5.86
H-keto
-Li
e
e
e
e
183.78 95.24
f,g
-Li2
H-enol
H-keto 191.94 61.22
e
e
2.96
3.40
5.33
h
203.25 42.59 191.95
-Li 180.37 107.31 194.01 42.77 179.81
60.10 204.52
106.10 192.72
41.89 4.12
5.07
3.80 2.76
5.13
h
a For 1H and 13C NMR measurements concentrations are ∼0.01 and ∼0.1 M, respectively. b Lithium bis(trimethylsilyl)amine was used
as base. c Lithium tetramethylsilane was used as base. d Percent of the enol form determined by 1H NMR. e Signals were not detected.
f Broad. g Major isomer. h Under DMSO-d6 signal.
FIGURE 4. Decoupled 13C NMR spectra of 1 (A), 1-Li (B), and 1-Li2 (C) in THF-d8 at 25 °C. Expanded decoupled (D) and coupled
(E) spectra of 1-Li2.
Information) and 6, respectively. In THF only the Câ and
Cδ carbons of the minor enol form (8%) were detected.
All of the other signals are either isochronous with those
of the ketone tautomer or too weak to be detected. On
the other hand, three sets of resonances were clearly
detected for the DMSO-d6 solution of 3 at 25 °C. This is
clearly shown in the expanded region (spectrum C) in
Figure 6, which corresponds to the CR and Cγ carbon
signals. The three equilibrating species can be associated
with the keto form of 3 and its nonequivalent enol
isomers (Scheme 4). The detection of both enol forms of
3 undoubtedly results from the much slower external
proton transfer involved with hydrogen bonding to sol-
vent.
J. Org. Chem, Vol. 69, No. 24, 2004 8351

Facchetti and Streitwieser
FIGURE 5. Decoupled 13C NMR spectra of 2 (A), 2-Li (B), and 2-Li2 (C) in THF-d8 at 25 °C. Expanded decoupled (D) and coupled
(E) 2-Li2 spectra.
FIGURE 6. Decoupled (A) (expanded in part C), and coupled (B) 13C NMR spectra of 3 in DMSO-d6 at 25 °C.
The 13C NMR spectra of 1-Li and 2-Li in Figures 4 and
5 also exhibit only one set of resonances. A number of
possible dimer structures such as that shown as 4 can
exist as cis and trans isomers. A single set of resonances
then means that such dimers must interconvert rapidly
on the NMR time scale, probably via a rapid equilibrium
with monomer. The 13C NMR spectra of 3-Li in both THF-
d8 and DMSO-d6 also exhibit only one clear set of signals
8352 J. Org. Chem., Vol. 69, No. 24, 2004

Aggregation of Metal Enolates to Dimers
at room temperature. Two isomeric dimer structures
can be proposed for this compound also, one of which
is shown in the suggested structure 5. The fact of a single
set of signals suggests a rapid equilibration between
them unless the isomers are isochronous. A single set
of resonances with minimal variation of the chemical
shifts ((3 ppm) was also recorded down to -40 °C in
THF-d8 suggesting that the dissociation of 5 is quite
facile.
In this structure, unlike that of 4, there is no role for
cis-trans isomerism. However, the monomer of 2-Li2 can
exist as E,Z isomers as shown in 7. Thus, a cubic dimer
The analysis of the 13C NMR spectra of the dilithium
dienediolates (1, 2)-Li2 in THF-d8 is more complex. For
both systems, CR and Cγ carbons move substantially to
high field whereas Câ shifts slightly to low field compared
to the corresponding conjugate acids. There are appar-
ently two effects operating here. The first is the second
deprotonation, which increases electron density on these
atoms and should shift their resonances to higher field.61-64
The second is the Cδ rehybridization and conjugation with
the unsaturated moiety to form a diene-like structure.
This has a random effect on the carbon olefin frame as
seen from the 13C NMR of the couples ethylene/butadiene,
1-pentene/1,3-pentadiene, and phenylethene/1-phenyl-
butadiene.65 The overall chemical shift variations are
therefore combinations of these two effects. Finally, there
are profound differences in the spectra of 1-Li2 and 2-Li2.
The spectrum of 1-Li2 (Figure 4C) exhibits one single set
of broad signals, indicating that at 25 °C the dimer has
a single structure or possible isomers undergo fast inter-
conversion (in the NMR scale) but not fast enough for
coalescence, which occurs at about 45 °C. Low-temper-
ature experiments, which would have provided informa-
tion on the number of isomers, could not be performed
due to the poor solubility of the enolate. Nevertheless, a
reasonable structure for the dimer based on the ubiqui-
tous cube of four oxygens and four lithiums frequently
found in X-ray structures of lithium enolates66-68 can be
proposed as shown in 6. A similar structure was sug-
gested from ab initio MO computations for the dilithium
enediolate formed from dideprotonation of a carboxylic
acid.33
as in 6 could consist of EE, EZ, and ZZ species. These
isomers could account for the extra signals in the 13C
NMR spectrum of 2-Li2. Moreover, this compound has a
relatively low Kd and some of the minor signals could
arise from the two isomeric E,Z monomers. This requires
that the dilithio species, unlike the monolithium enolates,
dissociate slowly on the NMR time scale. In this connec-
tion, we recall that proton-transfer equilibria of 1-Li2 and
2-Li2 are quite slow, taking hours to reach equilibrium.
Similarly, the dimer of the dilithium enediolate from a
carboxylic acid is also very tight and dissociates slowly.33
Discussion
The cesium ion pair pK values of the indicators in THF
are generally close to the pKa values in DMSO.41 For the
cesium salts of enolates, the pK values relative to such
indicators tend to be about 1 pK unit less than the DMSO
pKa values.23,25,28,69 In the present cases of the cesium
salts of â-diketones, the pK of the cesium salt of the enol
of 1, 10.9, is 3.3 units less than that of PhCOCH2COCH3
in DMSO, 14.2.70 The latter value is consistent with that
of dibenzoylmethane in DMSO, 13.35,71 which is expected
to be a little lower because of the inductive effect of the
second benzene ring. The cesium ion pair pK of 2 is
slightly lower than that of 1 again because of the
inductive effect of the second benzene ring now one atom
farther removed. The lower ion pair pK values compared
to ionic pKa values in DMSO show that even the large
cesium cation has a significant chelating stabilization
effect within the ion pairs. This result is also consistent
with the earlier finding by DePalma and Arnett72 that
cesium â-ketoenolates have significant ion pair associa-
tion constants even in DMSO. No such chelating stabi-
lization should apply to the cesium salt of 3 because of
the W-shape of its enolate ion. No comparable compounds
(61) Bradamante, S.; Pagani, G. A. J. Org. Chem. 1984, 49, 2863.
(62) Abbotto, A.; Bradamante, S.; Pagani, G. A. J. Org. Chem. 1993,
58, 449.
(63) Gatti, C.; Ponti, A.; Gamba, A.; Pagani, G. A. J. Am. Chem.
Soc. 1992, 114, 8634.
(64) Abbotto, A.; Bradamante, S.; Pagani, G. A. J. Org. Chem. 1993,
58, 444.
(65) Kalinowki, H.-O.; Berger, S.; Brau, S. In Carbon-13 NMR
Spectroscopy; John Wiley & Sons: New York, 1988.
(66) Seebach, D.; Amstutz, R.; Laube, T.; Schweizer, W. B.; Dunitz,
J. D. J. Am. Chem. Soc. 1985, 107, 5403-9.
(67) Seebach, D. Angew. Chem. 1988, 100, 1685-1715.
(68) Williard, P. G.; Salvino, J. M. Tetrahedron Lett 1985, 26, 3931-
4.
(69) Ciula, J. C.; Streitwieser, A. J. Org. Chem. 1992, 57, 431-432;
correction p 6686.
(70) Personal communication quoted in www.cem.msu.edu/∼reusch/
OrgPage/Tables.htm.
(71) Olmstead, W. N.; Bordwell, F. G. J. Org. Chem. 1980, 45, 3299-
305.
(72) DePalma, V. M.; Arnett, E. M. J. Am. Chem. Soc. 1978, 100,
3514-3525.
J. Org. Chem, Vol. 69, No. 24, 2004 8353

Facchetti and Streitwieser
have been studied in DMSO but the ion pair values in
THF, ∼12-13, are comparable to (and even somewhat
higher than) the ionic pKa values of 1,3-cyclohexanedione,
10.3, and dimedone, 11.2, in DMSO.73 These conclusions
will probably also apply to the more common potassium
salts.
but, as discussed above, reasonable structures are ex-
emplified in 4 and 5. An alternative structure 9 is a
These same considerations also apply to the lithium
ion pairs except that chelation effects are now much
stronger. Ion pair pK values of the lithium salts of the
delocalized indicators are comparable to the correspond-
ing Cs pK values because the lithium salts, as solvent-
separated ion pairs (SSIP), are effectively only somewhat
larger than the cesium contact ion pairs (CIP) and both
are relative to the same standard.43 Lithium ion pair pK
values of enolate ions are 6-9 pK units lower than the
corresponding cesium enolates,23,29,74 because the lithium
salts are now CIP electrostatically bound more firmly to
the negative enolate oxygens. One would therefore expect
that the corresponding ∆pK values for â-ketoenolates
would be still greater, but in fact the values shown in
Table 4, about 8, are of the same magnitude. The unequal
O-M bond distances in the â-ketoenolate chelates un-
doubtedly play a role together with the fact that the two
oxygens in the â-ketoenolate ions effectively share a
single negative charge.
The second pK values of the â-diketones are, of course,
much higher than the first. For the lithium and dilithium
salts of 1 the difference (Table 3) is about 19 pK units.
The difference drops to about 15 for 2; the pK2 difference
of 4.2 between 1-Li2 and 2-Li2 shows the effect of
conjugation; the second anion in 2-Li2 is now benzylic.
Indeed, the value of 17.7 for 2-Li2 is remarkably low for
a dianion; it is not much greater than the pK for the
monomer lithium salt of p-phenylisobutyrophenone, 15.9.31
The pK2 for the corresponding dicesium salt, 2-Cs2, 24.5,
is higher by an amount typical of the difference between
lithium and cesium salts of monoketones. For example,
the pK for the monomer of the cesium salt of p-phenyl-
isobutyrophenone is 25.1.25
twisted version of 4 and is a portion of a polymeric crystal
structure of unsolvated lithium acac. This structure,
however, seems less reasonable for the dimer in solution
because of the reduced possibilities for solvation of
lithium cation. In both structures 4 and 9 the lithiums
are probably less solvated than in the chelated monomer
and loss of solvation provides some driving force for
dimerization. In 5, however, the reduced solvation of the
lithiums compared to the nonchelated monomer of 3-Li
probably provides much of the driving force for the higher
value of Kd. Coordination solvation of the lithiums in
these structures has not been included in the proposed
structures but all of the lithiums are probably three- or
four-coordinate with inclusion of ether solvents.73
For the monomeric dimetal dienediolates a reasonable
structure would be 10, which features the cyclic Li2O2
bonding common in lithium enolate dimers.75 Such a
monomer structure could readily dimerize to the cubic
stucture 6. In this dimerization the reduced basicity of
2-Li2 and the greater steric hindrance provided by the
terminal phenyl group would both lead to a lower Kd
compared to that of 1-Li2.
The aggregation of lithium enolates in THF is affected
generally by steric effects and by basicity. For example,
6-phenyl-R-tetralone, 8 (R ) H), and its 2-benzyl deriva-
tive, 8 (R ) benzyl), have comparable ion pair lithium
basicities but the less hindered enolate aggregates to a
tetramer whereas the more hindered benzyl derivative
forms a dimer.30 The lithium salt of p-phenylisobuty-
rophenone aggregates to a tetramer31 but less basic
enolates that are also derived from ketones with a
tertiary R-hydrogen often form dimers instead. These
Finally, the structural speculations presented here
have been confirmed by ab initio computations to be
published separately.76
Conclusions
â-Diketones that can lead to chelated monoalkali salts
have relatively high ion pair acidities and dimerization
constants comparable to those of simple enolates, of the
order of 103 M-1. Thus, THF solutions of such compounds
in typical synthetic reactions or physical measurements
(such as NMR) of 0.01 M or more consist mostly of dimers
but the possible role of such dimers is rarely considered.
â-Diketones whose anions are W-shaped and cannot give
intramolecular chelates are less acidic but only by about
2 pK units. Their enolates, however, have larger aggrega-
tion constants by another order of magnitude. Derived
dilithium dienediolates have relatively low ion pair pK
dimers have Kd values of the order of 103 M-1 27-29
similar
,
to those of 1-Li and 2-Li. In the case of the lithium
enolates of monoketones, part of the driving force for
dimerization is the entropy gained by loss of solvation of
the lithium. For the chelated lithium enolates of â-dike-
tones, the lithiums are less solvated to begin with and
this driving force should be lower. The lithium in a
nonchelated enolate, such as 3-Li, is then more highly
solvated and has more solvation entropy to gain on
dimerization. Accordingly, Kd for 3-Li is much higher,
26 000 M-1. The structures of the dimers are not known
(73) Arnett, E. M.; Harrelson, J. A., Jr. J. Am. Chem. Soc. 1987,
109, 809-812.
(75) Pratt, L. M.; Streitwieser, A. J. Org. Chem. 2003, 68, 2830-
(74) Wang, D. Z.; Streitwieser, A. J. Org. Chem. 2004, 68, 8936-
8942.
2838.
(76) Manuscript in preparation.
8354 J. Org. Chem., Vol. 69, No. 24, 2004

Aggregation of Metal Enolates to Dimers
cuvette, and an excess of base was added to have the neutral
compound completely either mono- or dideprotonated. The
solution was successively diluted with known amounts of THF,
and the spectrum was obtained after each dilution with the
extinction coefficient at each concentration calculated in a
straightforward manner.
values and comparable dimerization constants. Corre-
sponding cesium salts generally are somewhat less
aggregated.
Experimental Section
Acidity Measurements. Single indicator method: A
known amount of hydrocarbon indicator was transferred into
a thermostated 0.1 or 1 cm quartz cell containing a known
amount of THF. The base was added and the UV-vis spectrum
was immediately recorded. A known amount of substrate (1-
3) was added and the mixture was left to reach equilibrium.
Dilution with known amounts of THF provided spectra at
different anion concentrations. Double indicator method:
A known amount of both an appropriate indicator and a
substrate (1-3) were transferred into a thermostated 0.1 or 1
cm quartz cell containing a known amount of THF. At this
point two procedures were followed depending on whether the
equilibrium between the indicator and the substrate was
reached rapidly or slowly. In the first case a small amount of
base was added and the spectrum of the solution was im-
mediately recorded. In the second, the base was added until
the absorbance of the solution reached 1.5-2.0 absorbance
units, then the mixture was left to reach equilibrium. Spectra
at different anion concentrations were obtained by successive
addition of base in the first case or by dilution with known
amounts of THF in the second case..
General. Melting points (Pyrex capillary) were determined
on a Buchi melting point apparatus and are uncorrected. UV-
vis spectra were recorded on a spectrophotometer equipped
with fiber optics cables connected to a thermostated cell holder
built into the floor of the glovebox.
Materials. Starting materials for synthesis were obtained
from commercial suppliers and purified by crystallization or
distillation prior to use. Purity of synthesized compounds was
monitored by combination of 1H NMR, GLC, mp, and elemental
analysis.
1-(4-Biphenylyl)butane-1,3-dione, 1: Prepared following
Spraugue,39 55% yield, white solid, mp 160 °C (after sublima-
tion) (lit.39 mp 156-7 °C). Anal. Calcd for C16H14O2: C, 80.65;
H, 5.92. Found: C, 80.42; H, 5.90.
1,4-Diphenylbutane-1,3-dione, 2: Prepared following
Padwa,40 61% yield, white solid, mp 56 °C (after sublimation)
(lit.40 mp 52-3 °C). Anal. Calcd for C16H14O2: C, 80.65; H, 5.92.
Found: C, 80.51; H, 5.98.
6,7,8,9-Tetrahydro-5H-benzocyclohepten-5,7-dione, 3.
Potassium tert-butoxide (0.34 g, 3.00 mmol) was added to a
solution of ethyl â-(2-acetylphenyl)propionate (0.30 g, 1.36
mmol) in dry tert-butyl alcohol (40 mL). After refluxing for 45
min, the reaction mixture was poured onto water (70 mL) and
H2SO4 (20%) was added until the pH was brought to about 1.
The aqueous solution was extracted with ether and the organic
layer was dried (MgSO4) and evaporated to afford a brown solid
(380 mg). The solid was chromatographed on silica gel (Et2O:
AcOEt 9:1) to afford the target compound as a white solid (0.17
g), which was obtained analytically pure after sublimation
NMR Measurements. THF-d8 and DMSO-d6 were dried
by distillation from Na/K alloy and CaH2, respectively. 1H and
13C NMR spectra were recorded at 25 °C with a spectrometer
operating at 125.70 MHz in the FT mode. The spectral
parameters and calibration were set according to known
procedures.78 When possible, the samples were investigated
as 0.01 and 0.1 M solutions for 1H and 13C spectra, respectively.
1
(0.15 g, 0.86 mmol, 63%); mp 61 °C. H NMR (CDCl3) δ 7.97
Acknowledgment. This research was supported in
part by grants from the National Science Foundation.
(d, J ) 7.8 Hz, 1H), 7.51 (t, J ) 7.5 Hz, 1H), 7.38 (t, J ) 7.4
Hz, 1H), 7.27 (d, 1H), 4.13 (s, 2H), 3.37 (t, J ) 6.2 Hz, 2H),
2.76 (t, 2H). Anal. Calcd for C11H10O2: C, 75.83; H, 5.80.
Found: C, 75.67; H, 5.58.
Supporting Information Available: Figures S1-S25
showing UV-vis spectra for 1, 2, and 3, plots of absorbance for
1, 2, and 3, plots of Kobs vs {enolate}/Kobs for 1-Li, 2-Li, 3-Li,
1-Cs, and 2-Cs, plots of Kobs vs {1-Li2}/Kobs, {2-Li2}/Kobs, and
Deprotonating Agents. The solutions of cesium diphenyl-
methane and cumyl cesium used to prepare the cesium
enolates were obtained according to known procedures.35
Trimethylsilylmethyllithium was prepared following the method
described by Saljoughian.77 After sublimation, it was trans-
ferred into the glovebox and used as a solid to prepare the
lithium enolates. Lithium diisopropylamide (Aldrich) was
sublimed three times before use.
{2-Cs2}/Kobs,
1H NMR spectra for 1, 2, and 3, and 13C NMR
spectra for 3; Tables S1 and S2 with properties of regression
lines and equilibrium constants of the enol/keto equilibrium
for 1-3. This material is available free of charge via the
Internet at http://pubs.acs.org.
JO0491915
Extinction Coefficient Measurements. A solution of
known concentration of 1-3 in THF was prepared in a UV
(78) Abbotto, A.; Alanzo, V.; Bradamante, S.; Pagani, G. A. J. Chem.
Soc., Perkin Trans. 2 1991, 481.
(77) Saljoughian, M. Monatsh. Chem. 1984, 115, 519.
J. Org. Chem, Vol. 69, No. 24, 2004 8355