Generation and detection of a relatively persistent carboxylic acid enol-2,2-bis(2′,4′,6′-triisopropylphenyl)ethene-1,1-diol

Article abstract of DOI:10.1021/ja9601090

Ditipyl ketene (tipyl = 2,4,6-triisopropylphenyl) (20), ditipylacetic acid (19), and 2,2-ditipyl bis-(trimethylsilyl) acetal 28 were prepared as potential precursors for the enol of 19, i.e., 2,2-ditipylethene-l,l-diol (25). Protonation of the dianion of

Full text of DOI:10.1021/ja9601090

J. Am. Chem. Soc. 1996, 118, 5169-5181
5169
Generation and Detection of a Relatively Persistent Carboxylic
Acid Enols2,2-Bis(2′,4′,6′-triisopropylphenyl)ethene-1,1-diol
Joseph Frey and Zvi Rappoport*
Contribution from the Department of Organic Chemistry, The Hebrew UniVersity,
Jerusalem 91904, Israel
ReceiVed January 12, 1996X
Abstract: Ditipyl ketene (tipyl ) 2,4,6-triisopropylphenyl) (20), ditipylacetic acid (19), and 2,2-ditipyl bis-
(trimethylsilyl) acetal 28 were prepared as potential precursors for the enol of 19, i.e., 2,2-ditipylethene-1,1-diol
(25). Protonation of the dianion of 19 or fluoride ion desilylation of 28 gave the radical Tip2CdC(OH)O• (27), and
only hydration of 20 in 42:5:3 DMF-d7:THF-d8:H2O or in THF-d8:H2O gave solutions of g 98% of the enediol 25.
1H and 13C NMR spectra of 25 and of mixtures of 25 with its O-mono- and dideuteriated derivatives 25-D and 25-D2
and linear δ(OH)[25] vs δ(OH) of 2,2-ditipyl-1-R-ethenols, R ) Me, H (30, 31) correlations corroborated the structure
and indicated that the two OH groups are identical on the NMR time scale. A propeller arrangement for the ditipylvinyl
moiety and anti arrangements for the CdC-O-H moieties where the OH groups are solvated by the dipolar aprotic
solvent (e.g., DMF) are suggested for the conformation of 25. Although solid 25 could not be isolated, it is the
longest-lived 1,1-enediol prepared so far in solution. The hydration rates of di(bulky)aryl ketenes decrease with the
increased bulk of the aryl groups.
Introduction
tions by Guthrie4 give pKenol () -log Kenol) of 21.24a and 19.44b
for the CH3COOH/H2CdC(OH)2 pair. MO calculations give
20.5,2 26,6 18.8,7 and 20.48 and deuterium exchange data give
19.3.4a,9 An earlier comparison with the acetaldehyde/vinyl
alcohol pair (where pKenol ) 6.23)10 gave ∆E(H2CdC(OH)2-
Enolization of aldehydes and ketones has been extensively
studied almost from the beginning of the century, and great
progress in the field was achieved in the last 15 years.2 Even
simple enols, i.e., those unsubstituted by strongly electron-
withdrawing substituents were observed, and the very small
equilibrium constants with their carbonyl tautomers (Kenol) and
their acidities were measured. In contrast, enols of carboxylic
acids (1,1-enediols) and their derivatives, such as esters, amides,
or acyl halides (1) were much less extensively studied3 and are
not even mentioned in textbooks. This is due to their very low
stability which is ascribed to participation of the lone pair on
the heteroatom X in stabilization of the acid (or derivative) by
electron donation (cf. 2a, eq 1).2 The anomeric interaction
(OH,X) in the 1,1-enediol derivatives is much lower and may
even be destabilizing.4 Consequently, in contrast to the large
number of enols which exist as stable species,2 no long-lived,
or thermodynamically stable enols of carboxylic acid derivatives
are known except for the amide enol (NC)2CdC(OH)NH2.5
CH3COOH) - ∆E (H2CdCHOH-CH3CHO) ) 10 kcal mol-1 2
.
Lower Kenol values were calculated for 1-cyclopropanecarboxylic
acid (pKenol ) 28.3) and 3-cyclopropenecarboxylic acid (pKenol
) 30.2).7
In spite of their very low thermodynamic stability, 1,1-
enediols and their anions were suggested as “unobserved”
intermediates in the decarboxylation of 2-phenyl-1,1-cyclohex-
ane dicarboxylic acid11a and in the reductive debromination of
1-bromo-2-phenylcyclohexanecarboxylic acid.11b 1,1-Enediol
derivatives were suggested as intermediates in the bromination
of acyl chlorides.12 Ester and amide “enols” (1, X ) OR and
NRR′) were suggested as intermediates in the addition of
alcohols to ketenes,13 the acid-catalyzed hydration of keten-
imines,14 and the electrophilic substitution of malonamide.15
Other suggested intermediates are an anhydride enol (1, X )
OCOR) in the electrophilic addition to Ac2O16 and acyl halide
(6) Heinrich, N.; Koch, W.; Frenking, G.; Schwarz, H. J. Am. Chem.
Soc. 1986, 108, 593.
(7) Andraos, J.; Kresge, A. J.; Peterson, M. R.; Csizmadia, I. G. J. Mol.
Struct. (Theochem) 1991, 232, 155.
(8) Duan, X.; Page, M. J. Am. Chem. Soc. 1995, 117, 5114.
(9) Bok, L. D. K.; Geib, K. H. Z. Phys. Chem. Abt. A 1939, 183, 353.
(10) Keeffe, J. R.; Kresge, A. J.; Schepp, N. P. J. Am. Chem. Soc. 1988,
110, 1993.
(11) (a) Zimmerman, H. E.; Cutshall, T. W. J. Am. Chem. Soc. 1958,
80, 2893. (b) Zimmerman, H. E.; Cutshall, T. W. J. Am. Chem. Soc. 1959,
81, 4305.
(12) (a) Harwood, H. J. Chem. ReV. 1962, 62, 99. (b)March, J. AdVanced
Organic Chemistry, 4th ed.; Wiley: New York, 1992; p 590.
(13) Ried, W.; Junker, P. Justus Liebigs Ann. Chem. 1967, 709, 85.
(14) (a) Nguyen, M. T.; Hegarty, A. F. J. Am. Chem. Soc. 1983, 105,
381. (b) Hegarty, A. F.; Kelly, J. G. 5th European Symposium on Organic
Reactivity (ESOR V), Santiago de Compostella, Spain, July 16-21, 1995;
Abstracts, C15A.
The reduced stabilities are also reflected by calculations which
show that the equilibrium constants Kenol for the carboxylic acid
(2, X ) OH)/1,1-enediol (1, X ) OH) are much lower than
those for the corresponding carbonyl/enol pairs. Recent estima-
X Abstract published in AdVance ACS Abstracts, May 1, 1996.
(1) Presented in part in the 12th Conference on Physical Organic
Chemistry, Padova, Italy, August 28-September 2, 1994, Abs. IC 2, p 35
and in the 2nd Symposium on Fundamental Organic Chemistry of Chemical
Society of Japan, Kyushu University, October 15-17, 1994, Abs. 6, p 15.
(2) The Chemistry of Enols; Rappoport, Z., Ed.; Wiley: Chichester, 1990.
(3) Hegarty, A. F.; O’Neill, P. in ref 2, Chapter 10, p 639. Nguyen, M.
T.; Hegarty, A. F. J. Am. Chem. Soc. 1984, 106, 1552.
(4) (a) Guthrie, J. P. Can. J. Chem. 1993, 71, 2123. (b) Guthrie, J. P.;
Liu, Z. Can. J. Chem. 1995, 73, 1395.
(5) Trofimenko. S.; Little, E. L., Jr.; Mower, H. F. J. Org. Chem. 1962,
27, 433.
(15) Williams, D. L. H.; Xia, L. J. Chem. Soc., Perkin Trans. 2 1993,
1429.
(16) Peno, N. C.; Billups, W.; Distepano, R.; Schnewer, S. J. Org. Chem.
1970, 35, 278.
S0002-7863(96)00109-6 CCC: $12.00 © 1996 American Chemical Society

5170 J. Am. Chem. Soc., Vol. 118, No. 22, 1996
Frey and Rappoport
enols (1, X ) halogen) in the addition of H-Hal to ketenes.17
None of these intermediates was isolated or even observed,
except for a recent brief report of the observation of amide
enols.14b
carbanion whose charges are resonatively delocalized on
2,17d,24
Câ
(eq 3). Consequently, electron withdrawing Câ sub-
stituents will stabilize the transition state of the hydration and
accelerate the reaction.
Several enediols were observed since 1987. Following an
early work of Kobayashi et al.18 the groups of Wirz19,20 and
Kresge20,21 generated the enediol of cyclopentadiene-1-carboxy-
lic acid (3, pKenol ) 6.7)19a and indene-1-carboxylic acid (4,
pKenol ) 7.26),21a PhC(OH)dC(OH)2 (5, pKenol ) 15.4),20a and
PhC(CN)dC(OH)2 (6, pKenol ) 7.22)20b which are electronically
stabilized at Câ. These were generated mostly by flash
photolysis of diazoketones which by a Wolff rearrangement22
led to short-lived ketenes which were then hydrated. Although
ketene hydration to carboxylic acids was extensively investigated
mechanistically, the intermediates were never observed until
Urwyler and Wirz19a generated 6-fulvenone 7 in water and
identified its hydration product as the enediol 3 which then
tautomerized rapidly to the carboxylic acid (eq 2). By using
fast kinetics techniques, pH-rate profiles, spectrophotometric
titrations, and thermodynamic cycles involving the enolate, the
carboxylate, and the enediolate anions, pKa’s of the enediols
and Kenol values were calculated.19-21 The pKenol values
demonstrate that substituents stabilize the 1,1-enediols by up
to >14 orders of magnitude compared with the value of 21
estimated for CH2dC(OH)2.4a Those for 3 and 6 are in the
range for simple stable enols such as acetaldehyde (6.23)10 or
acetone (8.33)10 in water at 25 °C. Kresge stated that “These
substances have also been called ‘ketene hydrates’. Their
chemistry, however, is quite different from other carbonyl, e.g.,
aldehyde and ketone, hydrates and more like that of the enol
isomers of aldehydes and ketones. We consequently prefer the
term ‘carboxylic acid enols’”.21a For evaluating this statement,
direct structural evidence on the structure of these species, which
is at present lacking, is necessary.
The thermodynamic stabilization is illustrated by the zwit-
terionic hybrid 3a of 3 (eq 2) in which the electron rich dioxy
function is positively charged and the five-membered ring
is negatively charged, i.e., aromatic. The enediols are acidic
(e.g., pKa(6) ) 0.99)20b and enolates which should ketonize
much more readily than the less nucleophilic enols25 were
suggested as intermediates in the ketonization of 1,1-enediols.21a
In all these cases the hydration and ketonization are very
fast.19-21
Another approach to observable 1,1-enediols is based on the
success in isolating both kinetically and thermodynamically
stable simple enols stabilized by bulky aromatic substituents
by Fuson26 and by us.27 O’Neill and Hegarty28 reported the
formation of the 2,2-diaryl-1,1-enediol 8 and the ester enol 9
when Ar ) pentamethylphenyl and of the enediol 10, Ar )
mesityl. Enediol 8 was generated by Bu4NF cleavage of its
bis(trimethylsilyl) acetal and characterized by UV and IR spectra
1
and the lack of H NMR CH signal of the isomeric acid. 9
was prepared similarly from the trimethylsilyl acetal of tert-
butyl bis(pentamethylphenyl)acetate. 8 and 9 decayed to the
acid and to the ester, respectively, with τ1/2 of several hours in
neutral solution and their air oxidation gave stable radicals.
Hegarty et al.29 hydrated bis(pentamethylphenyl)ketene 11
in MeCN-H2O mixtures and observed a long-lived intermediate
identified as enediol 8. Dimesitylketene 13 hydrates 3-4 times
faster than 11, but three orders of magnitude slower than
diphenylketene 14, demonstrating that bulky substituents in-
crease the kinetic stability of ketenes. The kinetic data are for
the decay rates of the ketenes, thus avoiding the question of
the stabilities of the 1,1-enediols.
The substituent effects are both kinetic and thermodynamic.
In order to observe the 1,1-enediols, their generation by ketenes
hydration ought to be much faster than their destruction by
tautomerization. This can be achieved either by accelerating
the hydration reaction or by decreasing the ketonization rate.
For simple enols conjugation of the enolic moiety to aromatic
rings reduces the ketonization rate.23 Ketene hydration proceeds
through an in-plane nucleophilic attack of water on the carbonyl
carbon, initially generating a zwitterion and afterwards a
The bulkier are the two â-aryl substituents, the more stable
is the enol.27 Hence, the best candidate for a stable 1,1-enediol
is 2,2-bis(2′,4′,6′-tri-tert-butylphenyl)ethene-1,1-diol, but we
(23) (a) Chiang, Y.; Kresge, A. J.; Krogh, E. T. J. Am. Chem. Soc. 1988,
110, 2600. (b) Keeffe, J. R.; Kresge, A. J.; Yin, Y. J. Am. Chem. Soc.
1988, 110, 8201.
(24) (a) Reference 17d, Chapter 5, pp 571-587. (b) Blake, P. In The
Chemistry of Ketenes, Allenes and Related Compounds; Patai, S., Ed.;
Wiley: Chichester, 1980; Chapter 9, p 311. (c) Poon, N. L.; Satchell, D. P.
N. J. Chem. Soc., Perkin Trans. 2, 1983, 1381.
(25) (a) Pruszynski, P.; Chiang, Y.; Kresge, A. J.; Schepp, N. P.; Walsh,
P. A. J. Phys. Chem. 1986, 90, 3760. (b) Chiang, Y.; Kresge, A. J.;
Santaballa, J. A.; Wirz, J. J. Am. Chem. Soc. 1988, 110, 5506.
(26) (a) Fuson, R. C.; Corse, J.; McKeever, C. H. J. Am. Chem. Soc.
1940, 62, 3250. (b)Fuson, R. C.; Armstrong, L. J.; Chadwick, D. H.;
Kneisley, J. W.; Rowland, S. P.; Shenk, W. J., Jr.; Soper, Q. F. J. Am.
Chem. Soc. 1945, 67, 386 and other papers in the series.
(27) (a) Hart, H.; Rappoport, Z.; Biali, S. E. in ref 2, Chapter 8, pp 481-
589. (b) Rappoport, Z.; Biali, S. E. Acc. Chem. Res. 1988, 21, 442.
(28) Hegarty, A. F.; O’Neill, P. J. Chem. Soc., Chem. Commun. 1987,
744.
(17) (a) Lillford, P. J.; Satchell, D. P. N. J. Chem. Soc., B. 1968, 897.
(b) Poon, N. L.; Satchell, D. P. N. J. Chem Res. Synop. 1983, 182. (c)
However, see: Seikaly, H. R.; Tidwell, T. T. Tetrahedron 1986, 42, 2587.
(d) Tidwell, T. T. Ketenes, Wiley, New York, 1995.
(18) Shibata, T.; Koseki, K.; Yamaoka, T.; Yoshizawa, M.; Uchiki, H.;
Kobayashi, T. J. Phys. Chem. 1988, 92, 6269.
(19) (a)Urwyler, B.; Wirz, J. Angew. Chem., Int. Ed. Engl. 1990, 29,
790. (b) Almstead, J. K.; Urwyler, B.; Wirz, J. J. Am. Chem. Soc. 1994,
116, 954.
(20) (a) Chiang, Y.; Kresge, A. J.; Pruszynski, P.; Schepp, N. P.; Wirz,
J. Angew. Chem., Int. Ed. Engl. 1990, 29, 792. (b) Andraos, J.; Chiang, Y.;
Kresge, A. J.; Pojarlieff, I. G.; Schepp, N. P.; Wirz, J. J. Am. Chem. Soc.
1994, 116, 73.
(21) (a) Andraos, J.; Chiang, Y.; Huang, C.-G.; Kresge, A. J.; Scaiano,
J. C. J. Am. Chem. Soc. 1993, 115, 10605. (b) Andraos, J.; Kresge, A. J.;
Popik, V. V. J. Am. Chem. Soc. 1994, 116, 961.
(22) (a) Wolff, L. Justus Liebigs Ann. Chem. 1944, 556, 65. (b) Meier,
H.; Zeller, K. P. Angew. Chem., Int. Ed. Engl. 1975, 14, 32.
(29) Allen, B. M., Hegarty, A. F.; O’Neill, P.; Nguyen, M. T. J. Chem.
Soc., Perkin Trans. 2 1992, 927.

Generation of a Carboxylic Acid Enol
J. Am. Chem. Soc., Vol. 118, No. 22, 1996 5171
Scheme 1
failed in our attempt to prepare a precursor for it. Smaller aryl
groups are therefore required. 2-Aryl-1-acenaphthyl enol is
more stable when Ar ) 2,4,6-triisopropylphenyl(tipyl) than
when Ar ) mesityl (Mes),30 and tipyl confers higher kinetic
stability than mesityl to 1,2-diaryl-1,2-ethenediols.31 The keto/
enol equilibrium constants Kenol are higher for tipyl- than for
mesityl-substituted enols,32,33 and the effect of the buttressing
in a pentamethylphenyl group compared with Mes on Kenol value
is small.34 Consequently, we assumed that two â-tipyl groups
would confer a higher stability to the 1,1-enediol moiety than
that hitherto achieved for the Mes and C6Me5 analogs. This
prediction was corroborated and the observation and study of
the corresponding 1,1-enediol is the subject of the present paper.
Figure 1. ORTEP drawing of 21.
Results and Discussion
Synthesis of Ditipylacetic Acid and Ditipylketene. The
relatively simple synthetic routes to the less crowded dimesityl
and bis(pentamethylphenyl) acetic acids29 failed to give ditipy-
lacetic acid. We therefore attempted the more tedious methods
suggested by Akkerman35 for the preparation of polyalkyl-
substituted diphenylacetic acids starting from the diarylaceto-
nitrile or from diarylmethyl methyl ether. In both routes
ditipylmethanol 15 is an intermediate. The alcohol was prepared
according to Scheme 1. Tipyllithium was prepared from tipyl
bromide and butyllithium in ether and when reacted with excess
of ethyl formate at 0 °C ditipylmethyl formate 16 was obtained
in a relatively low (27%) yield. Likewise, several other hindered
benzhydryl formates were isolated from Grignard reagents and
aldehydes.36 Reduction of 16 with LiAlH4 gave 15 in 92%
yield. This procedure is advantageous over the reaction of tipyl
bromide with BuLi and 0.5 equivalent of ethyl formate which
gave a low yield of 15 admixed with triisopropylbenzene,
triisopropylbenzenecarboxaldehyde, and 16. When gaseous HCl
Figure 2. ORTEP drawing of the ipso-attack product 24.
carbocation) and the formation of a methyl ether, unequivocally
identified the compound as ditipylmethyl chloride 17.
When solution of crude 17 in benzene was reacted with
cuprous cyanide in DMF (eq 4) by the method of Lock and
Riegler,37b no organic nitrile was isolated. Instead, two isomeric
was bubbled through the solution of 15 in dry benzene,37a
a
solid, which was not crystallized due to its high solubility in
all the solvents tried, was obtained. Its 1H and 13C NMR spectra,
the immediate precipitation of AgCl on addition of ethanolic
AgNO3 solution, the development of deep purple color on
absorption on SiO2 (suggesting the formation of the diarylmethyl
aryl tipyl methanes 21 and 22 in which the aryl group was 2,6-
diisopropyl-4-isopropenylphenyl and 2-isopropenyl-4,6-diiso-
propylphenyl, respectively, were isolated in a 1:3 ratio and
characterized by 1H,13C NMR, mass spectra, and microanalysis.
The structure of 21 was determined by X-ray diffraction, and
its ORTEP drawing is shown in Figure 1 (C(17)-C(18) ) 1.30
Å; dihedral angle between the two aryl groups 85.8°).38
(30) Miller, A. R. J. Org. Chem. 1976, 41, 3599.
(31) Fuson, R. C.; Horning, E. C. J. Am. Chem. Soc. 1940, 62, 2962.
Fuson, R. C.; Corse, J. J. Am. Chem. Soc. 1939, 61, 975. Fuson, R. C.;
McKeever, C. H.; Corse, J. J. Am. Chem. Soc. 1940, 62, 600. Fuson, R. C.;
Scott, S. L.; Horning, E. C.; McKeever, C. H. J. Am. Chem. Soc. 1940, 62,
11091.
(32) Frey, J.; Rappoport, Z. Unpublished results. Frey, J. Ph.D. Thesis,
The Hebrew University, Jersalem, 1996.
(33) Faraggi, E.; Frey, J.; Rappoport, Z. 11th IUPAC Conference on
Physical Organic Chemistry, Ithaca College, Ithaca, NY, August 2-7, 1992;
Abstr. C-13.
(34) Eventova, I.; Nadler, E. B.; Rochlin, E.; Frey, J.; Rappoport, Z. J.
Am. Chem. Soc. 1993, 115, 129.
In the reaction of 17 with a 18-crown-6/KCN complex in
benzene >90% of 17 was hydrolyzed to 16. The expected
nitrile 23 was formed in only a 0.8% yield together with 2.5%
of an cyclohexadienyl isomer 24 where the cyano group is
attached to the ring. Both compounds were fully characterized,
including an X-ray diffraction of 24 whose ORTEP drawing is
given in Figure 2 (The CdC bonds are C(1)-C(2) ) 1.329 Å;
C(4)-C(5) ) 1.322 Å; C(6)-C(7) ) 1.336 Å).38 We ascribe
the formation of 24 to a reaction of the Tip2CH+ carbocation,
formed from 17, with the cyanide ion at a ring position.
Precedents for ipso capture of relatively stable substituted benzyl
(35) Akkerman, O. S. Recl. TraV. Chim. Pays Bas 1967, 86, 1018.
(36) Rieker, A.; Butsugan, Y.; Shimizu, M. Tetrahedron Lett. 1971, 1905.
(37) (a) Nauta, W. Th.; Wuis, P. J. Recl. TraV. Chim. Pays Bas 1937,
56, 535. (b) Lock, G.; Rieger, V. Chem. Ber. 1953, 86, 74.

5172 J. Am. Chem. Soc., Vol. 118, No. 22, 1996
Frey and Rappoport
Figure 4. Stereoview of ditipyl ketene 20.
did not broaden on cooling (in acetone) down to 180 K. Hence,
the aryl groups in 15-20 freely rotate around the Ar-C bonds
at room temperature.
Synthesis of 1,1-Enediols. Most earlier studies on “observ-
able” carboxylic acid enols were kinetic measurements using
UV spectroscopy.19-21,28,29 The enediol was formed in very
low concentration (ca. 10-5 M), its isolation was impractical,
and its purity impossible to asses. Capon42 emphasized the
difference between “generation” and “synthesis” of unstable
enols which also applies for enediols. When generating unstable
enediols for spectroscopic and kinetic studies, Capon’s question
could be rephrased as “has the enediol been generated in the
absence of its acid-tautomer?”. In contrast, conventional
synthesis is interested in the purity and the yield of the
enediol.
The methods used so far to generate 1,1-enediols were
inspired by the two basic successful approaches for the
generation of stable simple enols: (a) using reactions of reactive
precursors containing the enediolic CdC(OR)2 skeleton
(e.g., O-protonation of endiolates or solvolysis of O-protected
enediols) and (b) a rapid generation from non-enediolic precur-
sors. In order to enable the isolation of 1,1-enediols, their
formation ought to be much faster than their destruction. As
discussed above, 2,2-ditipylethene-1,1-diol 25, i.e., the enol of
ditipylacetic acid 19, seems a proper candidate. Moreover, if
formed, it should be easily detected and monitored since a
pair of o-i-Pr methyl signals appears at a high field in the
1H NMR spectrum of gem-ditipylvinyl systems. Scheme 2
summarizes the three different synthetic methodologies at-
tempted.
Figure 3. ORTEP Drawing of ditipyl ketene 20.
cations are known.39 Due to the low yield of the nitrile an
alternative method based on the known benzylic C-O cleava-
geof benzyl and benzhydryl ethers by alkali metals35,40 was
applied.
Dry methanol and 17 in benzene gave the methyl ether 18 in
95% yield. Cleavage of 18 with Na/K alloy in ether with
sonication under argon, followed by passing dry carbon dioxide
into the red solution of the anion gave ditipylacetic acid 19 in
73% yield. The sonication, known to accelerate the reduction
reaction40c reduced the reaction time and increased the yield of
the acid over those of analogous reactions.35 When thionyl
chloride, trace of pyridine, and 19 were refluxed in dry benzene,
ditipylketene (20) was obtained in 89% yield. It was stable
enough for purification by chromatography on a silica column,
and it was identified by NMR, FTIR, mass spectra, microanaly-
sis, and X-ray diffraction. The ORTEP drawing of 20 is given
in Figure 3 and its stereoview in Figure 4. Selected bond lengths
and angles are given, together with those of dimesitylketene
1341 in Table 1.
(a) Careful Protonation of the Dimetal Enediolates
Tip2CdC(O-M+)2 26. Carboxylic acids yield metal ene-
diolates43 (Ivanov reagents44 ) with organometallic reagents.
Several structures of enediolates of phenylacetic acids have been
Scheme 2
In contrast to 1,1-ditipylvinyl systems which adopt a frozen
propeller conformation e room temperature,32,33 the derivatives
1
15-19 show two 12H o-i-Pr signals in the H NMR spectra
and the more symmetrical ketene 20 showed a 24H signal, which
(38) For crystallographic data see the supporting information section.
(39) (a) E.g., Richard, J. P. J. Am. Chem. Soc., 1989, 111, 6735.
Tetrahedron Lett. 1989, 30, 23. (b) Tidwell, T. T. J. Am. Chem. Soc. 1986,
108, 3470. (c) For reaction of photogenerated triarylvinyl cations with
cyanide ion, see: Kitamura, T.; Murakami, M.; Kobayashi, S.; Taniguchi,
H. Tetrahedron Lett. 1986, 27, 3885.
(40) (a) Maercker, A. Angew. Chem., Int. Ed. Engl. 1987, 26, 972. (b)
Fish, R. H.; Dupon, J. W. J. Org. Chem. 1988. 53, 5230. (c) Karaman, R.;
Kohlman, D. T.; Fry, J. L. Tetrahedron Lett. 1990, 31, 6155. (d) Azzena.
U.; Fenude, E.; Fina, C.; Melloni, G.; Pisano, L.; Sechi, B. J. Chem. Res.
(S) 1994, 108.
(41) Biali, S. E.; Gozin, M.; Rappoport, Z. J. Phys. Org. Chem. 1989,
2, 271.
(42) Capon, B. In The Chemistry of Enols; Rappoport, Z., Ed.; Wiley:
Chichester, 1990; Chapter 5, p 307.
(43) (a) Rasmussen, J. K. Synthesis 1977, 91. (b) Petragnani, N.;
Yonashiro, Synthesis 1982, 521.
(44) Ivanov, D.; Vassilev, G.; Panayotov, I. Synthesis 1975, 83 and
references cited therein.

Generation of a Carboxylic Acid Enol
J. Am. Chem. Soc., Vol. 118, No. 22, 1996 5173
standing for 1 h at room temperature in an argon atmosphere
and only after adding water the solution turns red, we suggest
that the monoenolate 25- rather than the enediolate 26 is the
species oxidized by air to the R-carboxy radical 27. Moreover,
1,1-enediols ionize substantially even in neutral solutions and
since a base is formed on quenching, we suggest that on
quenching 26-Li by water, the ditipylethene-1,1-diol monoanion
25- is formed and then oxidized by air to 27 (eq 5).
The ESR spectrum of 27 showed a broad signal with g )
2.0047, consistent49 with the carboxyl conjugated radical
structure proposed. We ascribe the relatively high g-value,
which resembles those reported for R-acyl radicals,28,49 to a spin
density delocalization on the oxygen50 as depicted in 27a and
27b (eq 6). 27 shows an IR νCdO at 1675 cm-1 (25 cm-1 lower
than in 19) indicating an increased single bond character of the
carbonyl moiety (cf. 27a and 27b).
Figure 5. 1H NMR spectra in CDCl3 of (A) the quenching product of
the enediolate 26-Li and (B) the acid 19.
determined by NMR,45 and they showed normal enolate
reactions with electrophiles.46
O-Protonation of enolates is widely used for generating enols
of â-ketoesters and â-diketones which ketonize relatively
slowly47 but less in the case of simple enols. Sufficiently rapid
mixing of the enolate solution with the water is required so
that the O-protonation will be completed before further keton-
ization. We prepared a solution of dilithium 2,2-ditipylethene-
1,1-diolate (26-Li) in THF by injecting 2 equiv of n-BuLi in
hexane into a THF solution of ditipylacetic acid. The bright
yellow solution was stirred under argon at room temperature
and when quenched with water or D2O after 10 min it turned
orange-red. The solvents were evaporated, CDCl3 was added,
and an 1H NMR spectrum of the deep red solution was
immediately taken. The spectrum which is compared with that
of acid 19 in Figure 5 displayed several persistent broad signals
even upon heating or cooling, thus excluding the occurrence of
a dynamic process on the NMR time scale as the reason for the
broad signals. When a catalytic amount of CF3COOH was
added to the red sample or when the enediolate solution was
quenched with an aqueous NH4Cl solution, the spectrum of acid
19 was immediately observed.
When solutions of 26-Na or 26-K generated by heating 19
with dimsyl sodium or with t-BuOK in DMSO-d6 were
quenched with water or D2O, a similar spectrum with broad
signals was obtained. Addition of iodine or bromine solutions
in CCl4 to ditipylethene-1,1-diol 25 resulted in a similar broad
1H NMR spectrum, suggesting the formation of some oxidized
form of 25. The broadening of only several of the signals is
reminiscent of spectra of stable radicals or of solutions display-
ing a rapid equilibrium of radicals with nonradical species.48
Hegarty49 has noted that the bulky R-acyldiaryl radicals formed
by oxidation of his enediols are indefinitely stable as solids and
do not dimerize or react with oxygen.
The 13C NMR spectrum of 27 is also affected by the
delocalization of the radical. The CdO and R-CH resonances
which appear at 180.9 and 48.3 ppm for 19 have disappeared,
presumably by broadening whereas the rest of the spectrum of
27 is slightly shifted upfield compared with that of 19. The
ipso- and ortho-ring carbons and the ortho-isopropyl carbons
are also broadened as expected from their proximity to the
radical center in line with the peak broadening observed in the
1H NMR spectrum.
Examination of Figure 5 raises questions. If the radical has
structure 27 which is the proton appearing at 5.7 ppm? Why
are some signals sharp if a radical should broaden the whole
NMR spectrum due to its paramagnetism? We suggest that this
is due to an association phenomenon. When an “NMR
invisible” radical associates with a “NMR detectable” nonradical
species, broadening of hydrogen signals in a close proximity
to the paramagnetic center should be observed. If radical 27
and acid 19 associate (eq 7), in analogy with the association of
carboxylic acids only the R- and the m-H and the o-i-Pr of the
“dimer” should broaden while the p-i-Pr will remain unchanged.
To test this hypothesis, we added dimesitylacetic acid to a
sample of 27 assuming that acid-radical association would
The solid obtained on quenching 26-Li by water was
chromatographed on a silica column, and the red solid isloated
displayed an identical NMR spectrum to that shown in Figure
5. Since the yellow color of a solution of 26-Li persists after
1
significantly broaden the usually sharp H NMR signals of
dimesitylacetic acid. Such a broadening was indeed observed.
Moreover, when the sample was kept for three days at room
temperature its color had almost faded and only a slight
broadening was observed in the 1H NMR spectrum. This
suggest that broadening is due to the presence of the radical 27
which is slowly lost by hydrogen abstraction from the solvent.
If radical 27 formed from enediol 25 abstract an hydrogen from
the solvent to yield ditipylacetic acid 19, this adds an additional
(45) Mladenova, M.; Blagoev, B.; Gaudemar, M.; Dardoize, F. Tetra-
hedron 1981, 37, 2153.
(46) (a) Mladenova, M.; Blagoev, B.; Gaudemar, M.; Gaudemar-Bardone,
F.; Lallemand, J. Y. Tetrahedron 1981, 37, 2157. (b) Toullec, J.; Mladenova,
M.; Gaudemar-Bardone, F.; Blagoev, B. J. Org. Chem. 1985, 50, 2563.
(47) (a) Knorr, L. Ann. Chem. 1899, 306, 363. (b) Meyer, K. H. Chem.
Ber. 1912, 45, 2864.
(48) Gu¨nther, H. NMR Spectroscopy; Wiley: Chichester, 1980; p 327
and references cited therein.
(49) (a) Hegarty, A. F.; O’Neill, P. Tetrahedron Lett. 1987, 28, 901. (b)
O’Neill, P.; Hegarty, A. F. J. Org. Chem. 1992, 57, 4421.
(50) Dobbs, A. J. Electron Spin Resonance Spectroscopy, Specialist
Periodical Reports, Vol. 2, The Chemical Society: London, 1974; Chapter
10, p 281.

5174 J. Am. Chem. Soc., Vol. 118, No. 22, 1996
Frey and Rappoport
route from 25 to 19 which does not involve keto-enol
tautomerization and should be considered in a complete kinetic
analysis of the 25 h 19 process.
even at 243 K. More important, 20 is soluble in mixtures of
THF-d8 (at least 10% v/v) with DMF-d7, DMSO-d6, Me2CO-
d6, or MeCN-d3 containing a few percent of water. However,
dissolving samples of 20 in either of these mixtures by heating
induced a rapid formation of 19, even before all the ketene had
reacted. Hence, the hydration was conducted at lower temper-
atures in order to retard the highly temperature-dependent
ketonization.
Both Kresge and Wirz reported problematic results in basic
solutions: (i) Analysis of the highly colored product mixtures
indicated that more than one product was formed in the
photoprocesses at pH > 8 or 9.19b (ii) In flash photolytic
processes conducted at high pH several transients are detected
as deduced from the different observed rates of decay.18a
Radical formation and destruction at these low concentrations
should be considered in these cases.
(b) Fluoride Ion Catalyzed Cleavage of the Bis(trimeth-
ylsilyl) Ketene Acetal 28. Tetrabutylammonium fluoride
(TBAF, “naked fluoride”),51 is a well-known silylation/desily-
lation reagent.52 We prepared ditipylketene bis(trimethylsilyl)
acetal 28 and attempted its in situ fluoride catalyzed cleavage
by using TBAF both in THF solutions or as a suspension while
adsorbed on SiO2, in an NMR tube in THF-d8. The NMR
spectrum of the red solution formed immediately after addition
of TBAF, resembled that shown in Figure 5. This is not
surprising since the F-/THF solutions are very basic53 and the
enediol is more prone to oxidize in basic media. We note that
this method was used by Hegarty in his first communication
on generation of 2,2-diarylethene-1,1-diols in solution28 but not
in his later studies.
(c) Uncatalyzed Hydration of Ditipylketene 20. We finally
succeeded in generating 2,2-ditipylethene-1,1-diol 25 by hydra-
tion of ditipylketene 20 under neutral conditions.54 The ketene
is insoluble in water in the 10-2 M concentration range, and an
organic-water mixture had to be used. When a drop of water
was added to a solution of 20 in CDCl3, no reaction was
observed even after 3 days, presumably due to the low solubility
of water in CDCl3. Hence, for a successful in situ hydration
the organic cosolvent should fulfill five prerequisites: (a)
dissolve the ketene, (b) be water miscible, (c) be unreactive
toward the ketene, (d) be commercially available in its poly-
deuteriated form, for enabling an NMR study, and (e) be a good
hydrogen-bond acceptor since the analogous 2,2-diarylethenols
are stabilized by hydrogen bonding to hydrogen bond accepting
solvents.55 The first prerequisite is very restrictive since the
solubility of 20 in the good aprotic hydrogen bond accepting
solvents DMSO-d6, DMF-d7, Me2CO-d6, or MeCN-d3 is so poor
that we were unable to prepare samples of 10 mg of 20 in 0.5
mL of these solvents at room temperature. Heating or sonication
achieved total dissolution, but on further addition of 30 µL of
water the ketene was immediately precipitated. If the sample
was then heated, the solution turned deep orange and its NMR
spectrum displayed signals of a mixture of 19, 20, and a new
compound having a doublet at 0.1-0.3 ppm which is 2,2-
ditipylethene-1,1-diol 25 (Vide infra). The products ratio was
dependent on the duration of the heating. However, after 5 min
some ketene had precipitated again. We finally found that the
water miscible THF-d8 dissolves 20 at room temperature and
Optimization of the reaction conditions led to the following
procedure: to a solution of 10 mg of 20 in 0.05 mL of THF-d8,
a mixture of 0.42 mL of DMF-d7, and 0.03 mL of water ([20]
) 0.045 M; 75-fold molar excess of water) was added at 273
K, and the reaction progress was monitored by 1H NMR. After
90 min the solution consisted of ca. 1.5% of 20 and >98% of
a new compound having an 1H NMR spectrum almost identical
with that of the bis(trimethylsilyl)ketene acetal 28 except that
the SiMe3 signals are replaced by a signal at δ 9.65 ppm, a
position reminiscent of the δ(OH) signals of Ar2CdCHOH in
hydrogen bond accepting solvents55 (Figure 6). The new
compound was therefore identified as the enediol 25. Its NMR
spectrum remained unchanged after 1 week at -18 °C, while
after standing overnight at room temperature or rapidly upon
addition of a catalytic amount of TFA, the spectrum of pure
acid 19 was obtained. Slower formation of 25 was observed
by its 1H NMR spectrum, in a binary 47:3 THF-d8:water mixture
at 273 K and in ternary solvent mixtures of THF-d8:water with
DMSO-d6 (at 298 K), Me2CO-d6, or MeCN-d3 (at 273 K) at a
solvent:water:THF v/v ratio of 42:5:3. Yet, only in DMF-d7
and in THF-d8 at 273 K, formation of 25 was almost complete
in <6 h without any apparent tautomerization; in Me2CO-d6
and MeCN-d3 the hydration at 273 K was very slow, e.g., in
acetone < 10% of 25 were formed after 6 h before tautomer-
ization to 19 had started. Tautomerization became important
at 298 K, and the ternary mixture 20:25:19 was observed in all
solvents. The reaction mixtures were orange-red, but no
additional spectral indication for formation of radical 27 was
found.
Spectroscopic Characterization of 2,2-Ditipylethene-1,1-
diol (25). Since 25 was not isolated its identification is based
on its way of preparation, its tautomerization to 19, and
unequivocally on its spectroscopic characterization.
(a)1H NMR. The 1H NMR spectrum of 25 in 42:5:3 DMF-
d7:THF-d8:H2O at 273 K (compared in Figure 6 with that of its
precursor 20) displays features characteristic of other sym-
metrical gem-ditipylvinyl systems Tip2CdCR2 (R ) H, OSi-
Me3). The presence of a longitudinal C2 symmetry axis reduces
the number of magnetically unequivalent sites by half. The
diastereotopic o-i-Pr methyls appear as four separate doublets,
2 Me each (J in Hz) at 0.14 (6.5), 0.93 (6.8), 1.26 (6.5), and
1.34 (6.8) ppm and a doublet with double integration at 1.19
ppm (J ) 6.8 Hz) is assigned to two accidentally equivalent
p-i-Pr methyls. The i-Pr methines appear as two heptets in a
2:1 ratio at 2.83 (J ) 6.8 Hz) and 3.43 (J ) 6.6 Hz) ppm, but
it is unclear whether the larger peak represents two accidentally
isochronous o-i-Pr methines or an overlap of one o- and one
p-methine heptets. Two different 2H aromatic protons appear
at 6.79 and 7.07 ppm and another 2H signal at 9.65 ppm is
assigned to the OH group. From the number of the o-i-Pr and
Ar-H signals we conclude that the rotation round the Tip-CdC
bonds of 25 is slow on the NMR time scale at 273 K.
Consequently, the molecule adopts a chiral conformation and
judging by the X-ray data of other diaryl and triarylethylenes,56
including gem-ditipyl derivatives32,33 this is a propeller confor-
(51) Concerning the controversy about this terminology, see: Seppelt,
K. Angew. Chem., Int. Ed. Engl. 1992, 31, 292.
(52) For a short review, see: Larson, G. L. In The Chemistry of Organic
Silicon Compounds; Patai, S., Rappoport, Z., Eds.; Wiley: Chichester, 1989;
Chapter 11, pp 789-793.
(53) (a) Hayami, J.; Ono, N.; Kaji, A. Tetrahedron Lett. 1968, 1385. (b)
Corey, E. J.; Venkateswarlu, A. J. Am. Chem. Soc. 1972, 94, 6190.
(54) We tried to apply this approach in 1984 for generating a 1,1-enediol
by adding water to dimesitylketene followed by neutralization with aqueous
NH4Cl solution. However, we isolated dimesitylacetamide, presumably
formed by the reaction of 13 with ammonia and we did not pursue the
study.
(55) (a) Biali, S. E.; Rappoport, Z. J. Am. Chem. Soc. 1984, 106, 5641.
(b) Rappoport, Z.; Nugiel, D. A.; Biali, S. E. J. Org. Chem. 1988, 53, 4814.
(c) Nadler E. B.; Rappoport, Z. J. Am. Chem. Soc. 1989, 111, 213.
(56) Kaftory. M.; Nugiel, D. A.; Biali, S. E.; Rappoport, Z. J. Am. Chem.
Soc. 1989, 111, 8181.

Generation of a Carboxylic Acid Enol
J. Am. Chem. Soc., Vol. 118, No. 22, 1996 5175
Figure 8. Plots of (A) δ(OH) of enol 31 in the pure solvent vs δ(OH)
of enediol 25 in 42:5:3 solvent:THF-d8:H2O solvents (b), (B) δ(OH)
[31] vs δ(OH)[25] in 42:5:3 solvent:THF-d8:H2O (9) and (c) δ(OH)-
[30] vs δ(OH)[25] in 42:5:3 solvent:THF-d8:H2O (2). All data at 298
K.
Tip2CdC(OH)(OD) (25-D) (eq 8a). Even after correcting the
integration ratio for the different numbers of protons in the two
species, i.e., [25]/[25-D] ) 0.65:1, the extent of exchange is
unknown since the doubly exchanged enediol Tip2CdC(OD)2
(25-D2) is not detected in the OH region.
Figure 6. 1H NMR spectra of (A) ketene 20 (obtained 3 min after its
dissolution in DMF:THF:D2O/residual H2O) and (B) enediol 25 in
DMF-d7:THF-d8:H2O at 273 K. The signals marked S are the solvent
signals.
This appearance of separate OH signals for 25, 25-D, and
water as well as the sharp water signal in all the 25-containing
mixtures corroborates the slow exchange on the NMR time scale
of the enediolic OH groups with water.
Figure 7. 1H NMR spectrum of the OH region of partially O-
deuteriated 25 in DMF:THF:H2O:D2O spectrum: (A) O-H of 25 and
(B) O-H of 25-D.
Whereas the general appearance of the gem-ditipylvinyl
moiety is identical in all solvents, the δ(OH) value is strongly
solvent and temperature dependent (Table 2), as observed for
other di- and triaryl ethenols55 including the structurally related
30 and ditipylethenol 31. A plot of the δ(OH) values of 31 in
five ternary solvent mixtures (42:5:3 solvent:THF-d8:water)
against the δ(OH) values of 25 in the pure solvents at 298 K is
linear with a slope of 1.59 (R ) 0.9812, SD ) 0.18) (Figure
8A). However, when 31 was dissolved in the same ternary
solvent mixtures its OH shifted downfield by 0.13-1.20 ppm.
A plot of δ(OH)[31] versus δ(OH)[25] in the five ternary
mixtures at 298 K gave an approximate linear correlation with
a slope of 1.10 (R ) 0.9622, SD ) 0.18) (Figure 8B).
mation, cf. 29. Since heating to 298 K does not cause peak
broadening we calculate that the activation barrier for enanti-
omerization is >14.2 kcal mol-1
. This value should be
compared with the barrier of 15.9 kcal mol-1 in DMSO for the
sterically close 1,1-ditipylpropen-2-ol 30. An accurate deter-
mination of the rotation barrier by DNMR is impractical since
the coalescence temperature (Tc) of the diastereotopic Tip-H
signals with ∆ν ) 108.9 Hz (at 400 MHz) in 42:5:3 DMF:
THF:H2O should be higher than 298 K (it is 331 K for 30 for
which ∆ν ) 101.8 Hz in DMSO at 400 MHz) and 25 starts to
isomerize to 19 > 273 K.
That the low field singlet is due to the enediolic hydroxyl
was corroborated by an OH/OD exchange experiment. To a
>95% pure sample of 25 in a 42:5:3 DMF-d7:THF-d8:H2O
mixture at 273 K, 0.05 mL of D2O was added, and the new
NMR spectrum was immediately recorded. Interestingly, in
contrast with the usual fast deuteron-hydron OH exchange of
alcohols leading to an immediate disappearance of the OH
signal, we did not detect any spectral changes. Only when the
spectrum was recorded again after keeping the sample at -18
°C for 3 days did a new signal at 0.027 ppm higher field appear
side by side with the old OH signal [integration ratio:1 (old):
0.65(new) (Figure 7)]. We ascribe the slow exchange to
hydrogen bonding of the OH groups to the good hydrogen bond
accepting solvent DMF. Since deuterium is electron donating
compared with hydrogen,57 the new high field signal was
assigned to the OH group of the monodeuteriated enediol
The sensitivity to changes of the medium are reflected by
the different slopes in the two media. A slope close to unity in
the ternary solvent mixture indicates a similar interaction of 25
and 31 with the medium. The addition of 6% water and 10%
THF to the solution of 31 results in small ∆δ(OH) values of
0.12-0.14 ppm in the good hydrogen bond accepting solvents
DMF and DMSO and a more substantial ∆δ(OH) value in
poorer hydrogen bond acceptors acetonitrile and acetone (0.81-
(57) March, J. AdVanced Organic Chemistry, 3rd ed.; Wiley: New York,
1985; p 18.

5176 J. Am. Chem. Soc., Vol. 118, No. 22, 1996
Frey and Rappoport
Table 1. Selected Structural Parameters for Ketenes 20 and 13a
Scheme 3
structural parameter
20
13
Bond Lengths (Å)
1.166(2)
C(1)-O
C(1)-C(2)
1.18(1)
1.29(1)
1.298(3)
Bond Angles (deg)
179.5(2)
O-C(1)-C(2)
176(1)
C(1)-C(2)-C(3)
C(1)-C(2)-C(9)
C(3)-C(2)-C(9)
118.1(2)
118.7(2)
121.3(1)
115.5(8)
119.3(8)
125.0
1.20 ppm). This is attributed to a combined change of the
polarity and the hydrogen bond accepting ability of the medium
caused by the addition of water which is more important for
the less hydrogen bond accepting solvent.58 The better an
hydrogen bond acceptor is the solvent, the more attenuated will
be the effect of the added water.
Dihedral Angles (deg)
Ar-C(2)-C(1)
Ar-C(2)-C(1)
Ar-C(2)-Ar
55.88b
54.13c
85.98
48.8
56.8
82.1
a The parameters for 13 are taken from ref 45 or calculated from it.
The numbering scheme has been adapted to that of 20 in Figure 3.b Ar
) C(3)-C(8). c Ar ) C(9)-C(14).
The slight difference in solvation of 31 and 25 may result
from the different steric bulk of the R-OH and R-H. Hence,
we measured the δ(OH) values for the R-Me enol 30 in the
five ternary solvent mixtures, since OH is closer in size to Me
than to H. The plot of δ(OH) [30] vs δ(OH)[25] is close to
linear with a slope of 1.20 (R ) 0.9891, SD ) 0.11) (Figure
8C) indicating only a slightly higher sensitivity of δ(OH)[30]
to the medium changes. The small steric effect is consistent
with a previous conclusion that R-substituents in R-alkyl-â,â-
dimesitylethenols influence the association with the solvent
mainly by a polar rather than a steric effect.55b
in a syn-type conformer, we believe that an anti arrangement
of both groups, i.e., as in the solvated aa conformation 32, is
implicated.
The solvent dependent shift of the δ(OH) values in various
di- and trimesityl substituted ethenols55 have been interpreted
in terms of different associations of the OH group. An anti-
type conformation of the enol moiety in hydrogen bond
accepting solvents is stabilized by hydrogen bonding, whereas
a syn-planar conformation is stabilized by an intramolecular
O-H‚‚‚π(Ar) hydrogen bond involving the aromatic ring cis
to the hydroxyl prevails in non-hydrogen-bond accepting
solvents. The same interactions should be also important for
25.
The two OH groups in the sa conformer should display two
different H NMR signals. The observation of a single OH
1
signal is consistent with either a symmetrical conformation, ss
(33) or aa (32), or with an sa (34) conformation whose two OH
hydrogens exchange rapidly on the NMR time scale. The
appearance of separate signals for enols55a or for 25 and water
suggest that such exchange is slow. A better tool for analyzing
the CdC-O-H dihedral angle of the dC(OH)2 moiety of 25
is the 4J(HOCOH) coupling constant. That it was not observed
could corroborate the existence of the symmetrical 32 confor-
mation. Since hydrogen bonding stabilize the CdC-O-H
moiety, we tentatively suggest that each OH group is hydrogen
bonded as depicted in 32. Furthermore, for steric reasons, and
in analogy with the enols,55 the CdC-O-H moieties are not
periplanar but clinal, cf. 32.
Conformation of the CdC-O-H Moieties. The appear-
ance of a single OH signal have a bearing on the conformation
of the CdC-OH moieties of the enediol. Since the OH group
was never hitherto experimentally “observed” a brief review is
in order. Several ab initio calculations on ethene-1,1-diol are
available. Early STO-3G/STO-3G calculations2 indicated that
out of three possible conformations of the hydroxy groups
(Scheme 3) the syn,syn (ss) and syn,anti (sa) conformations are
stable, the sa being preferred by 2 kcal mol-1 whereas the anti,-
anti (aa) confomer is a saddle point on the potential energy
surface. The relative destabilization of the ss conformer was
ascribed to repulsion between the vinylic and hydroxylic
hydrogens. The preference of the sa conformer is shown also
by 3-21G//3-21G, 6-31G*//3-21G,7 6-31G**/6-31G** 59 and
recently by high level 10-in-10/6-31G* CASSCF calculations.8
The sa conformer is also the most stable for 1-cyclopropane-
and 3-cyclopropenecarboxylic acids.7
The δ(OH) values of the 1,1-ethenediols 10 and 8 which were
prepared by the ketene hydration method in DMF:THF:water
mixtures at 273 K (see below) are 9.63 and 9.44 ppm,
respectively, compared with 9.65 ppm for 25.
(b)13C NMR. Additional corroboration for the structure of
25 is obtained from the 13C NMR spectra,60 at 273 K in THF-
d8:H2O and in DMF-d7:THF-d8:H2O which display 17 separate
signals completely assigned by a gated decoupled 13C NMR
spectrum (Table 3), in accord with a chiral propeller structure
29 having a C2 symmetry. In acetone-d6:THF-d7:H2O, however,
where hydration and tautomerization rates are comparable, the
spectrum display signals for a mixture of 19, 20, and 25 and a
complete assignment is hampered. Nevertheless, CR and Câ
were unequivocally assigned in a gated decoupled 13C NMR
spectrum.
These results could not be a priori assumed to apply for 25
for three reasons. (i) The steric interactions are much larger in
25 than in the parent enediol. (ii) Hydrogen bonding of each
OH group to the tipyl group cis to it could lower the energy of
the ss conformation. (iii) In the hydrogen bond accepting
solvents used, especially DMF, the conformation may be anti,
as suggested for the analogous enols.55 The position of the OH
signal and the δ(OH)[30] vs δ(OH)[31] correlation strongly
suggest that solvation is important. Since space filling models
show that there is no space for an associating solvent molecule
Comparison of δCR and δCâ of 25 with those of 31 (δ(CR)
) 144.02(143.69) and δ(Câ) )110.93 (110.48) at room tem-
61
perature (or 240 K) in CDCl3 and 144.70 and 108.62 ppm,
(60) For a review on NMR spectra of enols, see: Floris, B. In The
Chemistry of Enols; Rappoport, Z., Ed.; Wiley: Chichester, 1990; Chapter
4, p 147.
(58) Marcus, Y. Chem. Soc. ReV. 1993, 409.
(59) Rodler, M. Chem. Phys. 1986, 105, 345.

Generation of a Carboxylic Acid Enol
J. Am. Chem. Soc., Vol. 118, No. 22, 1996 5177
Table 2. δ(OH) Values (ppm) for Tip2CdC(OH)R in Several
Solvent Mixturesa
solventb
25
31
30
T (K)
273
298
273
298
273
298
DMSO-d6
c
8.89
c
c
9.06
c
8.20
8.92d
9.21
DMF-d7
THF-d8
9.65
8.50
8.58
7.92
9.33
8.28
8.32
7.65
9.45
9.31d
8.70
7.76d
8.76
7.95d
7.79
6.59d
8.68
7.50
7.67
6.87
8.45
7.18
7.34
6.54
9.08d
8.38
7.54d
7.98
Me2CO-d6
MeCN-d3
7.68d
7.46
6.59d
a Chemical shifts are relative to internal TMS standard. b Solvent
used: solvent given: THF-d8: water in a 42:5:3 v/v ratio. c Not
determined since the solvent freezes. d δ(OH) value in the neat solvent.
Table 3. 13C NMR Chemical Shifts (in ppm) for 25 in Several
Figure 9. The CR-OL (L ) H, D) region in the 13C NMR spectrum of
25: (A) 25; (B)25-D; and (C) 25-D2. Solvent: 42:5:3 DMF:THF:H2O.
The spectrum was obtained after 4 h acquisition time.
Solvents at 273 Ka
solvent
b
assignment
i-Pr-Me
DMF-d7
THF-d8
Me2CO-d6
decoupled 13C NMR spectrum of the mixture will then display
isotopic split of all the signals for carbons that are within three
bonds of the deuterium. The new signals are shifted upfield
by 0.09-0.12 ppm in the case of a â-isotope effect (i.e., C-XD
carbon) and by 0.07 ppm in the case of a γ-isotope effect (i.e.,
C-C-XD). This probe is applicable also for enols since when
we added 50 µL of a 1:1 H2O:D2O mixture to enol 31 in 1:5
THF-d8:CDCl3 at 240 K we found in the 13C NMR spectra two
new signals shifted upfield relative to the original signals by
0.11 ppm for CR and by 0.12 for Câ (Figure 9). Three
isotopomers are expected for 25, and when a 1:1 H2O:D2O
mixture was added to 20 three resolved 13CR(OL)2 (L ) H, D)
signals at 156.69, 156.60 and 156.52 ppm ascribed, respectively,
to 25, 25-D, and 25-D2 were indeed observed (Figure 9).63 In
contrast with 31, we did not observe the splitting of the signal
for Câ. The reason for the lower splitting of CR (and Câ) of 25
compared with 31 is unclear.
22.89
24.49
22.93
24.37
24.68
24.53
24.82
24.85
24.93
25.08
25.99
25.98
i-Pr-CH
25.99
30.65
30.91
31.08
34.42
34.95
Câ
80.40
80.45
81.16
m-Tip-C
121.63
122.63
137.48
148.98
146.24
149.11
157.02
121.64
122.72
136.80
149.18
146.49
149.30
156.68
ipso-Tip-C
p-Tip-C
o-Tip-C
CR
156.36
a All chemical shifts vs internal TMS standard. b See text.
(c) 17O NMR. Attempts to measure the 17O NMR spectrum
of 25 in aqueous THF, using natural abundance 17O which was
used successfully with 2,2-diarylethenols61 had failed. Only
signals of THF-17O and H217O were observed. This may be
due to the much larger concentration of the solvents compared
with 25 and perhaps to a large line width of the signal for
25, as found for the enols,61 which may result in overlap of
signals.64
respectively, in DMSO at room temperature61 ) shows that CR
undergoes a substantial downfield shift of ca. 14 ppm and Câ a
substantial upfield shift of ca. 29 ppm. This is ascribed to the
contribution of degenerate structures 25a and 25b (eq 9) in
which Câ is negatively charged and CR is attached to a positively
charged atom. In 31 there is only one parallel structure with a
charge on the single OH group. This observation is reminiscent
of the ∆δ(OH) that we have measured for CCl3CH(OH)2 and
Cl3CCH2OH in DMF-d7 at 298 K (respectively δ 96.46 vs 76.15
for CR and 104.95 vs 99.22 for Câ) in spite of the absence of
the double bond in these systems.
(d) FTIR. Measurement of the IR sectra of enediols in the
ν
O-H stretching region at 4000-3000 cm-1 is impractical since
even at very low water concentration and with short optical path,
the OH absorption around 3400 cm-1 is broad and total.65
Likewise, the CdC stretching region is covered by the solvent
peak.
Kinetic Stability of the 2,2-Diarylethene-1,1-diols. The low
thermodynamic stabilities of 1,1-enediols reviewed above also
applies to the bulky 1,1-enediol 25 which is transformed
irreversibly within the NMR detection limit to the acid 19.
Likewise, a sample of 19 which was kept for 3 months at room
temperature in DMSO-d6 both in the presence or the absence
Another corroborating evidence for the 1,1-enediolic stucture
with apparently identical two OH groups comes from a 13C
SIMPLE (Secondary Isotope Multiplets of Partially Labeled
Entities) NMR experiment.62 In a solvent containing a 1:1
mixture of H2O and D2O, the hydroxyl groups of ROH are 1:1
OH to OD due to proton-deuteron exchange (eq 8b). A noise-
(63) The relative intensities of the three signals depend on the acquisition
time of the spectra. This is probably associated with slow exchange rates
of 25, 25-D, and 25-D2 with H2O, D2O, and among themselves, which are
likely to differ for each species. The spectrum in Figure 9 was taken after
4 h.
(64) We thank Prof. G. Cerioni from the University of Cagliari, Italy
for this experiment.
(61) Frey, J.; Eventova, I.; Rappoport, Z.; Mu¨ller, T.; Takai, Y.; Sawada,
M. J. Chem. Soc., Perkin Trans. 2 1995, 621.
(62) (a) Christofides, J. C.; Davies, D. B. J. Am. Chem. Soc. 1983, 105,
5099. (b) Reuben, J. J. Am. Chem. Soc. 1985, 107, 1756.
(65) Perkins, W. D. J. Chem. Educ. 1987, 64, A296.

5178 J. Am. Chem. Soc., Vol. 118, No. 22, 1996
Frey and Rappoport
of a catalytic amount of TFA did not show any trace of 25.
Consequently, the “observed” stability conferred on 25 by the
tipyl rings is only kinetic. A thermodynamic stabilization by
several orders of magnitude compared with parent 1,1-ethenediol
is expected by analogy with the kinetically and thermodynami-
cally stabilized 2,2-dimesityl- and 2,2-ditipylethenols,27,33 but
it is unobservable by our detection technique.
10 were observed at 273 K and their tautomerization rates
measured (see above). Hence, 20 hydrates much slower than
its dimesityl and bis(pentamethylphenyl) analogs previously
studied.29
Increased steric bulk reduces significantly the hydration rate
as shown, e.g., by comparing the rates for di-tert-butylketene
with less hindered ketenes.69 The steric bulk caused by the two
geminal tipyl and mesityl rings on the ketene could be probed
by comparing the solid state structures of 20 and of 1341 given
in Table 1. The two different aryl-CdC dihedral angles of 13
differ slightly from those of 20, presumably due to the repulsion
of isopropyl groups at a spatial proximity. In both ketenes the
two ring planes are close to perpendicular. The deviation of
4° of the C-C-O bond of 13 from linearity presumably results
from crystal asymmetry.41
The difference in spatial proximity of the ketene’s carbons
and oxygen and the ortho methyls (in 13) or the isopropyls (in
20) could have been the reason for the substantial lower
hydration rate of 20 relative to 13 by an in-plane nucleophilic
attack at the ketene’s CR.24 The nonbonded distances between
the ortho methyls and the oxygen are 3.46-5.05 Å in 13 and
3.52-4.94 Å between the o-i-PrC-H carbon and the oxygen in
20. The distances from CR and Câ were 2.90-3.83 Å and 2.94-
2.99 Å for 13, 3.06-3.93 Å, and 2.97-3.03 Å for 20,
respectively. Since the tipyl C-H hydrogens in the crystal point
toward the ketene moiety and the two methyls are directed
backwards (Figure 3), the comparison of the nonbonded
distances between the ortho substituents and the ketene moiety
of the two ketenes is justified. These differences are not
sufficiently significant to account for the considerable hydration
rate difference between 20 and 13. Inspection of the space
filling models shows that the isopropyl methyls confer some
additional steric hindrance on the ketene skeleton from a more
remote position, and approach of water molecules in the CdO
bond plane may become more encumbered due to the presence
of some of the distant methyl groups.
The tautomerization t1/2 values of the short lived 1,1-enediols
of Kresge and Wirz were of the order of 10-6 s,19-21 whereas
those of Hegarty’s 1,1-enediols in MeCN-H2O were longer,29
although rate constants were not given. By starting from a
sample of >90% of 25 in 42:5:3DMF-d7:THF-d8:water at 303
K we conducted a qualitative kinetic analysis of the 25 f 19
transformation. From the build-up of the R-CH signal of 19
with time the reaction showed a first order behavior with a t1/2
of ca. 98 min, while the first order decay of one or two of the
aromatic signals of 25 gave t1/2 ) 93 min. Enediols 8 and 10
tautomerized faster with t1/2 of ca. 9 and 7 min, respectively.
The difference between them is much smaller than the reported29
three- to four-fold decrease in the ketonization rate of 8 relative
to 10. However, our results are only semiquantitative.
The tautomerization of 25 to 19 in THF-d8:water at 273 K is
very slow. However, the reaction is accelerated by addition of
a catalytic amount of TFA presumably due to a TFA-promoted
ketonization when the formation of 19 obey first order kinetics
with t1/2 ) ca. 60 min.
These lifetimes suggest that 25 should be regarded as the
first rather relatively “long lived” carboxylic acid enol. Its
kinetic stability, conferred by the bulky tipyl rings, allows the
hitherto impossible unequivocal spectroscopic characterization
of the 1,1-enediol intermediate in the ketene hydration reaction.
Solvent Effect on Hydration Rates and Equilibria. The
hydration rate of 20 is highly solvent dependent. At 273 K in
a DMF:THF:water mixture >98% of 20 had reacted after 90
min, in a THF:water mixture it took 6 h to obtain 95% 25 and
in MeCN:THF:water and Me2CO:THF:water <10% of 20
reacted after 10 h. The rate was higher at room temperature,
but both 25 and 19 were formed. Hegarty29 reported a low
solvent sensitivity of the hydration rate of dimesitylketene 13
in MeCN, 1,4-dioxane, and t-BuOH and following Bothe et al.,66
he suggested formation of a cyclic transition state of low
polarity. Likewise, we ascribe the differences in hydration rates
in different solvents to different solvation of both water and
20.
The hydration rate differences of several diarylketenes are
also consistent with their 13C and 17O NMR spectra.70 Table 4
summarizes the chemical shifts for four diaryl substituted
ketenes in CDCl3 at 298 K. The large 13C upfield shifts of Câ
relative to the average vinylic shifts at ca. 120 ppm reflect the
high negative charge density on Câ as illustrated by structure
35b. The bulkier the aryl substituent, the more upfield shifted
the 13Câ and 17O resonances, with the exception of 13Câ of 11
whose deviation is due to the inductive effect and/or the
buttressing effect of its four m-Me groups. δCR depends only
moderately on steric hindrance.
Since hydration of 20 is reversible and both 20 and its
“hydrate” 25 are present in equilibrium67 the more solvent
stabilized the enediol, the less ketene is expected to be in the
equilibrium. This is consistent with our finding that at 273 K
ketene 20 is hydrated almost completely in the solvents with
higher hydrogen-bond accepting ability parameter â68 DMF and
THF but not in MeCN and Me2CO.
Ground State Structure of 20 as a Reason for Its Slow
Hydration. The hydration rate of 20 was studied only
qualitatively since a comprehensive kinetic analysis was pro-
hibited by the high cost of the DMF-d7 and THF-d8. Half lives
(t1/2) of 130 min in a mixed THF-d8:water solvent and of 40
min in a 42:5:3 DMF-d7:THF-d8:water mixture were measured
at 273 K. Under the same conditions, hydrations of ketenes
11 and 13 were too fast to be followed although enediols 8 and
The shielding of 13Câ and 17O is rationalized by the different
oxygen hybridization in 35a and 35b.70 The two lone pairs of
the sp2 hybridized oxygen in 35a are relatively close to the
substituents, whereas the single lone pair of the sp-hybridized
oxygen of 35a is further away from them. The observed
shielding is therefore attributed to a larger contribution of 35b
having a lower steric requirement compared with 35a. Con-
sequently, the nucleophilic attack of water on CR will be
substantially slowed down the more significant the contribution
of 35b.
(66) Bothe, E.; Dessouki, A. M.; Schulte-Frohlinde, D. J. Phys. Chem.
1980, 84, 3270.
(67) Frey, J.; Rappoport, Z. J. Am. Chem. Soc. 1995, 117, 1161 and
1996, 118, 5182-5191.
(68) (a) Kamlet, M. J.; Abboud, J.-L. M.; Taft, R. Prog. Phys. Org. Chem.
1981, 13, 485. (b) For a list of recent â values, see: Kamlet, M. J.; Doherty,
R. M.; Abraham, M. H.; Carr, P. W.; Doherty, R. F.; Taft, R. J. Phys.
Chem. 1987, 91, 1996.
(69) Allen, A. D.; Tidwell, T. T. J. Am. Chem. Soc. 1987, 109, 2774.
(70) Cerioni, G.; Plumitallo, A.; Frey, J.; Rappoport, Z. Magn. Res. Chem.
1995, 33, 669.
(71) Nilsson, A. Acta Chem. Scand. 1967, 9, 21.

Generation of a Carboxylic Acid Enol
J. Am. Chem. Soc., Vol. 118, No. 22, 1996 5179
Table 4. 13C and 17O Chemical Shifts (ppm) of Diarylketenes
Ar2CdCdO70
at 200.13, 300.13, and 400.13 MHz for 1H and 50.32, 75.48, and 100.62
MHz for 13C, respectively. The ESR spectrum of radical 27 was
measured in the Institute for Fundamental Research of Organic
Chemistry in Kyushu University, Japan on a Bruker ESP 300
spectrometer in microcapillary tubes without prior degassing of the
sample in 9:1 THF:water. Electron impact (EI) and chemical ionization
(CI) mass spectra were recorded on a MAT-311 and on Finnigan MAT
4600 spectrometers. High resolution mass spectral analyses (HRMS)
were conducted at the Mass Spectrometry Center at the Technion, Haifa
on a Finnigan MAT 711 apparatus. X-ray diffraction data of the single
crystals were measured using a PW1100/20 Philips Four-Circle
computer-controlled diffractometer. Mo KR (λ ) 0.710 69 Å) radiation
with a graphite crystal monochromator in the incident beam was used.
All crystallographic computing was done on a VAX 9000 computer
using the TEXSAN structure analysis software.
Solvents and Materials. Tetrahydrofuran (THF) was distilled from
sodium benzophenone ketyl, ether was distilled from LiAlH4 and
benzene and ethyl formate were distilled from P2O5 immediately before
use. CCl4 was dried by standing over MgSO4 for 30 min and was
then filtered. All purchased reagents were the best commercial samples
and were used without purification. Polydeuteriated solvents for NMR
spectroscopy (Aldrich, except for DMF-d7 from Ferak, Berlin, Germany)
were dried over 4 Å (DMSO-d6 and DMF-d7) or 3 Å ((CD3)2CO and
CD3CN) molecular sieves. Small portions of THF-d8 were kept in
separate sealed ampules. The solvents used for chromatography were
not purified. 1,3,5-Triisopropylbenzene was kindly donated by Dr.
Hartig from Huels AG, Germany.
δ
13Ca
δ
17Ob,c
CdCdO
ketene
CdCdO
CdCdO
14
13
11
20
47.0
38.9
41.9
33.4
201.2
191.4
191.3
191.5
340(250)
310 (230)
301 (470)
292 (765)
a At 100.64 MHz. b At 40.66 MHz. c Values in parentheses are
estimated half-height line widths in Hz.
The possibility that different extents of the reversibility of
hydration for the various ketenes67 contribute to the difference
in hydration rates will be discussed in a future publication.
Could 1,1-Enediols be Isolated? Attempted Isolation of
25. The formation of g98% of enediol 25 in solution, coupled
with its relatively long lifetime raises the question whether it
could be isolated. We know that within the experimental error
of the NMR detection probe 25 is “completely” converted to
19, i.e., the pKenol value is >2, although we believe that the
value is much higher than 2. Nevertheless, we think that in
spite of this it is feasible to isolate enediols structurally related
to 25 although we failed so far in the attempted isolation of 25
due to a combination of inherent and practical reasons. The
ketene hydration approach seems advantageous over the two
other preparative approaches to 1,1-enediols of Scheme 2 due
to the mild reaction conditions and the minimal side products
formation when the only reagents are water and the ketene. The
difficulties encountered in many attempts to isolate 25 include
the need to work in the cold in order to prevent tautomerization
which sometimes results in precipitation of the water, and high
solubility of 25 which prevented its precipitation on adding
excess water to the reaction mixture in either DMF or THF.
Air oxidation to radical 27 took place under all conditions tested.
Likewise, attempted isolation of 25 from a THF-water mixture
by lyophilizing the solvents at low temperatures afforded only
27.
1-Bromo-2,4,6-triisopropylbenzene was prepared according to
Nilsson et al.71
Bis(2,4,6-triisopropylphenyl)methyl Formate (16). To a solution
of 1-bromo-2,4,6-triisopropylbenzene (25 g, 88.25 mmol) in dry ether
(150 mL) a 1.6 M solution of n-butyllithium in hexane (56 mL, 89.6
mmol) was added at room temperature. The reaction mixture was
stirred in an argon atmosphere for 1 h and then cooled to 0 °C. Dry
ethyl formate (20 mL, 248 mmol) was added dropwise over a period
of 15 min, and the resulting mixture was stirred at 0 °C for 30 min and
at room temperature for additional 12 h. The reaction was terminated
by adding 10% aqueous HCl (100 mL). The organic phase was
separated, washed with water, and dried (MgSO4), and the solvent was
evaporated. The residual yellowish oil was redissolved in methanol
(50 mL), and the white suspension that immediatly formed was
dissolved on heating. On cooling, white crystals of bis(2,4,6-triiso-
propylphenyl)methyl formate (5.5 g, 27%), mp 141-2 °C were isolated.
1H NMR (200 MHz, CDCl3) δ: 0.99 (12H, d, J ) 6.7 Hz, o-i-Pr-Me),
1.06 (12H, d, J ) 6.7 Hz, o-i-Pr-Me), 1.21 (12H, d, J ) 6.9 Hz, p-i-
Pr-Me), 2.84 (2H, m, J ) 6.9 Hz, p-i-Pr-CH), 3.40 (4H, m, J ) 6.7
Hz, o-i-Pr-CH), 6.97 (4H, s, Tip-H), 8.06 (1H, s, Tip2CH), 8.20 (1H,
s, CHO). 13C NMR: (50.32 MHz, CDCl3) δ: 23.89 (p-i-Pr-Me), 24.20
(o-i-Pr-Me), 29.01(o-i-Pr-CH), 33.95 (p-i-Pr-CH), 72.04 (Tip2CH),
122.31 (m-Tip-C), 131.45 (ipso-Tip-C), 147.51 (o-Tip-C), 148.29 (p-
Tip-C), 159.95(CdO). FTIR: νmax cm-1 (Nujol): 1726 (s, CdO), 1610
(m, CdC). MS (EI, 70 eV, m/z, relative abundance, assignment): 464
(32%, M), 421 (50%, M - i-Pr), 375 (B, M - i-Pr - HCOOH), 337
(49%, M - i-Pr - 2MeCHdCH2), 260 (59%, M - TipH), 232 (52%,
TipCHO). Anal. Calcd for C32H48O2: C, 82.70; H, 10.41. Found:
C, 82.41; H, 10.39.
These obstacles can be overcome, and, by using bulkier, or
strongly electron withdrawing substituents, isolation of a 1,1-
enediol is only a matter of time.
A General Comment. The unequivocal structural assign-
ment of the hydration product of ketene 20 as the 1,1-enediol
25 and the similar, only much less studied, behavior of other
diaryl ketenes, shows that hydration of these ketenes is initiated
by addition of water to the CdO bond. This has a bearing on
the mechanistic argument concerning the details of ketene
hydration.2,8,24 A relevant mechanistic point is that hydration
of diarylketenes is reversible as shown in detail for the 20/25
system.67 This and reactions of 1,1-enediol 25 which involve
several reaction sites of 25 are discussed in the following paper
(J. Am. Chem. Soc. 1996, 118, 5182-5191).67b
Conclusions. 2,2-Ditipylethene-1,1-diol was prepared in
solution in g98% purity at 273 K as a sufficiently persistent
species for spectroscopic characterization. Its structure was
2,4,6,2′,4′,6′-Hexaisopropylbenzhydrol (15). (a) From Bis(2,4,6-
triisopropylphenyl)methyl Formate. Bis(2,4,6-triisopropylphenyl)-
methyl formate (5.9 g, 12.7 mmol) was dissolved in dry ether (75 mL).
LiAlH4 (1 g, 26.4 mmol) was added, and the resulting suspension was
stirred under protection from air at room temperature for 2 h and then
cooled to 0 °C, and the excess LiAlH4 was destroyed with ethyl acetate
(20 mL), followed by 5% aqueous HCl (30 mL). The organic layer
was separated, washed with water (30 mL), dried (MgSO4) and
evaporated to dryness leaving a solid (5.6 g). Recrystallization from
isopropyl alcohol-water afforded 5.1 g (92%) of white crystals of
2,4,6,2′,4′,6′-hexaisopropylbenzhydrol, mp 88-9 °C. 1H NMR (200
MHz, CDCl3) δ: 1.01 (12H, d, J ) 6.8 Hz, o-i-Pr-Me), 1.07 (12H, d,
J ) 6.8 Hz, o-i-Pr-Me), 1.22 (12H, d, J ) 6.8 Hz, p-i-Pr-Me), 1.74
(1H, d, J ) 4 Hz, OH), 2.83 (2H, m, J ) 6.8 Hz, p-i-Pr-CH), 3.54
(2H, m, J ) 6.8 Hz, o-i-Pr-CH), 6.75 (1H, d, J ) 4 Hz, Tip2CH), 6.97
(4H, s, Tip-H). 13C NMR: (50.32 MHz, CDCl3) δ: 23.97 (p-i-Pr-
1
corroborated by H and 13C NMR techniques and exchange
experiments. A symmetric arrangement of the CdC-O-H
moieties, presumably anti, anti, with solvated OH groups, seems
likely. On raising the temperature it tautomerizes to ditipylacetic
acid. This is the longest-lived 1,1-enediol prepared so far.
Experimental Section
General Methods. Melting points were measured on a Thomas-
Hoover apparatus and are uncorrected. UV spectra were measured on
a Uvikon 930 spectrophotometer and FT infrared spectra on a Nicolet
Impact 400 spectrometer. NMR spectra were recorded on Bruker WP
200 SY, WH-300, and AMX-400 pulsed FT spectrometers operating

5180 J. Am. Chem. Soc., Vol. 118, No. 22, 1996
Frey and Rappoport
Me), 24.32 (o-i-Pr-Me), 24.60 (o-i-Pr-Me), 28.96 (o-i-Pr-CH), 34.00
(p-i-Pr-CH), 72.55 (Tip2CH), 122.20 (m-Tip-C), 135.46 (ipso-Tip-C),
147.18 (o-Tip-C), 147.57(p-Tip-C). FTIR: νmax cm-1 (Nujol): 3616
(w), 3411 (br, OH). MS (EI, 70 eV, m/z, relative abundance,
assignment): 436 (3%, M), 375 (B, M - i-Pr - H2O), 231 (90%,
TipCO), 189 (22%, TipH - Me). Anal. Calcd for C31H48O: C, 85.25;
H, 11.08. Found: C, 85.27; H, 10.93.
The 1H NMR of the crude product was identical with that obtained in
(a) above and no IR band at 2200 cm-1 was observed.
(c) With KCN/18-Crown-6 in Benzene. A mixture containing a
suspension of KCN (300 mg, 3.6 mmol), 18-crown-6 (100 mg, 0.38
mmol), and ditipylchloromethane (2 g, 4.4 mmol) in benzene (50 mL)
was refluxed under constant argon flow. The reaction was followed
by monitoring the disappearance of the purple color formed on
absorption of ditipylchloromethane onto the silica gel of a TLC plate.
After 6 h the mixture was cooled to room temperature and filtered, the
solvent was evaporated, and the residual yellowish oil was chromato-
graphed twice on silica gel using 30:1 high boiling petroleum ether:
ether as eluent. The first fraction afforded a white solid, mp 114 °C
after recrystallization from methanol (50 mg, 2.5%), identified as the
nitrile 24. 1H NMR (400 MHz, CDCl3) δ: 0.66 (3H, d, J ) 6.7 Hz,
i-Pr-Me), 0.79 (3H, d, J ) 6.8 Hz, i-Pr-Me), 0.97 (3H, d, J ) 6.9 Hz,
i-Pr-Me), 1.03 (3H, d, J ) 6.9 Hz, i-Pr-Me), 1.06 (3H, d, J ) 6.8 Hz,
i-Pr-Me), 1.08 (3H, d, J ) 6.7 Hz, i-Pr-Me), 1.09 (3H, d, J ) 6.9 Hz,
i-Pr-Me), 1.13 (3H, d, J ) 6.7 Hz, i-Pr-Me), 1.20 (3H, d, J ) 6.8 Hz,
i-Pr-Me), 1.23 (6H, d, J ) 6.9 Hz, i-Pr-Me), 2.00 (1H, m, J ) 6.7 Hz,
i-Pr-CH), 2.20 (1H, m, J ) 6.7 Hz, i-Pr-CH), 2.34 (1H, m, J ) 6.8
Hz, i-Pr-CH), 2.86 (2H, m, J ) 6.9 Hz, i-Pr-CH), 3.18 (1H, m, J )
6.7 Hz, i-Pr-CH), 5.68 (1H, d, J ) 1.2 Hz, HCdC-CdCH), 5.70 (1H,
d, J ) 1.2 Hz, HCdC-CdCH), 6.90 (1H, d, J ) 1.6 Hz, Tip-H), 6.93
(1H, d, J ) 1.6 Hz, Tip-H), 7.01 (1H, s, HCdC). 13C NMR: (100.64
MHz, CDCl3) δ: 18.43, 18.78, 20.33, 20.76, 21.95, 22.09, 22.78, 23.97,
24.13, 24.25, 25.48 (12 i-Pr-Me), 29.24, 29.48, 31.39, 32.76, 33.09,
34.24 (6 i-Pr-CH), 53.38 (C-CtN), 119.41, 120.02, 120.63, 121.20
(4HCdC-CdCH), 121.13(TipCHdC) 129.92(CtN), 131.52, 137.53,
142.58, 145.37, 145.70, 146.76, 148.20 (1 vinylic and 6 aromatic C).
IR: νmax cm-1 (Nujol): 1760 (m, CdN), 1620 (m, CdC). MS(CI,
i-C4H10) m/z (relative abundance, assignment) 446 (3%, MH+), 419
(B, M - CN), 402 (13%, M - i-Pr), 377 (5%, M - MeCHdCH2 -
CN). MS(EI) m/z (relative abundance, assignment): 402 (B, M - i-Pr),
360 (63%, M - i-Pr - MeCHdCH2), 318 (13%, M - i-Pr -
2MeCHdCH2), 200 (74%, M - Tip - i-Pr), 158 (14%, M - Tip -
i-Pr - MeCHdCH2). Anal. Calcd for C32H47N: C, 86.22; H, 10.62.
Found: C, 86.35; H, 10.43. Crystal data: C32H47N, space group P21/
n, a ) 30.047(4) Å, b ) 10.208(2) Å, c ) 9.625(1) Å, â ) 93.91(1),
(b) From 1-Bromo-2,4,6-triisopropylbenzene and Ethyl Formate.
To a stirred solution of 1-bromo-2,4,6-triisopropylbenzene (17 g, 60
mmol) in dry THF (150 mL) was added BuLi (1.2 M in hexane, 50
mL) in an argon atmosphere at -70 °C. After 20 min at -70 °C,
ethyl formate (2 mL, 30 mmol) was added dropwise, and at the end of
the addition the solution was allowed to reach room temperature. The
mixture was then refluxed for 18 h and then poured into 10% aqueous
NH4Cl solution (250 mL). The solvent was evaporated, the residual
suspension was extracted twice with CH2Cl2 (100 mL), the phases were
separated, and the organic phase was washed with water (100 mL),
brine (100 mL), water (100 mL), and dried (MgSO4). The solvent was
evaporated, and the residual oil was chromatographed on silica gel using
petroleum ether (40-60 °C):ether as the eluent. 1,3,5-Triisopropyl-
benzene (6 g, 48%) was eluted first, followed by 2,4,6,2′,4′,6′-
hexaisopropylbenzhydrol which was coeluted with a yellow unidentified
compound. After several crystallizations from aqueous i-PrOH the pure
1
benzhydrol (4.15 g, 32%) was obtained. The H NMR spectrum of
the crude product mixture indicated the presence of small quantities of
bis(2,4,6-triisopropylphenyl)methyl formate and (2,4,6-triisopropylphe-
nyl)carboxaldehyde, but these products were not isolated from the
mixture.
Bis(2,4,6-triisopropylphenyl)chloromethane (17). A stream of HCl
was passed for 30 min into a solution of 2,4,6,2′,4′,6′-hexaisopropyl-
benzhydrol 15 (5.6 g, 12.8 mmol) in dry benzene (75 mL) at room
temperature. Dry ice (3 g) was then added to the solution followed by
sodium carbonate (3 g). The mixture was allowed to reach room
temperature, the solids were filtered, and the solvent was evaporated
leaving a yellowish oily solid which we were unable to recrystallize
due to its high solubility in all solvents investigated. On adsorption
on SiO2 TLC plate, the compound turned deep purple, but this color
faded after ca. 1 min. When ethanolic AgNO3 was added to the
compound a white precipitate was immediately formed, indicating the
presence of an easily released halide ion. This and the spectral
properties suggest that the compound is bis(2,4,6-triisopropylphenyl)-
chloromethane, and the compound was used as such without further
purification. 1H NMR (200 MHz, C6D6) δ: 1.07 (12H, d, J ) 6.7 Hz,
o-i-Pr-Me), 1.14 (12H, d, J ) 6.7 Hz, p-i-Pr-Me), 1.16 (12H, d, J )
6.9 Hz, o-i-Pr-Me), 2.70 (2H, m, J ) 6.9 Hz, p-i-Pr-CH), 3.76 (4H, m,
J ) 6.7 Hz, 2H, o-i-Pr-CH), 7.06 (4H, s, Tip-H), 7.26 (1H, s, Tip2-
CH). 13C NMR: (50.32 MHz, C6D6) δ: 24.73 (p-i-Pr-Me), 25.10 (o-
i-Pr-Me), 25.42 (o-i-Pr-Me), 30.21 (o-i-Pr-CH), 34.39 (p-i-Pr-CH),
59.40 (Tip2CH), 123.57 (m-Tip-C), 134.57 (ipso-Tip-C), 148.66 (p-
Tip-C), 149.32 (o-Tip-C).
V ) 1513.1(6) Å3, Z ) 4, Fcalc ) 1.01 g cm-3, µ(Cu KR) ) 3.92 cm-1
,
no. of unique reflections ) 5594, no. of reflections with I > 3σI )
3841, R ) 0.053, Rw ) 0.080.
The second fraction contained 15 mg (0.8%) of a white crystalline
solid identified as ditipylacetonitrile, 23. 1H NMR (200 MHz, CDCl3)
δ: 1.05 (12H, d, J ) 6.75 Hz, o-i-Pr-Me), 1.12 (12H, d, J ) 6.75 Hz,
o-i-Pr-Me), 1.20 (12H, d, J ) 6.9 Hz, p-i-Pr-Me), 2.88 (2H, m, J )
6.9 Hz, p-i-Pr-CH), 3.49 (4H, d, J ) 6.75 Hz, o-i-Pr-CH), 6.04 (1H, s,
Tip2CH), 7.00 (4H, s, TipH). 13C NMR (50.32 MHz, CDCl3) δ: 23.87,
23.99, 24.38 (o-i-Pr-Me + Tip2CH), 29.75, 33.94 (p-i-Pr-CH), 122.0
(CtN), 122.65 (m-Tip-C), 128.88 (ipso-Tip-C), 147.15 (o-Tip-C),
148.53 (o-Tip-C). IR νmax cm-1 (Nujol): 2240 (m, CN), 1620 (m,
CdC), 1560 (m, CdC). MS (CI, i-C4H10) m/z (relative abundance,
assignment): 446 (39%, MH+), 404 (B, MH+ - MeCHdCH2), 362
(41%, MH+ - 2MeCHdCH2), 320 (21%, MH+ - 3MeCHdCH2),
294 (3%, M - 3MeCHdCH2 - CN), 215 (7%, M - Tip - HCN).
Bis(2,4,6-triisopropylphenyl)methyl Methyl Ether (18). Because
of the difficulties encountered during the attempted purification of bis-
(2,4,6-triisopropylphenyl)chloromethane, 17, dry methanol was added
in situ to the benzene solution of 17 at room temperature. After filtering
the Na2CO3 and evaporation of the solvent a residual white solid
remained. Recrystallization from methanol afforded 5.5 g (95%) of
white crystals of bis(2,4,6-triisopropylphenyl)methyl methyl ether, mp
105-7 °C. 1H NMR (400 MHz, CDCl3) δ: 0.99 (12H, d, J ) 6.8 Hz,
o-i-Pr-Me), 1.06 (12H, d, J ) 6.8 Hz, o-i-Pr-Me), 1.21 (12H, d, J )
7.2 Hz, p-i-Pr-Me), 2.83 (2H, m, J ) 7.2 Hz, p-i-Pr-CH), 3.43 (3H, s,
OMe), 3.44 (4H, m, J ) 6.8 Hz, o-i-Pr-CH), 6.13 (1H, s, Tip2CH),
6.95 (4H, s, Tip-H). 13C NMR: (100.64 MHz, CDCl3) δ: 24.0 (p-i-
Pr-Me), 24.4 (o-i-Pr-Me), 24.5 (o-i-Pr-Me), 28.8 (o-i-Pr-CH), 33.9 (p-
i-Pr-CH), 57.6 (OMe), 81.1 (Tip2CH), 122.1 (m-Tip-C), 133.7 (ipso-
Tip-C), 147.3 (p-Tip-C), 147.6 (o-Tip-C). MS(EI, 70 eV, m/z, relative
abundance, assignment): 449 (7%, M - H), 419 (34%, M - i-Pr -
MeOH - Me), 407 (22%, M - i-Pr), 375 (B, M - i-Pr - MeOH),
360 (30%, M - i-Pr - MeOH - Me), 247 (26%, M -Tip), 246 (21%,
Bis(2,4,6-triisopropylphenyl)acetonitrile (23). (a) With CuCN in
DMF. A solution of ditipylchloromethane (2 g, 4.4 mmol) and CuCN
(0.5 g, 5.6 mmol) in dry DMF (30 mL) was refluxed for 12 h. After
cooling to room temperature, water (100 mL) was added, and the
solution was extracted with benzene (5 × 50 mL). The combined
organic phases were washed ten times with water and dried (MgSO4),
and the solvent was evaporated, leaving a yellow oil which showed no
nitrile band around 2200 cm-1. The 1H NMR spectrum of the mixture
showed a 1:3 ratio of (2,6-diisopropyl-4-isopropenyl)phenyltipyl-
methane 21 to (2,4-diisopropyl-6-isopropenyl)phenyltipylmethane 22.
The X-ray structure of 21 is given in Figure 1. The compounds were
isolated in other experiments, and their characterization will appear
elsewhere.
(b) With CuCN in Pyridine. A solution containing ditipylchlo-
romethane (1 g, 2.2 mmol) and CuCN (0.5 g, 5.6 mmol) in dry pyridine
(30 mL) was refluxed for 5 h. After 5 min the solution turned deep
green, and the color changed gradually to deep red. The solution was
cooled to room temperature and poured into a 10% HCl solution (50
mL), and concentrated HCl was then added until the pH was 2. After
extraction thrice with ether (50 mL) the organic phase was washed
with water (30 mL) and dried (MgSO4), and the solvent was evaporated.

Generation of a Carboxylic Acid Enol
J. Am. Chem. Soc., Vol. 118, No. 22, 1996 5181
M- TipH), 231 (27%, M - TipH - Me), 215 (21%, TipC). Anal.
Calcd for C32H50O: C, 85.27; H, 11.18. Found: C, 85.24; H, 10.96.
Bis(2,4,6-triisopropylphenyl)acetic Acid (19). To a solution of bis-
(2,4,6-triisopropylphenyl)methyl methyl ether (4.56 g, 9.83 mmol) in
dry ether (120 mL) in a three-necked, round-bottomed flask fitted with
a reflux condenser, an argon inlet, and a closed gas bubbler, 4:1
potassium-sodium alloy (ca. 9 g) was added, and the resulting
suspension was sonicated in a dry argon atmosphere. The solution
became orange-red after ca. 1 h, and it turned darker with the progress
of the reaction. After 6 h the sonication of the deep red solution was
stopped, and a stream of carbon dioxide (dried over P2O5) was
introduced into the solution through the gas bubbler until total
decoloration. The solution was cooled to room temperature, ethanol
was added to destroy the unreacted alloy, the solution was acidified to
pH 2 with 10% aqueous HCl, and the organic phase was separated,
washed twice with water (50 mL), and dried (MgSO4). The solvent
was removed in vacuo, leaving a white solid. Recrystallization from
high boiling petroleum ether yielded 3.46 g (73%) of white crystals of
bis(2,4,6-triisopropylphenyl)acetic acid, mp 212-3 °C. 1H NMR (400
MHz, CDCl3) δ: 0.94 (12H, d, J ) 6.6 Hz, o-i-Pr-Me), 1.02 (12H, d,
J ) 6.6 Hz, o-i-Pr-Me), 1.20 (12H, d, J ) 6.9 Hz, p-i-Pr-Me), 2.83
(2H, m, J ) 6.9 Hz, p-i-Pr-CH), 3.06 (4H, m, J ) 6.6 Hz, o-i-Pr-CH),
5.82 (1H, s, Tip2CH), 6.93 (4H, s,Tip-H). 13C NMR (100.64 MHz,
CDCl3) δ: 23.93 (o-i-Pr-Me), 24.08 (o-i-Pr-Me), 24.17 (p-i-Pr-Me),
29.95 (p-i-Pr-CH), 33.92 (o-i-Pr-CH), 48.28 (Tip2C), 122.34 (m-Tip-
C), 131.83 (ipso-Tip-C), 147.43 (o-Tip-C), 147.48 (p-Tip-C), 180.86
(CdO). FTIR: νmax cm-1 (Nujol): 2640 (br, COOH), 1695 (s, CdO),
1608 (m, CdC). MS(CI, i-C4H10) m/z (relative abundance, assign-
ment): 464 (13%, M), 463 (7%, M - 1), 421 (3%, M - i-Pr), 303
(6%), 299 (4%), 261 (B, M - Tip), 260 (9%, M - TipH), 243 (4%,
M - Tip - H2O), 219 (3%, M - Tip - MeCHdCH2), 217 (8%, M -
TipH - i-Pr), 205 (5%), 203 (7%, Tip). Anal. Calcd for C32H48O2:
C, 82.70; H, 10.41. Found: C, 82.47; H, 10.36.
Bis(2,4,6-triisopropylphenyl)ketene (20). A solution containing
bis-(2,4,6-triisopropylphenyl)acetic acid (1 g, 2.15 mmol), thionyl
chloride (0.5 g, 4.3 mmol), and pyridine (two drops) in dry benzene
(25 mL) was refluxed for 1 h in a flask protected from moisture by a
CaCl2 drying tube. The mixture was then decanted from the solid
pyridinium chloride, and removal of the solvent left a yellow oil which
was purified by chromatography on silica gel using low boiling
petroleum ether as eluent. The first fraction, a bright yellow solid (880
mg), was recrystallized from hot acetonitrile, affording bright yellow
crystals of bis(2,4,6-triisopropylphenyl)ketene (850 mg, 89%), mp 97
°C. 1H NMR (400 MHz, CDCl3) δ: 1.02 (24H, d, J ) 6.8 Hz, o-i-
Pr-Me), 1.22 (12H, d, J ) 6.9 Hz, p-i-Pr-Me), 2.85 (2H, m, J ) 6.9
Hz, p-i-Pr-CH), 3.23 (4H, m, J ) 6.8 Hz, o-i-Pr-CH), 6.96 (4H, s,-
Tip-H). 13C NMR (50.32 MHz, CDCl3) δ: 23.99 (p-i-Pr-Me), 24.16
(o-i-Pr-Me), 30.52 (o-i-Pr-CH), 33.49 (Tip2C), 34.08 (p-i-Pr-CH),
122.18 (m-Tip-C), 125.85 (ipso-Tip-C), 148.25 (p-Tip-C), 148.73 (o-
Tip-C), 191.51 (CdO). λmax nm (ꢀ) (CH2Cl2): 249 (19500), 290sh
(530). FTIR: νmax cm-1 (Nujol): 2096 (s, CdCdO). MS (CI, i-C4H10)
m/z (relative abundance, assignment): 447 (24%, M + 1), 446 (9%,
M), 243 (B, M - Tip), 153 (5%), 134 (2%). Anal. Calcd for
C32H46O: C, 86.03; H, 10.38. Found: C, 86.06; H, 10.60. Crystal
data: C32H46O, space group C2/c, a ) 19.390(2) Å, b ) 17.978(4) Å,
c ) 15.597(2) Å, â ) 107.10(1), V ) 5863.0(8) Å3, Z ) 8, Fcalc
)
1.01 g cm-3, µ(Cu KR) ) 4.12 cm-1, no. of unique reflections ) 4480,
no. of reflections with I g 3σI ) 3738, R ) 0.052, Rw ) 0.089.
Bis(2,4,6-triisopropylphenyl)ketene Bis(trimethylsilyl)acetal (28).
To a solution of bis(2,4,6-triisopropylphenyl)acetic acid (253 mg, 0.54
mmol) in THF (20 mL) was added n-butyllithium (1.6 M solution in
hexane, 0.7 mL, 1.12 mmol) under argon at room temperature. The
moisture-protected mixture was stirred for 3 h, chlorotrimethylsilane
(0.3 mL, 2.6 mmol) was then added, and the resulting deep red mixture
was kept overnight at room temperature. After evaporation of the
solvent the residual red oil was redissolved in warm methanol (5 mL),
and the white precipitate formed on cooling was filtered. Recrystal-
lization from warm methanol afforded white crystals of bis(2,4,6-
triisopropylphenyl)ketene bis(trimethylsilyl) acetal (95 mg, 29%), mp
88-9 °C. 1H NMR (400 MHz, CDCl3) δ: 0.00 (18H, s, SiMe3), 0.13
(6H, d, J ) 6.7 Hz, o-i-Pr-Me), 1.03 (6H, d, J ) 6.8 Hz, o-i-Pr-Me),
1.18 (12H, d, J ) 6.9 Hz, p-i-Pr-Me), 1.22 (6H, d, J ) 6.8 Hz, o-i-
Pr-Me), 1.33 (6H, d, J ) 6.5 Hz, o-i-Pr-Me), 2.79 (4H, m, J ) 6.6 Hz,
o-i-Pr-CH), 3.35 (2H, m, J ) 6.5 Hz, p-i-Pr-CH), 6.71 (2H, d, J ) 0.7
Hz, Tip-H), 6.94 (2H, d, J ) 0.7 Hz, Tip-H). 13C NMR (50.32 MHz,
CDCl3) δ: 0.4 (SiMe3), 23.1, 24.1, 24.1, 24.2, 24.6, 25.7 (6 i-Pr-Me),
30.0, 30.5, 33.9 (3 i-Pr-CH), 95.2 (Tip2C), 120.8, 122.1 (2 m-Tip-C),
135.2 (ipso-Tip-C), 146.5 (p-Tip-C), 147.6, 148.1 (2 o-Tip-C), 149.5
(dC(OSiMe3)2). FTIR: νmax cm-1 (Nujol):1623 (m, CdC). MS (CI,
i-C4H10, m/z, assignment): 609 (M + 1). Anal. Calcd for C38H64O2:
C, 74.93; H, 10.59. Found: C, 74.77; H, 10.55.
Acknowledgment. We are indebted to Dr. Roy E. Hoffman
for helping to interpret the SIMPLE NMR experiment, to Dr.
Shmuel Cohen for the determination of the X-ray structures, to
Prof. H. Iwamura and Dr. Daniella Onicu for the ESR spectrum
of 27, to Professors S. E. Biali and Y. Marcus for discussions,
and to Dr. H. Hartig from Firma Huels AG for a generous gift
of triisopropylbenzene.
Supporting Information Available: Crystallographic sum-
mary for 20, 21, and 24 including tables of crystal data, bond
lengths and angles, and positional and thermal parameters and
stereoviews for 21 and 24 (28 pages). Ordering information is
given on any current masthead page.
JA9601090