Investigation of substituent effects on the selectivity of 4π-electrocyclization of 1,3-diarylallylic cations for the formation of highly substituted indenes

Article abstract of DOI:10.1021/jo100275q

(Figure Presented) Differentially substituted 1,3-diaryl-substituted allylic cations generated by ionization of the corresponding allylic alcohols in the presence of a Lewis acid undergo chemoselective and regioselective electrocyclization reactions to generate 1-aryl-1H-indenes. Electrocyclization only occurs for allylic cations bearing a 2-substituent, with 2-ester and 2-alkyl substituents both tolerated. In general, the presence of electron-withdrawing substituents deactivates the ring and disfavors cyclization. In contrast, the selectivity of cyclization of systems containing electron-donating substituents depends on the nature and position of the electron-donating group. Electron-donating substituents at the meta position particularly favor cyclization. There was no obvious correlation of cyclization selectivity with calculated electron densities as has been suggested for electrophilic aromatic substitution reactions. However, the calculated selectivities determined by a gas-phase (B3LYP/6-31G* + ZPVE) comparison of the relative rates of cyclization were in remarkably good agreement with the observed selectivities. Calculated transition-state structures for cyclization are consistent with a cationic π4_a conrotatory electrocyclization mechanism. In some cases involving more electron-deficient systems, the initially formed 1H-indene underwent subsequent alkene isomerization to the 3H-indene. In one example, an unusual dimerization reaction occurred to give a cyclopenta[a]indene via an unusual formal cationic 2π+2π cycloaddition of the allylic cation with the intermediate indene.

Full text of DOI:10.1021/jo100275q

pubs.acs.org/joc
Investigation of Substituent Effects on the Selectivity of
4π-Electrocyclization of 1,3-Diarylallylic Cations for the
Formation of Highly Substituted Indenes
Chris D. Smith, Gregory Rosocha, Leo Mui, and Robert A. Batey*
Department of Chemistry, University of Toronto, 80 St. George Street, Toronto, Ontario, Canada M5S 3H6
rbatey@chem.utoronto.ca
Received March 12, 2010
Differentially substituted 1,3-diaryl-substituted allylic cations generated by ionization of the
corresponding allylic alcohols in the presence of a Lewis acid undergo chemoselective and
regioselective electrocyclization reactions to generate 1-aryl-1H-indenes. Electrocyclization only
occurs for allylic cations bearing a 2-substituent, with 2-ester and 2-alkyl substituents both tolerated.
In general, the presence of electron-withdrawing substituents deactivates the ring and disfavors
cyclization. In contrast, the selectivity of cyclization of systems containing electron-donating sub-
stituents depends on the nature and position of the electron-donating group. Electron-donating
substituents at the meta position particularly favor cyclization. There was no obvious correlation
of cyclization selectivity with calculated electron densities as has been suggested for electrophilic
aromatic substitution reactions. However, the calculated selectivities determined by a gas-phase
(B3LYP/6-31G* þ ZPVE) comparison of the relative rates of cyclization were in remarkably good
agreement with the observed selectivities. Calculated transition-state structures for cyclization are
consistent with a cationic π4a conrotatory electrocyclization mechanism. In some cases involving
more electron-deficient systems, the initially formed 1H-indene underwent subsequent alkene
isomerization to the 3H-indene. In one example, an unusual dimerization reaction occurred to give
a cyclopenta[a]indene via an unusual formal cationic 2πþ2π cycloaddition of the allylic cation with
the intermediate indene.
Introduction
via phenyl-substituted allylic cations 23 to give indenes 5 (eq 1).
1
The allylic cations 2 can be generated and observed (by H
The cyclization of phenyl-substituted allylic cations is a
well-established method for the synthesis of indenes.1 Typi-
cally phenyl-substituted allylic alcohols 12 react in strongly
acidic media (i.e., FSO3H, FSO3H-SbF5-SO2ClF, H2SO4,etc.)
NMR) at low temperatures (-70 °C) in super acid media and
upon warming cyclize to give indenes 5.3a,4,5 Conventional
Lewis acids have also been used for the cyclizations of 1 into 5,
such as boron trifluoride etherate, aluminum trichloride, etc.6-8
The allylic alcohol precursors utilized have typically been
1-phenylallyl alcohols, 1-arylallyl alcohols, 1,3-diphenylallyl
alcohols, 1,1,3-triphenylallyl alcohols, and tetra- or pentaphenyl-
substituted allylic alcohols.
(1) (a) For review on indene formation, see: Ivchenko, N. B.; Ivchenko,
P. V.; Nifant’ev, I. E. Russ. J. Org. Chem. 2000, 36, 609–637. For reviews on
the synthesis of indenes and their use in transition metal chemistry, see:
(b) Enders, M.; Baker, R. W. Curr. Org. Chem. 2006, 10, 937–953.
(c) Halterman, R. L. Chem. Rev. 1992, 92, 965–994.
(2) Phenyl-substituted allylic alcohols are believed to be intermediates in
the cyclodehydration of phenyl-substituted diols: (a) Blum-Bergmann, O.
Ber. 1932, 65, 109–122. (b) Blum-Bergmann, O. J. Chem. Soc. 1935, 1020–
1022.
(3) (a) Olah, G. A.; Asensio, G.; Mayr, H. J. Org. Chem. 1978, 43, 1518–
1520. (b) Pittman, C. U., Jr.; Miller, W. G. J. Org. Chem. 1974, 39, 1955–
1956. (c) Bergmann, F. J. Org. Chem. 1941, 6, 543–549. (d) Koelsch, C. F.
J. Am. Chem. Soc. 1932, 54, 3384–3389.
(4) (a) Pittman, C. U., Jr.; Miller, W. G. J. Am. Chem. Soc. 1973, 95,
2947–2956. (b) Dytnerski, D.; Ranganayakulu, K.; Singh, B. P.; Sorensen,
T. S. J. Org. Chem. 1983, 48, 309–315. (c) Dytnerski, D.; Ranganayakulu, K.;
Singh, B. P.; Sorensen, T. S. Can. J. Chem. 1982, 60, 2993–3004. (d) Deno,
N. C.; Pittman, C. U., Jr.; Turner, J. O. J. Am. Chem. Soc. 1965, 87, 2153–
2157.
(5) For the isolation of the 1,3-diphenylallyl cation as a stable tetrafluoro-
borate salt, see: Hafner, K.; Pelster, H. Angew. Chem. 1961, 73, 342.
4716 J. Org. Chem. 2010, 75, 4716–4727
Published on Web 06/22/2010
DOI: 10.1021/jo100275q
r
2010 American Chemical Society

Smith et al.
JOCArticle
SCHEME 1. Selectivity of Cyclization of 1,3-Diarylallylic
Cations 7 into Indenes 8 and 9
The cyclization of 3 can be described as a 4π-conrotatory
electrocyclization to give indanyl cation 4, which is followed
by elimination of a proton to give 5.2b,3a,b,4a,6c,8,9 This
mechanism is reminiscent of the Nazarov reaction, which
occurs via cationic conrotatory 4π-electrocyclization and
proton loss to give cyclopentenones.10,11 The cyclization of
3 into 4 can also be envisaged as an intramolecular electro-
philic aromatic substitution (Friedel-Crafts alkylation) re-
action. The allylic cation must first rearrange from the
unreactive extended conformation 2 to adopt the conforma-
tion 3 suitable for cyclization. To promote reaction via this
conformation it is necessary for the precursor alcohols 1 to
contain at least one other substituent in addition to the
phenyl group onto which cyclization occurs.7
While the electrocyclization of 1,3-diphenylallylic cations
2/3 (R3 or R4 = Ph) is known, examples of cyclization of
differentially substituted 1,3-diarylallylic cations 7 have not
been rigorously explored (Scheme 1).12 Such reactions pose
an interesting chemoselectivity issue11,13 since cyclization of
the unsymmetrical cation 7 can occur from either of the
aromatic rings, leading to differentially substituted indenes 8
and 9. We were particularly interested in discovering the
factors that govern the chemoselectivity of electrocyclization
of 6 to either 8 or 9 and whether they could be rationalized in
terms of the well-known substituent effects for electrophilic
aromatic substitution or whether other factors would play a
role given the electrocyclic nature of the cyclization. We now
describe experimental and computational studies on the
conversion of 6 into functionalized indenes 8 and 9 and show
how the chemoselectivity of electrocyclization of the inter-
mediate cations 7 is dependent upon both the position and
electronic nature of the substituents.
Results and Discussion
Synthesis of 1,3-diarylallylic alcohol precursors 6 (R2 =
Me or COOEt) was achieved using either Knoevenagel or
aldol-based approaches from commercially available sub-
stituted benzaldehydes 10 (Scheme 2). Knoevenagel conden-
sation14 of 10 with ethyl benzoylacetate 11 (Ar = Ph, R2 =
COOEt) stereoselectively provided ethyl 3-aryl-2-benzoyl-
propenoates 12 (Ar=Ph, R2=COOEt) as the (E)-isomers.15
This method was also extended to aldol reactions of propio-
phenone 11 (Ar=Ph, R2 =Me) to stereoselectively prepare
(E)-3-aryl-2-methyl-1-phenylpropenones 12 (Ar=Ph, R2 =
Me). For the synthesis of 1,3-diaryl-2-methylpropenones 12
(Ar ¼ Ph, R2=Me) a three-step aldol protocol was required.
TiCl4-mediated aldol reaction16 between 10 and an appro-
priately substituted propiophenone 11 (Ar ¼ Ph, R2 = Me)
afforded syn-β-hydroxy ketones with high stereoselectivity.
These were then immediately converted to the corresponding
mesylates, which underwent elimination17 to give (E)-2-
propenones 12 (Ar ¼ Ph, R2 =Me). Selective 1,2-reduction
(6) See, for example: (a) Zhou, X.; Zhang, H.; Xie, X.; Li, Y. J. Org.
Chem. 2008, 73, 3958–3960. (b) Jeong, I. H.; Park, Y. S.; Kim, M. S.; Oh, S. T.
Bull. Korean Chem. Soc. 2001, 22, 1173–1174. (c) Jeong, I. H.; Park, Y. S.;
Kim, M. S.; Song, Y. S. J. Fluorine Chem. 2003, 120, 195–209.
(7) Solid-state thermolysis of magnesium alkoxides has also been re-
ported to lead to efficient and regioselective indene formation; see: Tolbert,
L. M. J. Org. Chem. 1979, 44, 4584–4588.
(8) (a) For BF3 OEt2-promoted cyclodehydration reactions of vinyl
3
sulfide benzyl alcohols, see: Singh, G.; Ila, H.; Junjappa, H. Synthesis
1986, 744–747. (b) For BF3 OEt2-catalyzed cyclizations of iodinated allylic
3
alcohols, see: Zhou, X.; Zhang, H.; Xie, X.; Li, Y. J. Org. Chem. 2008, 73,
3958–3960.
(9) For mild acid-promoted cyclization of dienostrol and related dienes to
the corresponding indenes, see: Hausmann, W.; Smith, A. E. W. J. Chem.
Soc. 1949, 1030–1032.
(10) For reviews of the Nazarov reaction, see: (a) Tius, M. A. Eur. J. Org.
Chem. 2005, 11, 2193–2206. (b) Frontier, A. J.; Collison, C. Tetrahedron
2005, 61, 7577–7606. (c) Pellissier, H. Tetrahedron 2005, 61, 6479–6517.
(d) Habermas, K. L.; Denmark, S. E.; Jones, T. K. Org. React. (N.Y.) 1994,
45, 1–158.
(11) The Nazarov reaction has also been used for the synthesis of indenes.
See, for example: Liang, G.; Xu, Y.; Seiple, I. B.; Trauner, D. J. Am. Chem.
Soc. 2006, 128, 11022–11023.
(12) (a) Wang, J.; Zhang, L.; Jing, Y.; Huang, W.; Zhou, X. Tetrahedron
Lett. 2009, 50, 4978–4982. (b) There is a single reported example of a
cyclization of methyl (2Z)-2-[hydroxy(4-chlorophenyl)methyl]-3-phenyla-
crylate into methyl 6-chloro-1-phenyl-1H-indene-2-carboxylate using Mon-
tmorillonite K10 clay under microwave conditions. However, no data were
reported for the product; see: Shanmugam, P.; Rajasingh, P. Chem. Lett.
2005, 34, 1494–1495.
(14) (a) Antonioletti, R.; Bovicelli, P.; Malancona, S. Tetrahedron 2002,
58, 589–596. (b) Horning, E. C.; Koo, J.; Fish, M. S.; Walker, G. N. Organic
Synthesis; Wiley: New York, 1963; Collect. Vol. IV, pp 408-409. (c) Allen,
C. F. H.; Spangler, F. W.; Shriner, R. L.; Neumann, F. W. Organic Synthesis;
Wiley: New York, 1955; Collect. Vol. III, pp 377-379.
(13) In general, there have been relatively few studies of substituent effects
for electrocyclic reactions. See, for example: (a) He, W.; Herrick, I. R.;
Atesin, T. A.; Caruana, P. A.; Kellenberger, C. A.; Frontier, A. J. J. Am.
Chem. Soc. 2008, 130, 1003–1011. (b) He, W.; Sun, X.; Frontier, A. J. J. Am.
Chem. Soc. 2003, 125, 14278–14279. (c) Marcus, A. P.; Lee, A. S.; Davis,
R. L.; Tantillo, D. J.; Sarpong, R. Angew. Chem., Int. Ed. 2008, 47, 6379–
6383. (d) For substituent effects on the selectivity of Nazarov-type cycliza-
tions of R-hydroxyallenes for the synthesis of benzofulvenes, see: Cordier, P.;
Aubert, C.; Malacria, M.; Lacote, E.; Gandon, V. Angew. Chem., Int. Ed.
2009, 48, 8757–8760.
(15) The stereochemistry of 12 was confirmed by the chemical shift of the
vinylic proton (see: Li, Z.; Li, H.; Guo, X.; Cao, L.; Yu, R.; Li, H.; Pan, S.
Org. Lett. 2008, 10, 803–805.). The vinylic proton of (E)-2-benzoyl-3-
arylpropenonates resonate further downfield (δ 7.8-8.2 ppm) in comparison
to the corresponding (Z)-isomers (δ 6.8-7.2 ppm).
(16) Harrison, C. R. Tetrahedron Lett. 1987, 28, 4135–4138.
(17) Modified from the original procedure by: Degraw, J. I.; Christie,
P. H.; Kisliuk, R. L.; Gaumont, Y.; Sirotnak, F. M. J. Med. Chem. 1990, 33,
212–215. Our attempts to eliminate the syn-β-hydroxy ketones directly using
POCl3 were unsuccessful.
J. Org. Chem. Vol. 75, No. 14, 2010 4717

Smith et al.
JOCArticle
SCHEME 2. Synthesis of 1,3-Diarylallylic Alcohols 6 Using Knoevenagel or Aldol-Based Approaches
with substrate 6c (R2 = COOEt) reacting to give the corre-
sponding indene 8c isolated in 89% yield.21
Reactions of the isomeric allylic alcohols 6d or 14 with
SCHEME 3. Effect of Substitution at the 2-Position for the
Cyclization of Alcohols 6
BF3 OEt2 (1.0 equiv) in dichloromethane at room tempera-
3
ture afforded an inseparable mixture of 1H-indenes 8d and
9d in combined 84% and 84% yield, respectively (Scheme 4).
As anticipated, an identical 3:7 ratio of the indenes 8d/9d was
obtained for both reactions, demonstrating that the product
ratio of the isomeric indenes is independent of the initial
position of the allylic alcohol in the precursor, which is
consistent with a mechanism involving the intermediacy of
a common allylic cation.22 The ratio of the products indicates
that cyclization is favored through the unsubstituted phenyl
ring over the p-methoxyphenyl ring. DIBAL-H reduction of
the mixture of isomeric esters 8d/9d gave the corresponding
mixture of allylic alcohols 15a/b in 73% yield.
of the enones 12 using Luche conditions18 generally afforded
alcohols 6 in high yields. However, the reduction of the
enones 12 (R2 = COOEt) was often problematic because of
incomplete conversion to the corresponding alcohols 6 and
the presence of over-reduced byproducts.
The results can be rationalized in terms of cyclization via a
common allyl cation intermediate, where cyclization requires
the presence of a substituent at the C-2 allylic position (i.e.,
R2 ¼ H) (Figure 1). For the symmetrical case (i.e., where
both of the C-1 and C-3 substituents are phenyl groups), the
allylic cation generated on ionization can exist in three forms,
7-EE, 7-ZE (in this case identical to 7-EZ), and 7-ZZ. The
low energy form 7-EE exists in an extended “W” conforma-
tion and cannot directly undergo electrocyclization. Isomer-
ization of 7-EE through bond rotation to give 7-ZE (or 7-
EZ), which exists in an “S” conformation, is necessary before
cyclization can occur. 7-ZZ is also capable of undergoing
cyclization, but it is even higher in energy due to the
unfavorable A1,3 strain between the two aromatic groups.
Electrocyclization of 7-ZE in a π4a conrotatory fashion leads
to intermediate indanyl cation 16-trans, which on proton loss
gives indene 8. Similarly, π4a conrotatory electrocyclization
of 7-ZZ leads to the indanyl cation 16-cis, which on proton
loss also gives 8. The failure of 6a to undergo cyclization to
indene 8a can be attributed to the difficulty of isomerization
of cation 7-EE (R2 = H) into the reactive conformations
7-ZE, 7-EZ, or 7-ZZ (R2 = H) necessary for cyclization.
Introduction of a substituent at the 2-position promotes
cyclization as for the reactions of 6b or 6c. In these cases,
A1,2 strain between the central 2-substituent and the two
Initial studies focused upon the effect of substituents at the
2-position through the reaction of alcohols 6a-c in the
presence of a Lewis acid (BF3 OEt2, 1.0 equiv) in dichloro-
3
methane at room temperature (Scheme 3).19 Reaction of
the unsubstituted precursor 6a (R2=H) did not result in the
formation of indene 8a, with only the reduced species 1,3-
diphenylpropene 13a isolated in low yield. Similar formation
of reduced products such as 13a from allylic alcohols has
been observed to occur via a disproportionation process.20
Conversely, reaction of 2-methyl-substituted precursor 6b
(R2=Me) afforded indene 8b in 85% yield. Interestingly, the
presence of an electron-withdrawing group at the 2-position
did not interfere with allylic cation formation or cyclization,
(18) (a) Gemal, A.; Luche, J. J. Am. Chem. Soc. 1981, 103, 5454–5459.
~
(b) Barluenga, J.; Fananas, F.; Sanz, R.; Garcı
Lett. 1999, 40, 4735–4736.
ꢀ
a, F.; Garcıa, N. Tetrahedron
´ ´
(19) Following addition of BF3 OEt2 to 6a-c at room temperature, an
3
immediate color change from colorless to red or dark purple occurred, which
we assume is due to the presence of the corresponding cations. In the case of
6b/c, the color dissipated to a pale yellow solution (within 5 min), indicating
that the allylic cation had been trapped and that indene formation was
complete. In order to determine the influence of temperature on allylic cation
formation, the reaction of 6b was performed at -78 °C and slowly warmed to
0 °C. Addition of BF3 OEt2 to the pale yellow solution of 6b at -78 °C did
3
not result in any change in color. Slow warming of the solution over several
minutes resulted in the solution darkening to an orange color at -55 °C, red
at - 45 °C, and finally a deep purple color at -30 °C. On further warming, the
color started to dissipate at -20 °C, turning yellow at -10 °C, and finally pale
yellow at 0 °C, at which point the reaction was quenched with water.
Quenching the reaction at lower temperature (-78 °C) did not lead to indene
formation. Similar color change observations were noted in the reaction of
6a, except that the dark red color did not dissipate even after warming to
room temperature.
(21) Indene formation was also observed with trifluoromethyl-substi-
tuted phenylallyl cations. See: Gassman, P. G.; Ray, J. A.; Wenthold, P. G.;
Mickelson, J. W. J. Org. Chem. 1991, 56, 5143–5146.
(22) (a) For a review of reactions with allylic cations, see: DeWolfe, R. H.;
Young, W. G. Chem. Rev. 1956, 56, 753–901. (b) For a kinetic study of the
acid-catalyzed equilibration of differentially substituted 1,3-diarylallylic
alcohols via allylic cations, see: Bernstein, S. C. J. Org. Chem. 1968, 33,
3486–3488.
(20) For an example of an FeCl3-catalyzed disproportionation of allylic
alcohols, see: Wang, J.; Huang, W.; Zhang, Z.; Xiang, X.; Liu, R.; Zhou, X.
J. Org. Chem. 2009, 74, 3299–3304.
4718 J. Org. Chem. Vol. 75, No. 14, 2010

Smith et al.
JOCArticle
FIGURE 1. Mechanistic scheme for the cyclization of allylic cations 7 into 2-substituted 1-phenylindenes 8.
SCHEME 4. Cyclization of Isomeric Alcohols 6d and 14 into Indenes 8d and 9d
flanking aromatic groups would be expected to result in
destabilization of 7-EE (R2 ¼ H), and facilitate isomeriza-
tion. This A1,2 strain would not occur for the unsubstituted
case 7-EE (R2 = H). These experimental results support the
hypothesis that a central substituent promotes indene for-
mation by facilitating reaction through the “S” conforma-
tion of the arylallyl cation.7
A DFT computational study (B3LYP/6-31G*) on this
cationic system 7/16 (R2 = H, Me, and COOH) was con-
sistent with this model (Tables 1 and 2).23 The calculations
were conducted in the gas phase and therefore do not take
into account solvation effects. However, since all of the
species are cationic, solvation effects could be anticipated
to be roughly similar within each series. Relative energies
were calculated both as ZPVE-corrected B3LYP/6-31G*
energies and as thermally corrected B3LYP/6-31G* free
energies (including ZPVE) at standard temperature and
pressure. In general, there was only a minor variance bet-
ween these relative energy values. 7-ZE (R2 =H) is calcula-
ted to be þ6.1 kcal/mol higher in energy than 7-EE (R2=H).
On the other hand, for the methyl-substituted case the energy
difference of þ1.8 kcal/mol between 7-ZE and 7-EE (R2=Me)
is much lower, reflecting the ground-state destabilization of
7-EE. The relative energy of the intermediate indanyl cation
is slightly lower for the 16-trans form over the 16-cis form in
both the unsubstituted and 2-Me-substituted cases. How-
ever, the 2-Me substituted cation 16-trans (R2=Me) is more
stable than the unsubstituted cation 16-trans (R2 =H), with
energy values relative to the corresponding allylic cations
7-EE of þ23.3 and þ10.2 kcal/mol, respectively. The greater
relative stabilization of the 2-Me-substituted intermediate
indanyl cation is consistent with the electron-donating abil-
ity of the methyl group, which can directly stabilize the δþ
charge at C-2. The electron-donating ability of the 2-methyl
group would have a lesser effect on the precursor cations 7,
since the δþ charge is principally localized at C-1 and C-3.
Analysis of the calculated transition states TS-trans and
TS-cis (for R2 = H and Me) show considerable C7a-C1
bond development consistent with late product-like transi-
tion states. The calculated bond lengths (C7a-C1) in TS-
trans and TS-cis decrease along the series R2 = Me > H >
COOH, revealing that bond formation is more well devel-
oped (i.e., a later transition state) for the 2-COOH over the
(23) For calculations on the cyclization of phenylallyl cation (B3LYP/6-
31G*), see: Suzuki, T.; Ohwada, T.; Shudo, K. J. Am. Chem. Soc. 1997, 119,
6774–6780.
J. Org. Chem. Vol. 75, No. 14, 2010 4719

Smith et al.
JOCArticle
TABLE 1. Relative DFT Energies [B3LYP/6-31G* þ ZPVE (B3LYP/6-31G* þ G)]a and Internuclear Distances (C7a-C1) for 7, 16, and TS-trans/cis
(R2 = H, Me, COOH)b
R2 = H (kcal mol-1
)
R2 = Me (kcal mol-1
)
R2 = COOH (kcal mol-1
)
7-EE
7-ZE
7-ZZ
16-trans
16-cis
0.0
þ6.1 (þ6.2) (C7a-C1 = 3.202 A)
0.0
þ1.8 (þ1.8) (C7a-C1 = 3.111 A)
þ9.6 (þ8.8)
0.0
þ0.5 (þ0.3) (C7a-C1 = 3.071 A)
þ4.2 (þ3.9)
þ15.5 (þ15.6)
þ23.3 (þ24.3) (C7a-C1 = 1.601 A)
þ24.9 (þ26.0) (C7a-C1 = 1.562 A)
þ27.1 (þ28.4) (C7a-C1 = 1.970 A)
þ34.1 (þ35.3) (C7a-C1 = 2.046 A)
þ10.2 (þ10.8) (C7a-C1 = 1.580 A)
þ11.0 (þ11.5) (C7a-C1 = 1.563 A)
þ17.8 (þ18.6) (C7a-C1 = 2.035 A)
þ22.7 (þ23.3) (C7a-C1 = 2.101 A)
þ14.2 (þ14.5) (C7a-C1 = 1.594 A)
þ13.8 (þ14.4) (C7a-C1 = 1.562 A)
þ28.0 (þ28.9) (C7a-C1 = 1.939 A)
þ31.7 (þ32.2) (C7a-C1 = 2.023 A)
TS-trans
TS-cis
aRelative energies are given as B3LYP/6-31G* þ ZPVE and in parentheses as B3LYP/6-31G* þ G (where G = free energy including ZPVE).
bTrue minima on the potential energy hypersurface were identified by the presence of no imaginary frequencies, and transition structures were confirmed
by identifying only one imaginary frequency.
TABLE 2. Calculated Geometries of 7-ZE, TS-trans, and 16-trans (R2 = Me) (B3LYP/6-31G*)
bond distance or dihedral angle
7-ZE (R2 = Me)
TS-trans (R2 = Me)
16-trans (R2 = Me)
C1-C2
1.404 A
1.403 A
1.431 A
1.421 A
3.111 A
179.2°
1.478 A
1.370 A
1.429 A
1.449 A
2.035 A
151.0°
1.522 A
1.373 A
1.419 A
1.489 A
1.580 A
141.2°
C2-C3
C3-C3a
C3a-C7a
C7a-C1
Ph-C1-C2-C3
2-H- and 2-Me-substituted transition states. The negative
vibrational mode for the transition states was animated to
ensure that the imaginary frequency correctly represented the
appropriate bond formation process, and revealed a conrota-
tory motion consistent with π4a electrocyclization for both TS-
trans and TS-cis, with the latter being significantly higher in
energy. The calculated transition barrier for the conversion of
7 into 16-trans via TS-trans is considerably lower for the 2-
Me-substituted case than for the unsubstituted case (17.8
and 27.1 kcal/mol, respectively) (Table 1). This can be attri-
buted to both the greater ground-state destabilization of
7-EE (R2 =H), and the additional stabilization of TS-trans
(R2 = Me) that occurs as a result of the electron-donating
2-Me substituent.
At this stage, we set out to study the chemo- and regiosele-
ctivity of electrocyclization of differentially substituted 1,3-
diarylallylic cations 7 (Scheme 1 and Table 3). The electro-
cyclizations of such unsymmetrical systems have not been
investigated but are of interest, since it is not known whether
substituent effects for cationic electrocyclization react-
ions would mirror those of electrophilic aromatic substitu-
tion.4a,8,24 Accordingly, we set about investigating the reac-
tions of monosubstituted substrates 6, containing either a
COOEt or methyl substituent at the 2-position of the allyllic
4720 J. Org. Chem. Vol. 75, No. 14, 2010

Smith et al.
JOCArticle
TABLE 3. Chemo- and Regioselectivity Study of the Cyclization of Differentially Substituted Allylic Alcohols 6
allylic alcohol 6
crude ratioa
yieldb (%)
entry
R1
R2
isolated indenes 8, 9, or 17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
6e
6f
6g
6h
6i
6j
6k
6l
6m
6d
6n
6o
6p
6q
6r
6s
6t
p-NO2
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
Me
17 (40-NO2)
c
c
c
70
67
69
61
56
64
77
77
79
84
96
69
54
57e
71
64
81
75
m-NO2
o-NO2
p-Cl
m-Cl
o-Cl
p-Me
m-Me
o-Me
p-OMe
m-OMe
o-OMe
m-CN
p-CN
17 (30-NO2)
17 (20-NO2)
9 (40-Cl)/17 (40-Cl)d
1:2.2d
1:1:1.5
e1:10
1:1
8 (5-Cl)/8 (7-Cl)/9 (30-Cl)
8 (4-Cl)/9 (20-Cl)
8 (6-Me)/9 (40-Me)
8 (5-Me)/8 (7-Me)/9 (30-Me)
8 (4-Me)/9 (20-Me)
11:10:1
9:1
3:7
8 (6-MeO)/9 (40-OMe)
8 (5-OMe)/8 (7-OMe)/9 (30-OMe)
8 (4-OMe)/9 (20-OMe)
8 (5-CN)/8 (7-CN)/9 (30-CN)
8 (6-CN)/9 (40-CN)
15:1:-
e1:25
e1:e1:20
e1:15:-e
e1:25
e1:e1:20
g25:1
1: 1:5
p-NO2
m-CN
m,m-di-OMe
m-Cl
8 (6-NO2)/9 (40-NO2)
8 (5-CN)/8 (7-CN)/9 (30-CN)
8 (5,7-OMe)/9 (30,50-OMe)
8 (5-Cl)/8 (7-Cl)/9 (30-Cl)
Me
Me
Me
6u
aCrude ratio calculated by 1H NMR analysis. bIsolated product yields are reported for the products after silica gel chromatographic purification.
cCompounds 8/9 were not detected by 1H NMR analysis, and only the isomerized product 17 was obtained (g25:1). d1H NMR analysis of the crude
product, prior to purification on silica gel, revealed a 1: 2.2 ratio of 9 and 17 (R1 = 40-Cl), with none of compound 8 detectable. However, only compound
17 was isolated following silica gel column chromatographic purification. eThe 1H-indene 9 was isolated prior to purification on silica gel. Purification
resulted in a 1:1 mixture of 9/17.
framework (i.e., R2). A systematic examination of methyl,
methoxy, chloro, and nitro groups at the ortho, meta, and
para positions of 6 (R2 = COOEt) was first undertaken
(Table 3, entries 1-12). The indene products incorporated
the substituents either in the indene ring as for 8, or in the
exocyclic aryl ring for 9 and 17, with 17 resulting through
alkene isomerization of 9. Formation of 17 rather than 9
generally occurred for the more electron-deficient cases.
For the para-substituted cases, cyclization occurred pre-
ferentially through the unsubstituted ring for the nitro-,
chloro-, and cyano-substituted compounds to give the 40-sub-
stituted products 9 and/or 17 (Table 3, entries 1, 4, and 14).
Isomerized product 17 (R1 =NO2) was obtained in the case
of the more electron-deficient nitro-substituted system, while
a 1:1 mixture of 9/17 (R2 = COOMe, R1 = 40-CN) was
obtained in the cyano-substituted case. In the latter case,
only the unisomerized product 9 was observed by 1H NMR
analysis of the crude reaction mixture prior to chromato-
graphic purification. Reaction of the p-methyl-substituted
substrate 6 was unselective, leading to a 1:1 mixture of the
6-substituted indene 8 and the 40-substituted indene 9
(Table 3, entry 7). Interestingly the more electron-donating
p-methoxy-substituted substrate 6 underwent cyclization
with a modest preference for formation of the 40-substituted
over the 6-substituted indenes (Table 3, entry 10). For the
o-nitro-, chloro-, and methoxy-substituted cases, cyclization
occurred preferentially through the unsubstituted phenyl
ring, leading to the 20-substituted compounds 17/9 (R1 =
20-NO2, Cl, OMe) (Table 3, entries 3, 6, and 12). However,
reaction of o-methyl-substituted 6 led to the selective forma-
tion of 8 (R1=4-Me) over 9 (R1=20-Me) (Table 3, entry 9).
For the meta-substituted cases, three possible cyclization
modes are possible. Cyclization onto the substituted aryl ring
leads to either the 5- or 7-substituted indenes 8, whereas
cyclization through the unsubstituted phenyl ring leads to
the 30-substituted indene 9. Once again, the presence of the
strongly electron-withdrawing nitro or cyano groups resulted
in selective cyclization to 17/9 (R1=30-NO2 or CN) (Table 3,
entries 2 and 13). Reaction of 6 (R1 =m-Cl) was unselective,
leading to a mixture of the 5-, 7-, and 30-substituted indenes
(Table 3, entry 5). Reaction of 6 (R1 = m-Me) showed poor
selectivitybetween 8 (R1=5-Me or 7-Me) and 9 (R1=30-Me),
although in this case the selectivity for the 7- over the 5-sub-
stituted indenes 8 was high (Table 3, entry 8). In contrast,
reaction of 6 (R1=m-OMe) was highly selective, leading to a
15:1 ratio of the 5- and 7-substituted indenes 8, with none of 9
(R1 =30-OMe) being detected (Table 3, entry 11).
Similar trends were observed in the cyclizations of 2-Me-
substituted precursors 6 (R2=Me) (Table 3, entries 15-18).
Once again, the presence of strongly electron-withdraw-
ing nitro or cyano groups led to preferential cyclization
through the unsubstituted phenyl ring (Table 3, entries
15 and 16). Reaction of the di-m-OMe substituted 6 (R2 =
Me) led to exclusive formation of 5,7-di-OMe-substituted
(24) For unsymmetrical substituted 1-phenylallyl and 1,3-diphenylallyl
systems containing alkyl and aryl substitutions along the allyl backbones that
undergo regioselective cyclizations via an allyl cation, see: (a) Guo, S.; Liu,
Y. Org. Biomol. Chem. 2008, 6, 2064–2070. (b) Xi, Z.; Guo, R.; Mito, S.; Yan,
H.; Kanno, K.; Nakajima, K.; Takahashi, T. J. Org. Chem. 2003, 68, 1252–
1257.
J. Org. Chem. Vol. 75, No. 14, 2010 4721

Smith et al.
JOCArticle
FIGURE 2. Mechanistic scheme for the formation of monosubstituted arylindenes 8/9.
8 (R2=Me) (Table 3, entry 17). The cyclization of m-chloro-
substituted 6 (R2 = Me, R1 = m-Cl) led to a mixture of the
5-Cl- and 7-Cl-substituted isomers of 8 and the 30-Cl-sub-
stituted 9 (Table 3, entry 18). The selectivity for this reaction,
favoring the exo-substituted product 9 (R2=Me, R1=30-Cl)
over 8 (R2 = Me, R1 = 5-Cl or 7-Cl), is higher than that
obtained for the corresponding m-Cl-2-COOEt-substituted
compound 6 (i.e., Table 3, entry 5).
Isomerization of the double bond in the indene ring of 8/9
to give 17 was observed only for those substrates containing
a central 2-COOEt ester substituent (Table 3, entries 1-4
and 14). One known mechanism for indene isomerization
involves consecutive 1,5-hydrogen shifts, although this
usually occurs at higher temperatures (>100 °C).25 However,
since isomerization only occurs for more electron-deficient
substrates, a more plausible alternative is that for the initially
formed indene products bearing a 2-COOEt ester substituent
group and an electron-deficient exocyclic aryl ring at the
1-position (i.e., 9, R2=NO2 or CN), the increased acidity of
the dibenzylic C(1)-H proton facilitates a base catalyzed
isomerization via an indenyl anion to give the more thermo-
dynamically stable isomerized adducts 17.26 This is sup-
ported by the observation that in unsuccessful attempts to
achieve conjugate addition reactions with the R,β-unsatu-
rated ester moiety of 8/9 (R1 =COOEt) using several tradi-
tional nucleophiles of varying basicity, only isomerization
products 17 were isolated.27
parallel those anticipated based on substituent effects for
electrophilic aromatic substitution, in this case an intramo-
lecular reaction, or whether other effects would be important
because of the electrocyclic nature of the cyclization. The
chemoselectivity of cyclization is presumably controlled by
the respective transition-state barriers for cyclization, since
subsequent proton loss to the product indenes should be
rapid and irreversible. A cursory examination of reaction
selectivity for formation of 8 versus 9 reveals that the
presence of electron-withdrawing nitro or cyano substituents
on 6 strongly disfavors cyclization onto the substituted ring,
resulting in formation of 9. This observation is perhaps not
unexpected, and is consistent with what would be expected
for an electrophilic aromatic substitution reaction. However,
the presence of electron donating groups such as methyl or
methoxy results in cyclization to either the substituted or
unsubstituted ring of 6, depending upon whether the sub-
stituent is at the ortho, meta, or para position. Overall, the
results obtained for the chloro-, methyl-, and methoxy-
substituted substrates 6 cannot be readily correlated with
trends in reactivity that would be expected from a compar-
ison with the substituent group effects on intermolecular
electrophilic aromatic substitution or through the use of
linear free energy relationships (using σ or σþ values).
An understanding of the chemoselectivity of cyclization
was established through consideration of a DFT (B3LYP/6-
31G*) computational study of the gas-phase transition state
barrier for electrocyclization of the substituted cation 18
(R2=Me, R1=o-, m-, and p-NO2, -Cl, -OMe, and -Me) into
19 via TSZE and for the cyclization of 18 into 20 via TSEZ
(Figure 2). Again, for this study it was assumed that since all
of the species are cationic the effects of solvation would be
anticipated to be roughly similar within each series. There-
fore, calculations were conducted in the gas phase and do not
explicitly include solvation effects. The transition-state bar-
riers leading to the formation of 19 and 20 are calculated as
the difference in energy (corrected for ZPVE or free energy/
ZPVE) in kcal mol-1 between TSZE and 18-EE and TSEZ
and 18-EE, respectively (Table 4). The difference in energy
between the calculated transition states TSZE and TSEZ
One of the main goals of this study was to establish
whether the selectivity of cyclization of substituted 6 would
(25) (a) Roth, W. R. Tetrahedron Lett. 1964, 5, 1009–1013. (b) Jones,
D. W.; Marmon, R. J. J. Chem. Soc., Perkin Trans. 1 1993, 6, 681–690. (c) 1,5-
Hydrogen shifts were observed with verbenindene derivatives at 115 °C in
pyridine. See: Rupert, K. C.; Liu, C. C.; Nguyen, T. T.; Whitener, M. A.;
Sowa, J. R. Organometallics 2002, 21, 144–149.
(26) (a) Friedrich, E. C.; Taggart, D. B. J. Org. Chem. 1975, 40, 720–723.
(b) Bergson, G. Acta Chem. Scand. 1963, 17, 2691–2700. (c) Bergson, G.;
Weidler, A.-M. Acta Chem. Scand. 1963, 17, 1798–1799. (d) Bergson, G.;
Weidler, A.-M. Acta Chem. Scand. 1963, 17, 862–864.
(27) Attempted conjugate addition of indene 8c with thiophenol (in the
presence of 12 mol % of NEt3) and sodium dimethylmalonate and indene 8d/
9d (p-OMe) with morpholine (in the presence of 150 mol % of i-Pr2NEt)
resulted in complete isomerization of the double bond. See the Supporting
Information for details.
4722 J. Org. Chem. Vol. 75, No. 14, 2010

Smith et al.
JOCArticle
TABLE 4. Transition-State Barrier (in kcal mol-1) Calculated [B3LYP/6-31G* þ ZPVE (B3LYP/6-31G* þ G)]a for the Difference in Energy between
TSb,c and the Respective 18-EE and Internuclear Distances (C7a-C1) for TS
transition state barriera,c
R = H
R = NO2
R = Cl
R = Me
R = OH
TSZE (p-6)
TSEZ (p-40)
TSZE (m-5)
TSZE (m-7)
TSEZ (m-30)d
TSZE (o-4)
TSEZ (o-20)d
17.8 (18.6) (2.035 A)
17.8 (18.6) (2.035 A)
17.8 (18.6) (2.035 A)
17.8 (18.6) (2.035 A)
17.8 (18.6) (2.035 A)
17.8 (18.6) (2.035 A)
17.8 (18.6) (2.035 A)
18.6 (19.7) (2.016 A)
16.9 (17.7) (2.076 A)
21.2 (22.1) (1.990 A)
18.8 (20.1) (1.956 A)
17.2 (18.0) (2.070 A)
19.6 (19.6) (2.019 A)
15.6 (17.6) (2.120 A)
20.2 (21.1) (2.009 A)
18.5 (19.3) (2.037 A)
16.7 (17.6) (2.074 A)
16.7 (17.8) (2.059 A)
17.4 (17.5) (2.062 A)
18.5 (19.6) (2.026 A)
16.8 (17.6) (2.054 A)
18.8 (19.6) (2.030 A)
18.8 (19.1) (2.018 A)
16.0 (16.9) (2.087 A)
16.6 (17.6) (2.090 A)
18.3 (19.4) (2.037 A)
16.3 (17.3) (2.037 A)
18.0 (19.2) (2.025 A)
22.6 (23.3) (1.961 A)
20.4 (21.3) (1.983 A)
13.5 (14.4) (2.159 A)
14.3 (14.5) (2.146 A)
17.9 (18.8) (2.055 A)
21.3 (22.2) (2.012 A)
17.0 (18.4) (2.060 A)
aTransition-state barriers were determined as the difference of calculated energies (B3LYP/6-31G*) between TS and 18-EE and are ZPVE corrected.
Values in parentheses were determined by the difference of calculated free energies between TS and 18-EE (B3LYP/6-31G* þ G, where G=free energy
including ZPVE correction). bTransition states were confirmed by identifying only one imaginary frequency. cEach transition state TS is followed by a
letter-number combination in brackets. The letter refers to the position of substitution (i.e., ortho, meta or para) on the corresponding precursor cation
18, and the letter refers to the position of substitution on either the indene ring system (i.e., 4-, 5-, 6- or 7-substituted) or the exocylic aromatic ring (i.e.,
20-, 30- or 40-substituted) of TS/19/8/20/9. dOnly data for the lowest energy conformation is shown (i.e., rotamer).
TABLE 5. Calculateda and Observedb Product Ratios for the Formation of 8 and 9 from Substituted Ortho-, Meta-, and Para-Monosubstituted 6
selectivity 6 f indene 8:9
R = NO2
R = Cl
R = Me
R = OHc (OMe)d
p-6 f 6-8: 40-9 (calcd)
p-6 f 6-8: 40-9 (obsd)
m-6 f 5-8: 7-8: 30-9 (calcd)
m-6 f 5-8: 7-8: 30-9 (obsd)
o-6 f 4-8: 20-9 (calcd)
o-6 f 4-8: 20-9 (obsd)
1:16 (1:28)
e1:25
1:61:870 (1:32:1100)
e1:e1:25
1:810 (1:210)
e1:25
1:17 (1:19)
1:18
3.2:2.8:1 (1.5:1:1.7)
1:1:1.5e
1:20 (1:26)
1:10
1.1:1 (2.2:1)
1:1
45:16:1 (62:22:1)
11:10:1
20: 1 (27:1)
9:1
1:44c (1:29)c
1:2.3d
1700:450:1c (1700:1400:1)c
15:1:0d
1:1400c (1:590)c
e1:25d
aCalculated ratio estimated on the basis of ΔΔEq which is equivalent to the energy difference between the ZPVE corrected transition-state energies for
TSZE and TSEZ leading to 8 (R2=Me) and 9 (R2=Me), respectively (see Table 4). Values in parentheses are determined on the basis of ΔΔGq, which is
equivalent to the free energy difference between TSZE and TSEZ leading to 8 (R2=Me) and 9 (R2=Me), respectively (see Table 4). The selectivity was
then estimated using exp(-ΔΔEq/RT) or exp(-ΔΔGq/RT) at 298.15 K. bObserved ratio determined for ester-derived products 8 (R2 =COOEt) and 9
(R2=COOEt). The observed ratios were determined by analyzing the integration of the vinylic protons (H-3) or benzylic protons (H-1) in the 1H NMR
of the crude reaction mixture. cCalculated ratio determined for OH substituted compounds. dObserved ratio for the OMe-substituted compounds (R2=
COOEt). eAn observed ratio of 1:1:5 was obtained for the R2 =Me substituted compounds.
(Table 4) can be used to estimate the kinetic selectivities for the
formation of the substituted compounds 8 versus 9 (Table 5).
In general, there is good agreement between the estimated
selectivities based upon the ZPVE-corrected transition-state
energies (or the free energy/ZPVE-corrected transition-state
energies) of TSZE and TSEZ and the experimentally deter-
mined ratios of 8/9 (Table 5). Despite the approximations
made in the calculations (i.e., the lack of solvent corrections
and the replacement of a COOEt for a methyl group for the
observed and computational studies), the computed selectiv-
ities are in generally good agreement with the observed
selectivities. This provides some support for the hypothesis
that the chemo- and regioselectivity of the reactions are
determined by the relative rates of the respective electrocy-
clization reactions (18-EE into 19 and 18-EE into 20).
Furthermore, there is a correlation between the internuclear
distance C7a-C1 and the calculated transition-state barrier
on varying substituents at a specific position for cyclization
(Table 4). Thus, as the activation energy increases the inter-
nuclear distance C7a-C1 in the respective TS decreases,
consistent with late transition states.
The ortho/para- or meta-selectivity of electrophilic aro-
matic substitution is often rationalized in terms of electron
density at the positions around the ring, with positions of
higher electron density reacting more rapidly. However,
calculated electron densities for the substituted ground-state
cations 18-EE, 18-EZ, or 18-ZE did not correlate with the
positional selectivity for cyclization to 8 versus 9. Similarly,
examination of the HOMOs for these species did not reveal
any obvious correlation with the selectivity of cyclization.
The calculated and observed selectivity for cyclization of
meta-substituted 6 to 5- or 7-substituted indenes 8 versus
30-substituted 9 decreases along the series OH (OMe)>Me>
Cl>NO2 (Table 5), which parallels the decreasing electron-
donating ability of these substituents. This can be rationa-
lized in terms of the decreasing ability of these groups to
stabilize intermediate 19 and transition state TSZE over 20
and TSEZ along the series. For each of the substituents
investigated, the highest selectivity for formation of 8 over 9
occurred from m-6 rather than o-6 or p-6. The calculated and
observed selectivities for cyclization of ortho-substituted 6 to
4-substituted indene 8 versus 20-substituted 9 decreases along
the series Me > Cl > NO2 J OH (Table 5). This trend does
not correlate with overall electron-donating ability of these
groups, presumably because resonance stabilization of 19
and TSZE is not significant. Instead, the trend in selectivity
can be rationalized in terms of the increasing inductive with-
drawing group ability along the series (group electronegativ-
ities28 for CH3=2.3, Cl=3.0, NO2=3.4, OH=3.7), which
would in turn lead to increasing destabilization of intermedi-
ate 19 and transition state TSZE (relative to 20 and TSEZ).
The selectivities for cyclization of para-substituted 6 to
6-substituted indene 8 versus 40-substituted 9 decrease along
the series (calculated) Me>Cl ∼ NO2 >OH or (observed)
Me > OMe > Cl > NO2 (Table 5). These trends parallel
those of the ortho-substituted series, although the selectivity
differences for the reactions of para-substituted 6 substrates
are less pronounced than for the corresponding ortho-sub-
stituted substrates. The observed cyclization selectivity for
p-6 never favors formation of 8 over 9, which also contrasts
with the reactions of the ortho-substituted substrates. The
highest selectivity for formation of 8 over 9 from o-6 or p-6
(28) Wells, P. R. Prog. Phys. Org. Chem. 1968, 6, 111–145.
J. Org. Chem. Vol. 75, No. 14, 2010 4723

Smith et al.
JOCArticle
occurs for the methyl-substituted cases, with ratios of 8:9 of
1:1 for cyclization of 6k (p-Me) and 9:1 for cyclization of 6m
(o-Me).
TABLE 6. Electrocyclization of Disubstituted 1,3-Diarylallyl Alcohols 6
Cyclization of meta-substituted 6 onto the substituted aryl
ring can give two possible regioisomers, 5-8 and 7-8. The
observed regioselectivity for 5-8 over 7-8 is about 1:1 for both
the chloro- and methyl-substituted examples but favors the
5-isomer for the methoxy-substituted case (15:1). The calcu-
lated regioselectivity decreases along the series OH J Me>
Cl>NO2, which correlates with decreasing electron-donating
ability of the substituents, rather than substituent size (e.g.,
A value) or calculated electron densities on the aromatic
carbons that undergo cyclization.
The results outlined in Tables 3 and 5 provide a guide to
the directing group effects of various aryl substituents.
Further study then focused upon substrates in which both
aryl rings of the 1,3-diarylallyl alcohols were substituted
(Table 6). 2-Methyl-substituted substrates 6v-z were chosen
since the synthetic protocol used for their preparation is
higher yielding than that used for 2-COOEt-substituted
precursors and the indene products formed would not under-
go double-bond isomerization. In addition, the results ob-
tained for the monosubstituted precursors 6 had shown that
the directing group influences of the aryl substituents were
not strongly affected by the central 2-substituent (i.e.,
2-COOEt vs 2-Me). o-OMe/p-NO2-substituted precursor
6v underwent regioselective cyclization, preferentially lead-
ing to 4-OMe-substituted indene 9v in moderate yield
(Table 6, entry 1). Although both substituents are deactivat-
ing toward electrocyclization, the directing group influence
of the p-NO2 group outweighs that of the o-OMe group,
leading to cyclization through the o-OMe-substituted ring.
Unexpectedly, even though a p-NO2 group might be expec-
ted to be substantially more deactivating than a p-Cl group,
cyclization of p-NO2/p-Cl-substituted 6w was less selective,
giving a 4:1 mixture of the 1H-indenes 9w and 8w prior to
chromatography (Table 6, entry 2). Introduction of a deac-
tivating p-Cl substituent in 6x enabled regioselective cycliza-
tion to 9x (Table 6, entry 3), even though a p-Me group had
been shown to exert little directing influence in the cycliza-
tion of 6k (Table 6, entry 7). Although both p-Cl and p-OMe
aryl substituents are deactivating, the greater deactivating
character manifested by the p-Cl over the p-OMe group
resulted in the regioselective formation of the 6-OMe-
substituted indene 9y from 6y (Table 6, entry 4). Finally,
reaction of p-NO2/p-OMe-substituted 6z at room tempera-
ture led to two products: 6-OMe-substituted 9z and an init-
ially unidentified dimeric compound 21. However, reaction
of 6z at 100 °C in toluene led to the formation of 9z in mode-
rate yield (Table 6, entry 5).
aCrude ratio calculated from crude 1H NMR analysis. bAddition of
BF3 OEt2 at 100 °C in toluene. cThe corresponding 1H-indenes 9v (6-
3
NO2, 20-OMe,) and 8v (4-OMe, 40-NO2) were observed in a 1:1.8 ratio
prior to purification. dThe isomerized 3H-indenes 17v (6-NO2, 20-OMe)
and (4-OMe, 40-NO2) were isolated in a 1:1.2 ratio following silica gel
column chromatographic purification. eThe corresponding 1H-indenes
9w (6-Cl, 40-NO2) and 8w (6-NO2, 40-Cl) were observed in a 4:1 ratio
prior to purification. fThe isomerized 3H-indenes 17w (6-Cl, 40-NO2 and
6-NO2, 40-Cl) were isolated in a 6:1 ratio following silica gel column
chromatographic purification.
Further investigation of the reaction of 6z at a lower
temperature (-20 °C) resulted in the exclusive formation
of the dimeric side product 21 in 88% yield (Scheme 5). The
cyclopenta[a]indene structure of dimer 21 was elucidated
from a combination of homonuclear and heteronuclear 2D
NMR experiments.29 A plausible mechanism for the forma-
tion of 21 involves a cationic domino or cascade annulation
sequence. Thus, regio- and diastereoselective trapping of the
allylic cation 2230 by in situ generated indene 9z to give an
intermediate tertiary carbocation 23, followed by intramo-
lecular cyclization to the cation 24, and selective elimination
would provide the 21. The formation of 21 is an example of a
formal [2π þ 2π] cycloaddition.31 The remarkable selectivity
exhibited in the formation of 21 can be rationalized, in part,
by attack of indene 9z from the less hindered face onto the
more electron-deficient position of the allylic cation 22.
(29) See the Supporting Information for a detailed analysis of the 2D
NMR experiments.
(30) For studies on regioselective additions to unsymmetrical 1,3-diary-
lallylic cations, see: Watts, W. E.; Easton, A. M.; Habib, M. J. A.; Park, J.
J. Chem. Soc., Perkin Trans. 2 1972, 15, 2290–2297.
4724 J. Org. Chem. Vol. 75, No. 14, 2010

Smith et al.
SCHEME 5. Plausible Mechanism for the Domino Electrocyclization/Formal [2π þ 2π] Cycloaddition/Dimerization of Alcohol 6z to 21
JOCArticle
SCHEME 6. Electrocyclization Pathways for 2-Naphthylallyl-Substituted Substrates
Subsequent cyclization of 23 to the cis-fused [3.3.0] ring
system 24 followed by proton loss would then give 21.32
The high regioselectivity exhibited for the trapping of cation
22 by 9z is surprising since kinetic studies of the substitution
of differentially substituted allylic cations 7 (R2 =H) reveal
little difference between the rates of trapping of the two
allylic positions of 7.22b The temperature-dependent forma-
tion of 21 at low temperature and 9z at high temperature
presumably results because the longer lifetime of the allylic
cation 22 at lower temperatures enables intermolecular
trapping by the in situ generated indene 6z, whereas at higher
temperatures the rate of electrocyclization of 22 is faster than
the more entropically disfavored intermolecular trapping
reaction.
three possible cyclization modes were isolated. In the case of
the 1-(2-naphthyl)-3-phenylallyl alcohol, products 8aa and
9aa were obtained in a 1:2 ratio, favoring cyclization through
the 1-position of the naphthyl ring. The modest selectivity
favoring the latter is consistent with the greater resonance
stabilization of the cation afforded by the naphthyl ring in
the transition state. This chemoselectivity preference was
much more pronounced for the reaction of substrate 6ab
bearing both a 2-naphthyl and a deactivated p-nitrophenyl
ring, resulting in exclusive formation of 9ab. For both
substrates, however, cyclization through the naphthyl ring
was highly regioselective, with the only observed products
9aa and 9ab obtained as a result of cyclization through the
1-position of the naphthyl ring, rather than through the
3-position of the naphthyl ring. This parallels the known
electrophilic aromatic substitution reaction of naphthalene
which kinetically favors substitution at the 1-position over
the 2-position (which is equivalent to the 3-position).
Reaction of naphthylallyl alcohol precursors 6aa/6ab
under the Lewis acidic protocol also led to electrocyclization
products (Scheme 6). Products corresponding to two of the
(31) See, for example: (a) Klein, H.; Mayr, H. Angew. Chem., Int. Ed.
Engl. 1981, 20, 1027–1029. (b) Miller, A.; Moore, M. Tetrahedron Lett. 1980,
21, 577–580. (c) Angle, S. R.; Arnaiz, D. O. J. Org. Chem. 1992, 57, 5937–
Finally, we were interested in investigating the selectivity
of a reaction of a substrate bearing both a phenyl and a
heterocyclic aromatic ring (Scheme 7). The use of a benzo-
furan-derived system was of particular interest, since elec-
trocyclization onto the benzofuran ring could potentially
lead to the tetrahydrocyclopenta[b]benzofuran core of the
rocaglamide series of natural products.33 However, attem-
pted reaction of the 3-substituted benzofuran 6ac (R=H) led
to a complex mixture of products, with 3H-indene 17ac
~
5947. (d) Alesso, E.; Torviso, R.; Lantano, B.; Erlich, M.; Finkielsztein,
L. M.; Moltrasio, G.; Aguirre, J. M.; Brunet, E. ARKIVOC 2003, 10, 283–
297.
(32) (a) The 1,2,3,3a,8,8a-hexahydrocyclopenta[a]indene core of 21 is
found in the natural product pallidol. It was recently synthesized by a 4π-
electrocyclization of a biaryl alcohol via a pentadienyl cation in the presence
of TFA or TsOH; see: Snyder, S. A.; Zografos, A. L.; Lin, Y. Angew. Chem.,
Int. Ed. 2007, 46, 8186–8191. (b) Snyder, S. A.; Breazzano, S. P.; Ross, A. G.;
Lin, Y.; Zografos, A. L. J. Am. Chem. Soc. 2009, 131, 1753–1765.
J. Org. Chem. Vol. 75, No. 14, 2010 4725

Smith et al.
JOCArticle
SCHEME 7. Electrocyclization Pathways for Benzofuran-
Substituted Substrates 6ac and 6ad
Experimental Section34
General Procedure for Indene Synthesis. To a solution of the
arylallyl alcohol in CH2Cl2 (0.1 M) at room temperature, unless
otherwise stated, was added BF3 OEt2 (1.0 equiv). The solution
3
was stirred until the alcohol was fully reacted, as determined by
TLC analysis. The solution was diluted with H2O (2 mL) and
saturated aqueous NaHCO3 (2 mL) and extracted with CH2Cl2
(3 ꢀ 5 mL). The combined organic extracts were dried with
Na2SO4, filtered, and concentrated in vacuo. Purification by
flash chromatography on silica gel afforded pure indene.
2-Methyl-1-phenyl-1H-indene (8b)35. The general procedure
for the synthesis of indenes from alcohol 6b was followed.
Without further purification, the indene (83 mg, 85%) was
isolated as a white solid: mp 39-41 °C (lit.35 mp 47.5-49 °C);
1H NMR (400 MHz, CDCl3) δ 7.18-7.29 (5 H, m), 7.11 (1H, d,
J=7.5 Hz), 7.04 (1H, dd, J=7.5, 1.0 Hz), 6.99-7.03 (2H, m),
6.54 (1H, s), 4.29 (1H, s), 1.92 (3H, s); 13C NMR (100 MHz,
CDCl3) δ 149.9 (C), 148.6 (C), 144.8 (C), 139.7 (C), 128.6 (CH),
128.2 (CH), 127.2 (CH), 126.72 (CH), 126.71 (CH), 124.2 (CH),
123.7 (CH), 119.8 (CH), 59.4 (CH), 15.2 (CH3).
Ethyl 1-Phenyl-1H-indene-2-carboxylate (8c). The general
procedure for the synthesis of indenes from alcohol 6c was
followed. Purification by flash chromatography on silica gel
using a gradient elution of hexanes to 5% ethyl acetate in
hexanes afforded the indene (30 mg, 89%) as a white solid: mp
87-88 °C; 1H NMR (300 MHz, CDCl3) δ 7.79 (1H, dd, J=2.0,
0.5 Hz), 7.50 (1H, d, J=7.0 Hz), 7.14-7.32 (6H, m), 7.03-7.10
(2H, m), 4.84 (1H, d, J=2.0 Hz), 4.04-4.22 (2H, m), 1.17 (3H,
t, J = 7.0 Hz); 13C NMR (75 MHz, CDCl3) δ 164.2, 150.4,
141.8, 141.2, 138.4, 128.4, 128.2, 127.9, 127.8, 127.2, 126.8,
124.4, 123.3, 60.2, 55.7, 14.1; MS (EI) m/z (rel intensity) 264
(42), 192 (27), 191 (100), 189 (24), 165 (11); HRMS (EI) m/z
Calcd for C18H16O2 [Mþ] 264.1150, found 264.1155.
isolated in 10% yield. Reaction of benzofuran 6ad (R=Ph),
either as the (E)- or (Z)-isomer, led to the indene product 9ad
(R = Ph) in good yield.
Conclusion
Selectivity effects in the cyclization reactions of 1,3-diaryl-
substituted allylic cations to indenes have been experimen-
tally and computationally examined. The presence of sub-
stituents on the aryl ring can affect the chemoselectivity of
cyclization and for reactions of meta-substituted cases the
regioselectivity of cyclization. In general, the presence of
electron-withdrawing substituents disfavors cyclization onto
this ring. The effects of electron-donating substituents are
more complex and depend on whether the substituent is
inductively or mesomerically electron donating and on the
position of the substituent (ortho, meta, or para). Substit-
uent effects are most pronounced at the meta position, with
electron-donating substituents promoting cyclization onto
the substituted ring. A simple gas-phase computational
model (B3LYP/6-31G* þ ZPVE or B3LYP/6-31G* þ G)
in which the selectivity is assumed to depend upon the
relative rates of electrocyclization showed generally good
agreement with experimental results. There was no obvious
correlation of selectivity with electron densities as has been
suggested for electrophilic aromatic substitution reactions.
Transition-state structures are consistent with a cationic π4a
conrotatory electrocyclization mechanism.
In some cases for more electron-deficient systems, the
product 1H-indenes underwent subsequent alkene isomer-
ization to 3H-indenes, which likely occurs through a base-
catalyzed isomerization rather than a consecutive 1,5-H shift
mechanism. In one case, in which both aromatic rings were
substituted, an interesting dimeric adduct resulting from a
formal 2π þ 2π cycloaddition of the allyl cation intermediate
to the initially formed indene occurred. Reaction at higher
temperatures prevented the formation of this adduct, sug-
gesting that the longer lifetime of the cation at lower tem-
peratures is required for the formal [2π þ 2π] cycloaddition
to occur. Further studies on electrocyclization reactions and
substituent effects will be reported in due course.
Ethyl 4-Methyl-1-phenyl-1H-indene-2-carboxylate (8m) and
Ethyl 1-(2-Methylphenyl)-1H-indene-2-carboxylate (9m). The
general procedure for the synthesis of indenes from alcohol 6m
was followed. Purification by flash chromatography on silica
gel using 10% ethyl acetate in hexanes afforded a 9:1 mixture
of the 4-Me and 20-Me isomer (75 mg, 85%) as white crystals,
mp 78-79 °C. Characterization data is reported for the 4-Me
isomer only: IR (thin film) νmax 3050, 3014, 2968, 1710, 1600,
1566, 1494, 1448, 1360, 1301, 1239, 1186, 1080, 1024, 771, 737,
693 cm-1; 1H NMR (400 MHz, CDCl3) δ 7.92 (1H, dd, J=2.5,
0.5 Hz), 7.15-7.25 (2H, m), 7.14 (2H, t, J=8.0 Hz), 7.04-7.14
(3H, m), 7.01 (1H, d, J = 7.0 Hz), 4.81 (1H, d, J = 0.5 Hz),
4.05-4.22 (2H, m), 2.50 (3H, s), 1.17 (3H, t, J = 7.0 Hz); 13
C
NMR (75 MHz, CDCl3) δ 164.6 (C), 150.7 (C), 141.3 (C), 140.5
(C), 139.9 (CH), 138.9 (C), 133.2 (C), 128.7 (CH), 128.6 (CH),
128.5 (CH), 128.1 (CH), 127.0 (CH), 122.1 (CH), 60.4 (CH2),
56.1 (CH), 18.7 (CH3), 14.3 (CH3); MS (EI) m/z (rel intensity)
279 (13), 278 (57), 206 (25), 205 (100), 203 (14), 202 (14), 189
(12); HRMS (EI) m/z calcd for C19H18O2 [Mþ] 278.1307,
found 278.1304.
5,7-Dimethoxy-2-methyl-1-phenyl-1H-indene (8t). The gener-
al procedure for the synthesis of indenes from alcohol 6t was
followed. Without any further purification, the indene (65 mg,
69%) was isolated as pale yellow crystals: mp 93-94 °C; IR (thin
film) νmax 3003, 2921, 2839, 1585, 1486, 1449, 1425, 1337, 1279,
1201, 1133, 1092, 1041, 871, 817, 702 cm-1; 1H NMR (300 MHz,
CDCl3) δ 7.12-7.25 (3H, m), 7.10 (2H, dd, J=8.0, 1.5 Hz), 6.50
(1H, d, J = 2.0 Hz), 6.39 (1H, t, J = 1.5 Hz), 6.21 (1H, d, J =
2.0 Hz), 4.32 (1H, s), 3.81 (3H, s), 3.57 (3H, s), 1.83 (3H, d, J=
1.0 Hz); 13C NMR (75 MHz, CDCl3) δ 161.2, 155.9, 152.2,
147.5, 139.7, 128.4, 128.2, 127.5, 126.8, 126.4, 98.1, 95.6, 57.3,
(33) (a) For chemistry and biological activity, see: Proksch, P.; Edrada,
R.; Ebel, R.; Bohnenstengel, F. I.; Nugroho, B. W. Curr. Org. Chem. 2001, 5,
923–938. (b) For isolation and structure elucidation, see: King, M. L.;
Chiang, C. C.; Ling, H. C.; Fujita, E.; Ochiai, M.; McPhail, A. T. J. Chem.
Soc., Chem. Commun. 1982, 1150. (c) For a recent total synthesis using a
Nazarov cyclization of a pentadienyl cation, see: Malona, J. A.; Cariou, K.;
Frontier, A. J. J. Am. Chem. Soc. 2009, 131, 7560–7561.
(34) Experimental data and spectra are given for a few selected examples
of the indenes. Experimental data and spectra for all other compounds and
precursors are given in the Supporting Information.
(35) (a) Tolbert, L. M. J. Org. Chem. 1979, 44, 4584–4588. (b) Ming-
Yuan, L.; Madhushaw, R. J.; Liu, R.-S. J. Org. Chem. 2004, 69, 7700–7704.
4726 J. Org. Chem. Vol. 75, No. 14, 2010

Smith et al.
JOCArticle
55.7, 55.6, 15.4; MS (EI) m/z (rel intensity) 266 (100), 251 (58),
235 (10), 189 (10), 165 (10), 84 (10); HRMS (EI) m/z calcd for
C18H18O2 [Mþ] 266.1307, found 266.1303.
IR (thin film) νmax 2935, 2847, 1605, 1516, 1476, 1346, 1292,
1279, 1223, 1106, 1031, 852, 803, 741, 698 cm-1 1H NMR
;
(400 MHz, CDCl3) δ 8.14 (2H, d, J=8.5 Hz), 7.19 (1H, d, J=
8.0 Hz), 7.18 (2H, d, J=8.5 Hz), 6.80 (1H, dd, J=8.0, 2.5 Hz),
6.66 (1H, d, J = 2.0 Hz), 6.53 (1H, t, J = 1.5 Hz), 4.37 (1H, s),
3.73 (3H, s), 1.88 (3H, d, J = 1.5 Hz); 13C NMR (75 MHz,
CDCl3) δ 157.9, 149.3, 148.8, 147.2, 146.4, 137.8, 129.2, 128.0,
124.3, 120.7, 112.6, 110.8, 59.0, 55.7, 15.3; MS (EI) m/z (rel
intensity) 281 (100), 266 (27), 235 (8), 220 (9), 189 (11), 165 (7),
115 (5); HRMS (EI) m/z calcd for C17H15NO3 [Mþ] 281.1052,
found 281.1057.
Ethyl 3-(2-Nitrophenyl)-1H-indene-2-carboxylate (17g). The
general procedure for the synthesis of indenes from alcohol 6g
was followed. Purification by flash chromatography on silica gel
using 20% ethyl acetate in hexanes afforded the indene (14 mg,
69%) as a yellow oil that solidified upon standing: mp 89-90 °C;
IR (thin film) νmax 3062, 2981, 1698, 1526, 1348, 1248, 1193,
1111, 1097, 853, 755, 740 cm-1; 1H NMR (400 MHz, CDCl3) δ
8.20 (1H, dd, J=8.5, 1.5 Hz), 7.70 (1H, td, J=7.5, 1.0 Hz), 7.59
(2H, t, J=7.0 Hz), 7.36-7.42 (2H, m), 7.30 (1H, td, J=7.5, 1.0
Hz), 7.07 (1H, d, J=7.5 Hz), 4.06 (2H, q, J=7.0 Hz), 3.88 (2H,
s), 1.06 (3H, t, J=7.0 Hz); 13C NMR (75 MHz, CDCl3) δ 164.4
(C), 149.3 (C), 148.9 (C), 144.2 (C), 143.4 (C), 132.9 (CH), 131.6
(C), 131.0 (C), 130.8 (CH), 128.8 (CH), 128.1 (CH), 127.0 (CH),
124.4 (CH), 124.3 (CH), 121.9 (CH), 60.3 (CH2), 39.1 (CH2),
13.8 (CH3); MS (ESIþ) m/z (rel intensity) 332 (72), 264 (100),
220 (38), 218 (20); HRMS (ESI) m/z calcd for C18H15NO4
[M þ Na]þ 332.0893, found 332.0895.
Ethyl 1-(2-Chlorophenyl)-1H-indene-2-carboxylate (9j). The
general procedure for the synthesis of indenes from alcohol 6j
was followed. Purification by flash chromatography on silica gel
using a gradient elution of 5-10% ethyl acetate in hexanes afforded
the indene (60 mg, 64%) as white crystals: mp 44-45 °C; IR (thin
film) νmax 3062, 2979, 2887, 1709, 1569, 1472, 1367, 1331, 1296,
1244, 1190, 1078, 1035, 751 cm-1; 1H NMR (300 MHz, CDCl3)
δ 7.86 (1H, d, J = 2.0 Hz), 7.52 (1H, dd, J = 8.0, 1.0 Hz), 7.46
(1H, d, J=8.0 Hz), 7.23-7.35 (3H, m), 7.12 (1H, t, J=7.0 Hz),
6.98 (1H, t, J = 7.0 Hz), 6.53 (1H, d, J = 7.0 Hz), 5.56 (1H, s),
4.04-4.23 (2H, m), 1.14 (3H, t, J=7.0 Hz); 13C NMR (75 MHz,
CDCl3) δ 164.3, 150.2, 142.4, 141.6, 141.3, 136.8, 134.8, 129.9,
128.6, 128.2, 127.6, 127.3, 127.2, 124.5, 123.8, 60.5, 51.6, 14.2;
MS (EI) m/z (rel intensity) 300 (12), 298 (31), 263 (65), 235 (27),
227 (44), 226 (30), 225 (100), 190 (28), 189 (79); HRMS (EI) m/z
calcd for C18H15ClO2 [Mþ] 298.0760, found 298.0766.
Ethyl 1-(2-Methoxyphenyl)-1H-indene-2-carboxylate (9o).
The general procedure for the synthesis of indenes from alcohol
6o was followed. Purification by flash chromatography on silica
gel using 10% ethyl acetate in hexanes afforded the indene (31
mg, 69%) as a colorless oil: IR (thin film) νmax 3067, 2976, 2935,
2835, 1707, 1596, 1567,1491, 1461, 1367, 1244, 1190, 1078, 1027,
756 cm-1 1H NMR (300 MHz, CDCl3) δ 7.83 (1H, d, J =
;
2.0 Hz), 7.49 (1H, d, J=8.0 Hz), 7.14-7.30 (4H, m), 6.95 (1H, d,
J=7.5 Hz), 6.74 (1H, td, J=7.5, 1.0 Hz), 6.63 (1H, dd, J=7.5,
2.0 Hz), 5.45 (1H, d, J = 2.0 Hz), 4.05-4.22 (2H, m), 3.92
(3H, s), 1.16 (3H, t, J=7.0 Hz); 13C NMR (75 MHz, CDCl3) δ
164.4 (C), 157.5 (C), 150.8 (C), 141.7 (C), 141.6 (C), 141.5 (CH),
141.1 (C), 128.0 (CH), 127.8 (CH), 127.1 (CH), 126.9 (CH),
124.2 (CH), 123.3 (CH), 120.7 (CH), 111.3 (CH), 60.1 (CH2),
55.8 (CH3), 48.8 (CH), 14.1 (CH3); MS (ESIþ) m/z (rel intensity)
331 (30), 317 (84), 295 (10), 263 (24), 249 (100); HRMS (ESI) m/z
calcd for C19H18O3 [M þ Na]þ 317.1148, found 317.1147.
6-Methoxy-2-methyl-1-(4-nitrophenyl)-1H-indene (9z). The
general procedure for the synthesis of indenes from alcohol 6z
was followed with the exception that the reaction was heated to
100 °C for 15 min. Purification by flash chromatography on
silica gel using a gradient elution of 10-20% ethyl acetate in
hexanes afforded the indene (27 mg, 45%) as a pale brown oil:
Acknowledgment. The Natural Science and Engineering
Research Council (NSERC) of Canada supported this work.
C.D.S. thanks NSERC for a postgraduate scholarship (PGS D3).
We thank Dr. A. B. Young for mass spectrometric analysis.
Supporting Information Available: Experimental proce-
dures for the synthesis of enones, allylic alcohols, indenes, and
21, characterization data for all new compounds, and copies of
1H NMR and 13C NMR spectra. Tables of optimized structures,
energies, and Cartesian coordinates for ground-state and transi-
tion structures. This material is available free of charge via the
Internet at http://pubs.acs.org.
J. Org. Chem. Vol. 75, No. 14, 2010 4727