Photochemical pinacol rearrangement

Article abstract of DOI:10.1021/jo980148p

Irradiation of 9,9′-bifluorene-9,9′-diol (1) gave 9-fluorenone and spiro[9H-fluorene-9,9′(10′-H)phenanthren]-10′-one (4), the latter arising from a pinacol rearrangement.- The distribution of fluorenone and ketone 4 was solvent dependent with the latter being the major product in trifluoroethanol, a solvent known to stabilize carbocation intermediates. Laser flash photolysis of diol 1 in 2,2,2-trifiuoroethanol or hexafluoro-2-propanol showed two transients absorbing at 350 and 505 nm with a weak band at 700 nm. The latter two peaks are assigned to the corresponding substituted 9-fluorenyl cation (5) formed from photoheterolysis of diol 1. Comparison of the decay kinetics between cation 5 and other 9-fluorenyl cations, the parent 9-fluorenyl and 9-phenyl-9fluorenyl cations, showed that the decay of 5 was relatively insensitive to the nature of the solvent as compared to the latter two carbocations suggesting that unimolecular rearrangement in 5 competes with nucleophilic quenching.

Full text of DOI:10.1021/jo980148p

7
168
J . Org. Chem. 1998, 63, 7168-7171
P h otoch em ica l P in a col Rea r r a n gem en t
†
‡
†
†
†
Mary Hoang, Timothy Gadosy, Hedieh Ghazi, Dong-Feng Hou, Alan C. Hopkinson,
,
‡
,†
Linda J . J ohnston,* and Edward Lee-Ruff*
Department of Chemistry, York University, Toronto, Ontario M3J 1P3, Canada, and Steacie Institute for
Molecular Science, National Research Council Canada, Ottawa, Ontario K1A 0R6, Canada
Received J anuary 27, 1998
Irradiation of 9,9′-bifluorene-9,9′-diol (1) gave 9-fluorenone and spiro[9H-fluorene-9,9′(10′-H)-
phenanthren]-10′-one (4), the latter arising from a pinacol rearrangement. The distribution of
fluorenone and ketone 4 was solvent dependent with the latter being the major product in
trifluoroethanol, a solvent known to stabilize carbocation intermediates. Laser flash photolysis of
diol 1 in 2,2,2-trifluoroethanol or hexafluoro-2-propanol showed two transients absorbing at 350
and 505 nm with a weak band at 700 nm. The latter two peaks are assigned to the corresponding
substituted 9-fluorenyl cation (5) formed from photoheterolysis of diol 1. Comparison of the decay
kinetics between cation 5 and other 9-fluorenyl cations, the parent 9-fluorenyl and 9-phenyl-9-
fluorenyl cations, showed that the decay of 5 was relatively insensitive to the nature of the solvent
as compared to the latter two carbocations suggesting that unimolecular rearrangement in 5
competes with nucleophilic quenching.
In tr od u ction
Destabilized carbocations in which one or more of the
R-substituents are electron-withdrawing and cations
incorporated in a 4n cyclic π-framework have been of
interest as intermediates in solvolysis,1
-4
as species
directly observable under stable ion conditions,5 or as
,6
transients obervable under laser flash photolysis in neu-
tral solutions.6
-9
The 9-fluorenyl cation can be readily
dication 2, a “doubly destabilized” cation.3 Dications of
the 1,1,2,2-tetraarylethane-1,2-diyl type have been ob-
served under stable ion conditions from oxidations of the
corresponding 1,1,2,2-tetraarylethylenes.12 In this study
we report on the photochemistry of 1 and the first
example of a photochemical pinacol rearrangement per-
formed in neutral solutions.13
produced by irradiation of 9-fluorenols unlike other di-
arylmethyl cations which require more photolabile leav-
ing groups. This is associated with the ease of photosol-
volysis processes of cyclic π-chromophores possessing 4n
electrons in the internal cyclic array (ICA).10 The pho-
tosolvolysis rates of fluorenols exceed those of dimethoxy-
benzyl alcohol which supports theoretical studies sug-
gesting that antiaromatic systems (in this case, the
9
-fluorenyl cation) in the ground state may possess
Resu lts a n d Discu ssion
1
1
aromatic stabilization in the excited state.
Diol 1 was prepared by reductive coupling of 9-fluo-
renone (3) according to a procedure described by Tana-
With the ease with which 9-fluorenyl cations can be
generated by irradiation of the corresponding alcohols,
we were interested in the possibility of using 9,9′-
bifluorene-9,9′-diol (1) as a precursor to the corresponding
14
ka.
Irradiation of 1 in acetonitrile gave two major
products, 9-fluorenone and the rearranged pinacolone 4,
22
in a ratio of 85:15, respectively. The characterization
of 4 was based on comparison of its spectral data with
those reported in the literature as well as comparison
with an authentic sample prepared from the acid-
catalyzed pinacol rearrangement of diol 1. Irradiation
of deaerated solutions of 1 in acetonitrile gave a slightly
†
York University.
National Research Council of Canada.
15
‡
(
1) Creary, X. Chem. Rev. 1991, 91, 1625-1677.
(2) Allen, A. D.; Fujio, M.; Mohammed, N.; Tidwell, T. T.; Tsuji, Y.
J . Org. Chem. 1997, 62, 246-252.
3) Allen, A. D.; Colombovakos, J . D.; Tee, O. S.; Tidwell, T. T. J .
Org. Chem. 1994, 59, 7185-7187.
4) Amyes, T. L.; Richard, J . P.; Novak, M. J . Am. Chem. Soc. 1992,
14, 8032-8041.
5) Creary, X.; Hopkinson, A. C.; Lee-Ruff, E. Adv. Carbocation
Chem. 1989, 1, 45-92.
6) Lew, C. S. Q.; Wong, D. F.; J ohnston, L. J .; Bertone, M.;
Hopkinson, A. C.; Lee-Ruff, E. J . Org. Chem. 1996, 61, 6805-6808.
7) Mecklenburg, S. L.; Hilinski, E. F. J . Am. Chem. Soc. 1989, 111,
471-5472.
8) McClelland, R. A.; Mathivanan, N.; Steenken, S. J . Am. Chem.
Soc. 1990, 112, 4857-4861.
9) Lew, C. S. Q.; Wagner, B. D.; Angelini, M. P.; Lee-Ruff, E.;
Lusztyk, J .; J ohnston, L. J . J . Am. Chem. Soc. 1996, 118, 12066-
2073.
(
(
(12) Olah, G.; Grant, J ..L.; Spear, R. L.; Bollinger, J . M.; Serianz,
A.; Sipos, G. J . Am. Chem. Soc. 1976, 98, 2501-2507.
(13) Other photochemical Wagner-Meerwein rearrangements have
been reported from irradiation of alkyl and benzyl halides or nucle-
ofugal groups more reactive than OH (see: Cristol, S. J .; Bindel, T. H.
Org. Photochem. 1983, 6, 327). However, in many such cases homolytic
dissociation followed by electron transfer appears to be the predomi-
nant mechanism. One report on the photolysis of a benzyl alcohol giving
products of 1,3-aryl migration appears to involve stabilized benzylic
carbocations (cf.: Lin, C.-I.; Singh, P.; Ullman, E. F. J . Am. Chem.
Soc. 1976, 98, 6711-6713).
1
(
(
(
5
(
(
(14) Tanaka, K.; Kishigami, S.; Toda, F. J . Org. Chem. 1990, 55,
2981-2983.
(15) Minami, T.; Matsuzaki, N.; Ohshiro, Y.; Agawa, T. J . Chem.
Soc., Perkin Trans. I 1980, 1731-1738.
1
(
10) Wan, P.; Krogh, E. J . Am. Chem. Soc. 1989, 111, 4887-4895.
(11) Mallar, E. J . P.; J ug, K. Tetrahedron 1986, 42, 417-426.
S0022-3263(98)00148-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/30/1998

Photochemical Pinacol Rearrangement
J . Org. Chem., Vol. 63, No. 21, 1998 7169
higher yield of 4 affording a 74:26 ratio of ketones 3 and
, respectively.2
2
4
Attempted trapping of carbocations generated by ir-
radiation of diol 1 in methanol gave only ketones 3 and
in a 67:33 ratio.22 The use of 2,2,2-trifluoroethanol
TFE), a solvent known to promote carbocation forma-
4
(
1
6
tion, gave the pinacolone 4 as the principal product
2
3
along with minor amounts of fluorenone 3 (vide infra).
No dark reactions of diol 1 in any of these solvents were
observed. The solvent dependence of the product distri-
bution suggests that pinacolone 4 and fluorenone 3 arise
via different mechanisms and that 4 is likely the product
from a carbocationic intermediate. To probe the nature
of these intermediates, we carried out laser flash pho-
tolysis (LFP) studies of diol 1.
Transient spectra obtained following 308 nm excitation
of diol 1 in 1:1 TFE/hexafluoro-2-propanol (HFIP) showed
a strong absorption at 505 nm, a weak band centered at
approximately 700 nm, and a third band at 350 nm
(Figure 1). Neither the intensity nor the decay kinetics
of the 505 and 700 nm bands was affected by oxygen;
the two bands decay with similar kinetics with measured
F igu r e 1. Transient spectra measured after 308 nm excitation
of diol 1 in 1:1 TFE/HFIP (A, nitrogen, and B, oxygen). Spectra
were measured at the following time delays after the laser:
b, immediately after laser pulse; O, 0.48 µs; 0, 1.3 µs; 9, 3.2
µs.
6
-1
rate constants of 1.4 and 2.0 × 10 s in nitrogen-purged
solution. These numbers are in good agreement, given
the small amplitude of the longer wavelength signal and
indicate that the two bands belong to the same species.
This is substantiated by the fact that similar rate
constants for reaction with nucleophiles are measured
at both 505 and 700 nm. Second-order rate constants
6
for reaction with water and methanol of k
w
) 1.8 × 10
-1
-1
7
-1 -1
M
s
and kMeOH ) 1.2 × 10 M
s , respectively, were
determined from the slopes of the plots of the observed
decay rate constant at 505 nm as a function of the
quencher concentration. The 350 nm signal shown in
Figure 1 decays much more slowly than the two longer
wavelength bands in nitrogen-purged solution and is
partially removed by oxygen, as shown in Figure 1A,B.
This clearly demonstrates that two transient species are
produced upon excitation of diol 1.
LFP of diol 1 was also examined in several other
solvents. Spectra that were virtually identical to those
shown in Figure 1 were obtained in HFIP. The 505/700
5
-1
nm species decayed with a rate constant of 4.4 × 10 s
(
measured at 510 nm for an oxygen-purged sample) and
F igu r e 2. Plots of the observed rate constant for the decay of
transient 5 (510 nm) generated from diol 1 as a function of
water (b) and methanol (9) concentration.
reacted with water and methanol with second-order rate
5
-1 -1
constants of k
w
) 2.7 × 10 M
s
and kMeOH ) 2.1 ×
6
-1
-1
1
0
M
s
(Figure 2). The lifetime of the 505 nm
two-thirds of the signal decayed within 100 ns, leaving
a long-lived residual absorption that did not decay over
hundreds of microseconds. Neither the intensity nor the
decay kinetics of the 350 nm transient is affected by the
addition of 0.028 M 2,3-dimethyl-1,3-butadiene, a triplet
state quencher.
The above results are consistent with the assignment
of the 505 nm transient to the 9-fluorenyl cation 5 formed
by photoheterolysis of diol 1. The assignment is based
on the observed reactivity of the transient to nucleophiles,
transient was substantially shorter in neat TFE, with
7
-1
an observed decay rate constant of 1.2 × 10
s . In
acetonitrile, only the 350 nm signal was observed and
there was no evidence for the 505/700 nm transient,
even on short time scales. The 350 nm signal decayed
slowly by second-order kinetics in nitrogen-purged ace-
tonitrile. In oxygen-saturated solution, approximately
(16) Eberson, L.; Hartshorn, M. P.; Persson, O.; Radner, F. J . Chem.
Soc., Chem. Commun. 1996, 2105-2112.

7
170 J . Org. Chem., Vol. 63, No. 21, 1998
Hoang et al.
its insensitivity to oxygen, and comparison of its spectral
features with known 9-fluorenyl cations. For example,
the parent 9-fluorenyl cation shows a λmax at 515 nm in
8
HFIP and both the 9-phenyl and 9-methyl derivatives
are slightly blue-shifted with λmax values of 490 and 485
nm, respectively, in alcohol solutions.7 The kinetic data
obtained for trapping by methanol and water are also in
good agreement with the data obtained for other substi-
tuted 9-fluorenyl cations. The modest decreases in rate
constants for HFIP vs TFE/HFIP are in line with kinetic
data for other cations and reflect the stronger solvating
ability of HFIP as compared to TFE.16
,17
The decay kinetics for the 9-fluorenyl and 9-phenyl-9-
fluorenyl cations were also measured in several solvents
for comparison with the data obtained for 5. Both cations
were obtained by photolysis of the corresponding alcohol.
The decay kinetics for the 9-fluorenyl cation decrease by
over 2 orders of magnitude when the solvent is changed
from 1:1 TFE/HFIP to HFIP, as shown in Figure 3. By
contrast the rate constant for decay of 5 decreases by only
a factor of 5. The 9-phenyl-9-fluorenyl cation decays with
F igu r e 3. Effect of solvent composition (XHFIP) in TFE/HFIP
mixtures on the rate constant for decay of 9-fluorenyl cation
(0) and the substituted 9-fluorenyl cation 5 (O).
4
-1
a rate constant of 1.6 × 10 s in TFE, in good agreement
readily arise from oxidation of 7 under ambient condi-
tions. Similar observations were obtained from other
with literature data,17 as compared to 1.2 × 10 s for
cation 5. If nucleophilic addition of solvent was respon-
sible for the decay of both cations then one would expect
cation 5 to decay with similar kinetics or even more
slowly than the 9-phenyl analogue, as a result of steric
effects. The rapid decay for cation 5 and the relative
insensitivity of its decay to variation in the solvent com-
position as compared to other 9-fluorenyl cations suggest
that nucleophilic addition is not the primary pathway.
It is not surprising that attempted nucleophilic trapping
of cation 5 by irradiation of diol 1 in methanol only gave
rearrangement product 4 and fluorenone. Both these
results and the product studies are consistent with
unimolecular decay of 5 via pinacol rearrangement to give
the stabilized cation 6.
7
-1
1
,1,2,2-tetraaryl-1,2-ethanediols in which the only prod-
ucts obtained were the corresponding benzophenones and
20
other oxidation products derived from ketyl radicals.
In conclusion we have shown that a pinacol rearrange-
ment occurs upon photoexcitation of 9,9′-bifluorene-9,9′-
diol (1). The reaction proceeds via initial C-O bond
heterolysis to give a substituted 9-fluorenyl cation that
has been observed by LFP. This cation undergoes
rearrangement and deprotonation giving the final prod-
uct. The fluorenone byproduct may arise from a compet-
ing homolysis process involving ketyl radical. There is
no evidence from either the photoproduct analysis or from
the LFP studies for the intervention of dication 2. The
possibility of double photon excitation of diol 1 to give
dication 2 through the monocation 5 will be investigated.
Exp er im en ta l Section
5
-1
The measured rate constant of 5 × 10 s for decay of
cation 5 in HFIP in the absence of quencher (Figure 3)
represents an upper limit for the pinacol rearrangement.
This suggests an activation barrier of ∆G ) 9 to 10 kcal/
mol in HFIP which is in line with activation barriers of
Ma ter ia ls a n d Gen er a l Tech n iqu es. All the solvents
used were reagent grade and were dried before use. Infrared
spectra were obtained on a Perkin-Elmer 1310 spectrometer
with KBr pellets. Proton NMR spectra were obtained on a
Bruker ARX-400 spectrometer using samples dissolved in
†
CDCl . Chemical shifts are measured in parts per million (δ)
3
3
-12 kcal/mol reported for 1,2-alkyl migrations in less
relative to TMS as internal standard. Mass spectra were
recorded at 70 eV on a Kratos profile mass spectrometer. The
continuous photolysis were performed using a Hanovia 450
W medium-pressure mercury arc lamp in a water-cooled
quartz immersion well. Quartz tubes containing the samples
were strapped around this well, and the assembly was im-
mersed in an ice-water bath. HPLC analyses were carried
out using a Waters 715 Ultra Wisp Sampler with a Lamda
Max 481 LC detector set at 254 nm. The column was a µ
Porasil analytical column using a 30:70 mixture of ethyl
acetate/hexane mixture as eluent. Standard solutions of diol
solvating media with the lower values associated with
alkyl migrations between tertiary carbocation centers.
1
8
The transient observed at 350 nm may be associated
with the ketyl radical 7 formed from C-C homolysis of
diol 1. The fluorenyl ketyl radical 7 has been reported
to show an absorption at λmax at 360 nm which is solvent
dependent.19 The observed fluorenone byproduct could
(
17) Cozens, F. L.; Mathivanan, M.; McClelland, R. A.; Steenken,
S. J . Chem. Soc., Perkin Trans. 2 1992, 2083-209018.
18) Olah, G.; Prakash, G. K. S.; Sommer, J . Superacids; J . Wiley
Sons: New York, 1985; pp 128-146.
19) Bicz o´ k, L; B e´ rces, T.; Linschitz, H. J . Am. Chem. Soc. 1997,
19, 11071-11077.
(
&
(20) Hoang, M. Preparation and Photochemistry of symmetrically
p-Substituted 1,1,2,2-tetraphenyl-1,2-ethanediols. M.Sc. Thesis, York
University, 1997.
(
1

Photochemical Pinacol Rearrangement
, 9-fluorenone (3), and ketone 4 were used for calibration
J . Org. Chem., Vol. 63, No. 21, 1998 7171
1
indicated 61% of ketone 4, 12% of 9-fluorenone (3), and 18%
of diol 1. Decomposition of diol 1 to 9-fluorenone and polar
products which could not be identified led to a less than
satisfactory mass balance in the recovery of products.
Photolysis of diol 1 in aerated solution was conducted in an
identical manner as above except for the purging with argon
prior to the photolysis.
purposes in the HPLC analyses of the photolysis mixtures.
Spir o[9H-flu or en e-9,9′(10′-H)-ph en an th r en ]-10′-on e (4).
Diol 1 (0.0691 g, 0.19 mmol) was dissolved in 10 mL of glacial
acetic acid containing 3 mL of sulfuric acid (5% aqueous) and
heated to reflux for 45 min. The reaction mixture was poured
onto ice (50 g) and the precipitate filtered off and washed with
water. The solid was recrystallized in methanol giving 0.0575
g of the title compound: mp 254-256 °C (lit. mp 254-255
La ser F la sh P h otolysis. The laser system has been
described in detail elsewhere.21 A Lumonics EX-510 excimer
laser (XeCl, 308 nm, 6-ns pulses, <40 mJ /pulse) was used for
1
5
-1 1
°
(
7
C ); IR 1683 cm ; H NMR δ 8.33 (d, 1H), 8.23 (d, 1H), 8.02
d, 1H), 7.83 (m, 4H), 7.49 (t, 1H), 7.42 (m, 4H), 7.32 (t, 1H),
.24 (t, 1H), 7.08 (d, 1H), 6.66 (d, 1H); MS m/z 344 (M), 316
2
sample excitation. Solutions were contained in 7 × 7 mm
Suprasil quartz cells and were either aerated or purged with
nitrogen or oxygen before laser excitaion.
(M - CO).
P r ep a r a tive P h otolysis. A solution of diol 1 (48 mg, 0.132
J O980148P
mmol) in 25 mL of 2,2,2-trifluoroethanol or acetonitrile was
purged with argon for 30 min. The solution was irradiated
for 45 min at 0 °C. After evaporation of the solvent, the
residue was purified by preparative thin-layer chromatography
(21) Kazanis, S.; Azarani, A; J ohnston, L. J . J . Phys. Chem. 1991,
9
5, 4430-4435.
(
22) By HPLC analysis; the fluorenone content is overestimated due
to decomposition of unreacted diol 1 to 9-fluorenone (3) on silica gel
chromatography.
(silica gel; 2:8 ethyl acetate/hexane) giving 10 mg of unreacted
1
, 4 mg (17%) of fluorenone, and 15 mg (40%) of ketone 4
identical in all respects with a sample prepared above. HPLC
analysis of the photolysis mixture in 2,2,2-trifluoroethanol
(23) The mixture contained 61% of ketone 4, 12% of 9-fluorenone
(3), and 18% of unreacted diol 1 by HPLC analysis.