Selective tandem enyne metathesis for the synthesis of functionalized cycloheptadienes

Article abstract of DOI:10.1016/j.tet.2008.03.027

The regio- and site-selective ring expansion of dienes and the regioselective ring expansion of substituted cyclopentenes provide 1,3-cycloheptadienes by enyne metathesis under methylene-free conditions. Site-selectivity results from differential ring strain among two different cycloalkenes in diene reactants. The high regioselectivity found in the ring expansion of tetrahydroindene (THI) is explained on the basis of a selective ring opening by the second generation Grubbs' ruthenium carbene complex. The ring opening of substituted cyclopentenes and cyclopentene contained in a bicyclic ring system was also achieved. The ring expansion of bicyclic dienes provided seven-membered dienes contained in the bicyclo[5.2.0]nonane ring system. Details of the structural analysis are also discussed. A mechanistic analysis is provided to account for the data presented herein.

Full text of DOI:10.1016/j.tet.2008.03.027

Tetrahedron 64 (2008) 6909–6919
Contents lists available at ScienceDirect
Tetrahedron
journal homepage: www.elsevier.com/locate/tet
Selective tandem enyne metathesis for the synthesis
of functionalized cycloheptadienesq
*
Steven T. Diver , Daniel A. Clark, Amol A. Kulkarni
Department of Chemistry, University at Buffalo, State University of New York, Buffalo, NY 14260-3000, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The regio- and site-selective ring expansion of dienes and the regioselective ring expansion of
substituted cyclopentenes provide 1,3-cycloheptadienes by enyne metathesis under methylene-free
conditions. Site-selectivity results from differential ring strain among two different cycloalkenes in diene
reactants. The high regioselectivity found in the ring expansion of tetrahydroindene (THI) is explained on
the basis of a selective ring opening by the second generation Grubbs’ ruthenium carbene complex. The
ring opening of substituted cyclopentenes and cyclopentene contained in a bicyclic ring system was also
achieved. The ring expansion of bicyclic dienes provided seven-membered dienes contained in the
bicyclo[5.2.0]nonane ring system. Details of the structural analysis are also discussed. A mechanistic
analysis is provided to account for the data presented herein.
Received 30 January 2008
Received in revised form 5 March 2008
Accepted 7 March 2008
Available online 12 March 2008
Keywords:
Ring synthesis
Ring expansion
Enyne metathesis
Grubbs’ catalyst
1,3-Cycloheptadiene
Regioselectivity
Site-selectivity
Ó 2008 Elsevier Ltd. All rights reserved.
1. Introduction
generation Grubbs’ ruthenium carbene 1 gives products pre-
dominantly with E-selectivity. This process occurs due to the
Metathesis has become a powerful synthetic transformation in
organic synthesis. In particular, the ring-closing metathesis has
been well used in complex molecule synthesis. More recently,
cross-metathesis methods have developed to conjoin two different
alkenes in alkene metathesis or an alkene and an alkyne in enyne
metathesis. Ring synthesis from two unsaturated reactants in an
intermolecular reaction is less common. Ring synthesis between an
alkene and an alkyne is actually an in situ sequence of cross-me-
tathesis followed by ring-closing metathesis occurring in a cascade.
As a result, the ring synthesis is both potentially complex and po-
tentially very powerful. In this report, we extend the ring synthesis
of 1,3-cycloheptadienes to furnish bicyclic ring systems (Scheme 1).
In this process, cycloalkene reactivity determines site-selectivity
and the factors controlling regioselectivity are addressed.
greater alkene reactivity of the metal carbene, and it can equilibrate
the products by a subsequent cross-metathesis reaction. For in-
stance, in enyne metathesis, the equilibration of initially produced
Z-diene products was found in mechanistic studies using the second
generation Grubbs’ carbene complex.1 This study also showed that
Z-dienes were kinetically produced along with the E-isomer and the
former underwent subsequent equilibration to give predominantly
the E-isomer. On the other hand, control of Z-selectivity is not
presently controlled simply through selection of an appropriate
metal carbene complex. This defines a major problem in olefin
metathesis today. One of the major problems in the intermolecular
ring synthesis is also one of stereoselection. To bring together the
ends and achieve a ring-closing metathesis, the intermediate must
have the Z-alkene double bond geometry. Scheme 2 illustrates the
intermediate carbene that can experience ring closure. The Z-vinyl
carbene holds the ruthenium carbene close to the pendant alkene
so that an intramolecular ring-closing metathesis can occur to give
diene 2 (Eq. 1). Since the ring closure is believed to be stereospecific
with respect to alkene geometry, the E-isomer cannot close to give
the cycloheptadiene product and instead proceeds to an
intermolecular reaction manifold (Eq. 2).
There are a number of concerns in achieving an effective cross-
metathesis. In general, cross-metathesis methods suffer from poor
control of alkene stereoselectivity. The selectivity often depends on
the catalyst being used. Usually, the more reactive second
q
The authors wish to congratulate Professor John Hartwig for recognition of his
In fact, the first example of ring synthesis by cross-enyne
metathesis produced two products illustrating no control of ster-
eoselectivity (Eq. 3). In this reaction, 1,5-hexadiene reacted with
1-alkynes to give two products in an approximately 1:1 ratio.2 The
outstanding accomplishments in mechanistic organometallic chemistry and
organic synthesis.
* Corresponding author. Tel.: þ1 716 645 6800x2201; fax: þ1 716 645 6963.
E-mail address: diver@buffalo.edu (S.T. Diver).
0040-4020/$ – see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tet.2008.03.027

6910
S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
6
7
(five-ring only)
R
5
7
Mes N
Cl
N Mes
Cl
Ru CHPh
+
R
6
5
1 (5 mol %)
PCy3
6
8
1 (Grubbs Ru gen-2)
8
5
(six-ring only)
R
Scheme 1. Site-selective ring synthesis from bicyclic diene.
RCM
- L Ru
(1)
(2)
Z
Ph
n
Ph
RuL
n
R
R
R
2
Ph
L Ru
Ph
RuL
Ph
n
n
R
oligomers
RuL
R
n
E
Scheme 2. Stereochemistry of intermediates leading to a successful ring synthesis.
desired 2-substituted-1,3-cyclohexadiene product arises from the
appropriate intermediate vinyl carbene with the Z-stereochemistry,
which has the right geometry for the ring-closing metathesis step.
The E-isomer does not have the right geometry for ring closure and
instead finds a CH2 group with which to react, supplied by the large
excess of 1,5-hexadiene reactant. Remarkably, this early example of
a cyclohexadiene ring synthesis gave good yields of products, albeit
in a 1:1 ratio. The 1:1 ratio shows that a nearly stereorandom
process led to the formation of the intermediate vinyl carbenes. The
rapid participation of the Z-diene in ring-closing metathesis pre-
sumably maintained the kinetic ratio of the intermediate vinyl
carbenes. Ultimately, our attention focused on improving the yield
of the desired 1,3-cyclohexadiene by removing the source of
methylene.
Our efforts toward the ring synthesis of both six- and seven-
membered ring dienes focused on the scenario where no methylene
(CH2) sources were present. On the most simplistic mechanistic
level, the E-intermediate in Scheme 3, which found a CH2 group to
give an unwanted byproduct, would not have this pathway avail-
able. Depriving the carbenes of CH2 sources led us to characterize
the reaction conditions as ‘methylene-free’.3 Early on, this drew an
important contrast between typical cross-enyne metatheses,4
which had exclusively used 1-alkenes in conjunction with either
terminal or internal alkynes (by far terminal alkynes saw the most
reaction development).5 In the earlier examples, there was always
a source of CH2 because the 1-alkene was used and always used in
molar excess. Our aim to eliminate the CH2 source was also some-
what unconventional at the time since an LnRu]CH2 was consid-
ered to be the reactive carbene intermediate, which reacted with
the alkyne (known as a ‘methylidene-first’ mechanism). We hy-
pothesized that in the absence of a CH2 source, the two vinyl car-
benes might equilibrate. The development of the ring synthesis was
largely based on our working model for the reaction mechanism,
which eventually led to more detailed kinetic analysis.6 The absence
of methylene source led to the development of a ring synthesis to
form six-membered rings.3 For the synthesis of cycloheptadienes 2,
the reaction mechanism is believed to involve a ring opening of
cyclopentene, alkyne insertion, equilibration of vinyl carbenes, and
a ring-closing metathesis to furnish the products (Scheme 2, above).
The six-membered ring synthesis has also been shown to be scal-
able, and when conducted on larger scale, the catalyst loading was
decreased.7
The ring expansion by enyne metathesis provides an effective
means for the synthesis of 2-substituted-1,3-cycloheptadienes.8
The cycloheptadiene can be fashioned from the five carbons
available from cyclopentene. This ring synthesis actually starts from
Ts
n-C4H9
N
n-C4H9
n-C4H9
Ru gen-2 (5 mol %)
N
Ts
N
Ts
+
(3)
(9 eq)
undesired
99 % (1:1.3)
L Ru
n
ring-closing metathesis
methylene transfer
Ts
n-C4H9
Ts
n-C4H9
N
N
Ts
n-C4H9
Z
N
E
+
Ru
n
RuLn
CH2
L
L Ru
n
Scheme 3. Ring synthesis (with methylene sources) using 1,5-hexadiene.

S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
6911
an existent ring, so it is a special case of methylene-free enyne
metathesis known as ring expansion. Some examples of
2-substituted 1,3-cycloheptadienes formed in this reaction are
shown in Scheme 4. The ring expansion using a variety of 1-alkynes
worked effectively giving yields that exceeded 50%. A variety of
heteroatom functionality could be carried out on the alkyne
reactant. In addition, substitution was tolerated at the propargylic
center. We noted lower isolated yields using alkynes with poten-
tially coordinating heteroatoms at the homopropargylic position,
though these reactions were not further optimized from the stan-
dard conditions. Fifty percent yields would have been predicted on
the basis of a completely stereorandom process, such as that
depicted in Eq. 3. We supposed that either the E- and Z-vinyl
carbenes could equilibrate or that additional back-biting could be
occurring from an oligomeric structure (or some combination of
the two). These possibilities are discussed in greater detail in the
previous report. We did not previously examine substituted
cyclopentenes though we did report one case of a silicon-contain-
ing ring that underwent expansion to give a silepin.8
cycloalkenes, can one be induced to react selectively over the other
(site-selectivity; Scheme 5, panel b)? If so, what controls the
differences in reactivity between cycloalkenes? Finally, can the
hypothesized equilibration of carbene intermediates under meth-
ylene-free conditions provide good yields of ring-containing
products in systems with additional complexity (rings or additional
alkene functionality)? These questions guided us in our in-
vestigation and have led us to preliminary conclusions in each of
these areas of inquiry.
2. Results and discussion
Given the expectation that different ring sizes might confer
differential reactivity in the ring expansion, we sought a diene that
contained two isolated alkenes in different rings. Tetrahydroindene
(THI) has a cyclopentene and a cyclohexene available for reaction
and it is commercially-available. The two alkenes are distant
enough from each other that little interaction was expected be-
tween the two. With a suitable test system in hand, we investigated
the cycloheptadiene synthesis under standard conditions, which
we had optimized previously, using 5 mol % Grubbs’ catalyst 1 and
fourfold excess of the cyclopentene.
1 (Ru gen-2, 5 mol %)
+
R
Under the standard conditions, a variety of 1-alkynes un-
derwent selective ring expansion. In these runs, we simply used
4 equiv of THI instead of cyclopentene. The results are summarized
in Table 1.
(4)
CH2Cl2
slow addition, 4-12 h
(4 equiv.)
AcO
R
2
Ph
Ts
OTBDPS
As it can be seen from Table 1, the reaction shows a wide sub-
strate scope with terminal alkynes. Propargylic esters as well as
silyl ethers are good substrates for this transformation (entries 1
and 2). In the next three entries, the presence of homopropargylic
oxygen functionality was evaluated. In general, the ester func-
tionality performed well, even at this homopropargylic position
(entry 3). The more Lewis basic benzyl ether in entry 4 gave slightly
lower isolated yields, possibly due to its ability to coordinate to the
intermediate vinyl carbene. Interestingly, the tosyl group can be
handled in the ring synthesis (entry 5) despite its tendency to
undergo E2 elimination to produce a conjugated enyne. Nitrogen in
the alkyne side chain is also tolerated whether it is spaced two or
three carbons away (entries 6 and 7). A more remote naphthoate
ester posed no difficulties, again despite the potential for chelation
at this position (which would form a six-membered chelate with
the proximal vinyl carbene). Finally, this procedure is well suited for
the incorporation of hydrocarbon side chains on the bicyclic triene
(entries 9 and 10) using the appropriate alkyne substrates. Due to
the effectiveness of the reaction, we did not evaluate additional
ruthenium carbenes for this transformation.
OBz
OBn
N
Bu
75%
62%
65%
53%
N(Ts)Boc
OTBDPS
Ph
45%
55%
77%
64%
Scheme 4. Ring expansion of five-membered rings by enyne metathesis (Ref. 8).
With the successful ring expansion to make seven-membered
rings, we were interested in extending the reaction scope further.
We had four questions that guided our efforts to improve the scope
of ring expansion by enyne metathesis. First, do substituted
cyclopentenes give the reaction? Second, in nonsymmetric cyclo-
pentenes, what factors will influence the regioselectivity? For
instance, in substituted cyclopentenes A, the carbene can approach
in two different ways giving rise to regioisomeric alkylidenes B and
C, which would produce a mixture of products (Scheme 5, panel a).
Third, in dienes where there are two different potentially reactive
Additional observations not included in the table also shed light
on the scope and effectiveness of the ring synthesis. In one case, an
a-substituted propargylic ester gave good conversion using the
conditions of Table 1. Unfortunately, the product proved to be an
R
(a) regioselectivity
(b) site-selectivity
Ph
A
A
A
RuL
Ph
n
+
+
L Ru=CHPh
n
top
RuL
n
C
B
A
Ru gen-2
R
R
A
A
bottom
and / or
R
R
R
Scheme 5. Possibilities in more complex systems: regio- and site-selection.

6912
S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
Table 1
Based on this, we suggest that the alkyne insertion represents
Site-selective ring expansion using the diene tetrahydroindene
a ‘high commitment’ step for the catalyst and the resulting vinyl
carbene intermediates do not revert to carbene D under the normal
reaction conditions.
H
H
Ru gen-2 (5 mol %)
CH2Cl2, reflux
slow addition 16 h
R
+
(5)
H
H
H
H
4
R
Ru gen-2 (5 mol %)
CH2Cl2, 40 °C
slow addition, 16 h
3, 4.0 equiv
+
R
H
H
4
R
Entry
Alkyne
Bicyclic triene 4
Yield
3 (4 equiv)
(isolated, %)
Ru gen-2
-Ru gen-2
1
2
4A, R¼CH2OBz
76
55
OBz
OTBDPS
4B, R¼CH2OTBDPS
H
H
H
H
E
OC(O)β-Naphth
Z
RuL
3
4
5
6
4C, R¼(CH2)2OC(O)b-Naphth 58
n
R
alkyne
R
+
OBn
4D, R¼(CH2)2OBn
46
45
64
62
R
RuL
H
insertion
n
Ph
H
Ph
RuL
n
OTs
4E, R¼(CH2)2OTs
Ph
D
Z-E
E-E
N(Ts)Boc
Scheme 6. Proposed mechanism of site- and regioselective ring synthesis.
4F, R¼(CH2)2N(Ts)Boc
4G, R¼(CH2)3N(Ts)Boc
7
8
N(Ts)Boc
The synthesis of small and medium rings by the enyne me-
tathesis-based ring synthesis depends on the strain in the cyclo-
alkene reactant. On one hand, strain is required to open the
cycloalkene in the first place. However, with too much strain, the
cycloalkene will participate in a rapid self-metathesis and lead to
ring-opening metathesis polymerization (ROMP).15 If we were to
use strained alkenes in the ring synthesis, ROMP will become
a competing process consuming the alkene. To minimize this
competing process, we have preferred not to use highly strained
cycloalkenes. For the seven-membered ring synthesis, cyclo-
pentene is used. Cyclopentene is considered to be a ‘low-strain’
cycloalkene (ca. 2–10 kcal/mol range).
Since different cycloalkenes have different ring strain, we
expected that selectivity might be possible among diene reactants
that possess two different cycloalkenes (as depicted in Scheme 1).
Our analysis focused on strain as a determinant for reactivity in the
ring-opening step. The ring opening is a reversible process that
generates a reactive alkylidene species F (Eq. 6). The alkylidene is
born with a coordinated alkene, which is poised to undergo
the reverse reaction, ring-closing metathesis. If there are two dif-
ferent cycloalkenes competing for reaction with the ruthenium
carbene catalyst, the one that is most kinetically accessible (steric
approach) and most strained will result in the highest population
of reactive alkylidene F. Greater cycloalkene ring strain will limit
the back reaction, thereby permitting the alkylidene time to bind
and react with the alkyne. Though the ring opening is not the
sole determinant of an effective overall ring synthesis, it is the
necessary first step. An inefficient reaction could be predicted in
the case of a very stable, unreactive alkene or in the case where the
back reaction is accelerated due to enforced proximity (e.g.,
Thorpe–Ingold effect). In either instance, any small amount of
alkylidene produced in the reaction of Eq. 6 would quickly fall back
to reactant.
4H, R¼(CH2)3OC(O)b-Naphth 59
OC(O) -Naphth
Ph
9
4I, R¼Ph
77
45
n-C6H13
10
4J, R¼n-C6H13
inseparable mixture of diastereomers. Most of the transformations
summarized in Table 1 (entries 2–10) were performed on a rela-
tively small scale (0.2 mmol of the alkyne or lower). On such a scale,
the isolation of the product by silica gel flash column chromatog-
raphy resulted in partial decomposition of the triene products 4.
It was observed that the reaction affords higher yield when run
on a larger scale (1 mmol of the alkyne). One such example is the
use of propargyl benzoate as alkyne (entry 1). When the reaction
was performed on a 0.2 mmol scale (alkyne), the product 4A was
obtained in 56% isolated yield; when the same transformation was
carried out on a 1.0 mmol scale (alkyne), the isolated yield of the
product 4A was found to be 76%. The other entries of Table 1 were
not duplicated on larger scale, but it is anticipated that material
losses will be proportionally less significant as the reactions are
conducted on larger reaction scale.
The syringe pump addition is a somewhat specialized procedure
used to suppress alkyne polymerization. Alkyne oligomerization is
believed to be particularly problematic with terminal alkynes and
more reactive carbene initiators. Slow addition is used in the ring
expansion to keep the alkyne concentration low to minimize
alkyne oligomerization. Alkyne oligomerization can occur by mul-
tiple alkyne insertion into carbene intermediates. In fact, we had
recently identified a process of alkyne oligomerization involving
the ruthenium vinyl carbene intermediates in a phosphine-free
system.9 The slow addition is also meant to allow the E-vinyl car-
bene intermediate to undergo equilibration.
The regiochemistry of alkyne insertion follows the ‘alkylidene-
first’ mechanism of enyne metathesis.6,10–12 We assume that the
cyclopentene reacts first with Grubbs’ benzylidene. Once the ring
opening has occurred, alkyne insertion takes place to give the
isomeric vinyl carbenes E-E and Z-E (Scheme 6). We suggest that
the ring closure arises from the Z-isomer and the E-isomer un-
dergoes partial equilibration under the reaction conditions.3,8
Based on the DFT calculations for enyne metathesis by Lippstreu
and Straub, the alkyne insertion is believed to be irreversible.13 Our
own experiments to effect reversal and crossover in 2-substituted-
1,3-cycloheptadienes failed, which indicated no net reversibility.14
L Ru
Ph
n
(6)
L Ru
+
n
Ph
n
n
1
F
Ring strain data is available from the ROMP literature. ROMP is
similar to the ring synthesis in many ways. First, it involves
a cycloalkene, which must be able to open by the metal carbene
initiator. Second, once the cycloalkene has opened, it needs to
participate in a cross-metathesis with additional cycloalkene rather

S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
6913
OTBS
OAc
AcO
OAc
7.4
7.8
6.8
5.0
3.4
2.9
2.3
ring strain in kcal/mole
ROMP
don't ROMP
Scheme 7. Some ‘low strain’ cycloalkenes (Refs. 16 and 17).
than being engaged in the reverse reaction (ring-closing metathe-
sis). When the ROMP reaction proceeds with another cycloalkene,
the metal carbene telomerizes and the process can repeat, each
time partitioning between forward reaction (ring opening) and
back reaction (ring closing). The ring-opening steps should collec-
tively release some degree of strain to keep ROMP productive. The
stability of products from ROMP and the ring expansion may also
dictate favorable reaction. ROMP produces a polymer, which can be
destabilized by too many substituents in the backbone due to
nonbonding interactions. As a result, monomers that are strained
but highly substituted do not engage in ROMP, not because they do
not open (kinetics), but because they do not form a thermody-
namically stable product. ROMP works best for strained cyclo-
alkenes such as norbornene. For comparison, the ring strain of
norbornene is 27.2 kcal/mol.16 Yet less strained cycloalkenes also
participate in ROMP. What is the cutoff? Grubbs’ group have
studied low strain cycloalkenes and evaluated their ability to un-
dergo ROMP. Grubbs et al. correlated ring strain with polymeriza-
tion ability.17 For example, all the neat cycloalkenes in the box of
Scheme 7 undergo ROMP. The cutoff is in the vicinity of 4–5 kcal/
mol of ring strain.
The data in Scheme 7 show that cyclopentene and cyclohexene
are divided by their ability to undergo ROMP. Focusing on cyclo-
pentene and cyclohexenes, there is ca. 4 kcal/mol difference in their
ring strain (6.8 kcal/mol vs 2.9 kcal/mol). In metathesis, the ring
opening of cyclohexene is difficult and rare.18 The ring strain data
above suggest that there should be a slight preference of the
cyclopentene ring to undergo productive opening as compared to
the cyclohexene ring, under nominal conditions of metathesis
(CH2Cl2, reflux, 1–24 h).
Given that 4-oxysubstituted cyclopentenes give ROMP, we
examined their ability to participate in ring synthesis. Based on the
data above, the 4-substituted cyclopentenes possess ca. 5 kcal/mol
ring strain, which is similar to the amount of ring strain found in
cyclopentene. The substrates 5A19 and 5B20 were prepared using
known procedures. In the event, the substituted cyclopentenes
5A and 5B underwent an efficient ring expansion giving the cor-
responding 1,3-cycloheptadienes 6 in good yield. Presumably
substituents located in the homoallylic position are too remote to
affect the stability of intermediate carbenes or participate in che-
lation. The products of Eq. 7 are interesting for a number of reasons.
First, the alkyne insertion process results in desymmetrization of a
prostereogenic meso-substrate 5. One can imagine that this process
might be controlled through the use of chiral catalysis. Second, the
resulting 1,3-cycloheptadienes can be readily functionalized, e.g.,
by Diels–Alder chemistry, and the rings feature substitution at the
remote 6-position. As a result, the use of functionalized cyclo-
pentene reactants permits access to more highly functionalized
cycloheptadienes with remote substituents that would be difficult
to introduce through conventional methods.
OR
OR
Ru gen-2 (5 mol %)
CH2Cl2, reflux
slow addition, 8 h
(7)
OBz
+
(3.0 equiv)
OBz
5A, R = Bz
6A, R = Bz, 64%
5B, R = TBS
6B, R = TBS, 59%
To further probe the effect of substitution on the cyclopentene,
we next investigated a cyclopentene with a fused ring. In this in-
stance, there is a single alkene reaction site, so there is no issue of
site selection. However, we thought this system was interesting
because of the potential to conduct further transformation on the
diene products. Moreover, this bicyclic ring system presents a dif-
ferent framework that could influence regiochemistry. Because the
regioselectivity had been high in the THI-alkyne ring expansion, we
also expected high regioselectivity in the present case. Bicy-
clo[3.2.0]heptene 8 is commercially available. Applying the condi-
tions in Table 1 to this ring expansion resulted in incomplete
conversions, so the number of alkene equivalents was increased to
eight. Under these conditions, complete consumption of phenyl-
acetylene was found, but two products were formed (Scheme 8).
The cycloheptadiene mixture was obtained as a 1.7:1.0 mixture
of 9A and 9B. Proton NMR spectroscopy was used to assign the
constitution of the major and minor regioisomers that were
formed. In Eq. 8, the major product bears phenyl substitution at the
3-position. The major isomer showed three vinyl protons: (1)
a doublet at d 6.48 (J¼4.5 Hz) for the H2 proton coupled to the
angular methine (H1), which appears as a broad multiplet at
d 3.94–3.87 ppm; (2) a doublet at d 6.16 (Jcis¼11.0 Hz, H4) coupled
to, (3) a ddd appearing at d 6.29 (J¼11.0, 7.0, 7.0 Hz, H5), which
shows an apparent triplet substructure due to 3J coupling with the
allylic methylene at C6. The minor isomer shows two distinct
vinylic resonances at 6.39 (app t, J¼7.0 Hz) and 6.37 (dd, J¼12.0,
4.0 Hz) for H5 and H2, respectively. The combined yields proved
comparable to those observed in the THI examples and in our
previous work. Though the scope of the reaction looked promising
with respect to yield, we wanted to understand the factors
accounting for the reduced regioselectivity.
6
5
O
O
O
4
Ru gen-2 (5 mol %)
Ph
4
+
Ph
+
(8)
CH2Cl2, 40 ºC
slow addition, 4h
3
Ph
1
3
2
7
8 (8 equiv)
9B
9A
(74% combined yield; 1.7 : 1.0 ratio 9A/9B)
Scheme 8. Ring-opening of bicyclo[3.2.0]heptene.

6914
S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
O
O
O
ROM
RCM
R
Ph
Ph
9B (R=Ph),
minor
Ph
R
R
RuL
R
LnRu
RuL
n
n
8 (R = H)
G
12 (R = Me)
O
O
O
ROM
RCM
R
9A (R=Ph),
major
RuLn
Ph
RuL
Ph
n
R
RuL
n
Ph
H
Scheme 9. The reversible formation of carbene intermediates in preequilibrium.
At this point, we considered possible reaction mechanisms to
explain the difference in regioselectivity seen in the two bicyclic
systems. There are two scenarios to explain the differences be-
tween the cyclopentene and THI above. First, it is conceivable that
the carbenes G and H are formed by reversible ring opening and
have different reactivity with alkynes in the next step (Scheme 9).
Although the mechanism of enyne metathesis is not completely
understood,21 the reaction of 1-hexene and 1-alkynes shows a fast
alkyne insertion and a slow turnover of the resultant E- and Z-vinyl
carbenes.6 Since the ring opening of the cyclopentene is believed to
occur before the rate-determining step, it is possible that the ring
opening is a preequilibrium that samples both carbene states G and
H. Once carbenes G and H react with the alkyne, they are com-
mitted to enyne metathesis and are thus fated to form the regio-
isomeric products (e.g., 9B and 9A, above) in a regiospecific manner.
In the preequilibrium state, where the energy barrier to RCM is
lower than the alkyne insertion step, a Curtin–Hammett situation
applies. In this case, the product ratio depends not on the relative
ratio of the two carbenes, but rather depends on the barrier for each
respective carbene to insert the alkyne. The results in Eq. 8 indicate
that the carbene H reacts predominantly with alkyne to account for
63% of product 9A and carbene G accounts for the remaining 37% of
product 9B. Though this analysis appears sound, it does not account
for the marked difference in insertion barriers seen in the THI ex-
ample above. That is to say, if we view the THI results above in
terms of the Curtin–Hammett principle, then the corresponding
carbenes I and D form under rapid preequilibrium conditions but
only D reacts (Scheme 10). Why would carbene D (THI) behave so
much differently than the corresponding carbene H?22 The Curtin–
Hammett analysis seems inconsistent because similarities of car-
bene substitution (e.g., I is like G and D is like H) predict similar
reactivity, which is inconsistent with the different product ratios
observed.
steric bulk. In this picture, the different ratios of products are in-
terpretable due to different steric factors of THI compared to the
bicyclo[3.2.0]heptene. This analysis posits that the orientation
depicted in Eq. 9 is unfavorable due to the steric bulk of the allylic
substituent. As a result, the carbene adds predominantly with the
orientation in Eq. 10 and carbene D forms exclusively. The most
populous carbene then dominates the regiochemical path of re-
action leading to product 4. In the case of bicyclo[3.2.0]heptene
shown in Scheme 9, the substituted allylic positions have less steric
bulk. The endocyclic CH2 group in the four-membered ring presents
less bulk due to its angle constraint in the rigid four-membered
ring. Loss of selectivity may be explained by nonselective opening
of the cyclopentene giving carbenes G and H.
Steric bulk is known to account for selectivity in ring-opening
metathesis. In pioneering work by Snapper et al.,23 it was found
that ring-opening cross-metathesis of strained cyclobutenes was
possible (Eq. 11). In conjunction with the cross-metathesis, they
discovered that the first generation Grubbs’ carbene complex
(Cy3P)2Cl2Ru]CHPh reacted with cyclobutene 10 to give a stable
product 11 that contained a new carbene signal in the proton NMR.
They isolated the new complex and determined its structure of
a carbene ‘caught in the act’ of ring opening. Importantly, due to the
unsymmetrical steric environment, Snapper observed a highly
stereoselective reaction, consistent with a selective approach of the
metal carbene. The observation of carbene 11 supports the notion
that selectivity in ring opening of an unsymmetrical cycloalkene
can be governed by the approach of Grubbs’ complex. It is a note-
worthy feature of this cyclobutenyl system that once ring opening
occurs, there is no reverse reaction, so there are no concerns about
a competing ring-closing metathesis obfuscating a kinetic result
(the cyclobutene ring strain is 30.6 kcal/mol).16 It should also be
noted that Snapper’s study was conducted with the first generation
Grubbs’ carbene, which has the bis(tricyclohexylphosphine) ligand
environment.
ROM
R
Ph
SNAPPER et al., 1997
Ph
N. R. (9)
RCM
LnRu
H
RuL
H
n
R
R
RuCl2PCy3
Ph
RuCl2(PCy3)2
Ph
I
+
(11)
-PCy3
Me
Me
O
O
ROM
RCM
R
10
11
RuL
n
RuL
n
(10)
4
Ph
Ph
D, reactive regioisomer
To probe the theory that less bulk in the bicyclo[3.2.0]heptene
accounts for the loss of regioselectivity, we tested the corollary
hypothesis that greater steric bulk should ‘restore’ selectivity. The
introduction of geminal methyl groups at the a-position increases
the steric bulk of the small endocyclic CH2 group. The larger group
would impose steric strain between the LnRu fragment and the
allylic R groups (8 vs 12, Scheme 9, above). Steric bulk was in-
troduced at the a-position and the analogous bicycloalkene 12 was
used in the ring expansion. The synthesis of the required substrate
Scheme 10. Reversible formation of carbene intermediates with tetrahydroindene.
Another possible explanation is that of stereoselective ring
opening by Grubbs’ carbene complex. The cycloaddition of the
metal carbene is sensitive to the steric environment of the alkene
and will position the ‘bulky’ metal fragment away from the greatest

S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
6915
Me
2
6
Ph
4
Me
1
Me
Ru gen-2 (10 mol %)
CH2Cl2, 40 °C
slow addition, 4 h
Me
Ph
+
(12)
O
5
O
12 (8 equiv)
13 (57%)
PTAD
CH2Cl2, rt, 1 h
O
O
H2
Ph
Ph
N
N
N
N
Ts
N
H
H
H
N
H
N
O
O
Me
Me
Me
Me
1
Ph
4
N(Ts)Boc
Boc
5
7
6
O
O
16 (55%, two steps from 15)
14 (51%)
15
Scheme 11. Regiochemical outcome with a,a-disubstitution on bicyclo[3.2.0]heptene 12.
followed a literature procedure24 (2þ2 cycloaddition of dime-
thylketene and cyclopentadiene). In the ring synthesis under the
conditions above, a good yield of diene 13 was obtained. Signifi-
cantly, the bicyclic product was obtained as a single regioisomer
with none of the minor isomer detected in the crude 1H NMR
spectrum (Scheme 11).
Similar results were obtained using alkyne 15, however, the
obtained diene proved unstable to chromatography on silica gel
and was trapped in situ with PTAD to afford the cycloadduct endo-
16 obtained in 55% yield over two steps. The adduct 16 was also
obtained as a single regioisomer. Though a large excess of the
bicycloalkene 12 was used, most could be recovered unchanged
from the reaction (95% recovery of theoretical 7.0 equiv).
Spectroscopic analysis firmly established regiochemistry. The
proton NMR of the bicyclodiene 13 showed the expected vinylic
doublet at d 6.39 (J¼5.0 Hz), coupled to the angular methine proton at
C1. Similarly, the vinylic proton H4 appears as a doublet at d 6.13
(J¼11.0 Hz)coupledtoadoubletoftriplets atd6.28(dt, J¼11.0, 6.5 Hz),
which also shows 3J allylic coupling to the methylene at C6. The crude
1H NMR spectrum of the diene from 15 showed a single isomer with
characteristic vinylic resonances at d 6.07 (dt, J¼13.0, 8.5 Hz), a broad
singlet at d 5.87, and a doublet at d 5.86 (J¼13 Hz). Similarly, the
cycloadduct 14 showed the contiguous H4–H5–H6–H7–H1–H2 con-
nectivity determined from 1H–1H COSY analysis. The structure of the
adduct14 corroborated thedieneregiochemistry, whichwasassigned
on the basis of coupling in the 1D 1H NMR spectrum.
Figure 1. ORTEP drawing of rac-14. Thermal ellipsoids are drawn at 50% probability.
One molecule of benzene was found in the unit cell. Note that the crystallographic
numbering is not the same as that used in Scheme 11.
We verified the assignment of regiochemistry through single
crystal X-ray structural analysis of the crystalline Diels–Alder
cycloadduct 14. Diene 13 was subjected to Diels–Alder cycloaddi-
tion with N-phenyl-1,3,5-triazolo-2,4-dione (1.1 equiv PTAD) to
provide the cycloadduct 14, mp 133–135 ꢀC, in good yield (Scheme
11, above). X-ray quality crystals were grown from benzene by slow
diffusion of pentanes. The ORTEP drawing of the resulting structure
is depicted in Figure 1. Not surprisingly, the cycloadduct was
formed exclusively by cycloaddition to the convex face of the
bicyclic diene and was produced solely as the endo-diastereomer.
The mechanism to account for the high regioselectivity in these
cases is consistent with a selective ring opening by Grubbs’ carbene
1. The ligand set on ruthenium is positioned away from the allylic
substituent to avoid steric strain (see 18, Eq. 13). The ring opening
accesses one major carbene species, which inserts an alkyne to give
vinyl carbene intermediates, e.g., 19. Progression through RCM
would then give the ring expansion product 13.
In conclusion, the regio- and site-selective ring synthesis of 1,3-
cycloheptadienes has been achieved using differential ring strain as
the basis for chemical reactivity in the ring expansion. The high
regioselectivity found in the ring synthesis derived from the diene
tetrahydroindene (THI) can be explained by a selective ring opening
by Grubbs’ ruthenium carbene complex. The ring opening of
substituted cyclopentenes and cyclopentene contained in a bicyclic
ring system was also achieved. In accordance with the regiose-
lectivity model, bulkier substituents at the allylic position lead to
high selectivity. Future studies are directed to extend the reaction
to other bicycloalkenes, expanding the repertoire of cyclopentenes
Ph
Ph
Me
Me
RuCl2(H2IMes)
Ph
RCM
Me
Me
13
(13)
RuCl2(H2IMes)(PCy3)
Ph
-1
O
O
18
Z-19

6916
S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
and alkynes participating in the reaction and further mechanistic
studies to better understand the role of ring strain in the overall
ring synthesis.
(57 mg, 55% yield) after purification of the crude reaction mix-
ture using silica gel flash column chromatography (elution with
1% ethyl acetate in hexanes). Analytical TLC: Rf 0.44 (10% ethyl
acetate in hexanes). 1H NMR (500 MHz, CDCl3, ppm) d 7.69–7.67
(m, 4H), 7.44–7.35 (m, 6H), 5.80–5.75 (m, 1H), 5.68–5.64 (m, 3H),
4.11 (s, 2H), 2.73–2.69 (m, 1H), 2.43–2.37 (m, 2H), 2.27–2.19 (m,
2H), 2.12–1.97 (m, 3H), 1.06 (s, 9H); 13C NMR (125 MHz, CDCl3,
ppm) d 135.6, 134.3, 133.9, 132.1, 131.1, 129.5, 127.6, 126.6, 124.9,
124.3, 68.9, 38.0, 34.4, 33.9, 31.6, 30.8, 26.8, 19.3; FTIR (thin film,
cmꢂ1) 2858, 1428, 1111, 824, 703, 613. High-resolution MS (EIþ,
3. Experimental
3.1. General information
Reactions were conducted under argon atmosphere unless oth-
erwise noted. Solvents were dried and degassed under argon by
a solvent purification system and drawn immediately prior to use.
Dichloromethane, tetrahydrofuran, and ether were dried by passage
through alumina, and benzene and toluene were dried and
deoxygenated using columns of alumina and Q5. Ruthenium [1,3-
bis-(2,4,6-trimethylphenyl)-2-imidazolidinylidine]dichloro(phenyl-
methylene)(tricyclohexylphosphine) (Grubbs’ second generation
catalyst) was obtained from Materia, Inc. (Pasadena, CA) or pur-
chased from Aldrich Chemical Co. 1,5-Hexadiene was purified by
distillation from sodium metal. All other chemicals were purchased
from Aldrich Chemical Co. and used as-received. Column chroma-
tography was carried out on Merck silica gel 60 (230–400 mesh).
1H NMR spectra were recorded at 300, 400, or 500 MHz and 13C
NMR spectra at either 75 or 125 MHz in the indicated solvent.
1H NMR spectra were referenced on the TMS signal. The 13C NMR
spectra were referenced at 77 ppm for CDCl3. Enantiomeric excesses
m/z) molecular ion calcd for
C28H34O1Si1: 414.2373, found:
414.2364.
3.2.3. 2{(5Z,7Z)-4,4a,9,9a-Tetrahydro-1H-benzo[7]annulen-6-
yl}ethyl 2-napthoate (4C)
The experiment was performed as per the general experimental.
The bicyclic triene 4C was obtained as pale yellow oil (50 mg, 58%
yield) after purification of the crude reaction mixture using silica gel
flash column chromatography (elution with 2% ethyl acetate in
hexanes). Analytical TLC: Rf 0.40 (10% ethyl acetate in hexanes).
1H NMR (500 MHz, CDCl3, ppm) d 8.59 (s,1H), 8.05 (d, J¼8.0 Hz,1H),
7.94 (d, J¼8.0 Hz, 1H), 7.88–7.66 (m, 2H), 7.60–7.52 (m, 2H), 5.85–
5.80 (m,1H), 5.75 (d, J¼12.5 Hz,1H), 5.67–5.51 (m, 3H), 4.42 (m, 2H),
2.68 (s, 1H), 2.53 (m, 2H), 2.32 (m, 2H), 2.24–2.20 (m, 2H), 2.01–1.97
(m, 3H); 13C NMR (125 MHz, CDCl3, ppm) d 166.7,135.5,133.3,132.9,
132.5, 131.9, 131.0, 129.3, 128.1, 128.0, 127.73, 127.68, 127.4, 126.7,
126.5, 125.3, 124.3, 64.6, 39.1, 38.0, 35.2, 34.6, 31.5, 30.6; FTIR (thin
film, cmꢂ1) 2899, 1716, 1283, 1228, 1196, 1097. High-resolution MS
(EIþ, m/z) molecular ion calcd for C24H24O2: 344.1771, found:
344.1769.
were determined by HPLC using
a Chiralcel OD-H column
(4.6 mmꢁ250 mm, 5 mm particle size) using UV detection.
3.2. General procedure for the THIdalkyne tandem
metathesis (Table 1)
An oven-dried 50 mL Schlenk tube equipped with a magnetic
stir bar and a cold finger condenser was charged with 10 mL CH2Cl2
followed by Grubbs’ second generation catalyst (10.5 mg, 5 mol %)
and this solution was immersed into an oil bath maintained at
45 ꢀC. Alkyne (0.25 mmol) and THI (120 mg, 1 mmol) were dis-
solved in 2.0 mL CH2Cl2. This solution was added to the solution of
the catalyst in CH2Cl2, maintained at 45 ꢀC over a period of 16 h by
means of a gas-tight syringe (syringe pump). After the addition was
complete, the reaction mixture was stirred at reflux for an addi-
tional 45 min. At this stage, the reaction mixture was concentrated
to yield dark brown oil. The latter was purified by flash column
chromatography using silica gel. Elution with the indicated eluent
yielded the corresponding bicyclic trienes.
3.2.4. (4a,5Z,7Z,9a)-6-{2-(Benzyloxy)ethyl}-4,4a,9,91-tetrahydro-
1H-benzo[7]annulene (4D)
The experiment was performed as per the general experimental.
The bicyclic triene 4D was obtained as colorless oil (32 mg, 46%
yield) after purification of the crude reaction mixture using silica
gel flash column chromatography (elution with 1% ethyl acetate in
hexanes). Analytical TLC: Rf 0.44 (10% ethyl acetate in hexanes). 1H
NMR (500 MHz, CDCl3, ppm) d 7.35–7.31 (m, 4H), 7.28–7.55 (m, 1H),
5.77–5.73 (m, 1H), 5.66 (dd, J¼12.0, 1.0 Hz, 1H), 5.62–5.60 (m, 1H),
5.47 (d, J¼5.5 Hz, 1H), 4.51 (s, 2H), 3.53 (t, J¼7.5 Hz, 2H), 2.67–2.63
(m, 1H), 2.36 (t, J¼7.5 Hz, 2H), 2.30 (t, J¼5.5 Hz, 2H), 2.24–2.19 (m,
2H), 2.09–2.04 (m, 1H), 2.00–1.96 (m, 2H); 13C NMR (125 MHz,
CDCl3, ppm) d 138.6, 133.4, 132.6, 132.3, 128.3, 127.8, 127.6, 127.4,
126.6, 124.3, 72.8, 70.4, 40.1, 38.0, 35.0, 34.4, 31.3, 30.8; FTIR (thin
film, cmꢂ1) 3021, 2898, 1451, 1361, 1101, 736. High-resolution MS
(ESIþ, m/z) molecular ion calcd for C20H24O1Na1: 303.1719, found:
303.1724 (MþNa)þ.
3.2.1. {(5E,7Z)-4,4a,9,9a-Tetrahydro-1H-benzo[7]annulen-6-
yl}methyl benzoate (4A)
The experiment was performed as per the general experimental.
The bicyclic triene 4A was obtained as pale yellow oil (53 mg, 76%
yield) after purification of the crude reaction mixture using silica
gel flash column chromatography (elution with 2% ethyl acetate in
hexanes). Analytical TLC: Rf 0.41 (10% ethyl acetate in hexanes). 1H
NMR (500 MHz, CDCl3, ppm) d 8.06–8.04 (m, 2H), 7.55 (t, J¼8.0 Hz,
1H), 7.43 (t, J¼8.0 Hz, 2H), 5.88–5.83 (m, 1H), 5.82–5.80 (m, 2H),
5.66–5.62 (m,1H), 5.57–5.54 (m,1H), 4.74 (m, 2H), 2.76 (d, J¼3.5 Hz,
1H), 2.47–2.40 (m, 1H), 2.37–2.23 (m, 3H), 2.15–2.09 (m, 1H), 2.04–
2.00 (m, 2H); 13C NMR (125 MHz, CDCl3, ppm) d 166.5, 136.4, 133.4,
132.8, 131.0, 130.4, 129.6, 128.3, 126.5, 124.9, 124.0, 70.9, 38.3, 34.3,
33.7, 31.0, 30.8; FTIR (thin film, cmꢂ1) 2939, 1719, 1237, 1097. High-
resolution MS (EIþ, m/z) molecular ion calcd for C19H20O2:
280.1458, found: 280.1466.
3.2.5. 2-{(4a,5Z,7Z,9a)-4,4a,9,9a-Tetrahydro-1H-benzo[7]annulen-
6-yl}ethyl-4-methylbenzenesulfonate (4E)
The experiment was performed as per the general experimental.
The bicyclic triene 4E was obtained as pale yellow oil (39 mg, 45%
yield) after purification of the crude reaction mixture using silica
gel flash column chromatography (elution with 10% ethyl acetate in
hexanes). Analytical TLC: Rf 0.37 (30% ethyl acetate in hexanes). 1H
NMR (500 MHz, CDCl3, ppm) d 7.78 (d, J¼8.0 Hz, 2H), 7.33 (d,
J¼8.0 Hz, 2H), 5.75–5.71 (m, 1H), 5.62–5.58 (m, 1H), 5.55–5.49 (m,
2H), 5.40 (d, J¼5.5 Hz, 1H), 4.05 (t, J¼7.0 Hz, 2H), 2.62–2.58 (m, 1H),
2.45 (s, 3H), 2.35 (t, J¼7.0 Hz, 2H), 2.27 (t, J¼5.5 Hz, 2H), 2.21–2.14
(m, 2H), 2.07–2.02 (m, 1H), 1.96–1.88 (m, 2H); 13C NMR (125 MHz,
CDCl3, ppm) d 144.6, 134.9, 133.3, 133.2, 130.3, 129.8, 127.9, 126.8,
126.5, 124.6, 70.0, 39.0, 38.0, 34.8, 34.3, 31.1, 30.7, 21.6; FTIR (thin
film, cmꢂ1) 2900, 1599, 1361, 1178, 1098, 963, 817, 773. High-reso-
lution MS (ESIþ, m/z) molecular ion calcd for C20H24O3S1Na1:
367.1338, found: 367.1339 (MþNa)þ.
3.2.2. tert-Butyldiphenyl{[(5E,7Z)-4,4a,9,9a-tetrahydro-1H-
benzo[7]annulen-6-yl]methoxy}silane (4B)
The experiment was performed as per the general experi-
mental. The bicyclic triene 4B was obtained as colorless oil

S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
6917
3.2.6. tert-Butyl 2-{(4a,5Z,7Z,9a)-4,4a,9,9a-tetrahydro-1H-
benzo[7]annulen-6-yl}ethyl(tosyl)carbamate (4F)
5.69–5.66 (m, 1H) 5.62–5.60 (m, 1H), 2.81–2.78 (m, 1H), 2.43–2.29
(m, 3H), 2.25–2.19 (m, 1H), 2.14–2.03 (m, 3H); 13C NMR (125 MHz,
CDCl3, ppm) d 143.8, 137.9, 134.9, 133.7, 128.1, 128.06, 127.0, 126.6,
125.5, 124.6, 38.5, 37.9, 34.7, 31.2, 30.8; FTIR (thin film, cmꢂ1) 3021,
2898, 1600, 1493, 1443, 937. High-resolution MS (EIþ, m/z) molec-
ular ion calcd for C17H18: 222.1403, found: 222.1406.
The experiment was performed as per the general experimental.
The bicyclic triene 4F was obtained as pale yellow oil (70 mg, 64%
yield) after purification of the crude reaction mixture using silica
gel flash column chromatography (gradient elution with 5% ethyl
acetate in hexanes, followed by 10% ethyl acetate in hexanes).
Analytical TLC: Rf 0.34 (30% ethyl acetate in hexanes). 1H NMR
(500 MHz, CDCl3, ppm) d 7.79 (d, J¼8.0 Hz, 2H), 7.29 (d, J¼8.0 Hz,
2H), 5.82–5.74 (m, 2H), 5.63–5.60 (m, 1H), 5.56–5.52 (m, 2H), 3.87–
3.83 (m, 2H), 2.69–2.65 (m, 1H), 2.47–2.45 (m, 2H), 2.43 (s, 3H),
2.34–2.30 (m, 2H), 2.27–2.19 (m, 2H), 2.11–2.06 (m, 1H), 2.01–1.94
(m, 2H), 1.35 (s, 9H); 13C NMR (125 MHz, CDCl3, ppm) d 150.8, 144.0,
137.5, 134.1, 132.7, 132.2, 129.1, 127.8, 127.4, 126.4, 124.2, 84.0, 47.3,
3.2.10. (4a,5Z,7Z,9a)-6-Hexyl-4,4a,9,9a-tetrahydro-1H-
benzo[7]annulene (4J)
The experiment was performed as per the general experimental.
The bicyclic triene 4J was obtained as pale yellow oil (26 mg, 45%
yield) after purification of the crude reaction mixture using silica
gel flash column chromatography (elution with 1% ethyl acetate in
hexanes). Analytical TLC: Rf 0.52 (5% ethyl acetate in hexanes). 1H
NMR (500 MHz, CDCl3, ppm) d 5.78–5.72 (m, 1H), 5.65–6.61 (m,
2H), 5.57–5.54 (m, 1H), 5.37 (d, J¼5.5 Hz, 1H), 2.66–2.62 (m, 1H),
2.30 (t, J¼5.5 Hz, 2H), 2.24–2.18 (m, 2H), 2.05–1.95 (m, 5H), 1.38–
1.26 (m, 8H), 0.88 (t, J¼7.0 Hz, 3H); 13C NMR (125 MHz, CDCl3, ppm)
d 136.2, 131.8, 131.4, 128.2, 126.6, 124.5, 39.9, 37.9, 35.2, 34.5, 31.8,
31.6, 30.8, 29.5, 28.8, 22.7, 14.1; FTIR (thin film, cmꢂ1) 3020, 2956,
2855, 1652, 1440, 653. High-resolution MS (EIþ, m/z) molecular ion
calcd for C17H26: 230.2029, found: 230.2030.
40.5, 38.1, 34.5, 34.3, 31.1, 30.9, 27.8, 21.5; FTIR (thin film, cmꢂ1
2901, 1727, 1357, 1289, 1158, 673. High-resolution MS (ESIþ, m/z)
)
molecular ion calcd for
C25H33O4N1S1Na1: 466.2023, found:
466.2017 (MþNa)þ.
3.2.7. tert-Butyl 3-{(4a,5Z,7Z,9a)-4,4a,9,9a-tetrahydro-1H-
benzo[7]annulen-6-yl}propyl(tosyl)carbamate (4G)
The experiment was performed as per the general experimental.
The bicyclic triene 4G was obtained as pale yellow oil (86 mg, 62%
yield) after purification of the crude reaction mixture using silica
gel flash column chromatography (gradient elution with 5% ethyl
acetate in hexanes followed by 10% ethyl acetate in hexanes).
Analytical TLC: Rf 0.34 (30% ethyl acetate in hexanes). 1H NMR
(500 MHz, CDCl3, ppm) d 7.77 (d, J¼8.0 Hz, 2H), 7.29 (d, J¼8.0 Hz,
2H), 5.80–5.76 (m, 1H), 5.67–5.62 (m, 2H), 5.58–5.55 (m, 1H), 5.43
(d, J¼5.0 Hz, 1H), 3.81–3.77 (m, 2H), 2.67–2.63 (m, 1H), 2.44 (s, 3H),
2.32 (t, J¼5.0 Hz, 2H), 2.26–2.20 (m, 2H), 2.11–2.05 (m, 2H), 2.03–
1.96 (m, 3H), 1.89–1.83 (m, 2H), 1.33 (s, 9H); 13C NMR (125 MHz,
CDCl3, ppm) d 150.9, 143.9, 137.5, 134.7, 132.4, 132.0, 129.2, 127.8,
127.7, 126.7, 124.4, 83.9, 46.9, 37.9, 37.1, 35.0, 34.4, 31.4, 30.8, 29.8,
27.9, 21.6; FTIR (thin film, cmꢂ1) 2904, 1727, 1358, 1287, 1158, 1090,
674. High-resolution MS (ESIþ, m/z) molecular ion calcd for
3.3. General procedure for the cycloheptadiene synthesis
using 4-substituted cyclopentene derivatives
Cyclopentenes 5A19 and 5B20 were synthesized as reported in
the literature.
Into an oven-dried 50 mL Schlenk tube equipped with magnetic
stir bar and a cold finger condenser was added dichloromethane
(5 mL) followed by Grubbs’ second generation catalyst (4.5 mg,
5 mol %) and the solution was heated to reflux. Propargyl benzoate
(16 mg, 0.1 mmol, 1.0 equiv) and the requisite cycloalkene
(0.3 mmol, 3.0 equiv) were dissolved in 2.0 mL of CH2Cl2 and this
solution was then added to the catalyst solution in CH2Cl2 at 45 ꢀC
over a period of 8 h by means of a gas-tight syringe (syringe pump).
After the addition was complete, the reaction mixture was stirred in
the oil bath for additional 45 min. The solvent was removed in
vacuo. The crude oil thus obtained was further purified by flash
chromatography using silica gel (using the indicated eluent) to
provide the corresponding dienes as colorless oils.
C
26H35O4N1S1: 480.2172, found: 480.2175 (MþNa)þ.
3.2.8. 3-{(5Z,7Z)-4,4a,9,9a-Tetrahydro-1H-benzo[7]annulen-6-
yl}propyl 2-napthoate (4H)
The experiment was performed as per the general experimental.
The bicyclic triene 4H was obtained as pale yellow oil (53 mg, 59%
yield) after purification of the crude reaction mixture using silica
gel flash column chromatography (elution with 2% ethyl acetate in
hexanes). Analytical TLC: Rf 0.40 (10% ethyl acetate in hexanes). 1H
NMR (500 MHz, CDCl3, ppm) d 8.61 (s, 1H), 8.07 (dd, J¼8.0, 1.5 Hz,
1H), 7.96 (d, J¼8.0 Hz, 1H), 7.88 (d, J¼8.0 Hz, 2H), 7.61–7.53 (m,
2H)5.82–5.78 (m, 1H), 5.69 (dd, J¼12.0, 1.5 Hz, 1H), 5.64–5.61 (m,
1H), 5.56–5.52 (m, 1H), 5.47 (d, J¼5.0 Hz, 1H), 4.39–4.34 (m, 2H),
2.67–2.63 (m, 1H), 2.33–2.31 (m, 2H), 2.25–2.19 (m, 3H), 2.11–1.89
(m, 5H); 13C NMR (125 MHz, CDCl3, ppm) d 166.8, 135.5, 134.6,
132.6, 132.5, 132.4, 130.9, 129.3, 128.1, 120.1, 127.74, 127.72, 127.6,
126.58, 126.56, 125.3, 124.4, 64.5, 38.0, 36.2, 35.0, 34.5, 31.4, 30.9,
28.4; FTIR (thin film, cmꢂ1) 2899, 1717, 1283, 1228, 1196, 1098, 779.
High-resolution MS (EIþ, m/z) molecular ion calcd for C25H26O2:
358.1927, found: 358.1927.
3.3.1. Cycloheptadiene-bis-benzoate (6A)
The experiment was performed as per the general experimental.
The cycloheptadiene derivative 6A was obtained as colorless oil
(23 mg, 66% yield) after purification of the reaction mixture
through silica gel column chromatography (elution with 5% ethyl
acetate in hexanes). Analytical TLC: Rf 0.27 (25% ethyl acetate in
hexanes). 1H NMR (500 MHz, CDCl3, ppm) d 8.07–8.02 (m, 4H),
7.58–7.53 (m, 2H), 7.46–7.40 (m, 4H), 6.04–5.99 (m, 2H), 5.96–5.92
(m, 1H), 5.43 (septet, J¼4.0 Hz, 1H), 4.82 (s, 2H), 2.76–2.59 (m, 4H);
13C NMR (125 MHz, CDCl3, ppm) d 166.4, 165.7, 134.3, 133.0, 132.9,
130.5, 130.2, 129.7, 129.6, 129.5, 128.4, 12.3, 127.4, 127.1, 74.7, 69.2,
35.9, 34.6; FTIR (thin film, cmꢂ1) 2953, 1715, 1450, 1273, 1111, 709.
High-resolution MS (EIþ, m/z) molecular ion calcd for C22H20O4Na1:
371.1257, found: 371.1263 (MþNa)þ.
3.2.9. (5E,7Z)-6-Phenyl-4,4a,9,9a-tetrahydro-1H-benzo[7]-
annulene (4I)
3.3.2. Cycloheptadiene TBS-OBz (6B)
The experiment was performed as per the general experimental.
The cycloheptadiene derivative 6B was obtained as colorless oil
(23 mg, 66% yield) after purification of the reaction mixture
through silica gel column chromatography (elution with 5% ethyl
acetate in hexanes). Analytical TLC: Rf 0.27 (15% ethyl acetate in
hexanes). 1H NMR (500 MHz, CDCl3, ppm) d 8.05 (d, J¼7.5 Hz, 2H),
7.56 (t, J¼7.5 Hz, 1H), 7.44 (t, J¼7.5 Hz, 2H), 5.92–5.89 (m, 2H),
5.86–5.82 (m,1H), 4.76 (s, 2H), 4.11 (septet, J¼4.0 Hz,1H), 2.52–2.36
The experiment was performed as per the general experimental.
The bicyclic triene 4I was obtained as pale yellow oil (43 mg, 77%
yield) after purification of the crude reaction mixture using silica
gel flash column chromatography (elution with 1% ethyl acetate in
hexanes). Analytical TLC: Rf 0.48 (5% ethyl acetate in hexanes). 1H
NMR (500 MHz, CDCl3, ppm) d 7.35–7.28 (m, 4H), 7.23–7.21 (m, 1H),
6.05 (d, J¼9.0 Hz, 1H), 6.05–6.00 (m, 1H), 5.94 (d, J¼5.5 Hz, 1H),

6918
S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
(m, 4H), 0.89 (s, 9H), 0.07 (s, 6H); 13C NMR (125 MHz, CDCl3, ppm)
d 166.4, 133.4, 132.9, 130.3, 130.2, 129.6, 128.3, 127.9, 126.6, 72.7,
69.5, 40.5, 39.3, 25.8, 18.2, ꢂ4.8; FTIR (thin film, cmꢂ1) 2954, 1222,
1462, 1272, 1069, 836. High-resolution MS (EIþ, m/z) molecular ion
calcd for C21H30O3Si1Na1: 381.1858, found: 381.1864 (MþNa)þ.
3.5.2. PTAD cycloadduct 14
To a 25 mL round bottom flask under argon gas were added
diene 13 (0.16 mmol) and dichloromethane (10 mL). PTAD
(0.17 mmol) was added portionwise over a 15 min period. When
a red color persisted, the reaction mixture was concentrated in
vacuo (rotatory evaporator) and the crude residue loaded directly
onto a silica gel column (100ꢁ600) purified by elution with 50% ethyl
acetate/petroleum ether to give 34 mg 14 (51%) as a white solid, mp
133–135 ꢀC. The cycloadduct was dissolved in a small portion of
benzene (0.25 mL) and crystallized in an atmosphere saturated
with pentane vapor at room temperature to afford crystals suitable
for X-ray structural analysis. Analytical TLC (50% ethyl acetate/
petroleum ether) Rf 0.25. 1H NMR (500 MHz, CDCl3) d 7.40–7.32 (m,
5H), 7.25–7.19 (m, 5H), 6.26 (d, J¼7 Hz, 1H), 5,54 (br s, 1H), 5.02 (dt,
J¼7.5, 7 Hz, 1H), 3.78 (t, J¼10 Hz, 1H), 2.78 (dd, J¼10.5, 3 Hz, 1H),
2.49 (dd, J¼15, 3 Hz,1H), 2.19–2.13 (m, 1H),1.27 (s, 3H), 0.50 (s, 3H);
13C NMR (75 MHz, CDCl3) d 211.8, 152.7, 151.4, 140.5, 137.7, 131.5,
129.2 (CH), 129.1 (CH), 128.6 (CH),128.1 (CH),127.8 (CH), 125.4 (CH),
125.1 (CH), 59.3, 54.6 (CH), 52.2 (CH), 51.2 (CH), 44.3 (CH), 29.1
(CH3), 27.7 (CH2),17.2 (CH3); FTIR (KBr, cmꢂ1) 2927,1775,1709,1512,
1420. High resolution ESI molecular ion calcd for C25H23O3N3:
413.1734, found: 413.1740 (C25H23O3N3).
3.4. Synthesis of bicyclic diene 9
3.4.1. (2E,4Z)-3-Phenylbicyclo[5.2.0]nona-2,4-dien-8-one (9A) and
(2Z,4E)-4-phenylbicyclo[5.2.0]nona-2,4-dien-8-one (9B)
To a 50 mL Schlenk tube fitted with a cold-finger condenser and
magnetic stirrer under argon gas were added 16 mg 1 (Grubbs’ Ru
gen-2, 0.02 mmol, 10 mol %), bicycle 8 (1.6 mmol, 8 equiv), and
benzene (2.5 mL). The reaction mixture was warmed to 60 ꢀC in
a pre-heated oil bath. Phenylacetylene (0.2 mmol) was dissolved in
benzene (2 mL) and the solution was loaded into a 5 mL gas tight
syringe. The alkyne solution was added via syringe pump over a 4 h
period to the Schlenk tube above. After complete addition, the
reaction mixture was allowed to stir at 60 ꢀC for 1 h before being
quenched with 1 mL of a methanolic solution of KO2CCH2NC25
(0.1 mmol, 5 equiv based on Grubbs’ catalyst). The resulting yellow
solution was filtered through a plug of silica gel, concentrated in
vacuo (rotatory evaporator), and purified by silica gel column
chromatography (100ꢁ700) using a gradient elution (100 mL each
petroleum ether, 2%, 8% ethyl acetate/petroleum ether), which
provided 31 mg of 9 (75%) as a clear liquid, obtained as an in-
separable 1.7:1 mixture of regioisomers. Analytical TLC (20% ethyl
acetate/petroleum ether) Rf 0.45. 1H NMR (500 MHz, CDCl3) d 7.32–
7.25 (m, 6.76H), 6.47 (d, J¼4.5 Hz, 1.42H), 6.37 (ddt, J¼12, 7, 4.5 Hz,
1H), 6.29 (ddt, J¼10.5, 7, 3 Hz, 1.54H), 6.16 (d, J¼10.5 Hz, 1.61H),
3.94–3.87 (m, 1.61H), 3.38–3.29 (m, 1.87H), 3.24–3.18 (m, 1.69H),
3.13–3.05 (m, 1.98H), 2.51–2.34 (m, 3.37H); 13C NMR (75 MHz,
CDCl3) d 210.9, 141.7, 138.5, 136.8, 133.2, 132.4, 130.8, 128.8,
128.5, 128.3, 128.2, 127.4, 126.4, 126.1, 73.5, 73.0, 54.3, 54.2, 28.0,
27.9, 26.1, 26.3; FTIR (thin film, cmꢂ1) 1786. High resolution ESI
3.5.3. PTAD cycloadduct 16
To a 50 mL Schlenk tube fitted with a cold-finger condenser and
magnetic stirrer under argon gas were added 16 mg 1 (Grubbs’ Ru
Table 2
Crystal data and data collection information for 14
Crystal data
Identification code
Moiety formula
Sum formula
Formula weight
PTAD cycloadduct 14
25H23N3O3, C6H6
C31H29N3O3
C
491.57
Crystal system, space group
Unit cell dimensions
Monoclinic, C2/C
a¼17.5848(9) [Å]
b¼8.7847(5) [Å]
c¼31.6268(14) [Å]
a¼90 [ꢀ]
molecular ion calcd for
(C15H14O).
C15H14O: 210.1045, found: 210.1050
b¼94.154(2) [ꢀ]
g¼90[ꢀ]
Volume
Z, calculated density
Radiation
Wavelength
q Range for data collection
sin q/lmax
Absorption coefficient
Temperature
F(000)
Crystal size
Crystal shape
Crystal color
4872.8(4) [Åꢂ3
]
3.5. Regioselective ring expansions
8, 1.340 [Mg/m3]
Mo Ka
3.5.1. (2E,4Z)-9,9-Dimethyl-3-phenylbicyclo[5.2.0]nona-2,4-dien-
8-one (13)
l¼0.71073 [Å]
1.29–27.12 [ꢀ]
0.64 [Åꢂ1
]
To a 50 mL Schlenk tube fitted with a cold-finger condenser and
magnetic stirrer under argon gas were added 16 mg 1 (Grubbs’ Ru
gen-2, 0.02 mmol, 10 mol %) and benzene (2 mL). The reaction
mixture was warmed to 60 ꢀC in a pre-heated oil bath. Phenyl-
acetylene (0.2 mmol) and bicycle 12 (1.6 mmol, 8 equiv) were dis-
solved in benzene (2 mL) and loaded into a 5 mL gas tight syringe.
This solution was added via syringe pump over a 4 h period to the
Schlenk tube above. After complete addition, the reaction mixture
was allowed to stir at 60 ꢀC for further 30 min before being
quenched with 1 mL of a methanolic solution of KO2CCH2NC
(0.1 mmol, 5 equiv based on Grubbs’ catalyst). The resulting yellow
solution was filtered through a plug of silica gel, concentrated in
vacuo (rotatory evaporator), and purified by chromatography on
silica gel (100ꢁ700) using gradient elution (100 mL each 2%, 4%, 10%
ethyl acetate/petroleum ether), which gave 27 mg of 13 (57%) as
a clear liquid. Analytical TLC (20% ethyl acetate/petroleum ether) Rf
0.60. 1H NMR (500 MHz, CDCl3) d 7.38–7.31 (m, 5H), 6.39 (d, J¼5 Hz,
1H), 6.28 (dt, J¼11, 6.5 Hz 1H), 6.13 (d, J¼11 Hz, 1H), 4.03 (dt, J¼11,
5 Hz 1H), 3.03 (dd, J¼11, 3.5 Hz, 1H), 2.44–2.37 (m, 2H), 1.21 (s, 3H),
1.12 (s, 3H); 13C NMR (75 MHz, CDCl3) d 217.8, 142.5, 138.9, 132.8,
130.7, 129.7, 128.3, 127.3, 126.5, 126.1, 68.4, 68.9, 41.5, 26.5, 24.8,
18.4; FTIR (thin film, cmꢂ1) 1775. High resolution ESI molecular ion
calcd for C17H18O: 238.1352, found: 238.1357 (C17H18O).
m¼0.087 [mmꢂ1
]
90(1) [K]
2080
0.20ꢁ0.15ꢁ0.10 [mm]
Prism
Colorless
Data Collection
Diffractometer
APEXII CCD area detector
diffractometer
6-Scan
Multi-scan (SADABS 2004/1;
Sheldrick, 2004)
Tmax¼0.9913, Tmin¼0.9828
34,770/5403/0.0261
100.0 [%]
Measurement method
Absorption correction
Max and min transmissions
Reflections collected/unique/Rint
Completeness to q¼25.68ꢀ
Limiting indices
h¼ꢂ22/22
k¼ꢂ11/11
l¼ꢂ40/40
Refinement
Refinement method
Data/restraints/parameters
R [I>2s(I)]
Full-matrix least-squares on F2
5403/0/336
R1¼0.0346/wR2¼0.0822
R1¼0.0416/wR2¼0.0868
1.025
R (all data)
Goodness-of-fit on F2¼S
Drmax
0.337 [e Åꢂ3
]
Drmin
ꢂ0.177 [e Åꢂ3
]

S.T. Diver et al. / Tetrahedron 64 (2008) 6909–6919
6919
gen-2, 0.02 mmol, 10 mol %) and benzene (2 mL). The reaction
Acknowledgements
mixture was warmed to 60 ꢀC in a pre-heated oil bath. Alkyne
(0.2 mmol) and bicycle 12 (1.6 mmol, 8 equiv) were dissolved in
benzene (2 mL) and loaded into a 5 mL gas tight syringe. This so-
lution was added via syringe pump over a 4 h period to the above
Schlenk tube. After complete addition, the reaction mixture was
allowed to stir at 60 ꢀC for 1 h before being quenched with 1 mL of
an ethanolic solution of isocyanide KO2CCH2NC (0.1 mmol, 5 equiv
based on Grubbs’ catalyst). The resulting yellow solution was
filtered through a plug of silica gel and concentrated in vacuo (ro-
tatory evaporator). The crude product was dissolved in dichloro-
methane (10 mL) and PTAD (0.2 mmol) was added portionwise
over a 15 min period at room temperature until a red color per-
sisted. The reaction mixture was subsequently concentrated in
vacuo (rotatory evaporator), and loaded directly onto a silica gel
column (100ꢁ600) and purified by elution with 50% ethyl acetate/
petroleum ether to give 70 mg 16 (55% over two steps). Analytical
TLC (50% ethyl acetate/petroleum ether) Rf 0.26. 1H NMR (500 MHz,
CDCl3) d 7.69 (d, J¼8.5 Hz, 2H), 7.49 (d, J¼8.5 Hz, 2H), 7.39 (t,
J¼7.5 Hz, 2H), 7.31 (t, J¼7 Hz, 1H), 7.25 (d, J¼7 Hz, 2H), 6.20 (d,
7.5 Hz, 1H), 5.08 (br s, 1H), 4.97–4.95 (m, 1H), 3.87–3.82 (m, 2H),
3.67 (dt, J¼14.5, 15 Hz, 1H), 2.48 (dd, J¼10.5, 3 Hz, 1H), 2.74–2.68
(m, 1H), 2.64–2.59 (m, 1H), 2.47 (dd, J¼15, 3 Hz, 1H), 2.40 (s, 3H),
2.18–2.12 (m, 1H), 1.44 (s, 3H), 1.27 (s, 9H), 1.23 (s, 3H); 13C NMR
(75 MHz, CDCl3) d 212.3, 151.2, 150.7, 150.0, 144.2, 137.0, 131.8, 129.2,
129.1, 128.9, 127.9, 127.7, 125.6, 84.6, 58.9, 54.4, 51.4, 51.1, 46.0, 44.7,
36.0, 28.8, 27.8, 27.1, 21.5, 18.2; FTIR (thin film, cmꢂ1) 2976, 2939,
1778,1716,1422,1360,1160. High resolution ESI molecular ion calcd
for C33H38O7N4SNa: 657.2353, found: 657.2372.
This work was supported by the Petroleum Research Fund (PRF
AC-44202) and the NSF (CHE-601206). D.A.C. acknowledges
Wyeth-Ayerst for a graduate stipend during the summer of 2006.
We thank Dr. Mateusz Pitak for determining the X-ray crystal
structure of 14, Materia for their generous catalyst support, and
Dr. Andrew Bell (Promerus, Cleveland, OH) for a gift of tetrahy-
droindene (THI).
Supplementary data
Supplementary data associated with this article can be found in
the online version, at doi:10.1016/j.tet.2008.03.027.
References and notes
1. Giessert, A. J.; Diver, S. T. J. Org. Chem. 2005, 70, 1046–1049.
2. Smulik, J. A.; Diver, S. T. Tetrahedron Lett. 2001, 42, 171–174.
3. Kulkarni, A. A.; Diver, S. T. J. Am. Chem. Soc. 2004, 126, 8110–8111.
4. Stragies, R.; Schuster, M.; Blechert, S. Angew. Chem., Int. Ed. 1997, 36, 2518–2520.
5. Diver, S. T.; Giessert, A. J. Chem. Rev. 2004, 104, 1317–1382.
6. Galan, B. R.; Giessert, A. J.; Keister, J. B.; Diver, S. T. J. Am. Chem. Soc. 2005, 127,
5762–5763.
7. Kulkarni, A. A.; Diver, S. T. Org. Synth. 2006, 83, 200–208.
8. Kulkarni, A. A.; Diver, S. T. Org. Lett. 2003, 5, 3463–3466.
9. Diver, S. T.; Kulkarni, A. A.; Clark, D. A.; Peppers, B. P. J. Am. Chem. Soc. 2007, 129,
5832–5833.
10. Schramm, M. P.; Reddy, D. S.; Kozmin, S. A. Angew. Chem., Int. Ed. 2001, 40,
4274–4277.
11. Hoye, T. R.; Donaldson, S. M.; Vos, T. J. Org. Lett. 1999, 1, 277–279.
12. Giessert, A. J.; Diver, S. T. Org. Lett. 2005, 7, 351–354.
13. Lippstreu, J. J.; Straub, B. F. J. Am. Chem. Soc. 2005, 127, 7444–7457.
14. Kulkarni, A. A. Ph.D. Dissertation, SUNY Buffalo, 2006.
15. Critical monomer concentration is another variable. As small molecule syn-
thetic chemists, we make the generalization that lower strain alkenes can be
employed in the ring synthesis at higher concentrations without triggering
self-ROMP. On the other hand, with strained cycloalkenes, a much lower alkene
concentration range must be observed otherwise self-ROMP will become
competitive.
16. Schleyer, P. v. R.; Williams, J. E., Jr.; Blanchard, K. R. J. Am. Chem. Soc. 1970, 92,
2377–2386.
17. Hejl, A.; Scherman, O. A.; Grubbs, R. H. Macromolecules 2005, 38, 7214–7218.
18. Randl, S.; Connon, S. J.; Blechert, S. Chem. Commun. (Cambridge, U.K.) 2001,
1796–1797.
19. Chatterjee, A. K.; Sanders, D. P.; Grubbs, R. H. Org. Lett. 2002, 4, 1939–1942.
20. Fu, G. C.; Grubbs, R. H. J. Am. Chem. Soc. 1993, 115, 3800–3801.
21. The rate law for cycloalkene–alkyne cross-metathesis has not yet been de-
termined. Since cyclopentene is nominally reactive, we liken it to 1-hexene and
make the tentative comparison tothe mechanism determined by kinetic analysis.
22. Another possibility is that of diminished reactivity for carbene H due to chelation
by the ketone. We considered this unlikely given the results below where the
reaction proceeds exclusively through the gem-dimethyl version of carbene H.
23. Tallarico, J. A.; Bonitatebus, P. J., Jr.; Snapper, M. L. J. Am. Chem. Soc. 1997, 119,
7157–7158.
3.6. Crystal structural analysis
X-ray diffraction data on 14 were collected at 90(1) K using
a Bruker SMART APEX2 CCD diffractometer installed at a rotating
anode source (Mo Ka radiation, l¼0.71073 Å), and equipped with
an Oxford Cryosystems nitrogen gas-flow apparatus. The data were
collected by the rotation method with 0.3ꢀ frame-width (u scan)
and 20 s exposure time per frame. Four sets of data (600 frames in
each set) were collected, nominally covering complete reciprocal
space. The data were integrated, scaled, sorted, and averaged using
the APEXII software package.26 The structure was solved by direct
method using SHELXS.2 The structure was refined by full-matrix
least squares against F2 using SHELXL.27 All non-hydrogen atoms
were refined anisotropically. X–H distances were set to idealized
values and not refined. Hydrogen atoms were refined with the
‘riding’ model with Uiso¼1.5Ueq for CH3 and Uiso¼1.2Ueq for the
remaining hydrogens. One benzene solvent molecule was found in
the crystal structure.
24. Cotterill, I. C.; Jaouhari, R.; Dorman, G.; Roberts, S. M.; Scheinmann, F.; Wake-
field, B. J. J. Chem. Soc., Perkin Trans. 1 1991, 2505–2512.
Crystal data and data collection information are summarized in
Table 2. Atomic coordinates, anisotropic displacement parameters,
bond lengths, angles, and torsion angles are given in Tables S1–S4
(Supplementary data).
25. Galan, B. R.; Kalbarczyk, K. P.; Szczepankiewicz, S.; Keister, J. B.; Diver, S. T. Org.
Lett. 2007, 9, 1203–1206.
26. APEX2 and SAINT-Plus, Area Detector Control and Integration Software, Ver. 2.0-2;
Bruker Analytical Xray Systems: Madison, Wisconsin, USA, 2005.
27. Sheldrick, G. M. SHELX97. Programs for Crystal Structure Analysis (Release 97-2);
University of Go¨ttingen: Go¨ttingen, Germany, 1997.
The data have been submitted to the CCDC.