Stable simple enols. Self-catalyzed E/Z-isomerization of the sterically crowded 2-(m-methoxymesityl)-1,2-dimesitylethenol

Article abstract of DOI:10.1021/jo020665e

(E)-2-(m-Methoxymesityl)-1,2-dimesitylethenol (3a) isomerizes in the absence of a catalyst in solution to a 1.0:0.9 E/Z (3a/3b) equilibrium mixture. In CDCl₃, the isomerization is first order in 3a within a run, but the plot of the rate constan

Full text of DOI:10.1021/jo020665e

Sta ble Sim p le En ols. Self-Ca ta lyzed E/Z-Isom er iza tion of th e
Ster ica lly Cr ow d ed 2-(m -Meth oxym esityl)-1,2-d im esityleth en ol1
Elimelech Rochlin and Zvi Rappoport*
Department of Organic Chemistry, The Hebrew University, J erusalem 91904, Israel
zr@vms.huji.ac.il
Received October 22, 2002
(E)-2-(m-Methoxymesityl)-1,2-dimesitylethenol (3a ) isomerizes in the absence of a catalyst in solution
to a 1.0:0.9 E/Z (3a /3b) equilibrium mixture. In CDCl3, the isomerization is first order in 3a within
a run, but the plot of the rate constant kobs vs the changing [3a ]0 in different runs is a half-parabola,
indicating self-catalysis by more than one enol molecule. At 0.09 M enol, the isotope effect k3a
/
k3a -OD ) 2.1. In the presence of 0.025-0.25 M pyridine-d5, the kobs vs [pyridine-d5] plot displays a
bell-shaped profile. The change in the shape of the OH signals of the 3a /3b mixture at 295-430 K
in C6D5NO2 was followed by DNMR. The four signals of the diastereomeric 3a /3b mixture observed
at 330 K coalesce at 350 K with barriers of 18.3 and 18.4 kcal mol-1 due to the diastereomerization
of the vinyl propellers. The resulting two signals observed at >360 K further coalesce at 425 K
with a barrier of 22.9 kcal mol-1 due either to oxygen-to-oxygen proton exchange or to E/Z
isomerization. The estimated upper limit for the rate of proton exchange of kex e (2-4) × 103 M-1
s-1 at 425 K between the enol molecules is sufficiently slow to be a rate-controlling step in the
isomerization. A process in which several enol molecules catalyze the isomerization is suggested,
and several mechanistic routes are analyzed.
In tr od u ction
double bond is twisted to a low degree (5-8°).4 (iv) When
X ) OH, the enols show high kinetic stability to keton-
ization.7 The uncatalyzed rotational barrier around the
double bond in these compounds should be lower than
that in alkenes due to the contribution of the hybrid form
1a (eq 1) with the expected longer than normal C1dC2
bond. However, the calculated CdC double bond lengths
An uncatalyzed internal rotation of the double bond
in sterically crowded alkenes which leads to E h Z
isomerization is well-known2 and usually takes place
when the double bond is partially twisted in the ground
state. Trimesitylvinyl-X systems 1 (Mes ) mesityl) where
X is resonatively electron donating, e.g., OR, OAc, OH,
are sterically crowded as reflected by different phenom-
ena. (i) They have a propeller conformation in the solid
state and in solution3 with high (>50°) Mes-CdC
torsional angles.4 (ii) They display hindered and cor-
related rotation around the dC-Ar bonds in solution
with activation barriers of 15-27 kcal mol-1 3,5,6
(iii) Their
.
(1) Presented in part in the 59th Meeting of the Israel Chemical
Society, Beer Sheva, Israel, J an 31-Feb 1, 1994, Abstract, p 183, and
at the 13th IUPAC Conference on Physical Organic Chemistry, Seoul,
Korea, Aug 25-29, 1996 (Rappoport, Z.; Frey, J .; Sigalov, M.; Rochlin,
E. Pure Appl. Chem. 1997, 69, 1933).
(2) (a) Eliel, E. L.; Wilen, S. H.; Mander, L. N. Stereochemistry of
Organic Compounds; Wiley: New York, 1994; see p 22, Chapter 9, and
references therein. (b) For reviews on crowded and push-pull ethyl-
enes, see: Sandstro¨m, J . Top. Stereochem. 1983, 14, 83. Sandstro¨m,
J . In The Chemistry of Double Bonded Functional Groups, Supplement
A3; Patai, S., Ed.; Wiley: Chichester, 1997; Chapter 21, p 1253. (c)
Biedermann, P. U.; Stezowski, J . J .; Agranat, I. In Advances in
Theoretically Interesting Molecules; Thummond, R. P., Ed.; J AI Press:
Stamford, CT, 1998; Vol. 4, p 245. (d) For a recent work, see: Tapuhi,
Y.; Suissa, M. R.; Cohen, S.; Biedermann, P. U.; Levy, A.; Agranat, I.
J . Chem. Soc., Perkin Trans. 2 2000, 93. (e) Feringa, B. L.; J ager, W.
F.; de Lange, B. Tetrahedron Lett. 1992, 33, 2887.
of vinyl alcohol and of ethylene are almost identical,8 and
(6) (a) Rochlin, E.; Rappoport, Z. J . Org. Chem. 1994, 59, 3857. (b)
Rochlin, E.; Rappoport, Z.; Kastner, F.; Pustet, N.; Mannschreck, A.
J . Org. Chem. 1999, 64, 8840. (c) Rochlin, E.; Rappoport, Z. Unpub-
lished results.
(3) For a review, see: Rappoport, Z.; Biali, S. E. Acc. Chem. Res.
1997, 30, 871.
(4) Kaftory, M.; Biali, S. E.; Rappoport, Z. J . Am. Chem. Soc. 1985,
107, 1701.
(5) Hart, H.; Rappoport, Z.; Biali, S. E. In The Chemistry of Enols;
Rappoport, Z., Ed.; Wiley: Chichester, 1990; Chapter 8, pp 481-590.
(7) Fuson, R. C.; Armstrong, L. J .; Kneisley, J . W.; Shenk, W. J .,
J r. J . Am. Chem. Soc. 1944, 66, 1464.
(8) Apeloig, Y. In The Chemistry of Enols; Rappoport, Z., Ed.;
Wiley: Chichester, 1990; Chapter 1, pp 1-74.
10.1021/jo020665e CCC: $25.00 © 2003 American Chemical Society
Published on Web 02/06/2003
J . Org. Chem. 2003, 68, 1715-1720
1715

Rochlin and Rappoport
even when the ethylene is substituted by bulky mesityl
groups this is not appreciably reflected in bond lengths.
The X-ray determined double bond lengths of systems 1,
X ) OH, OPr-i, are only slightly (by ca. 0.01 Å) longer
and those in the corresponding 1, X ) OAc are even
shorter than that in ethylene.4 Furthermore, there is no
spontaneous double-bond rotation when X * OH. For
example, (E)- and (Z)-2-m-methoxymesityl-1,2-dimesi-
tylvinyl acetates and isopropyl ethers do not undergo a
thermal E/Z-isomerization even when heated in solution
up to 150 °C.6 Likewise, an exchange between the â- and
â′-ring protons in the NMR spectra of most trimesitylvi-
nyl systems with identical â- and â′-rings was not
observed. However, triarylethenols isomerize relatively
rapidly. For example, only (Z)-1,2-dimesityl-2-phenyle-
thenol is known,9 and reactions which should initially
give the E-isomer, such as hydrolysis of the E-acetate,
lead only to the Z-enol, indicating a rapid E f Z enol
isomerization.9 Whereas both isomers of simple enols
with different â-substituents such as 1-propenol 2a 10 or
2-phenylethenol 2b11 can exist as an E/Z mixture and
their ratio was evaluated for 2b,11 they were not sepa-
rated or isolated and nothing is known about their
interconversions. Isomer Z-2c is more stable than E-2c
but on heating for 5 h in DMSO at 80 °C it isomerizes to
an E/Z mixture richer in Z-2c.12
2,2-dimesitylethenol17 as compared with its acetate and
other crowded 1, X * OH. It includes the formation of
the intermediate 1,2,2-triarylethylcarbenium ion gener-
ated by protonation at C1 of the enolic double bond. Such
an intermediate should also permit rotation around the
C1-C2 bond and hence enable an E/Z isomerization in
an enol with different â-aryl rings which will display
similar characteristics to those of the racemization
process. The barrier given above for the saturation
transfer in 1 leads to a t1/2 value which should enable
convenient kinetic measurements. A prerequisite for such
a study is a pair of E/Z enol isomers of not too different
energies but of sufficiently different and measurable
spectral properties. In the present work, we had studied
the E/Z isomerization in the E/Z pair of 2-(m-meth-
oxymesityl)-1,2-dimesitylethenols 3a /3b (eq 2) by a com-
bination of conventional kinetic and DNMR methods in
CDCl3 and C6D5NO2.
Resu lts a n d Discu ssion
Syn th esis. A 3a /3b mixture was synthesized by the
reaction of MesMgBr with (m-methoxymesityl) mesityl
ketene. The synthesis of the ketene was described
previously.6a The pure E-isomer 3a was isolated from the
3a /3b equilibrium mixture by slow crystallization from
petroleum ether, and its structure was established by
X-ray diffraction.6a Unfortunately, we failed to isolate the
pure Z-isomer 3b by crystallization from MeOH, which
only gave pure 3a . An attempted LC separation failed
due to the fast isomerization on the silica gel column.
Fortunately, for routine kinetic measurements either a
single isomer or a nonequilibrium mixture of 3a /3b is
sufficient.
The first observation pointing to E/Z-isomerization of
the enols was a saturation transfer observation in a pair
of â- and â′-p-Me groups in the NMR spectrum of 1,2,2-
trimesitylethenol (1, X ) OH).13 A lower limit of ∆Gq >
20.9 kcal mol-1 for the exchange barrier of this topomer-
ization was estimated. The MM2-calculated enthalpy of
activation for the double-bond rotation in 1 (X ) OH) is
6c
high (∆Hq ) 51.7 kcal mol-1
)
and is reminiscent of
rot
the observed ∆Gq > 48.7 kcal mol-1 for rotation around
the double bond of a tetramesitylethylene analogue14 and
of the observed values of 36-40 kcal mol-1 for similar
rotations in other tetraarylethylenes.15 These values are
well above the time scale which enables measurement
by NMR. Hence, the observed exchange in trimesityle-
thenol, as well as the rapid isomerization of an isolated
one geometric isomer of 1,2-dimesityl-2-[4-tert-butyl-2,6-
dimethylphenyl]ethenol16 to the E/Z mixture suggest that
an efficient catalytic isomerization process is operating.
The mechanism of E/Z isomerization of ethenols has
not been investigated. We recently reported an unusual
racemization mechanism of chiral 1-[9′-(2′-fluoroanthryl)]-
NMR Sp ectr a of 3a a n d 3a /3b. Solid 3a is stable
under ordinary conditions, while in CCl4, CHCl3, CH2-
Cl2, C6H6, C6H5NO2, AcOEt, Me2CO, DMSO, and MeOH
solutions it spontaneously isomerizes within 1-2 h at
room temperature to the 3a /3b equilibrium mixture and
no byproducts were detected. On a TLC plate, the
isomerization is complete within 10-20 min. We use the
term “spontaneous” in this work to describe an isomer-
ization in solution which occurs sufficiently rapidly
without an added catalyst. Unfortunately, the almost
identical UV spectra of 3a and the 3a /3b mixtures
excluded the use of the sensitive UV method for monitor-
ing the isomerization. HPLC analysis was also precluded
due to a catalyzed isomerization on the column. In
1
contrast, the H NMR spectra of 3a and 3b are different.
(9) (a) Biali, S. E.; Lifshitz, C.; Rappoport, Z.; Karni, M.; Mandel-
baum, A. J . Am. Chem. Soc. 1981, 103, 2896. (b) Yamataka, H.;
Aleksiuk, O.; Biali, S. E.; Rappoport, Z. J . Am. Chem. Soc. 1996, 118,
12580.
(10) (a) Chiang, Y.; Chawang, W. K.; Kresge, A. J .; Ying, Y. J . Am.
Chem. Soc. 1989, 111, 7185. (b) Turecˇek, F.; Havlas, Z. J . Org. Chem.
1986, 51, 4066. (c) Turecˇek, F.; Bradec, L.; Korvala, J . J . Am. Chem.
Soc. 1988, 110, 7984.
(11) Chiang, Y.; Kresge, A. J .; Walsh, P. A.; Ying, Y. J . Chem. Soc.,
Chem. Commun. 1989, 869.
(12) Nadler, E. B.; Ro¨ck, M.; Schmittel, M.; Rappoport, Z. J . Phys.
Org. Chem. 1993, 6, 233.
Due to the helicity of the propeller structure and since
the m-MeO group can lie either above (“up”) or below
(“down”) the plane of the double bond either 3a or 3b
should exist as two pairs of diastereomers. The four
stereoisomers and their inter-relations appear in our
earlier papers.3,6,17 At slow exchange and no accidental
signal overlap, each of them will display a maximum of
10 Ar-H, 2 OH, 2 OMe, and 18 Ar-Me signals.
1
Due to signal overlap, the observed H NMR spectrum
(13) Biali, S. E.; Rappoport, Z.; Hull, W. E. J . Am. Chem. Soc. 1985,
107, 5450.
of 3a at 200 MHz (Figure 1a) in CDCl3 at 295 K displays
fewer signals: 6 broaden Ar-H singlets, 2 OH, 2 OMe,
and 15 Ar-Me signals, which is still consistent with the
propeller structure. The 3a /3b equilibrium mixture,
(14) Maeda, K.; Okamoto, Y.; Morlender, N.; Haddad, N.; Eventova,
I.; Biali, S. E.; Rappoport, Z. J . Am. Chem. Soc. 1995, 117, 9686.
(15) (a) Kistiakowsky, G. B.; Smith, W. R. J . Am. Chem. Soc. 1934,
56, 538. (b) Leigh, W. J .; Arnold, D. R. Can. J . Chem. 1981, 59, 609.
(c) Leigh, W. J .; Frendo, D. T.; Klawunn, P. J . Can. J . Chem. 1985,
63, 2131.
(16) Biali, S. E.; Rappoport, Z. J . Am. Chem. Soc. 1984, 106, 477.
(17) Rochlin, E.; Rappoport, Z. J . Org. Chem., in press.
1716 J . Org. Chem., Vol. 68, No. 5, 2003

Stable Simple Enols
TABLE 1. Effect of En ol Con cen tr a tion [3a ]0 a n d
P yr id in e Con cen tr a tion Cp yr on th e P seu d o-F ir st-Or d er
Ra te Con sta n t kobs for th e E/Z-Isom er iza tion of 3a a t 295
K
a
[3a ]0, M
104 kobs, s-1
C
pyr, M
104kobs, s-1
0.028
0.049
0.078
0.09
0.104
0.114
0.67
0.83
1.8
2.8
5.0
0
2.8
7.0
0.025
0.050
0.080
0.160
0.175
0.200
0.225
0.25
14.3
32.3
31.7
25.0
15.0
9.9
6.5
F IGURE 1. 1H NMR spectra of (a) 3a immediately after
dissolution in CDCl3; (b) 3a /3b equilibrium mixture.
5.2
a
[3a ]0 ) 0.09 M.
which should display at most twice the number of signals
of a single isomer, shows under the same conditions a
smaller number [6 Ar-H, 3 OH (or 4 OH in the presence
of 1 drop of DMSO-d6), 4 OMe, and 19 Ar-Me signals]
(Figure 1b). Due to the overlap, not all the regions of the
NMR spectrum are convenient for monitoring the isomer-
ization. However, the remoteness of the four MeO signals
of 3a /3b mixture from the other signals and their spread
[δ 3.29 (major 3a ), 3.52 (minor 3a ), 3.59 (minor 3b), 3.75
(major 3b)] enables 1H NMR monitoring of the E/Z-
isomerization at 295 K in CDCl3 by following the decrease
and increase of the intensities of the MeO signals of 3a
and 3b, respectively.
The need for large quantities of the expensive deu-
teriated solvent for the kinetic investigation limited our
study only to the CDCl3 solvent. The use of D2O or of
D2O-organic media was also precluded due to insolubility
of 3a /3b in these media.
F IGURE 2. Plot of log kobs against log [3a ]0.
Kin etics. Solutions of 3a (0.09 M) in CDCl3 were first
used in the kinetic study, and the decrease in the molar
fraction of 3a with time was followed by using the
integrated sum of the MeO signals as an internal
standard. In each kinetic run, a reproducible plot for a
reversible first-order E h Z reaction was obtained by
using eq 3, where the P’s are the molar fractions of the
isomers (P3a + P3b ) 1). An observed first-order rate
constant kobs ) k1 + k-1 ) (2.8 ( 0.4) × 10-4 s-1 was
calculated from the slope of the ln[[P3a (K + 1) - 1]/K] vs
t plot. A K ) k1/k-1 value of 0.9 at 295 K was determined
thrice at the end of the reaction. By inserting these values
into eq 4 where the subscript “eq” indicate the concentra-
tions at equilibrium, we obtained k1 ) (1.3 ( 0.2) × 10-4
s-1 and k-1 ) (1.5 ( 0.2) × 10-4 s-1. The k1 value gives
an activation barrier of ∆Gq(295 K) ) 22.4 ( 0.2 kcal
mol-1 for the E/Z isomerization.
Supporting Information) with a catalytic order larger
than one in the enol, although the linearity according to
eq 1 is still retained in each run. The reaction does not
follow a simple rate law and as seen from Figure 2 the
order in the enol is not constant. At low [3a ]0, the overall
order in the enol is close to 1, whereas at higher [3a ]0 it
is close to 4 indicating a multiple term or complex rate
low.18 The curved dependence of kobs on the total enol
concentration means a reaction order higher than two
in the enol and that several enol molecules are present
in the transition state of the rate controlling step(s),
possibly in an associated form. Unfortunately, we know
only little about the association of such crowded enols in
chloroform and only at low temperature.19 The pseudo-
first-order reaction within a run could be understood if
3a and 3b have similar catalytic activities so that kobs is
sensitive only to the constant total enol concentration
within a kinetic run, and not to the nature of the specific
enol. Extrapolation of the kobs vs [3a ]0 plot to zero
concentration gives a residual kobs value of ca. 5 × 10-5
s-1, which can be tentatively ascribed either to a mono-
molecular self-catalyzed reaction or to catalysis by traces
of water or impurities in the CDCl3. The higher extent
of saturation transfer (vide supra) in a sample that stood
for several days in solution than for a freshly dissolved
ln{[P3a (K + 1) - 1]/K} ) -kobs
t
(3)
kobs ) k1 + k-1; K ) k1/k-1 ) [3b]eq/[3a ]eq (4)
To obtain more mechanistic details, we studied briefly
the effects of the total enol concentration, of the addition
of a base, and of the deuteriation of the enolic OH group
on the observed first-order rate constant.
On changing the total enol concentration [3a ]0 from
0.02 to 0.114 M, a range convenient for the NMR
measurements, the first-order rate constant kobs obtained
from eq 3 increases by ca. 10.5-fold (Table 1), suggesting
that catalysis by the enol takes place. The kobs vs [3a ]0
plot is a half-parabola-like curve (Figure S1 in the
(18) Bunnett, J . F. In Investigation of Rates and Mechanisms of
Reactions, 4th ed.; Bernasconi, C. F., Ed.; Wiley: Chichester, 1986;
Chapter 3, pp 177, 189-190.
(19) . Eventova, I.; Nadler, E. B.; Frey, J .; Rappoport, Z. J . Phys.
Org. Chem. 1994, 7, 28.
J . Org. Chem, Vol. 68, No. 5, 2003 1717

Rochlin and Rappoport
F IGURE 3. Plot of kobs against the pyridine-d5 concentration
Cpyr
.
sample13 may reflect the development of catalytic impuri-
ties on standing.
The isomerization of 3a -OD was studied under the
same conditions (295 K, [3a ]0 ) 0.09 M), and the kinetic
isotope effect (KIE) on kobs is kH/kD ) 2.1. Consequently,
a proton transfer is definitely involved in the rate-
determining step(s) of the isomerization. From the kobs
vs [3a ]0 plot, the proton transfer in the absence of added
acids or bases probably takes place in a transition state
containing several enol molecules. A third-order term was
reported for several reactions subject to general acid
(base) catalysis, such as enolization of ketones20 and
mutarotation of tetramethylglucose.21 It was ascribed to
a concerted mechanism in which two catalyst moleculess
one a proton donor and one a proton acceptorsare
present in the transition state. Bell suggested that in low
dielectric constant solvents such mechanism might be
preferable over the stepwise route.21 Self-catalysis by the
substrate, resulting in a second order in the substrate,
was recently observed for the enolization of malonate
anion.22
1
F IGURE 4. Temperature-dependent H NMR spectrum of the
OH region of 3a /3b equilibrium mixture in C6D5NO2.
Since temperatures above the boiling point of CDCl3 are
required in order to follow the exchange of the OH
protons, the exchange was studied in C6D5NO2 at [3a ]0
) 0.03 M. The temperature dependent changes of the OH
signals at 295-430 K are shown in Figure 4. Whereas
at 295 K only three OH signals are observed at 5.54,
5.515, and 5.49 ppm, due to their slight shift upfield on
raising the temperature all four singlets belonging to the
two diastereomeric pairs of 3a /3b are observed at 330
K. Broadening takes place on further temperature in-
crease, and each pair coalesces at 350 K, giving barriers
of 18.4 ( 0.2 and 18.3 ( 0.2 kcal mol-1. These values
resemble the barriers of 18.8 and 18.3 kcal mol-1 for the
3-ring flip diastereomerization process obtained earlier
for 3a /3b from the coalescence of the MeO signals6a and
identify the observed processes as the diastereomeriza-
tions of 3a and 3b. Two signals, one belonging to 3a and
one to 3b, then start to appear at 360 K, and they remain
relatively sharp and separate with ∆ν ) 5.8 Hz at the
380-400 K region (Figure 4). They then coalesce at 425
We were unable to quantitatively study the acid/base
catalysis in the E/Z-isomerization since our enols are
insoluble in D2O or aqueous-organic media, and hence,
the use of buffer solutions was precluded. Consequently,
we studied only the effect of 0.025-0.25 M pyridine-d5
on the isomerization rate in CDCl3. Due to deviations
from the first order plots at several pyridine concentra-
tions we used only the intercepts of the plots in the first
50-70% reaction in our qualitative analysis. A plot of
K, giving a barrier of ∆Gc ) 22.9 ( 0.1 kcal mol-1 which
q
is surprisingly close to the isomerization barrier derived
above from kobs at 295 K in CDCl3 from the kinetic study.
Consequently, the barrier for proton exchange between
3a and 3b is g22.9 kcal mol-1 regardless whether the
coalescence is due to a chemical exchange or to an E/Z
isomerization process. This barrier corresponds to a first-
order rate constant of kc e 13 s-1 at Tc ) 425 K. When
equilibrium enol concentrations [3a ] ≈ [3b] ≈ [3a ]0/2 of
ca. 0.01-0.02 M are used to estimate a second-order
proton exchange rate constant (kex) for 3a /3b, a value of
kex e 6.5-13 × 102 M-1 s-1 is obtained. The effective
proton exchange rate constant in 3a /3b solution will be
ca. 3 times larger, i.e., kex e (2-4) × 103 M-1 s-1, since
3a /3a and 3b/3b exchanges also occur but they are
unobservable by NMR. This value is unusually low for
k
obs vs [pyridine-d5] displays a bell-shaped profile (Figure
3): At the lower [pyridine-d5] of 0-0.1 M, kobs increases
and reaches a maximum at ca. 0.1-0.12 M and it then
decreases almost to its initial “uncatalyzed” value at ca.
0.25 M pyridine.
DNMR Stu d y. The proton exchange between two enol
molecules could be monitored by a dynamic NMR tech-
nique using as probes the four well-resolved OH singlets
in the NMR spectrum of the 3a /3b equilibrium mixture.
(20) Keeffe J . R.; Kresge, A. J . In The Chemistry of Enols; Rappoport,
Z., Ed.; Wiley: Chichester, 1990; Chapter 7, pp 434-446.
(21) Bell, R. P. The Proton in Chemistry, 2nd ed.; Cornell University
Press: Ithaca, NY, 1973; Chapter 8.
(22) Eberlin, A. R.; Williams, D. L. H. J . Chem. Soc., Perkin Trans.
2 1996, 883.
1718 J . Org. Chem., Vol. 68, No. 5, 2003

Stable Simple Enols
SCHEME 1
exchange of OH protons at such high temperature, and
it resembles exchange rate constants for CH-acids.23 In
contrast, the isotope exchange between the OH group of
3a /3b and D2O in CDCl3 is fast and is complete before
the NMR spectrum is immediately taken. Consequently,
the slow proton exchange is ascribed to a steric inhibition
to mutual approach of two enol molecules, especially in
the low dielectric aprotic medium. A similar steric
hindrance for hydrogen bonding of 2,2-24 and 1,2-diaryle-
thenols25 was previously suggested.
F IGURE 5. 1H NMR spectrum of Na-3a /3b enolates in (CD3)2-
SO at 444 K.
energy 2-ring flip processes.6a,b The important feature for
our study is that two separate MeO signals and six
separate p-Me signals are observed above 380 K. The
p-Me signals are still sharp above 420 K, and no
broadening occurs up to 444 K (Figure 5). Other signals,
i.e., o-Me and Mes-H, remain broad. Furthermore, satu-
ration transfer was neither observed between any of the
six p-Me nor between the two MeO signals, indicating
that E/Z-isomerization of the enolate is very slow on
NMR time-scale even at 444 K. This conclusion was
strengthened by monitoring the MeO signals of Na-3a
enolate prepared by rapid dissolution of 3a in (CD3)2SO
containing a nearly stoichiometric amount of sodium
hydride. At 295 K, the Na-3b MeO signals were not
observed even after 48 h. Assuming that the CdC bond
rotation in the enolate is not strongly solvent dependent,
this observation exclude intermediacy of the enolate in
the spontaneous E/Z-isomerization of the parent enol.
P ossible Mech a n ism s. The experimental data ex-
cludes the possibility of a spontaneous E/Z-isomerization
in 3a by a CdC bond rotation in the neutral enol
molecule. Five routes involving proton transfer that may
lead to the isomerization of 3a are known from the
literature and were discussed previously16,17 (Scheme
1): (i) ionization to the enolate ion f rotation f
neutralization; (ii) protonation of the OH group f ioniza-
tion to a linear vinyl cation f capture by water from both
CdC Bon d Rota tion in th e En ola te An ion . Force
field calculation showed that the CdC bond rotation in
the enolate anion of 1 (X ) OH) has an activation
enthalpy of ∆Hq ) 39.8 kcal mol-1. Although the value
rot
is 12 kcal mol-1 lower than ∆Hq ) 51.7 kcal mol-1 in 1
rot
(X ) OH),6c it is still too high to account for the
spontaneous E/Z isomerization. Consequently, we studied
briefly the dynamic behavior of Na-3a /3b enolates in
(CD3)2SO solution. The enolates were obtained by treat-
ing the 3a /3b equilibrium mixture with sodium hydride
in (CD3)2SO.
The dynamic behavior of the Na-3a /3b enolates in
(CD3)2SO at 295-380 K qualitatively resembles that of
the 3a /3b equilibrium mixture in C6D5NO2.7a At 295 K
and 400 MHz, the spectrum is consistent with the
propeller structure; i.e., both geometric isomers exist as
pairs of helicity diastereomers displaying 9 Ar-H, 4 MeO,
and 24 Mes-Me signals out of the possible maximum of
20, 4, and 36 signals, respectively. On raising the
temperature, many signals broaden and several coales-
cence processes take place above 320 K due first to the
threshold 3-ring flip process16 and then probably to higher
(23) Reference 21, Chapter 7.
(24) Rappoport, Z.; Nugiel, D. A.; Biali, S. E. J . Org. Chem. 1988,
53, 4814.
(25) Nadler, E. B.; Rappoport, Z. J . Am. Chem. Soc. 1989, 111, 213.
J . Org. Chem, Vol. 68, No. 5, 2003 1719

Rochlin and Rappoport
sides of the empty p-orbital f deprotonation; (iii) keton-
ization f rotation f re-enolization; (iv) protonation at
C2; or (v) at C1 of the double bond f rotation f
deprotonation.
Route (i) is unlikely since the isomerization of the Na-
3a enolate in the (CD3)2SO was much slower than that
of the neutral enol. It is also inconsistent with the bell
shape profile of the pyridine catalysis in CDCl3 since for
an isomerization involving an enol h enolate preequi-
librium the rate must increase on increasing the [pyri-
dine-d5].
Route (ii) is also excluded. First, vinyl cation formation
by initial protonation of the enolic OH group have no
precedent26 and it is highly likely that formation of the
ion will be slower than the observed isomerization.
Second, its rate should decrease consistently on increas-
ing the base concentration, in contrast with observation.
Concerning route (iii), we note that the ketone (m-
MeOMes)MesCHCOMes 4 was not detected in our reac-
tion mixtures by NMR or TLC neither at room temper-
ature in CDCl3 nor at 430 K in C6D5NO2. In the closely
related 1,2,2-trimesitylethenol/1,2,2-trimesitylethanone
system the equilibrium percentage of the ketone is 1.2%
in hexane at 353.6 K,27 and by analogy, we expect
formation of detectable concentration of 4 during the
isomerization. Although 4 may still be an “unstable
intermediate” in the E/Z-isomerization with lower than
equilibrium content in the mixture, the half-life for the
TFA-catalyzed ketonization of 1 in hexane at 353.6K is
several hours,27 much longer than the half-life for the
spontaneous isomerization of the structurally similar 3
at room temperature.
DNMR experiment shows that at 425 K the oxygen-to-
oxygen and hence more so the oxygen-to-carbon proton
transfer between the enol molecules is sufficiently slow
to be the rate controlling step. We regard our explanation
as tentative which requires further experimental support.
Con clu sion s. The main outcome of this study is that
E/Z-isomerization in sterically crowded 2-(m-methoxyme-
sityl)-1,2-dimesitylethenol 3 in CDCl3 does not proceed
via a CdC bond rotation in the neutral enol but includes
a proton-transfer step(s) as deduced from the kH/kD value
of 2.1 and the bell-shaped catalysis by pyridine-d5. In the
absence of added catalysts the reaction is self-catalyzed
by the enol as deduced from the curved kobs vs [3a ]0
dependence, which suggests higher than second order
terms in the enol. The apparently conflicting pseudo-first-
order dependence on [3a ] within a run is ascribed to a
similar catalytic activities of 3a and 3b. A DNMR study
involving the OH signals of 3a /3b mixture in C6D5NO2
gave an upper limit estimate at 425 K of kex e (2-4) ×
103 M-1 s-1 for the rate constant of oxygen-to-oxygen
proton exchange between the enol molecules. This un-
usually low value for an oxygen acid is ascribed to a steric
inhibition to a mutual approach of two enol molecules.
The enolate anion is excluded as an isomerization
intermediate since no saturation transfer was observed
between the six p-Me signals of Na-3a /3b enolates in
(CD3)2SO at 444 K and isomerization in a fresh solution
of Na-3a enolate was not observed. The most likely (but
still tentative) route for the isomerization is an internal
rotation in the C1-protonated carbenium ion intermediate
formed from the enol.
Exp er im en ta l Section
Route (iv) involves protonation at C2, which should be
rate determining for the ketonization and hence is also
excluded. Although C2 protonation may be electronically
favored over C1 protonation, the approach of a proton to
the sterically crowded C2 is slow, as deduced from the
slow protonation of a structurally related enol28 and the
protonation rate is likely to be slower than the isomer-
ization rate.
(v) A protonation at the sterically more accessible C1
by a second enol molecule, which generates a stabilized
tertiary diarylmethyl carbenium ion (route v) and is
slower than the subsequent C1-C2 bond rotation in the
carbenium ion is consistent with both the self-catalysis
and with the KIE. The higher order than 2 in the enol is
ascribed either to a contribution of a concerted route in
which one enol protonates and another deprotonates the
isomerizing enol molecule or as due to a reaction of a
cluster of several enol molecules preceding the protona-
tion-deprotonation step(s). The bell shape catalysis by
pyridine could also be consistent with this route. At low
[pyridine-d5] a pyridine molecule may catalyze the reac-
tion by serving as a deprotonating species in a concerted
transition state. At higher [pyridine-d5], deprotonation
of the enol species reduces the rate by lowering the
concentration of the protonating species. The relatively
low KIE is attributed to a composite rate constant, where
the proton transfer step is not fully rate controlling. The
Ma ter ia ls. 3a and mixtures of 3a /3b were prepared as
described previously.6a The 3a -OD was prepared as follows:
50 mg of 3a was dissolved in 10 mL of CHCl3, and after the
solution was shaken for 1-2 min with 1 mL of D2O, the phases
were quickly separated, the organic phase was filtered through
a sintered glass with ca. 5 mm layer of anhydrous oven-
activated MgSO4 and rapidly evaporated in vacuo, dried and
1
kept in a desiccator. Its H NMR spectrum was identical with
that of 3a except for the absence of OH signals. Since solutions
of Na-3a /3b enolates in (CD3)2SO are sensitive to oxygen, the
oxygen was removed by a stream of dry nitrogen and the
studies were conducted under argon. The deuteriated solvents
were the best available commercial grades and were used
without further purification. Commercial CDCl3 was stored
over Mg turnings before use.
Kin etics. In a typical run an 0.02-0.114 M solution of the
enol in 5.0 mL of CDCl3 was prepared and 0.5 mL aliquots
were transferred into the NMR tube and thermostated in the
probe of the spectrometer at 295 ( 0.2 K for 5 min. Acquisi-
tions of 16 pulses were done automatically at a predetermined
times (1, 5, 10, 15, 20, 30, 45, 60, 75, and 90 min). The
integrated sum of the MeO signals served as an internal
standard. The equilibrium constant K ) [3b]/[3a ] for the first-
order reversible isomerization reaction was measured after
several hours and several days for the same sample and the
values were identical within the experimental error (K ) 0.9
( 0.05). In the pyridine-catalyzed kinetics, 1-40 µL of pyri-
dine-d5 was added to 0.09 M solution of 3a in CDCl3.
The DNMR experiments were conducted similarly to those
described previously.6a,b
Su p p or tin g In for m a tion Ava ila ble: Figure S1: kobs vs
the enol concentration [3a ]. This material is available free of
charge via the Internet at http://pubs.acs.org.
(26) Dicoordinated Carbocations; Rappoport, Z., Stang, P. J ., Eds.;
Wiley: Chichester, 1997.
(27) Biali, S. E.; Rappoport, Z. J . Am. Chem. Soc. 1985, 107, 1007.
(28) Nadler, E. B.; Rappoport, Z. J . Org. Chem. 1990, 55, 2673.
J O020665E
1720 J . Org. Chem., Vol. 68, No. 5, 2003