Cyclizations of silyl enol ether radical cations - The cause of the stereoselectivity

Article abstract of DOI:10.1002/ejoc.200400128

We have used photoinduced electron transfer (PET) activation of silyl enol ethers for the synthesis of tricyclic hydrocarbons. The mechanism of this reaction was investigated by conducting independent radical-induced cyclizations of corresponding iodo ket

Full text of DOI:10.1002/ejoc.200400128

FULL PAPER
Cyclizations of Silyl Enol Ether Radical Cations ؊ The Cause of the
Stereoselectivity
Jens O. Bunte,[a] Eike K. Heilmann,[a] Birka Hein,[a] and Jochen Mattay*[a]
Keywords: Density functional calculation / Electron transfer / Radical cation / Silyl enol ether / Stereoselectivity
We have used photoinduced electron transfer (PET) activa- aim was to explain the nature of the reactive intermediate of
tion of silyl enol ethers for the synthesis of tricyclic hydrocar- the cyclization step and to find the causes of the various types
bons. The mechanism of this reaction was investigated by
conducting independent radical-induced cyclizations of cor-
responding iodo ketones and performing density functional
theory (DFT) calculations on the possible intermediates. Our
of selectivity observed in this process.
( Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2004)
Introduction
at the 2-position results from radical-type reactions, i.e.,
radical coupling or radical addition to double bonds.
Because of their electron-rich double bonds, silyl enol
ethers have gained widespread use as nucleophiles in or-
ganic synthesis.[1Ϫ3] In addition, the electron-rich character
of these masked enol ethers or enolates can be used for
oxidation reactions. A well-known example is the Rubot-
tom oxidation with peracids or related oxygen-containing
oxidants, which gives α-hydroxy ketones or silyl enol ether
epoxides as products.[4Ϫ13] One-electron oxidation can also
be performed with silyl enol ethers: it results in the forma-
tion of radical cations.[14] These highly reactive species can
be generated by several means. Electrochemical oxidation
has been applied for synthetic purposes[15Ϫ17] and for mech-
anistic investigations.[18] Reversible anodic oxidation during
fast-scan cyclovoltammetric experiments provided strong
evidence for the formation of radical cations.[19] One-elec-
tron oxidation with chemical oxidants has been performed
using cerium ammonium nitrate (CAN),[19,20] tetranitro-
methane,[21] and xenon difluoride,[22] respectively. Further-
more, photoinduced electron transfer (PET) has been ap-
plied successfully;[23Ϫ29] all transient species in the electron
transfer process between chloranil and several silyl enol
ethers have been investigated thoroughly by time resolved
spectroscopy.[24,25]
Scheme 1
The cleavage of the SiϪO bond leads to the formation of
α-carbonyl radicals 2 (path A), which also undergo radical
reactions. The products of radical follow-up reactions of the
primary radical cation 1 (path B), distonic radical cations
or cations 3, will also suffer loss of the silyl cation by nucle-
ophilic substitution. An obvious and fundamental question
is how these basic reaction steps are coupled in a reaction
mechanism.
Snider et al. compared their results on the CAN-me-
diated oxidative cyclization of silyl enol ether 4 with Cur-
ran’s results from atom transfer cyclization of 5.[20,30] They
assigned the radical cation 6 to be the reactive intermediate
for the cyclization reaction, because the cyclization mode
was completely 6-endo selective (8), in contrast to the results
from the atom transfer reaction, which proceeded via the
α-carbonyl radical 7 (Scheme 2).
The primary product of the oxidation process, radical
cation 1, undergoes competing follow-up reactions
(Scheme 1). The presence of nucleophiles (Nu) causes meso-
lytic cleavage of the SiϪO bond, whereas bond formation
[a]
Organische Chemie I, Fakultät für Chemie, Universität
Bielefeld,
Postfach 100131, 33501 Bielefeld, Germany
Fax: (internat.) ϩ49-521-106-6417
E-mail: mattay@uni-bielefeld.de
Supporting information for this article is available on the
WWW under http://www.eurjoc.org or from the author.
Eur. J. Org. Chem. 2004, 3535Ϫ3550
DOI: 10.1002/ejoc.200400128
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3535

J. O. Bunte, E. K. Heilmann, B. Hein, J. Mattay
FULL PAPER
Scheme 2
The influence of the reaction medium and, especially, its
nucleophilic character is decisive for the mesolytic SiϪO
cleavage, as has been pointed out by Schmittel[19] and Ko-
chi.[24,25] For example, addition of nucleophilic substances,
such as alcohols, to a solution of silyl enol ethers in aceto-
nitrile leads to an increase in the rate constant of SiϪO
cleavage. On the other hand, changing the solvent from po-
lar, nucleophilic acetonitrile to less-polar, non-nucleophilic
dichloromethane results in a slower SiϪO cleavage and, ad-
ditionally, in a different ion-pair dynamic, which leads to a
different product spectrum.[19,24,25]
Results and Discussion
PET Oxidative and Radical Cyclizations
As reported earlier, we have synthesized the silyl enol
ethers 11 and 12 from the corresponding cycloalkenones 17
and 18 by reaction with dimethyl cuprate and trimethylsilyl
chloride (Scheme 4) and subjected them to PET oxidative
cyclization with 1,4-dicyanonaphthalene (DCN).[32]
In the course of our own research on the PET oxidations
of silyl enol ethers, we have found high selectivities occur
during the formation of polycyclic hydrocarbons, like 15
and 16, from silyl enol ethers. In addition to complete
6-endo selectivity in the cyclization step, high stereoselectivity
has also been observed to occur in the formation of three
new stereogenic centers (Scheme 3).[31]
Scheme 4
Cyclization of 11 resulted in the formation of three dia-
stereoisomers, 15a, 15b, and 15c, respectively; their product
ratio was 41:31:28 (Scheme 5; as throughout the text, only
one enantiomer is shown for the sake of clarity). Obviously,
only 6-endo-cyclized products were formed from the pri-
mary radical cation of 11. Additionally, we observed a com-
plete cis-selectivity with respect to the methyl group and the
α-carbonyl hydrogen atom. Interestingly, the relative stereo-
chemistry at the remaining two stereogenic centers varies
with low selectivity, in contrast to the cyclization of 12, in
which only two diastereoisomers were observed with a
90:10 selectivity.
Scheme 3
In this paper, we present the results of our studies regard-
Focusing on silyl enol ethers derived from cycloalkenones
ing the nature of the reactive intermediates involved as well bearing a cyclohexenyl side chain, we were able to investi-
as the causes of the selectivities observed. gate three different aspects of the selectivity in the cycliza-
3536
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 3535Ϫ3550

Cyclizations of Silyl Enol Ether Radical Cations
FULL PAPER
the sensitizer, whereas the proton is transferred from re-
sidual water and the solvent (acetonitrile).
With the differences in selectivity of the cyclization of
silyl enol ethers 11 and 12 measured, we were faced with
the questions of how the ring size of the silyl enol ether
influences the product distribution and, in a broader con-
text, how the reaction proceeds mechanistically.
The complete transformation (11 Ǟ 15 or 12 Ǟ 16)
comprises four different steps: one-electron oxidation, ring
closure, mesolytic SiϪO cleavage, and saturation of the
radical position. The one-electron oxidation can be ident-
ified as the initiating step and the radical saturation as the
terminating step, but the sequence of ring closure and SiϪO
cleavage has not yet been determined (Scheme 6).
Generally speaking, the sequence of cyclization and
SiϪO cleavage determines the intermediate in the cycliza-
tion step and it is, therefore, essential for the understanding
of the processes involved in product formation. Because
6-endo selectivity within the cyclization step was considered
to be the result of radical cations being the reactive inter-
mediates, we tried to prove this hypothesis by the selective
generation of α-carbonyl radicals 32 and 33 and investiga-
ting the product distribution. α-Carbonyl radicals were gen-
erated by iodine abstraction from α-iodo ketones utilizing
the tin hydride method.[33,34] Consequently, iodo ketones 34
and 35 were prepared from silyl enol ethers 11 and 12 by
reaction with N-iodosuccinimide (NIS);[35,36] 34 and 35
were obtained in 77 and 83% yields, respectively, as light-,
air-, and temperature-sensitive oils (Scheme 6).
Both iodo ketones are 1:1 mixtures of diastereoisomers;
they differ with respect to the positioning of their methyl
group and iodine substituent. The structures were assigned
using 1H and 13C NMR spectroscopy and mass spec-
trometry. Because of their instability, mass spectra of the
iodo ketones could only be obtained by ESI/MS using their
hydrazone derivatives.
The reaction of 34 with tributyltin hydride afforded a
mixture of 15a and 15b in 42% yield after chromatographic
purification (Scheme 7). The diastereoisomeric ratio was
58:42, as determined by GC analysis of the crude reaction
mixture. Compound 35 was cyclized by the same means,
affording 16a and 16b as a 90:10 mixture of diastereoiso-
Scheme 5
tion reaction. As mentioned before, the ring-size selectivity mers in 33% yield.
in the formation of distonic radical cations 22 and 23 was
Comparison with the results from the PET oxidative
found to be 6-endo. This feature of the PET oxidative cycli- cyclization of these cyclohexane-derived species (Scheme 5)
zation of silyl enol ethers was observed earlier, where it was revealed several interesting aspects. Most remarkably, the
attributed to the radical cations being the reactive inter- same products were formed by the radical (from 35) and
mediates in the cyclization step (Scheme 6).[27,29] Addition- PET oxidative cyclization (from 12). Furthermore, the
ally, the relative stereochemistry of three of the four stereo- product ratios are identical. The follow-up reactions of
genic centers is determined in the cyclization step (22, 23). α-carbonyl radical 33 resulted only in the formation of
The third selectivity in the transformation is the saturation 6-endo-cyclized products. The complete lack of 5-exo prod-
reaction of the remaining radical center. This formal hydro- ucts was unexpected from this radical-type species. The
gen atom addition generates the fourth stereogenic center. stereoselectivity in the cyclization step was also fully cis-
The nature of this process is not fully understood yet. Re- transoid (radical 27) as it was in the PET reaction. The satu-
cent studies from our group, utilizing deuteration experi- ration process of radical 27 resulted in the same high dia-
ments, indicate that tertiary radical centers, like 22 and 23, stereoselectivity as that observed for the PET process, al-
are saturated via an electron transfer/protonation se- though the hydrogen source was completely different. Tri-
quence.[32] The source of the electron is the radical anion of butyltin hydride is significantly more bulky than the ex-
Eur. J. Org. Chem. 2004, 3535Ϫ3550
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3537

J. O. Bunte, E. K. Heilmann, B. Hein, J. Mattay
FULL PAPER
Scheme 6
The Radical Reaction Pathway
Further insight into the details of the radical cyclization
of iodo ketones 34 and 35 was obtained through calcu-
lations on the reaction profile.[37] Energy calculations of the
radical intermediates of the cyclization reaction, as well as
transition state geometry calculations, were performed on
the density functional theory (DFT) level, which is known
to produce good results for this task.[38,39] From the variety
of potential combinations of calculation methods and basis
sets, we found that a combination of the B3LYP func-
tional[40,41] and the 6Ϫ31G* basis set[42] gave good results
for these complex structures with an acceptable demand in
CPU time.[43]
Scheme 7
Calculations on the cyclopentanone system 32 started
with energy calculations on the four possible cyclized rad-
icals (26a, 26b, 28a, and 28b) as well the initial radical 32.
pected sources of hydrogen or protons in the PET case, and Additionally, we calculated the transition state geometries
the saturation occurs through a radical pathway.
leading to the cyclization products (Figure 1).
The cyclization of cyclopentanone-based iodo ketone 34
All energy values are displayed in kcal/mol and are refer-
led to the formation of only two diastereoisomers, 15a and enced to the starting radical 32. In addition, the energy val-
15b. In contrast to the PET-mediated cyclization, the cis- ues for the transition states are compared to the lowest-
cisoid product 15c was not formed in the radical cyclization energy transition state leading to 26a. To ensure that the
reaction. The ratio of 15a to 15b (1.38) was similar to the calculated energies are indeed the lowest possible energies
ratio of 15a to 15b obtained in the PET reaction (1.32), for these highly flexible intermediates, we applied a multis-
which displays a low selectivity in the saturation step. The tep calculation strategy. Starting with the cyclized radicals
missing third isomer gave a first hint that different reactive 26a, 26b, 28a, and 28b, we performed conformational
intermediates exist in both cyclization procedures (radical analyses on a semiempirical level (AM1). For each of the
vs. PET).
obtained conformers, we calculated a reaction path con-
3538
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org Eur. J. Org. Chem. 2004, 3535Ϫ3550

Cyclizations of Silyl Enol Ether Radical Cations
FULL PAPER
Figure 1. Calculated energy profile for radical 32
cerning the bond of interest. From these profiles, the tran-
The central six-membered rings of the 6-endo transition
sition states were calculated on a semiempirical level states can be described as stretched cyclohexane conform-
(AM1). Finally, from this collection, we calculated the low- ers. We observe in Figure 3 that the transition state leading
est transition states at the B3LYP/6Ϫ31G* level, applying to cis-transoid radical 27a shows a chair-like conformer,
frequency analysis to ensure that the intermediate is a tran- whereas 27b should be formed via a twist-like conformer in
sition state. Energy calculations of the cyclized radicals were the corresponding transition state. This finding is in good
conducted in the same way. For energy comparison, we ap- agreement with the energy difference of 3.45 kcal/mol be-
plied zero point energy correction. The analogous energy tween these two structures. In another view, both carbo-
profile for the cyclohexanone system 33 is shown in Fig- cyclic rings of the starting radical 33 approach in a coplanar
ure 2.
manner to reach the low-energy transition state leading to
As we observe from Figures 1 and 2, cyclization of start- 27a, whereas they must approach in a stack-like arrange-
ing radicals 32 and 33 cannot be described easily by means ment for the formation of the cis-cisoid radical 27b.
of kinetic control. The 6-endo cyclization pathways are con-
One of the questions we had to face was the origin of
siderably exothermic, whereas the 5-exo cyclization leads to the cis-cisoid cyclization product 15c in the PET oxidative
radical intermediates that are either slightly exothermic cyclization (Scheme 5); this compound is not formed in the
(28a, 28b) or endothermic (29a, 29b) when compared to the radical reaction and, furthermore, it occurred only in the
starting radicals. The reason for these thermodynamics lies cyclopentanone system. In agreement with the calculated
in the highly stabilized character of the α-carbonyl radicals energy values, we did not observe this stereochemistry in
32 and 33.
the radical cyclization of iodo ketone 34. At a reaction tem-
The geometries of the transition states leading to the cy- perature of 80 °C in refluxing benzene, an energy difference
clized radical intermediates shown in Figures 1 and 2 can of 3.45 (26a, 26b) or 3.72 kcal/mol (27a, 27b) between the
be described using the BeckwithϪHouk model for the cycli- transition states would lead to an amount of Ͻ1% of the
zation of hexenyl radicals.[44Ϫ46] The lengths of the bonds product from the high-energy transition state. Thus, the rea-
˚
˚
formed are 2.14 A for the 6-endo cyclizations and 2.19 A son for stereoselective 6-endo ring closure in the radical
for the 5-exo cyclizations, which are in the typical range for cyclization path can be seen in the clearly disfavored twist-
comparable radical cyclizations.[34] The transition states for like transition state for the cis-cisoid approach in 26b and
the cyclizations to 27a and 27b are shown in Figure 3.[47]
27b.
Eur. J. Org. Chem. 2004, 3535Ϫ3550
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3539

J. O. Bunte, E. K. Heilmann, B. Hein, J. Mattay
FULL PAPER
Figure 2. Calculated energy profile for radical 33
related cyclization of hexenyl radical 36. The rate constant
for the 5-exo cyclization was determined to be 2.6 ϫ 106
s
Ϫ1, whereas the 6-endo cyclization proceeds with a rate
constant of 2.0 ϫ 106 sϪ1 at 313 K (Scheme 8).[49]
Figure 3. Transition states of the formation of 27a and 27b
Scheme 8
Considering the same kinetic-control approach for the
low-energy transition states for the 5-exo cyclization (28b:
1.82 kcal/mol; 29b: 1.87 kcal/mol), 7% of the product
Assuming the radical concentration is equal for both the
should have been isolated. The lack of these products can cyclization and saturation reaction and setting the concen-
be explained by assuming a reversible 5-exo cyclization, tration of tin hydride to the concentration of 50% conver-
which is feasible when considering the energetic situation. sion (2.5 mmol·LϪ1), we obtained values of vcyc/[rad] ϭ 2
Besides suitable thermodynamics, a necessary requirement ϫ 106 sϪ1 for the cyclization and vsat/[rad] ϭ 4 ϫ 103 sϪ1
for a reversible ring closure is a lifetime of the intermediates for the saturation reaction.
that allows both cyclization and ring opening prior to a
Using this simplified expression with one constant con-
follow-up reaction. In this case, the follow-up reaction is the centration for all radical species and neglecting temperature
hydrogen transfer from tributyltin hydride. For secondary dependencies of the values known in the literature, we find
radicals, a rate constant of 1.5 ϫ 106 L molϪ1 Ϫ1 has been a cyclization rate 500 times faster than the saturation pro-
s
determined at 298 K; for tertiary radicals, the rate constant cess. Further evidence for the slow saturation is the lack of
is 1.7 ϫ 106 L molϪ1 sϪ1 at 303 K.[48] As is noticeable from uncyclized material. α-Carbonyl radicals 32 and 33 did not
these values, no sufficient differentiation between secondary react with tributyltin hydride, but underwent complete
and tertiary radicals in the saturation reaction can be cyclization. The assumed ratio between the cyclization and
achieved.
saturation kinetics results in a reversible 5-exo cyclization
A reasonable value for the rate constant of the cyclization and, consequently, the observed 6-endo selectivity of the
reaction of radicals 32 and 33 can be estimated from the radical cyclization of iodo ketones 34 and 35.
3540
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 3535Ϫ3550

Cyclizations of Silyl Enol Ether Radical Cations
FULL PAPER
In the case of tributyltin hydride, the stereoselectivity of by a comparison of the rate constants of the involved pro-
the saturation of the intermediately formed radicals 26a and cesses and is further supported by the fact that only cyclized
27a is determined by the energy differences of the transition products were isolated. (c) The exclusive cis-transoid stereo-
states. The calculation of these large, highly flexible struc- chemistry found in the cyclization products can be rational-
tures, or of simplified trimethyltin hydride species, could not ized by kinetic control; the competing cis-cisoid stereo-
be accomplished sufficiently on the DFT level. The strong chemistry could be formed only via a high-energy transition
exothermic character of the saturation reaction allows state. (d) Finally, the stereoselectivity of the saturation pro-
Hammond’s postulate to be used and, therefore, it should cess can be explained well using Hammond’s postulate and
be possible to describe the observed selectivity in the satu- comparing the energies of the conformers of the radicals
ration process using the conformer energies of the cyclized involved. Differences in the observed selectivities between
radicals. For the cyclohexanone species 27a, two conform- cyclopentanone and cyclohexanone systems were observed
ers are of major interest (Figure 4). Conformer 27a1, only in the last step, the radical saturation. Because of the
formed directly via the transition state shown in Figure 3, rigid cyclopentanone ring, the cyclized radical could not
can react with tributyltin hydride to form product 16b reach the stable conformation that formed in the cyclohexa-
through cis-saturation. Because of the slow reaction with none system.
tributyltin hydride, 27a1 can equilibrate with 27a2, which is
It should be noted that we were able to demonstrate that
2.02 kcal/mol more stable. Reaction of 27a2 with tributyltin an observed 6-endo selectivity cannot be attributed simply
hydride leads to the formation of the trans-saturated prod- to a radical cationic reaction intermediate, as it has been
uct 16a. According to Hammond’s postulate, this energy described previously.[29] The results of the radical cycliza-
difference results in the diastereoselectivity of the saturation tions demonstrate that the question concerning the reaction
process. At a reaction temperature of 353 K, the energy dif- mechanism of the PET oxidative cyclization has yet to be
ference of 2.02 kcal/mol leads to 6% of the cis-saturated answered.
isomer 16b. Indeed, this yield is close to the observed value
of 10% for 16b in the reaction product of the radical cycliza-
tion of 35.
The Radical-Cationic Reaction Pathway
Prior to investigating the cyclization reactions of silyl
enol ether radical cations, we tried to characterize the reac-
tive site of the radical cation generated by one-electron oxi-
dation by utilizing DFT calculations. We were interested in
the distributions of charge and spin, where a comparison
with the corresponding α-carbonyl radical was possible. As
a model system for calculation, we chose 3,3-dimethylcyclo-
pentanone. Comparison of the calculated structures (using
B3LYP/6Ϫ31G*) of silyl enol ether 39 and its correspond-
ing radical cation 40 revealed several changes in the molecu-
lar structure induced through oxidation (Figure 5).
Figure 4. Structures of 27a1, 27a2, and 26a
Unlike 27a, cyclopentanone radical 26a does not have a
preferred conformer. As a consequence of the less-flexible
cyclopentanone ring bearing substituents with dihedral
angles of nearly 0°, the central six-membered ring is forced
into twist- or boat-like conformations. In these very similar
conformers, the lowest-energy form of which is shown in
Figure 4, the radical position is flattened. This observation
contrasts with the geometric features of the conformations
of 27a, where the radical center is pyramidalized towards
Figure 5. Geometrical changes of 39 induced by one-electron oxi-
dation
As expected, one-electron oxidation weakens the CϪC
the site where the reaction with tin hydride occurs. The geo- double bond because this position is the most electron-rich.
metrical features of the conformers of 26a, in combination This process is accompanied by a strengthened CϪO bond,
with the similar energies for the conformers, explain the ob- as can be estimated by their bond lengths. The second im-
served lack of selectivity in the saturation process.
portant structural change concerns the silyl moiety. The
In summary, we were able to explain sufficiently all the OϪSi bond is weakened and the three methyl groups pos-
selectivities observed in the cyclization process. (a) The sess a more-planar orientation. The length of the OϪSi
˚
6-endo selectivity can be realized by a reversible 5-exo cycli- bond (1.80 A) can be compared directly with structures cal-
zation that becomes possible as a result of the stabilized culated by Olah for adducts between the trimethylsilyl cat-
α-carbonyl radicals. (b) The reversibility can be explained ion and several carbonyl compounds; for these systems,
Eur. J. Org. Chem. 2004, 3535Ϫ3550
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3541

J. O. Bunte, E. K. Heilmann, B. Hein, J. Mattay
FULL PAPER
[50]
˚
lengths of 1.8 A for OϪSi bonds were also found.
Ad-
ditionally, the X-ray structure of the cation formed from
the triisopropylsilyl cation and acetonitrile can be used for
comparison.[51,52] Here, a value of 1.82 A has been reported
˚
for the length of the OϪSi bond.
All the geometrical features found in these calculations
are in agreement with the expected properties of the radical
cations: the weakened OϪSi bond leads to an easy SN2-like
substitution of the silyl cation induced by the solvent or
other nucleophiles. The bond orders of 1.5 for both the
CϪC and CϪO bonds can be attributed to a heteroallylic
system corresponding to an α-carbonyl radical. To deter-
mine the influence of the attached trimethylsilyl cation on
the spin distribution, we calculated the distribution for the
α-carbonyl radical 41 for the sake of comparison (Figure 6).
Scheme 9
We calculated the energy profile for the possible cycliza-
tion products of radical cation 44 and the corresponding
transition states by the methodology described above. As
seen from Figure 7, the transition states for the 5-exo
products 45 and 47 are too high in energy to contribute to
the product distribution. The 5-exo cyclization also pro-
ceeds endothermally. As a direct consequence of the cat-
ionic character of the intermediates, additional minima
were found on the energy surface.
Structures 48 and 49 can be described as electron-de-
ficient compounds carrying a bond between the radical cat-
ionic center and the double bond.[48] The bond length was
˚
calculated to be 2.8 A and, thus, clearly it is less than
double the van der Waals distance of a typical p-system.
The α-carbonyl position is located in a 5-exo orientation
towards the double bond. Because of the weak character of
this bond, it can slip easily along the π-system to achieve a
6-endo cyclization. This process can be seen clearly from the
energy of the transition state leading to 45 (Figure 7). Thus,
the observed 6-endo selectivity of the PET oxidative cycliza-
tion reaction of 42 can be explained reasonably by the en-
ergy profile calculated on the DFT level.
We continued with calculations of the energy profiles of
radical cations 13 and 14. Unfortunately, it was not possible
to calculate the transition state geometries for these cycliza-
tions because the highly flexible character of the intermedi-
ates prohibited successful geometry optimizations for the
transition states. The energies of the stationary points are
displayed in Scheme 10.
Even without the accurate energy values for the tran-
sition states, it becomes obvious from Scheme 10 that the
5-exo cyclization is disfavored. The formation of 5-exo-
cyclized products is endothermic [6.7 (24a) and 8.5 (25a)
kcal/mol] and will, therefore, not contribute to product for-
mation. To explain the stereoselectivity in the cyclization
reaction, we chose to calculate simplified radical cations to
learn more about the transition state energies. The replace-
ment of the trimethylsilyl cation by a proton gave the rad-
ical cations shown in Scheme 11.
Figure 6. Calculated distributions in 40 and 41
As expected for this heteroallylic system, the spin density
is located at both ends of the three-atom arrangement.
About 70% of the spin density is located at the α-position,
which is the reactive position for follow-up reactions.
In contrast, the corresponding radical cation 40 shows a
delocalized spin distribution. Only 52% of the spin is lo-
calized at the reactive α-position. The carbonyl group con-
tributes further, showing a strong disturbance of the former
heteroallylic system. The silyl group does not carry any
spin.
To achieve a comparison between partial charges in the
neutral silyl enol ether 39 and its radical cation 40, we cal-
culated natural bond orbital (NBO) charges for both.[53Ϫ56]
The observed differences are shown in Figure 6. The tri-
methylsilyl group contributes to only 17% to the positive
charge; this finding contrasts the view that it is a cationic
leaving group. Obviously, the ease of SiϪO cleavage must
be explained by the substitution reaction mentioned above.
The remaining charge is delocalized over the carbonyl
group and its α-position. This cationic character of the
reactive site must be considered for the estimation of the
reactivity of the radical cation.
We calculated the transition state geometries 52؊55 and
found an energy difference of 1.77 kcal/mol between 52 and
54.[48] This value is clearly less than that (3.45 kcal/mol) in
The effect of the electrophilic reactive site can be seen the corresponding radical cyclization (Figure 1) and, thus,
clearly from calculations we conducted for one of our ear- it provides a first hint towards the cause of the observed
lier cyclization reactions.[29] As demonstrated experimen- product distribution. In contrast, the corresponding cis-
tally, butenyl-substituted cyclohexanone silyl enol ether 42 cisoid transition state 55 could not be calculated. Since we
cyclizes in a complete 6-endo fashion under PET oxidative could not identify a transition state for the reaction 51 Ǟ
condition to yield cis-α-decalone 43 (Scheme 9).
59, we calculated reaction profiles for both 51 Ǟ 59 and
3542
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org Eur. J. Org. Chem. 2004, 3535Ϫ3550

Cyclizations of Silyl Enol Ether Radical Cations
FULL PAPER
Figure 7. Calculated energy profile for radical-cation 44
Scheme 10. Energetics of possible cyclization products of 13 and 14 (in kcal/mol)
the corresponding cyclohexanone system 50 Ǟ 58. The re-
sults of these DFT calculations are shown in Figure 8.[57]
˚
The bond of interest was opened in steps of 0.10 A. Cyclo-
hexanone radical cation 59 does not represent a minimum
on the energy surface, which is in sharp contrast to the
cyclopentanone radical cation 58 (Figure 8). Although this
result explains why transition state 55 could not be calcu-
lated, it should be considered that these calculations were
conducted using a simplified system. The corresponding tri-
methylsilyl radical cation 23b (Scheme 10) is indeed a mini-
mum, although no follow-up products could be isolated.
According to the results in Figure 8, the reason for the
differences in the product distribution between cyclopenta-
none and cyclohexanone silyl enol ether radical cations 13
Scheme 11
Eur. J. Org. Chem. 2004, 3535Ϫ3550
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3543

J. O. Bunte, E. K. Heilmann, B. Hein, J. Mattay
FULL PAPER
tivity of ring closure, because no uncyclized ketone 60 could
be detected.
Reactions of radical cations are generally considered as
very fast processes as a result of the flattened energy sur-
face.[58] It is rare to find values for the rate constants of
cyclization reactions that are similar to those of our system.
Recently, Horner et al. investigated the cyclization of enol
ether radical cations and determined the rate constants for
these processes.[59] Examples for a 5-exo and a 6-exo cycli-
zation of radical cations 61 and 62 are shown in Scheme 13.
Figure 8. Calculated energy profiles for the cyclizations 50 Ǟ 58
and 51 Ǟ 59
and 14 could be a result of the different lifetimes of the
corresponding cis-cisoid cyclized radical cations 22b and
23b (Scheme 12). Both cyclizations from starting radical
cations 13 and 14 are endothermic, but only in case of 22b
was follow-up product 15c isolated.
As for the radical cyclization process described earlier,
the kinetics of the follow-up reactions must be considered
to describe the reaction mechanism properly. These reac-
tions are the mesolytic cleavage of the SiϪO bond and the
saturation of the remaining radical position. The possible
reaction pathways for radical cation 50 are displayed in Scheme 13
Scheme 12.
The rate constants were determined to be k5exo ϭ 2 ϫ
109 sϪ1 and k6exo ϭ 2 ϫ 107 Ϫ1. Because of the highly
s
stabilized double benzylic position in the product radical
cations 63 and 64 and the accompanying enhanced reaction
rate, we estimated the rate constants in our cyclization sys-
tem to kcyc-cc ϭ kcyc-ct ϭ ca. 2 ϫ 107 sϪ1. These rates are
about 10 times faster than those for the corresponding rad-
ical process.
The primary follow-up reaction, the mesolytic SiϪO
cleavage proceeds with a pseudo first-order rate constant of
kdesil ϭ 2.3 ϫ 106 sϪ1 in acetonitrile.[60]
As a consequence, the rate constants involved in this re-
action increases in the order kcyc-cc ϭ kcyc-ct Ͼ kdesil ϾϾ
ksat, which is in good agreement with our results.
If the cyclization reaction of the radical cation 13 pro-
ceeded under thermodynamic control, no cis-cisoid product
15c would be formed because of the endothermic character
of the cyclization process. If, on the other hand, the rate
constant for the desilylation reaction were higher, the cis-
cisoid product 15c would not have been formed, because the
α-carbonyl radical 32 cyclizes in only a cis-transoid fashion.
According to our model, the cyclization of 13 takes place
Scheme 12
Starting from 13, two 6-endo cyclization reactions are
possible, leading to either 22a or 22b. Additionally, desilyl- to yield 22a and 22b intermediately; the back reaction then
ation to 32 can occur, which cannot explain the formation competes with desilylation. Desilylation may also take place
of 15c, as shown for the radical cyclization of iodo ketones. at the noncyclized stage of 13, because the ratio of rate con-
Because of the energetics of both cyclization reactions (22a, stants of only 10 is not high enough to prevent this process
22b), reversibility must be taken into account. The final from occurring. The resulting α-carbonyl radical 32 cyclizes
saturation process seems to have no influence on the selec- in the described manner. A difference from the product dis-
3544
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 3535Ϫ3550

Cyclizations of Silyl Enol Ether Radical Cations
FULL PAPER
Scheme 14
tribution cannot be observed; both pathways afford 15a
and 15b.
In addition, 2-propanol undergoes hydrogen abstraction,
which should open competing side-reactions. For the pho-
The lack of cis-cisoid products from the cyclohexanone tochemically induced hydrogen abstraction of the triplet
silyl enol ether radical cation 14 can be explained by a state of 4,4Ј-bipyridine from 2-propanol, a rate constant has
shorter lifetime of the corresponding cyclized radical cation been determined to be kbipy ϭ 2.4 ϫ 105 L molϪ1 sϪ1 [61]
;
23b and a faster ring opening reaction.
the pseudo-first-order rate constant for a 15% mixture is
The saturation reaction of intermediate radicals 26a, 26b, k1stbipy ϭ 4.8 ϫ 105 sϪ1. This rate is only ca. five times
and 27a afforded a selectivity similar to that observed in slower than the rate of desilylation and only ca. 50 times
the radical cyclization of iodo ketones 34 and 35. Although slower than the rate of cyclization (107 sϪ1). Compared to
the mechanism of the saturation process is not been under- the reaction rate for the saturation with tributyltin hydride
stood fully,[33] conformers of the cyclized radical will have (k ϭ 4 ϫ 103 sϪ1), the hydrogen abstraction from 2-propa-
a great influence of the diastereoselectivity. Therefore, an nol is ca. 100 times faster. The saturation products at every
approach similar to the description of the radical saturation stage of the reaction are shown in Scheme 14: 68 is the satu-
by tributyltin hydride seems to be an appropriate expla- ration product prior to cyclization, 67 is the product from
nation.
saturation before the cyclized radicals were able to reach
If the rate constant for the saturation process increases, equilibrium (Figure 4), and 66 is the product observed orig-
the product distribution will change significantly. This find- inally after a slow saturation reaction.
ing parallels experimental results obtained earlier in our
The effect of adding 2-propanol is obviously different
group.[28,29] PET oxidative cyclization of silyl enol ether 65 from the original interpretation. The rate of mesolytic cleav-
in acetonitrile afforded 66 and 67 in a 95:5 ratio age is not altered by the addition of the alcohol, but the
(Scheme 14).
rate of radical saturation is enhanced drastically, resulting
This ratio is in good agreement with our results from the in a broader product spectrum. These findings give ad-
cyclization of 12. Cyclization in a mixture of 15% 2-propa- ditional support for the mechanism of PET oxidative cycli-
nol in acetonitrile resulted in a broader product spectrum. zation of silyl enol ethers that we have presented.
Uncyclized ketone 68 is produced in a considerable yield; it
has not been observed at all in PET oxidative cyclization
reactions previously. Furthermore, the selectivity in the
saturation reaction vanishes, as is noticeable from the 40:36
ratio between 66 and 67.
Conclusion
Using an independent synthetic approach to 15 and 16,
Originally, these results were interpreted as an enhanced by radical cyclizations of iodo ketones 34 and 35, ac-
mesolytic cleavage of the SiϪO bond in the primary radical companied by a variety of DFT calculations, we were able
cation by the nucleophilic alcohol. The product distribution to conduct an in-depth study of the results of our PET-
was thought to be the result of a cyclization of the resulting mediated oxidative cyclizations of silyl enol ethers 11 and
α-carbonyl radical.[29]
The rate constant for the reaction between 2-propanol
12.
Radical cyclizations of iodo ketones 34 and 35 proceed
and benzylic trimethylsilyl radical cations had been deter- with remarkable selectivity. The complete 6-endo selectivity
mined as k2PrOH ϭ 9.7 ϫ 105 L molϪ1 sϪ1 [61]
.
A 15% mix- of the cyclization can be explained by the stability of the
ture of 2-propanol has a concentration of ca. 2 . This situ- intermediate α-carbonyl radical and a reversible 5-exo cycli-
ation results in a pseudo-first-order rate constant of zation. We note that a simple correlation between 6-endo
k1st2PrOH ϭ 1.94 ϫ 106 sϪ1 and is very similar to the rate selectivity and radical cationic intermediates does not apply
constant obtained in pure acetonitrile (kdesil ϭ 2.3 ϫ 106 in this case.
s
Ϫ1). Therefore, the addition of 2-propanol should not en-
The stereoselectivity of the ring closure and saturation
steps can be explained by our results from DFT calcu-
hance mesolytic SiϪO cleavage.
Eur. J. Org. Chem. 2004, 3535Ϫ3550
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3545

J. O. Bunte, E. K. Heilmann, B. Hein, J. Mattay
FULL PAPER
(m, 1 H, H-4b), 5.82 (m, 1 H, H-4a) ppm. 13C NMR (125.8 MHz,
CDCl3): δ ϭ 22.26 (t, C-6), 22.58 (t, C-2), 22.74 (t, C-7), 25.06 (t,
C-5), 28.07 (t, C-8), 29.55 (t, C-1), 35.23 (t, C-9), 36.20 (t, C-10),
37.21 (t, C-3), 121.78 (d, 4b), 125.60 (d, 4a), 136.11 (s, 8a), 166.42
(s, 10a), 199.81 (s, C-4) ppm. GC/MS (EI, 70 eV): m/z (%) ϭ
205 (0.4), 204 (2), 148 (13), 111 (10), 110 (100), 95 (36), 91 (14),
79 (23), 77 (15), 67 (50), 65 (14), 55 (42), 53 (29), 41 (38), 39 (23),
27 (16). GC/MS (CI, isobutane): m/z (%) ϭ 206 (17), 205 (100).
HRMS: calcd. for C14H20O [Mϩ] m/z ϭ 204.1514; found 204.1522;
deviation, 3.9 ppm. C14H20O (204.31): calcd. C 82.30, H 9.87;
lations, in combination with rate constants known in the
literature, that support our model of a slow saturation pro-
cess.
With these results, it was possible to describe also the
radical cationic cyclization reaction. We could confirm that
radical cations are indeed reactive intermediates in the cycli-
zation step, although the contribution of a radical pathway
could not be excluded.
The kinetics of the fundamental steps are of major im-
portance, as is seen in the analysis of the PET reaction in found C 81.89, H 9.73. IR (film): ν˜ ϭ 3326, 2927, 2660, 1668, 1625,
1428, 1372, 1346, 1324, 1252, 1191, 1125, 1079, 1049, 1008, 964,
an acetonitrile/2-propanol medium. Increasing the satu-
ration rate leads to a completely different product spectrum.
Combining the known rate constants with our own results
917, 885, 859, 837, 800, 755 cmϪ1
.
from DFT calculations, we were able to present a compre- 3-(1-Cyclohexenylethyl)-3-methyl-1-cyclohexenyl
Trimethylsilyl
Ether (12): In a flame-dried flask, copper() iodide (5.14 g,
27.0 mmol) was suspended in dry THF (100 mL) and the mixture
was cooled in an ice/NaCl slurry. Methyllithium (33.8 mL,
54.1 mmol of a 1.6  solution in Et2O) was added dropwise and
the mixture was stirred until the yellow precipitate dissolved com-
pletely. The reaction mixture was cooled to Ϫ78 °C and a solution
of trimethylsilyl chloride (2.93 g, 27.0 mmol) and 3-(1-cyclohexenyl-
ethyl)-2-cyclohexen-1-one (18) (2.77 g, 13.5 mmol) in dry THF
(50 mL) was added dropwise. The cooled mixture was stirred for
2 h before being poured onto ice-cold 0.1  aq. HCl (200 mL) and
ice-cold pentane (250 mL) in a separating funnel. After shaking for
a short time, the layers were separated and the organic layer was
washed with ice-cold sat. aq. NaHCO3, dried (Na2SO4), and fil-
tered. Concentration in vacuo afforded 12 (3.39 g, 86%) as a color-
less oil. NMR experiments: 1H, H/H-COSY, APT, HMQC. 1H
NMR (500 MHz, CDCl3): δ ϭ 0.175 [s, 9 H, Si(CH3)3], 0.95 (s,
3 H, CH3), 1.23Ϫ1.31 (m, 1 H, H-1), 1.28Ϫ1.39 (m, 2 H, H-10),
1.37Ϫ1.44 (m, 1 H, H-1), 1.50Ϫ1.57 (m, 2 H, H-6), 1.57Ϫ1.64 (m, 2
H, H-7), 1.61Ϫ1.68 (m, 2 H, H-2), 1.80Ϫ1.92 (m, 2 H, H-9),
1.89Ϫ1.94 (m, 2 H, H-8), 1.91Ϫ1.94 (m, 2 H, H-3), 1.95Ϫ2.00 (m,
2 H, H-5), 4.65 (dd, J ϭ 1.5, 1.5 Hz, 1 H, H-4a), 5.38 (m, 1 H, H-
4b) ppm. 13C NMR (125.8 MHz, CDCl3): δ ϭ 0.13 [q, Si(CH3)3],
19.62 (t, C-2), 22.57 (t, C-6), 23.06 (t, C-7), 25.24 (t, C-5), 28.02 (q,
Me), 28.52 (t, C-8), 29.86 (t, C-3), 32.76 (t, C-9), 34.43 (s, C-10a),
34.46 (t, C-1), 41.64 (t, C-10), 114.45 (d, C-4a), 120.08 (d, C-4b),
138.63 (s, C-8a), 149.22 (s, C-4) ppm. GC/MS (EI, 70 eV): m/z
(%) ϭ 293 (0.4), 292 (1), 184 (15), 183 (100), 73 (39). HRMS: calcd.
for C18H32OSi [Mϩ], m/z ϭ 292.2222; found 292.2225; deviation,
0.9 ppm. C18H32OSi (292.54): calcd. C 73.90, H 11.03; found C
74.63, H 11.30. IR (film): ν˜ ϭ 2935, 2841, 1662, 1453, 1365, 1341,
1263, 1251, 1207, 1193, 1164, 1134, 1102, 1058, 964, 936, 909, 889,
hensive model for the PET-mediated oxidative cyclization
of silyl enol ethers.
Experimental Section
1
General Remarks: H NMR and 13C NMR spectra were recorded
at 300 K using either a Bruker DRX 500 or a Bruker Avance 600
spectrometer. Spectra were recorded in CDCl3 or C6D6; chemical
shifts were calibrated to the residual proton and carbon resonance
of the solvent: CDCl3 (δH ϭ 7.26 ppm, δC ϭ 77.00 ppm), C6D6
(δH ϭ 7.20 ppm, δC ϭ 128.0 ppm). IR spectra were recorded on a
PerkinϪElmer 841 spectrometer. HRMS were recorded on a Auto-
spec X (Vacuum Generators, Manchester). GC/MS were recorded
on a Shimadzu GC 17A / QP 5050A equipped with a 5MS capillary
column (HewlettϪPackard). Analytical thin-layer chromatography
was performed on silica gel 60 F254 (Merck). Column chromatog-
raphy was performed on silica gel MN60 (63Ϫ200µm;
MachereyϪNagel). HPLC was performed on an RT 250Ϫ25 silica
gel column [LiChrosorb Si60 (7µm); Merck] using a Merck L6000
pump and an RI Bischoff 8110 detector (Bischoff). Photochemical
reactions were performed using an RPR-100 Rayonet photochem-
ical chamber reactor (Southern New England Ultraviolet Com-
˚
pany) with RPR 4190-A lamps that show an emission maximum
at 419 Ϯ 15 nm at half band width. All reactions were carried out
under an atmosphere of argon. Starting materials and solvents were
purified using standard laboratory techniques.[62]
3-(1-Cyclohexenylethyl)-2-cyclohexen-1-one (18): tert-Butyllithium
(10.0 mL, 17.0 mmol of a 1.7  solution in pentane) was added to
a solution of 2,2Ј-bipyridine (2 mg) in dry THF (80 mL) cooled
to Ϫ78 °C and then a solution of 1-iodo-2-(1-cyclohexenyl)ethane
19[32][63] (1.75 g, 7.42 mmol) in dry THF (10 mL) was added drop-
wise. The reaction mixture was stirred for 1.5 h at Ϫ78 °C before
841, 801, 753, 686 cmϪ1
.
PET Cyclization of 3-(1-Cyclohexenylethyl)-3-methyl-1-cyclohexenyl
a solution of 3-ethoxycyclohex-2-enone 21 (1.00 g, 7.13 mmol) in Trimethylsilyl Ether (12): A solution of 12 (130 mg, 0.44 mmol),
dry THF (10 mL) was added dropwise. After stirring for 1 h at Ϫ78 DCN (42 mg, 0.42 mmol), and decane (66 µL) in dry acetonitrile
°C, the mixture was warmed to 0 °C and Et2O (50 mL) and 2  (66 mL) was filled into irradiation tubes (11 mL each, pyrex glass,
aq. HCl (50 mL) were added; the mixture was stirred for 20 min at 12 mm diameter), degassed with argon for 20 min, and irradiated
room temp. before the layers were separated. The organic layer was
washed successively with sat. aq. NaHCO3, water, and brine
(50 mL each), dried (MgSO4), filtered, and concentrated in vacuo.
Purification by column chromatography (cyclohexane/EtOAc,
90:10) and kugelrohr distillation afforded 18 (780 mg, 52%) as a
with light (350 nm) until complete consumption of the starting ma-
terial. The reaction mixture was concentrated in vacuo and the resi-
due was subjected to column chromatography (cyclohexane/EtOAc,
90:10). The diastereoisomers were separated by HPLC (cyclo-
hexane/EtOAc, 85:15) affording (4aR*,4bS*,8aR*,10aS*)-10a-
colorless oil. NMR experiments: 1H, H/H-COSY, 13C, DEPT135, methyldodecahydro-4(1H)-phenanthrenone (16a) (55.0 mg, 56%)
HMQC. 1H NMR (500 MHz, CDCl3): δ ϭ 1.45Ϫ1.53 (m, 2 H, H-
and (4aR*,4bS*,8aS*,10aS*)-10a-methyldodecahydro-4(1H)-
6), 1.53Ϫ1.60 (m, 2 H, H-7), 1.84Ϫ1.90 (m, 2 H, H-8), 1.92 (m, 2 phenanthrenone (16b) (5.5 mg, 5.6%) in 62% combined isolated
H, H-5), 1.93 (m, 2 H, H-2), 2.08 (dd, J ϭ 7.8, 7.8 Hz, 2 H, H-9), yield. The diastereoisomeric ratio was determined by GC directly
2.25 (m, 2 H, H-1), 2.27 (m, 2 H, H-10), 2.31 (m, 2 H, H-3), 5.37
from the reaction mixture (16a:16b, 90:10).
3546
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 3535Ϫ3550

Cyclizations of Silyl Enol Ether Radical Cations
FULL PAPER
(4aR*,4bS*,8aR*,10aS*)-10a-Methyldodecahydro-4(1H)-phenan-
THF was added to a THF solution (70 mL) of 3-(1-cyclohexenyl-
ethyl)-3-methyl-1-cyclohexenyl trimethylsilyl ether (12) (1.00 g,
threnone (16a): NMR experiments: 1H, H/H-COSY, APT, HMQC,
1
HMBC, NOESY. H NMR (500 MHz, CDCl3): δ ϭ 0.81 (d, J ϭ 3.42 mmol) in a flask wrapped with aluminum foil. After stirring
0.86 Hz, 3 H, CH3), 0.82 (ddddd, J ϭ 3.9, 3.9, 11.2, 11.2, 11.2 Hz,
1 H, H-8a), 0.83 (ddddd, J ϭ 3.5, 3.5, 11.4, 11.4, 11.4 Hz, 1 H, H-
for 1 h, sat. aq. Na2S2O3 (100 mL) and Et2O (80 mL) were added.
The layers were separated and the organic layer was dried (MgSO4)
5ax), 0.97 (dddd, J ϭ 3.9, 11.3, 13.0, 13.0 Hz, 1 H, H-8ax), 1.01 and concentrated in vacuo to afford 30 (911 mg, 77%) as a yellow-
(ddddd, J ϭ 1.7, 1.7, 2.4, 4.3, 13.7 Hz, 1 H, H-1eq), 1.10 (ddddd,
J ϭ 3.4, 3.4, 12.7, 12.7, 12.7 Hz, 1 H, H-6ax), 1.13 (ddddd, J ϭ
3.4, 3.4, 12.3, 12.3, 12.3 Hz, 1 H, H-7ax), 1.20 (ddd, J ϭ 4.4, 13.3,
13.3 Hz, 1 H, H-10ax), 1.27 (ddddd, J ϭ 1.5, 3.3, 3.3, 3.3, 13.4 Hz,
ish, light-sensitive oil. NMR experiments: 1H, H/H-COSY, 13C,
DEPT135. 1H NMR (500 MHz, CDCl3): δ ϭ 1.06 (s, 3 H, Me),
1.11 (s, 3 H, Me), 1.35Ϫ1.68 (m, 17 H), 1.74Ϫ2.06 (m, 17 H),
2.23Ϫ2.34 (m, 2 H), 3.18Ϫ3.28 (m, 1 H), 3.33Ϫ3.44 (m, 1 H), 4.261
1 H, H-5eq), 1.31 (dddd, J ϭ 3.5, 11.4, 13.3, 13.3 Hz, 1 H, H-9ax), (dd, J ϭ 1.6, 1.6 Hz, 1 H, H-4a), 4.43 (dd, J ϭ 1.4, 1.4 Hz, 1 H,
1.38 (dddd, J ϭ 3.5, 10.9, 10.9, 11.1 Hz, 1 H, H-4b), 1.39 (dddd,
H-4a), 5.39 (m, 1 H, H-4b), 5.43 (m, 1 H, H-4b) ppm. 13C NMR
J ϭ 2.6, 4.5, 4.5, 13.2 Hz, 1 H, H-9eq), 1.479 (ddd, J ϭ 3.1, 3.1, (500 MHz, CDCl3): δ ϭ 19.70 (q, Me), 21.01/21.18 (t, C-2), 22.38/
13.2 Hz, 1 H, H-10eq), 1.595 (ddddd, J ϭ 3.2, 3.2, 3.2, 3.2, 12.9 Hz,
1 H, H-8eq), 1.61 (ddddd, J ϭ 3.2, 3.2, 3.2, 3.2, 12.5 Hz, 1 H, H-
22.45 (t, C-6), 22.92/22.97 (t, C-7), 25.19/25.21 (t, C-5), 27.26 (q,
Me), 28.42/28.51 (t, C-8), 31.01/31.17 (t), 32.56/32.65 (t, C-9), 34.74
7eq), 1.63 (ddd, J ϭ 1.5, 1.5, 11.4 Hz, 1 H, H-4a), 1.64 (dddddd, (t), 35.04 (t), 35.32 (t), 39.74 (s, C-10a), 40.56 (s, C-10a), 41.54 (t),
J ϭ 1.80, 3.2, 3.2, 3.2, 3.2, 11.4 Hz, 1 H, H-6eq), 1.79 (ddddd, J ϭ
4.7, 4.7, 13.7, 13.7, 13.7 Hz, 1 H, H-2ax), 1.89 (ddddd, J ϭ 1.6,
2.1, 4.9, 7.3, 13.7 Hz, 1 H, H-2eq), 2.07 (ddddd, J ϭ 1.6, 1.6, 1.6,
4.9, 13.7 Hz, 1 H, H-3eq), 2.15 (ddd, J ϭ 4.9, 13.7, 13.7 Hz, 1 H,
H-1ax), 2.31 (ddd, J ϭ 7.3, 13.7, 13.7 Hz, 1 H, H-3ax) ppm. 13C
NMR (125.8 MHz, CDCl3): δ ϭ 22.52 (t, C-2), 25.99 (t, C-7), 26.29
(t, C-6), 27.77 (q, Me), 29.06 (t, C-9), 29.43 (t, C-1), 30.69 (t, C-5),
33.87 (t, C-8), 37.35 (t, C-3), 37.95 (s, C-10a), 39.37 (t, C-10), 41.27
(d, C-4b), 42.11 (d, C-8a), 65.79 (d, C-4a), 215.69 (s, C-4) ppm.
GC/MS (EI, 70 eV): m/z (%) ϭ 221 (5), 220 (14), 205 (25), 202 (11),
111 (100), 109 (12), 108 (18), 95 (26), 93 (15), 91 (12), 81 (18),
79 (22), 67 (26), 55 (22), 41 (26). GC/MS (CI, isobutane): m/z (%) ϭ
222 (14), 221 (100), 220 (4), 203 (3), 114 (3). HRMS: calcd. for
C15H24O [Mϩ], m/z ϭ 220.1827; found 220.1826; deviation,
0.7 ppm. C15H24O (220.35): calcd. C 81.76, H 10.98; found C 81.87,
H 11.24. IR (film): ν˜ ϭ 2963, 2932, 2860, 2298, 1698, 1453, 1374,
1345, 1301, 1273, 1254, 1230, 1202, 1153, 1123, 1107, 1063, 1040,
47.17 (d, C-4a), 49.31 (d, C-4a), 121.20/121.43 (d, C-4b), 137.06/
137.18 (s, C-8a), 205.87/205.37 (s, C-4) ppm.
Derivatization for ESI-MS: A TLC-plate spotted with 30 was
dipped into a solution of Girard’s reagent T [N-(hydrazinocarbon-
ylmethyl)trimethylammonium chloride (4 mg/mL) in 0.1% aq. for-
mic acid] and dried for 5 min in an oven at 80 °C. The plate was
then subjected to ESI-MS. MS (ESI): m/z ϭ 460 [Mϩ], 333 [M Ϫ
I]ϩ, 350, 332 [M Ϫ I Ϫ 1]ϩ, 132. MS/MS (ESI): 459 [M Ϫ 1]ϩ, 333
[M Ϫ I]ϩ, 274 [M Ϫ I Ϫ NMe3]ϩ, 224 [M Ϫ I Ϫ C8H13]ϩ.
Radical Cyclization of 3-(1-Cyclohexenylethyl)-2-iodo-3-methyl-1-
cyclohexanone (30): AIBN (16 mg, 0.1 mmol) was added to a solu-
tion of 3-(1-cyclohexenylethyl)-2-iodo-3-methyl-1-cyclohexanone
(30) (519 mg, 1.50 mmol) and tributyltin hydride (437 mg,
1.5 mmol) in dry benzene (400 mL). The solution was stirred for
4 h under reflux and then concentrated in vacuo. The residue was
purified by column chromatography (cyclohexane/EtOAc, 90:10) to
afford a mixture of 16a and 16b (109 mg, 33%). The diastereoiso-
meric ratio was determined by GC directly from the reaction mix-
ture (16a:16b, 90:10).
974, 961, 904, 886, 852, 837, 786, 776, 730 cmϪ1
.
(4aR*,4bS*,8aS*,10aS*)-10a-Methyldodecahydro-4(1H)-phenan-
threnone (16b): NMR experiments: 1H, H/H-COSY, 13C, DEPT135,
1
HMQC, HMBC, NOESY. H NMR (500 MHz, CDCl3): δ ϭ 0.89
3-(1-Cyclohexenylethyl)-2-cyclopenten-1-one (17): As described for
the synthesis of 18, 17 (660 mg, 47%) was synthesized from 1-iodo-
2-(1-cyclohexenyl)ethane (19) (1.73 g, 7.33 mmol) and 3-ethoxycy-
clopent-2-enone (20)[64Ϫ67] (1.11 g, 8.80 mmol) after purification by
column chromatography (cyclohexane/EtOAc, 90:10) and kugel-
rohr distillation. NMR experiments: 1H, H/H-COSY, APT,
HMQC. 1H NMR (500 MHz, CDCl3): δ ϭ 1.43Ϫ1.57 (m, 2 H, H-
8), 1.58Ϫ1.62 (m, 2 H, H-7), 1.82Ϫ1.90 (m, 2 H, H-6), 1.90Ϫ1.97
(m, 2 H, H-9), 2.16 (t, J ϭ 7.8 Hz, 2 H, H-5), 2.30Ϫ2.37 (m, 2 H,
H-2), 2.47 (t, J ϭ 7.8 Hz, 2 H, H-4), 2.51Ϫ2.57 (m, 2 H, H-3), 5.38
(m, 1 H, H-9a), 5.90 (m, 1 H, H-9b) ppm. 13C NMR (125.8 MHz,
CDCl3): δ ϭ 22.23 (t, C-8), 22.70 (t, C-7), 25.01 (t, C-9), 28.06 (t,
C-8), 31.42 (t, C-3), 31.51 (t, C-4), 35.11 (t, C-2/5), 35.14 (t, C-2/
5), 121.82 (d, C-9a), 129.37 (d, C-9b), 135.89 (s, C-5a), 182.87 (s,
C-3a), 210.04 (s, C-1) ppm. GC/MS (EI, 70 eV): m/z (%) ϭ
191 (0.8), 190 (3), 148 (55), 133 (17), 97 (14), 96 (100), 95 (90),
94 (10), 93 (16), 91 (23), 81 (13), 79 (39), 77 (26), 67 (92), 66 (18),
65 (26), 55 (58), 53 (46), 52 (11), 51 (15), 41 (69), 39 (46), 29 (13),
28 (10), 27 (30). GC/MS (CI, isobutane): m/z (%) ϭ 192 (13),
191 (100). HRMS: calcd. for C13H18O [Mϩ], m/z ϭ 190.1358; found
190.1351; deviation, 3.4 ppm. IR (film): ν˜ ϭ 2923, 2859, 2838,
1714, 1677, 1614, 1438, 1409, 1336, 1268, 1229, 1186, 1135, 1079,
(s, 3 H, CH3), 0.99 (ddddd, J ϭ 1.4, 1.4, 2.7, 4.3, 13.8 Hz, 1 H, H-
1eq), 1.13Ϫ1.46 (m, 9 H, H-5, H-5, H-6, H-6, H-7, H-8eq, H-9, H-
10, H-10), 1.52 (dddd, J ϭ 3.6, 13.0, 13.0, 13.0 Hz, 1 H, H-8ax),
1.68Ϫ1.77 (m, 2 H, H-7, H-8a), 1.83 (ddddd, J ϭ 4.7, 4.7, 13.6,
13.6, 13.6 Hz, 1 H, H-2ax), 1.83 Ϫ1.95 (m, 2 H, H-2eq, H-9), 2.10
(ddddd, J ϭ 1.7, 1.7, 1.7, 4.9, 13.8 Hz, 1 H, H-3eq), 2.16 (dddd,
J ϭ 3.7, 3.7, 3.7, 12.4 Hz, 1 H, H-4b), 2.266 (ddd, J ϭ 5.0, 13.7,
13.7 Hz, 1 H, H-1ax), 2.36 (ddd, 1.5, 1.5, 12.4 Hz, 1 H, H-4a), 2.38
(ddd, J ϭ 7.4, 13.7, 13.7 Hz, 1 H, H-3ax) ppm. 13C NMR
(500 MHz, CDCl3): δ ϭ 19.66 (t, C-6), 22.38 (t, C-2), 25.39 (t, C-
8), 26.41 (t, C-7), 27.14 (t, C-9), 28.20 (q, CH3), 28.57 (t, C-5),
28.79 (t, C-1), 33.86 (t, C-10), 34.38 (d, C-4b), 35.76 (d, C-8a),
37.06 (t, C-3), 37.81 (s, C-10a), 56.22 (d, C-4a), 216.43 (s, C-4) ppm.
GC/MS (EI, 70 eV): m/z (%) ϭ 221 (2), 220 (8), 112 (10), 111 (100),
109 (11), 108 (13), 95 (18), 93 (11), 91 (12), 81 (11), 79 (19), 67 (16),
55 (15), 41 (20). GC/MS (CI, isobutane): m/z (%) ϭ 222 (13),
221 (100), 220 (3), 203 (5), 111 (4). HRMS: calcd. for C15H24O
[Mϩ], m/z 220.1827; found 220.1824; deviation, 1.5 ppm. C15H24O
(220.35): calcd. C 81.76, H 10.98; found C 81.20, H 10.98. IR
(film): ν˜ ϭ 2963, 2931, 2869, 2852, 2673, 2328, 1846, 1796, 1774,
1753, 1740, 1729, 1692, 1546, 1453, 1443, 1427, 1383, 1345, 1309,
1286, 1261, 1240, 1210, 1156, 1123, 1071, 1054, 989, 953, 898, 876,
1049, 973, 918, 840, 801 cmϪ1
.
830, 820, 780, 730, 665 cmϪ1
.
3-(1-Cyclohexenylethyl)-3-methyl-1-cyclopentenyl
Trimethylsilyl
3-(1-Cyclohexenylethyl)-2-iodo-3-methyl-1-cyclohexanone
(30):
Ether (11): As described for the synthesis of 12, 11 (4.50 mg, 93%)
Freshly prepared N-iodosuccinimide (773 mg, 3.42 mmol) in dry was synthesized as a colorless oil from 3-(1-cyclohexenylethyl)-2-
Eur. J. Org. Chem. 2004, 3535Ϫ3550
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3547

J. O. Bunte, E. K. Heilmann, B. Hein, J. Mattay
FULL PAPER
cyclopenten-1-one (17) (3.34 g, 17.5 mmol). NMR experiments: 1H,
97 (100), 96 (16), 95 (17), 93 (17), 91 (12), 81 (39), 79 (26), 77 (10),
H/H-COSY, APT, HMQC. 1H NMR (500 MHz, CDCl3): δ ϭ 0.20 67 (28), 55 (30), 53 (17), 41 (33), 39 (13), 29 (14), 28 (10), 27 (14).
[s, 9 H, Si(CH3)3], 1.02 (s, 3 H, CH3), 1.34Ϫ1.46 (m, 2 H, H-4), GC/MS (CI, isobutane): m/z (%) ϭ 208 (25), 207 (100), 206 (29),
1.51Ϫ1.57 (m, 2 H, H-8), 1.55 (ddd, J ϭ 6.0, 9.1, 12.7 Hz, 1 H, H-
3), 1.57Ϫ1.63 (m, 2 H, H-7), 1.71 (ddd, J ϭ 5.8, 8.9, 12.7 Hz, 1 H, 206.1662; deviation, 4.4 ppm. IR (film): ν˜ ϭ 2923, 2857, 2243,
189 (18). HRMS: calcd. for C14H22O [Mϩ], m/z 206.1671; found
H-3), 1.81Ϫ1.91 (m, 2 H, H-5), 1.88Ϫ1.94 (m, 2 H, H-6),
1739, 1449, 1412, 1379, 1253, 1217, 1173, 1156, 1141, 1121, 1088,
1.94Ϫ2.00 (m, 2 H, H-9), 2.26 (dddd, J ϭ 1.6, 5.8, 8.7, 15.9 Hz, 1 1059, 1035, 988, 919, 857, 835, 665 cmϪ1
H, H-2), 2.31 (dddd, J ϭ 1.6, 5.8, 8.9, 15.9 Hz, 1 H, H-2), 4.49 (dd,
.
J ϭ 1.6, 1.6 Hz, 1 H, H-9b), 5.38 (m, 1 H, H-9a) ppm. 13C NMR
(125.8 MHz, CDCl3): δ ϭ Ϫ0.01 [q, Si(CH3)3], 22.60 (t, C-8), 23.07
(t, C-7), 25.25 (t, C-9), 28.02 (q, Me), 28.58 (t, C-6), 33.22 (t, C-2),
33.62 (t, C-5), 34.89 (t, C-3), 41.00 (t, C-4), 44.94 (s, C-3a), 112.23
(d, C-9b), 120.00 (d, C-9a), 138.66 (s, C-5a), 152.86 (s, C-1) ppm.
GC/MS (EI, 70 eV): m/z (%) ϭ 279 (1), 278 (5), 170 (27), 169 (100),
75 (12), 73 (48), 45 (12). HRMS: calcd. for C17H30OSi [Mϩ], m/z
278.2067; found 278.2067. C17H30OSi (278.51): calcd. C 73.31, H
10.86; found C 71.82, H 10.54. IR (film): ν˜ ϭ 2930, 2862, 1747,
1716, 1644, 1453, 1439, 1371, 1341, 1308, 1252, 1197, 1135, 1074,
(3aR*,5aS*,9aS*,9bR*)-3a-Methyldodecahydro-1H-cyclopenta[a]-
1
naphthalen-1-one (15b): NMR experiments: H; H/H-COSY; APT;
1
HMQC; HMBC at T ϭ 370 and 213 K; H NMR spectra in steps
1
of 10 K between T ϭ 213 and 370 K; NOESY at T ϭ 370 K. H
NMR (600 MHz, C2D2Cl4, T ϭ 370 K): δ ϭ 1.16 (s, 3 H, Me),
1.22Ϫ1.47 (m, 10 H), 1.51 (ddd, J ϭ 9.2, 9.2, 12.9 Hz, 1 H, H-3),
1.58Ϫ1.70 (m, 2 H, H-5, H-8), 1.67 (dd, J ϭ 1.1, 4.1 Hz, 1 H, H-
9b), 1.71 (ddddd, J ϭ 3.5, 3.5, 10.4, 10.4, 10.4 Hz, 1 H, H-9ax),
1.76 (ddd, J ϭ 6.6, 6.6, 13.0 Hz, 1 H, H-3), 2.04 (dddd, J ϭ 4.3,
4.3, 4.3, 10.5 Hz, 1 H, H-9a), 2.18 (ddd, J ϭ 6.9, 9.1, 14.7 Hz, 1
H, H-2), 2.20 (ddd, J ϭ 6.9, 9.1, 14.7 Hz, 1 H, H-2) ppm. 13C NMR
(150.96 MHz, C2D2Cl4, T ϭ 370 K): δ ϭ 22.95 (t, C-7), 23.98 (t,
C-5), 25.72 (t, C-8), 28.02 (t, C-9), 28.83 (q, Me), 30.65 (t, C-6),
33.89 (t, C-4), 34.08 (t, C-3), 34.77 (d, C-5a), 34.94 (d, C-9a), 35.03
1047, 997, 931, 871, 844, 803, 757, 689, 665 cmϪ1
.
PET Cyclization of 3-(1-Cyclohexenylethyl)-3-methyl-1-cyclopen-
tenyl Trimethylsilyl Ether (11) Using DCN as a Sensitizer: A solu-
tion of 3-(1-cyclohexenylethyl)-3-methyl-1-cyclopentenyl trimethyl-
1
(t, C-2), 38.38 (s, C-3a), 60.23 (d, C-9b), 219.83 (s, C-1) ppm. H
silyl ether (11) (196 mg, 0.95 mmol) and DCN (50 mg, 0.28 mmol) NMR (600 MHz, CD2Cl2, T ϭ 213 K): δ ϭ 1.01 (dddd, J ϭ 3.4,
in dry acetonitrile (55 mL) was filled into irradiation tubes (11 mL
each, pyrex glass, 12 mm diameter), degassed with argon for
20 min, and irradiated with light (350 nm) until complete consump-
tion of the starting material occurred. The reaction mixture was
concentrated in vacuo and the residue was subjected to column
chromatography (cyclohexane/EtOAc, 90:10) to afford a mixture of
(3aR*,5aR*,9aS*,9bR*)-3a-methyldodecahydro-1H-cyclo-
3.4, 3.4, 13.3 Hz, 1 H, H-5eq), 1.10Ϫ1.38 (m, 7 H), 1.194 (s, 3 H,
Me), 1.38Ϫ1.46 (m, 2 H), 1.50Ϫ1.65 (m, 4 H), 1.61 (m, 1 H, H-
9b), 1.69 (dddd, J ϭ 3.2, 13.2, 13.2, 13.2 Hz, 1 H, H-5ax), 2.08
(ddd, J ϭ 4.1, 4.1, 13.4 Hz, 1 H, H-9a), 2.13Ϫ2.25 (m, 2 H, H-2)
ppm. 13C NMR (150.96 MHz, CD2Cl2, T ϭ 213 K): δ ϭ 20.54 (t),
21.31 (t, C-5), 26.59 (t), 26.82 (t), 27.24 (q, Me), 31.32 (t), 33.38
(d, C-5a), 33.64 (t, C-4), 34.28 (t, C-3), 34.38 (d, C-9a), 34.44 (t, C-
2), 37.65 (s, C-3a), 61.46 (d, C-9b), 221.21 (s, C-1) ppm. GC/MS
(EI, 70 eV): m/z (%) ϭ 207 (2), 206 (13), 191 (12), 97 (100), 41 (13).
GC/MS (CI, isobutane): m/z (%) ϭ 208 (15), 207 (100), 206 (24),
205 (19), 189 (23). HRMS: calcd. for C14H22O [Mϩ], m/z ϭ
206.1671; found 206.16668; deviation, 0.78 ppm. IR (film): ν˜ ϭ
2928, 2863, 2667, 1736, 1452, 1410, 1381, 1350, 1286, 1264, 1202,
1181, 1163, 1147, 1111, 1091, 1073, 1023, 994, 953, 904, 870, 799
penta[a]naphthalen-1-one
(15a),
(3aR*,5aS*,9aS*,9bR*)-3a-
methyldodecahydro-1H-cyclopenta[a]naphthalen-1-one (15b), and
(3aR*,5aS*,9aR*,9bR*)-3a-methyldodecahydro-1H-cyclopenta-
[a]naphthalen-1-one (15c) (64 mg, 33%). The diastereoisomeric
ratio was determined by GC directly from the reaction mixture
(15a/15b/15c ϭ 41:31:28). Isomer 15c was separated by HPLC
(cyclohexane/EtOAc, 92.5:7.5); 15a and 15b were isolated by pre-
parative GC.
cmϪ1
.
(3aR*,5aR*,9aS*,9bR*)-3a-Methyldodecahydro-1H-cyclopenta[a]-
(3aR*,5aS*,9aR*,9bR*)-3a-Methyldodecahydro-1H-cyclopenta[a]-
1
naphthalen-1-one (15a): NMR experiments: H, H/H-COSY, APT,
naphthalen-1-one (15c): NMR experiments: 1H, H/H-COSY, 13C,
HMQC, HMBC, NOESY. 1H NMR (500 MHz, C6D6): δ ϭ 0.60 DEPT135, HMQC, HMBC. H NMR (600 MHz, C6D6): δ ϭ 0.85
1
(ddddd, J ϭ 3.3, 3.3, 10.5, 11.5, 11.5 Hz, 1 H, H-5a), 0.704 (dddd,
(dddd, J ϭ 3.8, 11.0, 13.0, 13.0 Hz, 1 H, H-6ax), 0.92 (d, J ϭ
J ϭ 3.3, 11.0, 11.0, 11.0 Hz, 1 H, H-9a), 0.73 (d, J ϭ 0.9 Hz, 3 H, 0.9 Hz, 3 H, CH3), 0.99 (dddd, J ϭ 3.6, 11.2, 13.3, 13.3 Hz, 1 H,
CH3), 0.85 (dddd, J ϭ 3.7, 11.4, 12.8, 12.8 Hz, 1 H, H-6ax), 0.90 H-5ax), 1.02 (ddd, J ϭ 2.2, 3.7, 13.5 Hz, 1 H, H-4eq), 1.07 (dddd,
(dddd, J ϭ 3.4, 11.3, 12.8, 12.8 Hz, 1 H, H-9ax), 1.00 (dddd, J ϭ
J ϭ 0.8, 3.7, 13.6, 13.6 Hz, 1 H, H-4ax), 1.16 (ddddd, J ϭ 3.3, 3.3,
1.4, 5.7, 5.7, 13.0 Hz, 1 H, Ha-3), 1.02 (dddd, J ϭ 4.1, 13.1, 13.1, 11.2, 13.0, 13.0 Hz, 1 H, H-5a), 1.18 (ddd, J ϭ 10.4, 10.4, 12.6 Hz,
13.1 Hz, 1 H, H-5ax), 1.09 (ddddd, J ϭ 3.3, 3.3, 12.8, 12.8, 12.8 Hz,
1 H, H-7ax), 1.16 (ddddd, J ϭ 3.4, 3.4, 12.8, 12.8, 12.8 Hz, 1 H,
1 H, H-3), 1.18 (ddddd, J ϭ 3.7, 3.7, 13.0, 13.0, 13.0 Hz, 1 H, H-
8ax), 1.19 (m, 1 H, H-9a), 1.29 (ddd, J ϭ 2.0, 8.9, 12.6 Hz, 1 H,
H-8ax), 1.18 (ddd, J ϭ 4.5, 13.2, 13.2 Hz, 1 H, H-4ax), 1.26 (dddd, H-3), 1.29 (m, 1 H, H-5eq), 1.37 (ddddd, J ϭ 3.6, 3.6, 13.1, 13.1,
J ϭ 2.8, 2.8, 4.3, 13.9 Hz, 1 H, H-5eq), 1.33 (dd, J ϭ 1.5, 11.1 Hz,
1 H, H-9b), 1.34 (ddd, J ϭ 2.5, 4.0, 13.3 Hz, 1 H, H-4eq), 1.54
(ddddd, J ϭ 1.8, 3.3, 3.3, 3.3, 13.03 Hz, 1 H, H-6eq), 1.64 (ddddd,
13.1 Hz, 1 H, H-7ax), 1.42 (ddd, J ϭ 1.4, 2.3, 3.8 Hz, 1 H, H-9b),
1.52 (ddddd, J ϭ 1.8, 3.3, 3.3, 3.3, 13.4 Hz, 1 H, H-9eq), 1.60
(ddddd, J ϭ 1.8, 3.3, 3.3, 3.3, 12.9 Hz, 1 H, H-6eq), 1.69 (dddddd,
J ϭ 1.72, 3.3, 3.3, 3.3, 12.7 Hz, 1 H, H-8eq), 1.70 (dddddd, J ϭ J ϭ 1.7, 3.3, 3.3, 3.3, 3.3, 13.0 Hz, 1 H, H-7eq), 1.88 (dddddd, J ϭ
1.8, 3.3, 3.3, 3.3, 3.3, 12.8 Hz, 1 H, H-7eq), 1.79 (ddd, J ϭ 10.4,
10.4, 12.9 Hz, 1 H, Hi-3), 2.06 (ddd, J ϭ 5.6, 10.4, 19.7 Hz, 1 H,
1.8, 3.3, 3.3, 3.3, 3.3, 12.7 Hz, 1 H, H-8eq), 1.92 (dddd, J ϭ 1.3,
2.0, 10.0, 19.2 Hz, 1 H, H-2), 2.00 (ddd, J ϭ 8.8, 10.5, 19.2 Hz, 1
H-2), 2.11 (ddd, J ϭ 5.6, 9.9, 19.7 Hz, 1 H, H-2), 2.31 (ddddd, J ϭ H, H-2), 2.79 (dddd, J ϭ 3.7, 11.8, 13.4, 13.4 Hz, 1 H, H-9ax) ppm.
1.7, 3.3, 3.3, 3.3, 13.3 Hz,
H, H-9eq) ppm. 13C NMR 13C NMR (150.96 MHz, C6D6): δ ϭ 24.86 (q, CH3), 26.63 (t, C-
(150.96 MHz): δ ϭ 26.53 (t, C-8), 26.73 (t, C-7), 29.28 (t, C-3), 7), 27.74 (t, C-8), 29.16 (t, C-9), 30.65 (t, C-5), 33.14 (t, C-4), 34.50
29.38 (q, CH3), 29.53 (t, C-5), 31.09 (t, C-9), 34.42 (t, C-6), 34.48 (t, C-3), 34.59 (t, C-6), 35.16 (t, C-2), 37.69 (d, C-5a), 39.65 (s, C-
(t, C-2), 35.03 (t, C-4), 38.56 (s, C-3a), 39.86 (d, C-9a), 40.97 (d, 3a), 40.12 (d, C-9a), 59.44 (d, C-9b), 216.90 (s, C-1) ppm. GC/MS
C-5a), 61.97 (d, C-9b), 217.15 (s, C-1) ppm. GC/MS (EI, 70 eV): (EI, 70 eV): m/z (%) ϭ 207 (2), 206 (10), 97 (100), 79 (10), 67 (10),
m/z (%) ϭ 207 (4), 206 (28), 191 (23), 108 (37), 107 (15), 98 (17), 55 (11), 41 (14). GC/MS (CI, isobutane): m/z (%) ϭ 207 (100),
1
3548
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 3535Ϫ3550

Cyclizations of Silyl Enol Ether Radical Cations
FULL PAPER
206 (26), 205 (21), 192 (12), 191 (11), 189 (74). HRMS: calcd. for
C14H22O [Mϩ], m/z ϭ 206.1671; found 206.1663; deviation,
3.6 ppm. IR (film): ν˜ ϭ 2913, 2854, 2822, 2668, 2307, 1813, 1735,
1449, 1407, 1379, 1344, 1317, 1288, 1261, 1229, 1199, 1169, 1127,
[1]
[2]
[3]
P. Brownbridge, Synthesis 1983, 1Ϫ28.
P. Brownbridge, Synthesis 1983, 85Ϫ104.
I. Fleming, A. Barbero, D. Walter, Chem. Rev. 1997, 97,
2063Ϫ2192.
G. M. Rubottom, M. I. Lopez Nieves, Tetrahedron Lett. 1972,
24, 2423Ϫ2425.
E. Friedrich, W. Lutz, Chem. Ber. 1980, 113, 1245Ϫ1263.
S. Sakaguchi, Y. Yamamoto, T. Sugimoto, H. Yamamoto, Y.
Ishii, J. Org. Chem. 1999, 64, 5954Ϫ5957.
S. Stankovic, J. H. Espenson, J. Org. Chem. 1998, 63,
4129Ϫ4130.
K. Kaneda, K. Nobuyoshi, K. Jitsukawa, S. Teranishi, Tetra-
hedron Lett. 1981, 22, 2595Ϫ2598.
G. M. Rubottom, J. M. Gruber, R. K. Boeckman Jr., M. Ra-
maiah, J. B. Medwid, Tetrahedron Lett. 1978, 19, 4603Ϫ4606.
P. A. Zoretic, M. Wang, Y. Zhang, Z. Shen, A. A. Ribiero, J.
Org. Chem. 1996, 61, 1806Ϫ1813.
Y. Horiguchi, E. Nakamura, E. Kuwajima, Tetrahedron Lett.
1989, 30, 3323Ϫ3326.
W. Adam, R. T. Fell, C. R. Saha-Möller, C.-G. Zhao, Tetra-
hedron: Asymmetry 1998, 9, 397Ϫ401.
J. P. McCormick, W. Tomasik, M. W. Johnson, Tetrahedron
Lett. 1981, 22, 607Ϫ610.
J. O. Bunte, J. Mattay in CRC Handbook of Organic Photo-
chemistry and Photobiology (Eds.: W. M. Horspool, F. Len-
ci),CRC Press, Boca Raton, 2nd ed., CRC Press, Boca Raton,
USA, 2004, chapters 10, 10.1Ϫ10.16.
D. L. Wright, C. R. Whitehead, E. H. Sessions, I. Ghiviriga,
D. A. Frey, Org. Lett. 1999, 1, 1535Ϫ1538.
K. D. Moeller, M. R. Marzabadi, D. G. New, M. Y. Chiang,
S. Keith, J. Am. Chem. Soc. 1990, 112, 6123Ϫ6124.
C. M. Hudson, M. R. Marzabadi, K. D. Moeller, D. G. New,
J. Am. Chem. Soc. 1991, 113, 7372Ϫ7385.
1047, 986, 954, 942, 887, 862, 842, 798 cmϪ1
.
[4]
PET Cyclization of 3-(1-Cyclohexenylethyl)-3-methyl-1-cyclopenten-
yltrimethylsilyl Ether (11) using DCA/Phenanthrene: A solution of
[5]
[6]
3-(1-cyclohexenylethyl)-3-methyl-1-cyclopentenyl
trimethylsilyl
ether (11) (521 mg, 1.87 mmol), DCA (60 mg, 0.26 mmol), and
phenanthrene (360 mg, 2.0 mmol) in dry acetonitrile (66 mL) was
filled into irradiation tubes (11 mL each, pyrex glass, 12 mm diam-
eter), degassed with argon for 20 min, and irradiated with light
(420 nm) until complete consumption of the starting material oc-
curred. The reaction mixture was concentrated in vacuo and the
residue was subjected to column chromatography (cyclohexane/
EtOAc, 90:10) to afford a mixture of 15a, 15b, and 15c (167 mg,
43%). The diastereoisomeric ratio was determined by GC directly
from the reaction mixture (15a/15b/15c, 41:31:28).
[7]
[8]
[9]
[10]
[11]
[12]
[13]
[14]
3-(1-Cyclohexenylethyl)-2-iodo-3-methyl-1-cyclopentanone (29): As
described for the synthesis of 30, 29 (285 mg, 83%) was synthesized
as a yellowish, light-sensitive oil from 3-(1-cyclohexenylethyl)-3-
methyl-1-cyclopentenyl trimethylsilyl ether (11). NMR experi-
ments: 1H, H/H-COSY, 13C. 1H NMR (500 MHz, CDCl3): δ ϭ
1.11 (s, 3 H, Me), 1.14 (s, 3 H, Me), 1.44Ϫ1.67 (m, 13 H), 1.74
(dddd, J ϭ 1.4, 3.1, 9.0, 13.0 Hz, 1 H), 1.84Ϫ2.08 (m, 15 H), 2.25
(dddd, J ϭ 1.2, 8.8, 8.8., 19.7 Hz, 1 H), 2.28Ϫ2.38 (m, 2 H), 2.44
(dddd, J ϭ 0.6, 3.1, 9.9, 19.8 Hz, 1 H), 4.22 (dd, J ϭ 1.5, 1.5 Hz,
1 H, H-9b), 4.48 (dd, J ϭ 0.8, 0.8 Hz, 1 H, H-9b), 5.41 (m, 1 H,
H-9a), 5.44 (m, 1 H, H-9a) ppm. 13C NMR (500 MHz, CDCl3):
δ ϭ 20.78 (q, Me), 22.39/22.44 (t, C-8), 22.91/22.95 (t, C-7), 25.18/
25.19 (t, C-9), 25.69 (q, Me), 28.46/28.49 (t, C-6), 31.40, 32.40/
32.42, 32.74/32.78, 32.89, 40.83, 42.17, 42.51, 44.66, 45.73, 121.16/
121.39 (d, C-9a), 137.03/137.16 (s, C-5a), 211.51/211.53 (s, C-1)
ppm. Derivatization for ESI-MS: A TLC-plate spotted with 29 was
dipped into a solution of Girard’s reagent T [N-(hydrazinocarbon-
ylmethyl)trimethylammonium chloride (4 mg/mL) in 0.1% aq. for-
mic acid] and dried for 5 min in an oven at 80 °C. The plate was
then subjected to ESI-MS. MS (ESI): m/z ϭ 478, 446 [M]ϩ, 355,
336 [M Ϫ C8H14]ϩ, 318 [M Ϫ I Ϫ 1]ϩ, 172, 132.
[15]
[16]
[17]
[18]
M. Schmittel, M. Keller, A. Burghart, J. Chem. Soc., Perkin
Trans. 2 1995, 2327Ϫ2333.
[19]
[20]
B. B. Snider, T. Kwon, J. Org. Chem. 1992, 57, 2399Ϫ2410.
`
J.-P. Gourves, R. Ruzziconi, L. Vilarroig, J. Org. Chem. 2001,
66, 617Ϫ619.
[21]
[22]
[23]
R. Rathore, J. K. Kochi, J. Org. Chem. 1996, 61, 627Ϫ639.
C. A. Ramsden, R. G. Smith, Org. Lett. 1999, 1, 1591Ϫ1594.
T. M. Bockman, D. Shukla, J. K. Kochi, J. Chem. Soc., Perkin
Trans. 2 1996, 1623Ϫ1632.
T. M. Bockman, J. K. Kochi, J. Chem. Soc., Perkin Trans. 2
1996, 1633Ϫ1643.
A. Heidbreder, J. Mattay, Tetrahedron Lett. 1992, 33,
1973Ϫ1976.
L. Ackermann, A. Heidbreder, F. Wurche, F.-G. Klärner, J.
Mattay, J. Chem. Soc., Perkin Trans. 2 1999, 863Ϫ869.
S. Hintz, R. Fröhlich, J. Mattay, Tetrahedron Lett. 1996, 37,
7349Ϫ7352.
[24]
[25]
[26]
[27]
[28]
Radical Cyclization of 3-(1-Cyclohexenylethyl)-2-iodo-3-methyl-1-
cyclopentanone (29): AIBN (16 mg, 0.1 mmol) was added to a solu-
tion of 3-(1-cyclohexenylethyl)-2-iodo-3-methyl-1-cyclopentanone
(29) (664 mg, 2.0 mmol) and tributyltin hydride (582 mg, 2.0 mmol)
in dry benzene (400 mL). The solution was stirred for 4.5 h under
reflux and then concentrated in vacuo. The residue was purified
by column chromatography (cyclohexane/EtOAc, 90:10) and the
isomers were separated by HPLC (cyclohexane/EtOAc, 92.5:7.5) to
afford 15a and 15b (170 mg, 42%). The diastereoisomeric ratio was
determined by GC directly from the reaction mixture (15a:15b,
52:48).
S. Hintz, J. Mattay, R. van Eldik, W.-F. Fu, Eur. J. Org. Chem.
1998, 1583Ϫ1596.
[29]
[30]
[31]
J. O. Bunte, S. Rinne, J. Mattay, Synthesis 2004, 619Ϫ633.
D. P. Curran, C.-T. Chang, J. Org. Chem. 1989, 54, 3140Ϫ3157.
J. O. Bunte, S. Rinne, C. Schäfer, B. Neumann, H.-G.
Stammler, J. Mattay, Tetrahedron Lett. 2003, 44, 45Ϫ48.
H. Rinderhagen, J. Mattay, Chem. Eur. J. 2004, 10, 851Ϫ874.
D. P. Curran, N. A. Porter, B. Giese, Stereochemistry of Radical
Reactions, VCH, Weinheim, 1996.
J. Fossey, D. Lefort, J. Sorba, Free Radicals in Organic Chemis-
try, Wiley, Chichester, New York, 1995.
B. B. Snider, B. O. Buckman, J. Org. Chem. 1992, 57,
4883Ϫ4888.
Y. D. Vankar, G. Kumaravel, Tetrahedron Lett. 1984, 25,
233Ϫ236.
All calculations were conducted using TITAN V1.05,
Schrödinger Inc., Wavefunction Inc., 2000, Schrödinger Devel-
opment Staff: M. D. Beachy, Y. Cao, R. B. Murphy, J. K. Perry,
W. T. Pollard, M. N. Ringnalda, G. R. Yacek, J. R. Wright;
[32]
[33]
Acknowledgments
[34]
[35]
[36]
[37]
Financial support by the Ministerium für Schule, Wissenschaft und
Forschung des Landes Nordrhein-Westfalen (MSWF-NRW), the
Deutsche Forschungsgemeinschaft (DFG), the Fonds der Chem-
ischen Industrie, and the Innovationsfonds der Universität Bielefeld
is gratefully acknowledged. We also thank the state of Nordrhein-
Westfalen for granting a predoctoral scholarship to J. O. B. We
wish to thank Prof. Dr. W. W. Schoeller for his support concerning
the calculations and Dr. Michael Oelgemöller for his help in pre-
paring the manuscript.
Eur. J. Org. Chem. 2004, 3535Ϫ3550
www.eurjoc.org
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3549

J. O. Bunte, E. K. Heilmann, B. Hein, J. Mattay
P. P. Gaspar, Science 2002, 297, 785Ϫ786.
J. P. Foster, F. Weingold, J. Am. Chem. Soc. 1980, 102,
7211Ϫ7218.
A. E. Reed, F. Weinhold, J. Chem. Phys. 1983, 78, 4066Ϫ4073.
A. E. Reed, R. B. Weinstock, F. Weinhold, J. Chem. Phys. 1985,
83, 735.
FULL PAPER
[52]
[53]
Wavefunction Development Staff: B. J. Deppmeier, A. J.
Driessen, W. J. Hehre, J. A. Johnson, P. E. Klunziger, M. Wana-
tabe, J. Yu.
H. Fischer, L. Radom, Angew. Chem. 2001, 113, 1380; Angew.
Chem. Int. Ed. 2001, 40, 1340Ϫ1371.
[54]
[55]
[38]
[39]
H. Zuilhof, J. P. Dinnocenzo, A. C. Reddy, S. Shaik, J. Phys.
[56]
[57]
J. E. Carpenter, F. Weinhold, J. Mol. Struct. (Theochem) 1988,
169, 41Ϫ62.
Chem. 1996, 100, 15774Ϫ15784.
A. D. Becke, Phys. Rev. A 1988, 38, 3098Ϫ3100.
C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785Ϫ789.
P. C. Hariharan, J. A. Pople, Chem. Phys. Lett. 1985, 82, 270.
[40]
[41]
Calculations for 50 Ǟ 58 were conducted using B3LYP/
6Ϫ31G*; for 51 to 59, calculations were conducted using
B3LYP/3Ϫ21G*.
N. L. Bauld, Radicals, Ion Radicals and Triplets, The Spin-
Bearing Intermediates of Organic Chemistry, Wiley-VCH, New
York, 1997, p. 159.
J. H. Horner, E. Taxil, M. Newcomb, J. Am. Chem. Soc. 2002,
124, 5402Ϫ5410.
J. P. Dinnocenzo, S. Farid, J. L. Goodman, I. R. Gould, W. P.
Todd, S. L. Mattes, J. Am. Chem. Soc. 1989, 111, 8973Ϫ8975.
O. Poizat, G. Buntinx, P. Valat, V. Wintgens, M. Bridoux, J.
Phys. Chem. 1993, 97, 5905Ϫ5910.
W. L. F. Armarego, D. D. Perrin in Purification of Laboratory
Chemicals, Butterworth-Heinemann, Oxford, 4th ed., 1996.
E. J. Kantorowski, S. W. E. Eisenberg, W. H. Fink, M. J. Kurth,
J. Org. Chem. 1999, 64, 570Ϫ580.
B. B. Kikani, J. T. McKee, M. Zanger, Synthesis 1990, 176.
E. Campaigne, T. P. Selby, J. Heterocycl. Chem. 1980, 17, 1255.
R. Fuchs, J. F. McGarrity, Synthesis 1992, 373.
L. Duc, J. F. McGarrity, T. Meul, A. Warm, Synthesis 1992,
391.
[42]
[43]
X. Fang, H. Xia, H. Yu, X. Dong, M. Chen, Q. Wang, F. Tao,
[58]
C. Li, J. Org. Chem. 2002, 67, 8481Ϫ8488.
A. L. J. Beckwith, C. J. Easton, A. K. Serelis, J. Chem. Soc.,
[44]
Chem. Commun. 1980, 482Ϫ483.
A. L. J. Beckwith, T. Lawrence, A. K. Serelis, J. Chem. Soc.,
[59]
[60]
[61]
[62]
[63]
[45]
Chem. Commun. 1980, 484Ϫ485.
D. C. Spellmeyer, K. N. Houk, J. Org. Chem. 1987, 52,
959Ϫ974.
[46]
[47]
The complete collection of calculated structures and transition
states is available in the supporting information.
J. Fossey, D. Lefort, J. Sorba, Free Radicals in Organic Chemis-
[48]
try, Wiley, Chichester, New York, 1995, p. 289.
Landolt-Börnstein, Numerical Data and Functional Realtionsh-
[49]
[64]
[65]
[66]
[67]
ips in Science and Technology, Group II: Molecules and Rad-
icals, Volume 13, Subvolume A: Carbon-Centered Radicals I,
Springer, http://www.landolt-boernstein.com, p. 289.
[50]
G. K. S. Prakash, C. Bae, G. Rasul, G. A. Olah, J. Org. Chem.
2002, 67, 1297Ϫ1301.
C. A. Reed, Acc. Chem. Res. 1998, 31, 325Ϫ332.
[51]
Received February 24, 2004
3550
 2004 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
Eur. J. Org. Chem. 2004, 3535Ϫ3550