Mechanistic investigations of the acid-catalyzed cyclization of a vinyl ortho-quinone methide

Article abstract of DOI:10.1002/chem.200800662

A study was conducted to perform mechanistic investigations of the acid-catalyzed cyclization of a vinyl ortho-Quinone methide (QM). The study focused on mechanistic investigations of the acid-catalyzed cyclization of a vinyl ortho-Quinone methide, as led to the formation of chromenes. Vinyl o-QM was prepared through oxidation of phenol 3, while the E configuration of the exo-alkylidene bond was established by detailed nuclear magnetic resonance (NMR) studies and X-ray crystallography. The structure of the dimer was also confirmed by X-ray crystallography. Kinetic investigation of the uncatalyzed cyclization of Vinyl o-QM and chromene revealed the sensitivity of the rate of the reaction to light, oxygen, and adventitious acid, studying the reactivity of o-QM.

Full text of DOI:10.1002/chem.200800662

COMMUNICATION
DOI: 10.1002/chem.200800662
Mechanistic Investigations of the Acid-Catalyzed Cyclization of a Vinyl
ortho-Quinone Methide
[
a]
[b]
[c]
[a]
Lee M. Bishop, Michael Winkler,* Kendall N. Houk, Robert G. Bergman,* and
[
a]
Dirk Trauner*
The reversible ring opening of chromenes to yield vinyl
ortho-quinone methides (vinyl o-QMs) has been the subject
of extensive studies. This reaction is one of the first photo-
chromic processes to be investigated and has important ap-
plications in certain materials, for example, in photochromic
[1]
ophthalmic lenses. Whereas the photochemical ring open-
ing is well understood, the thermal reverse reaction has re-
ceived relatively little attention. This is mostly due to diffi-
culties with isolating stable vinyl o-QMs. These typically
contain E-configured exo-alkylidene bonds preventing them
Scheme 1. Cyclization of vinyl ortho-quinone methides.
[2–4]
from undergoing oxa-6p electrocyclizations (Scheme 1).
The kinetics of vinyl o-QM cyclization or the possibility of
catalyzing this processes have not been investigated in
detail.
for the thermal cyclization of vinyl o-QMs to afford chro-
menes.
In the context of our ongoing studies on the catalysis of
electrocyclizations, we became interested in a report by Jurd
describing an acid-promoted cyclization of the isolable vinyl
o-QM 1 to chromene 2 (Scheme 1). We now present a de-
tailed study of this reaction, which has general implications
Vinyl o-QM 1 was prepared through oxidation of phenol
[2]
3 according to literature procedures (Scheme 2). The E
configuration of the exo-alkylidene bond was unequivocally
established by detailed NMR studies and X-ray crystallogra-
phy (see Supporting Information). Interestingly, attempts to
prepare a second vinyl o-QM from phenol 4 under similar
conditions resulted in the isolation of compound 5, the
result of a Diels–Alder dimerization of two vinyl o-QM mol-
[2]
[
a] L. M. Bishop, Prof. Dr. R. G. Bergman, Prof. Dr. D. Trauner
Department of Chemistry
[3]
ecules. The structure of this unusual dimer was confirmed
University of California, Berkeley
Berkeley, CA 94720-1460 (USA)
Fax : (+1)510-642-7714
by X-ray crystallography (Figure 1). Whereas o-QM dimeri-
[5]
zation is well established in the literature, this appears to
E-mail: rbergman@berkeley.edu
trauner@berkeley.edu
[
b] Dr. M. Winkler
Institut für Organische Chemie
Universität Würzburg
Am Hubland
9
7074 Würzburg
Fax : (+49)931-888-4606
E-mail: Michael.Winkler@uni-wuerzburg.de
[
c] Prof. Dr. K. N. Houk
Department of Chemistry and Biochemistry
University of California, Los Angeles
Los Angeles, CA 90095-1569 (USA)
Fax : (+1)310-206-1843
Supporting information for this article is available on the WWW
under http://www.chemistry.org or from the author.
Figure 1. ORTEP diagram of 5 at the 50% probability level. Some hydro-
gen atoms have been removed for clarity.
Chem. Eur. J. 2008, 14, 5405 – 5408
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
5405

be the first example that leaves an o-QM moiety intact. The
analogous dimerization was not observed with compound 1
under various conditions.
Scheme 2. Preparation of ortho-quinone methides.
Scheme 3. Proposed pathways for the thermal and acid-catalyzed reac-
tions.
With vinyl o-QM 1 in hand, we proceeded to study its re-
activity. Kinetic investigation of the uncatalyzed cyclization
of 1 to 2 revealed the sensitivity of the rate of this reaction
to light, oxygen, and adventitious acid. However, reproduci-
ble kinetic data could be obtained in the absence of light,
under air-free conditions and in the presence of Proton-
Table 1. Activation parameters measured in CD
catalyzed (9 equiv acid) cyclization of 1 to 2.
2
Cl
2
for the thermal and
Catalyzed
Thermal
1
°
À1
sponge. Monitoring the cyclization by H NMR spectroscopy
revealed quantitative formation of 2 with no observable in-
DH [kcalmol
]
13.3(4)
À38(5)
24.6(8)
20.1(4)
À4.1(1)
21.3(4)
°
DS [e.u.]
°
À1
DG 298 [kcalmol
]
termediates. Kinetic data fit a first order exponential decay
À5 À1
process, with an kobs of 4.2(1)”10
s
at 508C.
Given the E configuration of the exo-alkylidene bond of
1
, we hypothesized that the rate-limiting step in the cycliza-
tion of chloroacetic acid to 1, which suggested the vinyl o-
QM was being appreciably protonated with substoichiomet-
ric amounts of acid. Low temperature NMR titration
(Figure 2) revealed near quantitative protonation of 1 using
nine equivalents of chloroacetic acid. Interestingly, this
effect was not observed using acetic acid, which allows us to
tion of 1 to 2 was the double-bond isomerization and not
the subsequent electrocyclization. The activation parameters
for the thermal reaction support this hypothesis (Table 1). A
relatively modest enthalpy of activation was measured for
this process, which we attribute to the aromatization of the
quinoidal system in the exo-alkylidene bond isomerization
transition state (Scheme 3, TS1). The reaction also exhibits
a highly negative entropy of activation, revealing a high
degree of order in the transition state of the rate-limiting
step, relative to the vinyl o-QM reactant. Entropies of acti-
vation for oxa-6p electrocyclizations are typically around
À12 e.u., which is much less negative than the value mea-
sured for this vinyl o-QM cyclization. However, entropies of
activation for the isomerization of push-pull alkenes are
quite negative, an effect that has been attributed to solvent
reorganization on going from the relatively non-polar reac-
estimate the pK of the protonated vinyl o-QM 1H to be ap-
a
proximately 4. Furthermore, the unusually high pK of 1H
a
allowed us to observe substrate-saturation kinetics using
[10]
chloroacetic acid as catalyst.
In order to determine the activation parameters for the
rate-limiting step in the catalyzed reaction, kinetic data
were obtained under acid saturation to assure that we were
monitoring only cyclization of 1H with no effect from the 1-
1H pre-equilibrium. Both the enthalpy and entropy of acti-
vation for the catalyzed reaction increase dramatically as
compared to those of the thermal reaction (Table 1). The
less negative entropy of activation for the catalyzed reaction
suggests only a small change in polarity between the proton-
ated o-QM reactant and the isomerization transition state.
This observation is consistent with the small polarity change
that would be expected under acid-saturated conditions on
going from the cationic o-QM 1H to the cationic isomeriza-
tion transition state TS1H (Scheme 3). The increased en-
thalpy of activation must be understood in terms of the ba-
sicity of 1. Protonation of 1 results in aromatization of the
quinoidal system giving a doubly benzylic cation. This aro-
matization enthalpically stabilizes the vinyl o-QM reactant
[6]
[7]
tant to the highly polar transition state. This could also be
the case for the cyclization of 1, as DFT calculations suggest
[8]
the E/Z isomerization transition state is highly polar. Thus,
these activation parameters are consistent with a stepwise
process, wherein a rate-limiting double bond isomerization
step is followed by a faster oxa-6p electrocyclization.
Next, we proceeded to investigate the influence of Brønst-
ed acids on the rate of the cyclization. Catalytic amounts of
chloroacetic acid (0.23 equiv) led to a 200-fold rate accelera-
À3 À1 [9]
tion at 218C (kobs =1.31(2)”10 s ). Interestingly, a new
species was observed by H NMR spectroscopy upon addi-
1
5406
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5405 – 5408

Mechanistic Investigations on Vinyl ortho-Quinone Methide
COMMUNICATION
Figure 2. NMR titration of 1 with ClCH
2
COOH at À408C.
relative to the isomerization transition state. This is in con-
trast to the thermal reaction, where aromatization of the
quinoidal system occurs only once the transition state is
reached, resulting in a lower enthalpy of activation for the
to a strong increase in the entropy of activation that more
than outweighs the increased enthalpy of activation. The
gas-phase computations, on the other hand, that do not in-
clude solvent effects, indicate that the catalyzed reaction is
enthalpically more favorable than the thermal isomerization
with very minor contributions from entropy. Inasmuch as
the DFT calculated energies of TS1 and TS1H are accurate,
the comparison of calculated and measured activation pa-
rameters thus gives another clue to the decisive role of sol-
vent effects in the E/Z isomerization barrier of push–pull
substituted alkenes. The computations also show that upon
protonation of 1, the s-cis/s-trans isomerization and electro-
cyclization energy barriers increase, relative to the unproto-
nated reaction (TS2H/TS3H vs TS2/TS3 and TS5H vs TS5,
respectively). The former remain slightly below the E/Z iso-
merization barrier, whereas the latter increases drastically.
We propose that the catalyzed reaction proceeds first by
rate-limiting isomerization of 1H to yield protonated Z o-
QM 6H. Because the s-cis/s-trans isomerization and electro-
[11]
thermal process.
In parallel to our experimental work, the energetics of the
thermal and catalyzed reactions were modeled by density
functional theory calculations. The pathway whereby E/Z
isomerization precedes s-cis/s-trans isomerization was mod-
eled along with the pathway whereby E/Z isomerization is
preceded by s-cis/s-trans isomerization at the UB3LYP/6-
[7]
31G* level (Figure 3). According to these calculations, the
exo-alkylidene bond isomerization is the rate-limiting step
in the thermal cyclization, having a Gibbs free energy barri-
À1
À1
er of 29.3 kcalmol , which is over 10 kcalmol higher than
the electrocyclization barrier. The lowest energy pathway
for the thermal reaction occurs through exo-alkylidene bond
isomerization of 1 to give the Z o-QM 6, which can then un-
°
À1
dergo s-cis/s-trans isomerization (DG =13.8 kcalmol ) to
give 7, which is poised to under-
go an oxa-6p electrocyclization
°
À1
(
DG =19.8 kcalmol ) to give
chromene 2. Thus, these calcu-
lations mirror our experimental
evidence that the isomerization,
not the electrocyclization, is the
rate-limiting step of the thermal
cyclization.
The catalyzed reaction was
modeled by protonation of the
carbonyl oxygen of 1 anti to the
exo-alkylidene bond. The calcu-
lated lowering of the catalyzed
exo-alkylidene E/Z isomeriza-
À1 [12]
tion barrier by 4–5 kcalmol ,
relative to the thermal reaction
barrier, is in reasonable agree-
ment with the experimental ob-
À1
servation of
a 3.3 kcalmol
lowering in the Gibbs free
energy of activation. The mea-
sured activation parameters in-
À1
Figure 3. Relative Gibbs free energies [kcalmol ] of the thermal (above) and carbonyl-protonated (below)
pathways (conversion of 1 to 2) computed at the UB3LYP/6-31G* level of theory. Structures shown refer to
dicate that acid-catalysis is due the thermal reaction. [a] Consult ref. [12].
Chem. Eur. J. 2008, 14, 5405 – 5408
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5407

D. Trauner, M. Winkler, R. G. Bergman et al.
cyclization energy barriers increase for 6H relative to 6, we
believe that the transformation from 6H to product then
occurs on the thermal reaction pathway via 6 to 7 to 2
[
1] a) S. Delbaere, J.-C. Micheau, G. Vermeersch, J. Org. Chem. 2003,
6
8, 8968–8973; b) B. Van Gemert, Benzo and Naphthopyrans
(
Chromenes) in Organic Photochromic and Thermochromic Com-
(
Scheme 3).
pounds, Vol. 1 (Eds.: J. C. Crano, R. Guglielmetti), Plenum Press:
New York, 1999, pp. 111–140; c) J. D. Hepworth, B. M. Heron,
Prog. Heterocycl. Chem. 2005, 17, 33–62; d) A. Samat, R. Gugliel-
metti, Chromogenic Materials, Photochromic in Kirk Othmer ency-
clopedia of chemical technology, Vol. 6, 5th ed. (Ed.: A. Seidel),
Wiley, Hoboken, 2004, pp. 587–606.
Interestingly, calculations indicate that the chromene
product is unstable towards protonation. The protonated
chromene 2H occupies a very flat minimum on the potential
energy surface and after inclusion of zero-point vibrational
energy corrections, the corresponding ring-opening transi-
tion state TS5H disappears as a stationary point, so that
ring-opening of the protonated chromene becomes barrier-
less. Once formed, however, the chromene product is pre-
sumably not basic enough to be protonated and is therefore
stable under the conditions used in our experiments.
In summary, we have provided evidence that the cycliza-
tion of E-configured vinyl o-QMs involves a relatively slow
acid-catalyzed exo-alkylidene bond isomerization followed
by faster oxa-6p electrocyclization. Computations indicate
that the overall reaction is catalyzed by Brønsted acids, but
the oxa-6p electrocyclization step itself is not prone to such
catalysis. The vinyl o-QM was found to be surprisingly basic,
allowing for quantitative protonation with relatively weak
acids. In addition, we have identified a new mode of Diels–
Alder dimerization of vinyl o-QMs. The factors determining
the stability of o-QMs towards dimerization are currently
under investigation in our laboratories.
[2] L. Jurd, Tetrahedron 1977, 33, 163–168.
[
3] M. R. Iyer, G. K. Trivedi, Bull. Chem. Soc. Jpn. 1992, 65, 1662–
664.
1
[
4] a) A. Arduini, A. Pochini, R. Ungaro, P. Domiano, J. Chem. Soc.
Perkin Trans. 1 1986, 1391–1395; b) T. Inoue, S. Inoue, K. Sato,
Bull. Chem. Soc. Jpn. 1990, 63, 1062–1068; c) M. G. Banwell, J. N.
Lambert, G. L. Gravatt, J. Chem. Soc. Perkin Trans. 1 1993, 2817–
2
830.
5] R. W. Van De Water, T. R. R. Pettus, Tetrahedron 2002, 58, 5367–
405.
[
[
5
6] P. Schiess, R. Seeger, C. Suter, Helv. Chim. Acta 1970, 53, 1713–
1722.
7] J. Sandstrom, Top. Stereochem. 1983, 14, 83–181.
[
[
8] The calculated dipole moments of TS1 and 1 are 4.56 and 4.52 D, re-
spectively. The intrinsic molecular charge distribution, as conjec-
tured from natural population analysis, is also comparable for both
stationary points. The strongly negative entropy of activation for the
E/Z isomerization step might be attributable to the fact that the
transition state is polar in all three spatial directions, whereas 1 is a
planar species, resulting in stronger solvation of the former. All cal-
culations have been carried out at the B3LYP/6-31G* level using
Gaussian 03: M. J. Frisch, et al. Gaussian 03; Gaussian, Inc.: Pitts-
burgh PA, 2003. See the Supporting Information for complete cita-
tion. All RDFT stationary points have been checked for triplet in-
stabilities by calculating the eigenvalues of the Hermitian stability
matrices. UDFT energies are reported throughout whenever an ex-
ternal instability was found.
Experimental Section
Experimental procedures, characterization data for products, and com-
plete citation of reference [8] are available in the Supporting Informa-
tion.
[
9] Kinetic data were not obtained for the thermal reaction at 218C.
This rate acceleration is based on extrapolation of the thermal
Eyring plot.
[
2
10] A kinetic isotope effect of 1 using ClCH COOD was measured
under acid-saturation; see Supporting Information.
11] From the available data it cannot be determined conclusively that
the enthalpy-entropy compensation effect observed is not simply a
result of the fact that the enthalpy and entropy are both derived
from the same data set and are thus not statistically independent.
See also: W. Linert, R. F. Jameson, Chem. Soc. Rev. 1989, 18, 477–
505.
[12] Despite several attempts, TS1H could not be located as a stationary
point on the PES at the RB3LYP or UB3LYP level. Given the
almost identical energies of TS1 and TS4, and according to single
[
Acknowledgements
We are indebted to Dr Allen Oliver at the UC Berkeley CHEXRAY X-
ray crystallographic facility for X-ray crystal structure determinations
and Dr. Kathleen Durkin and Dr. Jamin Krinsky at the Molecular
Graphics Facility (NSF grant CHE-0233882). We acknowledge financial
support from NSF grant CHE-0345488 (to R.G.B.) and CHE-0548209 (to
K.N.H.). M.W. thanks the Alexander von Humboldt foundation for a
Feodor-Lynen fellowship and the Fonds der Chemischen Industrie for a
Liebig fellowship.
point energy calculations, the energy of TS1H relative to 1H is esti-
À1
mated to be 24–25 kcalmol
.
Keywords: activation parameters
·
alkenes
·
density
Received: April 8, 2008
functional calculations · dimerization · isomerization
Published online: April 30, 2008
5408
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 5405 – 5408