Superacid-catalyzed electrocyclization of diphenylmethyl cations to fluorenes. Kinetic and theoretical revisit supporting the involvement of ethylene dications

Article abstract of DOI:10.1021/ja980090w

In superacid media, diphenylmethyl cations bearing an α-carbonyl group generate fluorene compounds. We have obtained chemical evidence showing the acidity dependence of this fluorene cyclization process. A linear relationship was found between the rate of

Full text of DOI:10.1021/ja980090w

J. Am. Chem. Soc. 1998, 120, 4629-4637
4629
Superacid-Catalyzed Electrocyclization of Diphenylmethyl Cations to
Fluorenes. Kinetic and Theoretical Revisit Supporting the
Involvement of Ethylene Dications
Tomohiko Ohwada,* Takayoshi Suzuki, and Koichi Shudo
Contribution from the Graduate School of Pharmaceutical Sciences, UniVersity of Tokyo, 7-3-1 Hongo,
Bunkyo-ku, Tokyo 113, Japan
ReceiVed January 8, 1998
Abstract: In superacid media, diphenylmethyl cations bearing an R-carbonyl group generate fluorene
compounds. We have obtained chemical evidence showing the acidity dependence of this fluorene cyclization
process. A linear relationship was found between the rate of the fluorene cyclization and the acidity of the
reaction media in kinetic studies, strongly supporting the intervention of an additional proton transfer to the
diphenylmethyl monocation in the fluorene cyclization, that is, the involvement of a dication in the form of
the diphenylmethyl cation bearing an O-protonated carbonyl group. Ab initio calculations employing Becke3-
LYP density functional theory provided theoretical support for the idea that the dicationic fluorene cyclization
is energetically more favorable than monocationic electrocyclization to give the benzofuran or the fluorene.
The magnitudes of the predicted activation barriers of the dicationic fluorene cyclization are compatible with
the experimental values for the fluorene cyclization and with the experimental dependence of the reaction
aptitude on the R-carbonyl substituent (ester, acid, or acetyl group).
Introduction
Scheme 1
Diphenylmethyl cations (1) and related diarylmethyl cations
substituted with electron-withdrawing groups constitute a typical
class of destabilized cations, i.e., carbocations bearing electron-
withdrawing groups directly attached to the carbenium center.1
The prototype diphenylmethyl cations (1 (R2 ) H)) bearing
electron-withdrawing substituents have been generated in the
presence of acids,3 or more recently by flash-laser photolysis,4
and characterized as discrete stable ions. The cations 1 were
proposed to undergo three modes of reaction depending on the
electron-withdrawing substituent (Scheme 1): (1) cyclization to
form benzofurans, (2) cyclization to give fluorenes, and (3)
electrophilic reaction onto the carbenium center. Unresolved
discrepancies were encountered in the former two cyclization
reactions with respect to the reaction conditions, the substituent
effect, and in particular, the reactive intermediates. In the case
of the cations substituted with a ketone (for example, 1a and
1i), it was reported that R-benzoyldiphenylmethanol 4a and
R-acetyldiphenylmethanol 4i in CHCl3/H2SO4 underwent elec-
trocyclization between the benzene ring and the carbonyl group
(Scheme 2), resulting in the formation of benzofurans 2a and
2i, respectively,3,5 although the formation of the fluorene
derivative 3a has also been reported from 4a under similar
(1) (a) Creary, X. Chem. ReV. 1991, 91, 1625-1678. (b) Creary, X.;
Hopkinson, A. C.; Lee-Ruff, E. AdVances in Carbocation Chemistry; Creary,
X., Ed.; JAI Press: Greenwich, CT, 1989; Vol. 1, p 45. (c) Gassman, P.
G.; Tidwell, T. T. Acc. Chem. Res. 1983, 16, 279. (d) Tidwell, T. T. Angew.
Chem., Int. Ed. Engl. 1984, 23, 20-32.
(2) Takeuchi, K.; Kitagawa, T.; Okamato, K. J. Chem. Soc., Chem.
Commun. 1983, 7.
(3) (a) Dao, L. H.; Maleki, M.; Hopkinson, A. C.; Lee-Ruff, E. J. Am.
Chem. Soc. 1986, 108, 5237-5242. (b) Hopkinson, A. C.; Lee-Ruff, E.;
Maleki, M. Synthesis 1986, 366-371.
(4) Johnston, L. J.; Kwong, P.; Shelemay, A.; Lee-Ruff, E. J. Am. Chem.
Soc. 1993, 115, 1664-1669.
(5) Hopkinson, A. C.; Dao, L. H.; Duperrouzel, P.; Maleki, M.; Lee-
Ruff, E. J. Chem. Soc., Chem. Commun. 1983, 727.
(6) Maleki, M.; Hopkinson, A. C.; Lee-Ruff, E. Tetrahedron Lett. 1983,
4911.
conditions.6 In contrast to this heteroatom cyclization pro-
cess, the acid, ester, and amide analogues underwent electro-
cyclic coupling of the two aromatic rings, leading to the fluorene
S0002-7863(98)00090-0 CCC: $15.00 © 1998 American Chemical Society
Published on Web 04/30/1998

4630 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Ohwada et al.
Scheme 2
Scheme 3
superacid-catalyzed cyclization of the diphenylmethyl cations
1, aiming (1) to establish chemically and kinetically the acidity
dependence of the fluorene cyclizations, and the intervention
of an additional proton transfer to the monocation 1, and (2) to
demonstrate theoretically the energetic favorability of the
dicationic electrocyclization to give the fluorene over the
monocationic electrocyclization to give the benzofuran or the
fluorene. We also show that the magnitudes of the predicted
activation barriers of the dicationic fluorene cyclization are con-
sistent with the experimental values for the fluorene cyclization
and with the observed dependence of the reaction aptitude on
the R-carbonyl substituent (ester, acid, or acetyl group).
derivatives: for example, R-(methoxycarbonyl)diphenylmetha-
nol 4c gave 9-(methoxycarbonyl)fluorene 3c.3,7 These cycliza-
tions were proposed to take place through the monocation 1.3
The isolation of the relevant carbocation, the di-p-anisyl(p-
methoxybenzoyl)methyl cation 1b (R1 ) p-methoxyphenyl, R2
) OCH3), as a stable crystalline antimony pentafluoride salt
was reported by Takeuchi et al., who showed that heating of
the salt 1b in a neutral solvent such as 1,2-dichloroethane gave
the benzofuran 2b (Scheme 3).2 This study clearly demonstrated
that the benzofuran cyclization of the cation 1 does not require
acid catalysis.
Results
Acidity Dependence of the Reactions. Formation of
9-(methoxycarbonyl)fluorene 3c was observed when R-(meth-
oxycarbonyl)diphenylmethanol 4c was dissolved in TFSA at
-45 °C.8 The cyclization depends on the acidity of the reaction
medium, as judged from the yields of the product (Table 1). In
TFA (H0 ≈ -2.7), the cyclization of 4c to give the fluorene 3c
did not take place. As the acidity was increased to H0 ≈ -11.8
(56% TFSA- 44% TFA), the reaction of 4c started, though
slowly. The reaction in TFSA was rapid, and the fluorene 3c
was formed in high yield. A similar acceleration of the reaction
with increase of the acidity was observed in the cases of the
bis(p-bromophenyl) (4d) and bis(p-methylphenyl) (4e) deriva-
tives. The cyclization of R-(methoxycarbonyl)bis(p-fluorophe-
nyl)methanol 4f in TFSA was slow, and the yield was moderate.
A decade ago, the chemical behaviors of the monocations 1,
generated under neutral and strongly acidic conditions were
reported from this laboratory8 and were divergent from those
described above: R-benzoyldiphenylmethanol 4a (R1 ) Ph, R2
) H), a precursor of the unsubstituted cation (1a), reacted in
trifluoromethanesulfonic acid (TFSA) at -50 °C to afford the
fluorene 3a, along with the phenanthrene derivative, 9-phenyl-
phenanthr-10-ol. No benzofuran was obtained in TFSA. This
superacid-catalyzed fluorene cyclization of 4a was also reported
by Olah and Wu in the same acid,9 although the yields were a
little different. The monocation 1a could be generated as a
stable entity at -50 °C by the reaction of the R-chloro ketone
5 with silver salts. However, no fluorene or benzofuran was
formed under these neutral conditions. When the stable cation
was added to TFSA at -50 °C, the fluorene and the phenan-
threne were produced. This led to the proposals that the
fluorenes 3 do not arise directly from the monocation 1 and
that the real intermediate is the dication 6 (Scheme 4), formed
by protonation of the R-carbonyl group by TFSA, although an
attempt at direct spectroscopic observation of the dication (6a
and 6c) was unsuccessful.8 Since the review of these discrep-
ancies in the behavior of the cations 1 in Chemical ReViews by
Creary,1a there has been no advance toward achieving a
consistent understanding of the features of the acid-catalyzed
fluorene cyclization. This cyclization represents an acid-
catalyzed electrocyclization wherein two aromatic rings par-
ticipate and is of synthetic interest.3b Here we revisit the
+
The pKR values of the R-carbonyl-substituted monocations
1 are not available,4 while those of the parent diphenylmethyl
cations are -6.9 (diphenylmethyl cation), -7.2 (bis(p-chloro-
phenyl)methyl cation), and -5.1 (bis(p-methylphenyl)methyl
cation),12 as calibrated in terms of the H0 scale.13 The pKR
+
value of the dibromo derivative bearing an R-ester (4d) is
estimated to be around -10 (H0) on the basis of the following
observation: a solution of the bis(p-bromophenyl)methanol 4d
in 40% TFSA-60% TFA (H0 ≈ -11, 1000 equiv of the acid),
prepared at -45 °C, was quenched with a large excess of
methanol to give, in 94% yield, a methoxy derivative, which is
formed by the addition of methanol to the corresponding bis-
(p-bromophenyl)methyl cation 1d. This experiment suggests
at least 90% ionization of the alcohol 4d to the cation 1d at H0
≈ -11. If substitution of an ester group in the diphenylmetha-
nols has a consistent retarding effect upon ionization to the
+
diphenylmethyl cations, the pKR values of the alcohols bearing
an R-ester group (4) should be around -10 (4c) and -8 (4e),
respectively.
(10) The acidity (H0) of the trifluoromethanesulfonic acid (TFSA)-
trifluoroacetic acid (TFA) system has been described: Saito, S.; Saito, S.;
Ohwada, T.; Shudo, K. Chem. Pharm. Bull. 1991, 39, 2718-2720. See
also Saito, S.; Sato, Y.; Ohwada, T.; Shudo, K. J. Am. Chem. Soc. 1994,
116, 2312-2317, footnote 8.
(7) Vorlander, D.; Pritzsche, A. Ber. Dtsch. Chem. Ges. 1913, 46, 1793.
Britzrzycki, A.; Herbst, C. Ber. Dtsch. Chem. Ges. 1903, 36, 145. Dobeneck,
H. v.; Kiefer, R. Justus Liebigs Ann. Chem. 1965, 684, 115. Arnold, R. T.;
Parham, W. E.; Dodson, R. M. J. Am. Chem. Soc. 1949, 71, 2439.
Hopkinson, A. C.; Khazanie, P. G.; Dao, L. H. J. Chem. Soc., Perkin Trans.
2 1979, 1395. Delacre, M. Bull. Soc. Chim. Fr. 1918, 23, 229.
(8) (a) Ohwada, T.; Shudo, K. J. Am. Chem. Soc. 1988, 110, 1862-
1870. (b) Ohwada, T.; Shudo, K. J. Org. Chem. 1989, 54, 5227-5237.
(9) Olah, G. A.; Wu, A.-h. J. Org. Chem. 1991, 56, 2531-2534. For a
review of superelectrophiles, see Olah, G. A. Angew. Chem., Int. Ed. Engl.
1993, 32, 767-788.
(11) Suzuki, T.; Ohwada, T.; Shudo, K. J. Am. Chem. Soc. 1997, 119,
6774-6780.
(12) Deno, N. C.; Jaruzelski, J. J.; Schriesheim, A. J. Am. Chem. Soc.
1955, 77, 3044-3051. Deno, N. C.; Schriesheim, A. J. Am. Chem. Soc.
1955, 77, 3051-3054.
(13) Jorgenson, M. J.; Hartter, D. R. J. Am. Chem. Soc. 1963, 85, 878-
883.

Superacid-Catalyzed Electrocyclization
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4631
Scheme 4
Table 1. Acid-Catalyzed Cyclization of Diarylmethanols Substituted with a Carbonyl Group and Its Acidity Dependence
a
R1
R2
acid
-H0
time (h)
temp (°C)
recovery (%)
yield (%)
4c
OCH3
H
TFA
2.7
11.8
13.2
12.0
13.2
11.0
13.2
13.2
13.2
13.2
12.0
13.2
10.9
13.2
1
1
1
5
5
5
5
144
1
2
15 min
15 min
15 min
15 min
-20
-45
-45
10
10
0
0
20
0
-20
-45
-45
-45
-45
100
75
0
47
0
80
0
0
0
0
65
0
79
0
0
20
56% TFSA-44% TFA
TFSA
54% TFSA-46% TFA
TFSA
31% TFSA-69% TFA
TFSA
TFSA
TFSA
TFSA
55% TFSA-45% TFA
TFSA
30% TFSA-70% TFA
TFSA
94b
49
4d
4e
OCH3
OCH3
Br
100
15
CH3
94c
44
4f
OCH3
OH
OH
F
H
CH3
H
4g
4h
4i
65d
82e
19f
86g,h
0i
CH3
4j
CH3
CH3
65j
a Corrected values of acidity function of the reaction media (see Table 7 and ref 10). b 81% yield (ref 8a). c 94% (ref 8b). d Reference 8a. e Reference
8b. f The fluorene dimer (7) (see the Experimental section) (16%) and fluorene (3%). g 70% yield (ref 8a). h The fluorene dimer (7) is also formed
(1.8%). i Benzofuran (2j) 1.5%. j Benzofuran (2j) 6.5%.
Table 2. Rate Constants for Acid-Catalyzed Cyclizations of
4c-ea,b
Similarly, benzilic acid (4g) was reported to give fluorene-
9-carboxylic acid in 65% yield in TFSA8a and R-carboxybis-
(p-methylphenyl)methanol (4h) gave the fluorene 3h in 82%
yield in TFSA.8b The reaction of 4h was faster than that of the
corresponding ester 4e under similar reaction conditions (in
TFSA), the ratio being approximately 9 (4h/4e).8b The observed
rate ratio is due to the inductive stabilization of the cationic
intermediate by the substituent (CH3) at the carbonyl group.
R-Acetyldiphenylmethanol 4i also gave 9-acetylfluorene 3i
(86% yield) in TFSA at -45 °C, indicating that the benzoyl
group is not necessary for the fluorene formation (Table 1).8,14
At H0 ≈ -12.0 (in 55% TFSA-45% TFA), the reaction is very
slow. In TFSA (H0 ≈ -13.2) the reaction was accelerated,
being completed within 15 min. The reaction of R-acetylbis-
(p-methylphenyl)methanol 4j is also dependent on the acidity
of the medium: in 30% TFSA-70% TFA (H0 ≈ -10.9), the
reaction is too slow to detect, although ionization to the bis(p-
methylphenyl)methyl cation 1j should occur to a large extent,
4cc
105k (s-1
4dc
105k (s-1
4ec
105k (s-1
)
H0
)
H0
)
H0
-12.7
-12.6
-12.2
-11.9
-11.7
82.55
70.96
39.53
21.83
15.58
-13.2
-12.7
-12.6
-12.2
-11.9
-11.6
-11.3
11.34
6.35
5.55
3.38
1.77
1.19
0.75
-13.2
-12.7
-11.7
-11.5
-11.1
-10.7
37.20
24.04
6.32
2.78
1.46
0.61
a Corrected values of acidity function of the reaction media (see Table
7 (Experimental Section) and refs 10 and 11). b Errors of rates, (2%.
c At 0 °C.
phenyl)methyl monocation 1j is much slower at -45 °C than
the fluorene cyclization. The fluorene cyclization reactions of
the R-acetyl derivatives were too fast even at -45 °C to allow
kinetic measurements to be made (vide infra), in contrast to
those of the ester analogues. Therefore the experimental
reaction aptitude for the fluorene cyclization is in the order ester
< acid , acetyl. Nevertheless, the similar acidity dependences
of the reaction of the R-acetyl cations (1i and 1j) strongly
suggest a similar reaction mechanism of the cyclization to that
in the case of the ester derivatives.
Kinetic Evidence for Proton Transfer in the Fluorene
Cyclization. We measured the reaction rates of the fluorene
cyclization of several R-methoxycarbonyldiarylmethanols (4c,
4d, and 4e) in the TFSA-TFA acid system (Table 2). The
acidity rate profiles of the reactions of the R-(methoxycarbonyl)-
diphenylmethyl cation 1c and the R-(methoxycarbonyl)bis(p-
bromophenyl)methyl cation 1d were linear, with slopes of 0.71
and 0.63, respectively (Figure 1). This linearity supports the
intervention of the additional proton transfer to the diphenyl-
+
as judged from the estimated pKR value (-8) of the bis(p-
methylphenyl)methanol 4e. Under this condition, a trace (1.5%
yield) of the benzofuran derivative 2j was detected. In TFSA
(H0 ≈ -13.2), the fluorene cyclization was very rapid (the yield
was increased to 65% yield) and was accompanied with the
formation of the benzofuran derivative 2j (6.5% yield). The
isolated benzofuran derivative 2j was stable in TFSA at -30
1
°C for 1 h (monitored by H NMR spectroscopy), showing no
transformation to the fluorene derivative 3j. These results
indicate that the benzofuran cyclization of the bis(p-methyl-
(14) Carter, J. P.; Noronha-Blob, L.; Audia, V. H.; Dupont, A. C.;
McPherson, D. W.; Natalie, K. J.; Rzeszotarski, W. J.; Spagnuola, C. J.;
Waid, P. P.; Kaiser, C. J. Med. Chem. 1991, 34, 3065-3074. Koenigkramer,
R. E.; Zimmer, H. J. Org. Chem. 1980, 45, 3994-3998. Kaiser, C.; Audia,
V. H.; Carter, J. P.; McPherson, D. W.; Waid, P. P.; Lowe, V. C.; Noronha-
Blob, L. J. Med. Chem. 1993, 36, 610-616.

4632 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Ohwada et al.
Figure 1. Acidity rate profiles of the fluorene cyclizations.
methyl cations in the cyclization, in accordance with the
Zucker-Hammett criteria.11,15-17 The R-(methoxycarbonyl)-
bis(p-methylphenyl)methyl cation 1e also showed a linear acidity
rate profile in the region between -10.6 and -11.7 (H0), the
slope being 0.92. This linear acidity rate dependence again
indicates the intervention of the additional proton transfer.
However, in the dimethyl case, distinct curvature of the plot
appears in the acidity region higher than -12 (Figure 1). This
can be interpreted in terms of a leveling effect, i.e., the
concentration of the dicationic species 6e is increased signifi-
cantly in the highly acidic medium, as judged from the estimated
Figure 2. B3LYP/6-31G*-optimized structures of diphenylmethyl
monocations substituted with an R-carbonyl group (R) (above, top view;
below, side view).
18
+
pKBH value (vide infra). In TFSA-SbF5 (2.5:1) a proton on
the carbonyl group of the bis(p-methylphenyl)methyl cation 1e
was observed in the NMR spectrum, supporting the formation
of the dication 6e.8 A comparison of the chemical shifts of the
ions formed in TFSA with those of the dication formed in
TFSA-SbF5 acid indicated the discrete formation of the dication
6e in TFSA at -30 °C.8 The acidity of TFSA-SbF5 (2.5:1) is
hydroxyl group in the carboxyl substituent, the species 1g has
s-cis conformation. The s-trans conformation with respect to
the hydroxyl group is higher in energy by 11.2 kcal/mol (HF/
3-21G, data not shown) than the s-cis conformation 1g. Thus,
we focus on the s-cis conformer 1g hereafter. These energy-
minimized structures (1c,1g, 1i, and 1k) show a tendency to
take bisected forms with respect to the central C-C bond. The
transition structures (ts1) were also identified at the Becke3-
LYP/6-31G* level (Figure 3) and represent those of the
conrotatory electrocyclization reactions of the diphenylmethyl
monocations to yield the fluorenes in the cases of the meth-
oxycarbonyl (1c-ts1), the carboxyl (1g-ts1), the acetyl (1i-ts1),
and the formyl (1k-ts1) substituents. The transition structures
(ts2) which represent the cyclization to give the benzofurans
were also located in the cases of the acetyl (1i-ts2) and the
formyl substituents (1k-ts2). The predicted activation energies
(B3LYP/6-31G*+ZPE) for the fluorene cyclizations are 22.7
kcal/mol (for the ester, 1c f 1c-ts1), 21.6 kcal/mol (for the
carboxyl group, 1g f 1g-ts1), 25.0 kcal/mol (for the acetyl
group, 1i f 1i-ts1), and 21.9 kcal/mol (for the formyl group,
1k f 1k-ts1), respectively (Tables 3 and 4). The distances
between the cyclizing aromatic carbon atoms (1i-ts1, 1.951 Å
< 1c-ts1, 1.963 Å < 1k-ts1, 1.966 Å < 1g-ts1, 1.972 Å) also
correspond well to an early location of the transition states along
the reaction coordinate (the earlier the transition state, the less
exothermic the reaction, as represented in terms of the predicted
activation barrier) (see also Figure 5). The activation energies
of the benzofuran cyclization are 17.0 kcal/mol (for the acetyl
group, 1i f 1i-ts2, B3LYP/6-31G* + ZPE), and 17.9 kcal/
+
stronger than -16 (H0 < -16), so the pKBH of the diphenyl-
methyl cation 1e is estimated to be around -12 to -13. This
stable formation of the dication 6e can be attributed to
conjugative stabilization by the two methyl substituents on the
phenyl groups. Such acidity rate behavior is also compatible
with the Zucker-Hammett criteria, which predict linearity when
the concentration of the active species is low.15
The fact that the reaction of the dimethyl substrate 4e was
slower than that of unsubstituted 4c despite the higher basicity
of the monocation 1e (as an oxygen base) supports the
involvement of a rate-determining cyclization process, rather
than a rate-determining protonation.19
Theoretical Evaluation of the Fluorene and Benzofuran
Cyclizations. The effects of protonation on the structures and
electrocyclization reactions were evaluated on the basis of the
ab initio calculations employing the density functional theory.20,21
Diphenylmethyl Monocations. Becke3-LYP/6-31G*-opti-
mized structures of the experimentally examined diphenylmethyl
cations bearing R-methoxycarbonyl (1c), R-carboxyl (1g),
R-acetyl (1i), and the simplest substituent R-formyl (1k), are
shown in Figure 2.20 With respect to the isomerism of the
(15) Zucker, L.; Hammett, L. P. J. Am. Chem. Soc. 1939, 61, 2791-
2798. Bonner, T. G.; Thorne, M. P.; Wilkins, J. M. J. Chem. Soc. 1955,
2351-2358. Bonner, T. G.; Barnard, M. J. Chem. Soc. 1958, 4181-4186.
(16) (a) Saito, S.; Sato, Y.; Ohwada, T.; Shudo, K. J. Am. Chem. Soc.
1994, 116, 2312-2317. (b) Ohwada, T.; Yamazaki, T.; Suzuki, T.; Saito,
S.; Shudo, K. J. Am. Chem. Soc. 1996, 118, 6220-6224.
(20) Gaussian 94, Revision E.2 Frisch, M. J.; Trucks, G. W.; Schlegel,
H. B.; Gill, P. M. W.; Johnson, B. G.; Robb, M. A.; Cheeseman, J. R.;
Keith, T.; Petersson, G. A.; Montgomery, J. A.; Raghavachari, K.;
Al-Laham, M. A.; Zakrzewski, V. G.; Ortiz, J. V.; Foresman, J. B.;
Cioslowski, J.; Stefanov, B. B.; Nanayakkara, A.; Challacombe, M.; Peng,
C. Y.; Ayala, P. Y.; Chen, W.; Wong, M. W.; Andres, J. L.; Replogle, E.
S.; Gomperts, R.; Martin, R. L.; Fox, D. J.; Binkley, J. S.; Defrees, D. J.;
Baker, J.; Stewart, J. P.; Head-Gordon, M.; Gonzalez, C.; Pople, J. A.
Gaussian, Inc.: Pittsburgh, PA, 1995.
(17) Buncel, E. Acc. Chem. Res. 1975, 8, 132-139.
(18) A referee pointed out that a straight line (slope 0.72) rather than a
curved line is obtained even in the case of 4e, if all six data are used.
Although the correlation coefficient is high (regression coefficient r )
0.986), we prefer to interpret the kinetic data in the high-acidity region in
terms of leveling-off, because of our experimental observations by NMR
spectroscopy of the facile formation of the relevant dication (ref 8).
(19) Ritchie, C. D.; Lu, S. J. Am. Chem. Soc. 1989, 111, 8542-8543.
(21) (a) Stephens, P. J.; Devlin, F. J.; Chabalowski, C. F.; Frisch, M. J.
J. Phys. Chem. 1994, 98, 11623. (b) Becke, A. D. J. Chem. Phys. 1993,
98, 1372; 1992, 96, 2155; 1992, 97, 9173.

Superacid-Catalyzed Electrocyclization
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4633
Figure 3. B3LYP/6-31G*-optimized transition structures of the fluorene cyclizations (ts1) and the benzofuran cyclizations (ts2) of the diphenylmethyl
monocations bearing an R-carbonyl group. Distances between the cyclizing atoms are shown in parentheses. Relative energy differences (in kcal/
mol) of the two modes of cyclization are shown in brackets.
Table 3. Total Energies (in hartrees) and Zero-Point Energies (in kcal/mol) of Diphenylmethyl Cation Derivatives and Transition Structures
for Cyclizationsa
structure
symm
HF/3-21G
HF/6-31G*
ZPEb
B3LYP/6-31G*
Monocations and Transition Structures
1c
1c-ts1
1g
1g-ts1
1i
1i-ts1
1i-ts2
1k
C1
C1
C1
C1
C1
C1
C1
C1
C1
C1
-721.03529 (0)
-725.09019 (0)
-725.04000 (1)
-686.05919 (0)
-686.01064 (1)
-650.21687 (0)
-650.16116 (1)
-650.17401 (1)
-611.16644 (0)
-611.11692 (1)
-611.12263 (1)
162.98
162.64
144.44c
143.73
158.68
158.57
158.59
140.24
139.76
140.17
-729.61709
-729.58042
-690.30621
-690.27078
-654.38302
-654.34296
-654.35582
-615.05392
-615.01828
-615.02538
-720.99653 (1)
-682.21700 (0)
-682.18008 (1)
-646.57484 (0)
-646.52849 (1)
-646.54081 (1)
-607.74142 (0)
-607.69836 (1)
-607.70315 (1)
1k-ts1
1k-ts2
Dications and Transition Structures
6c
6c-ts1
6g
6g-ts1
6i
6i-ts1
6k
C1
C1
C1
C1
C1
C1
C1
C1
-721.23660 (0)
-725.28782 (0)
-725.25354 (1)
-686.24278 (0)
-686.21215 (1)
-650.40846 (0)
-650.37996 (1)
-611.34357 (0)
-611.32520 (1)
170.49
170.19
151.77
151.32
166.45
166.43
148.36
147.80
-729.81890
-729.79529
-690.49531
-690.47485
-654.57939
-654.56579
-615.24227
-615.23155
-721.21044 (1)
-682.40262 (0)
-682.37998 (1)
-646.76592 (0)
-646.74669 (1)
-607.91670 (0)
-607.90473 (1)
6k-ts1
a The number of imaginary frequencies is shown in parentheses. b Unscaled (in kcal/mol). c HF/3-21G ZPE; 184.02 kcal/mol (HF/6-31G*).
Table 4. Energy Barriers (in kcal/mol) of Conrotatory Electrocyclizations of Diphenylmethyl Cation Derivatives
process
HF/3-21G
HF/6-31G*
B3LYP/6-31G*
B3LYP/6-31G*+ZPEa
Monocationic Fluorene Cyclizations
1c f 1c-ts1
1g f 1g-ts1
1i f 1i-ts1
1k f 1k-ts1
24.33
23.17
29.09
27.02
31.50
30.47
34.96
31.07
23.01
22.23
25.14
22.36
22.67
21.60
25.04
21.94
Monocationic Benzofuran Cyclizations
1i f1i-ts2
1k f 1k-ts2
21.35
24.01
26.90
27.49
17.07
17.91
16.99
17.85
Dicationic Fluorene Cyclizations
6c f 6c-ts3
6g f 6g-ts3
6i f 6i-ts3
6k f 6k-ts3
16.42
14.21
12.07
7.51
21.51
19.22
17.88
11.53
14.82
12.84
8.53
14.52 (13.99)b
12.44
8.52
6.73
6.23
a Scaled by 0.89, based on the values obtained in HF/6-31G* calculations. b The value thermally corrected to 273.15 K is shown in parentheses.
mol (for the formyl group, 1k f 1k-ts2, B3LYP/6-31G* +
ZPE). The distances between the carbonyl oxygen atom and
the cyclizing aromatic carbon atom (1i-ts2, 1.928 Å; 1k-ts2,
1.898 Å) also indicate that the transition states lie at an early
point along the reaction coordinate. The calculated activation
barriers for the monocationic fluorene cyclizations are apparently

4634 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Ohwada et al.
Figure 5. B3LYP/6-31G*-optimized transition structures of the
fluorene cyclizations (ts3) of the dications of the diphenylmethyl cations
bearing an O-protonated R-carbonyl group. Distances between the
cyclizing atoms are shown in parentheses.
Figure 4. B3LYP/6-31G*-optimized structures of the dications of the
diphenylmethyl cations bearing an O-protonated R-carbonyl group
(R′H+) (above, top view; below, side view).
groups and (an) oxygen group(s) or a methyl group or a
hydrogen atom. These structural preferences are consistent with
the theoretically predicted effects of substitutions of the ethylene
dications with an electron-donating group such as a phenyl,
hydroxyl, methyl, or methoxyl group:22 as the number of
substituents is increased, the favored geometry of the ethylene
dication (CH2+-CH2+) changes from the intrinsic perpendicular
structure toward the planar state. However, a completely planar
structure of tetra-substituted ethylene dications is unfavorable
owing to steric repulsion.22b The transition structures (ts3)
which represent the conrotatory electrocyclization of the dica-
tions to give the fluorenes were also located at the B3LYP/6-
31G* level (Figure 5). The activation energies (B3LYP/6-
31G*+ZPE) are 14.5 kcal/mol (6c f 6c-ts3), 12.4 kcal/mol
(6g f 6g-ts3), 8.5 kcal/mol (6i f 6i-ts3), and 6.2 kcal/mol
(6k f 6k-ts3), respectively (Tables 3 and 4). These barriers
are greatly attenuated as compared with those of the monoca-
tionic fluorene cyclizations, overwhelming the monocationic
benzofuran cyclization. This energy difference can explain the
experimental reality that the R-acetyl substrates 4i and 4j gave
predominantly the fluorenes over the benzofurans in TFSA. The
activation barriers for the dicationic fluorene cyclization of the
cations bearing the ester (6c) and the carboxylic acid group (6g)
are larger than those of the ketone analogues (6i and 6k): the
reaction aptitude is predicted to be in the order ester < acid ,
acetyl. This order is consistent with the experimental results.
Furthermore, the fact that the ketone substrates (4i and 4j)
reacted faster than the corresponding ester substrates (4c and
4e) is consistent with rate-determining cyclization rather than
rate-determining protonation, because the oxygen basicity of
the carbonyl group of the monocation 1 is in reverse order as
judged from the calculated proton affinities (B3LYP/6-
31G*+ZPE; 1c f 6c, 120.0 kcal/mol; 1i f 6i, 116.3 kcal/
mol). The stability of the dication 6 is relevant to the reaction
aptitude. Finally it is worthwhile to compare these predicted
activation barriers with the experimentally available thermo-
chemical parameters (activation energies Ea and enthalpy of
activation ∆Hq), obtained from the rates at different temperatures
in 55% TFSA-45% TFA (H0 ) -11.8) (Table 5).23,24 After
larger than those of the corresponding benzofuran cyclizations:
in the case of the acetyl group, the energy difference between
1i-ts1 and 1i-ts2 is 8.05 kcal/mol (B3LYP/6-31G*+ZPE); in
the case of the formyl group, that between 1k-ts1 and 1k-ts2 is
4.09 kcal/mol (B3LYP/6-31G*+ZPE). These energetic differ-
ences of the two modes of cyclization are in complete agreement
with the experimentally observed exclusive formation of the
benzofuran derivative 2b upon heating of the di-p-anisyl(p-
methoxybenzoyl)methyl cation 1b in a neutral solvent.2 Thus,
the cyclization to the benzofuran is favored, but nevertheless
has a high activation barrier, which may be compatible with
the experimental observation that the formation of the benzo-
furan from the acetyl substrate (4i) is very slow at -45 °C even
though the monocation (1i) is discretely formed. On the basis
of the calculated activation energies, the reaction aptitude for
the monocationic fluorene cyclization is predicted to be in the
order acetyl < ester < acid. However, the acetyl substrates
(4i and 4j) cyclized much faster than the corresponding ester
derivatives (4c and 4e).
Diphenylmethyl Cations Substituted with O-Protonated
Carbonyl Groups. The structures of the corresponding dica-
tions 6, formed upon O-protonation of the carbonyl group of
R-(methoxycarbonyl)- (1c), R-carboxy- (1g), R-acetyl- (1i), and
R-formyl- (1k) diphenylmethyl cations, were also optimized with
Becke3-LYP/6-31G* basis sets (Figure 4). With respect to
possible conformations due to the geometries of the O-
protonated carboxy group, the sickle-shaped conformation 6g
was chosen, because the w-shaped and u-shaped conformers
are higher in energy than the sickle 6g by 7.75 kcal/mol (HF/
3-21G, data not shown) and 16.10 kcal/mol (HF/3-21G, data
not shown), respectively. We also assumed the sickle-shaped
conformation (as in 6c) of the O-protonated methoxycarbonyl
group in the case of the dication bearing an O-protonated ester
group. We focused on the s-cis conformations of the O-
protonated acetyl group of 6i and the O-protonated formyl group
of 6k, because the s-trans counterparts are less stable (1.33 kcal/
mol over the s-cis 6i; 1.81 kcal/mol over the s-cis 6k, HF/3-
21G, data not shown). The dications (6c, 6g, 6i, and 6k) can
be regarded as ethylene dications substituted with two aryl
(22) (a) Lammertsma, K.; Barzaghi, M.; Olah, G. A.; Pople, J. A.; Kos,
A. J.; Schleyer, P. v. R. J. Am. Chem. Soc. 1983, 105, 5252. (b) Frenking,
G. J. Am. Chem. Soc. 1991, 113, 2476-2481.

Superacid-Catalyzed Electrocyclization
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4635
Table 5. Thermochemical Data for Acid-Catalyzed Cyclizations of 4d and 4e
rate constants 105k (s-1) at T (°C)
Ea (kcal/mol)b
∆Gq (kcal/mol)c
∆Hq (kcal/mol)d
∆Sq (eu)e
a
-H0
-10
0
10
4d
4e
4e
11.8
11.8
13.2
0.53
1.83
14.0
1.39
4.55
37.2
3.21
10.02
75.0
13.3
12.7
12.5
22.0
21.4
20.3
12.8
12.2
11.9
-33.8
-33.6
-30.5
a Corrected values of acidity function of the media (Table 7). b Errors (0.3 kcal/mol. c Errors (0.5 kcal/mol. d Errors (0.3 kcal/mol. e Errors
(1.0 eu, at 273.15 K (0 °C).
Table 6. Natural Bond Orbital Populations (Becke3-LYP/6-31G*)
of Ion Precursors and Transition Structures
thermal correction to 273.15 K, the predicted barrier of the ester
substrates (6c f 6c-ts3) is 14.0 kcal/mol.
aromatic ringb
R-carbon center substituent
total
The magnitudes of the experimental enthalpies of activation
of the ester substrates (∆Hq, for 4d, 12.2 kcal/mol and for 4e,
12.8 kcal/mol) agree well with the predicted activation energy
of the dicationic fluorene cyclization (6c f 6c-ts3) rather than
with that of the monocationic fluorene cyclization (1c f 1c-
ts1).24 This coincidence also supports the intervention of
dicationic electrocyclization in the fluorene cyclization of the
diphenylmethyl cations.
1c
+0.347, +0.354
(+0.702)
+0.163
-0.057
+0.151
-0.070
+0.176
-0.046
+0.135
+0.086
+0.126
+0.079
+0.143
+0.090
+1.000
1c-ts1 +0.496, +0.474
(+0.970)
1g
+1.000
+1.000
+1.000
+1.000
+1.000
+0.364, +0.359
(+0.723)
1g-ts1 +0.505, +0.486
(+0.991)
1i
+0.316, +0.364
(+0.680)
Charge Delocalization. The charge distributions of the
energy minima (1c, 1g, 1i, 1k, 6c, 6g, 6i, and 6k) and transition
structures (ts1, ts2, and ts3) of ions, based on natural bond
orbital (NBO) population analysis are shown in Table 6.25 The
total aromatic ring charge was obtained by summing the NBO
population net charges associated with the carbon atoms and
the hydrogen atoms belonging to the benzene ring. The total
substituent charge was obtained by summing the NBO charges
of the carbonyl substituent. In terms of aromatic ring charges
(the summation of the two aromatic ring charges of the dications
6 is +1.04 to +1.27; that of the monocations 1 is +0.68 to
+0.75), the dications 6 (6c, 6g, 6i, and 6k) can be regarded as
carbodications with positive charge substantially delocalized
over the aromatic rings, which suggests enhanced pentadienium
character of the ground-state dications 6 as compared with the
corresponding monocations 1.26 More positive charges are
localized into the aromatic rings of the transition structures for
both mono- (ts1) and dicationic (ts3) fluorene cyclization as
compared with the starting minima structures, suggesting
enforced pentadienium character in the transition structures. A
similar increase of the positive charge is also found in the
cyclizing aromatic ring of the transition structures (ts2) in the
case of the benzofuran cyclization. In the transition structures
1i-ts1 +0.508, +0.448
(+0.956)
1i-ts2 +0.578,a +0.190
-0.008
+0.121
+0.240
+0.123
+1.000
+1.000
1k
+0.347, +0.408
(+0.756)
1k-ts1 +0.515, +0.469
(+0.985)
-0.067
+0.082
+1.000
1k-ts2 +0.607,a +0.209
-0.036
+0.031
+0.221
+0.929
+1.000
+2.000
6c
+0.466, +0.575
(+1.041)
6c-ts3 +0.596, +0.677
(+1.272)
6g
-0.130
+0.002
-0.143
+0.033
-0.096
-0.029
-0.113
+0.858
+0.893
+0.811
+0.870
+0.726
+0.759
+0.675
+2.000
+2.000
+2.000
+2.000
+2.000
+2.000
+2.000
+0.619, +0.485
(+1.104)
6g-ts3 +0.711, +0.621
(+1.332)
6i
+0.521, +0.576
(+1.097)
6i-ts3 +0.707, +0.662
(+1.370)
6k
+0.639, +0.630
(+1.269)
6k-ts3 +0.709, +0.729
(+1.438)
a The charge of the cyclizing aromatic rings. b The summation of
the charges of the two aromatic rings is shown in parentheses.
(23) The rates were measured in 55% TFSA-45% TFA (H0 ) -11.8)
because leveling-off of the reaction rates of 4e was found in TFSA (see
Table 2 and Figure 1). A preliminary report of the thermochemical
parameters of the reaction of 4e in TFSA (ref 8b) gave values in reasonable
agreement with those obtained in this work. The activation parameters of
4e in a higher-acidity region (in TFSA) are similar to those obtained at H0
) -11.8. See also ref 18.
(24) The observed rate constant kobs is related to kr/Ke where kr is the
rate constant for cyclization and Ke is the equilibrium constant associated
with the protonation of the monocation (i.e., the ionization constant of the
carbodication). If kr and Ke are temperature-dependent in different ways,
of the monocationic fluorene (ts1) and benzofuran (ts2) cy-
clizations, the R-carbon atom has a fractional negative charge,
while the carbonyl carbon and the aromatic ipso carbons bear
large positive charges. In the cases of the transition structures
(ts3) of the dicationic fluorene cyclizations, the R-carbon atom
has a definite negative charge, and therefore, charge alternation
is encouraged, which leads to internal Coulombic stabilization
of the ions.27 This charge alternation mechanism provides at
least a partial rationale for the stabilization of the transition
structures of the dicationic fluorene cyclization (ts3), resulting
in the reduction of the activation energies.
kobs will be affected by temperature in a complex manner. It has been shown
that the H0 is temperature-dependent in a H2SO4-H2O solvent system
(see: Johnson, C. D.; Katritzky, A. R.; Shapiro, S. A. J. Am. Chem. Soc.
+
1969, 91, 6654-6662). Therefore, the validity of the pKBH of bases
(indicator and the relevant monocations) and H0 values at different
+
temperatures may be questioned. However, the changes of H0 and pKBH
should essentially cancel each other out, and the protonation states of the
bases should not be significantly affected by the temperature. The excellent
linear relationships (regression coefficient r > 0.99) obtained in the
estimation of activation parameters support the postulation that the changes
of H0 and pKBH are canceled out in the TFSA-TFA acid system and,
thus, that kobs is affected by kr and not Ke at the different temperatures
(Table 5).
Conclusion
The chemical and kinetic characterizations of the fluorene
cyclization of diphenylmethyl cations bearing an R-carbonyl
group described herein provide conclusive support for the idea
that an additional proton transfer is required to induce efficient
electrocyclization of the diphenylmethyl cations.9 The benzo-
+
(25) NBO Version 3.1: Glendening, E. D.; Reed, A. E.; Carpenter, J.
E.; Weinhold, F. A. Reed, E.; Curtiss, L. A.; Weinhold, F. Chem. ReV.
1988, 88, 899-926.
(27) (a) Klein, J. Tetrahedron 1983, 39, 2733. (b) Klein, J. Tetrahedron
1983, 48, 2928. (c) Wiberg, K. J. Am. Chem. Soc. 1990, 112, 4177-4182.
(26) Ohwada, T.; Shudo, K. J. Am. Chem. Soc. 1989, 111, 34-40.

4636 J. Am. Chem. Soc., Vol. 120, No. 19, 1998
Ohwada et al.
Acid-Catalyzed Fluorene Cyclization Reactions in TFSA. Gen-
eral Procedure. The diphenylmethanol (4) (1 mmol) was added to
precooled TFSA ((44 mL, 500 mmol) at the specified temperature (-45
or 0 °C) with stirring. The resultant solution was stirred for the
specified time. The whole was poured into ice-water and extracted
with CH2Cl2 (150 mL). The organic layer was washed with brine (100
mL), dried over Na2SO4, filtered, and evaporated to give a residue,
which was flash-chromatographed (CH2Cl2-n-hexane 2:3) to give the
fluorene compounds (3).
furan cyclization was much slower than the fluorene cyclization
at -45 °C in TFSA in the case of the acetyl substrates. This
can be at least partially rationalized in terms of the predicted
activation barriers of the dicationic fluorene cyclization, which
are smaller than those of the monocationic benzofuran or
fluorene cyclization reactions. We hypothesize a similar
contribution of the dicationic species in the alternative cycliza-
tion of the cation 1a to the phenthrol9 and probably also in the
superacid-catalyzed cyclization of pinacols to phenanthrenes.28
Acid- Catalyzed Reaction of r-(Methoxycarbonyl)bisphenyl-
methanol (4c). 9-(Methoxycarbonyl)fluorene 3c: mp 63.5-64.5 °C
Experimental Section
1
(recrystallized from n-hexane); H NMR 7.76 (2H, d, J ) 7.5 Hz),
Materials. R-(Methoxycarbonyl)bisphenylmethanol (4c) and R-(meth-
oxycarbonyl)bis(p-methylphenyl)methanol (4e) were prepared as de-
scribed previously.8a,b Other R-(methoxycarbonyl)bisarylmethanols
were prepared similarly by esterification (with catalytic sulfuric acid
in methanol at ambient temperature) of the corresponding R-carboxy-
bisarylmethanols, which were prepared by utilizing the benzilic acid
rearrangement reactions of the substituted benzils.
R-(Methoxycarbonyl)bis(p-bromophenyl)methanol (4d): mp 73.5-
74.0 °C (recrystallized from n-hexane); 1H NMR 7.47 (4H, d, J ) 8.4
Hz), 7.27 (4H, d, J ) 8.4 Hz), 4.21 (1H, s), 3.86 (3H, s). Anal. Calcd
for C15H12Br2O3: C, 45.03; H, 3.02; N, 0.00. Found: C, 44.83; H,
2.92; N, 0.00.
R-(Methoxycarbonyl)bis(p-fluorophenyl)methanol (4f): colorless oil
(molecular distillation 190 °C/2 mmHg); 1H NMR 7.37 (4H, ddd, J )
7.0, 5.1, 2.2 Hz), 7.02 (4H, t, J ) 8.8 Hz), 4.24 (1H, s), 3.86 (3H, s).
Anal. Calcd for C15H12F2O3: C, 64.75; H, 4.35; N, 0.00. Found: C,
64.45; H, 4.24; N, 0.00.
7.66 (2H, dd, J ) 0.74, 7.5 Hz), 7.43 (2H, ddd, J ) 0.74, 7.5, 7.5 Hz),
7.34 (2H, ddd, J ) 1.3, 7.5, 7.5 Hz), 4.88 (1H, s), 3.76 (3H,s).
Acid-Catalyzed Reaction of r-(Methoxycarbonyl)bis(p-bromo-
phenyl)methanol (4d). 9-(Methoxycarbonyl)-3,6-dibromofluorene
1
3d: mp 182-183 °C (recrystallized from n-hexane); H NMR 7.84
(2H, d, J ) 1.5 Hz), 7.54 (2H, d, J ) 8.4 Hz), 7.49 (2H, dd, 1.8, 8.1
Hz), 4.76 (1H, s), 3.98 (3H, s); HRMS calcd for C15H10Br2O2 379.905,
found 379.907.
Acid-Catalyzed Reaction of r-(Methoxycarbonyl)bis(p-meth-
ylphenyl)methanol (4e). 9-(Methoxycarbonyl)-3,6-dimethylfluorene
1
3e: mp 104.0 °C (recrystallized from n-hexane); H NMR 7.54 (2H,
s), 7.50 (2H, d, J ) 7.7 Hz), 7.13 (2H, d, J ) 8.1 Hz), 4.78 (1H, s),
3.70 (3H, s), 2.44 (6H, s). Anal. Calcd for C17H16O2: C, 80.92; H,
6.39; N, 0.00. Found: C, 80.63; H, 6.19; N, 0.00.
Acid-Catalyzed Reaction of r-(Methoxycarbonyl)bis(p-fluoro-
phenyl)methanol (4f). 9-(Methoxycarbonyl)-3,6-difluorofluorene 3f:
1
mp 134-135 °C (recrystallized from n-hexane); H NMR 7.61 (2H,
R-Acetylbis(phenyl)methanol (4i) and R-acetylbis(p-methylphenyl)-
methanol (4j) were prepared by the acid-catalyzed desilylation of the
corresponding trimethylsiloxy derivatives, which were prepared by
addition of the anion of [1-(diethoxyphosphinyl)ethoxy]trimethylsilane
to benzophenone or 4,4-dimethylbenzophenone:14 that is, to a stirred
solution of 12 mL of diisopropylamine in 60 mL of dry THF at -78
°C was added 52 mL of n-butyllithium (1.6 M in hexane). The mixture
was stirred at -78 °C for 20 min, and then the cooling bath was
replaced with an ice-water bath. The mixture was stirred for 15 min
and then cooled to -78 °C, and 27.1 g of [1-(diethoxyphosphinyl)-
ethoxy]trimethylsilane (prepared from trimethylsilyl chloride, triethyl
phosphite, and acetaldehyde (Merck, 99%)) was added dropwise over
10 min at -78 °C. The mixture was stirred at -78 °C for 75 min, and
then a solution of benzophenone (13.46 g) in 10 mL of THF was added
over 10 min. The resultant yellow solution was stirred and warmed to
room temperature and then further stirred at room temperature for 37
h. It was poured into 400 mL of water, and the mixture was extracted
with methylene chloride. The organic layer was washed with 1 N
aqueous hydrogen chloride, followed by brine, and dried over sodium
sulfate. Evaporation of the solvent gave 7.40 g of the residue, which
was flash-chromatographed (ethyl acetate-n-hexane 2:98) to give 5.66
g of 1,1-diphenyl-1-(trimethylsiloxy)propan-2-one. A solution of this
trimethylsiloxy derivative in 60 mL of methanol and 0.2 mL of
concentrated hydrochloric acid was heated at 43 °C for 30 min.
Evaporation of methanol gave the residue, which was extracted with
methylene chloride. The organic layer was washed with brine and dried
over sodium sulfate. Evaporation of the solvent gave the residue (4.296
g), which was flash-chromatographed (ethyl acetate-n-hexane 1:12)
to give 2.77 g of R-acetylbisphenylmethanol 4i. For 4i: mp 66.0-
66.5 °C (colorless flakes, recrystallized from n-hexane); 1H NMR 7.363
(10H, m), 4.844 (1H, s, OH), 2.258 (3H, s). Anal. Calcd for
C15H14O2: C, 79.62; H, 6.24; N, 0.00. Found: C, 79.38; H, 6.24; N,
0.00.
dd, J ) 5.0, 8.3 Hz), 7.37 (2H, dd, J ) 8.6, 2.4 Hz), 7.05 (2H, ddd, J
) 10.6, 8.6, 2.4 Hz), 4.80 (1H, s), 3.76 (3H, s); HRMS calcd for
C15H10F2O2 260.065, found 260.065.
Acid-Catalyzed Reaction of r-Acetylbisphenylmethanol (4i). A
solution of the diphenylmethanol (4i) (1 mmol, 227 mg) in 1.5 mL of
methanol-free methylene chloride was added to precooled TFSA ((44
mL, 500 mmol) at -45 °C (acetonitrile-dry ice) with stirring over 1
min. The resultant solution was stirred for 15 min. The whole was
poured into ice-water (300 mL) and extracted with CH2Cl2 (300 mL).
The organic layer was washed with brine (100 mL), dried over Na2-
SO4, filtered, and evaporated to give a residue (232 mg), which was
flash-chromatographed (CH2Cl2-n-hexane 1:1) to give 9-acetylfluorene
3i (180 mg, 86%), together with the fluorene dimer 7 (4 mg, 1.8%).
9-Acetylfluorene 3i: mp 73.5-74.0 °C (colorless cubes, recrystallized
1
from n-hexane); H NMR 7.828 (2H, d, J ) 7.5 Hz), 7.809 (2H, d, J
) 7.5 Hz), 7.467 (2H, dd, J ) 7.5, 7.5 Hz), 7.356 (2H, dd, J ) 7.5,
7.5 Hz), 4.800 (1H, s), 1.617 (3H, s). Anal. Calcd for C15H12O: C,
86.51; H, 5.80; N, 0.00. Found: C, 86.80; H, 5.89; N, 0.00. The
fluorene dimer 7: 1H NMR 7.826 (2H, d, J ) 7.7 Hz), 7.481-7.110
(15H, m), 4.847 (1H, s), 4.799 (1H, s), 2.271 (3H, s), 1.615 (3H, s);
MS m/e 416 (M+).
Acid-Catalyzed Reaction of r-Acetylbis(p-methylphenyl)metha-
nol (4j). A solution of the diphenylmethanol (4j) (1 mmol, 260 mg)
in 1.5 mL of methanol-free methylene chloride was added to precooled
TFSA ((44 mL, 500 mmol) at -45 °C (acetonitrile-dry ice) with
stirring over 1 min. The resultant solution was stirred for 15 min. The
whole was poured into ice water (300 mL) and extracted with CH2Cl2
(300 mL). The organic layer was washed with brine (100 mL), dried
over Na2SO4, filtered, and evaporated to give a residue (246 mg), which
was flash-chromatographed (CH2Cl2-n-hexane 2:3) to give 9-acetyl-
3,6-dimethylfluorene 3j (158 mg, 65%), together with 3-(p-methyl-
phenyl)-2,6-dimethylbenzofuran 2j (15.8 mg, 6.5%). 9-Acetyl-3,6-
dimethylfluorene 3j: mp 100.5-101.5 °C (colorless plates, recrystallized
R-Acetylbis(p-methylphenyl)methanol 4j was prepared similarly
from 4,4′-dimethylbenzophenone in 62% (total yield). The crude
alcohol was purified by flash column chromatography (ethyl acetate-
n-hexane 1:15) to give a viscous, colorless oil of 4j: 1H NMR 7.251
(4H, d-like, J ) 8.43 Hz), 7.174 (4H, d, J ) 8.06 Hz), 4.772 (1H, s,
OH), 2.362 (6H, s), 2.249 (3H, s); HRMS calcd for C17H18O2 254.1306,
found 254.1320.
(28) Olah, G. A.; Klumpp, D. A.; Wang, N. Q. Synthesis 1996, 321-
323.
1
from n-hexane); H NMR 7.611 (2H, s), 7.361 (2H, d, J ) 7.69 Hz),

Superacid-Catalyzed Electrocyclization
J. Am. Chem. Soc., Vol. 120, No. 19, 1998 4637
Table 7. Corrected Values of Acidity Function of the Acid Media
in the Presence of Water
to give R-(methoxycarbonyl)-R-methoxybis(p-bromophenyl)methane as
a colorless oil (78.3 mg, 94% yield): 1H NMR 7.46 (4H, d, J ) 8.8
Hz), 7.29 (4H, d, J ) 8.8 Hz), 3.77 (3H, s), 3.16 (3H, s).
c
d
acida
acid/waterb uncorrd H0 corrd H0
Kinetic Measurement. Acid mixtures of TFSA and TFA were
prepared in a polyethylene glovebox (AtmosBag, Aldrich) under an
argon atmosphere. A precooled acid mixture (-45 °C, dry ice-
acetonitrile) (1000 equiv (50 mmol), typically, 4.4 mL) was added to
a weighed diphenylmethanol (1, 0.05 mmol), in the dry glovebox. A
good solution was obtained, and a portion (ca. 0.6 mL) of it was
transferred to an NMR tube, which contained a small quantity of
methanol-free methylene chloride as an internal standard for signal
integrations. The spectra were recorded at a specified temperature.
The NMR probe temperature was controlled to within (0.1 °C using
a variable-temperature apparatus, NM-ALTAS/L and NM-AVT1A
(JEOL, Japan). The disappearance of the starting material was
monitored in terms of the integrated signal intensity with a GSX 500-
MHz NMR spectrometer. Errors of rates, derived from the signal
integration, are (2%.
Computational Methods. Computations were carried out with the
GAUSSIAN 94 ab initio programs. Geometries were initially optimized
without any symmetry restriction with the split valence HF/3-21G basis
set, and with the heavy atom d-polarized HF/6-31G* basis set.29 The
geometries were further optimized with a hybrid density-functional
theory method, Becke3-LYP (B3LYP) with the HF/6-31G* optimized
geometries.21 All the structures of the reactants shown in Figures 2
and 4 were minima, and the transition structures (Figures 3 and 5),
which represent cyclization processes, were characterized by frequency
calculations at the HF/6-31G* level. Zero-point vibrational energy
(ZPE) corrections were made on the basis of the values scaled by 0.89
obtained in the HF/6-31G* calculations.30 For comparison with
experimental data (∆Hq) (Table 5), the activation barrier (6c f 6c-
ts3) was thermally corrected to 273.15 K.
100% TFSA
511 equiv
-14.12
-12.93
-12.51
-12.10
-12.06
-10.69
-13.18
-12.59
-12.18
-11.80
-11.76
-10.66
84.2% TFSA-15.8% TFA 503 equiv
71.2% TFSA-28.8% TFA 497 equiv
57.2% TFSA-42.8% TFA 492 equiv
55.8% TFSA-44.2% TFA 490 equiv
24.3% TFSA-75.7% TFA 516 equiv
a Weight percent of the composite acids. b Mole ratios of acid and
water. c Reference 10. d See also ref 11.
7.153 (2H, d, 8.43 Hz), 4.697 (1H, s), 2.471 (6H, s), 1.595 (3H, s).
Anal. Calcd for C17H16O: C, 86.40; H, 6.83; N, 0.00. Found: C,
86.46; H, 7.11; N, 0.00. 3-(p-Methylphenyl)-2,6-dimethylbenzofuran
1
2j: mp 77-80 °C; H NMR 7.434 (1H, d, J ) 8.06 Hz), 7.392 (2H,
d, J ) 8.06 Hz), 7.285 (2H, d, J ) 8.06 Hz), 7.247 (1H, s), 7.037 (1H,
d, J ) 7.32 Hz), 2.505 (3H, s), 2.467 (3H, s), 2.417 (3H, s); HRMS
calcd for C17H16O 236.1201, found 236.1201.
The Acid Systems. In strong acid of more than H0 ) -10 (see the
text), the alcohols 4 is ionized almost completely to yield one equivalent
amount of water and the diphenylmethyl cation 1 and the resultant water
behaves as an oxygen base, thus reducing the acidity of the reaction
medium. The carbonyl group of the diphenylmethyl cations 1 may
also behave as an oxygen base, particularly in the region of high acidity.
We therefore estimated the lowering of the acidity of the reaction
medium by measuring the acidity of TFSA-TFA (500 equiv with
respect to water) in the presence of water at acidity levels stronger
than H0 ≈ -9 (Table 7).10,11 Naturally, addition of water to 100%
TFSA caused the largest lowering of the acidity, and the effect of water
was attenuated as the acidity of the medium decreased upon the addition
of TFA.
Ionization of r-(Methoxycarbonyl)bis(p-bromophenyl)methanol
(4d). A solution of R-(methoxycarbonyl)bis(p-bromophenyl)methanol
(4d) (80.3 mg, 0.2 mmol) in 18 mL (1000 equiv) of 40.2% w/w TFSA-
59.8% w/w TFA was prepared at -45 °C. The solution was poured
into methanol (300 mL), followed by the addition of methylene chloride
(300 mL). The organic layer was washed with three portions of water
(200 mL) and dried over sodium sulfate. The solvent was evaporated
JA980090W
(29) Hehre, W. J.; Radom, L.; Schleyer, P. v. R.; Pople, J. A. Ab Initio
Molecular Orbital Theory; Wiley and Sons: New York, 1986.
(30) (a) Pople, J. A.; Scott, A. P.: Wong, M. W.; Radom, L. Isr. J.
Chem. 1993, 33, 345-350. (b) Foresman, J. B.; Frisch, Æ. Exploring
Chemistry with Electronic Structure Methods, 2nd ed.; Gaussian Inc.:
Pittsburgh, PA, 1993.