Photochemical C2-C6 cyclization of enyne-allenes: Detection of a fulvene triplet diradical in the laser flash photolysis

Article abstract of DOI:10.1021/jo801689w

(Chemical Equation Presented) A series of enyne-allenes, with and without benzannulation at the ene moiety and equipped with aromatic and carbonyl groups as internal triplet sensitizer units at the allene terminus, was synthesized. Both sets, the cyclohexenyne-allenes and benzenyne-allenes, underwent thermal C²-C⁶ cyclization exclusively to formal ene products. In contrast, the photochemical C²-C⁶ cyclization of enyne-allenes provided formal Diels-Alder and/or ene products, with higher yields for the benzannulated systems. A raise of the temperature in the photochemical cyclization of enyne-allene 1b′ led to increasing amounts of the ene product in relation to that of the formal Diels-Alder product. Laser flash photolysis at 266 and 355 nm as well as triplet quenching studies for 1b,b′ indicated that the C²-C⁶ cyclization proceeds via the triplet manifold. On the basis of a density functional theory (DFT) study, a short-lived transient (r = 30 ns) was assigned as a triplet allene, while a long-lived transient (τ = 33 μs) insensitive to oxygen was assigned as fulvene triplet diradical. An elucidation of the reaction mechanism using extensive DFT computations allowed rationalization of the experimental product ratio and its temperature dependence.

Full text of DOI:10.1021/jo801689w

Photochemical C2-C6 Cyclization of Enyne-Allenes: Detection of a
Fulvene Triplet Diradical in the Laser Flash Photolysis
Go¨tz Bucher,*,†,‡ Atul A. Mahajan,§ and Michael Schmittel*,§
Lehrstuhl fu¨r Organische Chemie II, Ruhr-UniVersita¨t Bochum, UniVersita¨tsstrasse 150,
D-44801 Bochum, Germany, and Organische Chemie I, FB 8 (Chemie-Biologie), UniVersita¨t Siegen,
Adolf-Reichwein-Strasse, D-57068 Siegen, Germany
goebu@chem.gla.ac.uk; schmittel@chemie.uni-siegen.de
ReceiVed July 30, 2008
A series of enyne-allenes, with and without benzannulation at the ene moiety and equipped with aromatic
and carbonyl groups as internal triplet sensitizer units at the allene terminus, was synthesized. Both sets,
the cyclohexenyne-allenes and benzenyne-allenes, underwent thermal C2-C6 cyclization exclusively
to formal ene products. In contrast, the photochemical C2-C6 cyclization of enyne-allenes provided
formal Diels-Alder and/or ene products, with higher yields for the benzannulated systems. A raise of
the temperature in the photochemical cyclization of enyne-allene 1b′ led to increasing amounts of the
ene product in relation to that of the formal Diels-Alder product. Laser flash photolysis at 266 and 355
nm as well as triplet quenching studies for 1b,b′ indicated that the C2-C6 cyclization proceeds via the
triplet manifold. On the basis of a density functional theory (DFT) study, a short-lived transient (τ ) 30
ns) was assigned as a triplet allene, while a long-lived transient (τ ) 33 µs) insensitive to oxygen was
assigned as fulvene triplet diradical. An elucidation of the reaction mechanism using extensive DFT
computations allowed rationalization of the experimental product ratio and its temperature dependence.
Introduction
clization)2 to highly reactive diradicals characterize the modus
operandi of the so-called natural enediyne antitumor antibiotics,
which through the abstraction of hydrogen atoms from DNA
ultimately ignite cell death.3 As a consequence, the design of
new trigger protocols for the activation of simple model
compounds has become an active field of research. Among the
various trigger mechanisms, the photochemical activation of
enediynes along the Bergman pathway4 has lately received great
Thermal cycloaromatizations of enediynes (Bergman cycliza-
tion)1 and enyne-allenes/enyne-cumulenes (Myers-Saito cy-
† Ruhr-Universita¨t Bochum.
‡ Current address: WestCHEM Research School, Department of Chemistry,
University of Glasgow, Joseph-Black-Building, University Avenue, Glasgow G12
8QQ, United Kingdom.
§ Universita¨t Siegen.
10.1021/jo801689w CCC: $40.75
Published on Web 10/15/2008
2008 American Chemical Society
J. Org. Chem. 2008, 73, 8815–8828 8815

Bucher et al.
SCHEME 1. Thermal and Photochemical Myers-Saito
(C2-C7) and C2-C6 Cyclizations of Enyne-Allenes
CHART 1. Model Compounds 1a,a′-1c,c′
interest due to its obvious potential for photodynamic therapy
(PDT).5 In contrast, despite a first report on the photochemical
Myers-Saito and C2-C6 cyclization in 2005 (Scheme 1),6 the
photochemical activation of enyne-allenes/enyne-cumulenes
is still a terra incognita.
Our original study on the photochemical Myers-Saito and
C2-C6 cyclization of enyne-allenes led to cyclohexenyne
allenes as model compounds,6 as theoretical investigations by
Engels et al.7 had suggested a high vertical excitation T1(S0)
energy for a benzannulated enyne-allene (101 kcal mol-1) and
a much lower one for (Z)-1,2,4-heptatrien-6-yne (59 kcal
mol-1).8 By utilizing proper internal triplet sensitizer units at
the allene terminus, we were able to ignite for the first time
photochemical cyclizations of cyclohexenyne-allenes.6 Parallel
first laser flash photolysis results on the photochemical C2-C6
cyclization pointed to the occurrence of two intermediates. As
the long-lived intermediate (τ ) 33 µs) did not react with 3O2,
n-Bu3SnH, and 1,4-cyclohexadiene, it was assigned to the singlet
fulvenyl diradical.6 However, it was difficult to reconcile such
a long lifetime with the low barriers computed for the possible
follow-up reactions.9 Hence, an extensive study on the identity
of the observable intermediates and on the details of the reaction
course seemed to be warranted.
thermal Myers-Saito and C2-C6 cyclization, we decided to
enlarge our original set of compounds to study the reaction
course thoroughly by laser flash photolysis and to clarify the
reaction course through extensive DFT computations. The
present combination of laser flash photolysis (LFP) and DFT
calculations provides a comprehensive picture of the mechanism
of the photochemical C2-C6 cyclization.
For the present investigations, we chose six enyne-allenes
with and without benzannulation at the ene moiety (Chart 1) to
elucidate whether photochemical cyclization would become
difficult with benzannulated representatives as predicted by
Engels.7 The TIPS group at the alkyne is important to avoid
thermal reactions during photolysis.13 While parts of the results
on the cyclohexenyne-allenes 1a-c have been reported in our
communication,6 all results concerning the benzannulated en-
yne-allenes 1a′-c′ are new.
Encouraged by the successful implementation of the photo-
chemical pathway and the wide knowledge about the effect of
benzannulation,10 regioselectivity,11 and aromaticity12 on the
Results and Discussion
(1) (a) Bergman, R. G.; Jones, R. R. J. Am. Chem. Soc. 1972, 94, 660–661.
(b) Bergman, R. G. Acc. Chem. Res. 1973, 6, 25–31.
(2) (a) Myers, A. G.; Kuo, E. Y.; Finney, N. S. J. Am. Chem. Soc. 1989,
111, 8057–8059. (b) Nagata, R.; Yamanaka, H.; Okazaki, E.; Saito, I. Tetrahedron
Lett. 1989, 30, 4995–4998.
(3) (a) Nicolaou, K. C.; Dai, W.-M. Angew. Chem., Int. Ed. Engl. 1991, 30,
1387–1416. (b) Borders, D. B.; Doyle, T. W. Enediyne Antibiotics as Antitumor
Agents; Marcel Dekker: New York, 1995.
(4) (a) Turro, N. J.; Evenzahav, A.; Nicolaou, K. C. Tetrahedron Lett. 1994,
35, 8089–8092. (b) Evenzahav, A.; Turro, N. J. J. Am. Chem. Soc. 1998, 120,
1835–1841. (c) Funk, R. L.; Young, E. R. R.; Williams, R. M.; Flanagan, M. F.;
Cecil, T. L. J. Am. Chem. Soc. 1996, 118, 3291–3292. (d) Kaneko, T.; Takahashi,
M.; Hirama, M. Angew. Chem., Int. Ed. 1999, 38, 1267–1268. (e) Choy, N.;
Blanco, G.; Wen, J.; Krishan, A.; Russell, K. C. Org. Lett. 2000, 2, 3761–3764.
(f) Clark, A. E.; Davidson, E. R.; Zaleski, J. M. J. Am. Chem. Soc. 2001, 123,
2650–2657. (g) Chandra, T.; Kraft, B. J.; Huffman, J. C.; Zaleski, J. M. Inorg.
Chem. 2003, 42, 5158–5172. (h) Benites, P. J.; Holmberg, R. C.; Rawat, D. S.;
Kraft, B. J.; Klein, L. J.; Peters, D. G.; Thorp, H. H.; Zaleski, J. M. J. Am.
Chem. Soc. 2003, 125, 6434–6446.
(5) (a) Boyle, R. W.; Dolphin, D. Photochem. Photobiol. 1996, 64, 469–
485. (b) Mody, T. M., Sessler, J. L., Eds. Supramolecular Materials and
Technologies; Wiley: Chichester, U.K., 1999; Vol. 4. (c) Ali, H.; van Lier, J. E.
Chem. ReV. 1999, 99, 2379–2450. (d) Bonnett, R.; Martinez, G. Tetrahedron
2001, 57, 9513–9547. (e) Ackroyd, R.; Kelty, C.; Brown, N.; Reed, M.
Photochem. Photobiol. 2001, 74, 656–669.
Synthesis of Enyne-Allenes. Sonogashira coupling of 214
and 2′ with triisopropylsilyl acetylene furnished aldehydes 3
and 3′ in good yield.15 Addition of BrMgCt C-n-Bu (prepared
from the reaction of EtMgBr with 1-hexyne) to 3,3′, provided
the corresponding propargyl alcohols 4,4′ in moderate yields16
that were subsequently subjected to acetylation using acetic
anhydride and DMAP in dry dichloromethane at room temper-
ature.17 Propargyl acetates 5,5′ were further reacted with ArZnCl
in the presence of Pd(PPh3)4 to afford the desired enyne-allenes
1a,a′ and 1b,b′ (Scheme 2) in moderate yields.18 Enyne-allenes
1c,c′ were obtained by another route as shown in Scheme 3.
Reaction of enyne-aldehyde 3 with PPh3, zinc, and carbon
tetrabromide in dry dichloromethane at room temperature
afforded the dienyne 6, which was further treated with n-
butyllithium (2.5 M in hexane) to afford enediyne 7.19 The
desired enyne-allene 1c was obtained from enediyne 7 after a
deprotonation using LDA and subsequent treatment with
2-bromo-3-pentanone at -78 °C. After an aqueous workup and
(6) Schmittel, M.; Mahajan, A. A.; Bucher, G. J. Am. Chem. Soc. 2005, 127,
5324–5325.
(7) Spo¨ler, C.; Engels, B. Chem.sEur. J. 2003, 9, 4670–4677.
(8) Private communication by Prof. B. Engels.
(9) (a) Musch, P. W.; Engels, B. J. Am. Chem. Soc. 2001, 123, 5557–5562.
(b) Bekele, T.; Christian, C. F.; Lipton, M. A.; Singleton, D. A. J. Am. Chem.
Soc. 2005, 127, 9216–9223. (c) Schmittel, M.; Vavilala, C.; Jaquet, R. Angew.
Chem., Int. Ed. 2007, 46, 6911–6914.
(10) Wenthold, P. G.; Lipton, M. A. J. Am. Chem. Soc. 2000, 122, 9265–
9270.
(13) Schmittel, M.; Strittmatter, M.; Mahajan, A. A.; Vavilala, C.; Cinar,
M. E.; Maywald, M. ARKIVOC 2007, 66–84 (Part viii).
(14) For the synthesis of compound 2 see: (a) Arnold, Z.; Holy, A. Collect.
Czech. Chem. Commun. 1961, 26, 3059–3073. (b) Salem, B.; Delort, E.; Klotz,
T.; Suffert, J. Org. Lett. 2003, 5, 2307–2310.
(15) Schmittel, M.; Steffen, J.-P.; Maywald, M.; Engels, B.; Helten, H.;
Musch, P. J. Chem. Soc., Perkin Trans. 2 2001, 1331–1339.
(16) Schmittel, M.; Strittmatter, M.; Schenk, W. A.; Hagel, M. Z. Naturforsch.
B 1998, 53, 1015–1020.
(11) Musch, P. W.; Remenyi, C.; Helten, H.; Engels, B. J. Am. Chem. Soc.
2002, 124, 1823–1828.
(12) Stahl, F.; Moran, D.; Schleyer, P. v. R.; Prall, M.; Schreiner, P. R. J.
Org. Chem. 2002, 67, 1453–1461.
(17) Fleming, I.; Terrett, N. K. J. Organomet. Chem. 1984, 264, 99–118.
(18) Elsevier, C. J.; Stehouwer, P. M.; Westmijze, H.; Vermeer, P. J. Org.
Chem. 1983, 48, 1103–1105. For the synthesis of 1a, see: ref 13.
(19) Herndon, J.; Wang, H. J. Org. Chem. 1998, 63, 4562–4563.
8816 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical C2-C6 Cyclization of Enyne-Allenes
SCHEME 2. Synthesis of Enyne-Allenes 1a,a′ and 1b,b′a
TABLE 1. Tonset from DSC Measurements of the Thermal
Cyclization of Enyne-Allenes 1a-c and 1a′-c′ and the Yield of
Their Cyclization Products
enyne-allene
Tonset (°C)
1b, 143
% yield
cyclohexenyne-
allenes
benzannulated
enyne-allenes
1a, 152
1a′, 128
1c, 151
1c′, 126
8a, 67; 8b,
68; 8c, 42
8a′, 91; 8b′,
83; 8c′, 80
1b′, 108
SCHEME 4. Thermal Cyclization of Enyne-allenes
1a,a′-c,c′
a Reagents and conditions: (i) H-Ct C-TIPS, Pd(PPh3)2Cl2, CuI, Et3N,
rt, 16 h, 3: 94%, 3′: 92%; (ii) EtMgBr, H-Ct C-nBu, THF, rt, 4: 78%,
4′: 88%; (iii) Ac2O; DMAP, Et3N, rt, 16 h, 5: 80%, 5′: 98%; (iv) Ar-Br,
nBuLi/ ZnCl2 (1 M in diethyl ether)/ Pd(PPh3)4, -60 °C to rt, 16 h, 1a:
40%, 1a′: 30%, 1b: 62%, 1b′: 48%.
SCHEME 3. Synthesis of Enyne-Allenes 1c,c′a
exothermic signal in differential scanning calorimetry (DSC)
experiments.
The DSC experiments showed Tonset for the cyclization of
cyclohex-enyne-allenes to be reproducibly higher than for its
benzannulated analogs (Table 1). For example, the temperature
for the thermal cyclization of 1b′ is lower by 35 °C than that
of 1b, thus emphasizing that benzannulation decreases the
activation barrier in agreement with the earlier theoretical
findings by Wenthold and Lipton10 and Engels.22
Photochemical Cyclization of Enyne-allenes 1a-c and
1a′-c′. Photolysis of enyne-allene 1a′ in dry degassed toluene
(1-2 mM; details see Table 2 and Scheme 5) in the presence
of 1,4-cyclohexadiene (1,4-CHD; 100-fold excess) provided the
formal Diels-Alder product 9a′ (8%) as the only isolable low
molecular weight compound. The 1H NMR spectrum showed a
characteristic singlet at 4.29 ppm for the CH2 unit of the fluorene
subunit and a benzylic triplet at 2.73 ppm of the -CH2Pr unit
(3J ) 7.9 Hz). The analogous irradiation of 1a did not lead to
any isolable product.6 Irradiation of enyne-allene 1b′ (1.27
mM) in dry degassed toluene furnished a mixture of two
isomeric ene products23 8b′ and (E)-8b′ in 32% yield along with
18% of the formal Diels-Alder product 10b′. In contrast, the
photochemical cyclization of cyclohexenyne-allene 1b provided
ene product 8b and formal Diels-Alder product 10b in a 1:1
ratio.6 In both 10b and 10b′, the TIPS group was lost, contrary
to the situation in 9a′. Recent experimental and computational
studies showed that the desilylation in the formal Diels-Alder
product takes place after a silyl shift in the primary Diels-Alder
adduct.24
a Reagents and conditions: (i) PPh3, Zn, CBr4, DCM, 40 h, rt, 6: 86%,
6′: 92%; (ii) n-BuLi, THF, -78 °C to rt, 3 h, 7: 75%, 7′: 71%; (iii) LDA,
HMPA, 2-bromo-3-pentanone, -78 °C, THF; (iv) SiO2, 1c: 42%, 1c′: 32%.
column chromatography, enyne-allene 1c was received in 42%
yield (Scheme 3). Mechanistic aspects of this tranformation have
been addressed earlier.20 The benzannulated enyne-allene 1c′
was prepared analogously.
Thermal Cyclization. In accord with the thermolysis of other
enyne-allenes containing bulky groups at the alkyne terminus,21a-c
the thermal cyclization of cyclohexenyne-allenes 1a-c and
benzenyne-allenes 1a′-c′ in dry toluene at 110 °C furnished
exclusively the formal ene products 8a-c and 8a′-c′ in mostly
good yields (Table 1, Scheme 4). In general, thermal product
formation from the benzannulated enyne-allenes was distinctly
cleaner. Only in the thermolysis of enyne-allene 1b′ a trace
amount (<1%) of the formal Diels-Alder product 10b′ was
formed.21c
The thermal reactivity of enyne-allenes 1a-c and 1a′-c′
was evaluated by measuring the temperature Tonset of the
Under analogous conditions, the photolysis of enyne-allene
1c′ furnished four isomers of the C2-C6 product in moderate
total yield (51%) (Scheme 6 and Table 2) that were separated
(20) Schmittel, M.; Strittmatter, M. Tetrahedron 1998, 54, 13751–13760.
(21) (a) Schmittel, M.; Strittmatter, M.; Kiau, S. Tetrahedron Lett. 1995,
36, 4975–4978. (b) Schmittel, M.; Kiau, S.; Strittmatter, M. Angew. Chem., Int.
Ed. 1996, 35, 1843–1845. (c) Schmittel, M.; Keller, M.; Kiau, S.; Strittmatter,
M. Chem.sEur. J. 1997, 3, 807–816. (d) Schmittel, M.; Kiau, S.; Siebert, T.;
Strittmatter, M. Tetrahedron Lett. 1996, 37, 7691–7694. (e) Engels, B.; Lennartz,
C.; Hanrath, M.; Schmittel, M.; Strittmatter, M. Angew. Chem., Int. Ed. 1998,
37, 1960–1963.
(22) Musch, P. W.; Engels, B. J. Am. Chem. Soc. 2001, 123, 5557–5562.
(23) E-8b′ arises most likely from the irradiation of the photolabile C2-C6
product 8b′. This was tested by independent irradiation of 8b′ (isolated from
the thermal reaction).
(24) Schmittel, M.; Mahajan, A. A.; Bucher, G.; Bats, J. W. J. Org. Chem.
2007, 72, 2166–2173.
J. Org. Chem. Vol. 73, No. 22, 2008 8817

Bucher et al.
TABLE 2. Yields of Photochemical C2-C6 Product after Irradiation at 300 nm
compd (conc, mM)
solvent (irradiation time, h)
λ
max/nm (in hexane)
ene product (yield, %)d
DA product (yield, %)d
1a (1.78)a
1a′ (1.14)
1b (1.36)a
1b′ (1.27)
1c (5.28)a
1c′ (1.69)
toluene + 1,4-CHDb (6)
toluene + 1,4-CHDb (1)
toluene + 1,4-CHDb (6)
toluene + 1,4-CHDb (6)
n-hexane + 1,4-CHDc (14)
n-hexane + 1,4-CHDb (36)
280, 297
277, 283, 293
281, 318
266, 318
279, 290 (sh)
236, 272, 282 (sh)
8
15
18
15
32
32e
51
a Reference 6. b In the presence of 1,4-CHD (100-fold excess); DA- formal Diels-Alder. c 2:1 mixture. d Isolated yield. e Isolated yield: 8c′ (29%),
(E)-8c′ (16%), 8c′′ (6%).
SCHEME 5. Photochemical Cyclization of Enyne-Allenes 1a′, 1b, and 1b′
SCHEME 6. Photochemical Cyclization of Enyne-Allenes 1c and 1c′
into three fractions representing 8c′, (E)-8c′, and 8c′′ (the latter
as a mixture of two isomers). Compound 8c′ was the main
product with 29% yield. Independent irradiation of 8c′, readily
produced from thermal cyclization of 1c′, at 300 nm in dry
degassed n-hexane furnished the photoproducts (E)-8c′ and 8c′′.
This suggests that products (E)-8c′ and 8c′′ emerging during
photolysis of 1c′ equally arose in a follow-up photo process
from the photolabile product 8c′. The geometry at the exo-
methylene double bond of 8c′ was clearly assigned on the basis
of NOESY spectra. The aromatic proton Ha in 8c′ showed a
diagnostic NOE signal with the TIPS group; consequently,
compound 8c′ has a (Z)-geometry at the fulvenyl exo-double
bond. While the configurations at all stereogenic double bonds
in (E)-8c′ could be clarified by NOE investigations, this was
not possible for the isomeric mixture of 8c′′.25
Since the laser flash photolysis experiments (vide infra) po-
inted toward a wavelength dependence of the reaction mecha-
nism, we studied also the preparative 350-nm photolysis of
enyne-allene 1b′ in dry toluene in the presence of 1,4-
(25) Despite the use of g-NOESY and g-ROESY NMR techniques the precise
geometry at the exo-methylene double bond of C2-C6 product (9d-1) could not
be clarified.
8818 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical C2-C6 Cyclization of Enyne-Allenes
TABLE 3. Triplet-Quenching Reaction of Enyne-Allenes (1 mM)
TABLE 4. Photoreactivity of Enyne-Allene 1b′ at Different
Temperatures
irradiation
time (h)
C2-C6 product
(yield, %)
compd
triplet quencher
irradiation
time (h)
ene producta DielssAlder producta
solvent
T/°C
8b′
10b′ (%)
1b
1b
1b′
1,4-diphenylbutadienea
7
7
3
0
0
b
3O2
toluene
toluene
toluene
6
3
1.5
20 ( 1
40 ( 1
60 ( 1
32
27
27
18
8
trace
b
3O2
12 (ene)
b
a 15 equiv of 1,4-diphenyl-1,3-butadiene was used. 3O2 was con-
tinuously bubbled through the reaction mixture.
a Isolated yield.
study (vide infra), the preparative results suggest that at higher
temperature two rotameric fulvenyl diradicals may interconvert
(Scheme 7).
Laser Flash Photolysis of 1b. In order to gain information
on the reaction mechanism and to identify possible intermediates
in photochemical enyne-allene C2-C6 cyclizations, we have
performed a LFP (laser flash photolysis) study of 1b and 1b′.
Results obtained with enyne-allene 1b had previously been
published in preliminary form;6 they will be presented in greater
detail here. As reported earlier, LFP of 1b (λexc ) 355 nm,
cyclohexane solution, 1 atm of Ar) resulted in the observation
of transient phenomena on both short (ns) and longer (µs) time
scales. On the short time scale, a transient (λmax ) 440 and 505
nm, τ ) 30 ns; see Figure S1 in the Supporting Information
and also Figure 3) was observed that was quenched by 3O2 (kq
) (1.8 ( 0.3) × 109 l mol-1 s-1; see Figure S2 in the Supporting
Information). It is not quenched by 1,3-cyclohexadiene. Based
on its high reactivity toward triplet oxygen, we had previously
assigned this short-lived transient to 31b, the first triplet excited-
state of 1b.6
FIGURE 1. Transient spectra, observed during different time intervals
after LFP (355 nm) of enyne-allene 1b′ in degassed cyclohexane. Black
circles: 8 ns after LFP, light circles: 43 ns after LFP, black triangles:
380 ns after LFP. Inset: transient trace, monitored at λ ) 410 nm.
On a longer time scale, a second transient was observed (λmax
) 480 nm, τ ) (33 ( 5) µs, see Figure 5). This transient was
not measurably quenched by 1,4-cyclohexadiene (up to 0.5%
by volume) or tri-n-butylstannane. Purging the solution with
oxygen resulted in a decrease in transient intensity, while its
lifetime remained constant. A plot of the inverse transient
intensity(1/∆OD)vs.oxygenconcentration(SVplot:Stern-Volmer
plot) was rather noisy because of poor signal-to-noise ratio due
to the small transient intensity and the use of a baseline
compensator. Nevertheless, a linear correlation was observed,
with kq τ ) 17550/169 ) 104. With kq ) 1.8 × 109 L mol-1
s-1, the lifetime of the precursor transient to the long-lived
transient is obtained as τ ) 58 ns. Given the rather noisy SV
plot, this is in reasonable agreement with the directly measured
cyclohexadiene. The reaction furnished both ene (38%, two
isomers) and DA (16%) products, an outcome very similar to
that after irradiation at 300 nm. Additionally, two side products
were isolated whose H NMR and ESI-MS unfortunately did
not allow for a final structural assignment.
1
Mechanistic Experiments. When the photochemical cycliza-
tion of enyne-allene 1b was carried out in the presence of triplet
26
quenchers such as 1,4-diphenyl-1,3-butadiene (42 kcal mol-1
)
or 3O2, the cyclization of 1b was completely suppressed. Along
the same line, the photochemical cyclization of enyne-allene
1b′ in the presence of O2 furnished a lower yield of the ene
3
product 12% (Table 3). Hence, the triplet quenching studies
strongly suggest that the photochemical C2-C6 cyclization
proceeds along the triplet manifold.
3
value τ ) 30 ns for the short-lived transient assigned to 1b.
The effect of temperature (20-60 °C) on the photochemical
C2-C6 cyclization was studied with enyne-allene 1b′. With
cyclohexenyne-allene 1b, the same investigation was not
feasible due to its thermal instability at 40-60 °C. Photolysis
(300 nm) of 1b′ at 20 °C afforded ene and Diels-Alder products
in 32% and 18% yield. The data (Table 4) show that with
increasing temperature the yield of 8b′ roughly remained
constant, while the yield of the formal DA product 10b′ declined
drastically from 18% (20 °C) to 8% (40 °C) and finally <2%
(60 °C). While the decreasing yield of the Diels-Alder product
might imply that in compensation the yield of the ene product
should augment with increasing temperature, we did not observe
such trend. This is possibly due to an instability of the ene
product at 40-60 °C under photochemical conditions. Although
a final analysis has to await the outcome of the laser photolysis
This observation suggests that the short-lived transient is a direct
precursor to the long-lived transient.6 (Figure S3 shows transient
spectra in the microsecond regime observed upon LFP at 355
nm of 1b, Figure S4 displays a transient difference spectrum,
and Figure S5 shows the Stern-Volmer plot; see the Supporting
Information).
Previously, we had assigned the 480 nm transient to a singlet
6
3
diradical formed from 1b by C2-C6 cyclization. This pre-
liminary assignment had been based on the lack of direct
reactivity of the 480 nm - transient toward oxygen. However,
calculations now indicate that the ground-state of silyl-func-
tionalized fulvene diyls possibly is a triplet; see below.
LFP of 1b in cyclohexane at λexc ) 266 nm employing the
fourth harmonic of a Nd:YAG laser gave results that were
similar to the results obtained using λexc ) 355 nm.
Laser Flash Photolysis of 1b′. As in the case of cyclohex-
enyne-allene 1b, two transients were observed for the ben-
zannulated enyne-allene 1b′ using λexc ) 355 nm. The first
(26) Handbook of photochemistry, 2nd ed.; Murov, S. L., Carmichael, I.,
Hug, G. L., Eds.; Marcel Dekker/CRC Press: Boca Raton, 1993.
J. Org. Chem. Vol. 73, No. 22, 2008 8819

Bucher et al.
FIGURE 3. Comparison of calculated UV/vis spectra (TD-UB3LYP/
6-31+G*//UB3LYP/6-31G*) of the four isomeric triplet states of 1B
with the experimental transient spectrum observed 16 ns after LFP (355
nm) of 1b in cyclohexane (cf. Figure S1, Supporting Information). Red:
syn-E-1B-triplet II. Black: syn-Z-1B-triplet II. Magenta: anti-E-1B-
triplet II. Light blue: anti-Z-1B-triplet II. Green: anti-lk-1B-triplet I.
Yellow: anti-ul-1B-triplet I. Dark blue: experimental spectrum, scaled
to match the calculated spectra in intensity.
FIGURE 2. Transient spectra, observed during different time intervals
after LFP (355 nm) of enyne-allene 1b′ in degassed cyclohexane. Black
circles: 7.9 µs after LFP, light circles: 34 µs after LFP, black triangles:
150 µs after LFP. Inset: transient difference spectrum (7.9 µs minus
150 µs).
short-lived transient (τ ∼ 30 ns) showed maxima at λ ) 410
and 490 nm along with a very weak maximum at 660 nm
(Figure 1). This transient was not measurably quenched by 3O2.
Due to the very fast decay kinetics in the absence of oxygen,
this observation places an upper limit of kO2 ≈ 5 × 108 L mol-1
s-1 on the oxygen-quenching reaction, while a kO2 smaller than
this value is not ruled out. The second transient exhibited a much
at the reactivity pattern and transient spectra, it appears that
the transients observed upon 266 nm excitation of 1b′ are similar
to the transients observed upon LFP of 1b. In both cases, a
transient with a lifetime in the microsecond regime is observed,
with spectroscopic properties that are similar (compare Figures
S3 and S6; see the Supporting Information). Both 1b and 1b′
(266 nm excitation) also yield a short-lived transient with a
lifetime of the order of 30-60 ns. In both cases, this transient
is quenched by oxygen, as is shown both directly and by the
Stern-Volmer quenching observed. It is thus tempting to
conclude that 1b′ upon 266 nm excitation, like 1b, yields a
fulvenediyl-type diradical whose formation is preceded by decay
of an allene-type triplet state. What, however, would then be a
possible assignment for the transients observed upon 355 nm
LFP of 1b′? We will discuss this issue later in context with the
results of our calculations.
longer lifetime (τ ) 33 µs). Its absorption maximum (λmax
)
520 nm) was red-shifted compared to that of the long-lived
transient (λmax at 470 nm) observed upon excitation of cyclohex-
enyne-allene 1b. The long-lived transient also did not show
3
any measurable direct reactivity toward O2.
Figure 1 displays the spectrum of the short-lived transient
observed upon LFP of 1b′. Figure 2 shows a transient
spectrum of the long-lived transient. Its absorption maximum
is most easily discerned in the transient difference spectrum
(black circles minus black triangles) as provided in the inset
of Figure 2.
We also investigated the photochemistry of 1b′ in cyclohex-
ane using the fourth harmonic of a Nd:YAG laser (λexc ) 266
nm) as excitation light source. Under these conditions, two
weakly absorbing transient species were observed which were
different from the ones detected upon 355 nm excitation (Figures
S6-S8, Supporting Information). A short-lived transient was
found to have a lifetime τ ) (60 ( 10) ns with λmax ) 415 nm
and a broad band extending to λ ) 700 nm and probably
beyond. (Figure S7, see the Supporting Information). The
second, longer-lived transient (τ ) (3.0 ( 0.5) µs, λmax ) 395,
485 nm and a broad band extending beyond 650 nm, see Figure
S6 in the Supporting Information) was not directly quenched
by oxygen. Stern-Volmer (SV) quenching of this species by
3O2, on the other hand, was observed (Figure S9, see the
Supporting Information). From the SV plot, the product of
lifetime and rate constant of decay of the precursor to the 395/
485 nm transient was derived as kq * τ ) 86. Assuming τ ) 60
ns, kq follows as kq ) 1.4 * 109 L mol-1 s-1. As this is a
reasonable value for oxygen quenching of an excited-state or a
reactive diradical, it appears likely that the precursor transient
detected indirectly by SV-quenching in fact is the short-lived
415 nm transient also monitored directly.
Calculations of Triplet Allenes. Energies and UV/vis Spec-
tra. In order to come to more reliable assignments of the
transient species observed in the two sets of experiments,
we have employed density functional theory ((U)B3LYP/6-
31G(d)) to optimize the geometries of a number of species
and to predict some of their properties of interest. Calculated
UV/vis spectra should prove particularly helpful in confirming
or disproving an assignment. We therefore employed time-
dependent DFT (TD-UB3LYP/6-31+G(d)//UB3LYP/6-31G(d))
to calculate the UV/vis spectra of the triplet excited states
of a series of model compounds for transients derived from
1b and 1b′. In order to stay as close as possible to the
experimental system, only the remote bromine atom substit-
uents were omitted in our model system, and the n-butyl
substituent was replaced by methyl, resulting in compounds
1B (as model for 1b) and 1B′ (as model for 1b′). DFT
calculations on the model precursors showed that the syn-
conformers were higher in energy than the anti-conformers,
as expected (Scheme 8). However, RIMP2 single-point ene-
rgy calculations (based on the geometries obtained by DFT)
indicated that DFT underestimates the stability of the syn-
conformers by ca. 6 kcal mol-1. At the RIMP2 level of
theory, the syn-conformers were actually predicted to be
energetically more favorable than the anti-conformers. This
The observation of two entirely different sets of transients
upon 355 and 266 nm excitation of 1b′ is puzzling. If one looks
8820 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical C2-C6 Cyclization of Enyne-Allenes
SCHEME 7. Preliminary Summary of LFP Results of 1b and 1b′
kcal mol-1 for the endothermic conversion of anti-Z-1B triplet
II into syn-Z-1B triplet II. Attempts to optimize the transition
states of a direct interconversion of E- and Z-isomers yielded
the alkene-type triplet I isomers. While their relatively low
energy may thus facilitate a direct Z-/E-equilibration of triplet
excited 1B, we note that barriers of the order of 10 kcal mol-1
for such a reaction are significant considering the inherently
short lifetimes of the triplet states involved. Finally, we note
that an optimization of triplet 1B using the optimized geometry
of singlet ground state anti-1B as starting geometry yielded the
higher energy twisted alkene triplet I state of 1B rather than
the allene-type triplet II state.
SCHEME 8. Conformers of Model Precursors 1B and 1B′;
Energies in kcal mol-1 Relative to the Anti Conformers (Top
(roman font): B3LYP/6-31G(d). Bottom (Italics): RIMP2/
TZVP//B3LYP/6-31G(d))
In case of the benzannulated system 1B′, no twisted alkene-
type triplet states could be located, which is unsurprising con-
sidering the loss of aromaticity connected with such a structure.
As in the case of 1B, four stereoisomeric allene-type triplet states
(triplet II) could be located. Again, the triplet enthalpies are
consistently around 40 kcal mol-1 (Scheme 10). Due to the
absence of accessible alkene-type triplet states, interconversion
of Z- and E-stereoisomers is predicted to be a single-step
reaction in this system. Because of its significant barrier an E-/
Z-interconversion is unlikely to take place during the very short
lifetime of triplet 1B′. In order to test for the influence of the
DFT method, additional calculations using the UBLYP method
were performed here. As a result, we note that changing the
functional from B3LYP to BLYP does essentially nothing to
the enthalpies calculated.28b
is likely due to the fact that weak long-range interactions
are poorly described by DFT.
In the case of 1B, two fundamentally different triplet excited
states could be located (Scheme 9). The first of the two is of a
triplet alkene-type (triplet I), with a strongly twisted double bond
(θ ) -56.9°) in the cyclohexene moiety. Excitation to triplet I
hence introduces helicity into the molecule, which combines
with the axial chirality of the allene unit. Hence, lk and ul
diastereomers of 1B-triplet I should exist in both syn- and anti-
conformations. According to our calculations, the triplet energy
is ET ≈ 43 kcal mol-1 for all diastereomers. The second triplet
excited-state located for 1B (triplet II) is of the triplet allene-
type (allyl-vinyl diradical) with a nonlinear allene moiety. Four
stereoisomers of triplet II are predicted to exist, with calculated
triplet energies somewhat below 40 kcal mol-1 (Scheme 9).
Triplet energies somewhat below 40 kcal mol-1 are consistent
with published experimental triplet energies of arylallenes, e.g.,
the triplet energy of tetraphenylallene27 has been reported as
Assignment of the Short-Lived Intermediates of 1b and
1b′. The calculated UV/vis spectra of six isomeric triplet states
of 1B are given in Figure 3, along with a scaled experimental
spectrum (in dark blue). It is clearly seen that only the
energetically lower lying allene-type triplets (triplet II, see
SCHEME 9) yield a reasonable fit with the experimental
spectrum, while the alkene-type triplets (triplet I) are not
predicted to show any significant absorption beyond λ ) 440
nm. As the calculated spectra for all stereoisomers of 1B-triplet
(28) (a) Errors in calculated spectra are larger at long wavelengths (an energy
difference corresponding to 20 nm wavelength difference at λ ) 200 nm amounts
to 100 nm wavelength difference at λ ) 1000 nm). Our omission of the remote
bromine atom substituents may also contribute to deviations between calculated
and experimental spectra. (b) This statement also holds, if the biradicals and
final products are included into the consideration. For the SP energies, see the
Supporting Information. (c) Morina, V. F.; Sveshnikova, E. B. Opt. Spectrosc.
1973, 34, 359–360.
ET ) (40.2 ( 0.7) kcal mol-1
.
The activation enthalpy for interconversion of the syn- and
anti-stereoisomers of 1B triplet II is calculated as ∆Hq ) 12.3
(27) Brennan, C. M.; Caldwell, R. A.; Elbert, J. E.; Unett, D. J. J. Am. Chem.
Soc. 1994, 116, 3460–3464.
J. Org. Chem. Vol. 73, No. 22, 2008 8821

Bucher et al.
SCHEME 9. Different Triplet Excited State Isomers of Allene 1B, Calculated (UB3LYP/6-31G(d)) Triplet Enthalpies in kcal
mol-1, Relative to the Singlet Ground State of anti-1B ) 0.0 kcal mol-1 (The Rightmost Arrow Connects anti-ul-1B-Triplet I
and syn-ul-1B-Triplet I)
SCHEME 10. Different Allene-Type Triplet Excited State
(Triplet II) Isomers of 1B′, Calculated (UB3LYP/6-31G(d))
Triplet Enthalpies in kcal mol-1, Relative to the Singlet
Ground State of anti-1B′ ) 0.0 kcal mol-1 (Italics:
Calculated at the UBLYP/6-31G(d) Level of Theory)
upon 355 nm LFP of 1b′ to one or two anti-stereoisomers of
triplet-II-1b′ seems plausible.
The agreement between the calculated triplet II spectra shown
in Figure 4 and the experimental transient spectrum (Figure S6,
Supporting Information) of the short-lived transient observed
upon 266 nm excitation of 1b′, on the other hand, is poor. First,
the 400 nm band is calculated to be extremely intense for all
triplet II state stereoisomers of 1B′. While the short-lived
transient observed upon 266 nm excitation of 1b′ does have a
band in this region, it is rather weak. Second, all triplet II state
stereoisomers of 1B′ are predicted to show a prominent
absorption either around 480 nm (anti) or around 550 nm (syn).
There is no such band in the experimental spectrum shown in
II are similar, an unambiguous assignment cannot be made based
on the TD-DFT calculations.
A comparison of the calculated UV/vis spectra of the triplet
II state stereoisomers of 1B′ with the experimental transient
spectrum recorded upon LFP (355 nm excitation) of 1b′ shows
that the agreement is reasonable with anti-stereoisomers (Figure
4). The blue (exptl) and yellow or green (calcd) spectra are
similar in overall shape (strong absorption at λ ≈ 400 nm with
a weaker, not baseline-separated band at λ ≈ 500 nm), only
the long-wavelength near-IR absorption calculated is not
observed experimentally.28a The syn-conformers, on the other
hand, should show a well-separated band between λ ) 500 and
600 nm, which is not observed in our experiments. Given this
agreement, an assignment of the short-lived transient observed
FIGURE 4. Comparison of the experimental transient spectrum of the
short-lived intermediate (see also Figure 1) observed upon 355 nm LFP
of 1b′ (blue) with the calculated UV/vis spectra of triplet-state
stereoisomers of 1B′. Black: syn-E-1B′ triplet II. Red: syn-Z-1B′ triplet
II. Green: anti-E-1B′ triplet II. Yellow: anti-Z-1B′ triplet II. Blue:
experimental spectrum (cf. Figure 1), scaled to match the calculated
spectra in intensity.
8822 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical C2-C6 Cyclization of Enyne-Allenes
may argue against such an assignment, we note that certain (π
f π*) triplet states of aromatic ketones have been reported to
3
show reduced reactivity toward O2. Thus, triplet (π f π*)
4-phenylbenzophenone, with a triplet energy of 60.8 kcal
mol-1
,
25 reacts with molecular oxygen with a rate constant kO2
28c
) 7.8 × 107 M-1 s-1
.
This value is well below the upper
limit of kO2 ≈ 5 × 108 M-1 s-1 determined for the reaction of
the short-lived transient observed upon LFP (355 nm) of 1b′
with oxygen.
We also conclude that the spectrum of the short-lived transient
observed upon LFP (266 nm) of 1b′ is not reproduced well by
the calculations of the triplet state stereoisomers of 1B′ (see
also Figure 6). This issue will be taken up again later in the
manuscript.
Calculations: Reactivity of Triplet Allenes and Follow-
Up Reactions. In order to understand the role of the allene triplet
II state for the formation of 8b,b′ and 10b,b′ via the putative
intermediates 11b,b′, we have also done extensive computations
on the follow-up reactions of isomers E-1B,B′-triplet II and
Z-1B,B′-triplet II. In all cases, formation of the fulvene-diyl type
diradicals should occur from the syn- isomers of the 1B,B′-
triplet II, since the vinyl-type radical is in a good position to
add to the triple bond of the alkyne moiety. Syn-E-1B-triplet II
is expected to yield the fulvene diyl E-11B, in either the triplet
or singlet spin state, while isomer syn-Z-1B-triplet II should
yield fulvene diyl Z-11B, again either as singlet or as triplet
species. Diradical Z-11B in its singlet spin manifold should
further undergo ring closure to the Diels-Alder product 12B,
which is a well-known precursor24 to 10B (equivalent to 10b),
while E-11B is set to undergo a 1,5-hydrogen shift to yield the
ene-product 8B. Analogous reactions can be formulated for
precursor 1B′ (Scheme 11). Scheme 11 does not include the
anti-conformers of the allene-type triplet states. A similar
scheme for 1B′, calculated at the BLYP/6-31G(d) level of theory
gives results that are very similar to the B3LYP results (see the
Supporting Information, Scheme S1).
FIGURE 5. Comparison of calculated UV/vis-spectra (TD-UB3LYP/
6-31+G*//UB3LYP/6-31G*) of the singlet and triplet states of 11B
with the experimental difference spectrum of the long-lived transient
observed after LFP (355 nm) of 1b in cyclohexane (cf. Figure S4, see
the Supporting Information). Red: Z-11B-triplet. Black: Z-11B-singlet.
Green: experimental transient difference spectrum, scaled to match the
calculated spectra in intensity.
In cyclohexenyne-allene 1B, the syn-E-conformer of the
precursor triplet II excited-state is predicted to be exceedingly
short lived, with a calculated activation enthalpy for ring-closure
to the E-fulvene diyl triplet diradical of only 2.9 kcal mol-1
.
The singlet state of E-11B is predicted to be slightly higher in
energy than the triplet state (∆HS-T ) 1.8 kcal mol-1). A
transition state for the hydrogen transfer reaction yielding
fulvene 8B could be located, but the activation enthalpy is very
small. If a spin-projection correction is employed, it disappears
completely. Given the very small barrier for formation of E-11B
and the negligible barrier for its decay, it appears highly unlikely
that any intermediate along the route from syn-E-1b triplet to
the fulvene product 8b should be detectable in our experiments.
The cyclization of syn-Z-1B-triplet II, on the other hand, is
predicted to be hindered by an activation enthalpy. As before
in the E-manifold, the singlet state of diradical Z-11B is foreseen
to be higher in energy than the triplet state (∆HS-T ) 3.5 kcal
mol-1). From the singlet diradical, cyclization to 12B can occur
FIGURE 6. Comparison of calculated UV/vis spectra (TD-UB3LYP/
6-31+G*//UB3LYP/6-31G*) of the first triplet excited-state of 15 with
the experimental transient difference spectrum observed after LFP (266
nm) of 1b′ in cyclohexane (cf. Figure S7, Supporting Information).
Black circles: Calculated spectrum of triplet 15. Light circles: experi-
mental transient spectrum, scaled to match the calculated spectra in
intensity.
Figure 6. Third, the relative intensities at λ ) 400 and 650 nm
are very different in the calculated and experimental spectra.
To conclude this section, we may thus state that TD-DFT
calculations give good support to the assignment of the short-
lived transient observed upon LFP of 1b and 1b′ (at 355 nm)
to an allene-type triplet excited state. Furthermore, the good
agreement of the calculated spectra of anti-1B′ triplets II with
the experimental spectrum of the short-lived transient observed
upon LFP (355 nm) of 1b′ suggests the existence of the anti-
triplet II conformers. While the lack of reactivity toward oxygen
with a calculated activation enthalpy of ∆Hq ) 5.3 kcal mol-1
.
Interconversion of E- and Z-11B diradicals should be relatively
facile on the singlet hypersurface, with activation enthalpies ∆Hq
) 6.4 kcal mol-1 from the energetically lower-lying E-
conformer and ∆Hq ) 4.8 kcal mol-1 from the Z-conformer,
while the barrier is predicted to be higher on the triplet hyper-
surface (∆Hq ) 9.1 kcal mol-1). Detectable intermediates in
our experiments are thus possibly the fulvenediyls Z-11b.
J. Org. Chem. Vol. 73, No. 22, 2008 8823

Bucher et al.
SCHEME 11. Reactions of Enyne-allene Triplet States with Calculated Reaction Enthalpies and Activation Enthalpies
((U)B3LYP/6-31G(d)). Energies (in kcal mol-1) Relative to the Z-syn Precursor Allene-Type Triplet II (Upper Numbers Refer to
the Cyclohexenyne-Allene 1B; Bottom Numbers (Underlined) Refer to the Benzene Derivative 1B′)
In case of 1B′, the energetically lower-lying syn-Z-1B′-triplet
II is predicted to show a somewhat higher barrier (∆Hq ) 5.5
kcal mol-1) toward addition to the alkyne Ct C bond, as
compared to the E-isomer syn-E-1B′-triplet II (∆Hq ) 3.7 kcal
mol-1). In this system, the singlet- and triplet-fulvenediyl-type
diradicals are practically degenerate in energy. Hence, irrespec-
tive of the spin manifold, the barrier for interconversion of the
Z- and E-diradicals is calculated as ∆Hq ) 8.0 kcal mol-1 from
the energetically lower-lying E-conformer and ∆Hq ) 4.7 kcal
mol-1 from the Z-conformer. The activation enthalpies for the
follow-up reactions of the singlet diradicals parallel those
predicted for system 1B; cyclization of Z-11B′ yielding 12B′ is
predicted to be much slower (∆Hq ) 8.8 kcal mol-1) than
H-transfer from E-11B′ yielding 8B′ (∆Hq ) 3.3 kcal mol-1).
According to the calculations, the preferred reaction path upon
triplet excitation of 1B′ would thus involve formation of syn-
Z-1B′-triplet II, followed by intramolecular addition to the Ct C
triple bond yielding Z-11B′ with nearly degenerate triplet and
singlet states. Diradical Z-11B′ is expected to relax to the more
favorable E-11B′. Due to the small barrier for H-transfer E-11B′
(and thus likely also E-11b′) should be only present in
quasistationary concentrations. Detectable intermediates in our
experiments are thus only anti-conformers of 1b′-triplet II and
Z-11b′.
It is interesting to note that the DFT results also agree in a
qualitative manner with the experimental ene to Diels-Alder
product ratios and their temperature dependence. The kinetically
controlled product 8b,b′: 10b,b′ ratio should be related to the
barriers of the rate determining steps, i.e., that of Z-11B,B′
singlet f E-11B,B′ singlet versus that of Z-11B,B′ singlet f
12B,B′. Indeed, for 11B the computed barriers (4.8 vs. 5.3 kcal
mol-1) are almost identical (expt: 8b:10b ) 1:1), while for 11B′
the computed barrier of the Diels-Alder product formation is
significantly higher by 4.1 kcal mol-1 (expt: 8b′:10b′ ) ca. 2:1).
Equally, the temperature dependence of the ene to Diels-Alder
product ratio 8b′:10b′ is reproduced by the computations. The
increasing amount of ene product 8b′ formed at higher tem-
perature is due to a positive activation entropy of the Z-11B′
singlet f E-11B′ singlet reaction and a negative activation
entropy of the Z-11B′ singlet f 12B′ step (see Table S1,
Supporting Information).
Assignment of the Long-Lived Transient in the Photoly-
sis of 1b and 1b′ (266 nm). Our calculations suggest that the
short-lived species observed in the experiments with 1b ought
to be an anti-conformer of an allene-type triplet II excited state
of 1b, while the long-lived transient would be either the singlet
or the triplet state of Z-11b. Using the calculated reactivity data,
the long-lived transient can be assigned to triplet Z-11b because
its effective barrier to 12b, consisting of the S-T energy gap
and the barrier for ring closure, would be of the order of 9 kcal
mol-1, in agreement with a lifetime of 33 µs. Assigning this
transient to triplet Z-11b, however, holds the problem that a
triplet diradical would be expected to react rapidly with oxygen.
In order to address the issue of spin multiplicity of the diradical
observed, we have performed a series of calculations (UB3LYP/
cc-pVTZ and UBLYP/cc-pVTZ) on fulvene diyl diradicals 13
bearing a variety of substituents (R ) F, Cl, Br, H, Me, H3Si,
Li) and compared those with calculations on Z-11B.
The fact, that carbene-type mesomeric structures exist11 is
of particular relevance, for it implies that rules pertaining to
singlet-triplet energy gaps in carbenes might also apply. Table
5 gives the calculated singlet - triplet energy gaps along with
the important angle around the carbenic carbon atom. The data
clearly indicate that the fulvene diyls investigated here follow
the rules accepted for carbenes.29a σ-Acceptor substituents like
the halogen atoms result in stabilization of the singlet spin
manifold, while electron-donating substituents stabilize the
triplet. The angles around the “carbenic” carbon atom also
follow the trend observed for carbenes, with small angles for
ground-state singlets and larger angles for ground-state triplet
species. The data in Table 5 also show that B3LYP, in
comparison to BLYP, places the triplets lower in energy. The
preference for the triplet state is calculated to be smaller in
Z-11B than in the model diradicals. Nevertheless, both func-
tionals employed yield a triplet ground-state for Z-11B.
8824 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical C2-C6 Cyclization of Enyne-Allenes
TABLE 5. Singlet-Triplet Energy Gap for Fulvene Diyls, As
Calculated at the UB3LYP/cc-pVTZ or UBLYP/cc-pVTZ Levels of
Theory, Except for Z-11B (6-31G(d)-Basis) (The Singlet Energies
Have Been Corrected for Spin Projection Error. Positive
Singlet-Triplet Energy Gaps Imply a Triplet Ground State. Angle
around the “Carbene” Atom (UBLYP))
upon LFP of 1b. While the essential features of the experimental
spectrum (double band with λ ) 410, 480 nm in the visible
range of the spectrum) in principle are in agreement with the
calculated spectra, the relative intensities of the two bands and
their spacing are not reproduced well. The comparison does not
allow for a conclusive assignment of the spin state of the
diradical observed. It must be noted, however, that the experi-
mental spectrum shown in green in fact is a transient difference
spectrum (cf. Figure S4, see the Supporting Information). Any
absorption of a product formed concomitantly with the decay
of Z-11b would change relative intensities or shift maxima.
The long-lived transient observed upon LFP (266 nm) of 1b′
has a UV/vis-spectrum (see Figure S6, Supporting Information)
that is very similar to the spectrum of the long-lived transient
observed upon LFP of 1b (Figure S4, Supporting Information).
For that reason, it appears very likely that this species can be
assigned to a fulvene diyl-type diradical. Our calculations
indicate that this should be Z-11b′, as the E-conformer should
be too short-lived for detection. Again, the spin multiplicity
cannot be derived by comparing calculated vs. experimental
spectra, as the spectra of the singlet and triplet states of Z-11B′
are predicted to be very similar (Figure S10, see the Supporting
Information).
∆H(S-T) (kcal mol-1
B3LYP (BLYP)
)
angle around “carbene”
carbon (deg) (BLYP)
R, compd 13
H
F
Cl
Br
CH3
SiH3
Li
singlet not converged (+4.0)
-6.7 (-8.0)
S: 139.1 T: 138.5
S: 107.8 T: 128.7
S: 111.9 T: 134.0
S: 112.5 T: 134.4
S: 124.2 T: 141.6
S: 145.4 T: 165.9
S: 178.7 T: 178.9
S: 166.3 T: 165.2
-1.3 (-2.5)
-0.8 (-2.4)
singlet not converged (+7.7)
+17.8 (+9.6)
+32.4 (+18.2)
+3.5 (+1.2)
Z-11B
SCHEME 12. Calculated Activation and Reaction Enthalpy
for the Reaction of Model Diradical Z-11ꢀ with Oxygen
While we have been able to identify some of the transient
species with confidence, a number of open questions remain at
this point.
Our data suggest that triplet-11B (and thus triplet-11b) may
be compared to triplet carbenes that bear very space-demanding
substituents. Less than diffusion-controlled reactivity toward
molecular oxygen has frequently been observed for such species.
Didurylcarbene, which is significantly less shielded than 11B,
reacts with oxygen only at a fraction of the diffusion controlled
rate limit (kO2 ) 4.1 × 107 M-1 s-1).29b Triplet di-9-
Assignment of the Short-Lived Transient of 1b′ after
Excitation at 266 nm. Experimental evidence suggests an
analogy between the photochemical behavior of 1b and the
photoreactivity of 1b′ upon 266 nm excitation. However, the
agreement between the experimental transient UV/vis-spectrum
(see Figure S7 (Supporting Information) and the experimental
spectrum shown in Figure 6) and the calculated spectrum of
any of the stereoisomers of triplet 1B′ (cf. Figure S10, Sup-
porting Information) is poor.
anthrylcarbene has been reported to react with oxygen with a
30
rate constant kO2 ≈ 5 × 105 M-1 s-1
.
Hence, the absence of
reactivity toward oxygen does not rule out a triplet ground-
state for 11b.
In order to address this point of concern, we need to discuss
some alternative mechanistic pathways that are known to be of
importance in photochemical transformations of allenes and
triarylamines. Di- and triphenylamine have been reported to
undergo a photochemical cyclization to carbazole derivatives
Via dihydrocarbazole diradicals.32 The triplet energy of
triphenylamine26,33 has been reported as 69.6 kcal mol-1, which
is far higher than the triplet energies of the precursors
investigated in this work. As a consequence, the spin density
in the triphenylamine moiety is predicted to be small for the
first triplet states of 1B and 1B′, and any carbazole-forming
reaction starting from the lowest excited triplet state of 1b′
would be expected to have a very low quantum yield for that
reason. Our product studies gave no evidence for carbazole-
derived products. Also, the characteristic and intense absorption
of the rather long-lived singlet dihydrocarbazole diradical at λ
) 610 - 62032a nm was not observed in our experiments. On
the other hand, in our experiments employing 266 nm excitation,
higher excited precursor states may play a role, which might
Calculating the barrier for the reaction of a triplet diradical
like Z-11B with triplet molecular oxygen is difficult. In order
to reduce the computational cost we have studied this reaction
for a simplified model compound, triplet Z-11ꢀ (Scheme 12),
which is oxidized to the peroxy-diradical 14ꢀ. Including a spin
projection correction for the singlet transition state, the activation
31
enthalpy obtained by UB3LYP/6-31G(d) is 5.4 kcal mol-1
.
We note that the singlet peroxydiradical 14ꢀ that would initially
be formed in this reaction is predicted to have a closed-shell
ground state (S2 ) 0).
The reason for a barrier of 5.4 kcal mol-1 likely lies in the
steric shielding of the reactive center, which is positioned in a
deep pocket provided by the triisopropylsilyl group and the
aniline moiety. Given the rather low concentration of oxygen
in oxygen-saturated cyclohexane ([O2]sat ) 11.5 mmol),26 even
this very modest activation barrier would be sufficient to prevent
reaction of Z-11ꢀ, and thus likely also Z-11b, with oxygen.
Time-dependent DFT was also employed to aid in the spectral
assignment of the long-lived transients in the photochemistry
of 1b. Figure 5 shows a comparison of the calculated spectra
of the singlet and triplet states of Z-11B with the experimental
difference spectrum of the slowly decaying transient observed
(32) (a) Suzuki, T.; Kajii, Y.; Shibuya, K.; Obi, K. Bull. Chem. Soc. Jpn.
1992, 65, 1084–1088. (b) Chattopadhyay, N.; Serpa, C.; Purkayastha, P.; Arnaut,
L. G.; Formosinho, S. J. Phys. Chem. Chem. Phys. 2001, 3, 70–73. (c) Fo¨rster,
E. W.; Grellmann, K. H. Chem. Phys. Lett. 1972, 14, 536–538. (d) Compared to
the experimental system, only the n-butyl group has been replaced by a methyl
group in 15. An alternative dihydrocarbazole triplet 15b, where the ring closure
reaction involves the benzene ring connected to the allene moiety, is calculated
to be higher in energy than 15.
(29) (a) Nemirowski, A.; Schreiner, P. R. J. Org. Chem. 2007, 72, 9533–40.
(b) Tomioka, H.; Okada, H.; Watanabe, T.; Banno, K.; Komatsu, K.; Hirai, K.
J. Am. Chem. Soc. 1997, 119, 1582–1593.
(30) Astles, D. J.; Girard, M.; Griller, D.; Kolt, R. J.; Wayner, D. D. M. J.
Org. Chem. 1988, 53, 6053–6057.
(31) Oxygen attack at other positions is less favorable.
(33) McGlynn, S. P.; Azumi, T.; Kinoshita, M. Molecular Spectroscopy of
the Triplet State; Prentice Hall: Englewood Cliffs, NJ, 1969.
J. Org. Chem. Vol. 73, No. 22, 2008 8825

Bucher et al.
open up channels not available from the lowest excited precursor
singlet states. A comparison of the experimental transient UV/
vis spectrum of the short-lived transient observed upon 266 nm
LFP of 1b′ with the calculated UV/vis spectrum of a model
dihydrocarbazole triplet excited state 1532c reveals a general
agreement in the spectral pattern (Figure 6). Hence a dihydro-
carbazole triplet, likely accessible only from an upper singlet
excited-state only populated upon 266 nm excitation, may indeed
play the role of internal triplet sensitizer. In agreement with
the rather weak transient intensities monitored upon 266 nm
LFP of 1b′, the calculated oscillator strengths of the transitions
of triplet 15 are also smaller than the oscillator strengths of
transitions of allene-type triplet II states of 1b′. Given this
agreement, a tentative assignment of the short-lived transient
observed upon 266 nm, excitation of 1b′ to a dihydrocarbazole
triplet state appears justified.
Long-Lived Transient Received after Excitation of 1b′
at 355 nm. The second open question relates to the nature of
the long-lived transient observed upon 355-nm excitation of 1b′,
which has not yet been identified. Given the fact that two minor
unidentified products are formed if a preparative irradiation of
1b′ is performed at λexc ) 350 nm, it appears plausible that the
transient observed in the LFP experiment and the unidentified
trace products are linked to each other.
Arylallenes, if irradiated in the absence of external quenchers,
have been reported to undergo hydrogen- or aryl-shift reactions
resulting in the formation of carbene-derived products.34 Thus,
irradiation of triphenylallene in degassed cyclohexane yields
diphenylindenes, whose formation can be rationalized by
intramolecular addition of a phenylvinyl carbene to one of the
benzene rings.34 This reaction occurs rather inefficiently in the
photochemistry of triphenylallene.34 The formation of two minor
unidentified products upon long-wavelength excitation of 1b′
therefore could be related to photochemical phenylvinylcarbene
formation. Triplet phenylvinylcarbenes thus remain a possible,
but as yet unproven assignment for the long-lived transient
observed upon 355 nm excitation of 1b′.
conformers on the triplet hypersurface into syn-conformers is
highly unlikely given the calculated barriers of 9 - 12 kcal
mol-1 (cf. Schemes 9 and 10). Such barriers cannot be
reconciled with the observed very short triplet lifetimes.
Effect of Benzannulation. What is the effect of benzannu-
lation on the photochemical C2-C6 cyclization of enyne-allenes?
Our data indicate that the impact of benzannulation is quite small
as far as the mechanism, thermodynamics and kinetics of the
cyclization reaction itself are concerned. Equally, an evaluation
of the yields (see Table 2) obtained in the preparative photolyses
does not show vast differences, although the yields are
consistently a little bit higher for all benzannulated representa-
tives. The most notable difference is thus the wavelength
dependence seen in the LFP study, suggesting that the observed
discrepancies arise from different competing excited states. For
simple enyne-allenes, for example, twisted-alkene type triplet
states (triplet I) are almost degenerate in energy with the lowest
allene-type triplet (triplet II) states, while for benzannulated
enyne-allenes only the triplet II states matter. Moreover, the
degree of Franck-Condon overlap between singlet and triplet
excited states may be a factor that is crucial in determining ISC
efficiency and also the type of triplet state being formed.
Additional flexibility in accommodating triplet energy, such as
it is available to 1b (but not 1b′) in form of the triplet I state,
will certainly help in this respect. The formation of a dihydro-
carbazole-type triplet state of rather high energy upon 266 nm;
LFP of 1b′ might find its explanation along these lines.
However, since no photophysical study was conducted, an in-
depth discussion of the role of the triplet I and possibly
competing singlet excited states certainly is premature at the
present stage.
Conclusions
The results obtained from product, LFP, and DFT studies on
the photochemical C2-C6 cyclization of benzannulated and
nonbenzannulated enyne-allenes now provide a conclusive
picture. Accordingly, the photochemical reaction sequence
involves the excitation of the syn-enyne-allene to a highly
reactive syn-enyne-allene triplet II species. Due to its rapid
follow-up cyclization (∆Hq ) 4-5 kcal mol-1) to a triplet
fulvene diyl, the syn-enyne-allene triplet is not visible in the
LFP experiment. Rather, the unproductive anti-enyne-allene
triplet, not likely to undergo a energetically demanding bond
rotation during its short lifetime, was assigned to the short-
lived transient. The long-lived transient was assigned to the
triplet fulvene diyl based on the good agreement of experimental
data (spectra, product ratios) and DFT results. This intermediate,
generated in two conformations (E and Z), will afford the
observed products after a triplet to singlet interconversion. The
thus derived reaction mechanism, exhaustively evaluated by
DFT, is in full agreement with experimental data, such as
product ratio and its temperature dependence.
Formation of the syn-Enyne-Allene Triplets. According
to our product studies and LFP experiments, C2-C6 cyclization
does occur upon excitation of allenes 1b and 1b′. This reaction
requires the formation of syn-conformers of allene-type triplet
states of the two precursors. Observation of Stern-Volmer
quenching of diradical Z-11b upon addition of oxygen indicates
that at least a significant portion of Z-11b must be formed Via
an observable triplet state of 1b with a lifetime τ ∼ 58 ns, which
is in qualitative agreement with the value measured directly for
the short-lived transient in this system. As a consequence, we
would have to postulate that at least a significant portion of the
allene triplet state formed upon excitation of 1b must be in a
syn-conformation. The latter, i.e., triplet syn-31b, is formed either
directly from ground state syn-1b or via facile bond rotation
from triplet anti-31b. In contrast to the DFT results, our RIMP2
calculations indeed predict that the syn-conformers of ground
state 1B and 1B′ should be favored by 0.9 and 2.9 kcal mol-1
,
respectively, relative to the anti-conformers. Equilibrium ratios
can thus be estimated as anti-/syn-1b ) 22:78 and as anti-1b′/
syn-1b′ ) 0.7:99.3. If the conclusion from the RIMP2 calcula-
tions holds that the syn conformers should be preferred, the
experimental observations are therefore easily rationalized.
On the other hand, an alternative mechanistic scenario
involving a bond rotation of the primarily formed anti-
A DFT investigation on a series of fulvene diyls bearing a
variety of substituents indicates substituent effects on the
singlet-triplet energy gap that are similar to the effects generally
observed in the case of carbenes. The calculations indicate a
slight preference of triplet E-11B and Z-11B over the singlet
state of these diradicals, whereas the singlet and triplet states
are predicted to be almost degenerate for the benzannulated
derivatives E-11B′ and Z-11B′.
(34) Klett, M. W.; Johnson, R. P. J. Am. Chem. Soc. 1985, 107, 3963–3971.
8826 J. Org. Chem. Vol. 73, No. 22, 2008

Photochemical C2-C6 Cyclization of Enyne-Allenes
3
(m, 3H), 7.77 (d, J ) 7.6 Hz, 1H), 8.22 (dd, 3,4J ) 8.2 and 1.0
Experimental Section
Hz, 1H), 8.28 (dd, 3,4J ) 8.2 and 1.0 Hz, 1H); 13C NMR (100 MHz,
CDCl3) δ 11.3, 13.9, 18.7, 22.5, 30.1, 34.6, 93.5, 95.7, 105.0, 107.7,
121.4, 122.2, 125.9, 126.0, 126.3, 126.5, 126.8, 127.2, 127.7, 128.5,
129.5, 132.3, 132.6, 132.9, 136.1, 136.6, 205.3 ppm; MS-EI (70
eV) m/z 556.2 (0.2, M+); HRMS calcd for C34H41BrSi 556.216,
found 556.216.
The photochemical reactions of enyne-allene 1a-1c and 1a′-c′
were performed under nitrogen atmosphere, using a Rayonet RPR-
100 Photochemical reactor (16 lamps) at 300 nm wavelength in
dry degassed toluene/n-hexane (at 17 ( 2 °C).
Laser Flash Photolysis. The system employed has already been
described.35 The precursor allenes were excited using the third
harmonic (λ ) 355 nm) or fourth harmonic (λ ) 266 nm) of a
Nd:YAG laser (Quanta-Ray Laboratory 130, 120 mJ/pulse (third
harmonic), 50 mJ/pulse (fourth harmonic), 7 ns pulse duration).
The precursor concentrations were adjusted to an optical density
OD ) 0.3 at the laser wavelength. A flow cell was used in order
to avoid buildup of photoproducts. Prior to experiments, the
solutions were purged with argon for ca. 20 min. The cyclohexane
used in the experiments was of spectroscopic grade.
Bis(4-bromophenyl)-[4-(1-{2-[(triisopropylsilanyl)ethynyl]phen-
ylvinylidene}pentyl]phenyl]amine (1b′). In 30 mL of dry diethyl
ether tris-(p-bromophenyl)amine (1.41 g, 2.92 mmol) was cooled
to 0 °C in an ice bath. n-BuLi (2.5 M) (1.17 mL, 2.92 mmol) was
added dropwise and stirred. After 4 h, this reaction mixture was
added dropwise to a 1 M ZnCl2 solution (398 mg in 2.92 mL of
diethyl ether) and stirred for 60 min at room temperature. The
reaction mixture was then cooled to -60 °C, and Pd(PPh3)4 (84.0
mg, 72.7 µmol) in dry THF (5 mL) was added dropwise. After the
mixture was stirred for 30 min at the same temperature, propargyl
acetate 5′ (300 mg, 0.73 mmol) in dry THF (10 mL) was added
dropwise. After being stirred for 16 h at room temperature, the
reaction mixture was quenched with aqueous saturated ammonium
chloride solution. The aqueous layer was washed with pentane (2
× 50 mL). The combined organic layers were dried over sodium
sulfate and evaporated under reduced pressure. After purification
by column chromatography (silica gel, n-pentane, Rf ) 0.58)
compound 1b′ was isolated as light yellow oil: yield 265 mg, 48%;
IR (NaCl) ν 3053 (s), 2987 (m), 2865 (w), 2685 (w), 2151 (w),
1923 (w), 1581 (w), 1486 (m), 1421 (m), 1313 (w), 1265 (s), 1072
(w), 1007 (w), 896 (s), 824 (w), 739 (s) cm-1; 1H NMR (400 MHz,
acetone-d6) δ 0.91 (t, 3J ) 7.3 Hz, 3H), 1.17 (brs, 21H), 1.40-1.50
Computational. The program package Gaussian 0336 was used
for the DFT calculations. All calculated minima and transition states
were characterized as such by performing a vibrational analysis.
All calculated reaction and activation enthalpies include a zero point
vibrational energy correction. The energies of singlet diradicals and
transition states of singlet diradical reactions were corrected by
performing a spin projection correction.37 We used the pure density
functional BLYP38 and the hybrid B3LYP38 method in combination
with 6-31G(d) or 6-31+G(d)39 and cc-pVTZ40 basis sets. UV/vis
spectra were calculated using time-dependent B3LYP/6-31+G(d)41
based on B3LYP/6-31G(d) geometries. RIMP2 single-point energy
calculations42 with a TZVP43 basis set were performed using
Turbomole software.44 The UV/vis spectra shown were generated
from the calculated transition intensities and positions by overlaying
a Gaussian function over each transition (program kindly provided
by Koop Lammertsma).
3
(m, 2H), 1.55-1.64 (m, 2H), 2.52-2.67 (m, 2H), 6.98 (d, J )
9.1 Hz, 4H), 7.05 (d, 3J ) 8.8 Hz, 2H), 7.20 (t, 5J ) 2.9 Hz, 1H),
7.22 (td, 3,4J ) 7.5 and 1.3 Hz, 1H), 7.33 (td, 3,4J ) 7.8 and 1.3
{2-[3-(4-Bromonaphthalen-1-yl)hepta-1,2-dienyl]phenyleth-
ynyl}triisopropylsilane (1a′). To a solution of 1,4-dibromonaph-
thalene (1.53 g, 5.36 mmol) in dry diethyl ether (60 mL), cooled
to 0 °C by an ice bath, was added n-BuLi (2.5 M, 2.14 mL, 5.36
mmol) dropwise. After being stirred for 4 h, the reaction mixture
was added dropwise to 1 M ZnCl2 solution (731 mg in 5.4 mL of
diethyl ether) and stirred for 30 min at room temperature. After
the reaction mixture was cooled to -60 °C, Pd(PPh3)4 (155 mg,
134 µmol) in dry THF (5 mL) was added dropwise, and after the
mixture was stirred for 30 min at the same temperature, propargyl
acetate 5′ (550 mg, 1.34 mmol) in dry THF (10 mL) was added
dropwise. After being stirred for 16 h at room temperature, the
reaction mixture was quenched with aqueous saturated ammonium
chloride solution. The aqueous layer was washed with pentane (2
× 100 mL). The combined organic layer was dried over sodium
sulfate and evaporated under reduced pressure. After purification
by column chromatography (silica gel, n-pentane, Rf ) 0.53)
compound 1a′ was isolated as colorless oil: yield 220 mg, 30%;
IR (film) ν 3051 (w), 2943 (s, C-H), 2865 (s), 2253 (s), 2151 (s),
1943 (m), 1583 (w), 1504 (m), 1484 (m), 1465 (s), 1376 (s), 1265
(s), 1091 (m), 996 (m), 909 (s), 832 (m), 734 (s), 651 (s), 546 (w)
3
3
Hz, 1H), 7.42 (d, J ) 8.8 Hz, 2H), 7.43 (d, J ) 9.1 Hz, 4H),
7.49 (dd, 3,4J ) 7.5 and 1.1 Hz, 1H), 7.51 (td, 3,4J ) 7.8 and 1.1
Hz, 1H); 13C NMR (100 MHz, acetone-d6): δ 11.9, 14.2, 19.0, 23.1,
30.3, 30.8, 96.1, 96.6, 106.1, 110.6, 116.0, 121.9, 125.4, 126.5,
126.9, 127.9, 128.2, 129.9, 131.8, 133.3, 133.8, 137.3, 146.8, 147.4,
208.1 ppm; MS-EI (70 eV) m/z 751.2 [M]+; HRMS calcd for
C42H47Br2NSi 751.184, found 751.184.
3-Ethyl-5-{2-[triisopropylsilylethynyl]phenyl}penta-3,4-dien-
2-one (1c′). A mixture of LDA (2.33 mmol) and HMPA (0.41 mL,
2.33 mmol) in THF (30 mL) was treated at -80 °C dropwise with
a solution of 7′ (550 mg, 1.94 mmol) in 15 mL of THF. After the
mixture was stirred for 40 min at the same temperature, a solution
of 2-bromo-3-pentanone (476 mg, 2.92 mmol) in THF (10 mL)
was added dropwise. After 30 min, the reaction mixture was allowed
to warm to room temperature and quenched with water (40 mL).
The aqueous layer was extracted with diethyl ether (2 × 50 mL).
The combined layers were dried over sodium sulfate and concen-
trated under reduced pressure. After purification by column
chromatography (silica gel, n-hexane, Rf ) 0.70 n-pentane/diethyl
ether 9:1) 1c′ was isolated as yellow oil: yield 230 mg, 32%; IR
(film) ν 2944 (m, C-H), 2866 (m), 2254 (m), 2152 (m), 1933 (m),
1677 (s), 1461 (m), 1360 (w), 1265 (s), 1238 (m), 997 (w), 908
1
3
cm-1; H NMR (400 MHz, CDCl3) δ 0.91 (t, J ) 7.3 Hz, 3H),
1.16 (s, 21H), 1.43 (sextet, 3J ) 7.3 Hz, 2H), 1.52-1.62 (m, 2H),
2.52-2.67 (m, 2H), 6.99 (t, 5J ) 3.0 Hz, 1H), 7.13 (td, 3,4J ) 7.6
(s), 850 (w), 734 (s), 651 (m) cm-1; H NMR (200 MHz, CDCl3)
1
δ 1.07 (t, 3J ) 7.3 Hz, 3H), 1.17 (s, 21H), 2.32 (s, 3H), 2.33-2.47
(m, 2H), 7.16-7.29 (m, 3H, Ar-H), 7.37 (td, 3,4J ) 7.5 and 1.5
Hz, 1H), 7.53 (dd, 3,4J ) 7.5 and 1.5 Hz, 1H); 13C NMR (50 MHz,
CDCl3) δ 11.3, 12.3, 18.7, 20.3, 27.4, 96.6, 97.4, 104.5, 114.8,
121.9, 126.2, 127.4, 128.8, 133.1, 134.2, 198.0, 215.4 ppm; MS-
EI (70 eV) m/z 366.2 [M]+·; HRMS calcd for C24H34OSi 366.238,
found 366.238.
and 1.2 Hz, 1H), 7.28 (td, 3,4J ) 7.6 and 1.2 Hz, 1H), 7.35 (d, J
3
) 7.6 Hz, 1H), 7.45 (dd, 3,4J ) 7.6 and 1.2 Hz, 1H), 7.53-7.63
(35) Bucher, G. Eur. J. Org. Chem. 2001, 2463–2475.
(36) Frisch, M. J. et al. Gaussian03; Gaussian, Inc.: Wallingford, CT, 2004.
(37) Goldstein, E.; Beno, B.; Houk, K. N. J. Am. Chem. Soc. 1996, 118,
6036–6043.
(38) Becke, A. D. J. Chem. Phys. 1993, 98, 5648–5652.
(39) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971, 54, 724–
728.
(40) Dunning, T. H., Jr. J. Chem. Phys. 1989, 90, 1007–1023.
(41) Bauernschmitt, R.; Ahlrichs, R. Chem. Phys. Lett. 1996, 256, 454–464.
(42) Weigend, F.; Ha¨ser, M. Theor. Chem. Acc. 1997, 97, 331–340.
(43) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7, 3297–
3305.
{(1E)-(2-[(E)-1-(4-Bromonaphthalen-1-yl)pent-1-enyl]-1H-in-
den-1-ylidenemethyl}triisopropylsilane (8a′). In 10 mL of dry
toluene, 36.0 mg (76.0 µmol) of enyne-allene 1a′ was refluxed
for 40 h. After removal of toluene under reduced pressure and
purification by chromatography (aluminum sheet, silica gel 60 F254
n-pentane, Rf ) 0.53), compound 8a′ was isolated as yellow oil:
yield:33 mg, 91%; IR (film) ν 2945 (s), 2866 (s), 2253 (m), 1581
,
(44) TURBOMOLE V5-9-1; University of Karlsruhe: Karlsruhe, 2007.
J. Org. Chem. Vol. 73, No. 22, 2008 8827

Bucher et al.
(w), 1504 (m), 1461 (m), 1379 (m), 1256 (w), 1199 (w), 1105 (w),
3H), 2.19 (s, 3H), 6.13 (d, 4J ) 1.0. Hz, 1H), 6.69 (s, 1H), 7.14 (q,
3J ) 7.0 Hz, 1H), 7.16-7.20 (m, 1H), 7.24-7.29 (m, 2H), 7.72
(dd,3,4J ) 7.4 and 0.7 Hz, 1H); 13CNMR (100 MHz, CDCl3) δ
12.9, 15.9, 19.1, 27.9, 120.8, 122.3, 125.1, 128.3, 131.2, 135.7,
137.6, 138.1, 139.6, 140.7, 143.8, 155.4, 198.5 ppm; MS-EI (70
eV) m/z 366.2 [M]+; HRMS calcd for C24H34OSi 366.238, found
366.238.
1
1015 (w), 908 (s), 832 (w), 734 (s), 651 (s), 577 (w) cm-1; H
3
3
NMR (400 MHz, CDCl3) δ 0.97 (t, J ) 7.3 Hz, 3H), 1.02 (d, J
) 7.4 Hz, 18H), 1.35 (septet, 3J ) 7.4 Hz, 3H), 1.54 (sextet, 3J )
7.3 Hz, 2H), 2.32 (q, 3J ) 7.4 Hz, 2H), 6.08 (t, 3J ) 7.4 Hz, 1H),
5
6.56 (d, J) 0.8 Hz, 1H), 6.67 (s, 1H), 7.12-7.17 (m, 2H), 7.21
(dd, 3,4J )7.2 and 0.9 Hz, 1H), 7.25 (d, 3J)7.7 Hz, 1H), 7.40-7.44
(m, 1H), 7.52-7.56 (m, 1H), 7.68 (d,3J ) 7.3 Hz, 1H), 7.71 (d, 3J
(5-Bromo-13-butyl-12H-indeno[1,2-b]phenanthren-7-yl) triiso-
propylsilane (9a′):1H NMR (400 MHz, CDCl3) δ 0.90 (t, 3J ) 7.2
Hz, 3H), 1.16 (s, 18H), 1.34-1.45 (m, 3H), 1.58-1.64 (m, 4H),
3
3
) 7.8 Hz, 1H), 8.25 (dd, J ) 8.3 and 0.6 Hz, 1H), 8.30 (d, J )
8.6 Hz, 1H); 13CNMR (100 MHz, CDCl3) δ 12.9, 13.9, 19.0, 23.0,
32.9, 120.6, 121.7, 122.1, 124.8, 126.4, 126.6, 126.8 (2C), 127.4,
128.1, 129.3, 129.4, 131.1, 132.2, 132.3, 132.8, 133.4, 136.0, 137.7,
142.2, 143.8, 155.0 ppm; MS-EI (70 eV) m/z 556.2 [M]+; HRMS
calcd for C34H41BrSi 556.216, found 556.216.
3
3
2.73 (t, J ) 7.9 Hz, 2H), 4.29 (s, 2H),7.34 (d, J ) 7.7 Hz, 1H),
3
3
7.65-7.73 (m, 2H), 7.79 (d, J ) 8.0 Hz, 1H), 7.90 (d, J ) 7.5
3
3
Hz, 1H), 7.96 (t, J ) 7.8 Hz, 2H), 8.35 (d, J ) 8.3 Hz, 1H);
MS-EI (70 eV) m/z 556.2 (M+); HRMS calcd for C34H41BrSi
556.216, found 556.216.
Bis(4-bromophenyl)-[4-((1E)-1-{3-[1-triisopropylsilanylmeth-
(E)-ylidene]-3H-inden-1-yl)pent-1-enyl)phenyl]amine (8b′). In 15
mL of dry toluene, 30.0 mg (39.0 µmol) of enyne-allene 1b′ was
refluxed for 5 h. After removal of toluene under reduced pressure
and purification by chromatography (preparative TLC, silica gel
60 F254, n-pentane, Rf ) 0.35) compound 8b′ was isolated as yellow
oil: yield 25 mg, 83%; IR (film) ν 2941 (s), 2863 (s), 1580 (m),
1505 (m), 1485 (s), 1312 (s), 1285 (m), 1072 (w), 1007 (w), 881
(w) cm-1; 1H NMR (400 MHz, C6D6) δ 0.82 (t, 3J ) 7.2 Hz, 3H),
Bis(4-bromophenyl)(10-butyl-11H-benzo[b]fluoren-7-yl)-
amine (10): IR (film) ν 3054 (m), 2933 (m), 1580 (m), 1486 (m),
1
1262 (s), 1072 (m), 896 (m), 823 (m), 740 (s), 705 (s) cm-1; H
NMR (400 MHz, CDCl3) δ 1.00 (t, 3J ) 7.2 Hz, 3H), 1.54 (sextet,
3J )7.3 Hz, 2H), 1.73 (quintet, 3J ) 7.7 Hz, 2H), 3.13 (t, 3J ) 7.9
3
Hz, 2H), 4.03 (s, 2H), 7.02 (d, J ) 8.9 Hz, 4H), 7.22 (dd, 3,4J )
3
9.1 and 2.4 Hz, 1H), 7.32-7.40 (m, 2H), 7.37 (d, J ) 8.9 Hz,
4H), 7.52 (d, 4J ) 2.4 Hz, 1H), 7.56 (d, 3J ) 6.8 Hz, 1H), 7.83 (d,
3
4J ) 6.9 Hz, 1H), 7.87 (s, 1H), 7.96 (d, J ) 9.0 Hz, 1H); 13C
3
3
1.04 (d, J ) 7.5 Hz, 18H), 1.31 (septet, J ) 7.5 Hz, 3H), 1.39
(sextet, 3J ) 7.3 Hz, 2H), 2.17 (q, 3J ) 7.3 Hz, 2H), 6.25 (t, 3J )
7.4 Hz, 1H), 6.44 (d, 5J ) 0.8 Hz, 1H), 6.62 (d, 3J ) 8.8 Hz, 4H),
6.65 (s, 1 H), 6.81 (d, 3J ) 8.8 Hz, 2H), 7.08-7.14 (m, 7H), 7.28
(d,3J ) 8.6 Hz, 2H), 7.87 (dd, 3,4J ) 6.8 and 1.2 Hz, 1H); 13CNMR
(100 MHz, C6D6) δ 13.2, 13.9, 19.3, 23.3, 32.6, 115.7, 121.2, 122.5,
124.7, 125.2, 125.6, 127.9, 128.8, 131.3, 131.9, 132.2, 132.6, 135.8,
136.3, 138.5, 143.5, 144.7, 145.9, 146.7, 156.6 ppm; MS-EI (70
eV) m/z 751.2 (M+); HRMS calcd for C42H47Br2NSi 751.184, found
751.184.
NMR (50 MHz, CDCl3) δ 14.1, 23.3, 29.7, 32.4, 35.7, 115.5, 115.6,
120.6, 122.4, 123.4, 125.2, 125.3, 125.5, 126.9, 127.6, 128.3, 132.4,
134.5, 134.8, 138.6, 140.6, 141.4, 143.6, 143.7, 146.6 ppm; MS-
EI (70 eV) m/z 595.05 (M+); HRMS calcd for C33H27Br2N 595.051,
found 595.051.
Acknowledgment. We are very much indebted to the Deut-
sche Forschungsgemeinschaft and the Fonds der Chemischen
Industrie for continued support of our research.
(3Z)-3-[(1E)-1-(Triisopropylsilylmethylene)-1H-inden-2-yl]pent-
3-en-2-one (8c′). Enyne-allene 1c′ (46.0 mg, 0.12 mmol) was
refluxed for 48 h in 15 mL of dry toluene. After removal of toluene
under reduced pressure and purification by chromatography (pre-
parative TLC: silica gel 60F254, n-hexane/diethyl ether, 100:10,
n-hexane/diethyl ether 9:1, Rf ) 0.45) 8c′ was isolated as yellow
oil: yield 37 mg, 80%; IR (film) ν 2943 (s), 2866 (s), 1690 (s),
1625 (m), 1462 (m), 1356 (m), 1265 (s), 881 (m), 740 (s), 704
Supporting Information Available: Experimental proce-
dures for the preparation of and characterizations of 3, 3′, 4, 4′,
1
5, 5′, 6, 6′, 7, and 7′; H and 13C spectra of 1a′-c′, 8a′-c′, 9,
and 10′ and LFP results; UV-vis spectra of 1a′-c′. Cartesian
coordinates and energies of the stationary points optimized. This
material is available free of charge via the Internet at
http://pubs.acs.org.
3
(m), 659 (s) cm-1;1H NMR (400 MHz, CDCl3): δ 1.10 (d, J )
7.5 Hz, 18H), 1.42 (septet, 3J ) 7.5 Hz, 3H), 1.72 (d, 3J ) 7.0 Hz,
JO801689W
8828 J. Org. Chem. Vol. 73, No. 22, 2008