Remote substituent effects in ruthenium-catalyzed [2+2] cycloadditions: An experimental and theoretical study

Article abstract of DOI:10.1021/jo060103l

The effects of a remote substituent on the regioselectivity of ruthenium-catalyzed [2+2] cycloadditions of 2-substituted norbornenes with alkynes have been investigated experimentally and theoretically using density functional theory. Most of the cycloadditions occurred smoothly at room temperature, giving the exo cycloadducts in excellent yields. Regioselectivities of 1.2:1 to 15:1 were observed with various substituents on the C-2 position of the norbornenes. Exo-C-2-substituents usually showed greater remote substituent effects on the regioselectivities of the cycloadditions than the corresponding endo-C-2-substituents. The regioselectivity of the cycloadditions with C-2 substituents containing an exocyclic double bond (sp² hybridized carbon at C-2) are much higher than the cycloadditions with the exo and endo 2-substituted norbornenes. Theoretical studies predicted the same trends as experiment and matched the experimental product ratios well. The nature of the regioselectivity in this reaction is discussed. Different strengths of the π(C⁵-C⁶)→π*(C²-Y) or π(C⁵-C⁶)→σ*(C²-Y) orbital interactions in 2-substituted norbornenes result in different degrees of C ⁵-C⁶ double bond polarization. Stronger C ⁵-C⁶ polarization will increase the difference in the activation energies between the major and minor pathways and thus lead to greater regioselectivities.

Full text of DOI:10.1021/jo060103l

Remote Substituent Effects in Ruthenium-Catalyzed [2+2]
Cycloadditions: An Experimental and Theoretical Study
Peng Liu, Robert W. Jordan, Steven P. Kibbee, John D. Goddard,* and William Tam*
Guelph-Waterloo Centre for Graduate Work in Chemistry and Biochemistry, Department of Chemistry,
UniVersity of Guelph, Guelph, Ontario, Canada N1G 2W1
wtam@uoguelph.ca; jgoddard@uoguelph.ca
ReceiVed January 18, 2006
The effects of a remote substituent on the regioselectivity of ruthenium-catalyzed [2+2] cycloadditions
of 2-substituted norbornenes with alkynes have been investigated experimentally and theoretically using
density functional theory. Most of the cycloadditions occurred smoothly at room temperature, giving the
exo cycloadducts in excellent yields. Regioselectivities of 1.2:1 to 15:1 were observed with various
substituents on the C-2 position of the norbornenes. Exo-C-2-substituents usually showed greater remote
substituent effects on the regioselectivities of the cycloadditions than the corresponding endo-C-2-
substituents. The regioselectivity of the cycloadditions with C-2 substituents containing an exocyclic
double bond (sp2 hybridized carbon at C-2) are much higher than the cycloadditions with the exo and
endo 2-substituted norbornenes. Theoretical studies predicted the same trends as experiment and matched
the experimental product ratios well. The nature of the regioselectivity in this reaction is discussed. Different
strengths of the π(C5-C6)fπ*(C2-Y) or π(C5-C6)fσ*(C2-Y) orbital interactions in 2-substituted
norbornenes result in different degrees of C5-C6 double bond polarization. Stronger C5-C6 polarization
will increase the difference in the activation energies between the major and minor pathways and thus
lead to greater regioselectivities.
Introduction
Unlike the remote substituent effects on nucleophilic and
electrophilic additions that have been studied extensively,
usually showing strong remote stereoelectronic effects in
controlling the regio- and stereoselectivities, examples of remote
substituent effects on transition-metal-catalyzed cycloadditions
are very rare in the literature and usually only low levels of
selectivities were observed.3f,7c,8 For example, in the Co-
catalyzed Pauson-Khand [2+2+1] cycloadditions of 2-substi-
tuted norbornenes, the C-2 remote substituents showed only very
The study of remote electronic effects in controlling the regio-
and stereoselectivities of nucleophilic and electrophilic additions
to π-bonds has attracted considerable interest.1,2 The most
common systems for these studies are 7-oxabicyclo[2.2.1]hept-
5-ene derivatives (1),3 7-norbornanones (2),4 7-methylenenor-
bornanes (3),5 and 2-substituted norbornenes (4)6,7 (Figure 1).
We have recently studied the effect of a remote substituent on
the regioselectivity of oxymercuration reactions of 2-substituted
norbornenes (4).7b Moderate to high levels of regioselectivity
were observed with both exo- and endo-substituents at C-2 of
norbornenes (Scheme 1).
(3) (a) Black, K. A.; Vogel, P. J. Org. Chem. 1986, 51, 5341. (b) Arjona,
O.; de la Pradilla, R. F.; Plumet, J.; Viso, A. Tetrahedron 1989, 45, 4565.
(c) Arjona, O.; de la Pradilla, R. F.; Garcia, L.; Mallo, A.; Plumet, J. J.
Chem. Soc., Perkin Trans. 2 1989, 1315. (d) Arjona, O.; de la Pradilla, R.
F.; Pita-Romero, I.; Plumet, J.; Viso, A. Tetrahedron 1990, 46, 8199. (e)
Arjona, O.; Manzano, C.; Plumet, J. Heterocycles 1993, 35, 63. (f) Arjona,
O.; Csa´ky¨, A. G.; Murcia, M. C.; Plumet, J. J. Org. Chem. 1999, 64, 7338.
(g) Arjona, O.; Csa´ky¨, A. G.; Murcia, M. C.; Plumet, J. J. Org. Chem.
1999, 64, 9739.
(1) For recent reviews, see: (a) Cieplak, A. S. Chem. ReV. 1999, 99,
1265. (b) Ohwada, T. Chem. ReV. 1999, 99, 1337. (c) Mehta, G.;
Chandrasekhar, J. Chem. ReV. 1999, 99, 1437.
(2) (a) Gung, B. W. Tetrahedron 1996, 52, 5263. (b) Fraser, R. R.;
Faibish, N. C.; Kong, F.; Bednarski, J. J. Org. Chem. 1997, 62, 6167.
10.1021/jo060103l CCC: $33.50 © 2006 American Chemical Society
Published on Web 04/07/2006
J. Org. Chem. 2006, 71, 3793-3803
3793

Liu et al.
catalyzed reactions are much rarer because of the complexity
of the reaction mechanisms and the higher computational cost
of including the metal at reasonably high levels of theory and
with adequate basis sets. The energy differences between
different pathways are often less than a few kilocalories per
mole in the reactions, and a suitable theoretical model is required
to rationalize any computed energy differences. In this paper,
we report the use of density functional methods both to explain
qualitatively and to predict quantitatively the ratios of isomeric
products of the ruthenium-catalyzed [2+2] cycloadditions of
2-substituted norbornenes and unsymmetrical alkynes.
FIGURE 1. Study of remote substituent effects in different bicyclic
systems.
SCHEME 1. Remote Substituent Effects on
Oxymercuration
Theoretical interpretations of the substituent effects have been
proposed on the basis of stereoelectronic arguments in previous
studies of regioselective reactions of 2-substituted norbornenes.7
In this paper, in addition to the electrostatic analysis, we use a
homoconjugation model and a theoretical picture of through-
space orbital interactions to rationalize the remote substituent
effects. The orbital interactions in norbornene and norbornadiene
derivatives have been studied extensively to understand the π
facial reactivities and selectivities.1b,4b,f,5b,10 However, there are
few previous studies which deal with the effects of remote
substituents at the C-2 position that control the regioselectivity
of reactions of the C5-C6 double bond.6,7 In this paper, we
examine the orbital interactions which connect C-2 substituent
effects to the polarization of the C5-C6 double bond, the relative
stability of the transition-state structures, and finally the
regioselectivities.
SCHEME 2. Remote Substituent Effects on the
Pauson-Khand Cycloaddition
weak remote stereoelectronic effects and low levels of regio-
selectivities (1:1 to 2.8:1) were observed (Scheme 2).7c In this
paper, we report our experimental and theoretical results on
remote substituent effects in ruthenium-catalyzed [2+2] cyclo-
additions of 2-substituted norbornenes with unsymmetrical
alkynes. To the best of our knowledge, no example of con-
trolling regioselectivity of any metal-catalyzed [2+2] cyclo-
additions by a remote substituent has been reported in the
literature prior to this study.
Quantum chemical methods have been widely used to study
the mechanisms and selectivities of organic reactions. Recently,
quantitative predictions of stereoselectivities have been applied
to organic reactions, such as aldol reactions,9a 1,3-dipolar
cycloadditions,9b alkylation reactions,9c and benzoin condensa-
tions.9d However, theoretical predictions of transition-metal-
Results and Discussion
Remote Substituent Effects of Exo and Endo 2-Substituted
Norbornenes on Ru-Catalyzed [2+2] Cycloadditions with
Alkyne 5a. Four different [2+2] cycloadducts are theoretically
possible in the cycloaddition between a 2-substituted norbornene
4 and an unsymmetrical alkyne 5 (Scheme 3). On the basis of
our previous work and others,11 Cp*RuCl(COD)-catalyzed
[2+2] cycloadditions between bicyclic alkenes and alkynes
usually produced only the exo cycloadducts. Thus, although four
possible cycloadducts could be formed, we anticipated that only
regioisomers of the exo cycloadducts, 6 and 7, would be formed
in the cycloadditions.
(4) (a) Mehta, G.; Khan, F. A. J. Am. Chem. Soc. 1990, 112, 6140. (b)
Mehta, G.; Khan, F. A.; Ganguly, B.; Chandrasekhar, J. J. Chem. Soc.,
Chem. Commun. 1992, 1711. (c) Ganguly, B.; Chandrasekhar, J.; Khan, F.
A.; Mehta, G. J. Org. Chem. 1993, 58, 1734. (d) Mehta, G.; Khan, F. A. J.
Chem. Soc., Perkin Trans. 1 1993, 1727. (e) Mehta, G.; Khan, F. A.;
Ganguly, B.; Chandrasekhar, J. J. Chem. Soc., Perkin Trans. 2 1994, 2275.
(f) Mehta, G.; Khan, F. A.; Adcock, W. J. Chem. Soc., Perkin Trans. 2
1995, 2189. (g) Mehta, G.; Khan, F. A.; Mohal, N.; Narayan, I. N.;
Kalyanaraman, P.; Chandrasekhar, J. J. Chem. Soc., Perkin Trans. 1 1996,
2665. (h) Mehta, G.; Ravikrishna, C.; Kalyanaraman, P.; Chandrasekhar,
J. J. Chem. Soc., Perkin Trans. 1 1998, 1895. (i) Mehta, G.; Mohal, N. J.
Chem. Soc., Perkin Trans. 1 1998, 505.
SCHEME 3. Possible [2+2] Cycloadducts between a
2-Substituted Norbornene and an Unsymmetrical Alkyne
(5) (a) Mehta, G.; Khan, F. A. J. Chem. Soc., Chem. Commun. 1991,
18. (b) Mehta, G.; Khan, F. A.; Gadre, S. R.; Shirsat, R. N.; Ganguly, B.;
Chandrasekhar, J. Angew. Chem., Int. Ed. Engl. 1994, 33, 1390. (c) Mehta,
G.; Gunasekaran, G. J. Org. Chem. 1994, 59, 1953.
(6) (a) Arjona, O.; de la Pradilla, R. F.; Plumet, J.; Viso, A. J. Org.
Chem. 1991, 59, 6227. (b) Brands, K. M. J.; Kende, A. S. Tetrahedron
Lett. 1992, 33, 5887.
When an equimolar amount of norbornene 4a (Y ) exo-OAc)
and alkyne 5a (R1) Ph, R2 ) COOEt) was treated with 5 mol
% of Cp*RuCl(COD) at 80 °C for 60 h without solvent, only
(7) (a) Mayo, P.; Poirier, M.; Rainey, J.; Tam, W. Tetrahedron Lett.
1999, 40, 7727. (b) Mayo, P.; Orlova, G.; Goddard, J. D.; Tam, W. J. Org.
Chem. 2001, 66, 5182. (c) Mayo, P.; Hecnar, T.; Tam, W. Tetrahedron
2001, 57, 5931. (d) Mayo, P.; Tam, W. Tetrahedron 2001, 57, 5943. (e)
Mayo, P.; Tam, W. Tetrahedron 2002, 58, 9513. (f) Mayo, P.; Tam, W.
Tetrahedron 2002, 58, 9527.
(8) (a) MacWhorter, S. E.; Sampath, V.; Olmstead, M. M.; Schore, N.
E. J. Org. Chem. 1988, 53, 203. (b) Lautens, M.; Tam, W.; Edwards, L. E.
J. Chem. Soc., Perkin Trans. 1 1994, 2143.
(9) (a) Bahmanyar, S.; Houk, K. N.; Martin, H. J.; List, B. J. Am. Chem.
Soc. 2003, 125, 2475. (b) Diaz, J.; Silva, M. A.; Goodman, J. M.; Pellegrinet,
S. C. Tetrahedron 2005, 61, 10886. (c) Gordillo, R.; Carter, J.; Houk, K.
N. AdV. Synth. Catal. 2004, 346, 1175. (d) Dudding, T.; Houk, K. N. Proc.
Natl. Acad. Sci. U.S.A. 2004, 101, 5770.
(10) Mazzocchi, P. H.; Stahly, B.; Dodd, J.; Rondan, N. G.; Domelsmith,
L. N.; Rozeboom, M. D.; Caramella, P.; Houk, K. N. J. Am. Chem. Soc.
1980, 102, 6482.
3794 J. Org. Chem., Vol. 71, No. 10, 2006

Ruthenium-Catalyzed [2+2] Cycloadditions
11% of a mixture of exo regioisomers 6a and 7a was obtained
in a ratio of 3.1:1 (measured by GC and H NMR). No endo
4a-j (Y ) COOMe, OH, OTBS, OBn, OAc) were prepared7e
and their ruthenium-catalyzed [2+2] cycloadditions with alkyne
5a were studied (Tables 1 and 2). Under optimized reaction
conditions, both exo- and endo-norbornenes undergo ruthenium-
catalyzed [2+2] cycloadditions with alkyne 5a smoothly, giving
the corresponding cycloadducts in excellent yields. The regio-
selectivities were only modest, and generally, exo-norbornenes
gave higher regioselectivities than their corresponding endo
isomers. For exo-norbornenes 4a-e, regioselectivities of 1.5:1
to 4.0:1 were observed, with 4a (Y ) OAc) giving the highest
regioselectivity. In the case of endo-norbornenes 4f-j, very little
difference in regioselectivities was observed with different Y
substituents.
1
isomers, 8 or 9, were detected. Using hexanes, toluene, DMF,
or 1,2-dichloroethane as solvent, low yields of the cycloadditions
were observed (11-38%). However, with Et3N or THF as
solvents, the yields increased considerably to 65 and 89%. Very
small changes in the regioselectivities were observed with the
use of different solvents. To optimize the yield and the
regioselectivity of the cycloaddition, several reaction conditions
were varied. Altering the equivalence of alkyne 5a led to
minimal changes in the regioselectivity. However, a dramatic
reduction in the yield was observed as the equivalence of alkyne
5a was increased. When 2 equiv of alkyne 5a were used, the
yield decreased to 29%, and when 5 equiv of alkyne 5a were
used, the yield was reduced to only 9%. On the other hand,
increasing the number of equivalents of norbornene 4a leads to
higher yields and slightly greater regioselectivities. When 2
equiv of norbornene 4a were used, the yield was 90% with
regioselectivity increased to 4.2:1. When 5 equiv of norbornene
4a were used, a quantitative yield was obtained with the same
regioselectivity of 4.2:1. When we carried out the cycloadditions
at different temperatures, we noticed that the cycloadditions
occurred even at room temperature with essentially the same
yield and regioselectivity. Thus, under optimized cycloaddition
conditions, when 1 equiv of alkyne 5a and 5 equiv of norbornene
4a were treated with 5-10 mol % of Cp*RuCl(COD) in THF
at room temperature, cycloadducts 6a and 7a were produced in
94% isolated yield in a ratio of 4.0:1 (Table 1, entry 1).
TABLE 2. Ruthenium-Catalyzed [2+2] Cycloadditions of Endo
2-Substituted Norbornenes
entry
norbornene
Y
yield (%)a
6/7b
1
2
3
4
5
4f
4g
4h
4i
OAc
OBn
OTBS
OH
COOMe
94
92
93
90
91
1.4:1
1.2:1
2.0:1
1.4:1
2.0:1
TABLE 1. Ruthenium-Catalyzed [2+2] Cycloadditions of Exo
2-Substituted Norbornenes
4j
a Isolated yields after column chromatography. b Ratios measured by GC
and/or by 1H NMR and based on the average number from two to four
runs.
Remote Substituent Effects of 2-Substituted Norbornenes
with the C-2 Substituent Containing an Exocyclic Double
Bond (sp2-Hybridized Carbon at C-2) in Ru-Catalyzed [2+2]
Cycloadditions with Alkynes 5a-g. During our studies on the
remote substituent effects on the cobalt-catalyzed Pauson-
Khand [2+2+1] cycloadditions of 2-substituted norbornenes,7d
although the C-2 remote substituents only showed very weak
remote stereoelectronic effects and low levels of regioselec-
tivities (1:1 to 2.8:1) were observed, the highest regioselectivity
was observed with 2-norbornenone 4k, with a sp2-hybridized
carbon at C-2. When 2-norbornenone 4k was subjected to the
Ru-catalyzed cycloaddition conditions with alkyne 5a, an
excellent yield (96%) and a better regioselectivity (7.5:1) were
observed (Table 3, entry 1; compare with Tables 1 and 2). We
then studied the ruthenium-catalyzed [2+2] cycloadditions of
2-norbornenone 4k with several different alkynes, and the results
are shown in Table 3. When an ester moiety is attached to the
acetylenic carbon (R2 ) COOEt, 5a-c, entries 1-3), the Ru-
catalyzed cycloadditions occurred smoothly at room temperature
giving excellent yields of the [2+2] cycloadducts and the
regioselectivity improved from 7.5:1 (when R1 ) Ph) to 8.2:1
entry
norbornene
Y
yield (%)a
6/7b
1
2
3
4
5
4a
4b
4c
4d
4e
OAc
OBn
OTBS
OH
COOMe
94
93
92
91
96
4.0:1
3.0:1
1.5:1
2.0:1
2.4:1
a Isolated yields after column chromatography. b Ratios measured by GC
and/or by 1H NMR and based on the average number from two to four
runs.
To study the effect of the remote Y substituent of norbornene
4 on the cycloaddition, exo and endo 2-substituted norbornenes
(11) For ruthenium-catalyzed [2+2] cycloadditions, see: (a) Mitsudo,
T.; Naruse, H.; Kondo, T.; Ozaki, Y.; Watanabe, Y. Angew. Chem., Int.
Ed. Engl. 1994, 33, 580. (b) Jordan, R. W.; Tam, W. Org. Lett. 2000, 2,
3031. (c) Jordan, R. W.; Tam, W. Org. Lett. 2001, 3, 2367. (d) Jordan, R.
W.; Tam, W. Tetrahedron Lett. 2002, 43, 6051. (e) Villeneuve, K.; Jordan,
R. W.; Tam, W. Synlett 2003, 2123. (f) Villeneuve, K.; Tam, W. Angew.
Chem., Int. Ed. 2004, 43, 610. (g) Jordan, R. W.; Khoury, P. K.; Goddard,
J. D.; Tam, W. J. Org. Chem. 2004, 69, 8467. (h) Villeneuve, K.; Riddell,
N.; Jordan, R. W.; Tsui, G. C.; Tam, W. Org. Lett. 2004, 6, 4543. (i) Riddell,
N.; Villeneuve, K.; Tam, W. Org. Lett. 2005, 7, 3681. (j) Mitsudo, T.;
Kokuryo, K.; Takegami, Y. J. Chem. Soc., Chem. Commun. 1976, 772. (k)
Mitsudo, T.; Kokuryo, K.; Shinsugi, T.; Nakagawa, Y.; Watanabe, Y.;
Takegami, Y. J. Org. Chem. 1979, 44, 4492.
n
(R1 ) Bu); the highest regioselectivity of 10:1 was observed
when R1 ) Cy (cyclohexyl). Comparing the alkynes with a Ph
group attached to the acetylenic carbon (entries 1 and 4-7),
alkynes with ester (5a, R2 ) COOEt) and carboxylic acid (5d,
R2 ) COOH) moieties gave excellent yields in the cycloaddi-
tions at room temperature, whereas alkynes 5e-g (with R2 )
Cl, CH2OH, and Me) required a higher reaction temperature
J. Org. Chem, Vol. 71, No. 10, 2006 3795

Liu et al.
TABLE 3. Ruthenium-Catalyzed [2+2] Cycloadditions of
2-Norbornenone 4k with Different Alkynes
cycloadducts can easily be distinguished by the coupling
constants of H2 and H5 (Figure 2 and Scheme 3) in the H
1
NMR.13 Several exo and endo cycloadducts were modeled for
energy minimization at the PM3 level (CS Chem 3D Pro,
Version 3.5.1) using MOPAC for the assessment of the dihedral
angles between H2 and H5. These dihedral angles were then
compared to the Karplus curve for the determination of the
theoretical coupling constants. On the basis of these computa-
tional studies, the dihedral angles between H5 and H6 (or
between H1 and H2) in the exo cycloadducts 6 and 7 (Scheme
3 and Figure 2) are in the range of 70°-75°; the coupling
constant between H5 and H6 (or between H1 and H2) would be
very small (J ∼ 0-2 Hz). On the other hand, in the endo
cycloadducts 8 and 9, the dihedral angles between H5 and H6
(or between H1 and H2) are in the range of 35°-40° and would
give a coupling constant of 6-9 Hz between H5 and H6 (or
between H1 and H2). In all the cycloadducts that we obtained,
the coupling constants between H5 and H6 (or between H1 and
H2) are always less than 1 Hz, and therefore, all the cycloadducts
must have exo stereochemistry.
entrya
alkyne
R1
R2
yield (%)b
11/12c
1
2
3
4
5
6
7
5a
5b
5c
5d
5e
5f
Ph
nBu
Cy
Ph
Ph
Ph
Ph
COOEt
COOEt
COOEt
COOH
Cl
96
94
96
95
34d
63d
95
7.5:1
8.2:1
10:1
7.9:1
2.7:1
1.6:1
1.4:1
CH2OH
Me
5g
a For entries 1-4, the reactions were run at 25 °C for 45-70 h; for
entries 5-7, the reactions were run at 60-80 °C for 115-165 h. b Isolated
yields after column chromatography. c Ratios measured by GC and/or by
1H NMR and based on the average number from two to four runs.
d Incomplete reaction; unreacted alkyne was recovered.
(60-80 °C). Although alkyne 5g (R2 ) Me) gave an excellent
yield in the cycloaddition, alkynes 5e and 5f are both less
reactive and the cycloadditions were incomplete after stirring
at 80 °C for 165 h. Comparing the regioselectivities of these
Ph-substituted alkynes (5a and 5d-g), we observed good and
similar regioselectivities for 5a (R2 ) COOEt, 7.5:1) and 5d
(R2 ) COOH, 7.9:1) but low levels of regioselectivities were
observed with 5e (R2 ) Cl, 2.7:1), 5f (R2 ) CH2OH, 1.6:1),
and 5g (R2 ) Me, 1.4:1).
To determine if other 2-substituted norbornenes, with the C-2
substituent being an exocyclic double bond (sp2-hybridized
carbon at C-2), also give higher regioselectivity in the Ru-
catalyzed cycloaddition than exo and endo 2-substituted nor-
bornenes 4a-j, as in 2-norbornenone 5k, we have synthesized
norbornene 4m12 and studied its Ru-catalyzed cycloaddition with
alkyne 5c (Scheme 4). To our delight, similar to 2-norbornenone
4k, the Ru-catalyzed cycloaddition between 2-norbornene 4m
and alkyne 5c occurred smoothly at room temperature, giving
the cycloadducts in good yield and an excellent regioselectivity
of 15:1. Thus, the dicyanomethylene group gave the strongest
remote stereoelectronic effects in the Ru-catalyzed cycloaddition.
FIGURE 2. Determination of regio- and stereochemistry of the
cycloadducts.
For the cycloadducts 6a (exo-OAc, Table 1), 6f (endo-OAc,
Table 2), 6j (endo-COOMe, Table 2), 11a and 11c (ketone
group, Table 3), and 11m (dicyanomethylene group, Scheme
4), we were able to isolate and separate the major regioisomers
from the mixture of regioisomers in the cycloadditions either
by fractional recrystallization or by column chromatography.
To determine the regiochemistry of the cycloadducts, we first
identified all the protons and carbons in the cycloadducts
unambiguously by using several NMR techniques (HCOSY,
HSQC, HMBC, and NOESY or GOESY experiments).14,15 To
distinguish the two regioisomers 6a and 7a (Figure 2), NOESY
or GOESY experiments were used. In the GOESY experiments
of the major regioisomer (6a), H5 (at 2.79 ppm) showed +ve
NOE with H2 (2.64 ppm), H7 (4.63 ppm), H6 (2.35 ppm), and
Ha (Ar-H, at 8.02 ppm); H2 (at 2.64 ppm) in 6a showed a +ve
NOE effect with H5 (2.79 ppm), H1 (2.33 ppm), and H8n (1.73
SCHEME 4. Ruthenium-Catalyzed [2+2] Cycloadditions of
Norbornene 4m with Alkyne 5c
(13) Similar methods have been used for the assignment of exo and endo
stereochemistries of bicyclic alkanes and alkenes. See: (a) Yip, C.;
Handerson, S.; Jordan, R.; Tam, W. Org. Lett. 1999, 1, 791. (b) Tranmer,
G. K.; Keech, P.; Tam, W. Chem. Commun. 2000, 863. (c) Yip, C.;
Handerson, S.; Tranmer, G. K.; Tam, W. J. Org. Chem. 2001, 66, 276. (d)
Tranmer, G. K.; Tam, W. J. Org. Chem. 2001, 66, 5113. (e) Flautt, T. J.;
Erman, W. F. J. Am. Chem. Soc. 1963, 85, 3212. (f) Meinwald, J.; Meinwald,
Y. C.; Baker, T. N., III. J. Am. Chem. Soc. 1964, 86, 4074. (g) Smith, C.
D. J. Am. Chem. Soc. 1966, 88, 4273. See also: refs 11a-k.
(14) HCOSY: 1H-1H correlated spectroscopy. HSQC: heteronuclear
single quantum coherence. HMBC: heteronuclear multiple bond correlation.
NOESY: nuclear overhauser enhancement spectroscopy. See: Crews, P.;
Rodriguez, J.; Jaspars, M. Organic Structure Analysis; Oxford University
Press: Oxford, 1998.
(15) GOESY: gradient enhanced nuclear Overhauser enhancement
spectroscopy. See: (a) Stonehouse, J.; Adell, P.; Keeler, J.; Shaka, A. J. J.
Am. Chem. Soc. 1994, 116, 6037. (b) Stott, K.; Stonehouse, J.; Keeler, J.;
Hwang, T.-L.; Shaka, A. J. J. Am. Chem. Soc. 1995, 117, 4199. (c) Dixon,
A. M.; Widmalm, G.; Bull, T. E. J. Magn. Reson. 2000, 147, 266.
Determination of Regio- and Stereochemistry of the
Cycloadducts. The exo and endo stereochemistry of the
(12) Carrupt, P. A.; Vogel, P. HelV. Chim. Acta 1989, 72, 1008.
3796 J. Org. Chem., Vol. 71, No. 10, 2006

Ruthenium-Catalyzed [2+2] Cycloadditions
SCHEME 5. Proposed Mechanism
ppm), but no NOE was observed with any Ar-H (Ha). Similar
methods were used to determine the regiochemistry of most of
the cycloadducts (both major and minor isomers). Cycloadducts
for which we had difficulty in assigning the regiochemistry by
the above NMR techniques were chemically converted to
cycloadducts with known regiochemistries, and their NMR
spectra were compared. In all cases, all the major regioisomers
were found to have similar regiochemistry, having the C-2
remote substituent “syn” to the Ph group and “anti” to the
COOEt group (as in 6a, Figure 2). These assignments also were
supported by X-ray crystallography.16
Discussion of the Mechanism. Previous studies have shown
that formation of a cationic [Cp*Ru]+ species from Cp*RuCl-
(COD) by treatment with AgOTf decreased the catalytic activity,
and low yields (<25%) of the cycloadducts were observed.11
This result suggested that a neutral [Cp*RuCl] moiety is likely
to be the active catalytic species in the cycloadditions. A
proposed mechanism of the Ru-catalyzed [2+2] cycloadditions
between 2-substituted norbornenes 4 and alkyne 5a is shown
in Scheme 5.
Dissociation of one of the double bonds of the cyclooctadiene
(COD) ligand from the catalyst Cp*RuCl(COD) followed by
ligand association with alkyne 5a will provide complex 14
(Scheme 5). Upon dissociation of the COD ligand to form the
coordinatively unsaturated complex 15, either another molecule
of alkyne 5a or 2-substituted norbornene 4 could complex with
15. In the studies of the effect of the number of equivalents of
norbornene 4a and alkyne 5a, we noticed that the use of an
(16) X-ray structures of the major regioisomers of the cycloadducts 6a,
6j, 11a, and 11m were obtained, and their structures were reported: (a)
Schlaf, M.; Jordan, R. W.; Tam, W. Acta Crystallogr. 2002, E58, o629.
(b) Lough, A. J.; Jordan, R. W.; Tam, W. Acta Crystallogr. 2003, E59,
o1675. (c) Tam, W.; Lough, A. J.; Jordan, R. W. Acta Crystallogr. 2003,
E59, o1685. (d) Jordan, R. W.; Tam, W.; Lough, A. J. Acta Crystallogr.
2003, E59, o1687.
J. Org. Chem, Vol. 71, No. 10, 2006 3797

Liu et al.
FIGURE 3. Potential energy profile of the solvent-free model for Ru-catalyzed [2+2] cycloaddition between norbornene 4a (Y ) exo-OAc) and
alkyne 5a (R1 ) Ph, R2 ) COOEt). Energies are Gibbs free energies in kcal/mol, with respect to separated reactants 4a + 5a + Cp*RuCl(COD).
excess of the alkene component improves the yields of the
cycloadditions but, on the other hand, the use of excess alkyne
decreases the yield dramatically. When an excess of alkyne 5a
was used, complex 16 would form preferentially. Because Ru
is known to form stronger π-complexes with alkynes than with
alkenes,17 the formation of complex 16 inhibits the cycloaddi-
tions. This explains why the use of an excess of the alkyne
leads to low yields in the cycloadditions. When an excess of
2-substituted norbornene 4 was used, complexation of 15 to the
π bond on the exo face of 2-substituted norbornene 4 would
give complex 17. In complex 17, there are several ways that
the four different groups (Cl, Cp*, alkene 4, and alkyne 5a)
attached to Ru can be arranged. On the basis of our theoretical
studies in Ru-catalyzed [2+2] cycloadditions between bicyclic
alkenes and alkynes,9g the most favorable arrangement (with
the lowest energy) is with the Cl and the alkyne located above
C5 and C6 and the Cp* pointing away from the bicyclic alkene.
Because both the bicyclic alkene (2-substituted norbornene 4)
and the alkyne (5a) are unsymmetrical, four different Ru-
alkene-alkyne π complexes are possible (17A-D). The Cl can
be located above C6 (syn to the C2 substituent, Y), and the
alkyne can be located above C5, as in 17A and 17D. In 17A,
the ester COOEt group of the unsymmetrical alkyne is pointing
toward the bicyclic alkene and the Ph group of the alkyne is
pointing away from the bicyclic alkene. In 17D, the Ph group
of the unsymmetrical alkyne is pointing toward the bicyclic
alkene and the COOEt group of the alkyne is pointing away
from the bicyclic alkene. On the other hand, the Cl can be
located above C5 (anti to the C2 Y substituent) and the alkyne
can be located above C6, as in 17B and 17C. Oxidative addition
of these Ru-alkene-alkyne π complexes 17A-D would
provide the metallacyclopentenes 19A-D. Reductive elimina-
tion of both the metallacyclopentenes 19A and 19B would give
the major regioisomer 6, whereas reductive elimination of the
metallacyclopentenes 19C and 19D would give the minor
regioisomer 7.
Theoretical Studies. To gain further understanding of remote
substituent effects on the regioselectivities in the cycloadditions
between 2-substituted norbornenes and alkynes, a density
functional study was performed. We studied the reactions of
five different norbornenes, 4a (Y ) exo-OAc), 4d (Y ) exo-
OH), 4f (Y ) endo-OAc), 4i (Y ) endo-OH), and 4k (Y ) O),
with alkyne 5a (R1 ) Ph, R2 ) COOEt). Structures of the π
complexes 17, oxidative addition transition states 18, metalla-
cyclopentenes 19, reductive elimination transition states 20, and
products 6 and 7 for all four possible pathways A-D were
studied. The predicted potential energy profiles of the reaction
between norbornene 4a and alkyne 5a are shown in Figure 3.
For all four pathways, the oxidative addition transition states
18a have energies 2-4 kcal/mol higher than the reductive
elimination transition states 20a. This indicates that the oxidative
addition is rate-determining for the overall reaction. The initial
π complexes 17a in pathways A and C are ca. 3 kcal/mol more
stable than those in pathways B and D. The oxidative addition
transition states 18a in pathways A and C are ca. 4 kcal/mol
more stable than those in pathways B and D. With the oxidative
addition determining the rate, pathways A and C are preferred
leading to products 6a and 7a, respectively. The oxidative
addition transition state 18a-A is 0.6 kcal/mol more stable than
18a-C. This small energy difference in the rate-determining step
could explain the very moderate product regioselectivity.The
metallacyclopentenes 19a-A and 19a-D are ca. 2 kcal/mol more
stable than 19a-B and 19a-C. The reductive elimination
transition states 20a-A and 20a-C are ca. 2 kcal/mol more stable
than 20a-B and 20a-D. The major product 6a is 0.2 kcal/mol
more stable than the minor product 7a.
The potential energy profiles for the other four reactions are
given in the Supporting Information. For all five reactions, the
oxidative addition transition states 18 have higher energies than
(17) (a) Naota, T.; Takaya, H.; Murahashi, S.-I. Chem. ReV. 1998, 98,
2599. (b) Trost, B. M.; Toste, F. D.; Pinkerton, A. B. Chem. ReV. 2001,
101, 2067.
3798 J. Org. Chem., Vol. 71, No. 10, 2006

Ruthenium-Catalyzed [2+2] Cycloadditions
TABLE 4. Relative Free Energies (in kcal/mol) at 298 K for the Oxidative Addition Transition States 18 with Respect to the Separated
Reactants
free energies, ∆Gq
298(theory)
a
∆∆Gq
∆∆Gq
298(exptl)
(exptl selectivity)
298(theory)
entry
norbornene
Y
alkyne
R1
R2
18-A
18-B
18-C
18-D
(calcd selectivity)
1
2
3
4
5
6
7
8
4f
4i
endo-OAc
endo-OH
exo-OH
exo-OAc
O
5a
5a
5a
5a
5a
5b
5c
5c
Ph
Ph
Ph
Ph
Ph
nBu
Cy
Cy
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
35.0
32.0
32.5
35.5
32.9
29.6
33.2
34.7
38.8
36.9
37.7
39.8
38.4
35.7
37.7
39.9
35.6
32.4
33.5
36.1
34.2
31.1
34.9
35.5
39.3
37.3
37.7
39.3
36.8
35.7
37.4
40.0
0.6 (2.8:1)
0.4 (2.0:1)
1.0 (5.4:1)
0.6 (2.8:1)
1.3 (9.0:1)
1.5 (12:1)
1.7 (18:1)
0.8 (3.9:1)
0.20 (1.4:1)
0.20 (1.4:1)
0.41 (2.0:1)
0.82 (4.0:1)
1.19 (7.5:1)
1.24 (8.2:1)
1.36 (10:1)
1.60 (15:1)
4d
4a
4k
4k
4k
4m
O
O
C(CN)2
a Energy differences between preferred structures 18-A and 18-C.
TABLE 5. Relative Free Energies (in kcal/mol) in THF Solution at 298 K for the Oxidative Addition Transition States 18 with Respect to the
Separated Reactants
free energies in THF, ∆Gq
298(THF, theory)
∆∆Gq
∆∆Gq
298(exptl)
(exptl selectivity)
298(THF, theory)
entry
norbornene
Y
alkyne
R1
R2
18-A
18-B
18-C
18-D
(exptl selectivity)
1
2
3
4
5
6
7
8
4f
4i
endo-OAc
endo-OH
exo-OH
exo-OAc
O
5a
5a
5a
5a
5a
5b
5c
5c
Ph
Ph
Ph
Ph
Ph
nBu
Cy
Cy
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
30.0
26.0
26.6
29.2
26.3
25.5
29.6
33.2
32.8
29.2
30.1
33.6
30.6
28.9
32.8
36.9
30.2
26.6
27.0
30.6
28.0
26.5
32.2
36.2
33.1
29.5
29.7
32.9
30.1
28.5
30.9
34.8
0.2a (1.4:1)
0.6a (2.8:1)
0.4a (2.0:1)
1.4a (11:1)
1.7a (18:1)
1.0a (5.4:1)
1.3b (9.0:1)
1.6b (15:1)
0.20 (1.4:1)
0.20 (1.4:1)
0.41 (2.0:1)
0.82 (4.0:1)
1.19 (7.5:1)
1.24 (8.2:1)
1.36 (10:1)
1.60 (15:1)
4d
4a
4k
4k
4k
4m
O
O
C(CN)2
a Energy differences between preferred structures 18-A and 18-C. b Energy differences between preferred structures 18-A and 18-D.
the reductive elimination transition states 20. The oxidative
additions are rate-determining.
predicted to be slightly favored, which means that the regio-
isomers 6 with a C2 substituent anti to the electron-withdrawing
group COOEt are always the major products. Second, in the
experiments, norbornenes with exo-substituents (Table 4, entries
3 and 4) usually have higher selectivities than those with endo-
substituents (Table 4, entries 1 and 2). The theoretically
predicted activation energy differences of the norbornenes with
the exo-substituents are greater (Y ) OH) or equal (Y ) OAc)
to those with the endo-substituents. Third, except for entry 8
which has a slightly smaller activation energy difference than
entry 3, the regioselectivities of the cycloadditions of nor-
bornenes with C2 substituents containing an exocyclic double
bond (Table 4, entries 5-8) are predicted to be higher than those
with the exo and endo 2-substituted norbornenes (Table 4,
entries 1-4).
For greater completeness and reliablity, we studied three more
n
reactions: 2-norbornenone 4k with alkynes 5b (R1 ) Bu, R2
) COOEt) and with 5c (R1 ) Cy, R2 ) COOEt) and norbornene
4m (Y ) C(CN)2) with alkyne 5c. Because the theoretical results
discussed above showed the oxidative addition steps to be rate-
determining and to control regioselectivities, only the reactants
and oxidative addition transition states 18 were studied for these
three reactions. For all eight reactions, the relative Gibbs free
energies at 298 K, ∆Gq298(theory), for the oxidative addition
transition states 18-A-D with respect to the separated reactants
(norbornene, alkyne, and Cp*RuCl(COD)) are given in Table
4. For all the reactions, the 18-A transition states are more stable
than 18-B, and the 18-C transition structures are more stable
than 18-D. Thus, pathway A is preferred leading to product 6,
and pathway C is preferred leading to product 7. The energy
differences between the two preferred oxidative addition transi-
tion states 18-A and 18-C, ∆∆Gq298(theory), control the product
regioselectivity. The theoretically predicted regioselectivities and
To include solvation effects in these studies, free energies in
THF solution at 298 K of the reactants, products, and transition
states were calculated with the PCM solvation model.18 The
potential energy profile of the reaction between norbornene 4a
and alkyne 5a with the PCM model is shown in Figure 4. Results
with this solvation model again indicate that the oxidative
addition transition states 18a have higher energies than the
reductive elimination transition states 20a. The oxidative
addition again is predicted to be the rate-determining step, as
in the gas-phase results. Free energies in THF solution
experimental activation free energy differences, ∆∆Gq
,
298(exptl)
are calculated using absolute rate theory: ln(k1/k2) ) -∆∆Gq/
RT. These two sets of energy differences, ∆∆Gq
and
298(theory)
∆∆Gq298(exptl), are given in Table 4. The theoretically predicted
activation energy differences are in good agreement with the
experimental free-energy differences derived from the product
∆Gq
for the oxidative addition transition states 18
298(THF,theory)
ratios. The mean absolute error between ∆∆Gq
and
298(theory)
of the different reactions and pathways and the free-energy
differences between the preferred pathways leading to the two
regioisomers, ∆∆Gq298(THF,theory), are given in Table 5. It should
be noted that, in entries 7 and 8, pathways D instead of pathways
∆∆Gq298(exptl) is 0.36 kcal/mol. Entries 3 and 8 have the greatest
absolute errors of 0.6 and 0.8 kcal/mol, respectively. The
experimental trends in regioselectivity could be explained well
by these theoretical predictions (except for entries 3 and 8; see
Table 5 for improvement to these exceptions when solvation
effects are included). First, for all reactions, the transition states
18-A are slightly more stable than 18-C. Thus, pathway A is
(18) (a) Miertus, S.; Scrocco, E.; Tomasi, J. Chem. Phys. 1981, 55, 117.
(b) Miertus, S.; Tomasi, J. Chem. Phys. 1982, 65, 239. (c) Cossi, M.; Barone,
V.; Cammi, R.; Tomasi, J. Chem. Phys. Lett. 1996, 255, 327.
J. Org. Chem, Vol. 71, No. 10, 2006 3799

Liu et al.
FIGURE 4. Potential energy profile by the PCM solvation model for Ru-catalyzed [2+2] cycloaddition between norbornene 4a (Y ) exo-OAc)
and alkyne 5a (R1 ) Ph, R2 ) COOEt). Energies are free energies in solution (in kcal/mol), with respect to separated reactants 4a + 5a +
Cp*RuCl(COD). THF (ꢀ ) 7.58) was used as solvent in the PCM calculations.
TABLE 6. NPA Charges with Hydrogens Summed into the Heavy
C are preferred, leading to product 7. Compared to the gas-
phase results for ∆∆Gq298(theory) in Table 4, the free energies in
solution have a smaller mean absolute error, 0.22 kcal/mol. For
entries 3 and 8 that showed a greater deviation from experiment
in the gas-phase calculations, selectivities predicted in solution
are in better agreement with the experimental results. In the
solvation model, entries 4 and 5 have the greatest deviation from
experiment results, 0.6 and 0.5 kcal/mol, respectively. (Thus,
similar trends in the predicted regioselectivities can be observed
in the PCM model, in the gas phase, and in the experiments.)
Atoms for C5, C6, Ca, and Cb in Structures 18a-A to 18a-D
Natural population analyses (NPA)19b,c on the rate-determin-
ing transition-state structures 18a were carried out to study the
electronic effects of the remote substituents. The NPA charges
with hydrogens summed into the heavy atoms on C5, C6, Ca,
and Cb in structures 18a-A to 18a-D are given in Table 6.
Because of the electron-withdrawing ability of COOEt, for all
four structures, Cb, which is connected to COOEt, is more
negatively charged than Ca.20 Because of the inductive effect
of the ruthenium, for 18a-A and 18a-D where the ruthenium is
structure
C5
C6
Ca
Cb
18a-A
18a-B
18a-C
18a-D
0.071
-0.028
-0.019
0.019
-0.022
0.014
0.065
0.059
-0.003
0.052
-0.202
-0.132
-0.193
-0.128
-0.031
-0.009
(19) (a) Glendening, E. D.; Reed, A. E.; Carpenter, J. E.; Weinhold, F.
NBO, version 3.1. (b) Reed, A. E.; Weinhold, F. J. Chem. Phys. 1983, 78,
4066. (c) Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985,
83, 735.
adjacent to C6, C6 is more negative than C5, and for 18a-B and
18a-C where ruthenium is adjacent to C5, C6 is more positive
than C5. In the transition state 18a-A in which the C5-Cb bond
is forming, C5 is positive and Cb is negative. Thus, 18a-A is
stabilized by the electrostatic attraction. In the transition state
18a-B in which the C6-Ca bond is forming, C6 is positive and
Ca is nearly neutral. Thus, there is no such electrostatic
stabilization in 18a-B. For the other two transition states, 18a-C
benefits from this electrostatic stabilization and 18a-D does not.
Thus, considering the orientation of the alkynes, pathways A
and C, in which the electron-donating phenyl group is adjacent
to ruthenium, are favored. With the same orientation of alkynes,
(20) It should be noted that the polarization of the Ca-Cb bond is also
important for the regioselectivity. If Ca and Cb are both connected to
electron-donating groups, such as in the alkyne 4g (R1 ) Ph, R2 ) CH3),
the Ca and Cb will have similar electrostatic properties. In this case, if there
is no obvious steric repulsions or through-space interactions between the
substituent R1 or R2 and the catalyst, or in other words, if the orientation
of the alkyne does not obviously affect the stability of the transition state
18, the activation energies for 18-A and 18-D (or 18-B and 18-C) should
be similar. Thus, no obvious regioselectivity could exist. However,
interactions between substituents on the alkyne and the catalyst might be
masked. Theoretical studies of the effects of different alkynes on reactivity
and selectivity will need to be pursued.
3800 J. Org. Chem., Vol. 71, No. 10, 2006

Ruthenium-Catalyzed [2+2] Cycloadditions
TABLE 7. Selected Bond Lengths (Å) and Percentage Changes from Unsubstituted Norbornene 4n, Wiberg Indices, and Orbital Interactions
(kcal/mol) for Norbornenes with Different 2-Substituents
entry
norbornene
Y
r(C2-C6) (∆r %)
r(C3-C5) (∆r %)
WI(C2-C6)
WI(C3-C5)
orbital interaction
1
2
3
4
5
6
7
4k
4m
4a
4d
4f
O
2.407 (-2.7)
2.393 (-3.3)
2.455 (-0.8)
2.476 (+0.1)
2.480 (+0.2)
2.483 (+0.4)
2.474 (0)
2.466 (-0.3)
2.472 (-0.1)
2.472 (-0.1)
2.471 (-0.1)
2.473 (-0.0)
2.470 (-0.2)
2.474 (0)
0.0288
0.0338
0.0142
0.0123
0.0109
0.0107
0.0112
0.0101
0.0088
0.0100
0.0103
0.0108
0.0110
0.0112
3.40a
2.98a
1.98b
1.59b
0.62c
0.74c
0.88c
C(CN)2
exo-OAc
exo-OH
endo-OAc
endo-OH
H
4i
4n
a π(C5-C6)f π*(C2-Y) orbital interaction. b π(C5-C6)fσ*(C2-Y) orbital interaction. c π(C5-C6)fσ*(C2-H) orbital interaction.
these two structures. Wiberg indices in exo 2-substituted
norbornenes (entries 3 and 4) are slightly larger than those in
endo 2-substituted norbornenes (entries 5 and 6) and in
norbornene itself (entry 7). The Wiberg indices of endo
2-substituted norbornenes (entries 5 and 6) are quite similar to
norbornene (entry 7). Norbornenes with different substituents
show similar Wiberg indices for C3-C5.
FIGURE 5. Orbital interactions between C2 and C6.
The interactions between the filled π(C5-C6) orbital and the
empty π*(C2-Y) or σ*(C2-Y) orbital lead to loss of electron
density from the donor orbital, the π(C5-C6) orbital, into the
acceptor orbital, π*(C2-Y) or σ*(C2-Y). The strength of these
interactions can be used as a measure of delocalization.22 The
π(C5-C6)fπ*(C2-Y) interaction is 3.40 kcal/mol for 4k (Table
7, entry 1) and 2.98 kcal/mol for 4m (Table 7, entry 2). The
π(C5-C6)fσ*(C2-Y) interaction is smaller than the πfπ*
interaction and is 1.98 kcal/mol for 4a (Table 7, entry 3) and
1.59 kcal/mol for 4d (Table 7, entry 4). For endo 2-substituted
norbornenes, no π(C5-C6)fσ*(C2-Y) interactions were ob-
served in the computed results. For the exo-H at the C-2 position,
the π(C5-C6)fσ*(C2-H) interaction is smaller than 1 kcal/
mol (Table 7, entries 5-7).
In conclusion, bond lengths, Wiberg indices, and orbital
interaction analyses all suggest that the norbornenes with an
exocyclic double bond (Table 7, entries 1 and 2) have the
greatest degree of homoconjugation. Norbornenes with exo-2-
substituents (Table 7, entries 3 and 4) have weaker homo-
conjugation than those with an exocyclic double bond. No
homoconjugation could be observed for the norbornenes with
endo-2-substituents (Table 7, entries 5 and 6). The homocon-
jugations led to electron delocalization from the π(C5-C6)
orbital into the π*(C2-Y) orbital or the σ*(C2-Y) orbital. Thus,
the C5-C6 is polarized. C5 has a more positive charge than C6
(Figure 5). A greater degree of delocalization leads to a greater
C5-C6 polarization. C5 will be more negative, and C6 will be
more positive.
To study the effect of orbital interactions and C5-C6
polarizations on the oxidative addition transition state 18,
Mulliken population analyses were carried out for C5 and C6
of 18-A and 18-C (Table 8). The charges and the changes in
the charges relative to the unsubstituted transition states are
given in Table 8. The changes in the charges at C5 for the
substituted relative to the unsubstituted transition states are not
large (less than 0.010 electrons, except for entries 6-8 for 18-
C). Greater distances from the C2 substituents may have caused
the smaller changes. For C6, the changes in the charges are in
agreement with the discussion above. For endo-substituents
(entries 1 and 2) where no obvious orbital interactions exist,
the energy and charge distribution differences between pathways
A and C are due solely to the orientations of the norbornenes.
In pathway A, the C5-Cb bond is formed. In pathway C, the
C6-Cb bond is formed. C5 in pathway A is slightly more
positive than C6 in pathway C. Thus, pathway A is slightly
favored over pathway C.
Having analyzed electron densities and their effects on the
stability of the rate-determining transition states 18 on different
pathways, two questions are: (i) How does the 2-substituent
affect the electronegativities at C6? (ii) Why do exo, endo, and
unsaturated C-2 substituents show different levels of effect? The
electronic effects of the 2-substituent on C6 can be rationalized
by a homoconjugation model. The bent geometry of the bicyclic
frameworks orients the orbitals in such a manner that through-
space interactions are feasible (Figure 5). For the norbornenes
with exocyclic double bonds, 4k and 4m, there is homoconju-
gation between the π*(C2-Y) bond and the π(C5-C6) bond.
For norbornenes with exo-2-substituents, there is homoconju-
gation between the σ*(C2-Y) bond and the π(C5-C6) bond.
For norbornenes with endo-2-substituents, no strong homocon-
jugation could exist. The bond distance changes in r(C2-C6)
of the different 2-substituted norbornenes with respect to
norbornene 4n are shown in Table 7. The distances between
C2-C6 decrease ca. 3% for norbornenes with an exocyclic
double bond (entries 1 and 2). These shorter distances suggest
stronger orbital interactions. The C2-C6 distances of exo-
substituted norbornenes (entries 3 and 4) are slightly shorter
than those of endo-substituted norbornenes (entries 5 and 6).
C2-C6 distances of endo-substituted norbornenes (entries 5 and
6) are slightly longer than that of norbornene itself (entry 7).
C3-C5 distances of all substituted norbornenes are quite similar
to that of norbornene itself.
More information about orbital interactions of norbornenes
with different substituents was obtained using natural bond
orbital (NBO) analysis.19 Wiberg bond indices21 and orbital
interaction energies from NBO analyses are shown in Table 7.
The Wiberg index (WI) is usually considered as the order of
the corresponding bond in the molecule. It is more sensitive to
a change in the bond order than in the bond distance. Entries 1
and 2 in Table 7 have the largest Wiberg indices for C2-C6,
which means the strongest interaction between C2 and C6 in
(22) (a) Weinhold, F. Nature 2001, 411, 539. (b) Schreiner, P. R. Angew.
Chem., Int. Ed. 2002, 41, 3579. (c) Pophristic, V.; Goodman, L. Nature
2001, 411, 565.
(21) Wiberg, K. B. Tetrahedron 1958, 24, 1083.
J. Org. Chem, Vol. 71, No. 10, 2006 3801

Liu et al.
TABLE 8. Mulliken Charges and Changes in These Charges Relative to the Transition State Structures of the Parent Norbornene (entry 9)
for C5 and C6 in 18-A and 18-C
18-A
18-C
entry
norbornene
Y
alkyne
R1
R2
C5
C6
C5
C6
1
2
3
4
5
6
7
8
9
4f
4i
endo-OAc
endo-OH
exo-OH
exo-OAc
O
5a
5a
5a
5a
5a
5b
5c
5c
5a
Ph
Ph
Ph
Ph
Ph
nBu
Cy
Cy
Ph
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
COOEt
-0.099 (-0.002)
-0.100 (-0.003)
-0.098 (-0.001)
-0.100 (-0.003)
-0.095 (+0.002)
-0.095 (+0.002)
-0.108 (-0.011)
-0.107 (-0.010)
-0.097 (0)
-0.155 (-0.004)
-0.145 (+0.006)
-0.163 (-0.012)
-0.169 (-0.018)
-0.185 (-0.034)
-0.198 (-0.047)
-0.197 (-0.046)
-0.189 (-0.038)
-0.151 (0)
-0.154 (-0.003)
-0.150 (+0.001)
-0.154 (-0.003)
-0.160 (-0.009)
-0.159 (-0.008)
-0.171 (-0.020)
-0.170 (-0.019)
-0.166 (-0.015)
-0.151 (0)
-0.103 (-0.006)
-0.086 (+0.011)
-0.110 (-0.013)
-0.115 (-0.018)
-0.117 (-0.020)
-0.118 (-0.021)
-0.129 (-0.032)
-0.132 (-0.035)
-0.097 (0)
4d
4a
4k
4k
4k
4m
4
O
O
C(CN)2
H
substituent effects observed in the ruthenium-catalyzed [2+2]
cycloadditions are significant because such a high level of
remote substituent effect is seldom observed in any transition-
metal-catalyzed cycloaddition reactions.
Theory predicted the same trends as experiment and matched
the experimental product ratios well. The different regioselec-
tivities were interpreted by comparison of activation energies,
stereoelectronic analysis, and orbital interactions. π(C5-
C6)fπ*(C2-Y) is the strongest orbital interaction which exists
in the norbornenes containing an exocyclic double bond. π(C5-
C6)fσ*(C2-Y) is a weaker interaction which exists in the exo
2-substituted norbornenes. The endo 2-substituted norbornenes
do not have such orbital interactions. A stronger orbital
interaction will not only change the geometric and electrostatic
properties of the 2-substituted norbornene reactants but also
increase the degree of C5-C6 bond polarization in the rate-
determining transition-state structure 18. This bond polarization
gives C6 a more negative charge at transition state 18, which is
not favorable to bond formation between C6 and the negatively
charged Cb in 18-C, the transition state leading to the minor
isomeric product. A stronger destabilization of 18-C will
increase the difference of activation energies between the two
pathways leading to the major and minor isomeric products and
finally will result in higher regioselectivity.
FIGURE 6. The orbital interactions give a more positive charge to
C5 and a more negative charge to C6. This will stabilize 18-A and
destabilize 18-C because of the electrostatic interactions.
the charges on C6 are about the same as those in the
unsubstituted transition states (entry 9). For exo-substituents
(entries 3 and 4) where weak orbital interactions exist, the
charges for C6 are slightly more negative (0.010-0.020
electrons) than those in the unsubstituted transition states. For
the substituents with an exocyclic double bond (entries 5-8)
where greater orbital interactions exist, the charges for C6 are
much more negative (0.020-0.050) than those for the unsub-
stituted transition states. These changes in the charges in
particular may be viewed in light of the homoconjugation model
discussed above (Figure 5).
The Mulliken charge analyses suggest that the orbital
interactions in the substituted norbornenes will also affect the
C5-C6 bond polarization of the oxidative addition transition
state 18. A stronger orbital interaction will lead to a more
negative C6. For the transition states 18-C in which C6 is
forming a bond with the negatively charged Cb, a more negative
C6 will destabilize the transition state. In 18-A, this destabiliza-
tion does not exist (Figure 6). Thus, a stronger orbital interaction
will increase the activation energy difference between pathways
A and C and finally leads to a greater regioselectivity.
Experimental Section
Only representative procedures and characterizations of the
products are described here. Full details can be found in the
Supporting Information.
General Procedure for Ruthenium-Catalyzed [2+2] Cyclo-
additions. A mixture of a 2-substituted norbornene (4a-k, 4m)
(1.0 mmol, 5 equiv), an alkyne (5a-g) (0.2 mmol, 1 equiv), and
THF (0.25 mL) in an oven-dried vial was added via a cannula to
an oven-dried screw-cap vial containing Cp*RuCl(COD) (weighed
out from a drybox, 0.01 mmol, 5 mol %) under nitrogen. The
reaction mixture was stirred in the dark at 25-80 °C for 45-165
h. The crude product was purified by column chromatography
(EtOAc-hexane mixtures) to give the cycloadduct.
Cycloadducts 6a and 7a (Table 1, entry 1, Y ) OAc). Yield:
94% (white solid, 6a/7a ) 4:1 measured by GC and 1H NMR). Rf
) 0.39 (EtOAc/hexanes ) 1:9). GC (HP-1 column): retention time
for major isomer 6a ) 30.571 min, and retention time for minor
isomer 7a ) 30.695 min. IR (CH2Cl2) νmax (cm-1) 3066, 2937,
2969, 1734, 1704, 1617, 1492, 1448, 1374, 1297, 1244, 1217, 1204,
1136, 1109. 1H NMR (CDCl3, 400 MHz) δ 8.02 (m, 2H), 7.36 (m,
3H), 4.65 (dm, 1H, J ) 4.6 Hz), 4.23 (q, 2H, J ) 7.1 Hz), 2.81 (d,
0.8H, J ) 3.3 Hz), 2.75 (d, 0.2H, J ) 3.3 Hz), 2.69 (d, 0.2H, J )
3.3 Hz), 2.66 (d, 0.8H, J ) 3.3 Hz), 2.40 (br. s, 0.2H), 2.34-2.36
(m, 1.8H), 2.02 (s, 3H), 1.75 (dd, 1H, J ) 13.3, 7.3 Hz), 1.60 (ddd,
1H, J ) 13.3, 3.9, 2.7 Hz), 1.39 (br. s, 2H), 1.33 (t, 3H, J ) 7.1
Hz). 13C NMR (APT, CDCl3, 100 MHz) major isomer 6a: δ 170.7,
Conclusion
In conclusion, we have demonstrated the first examples of
the remote substituent effect on the regioselectivity of ruthenium-
catalyzed [2+2] cycloadditions between 2-substituted nor-
bornenes and unsymmetrical alkynes. Most of the cycloadditions
occurred smoothly at room temperature, giving the cycloadducts
in excellent yields (usually above 90% except for less reactive
alkynes 5e and 5f). Regioselectivities of 1.2:1 to 15:1 were
observed with various substituents at the C-2 position of
norbornenes. Exo-2-substituents usually showed a greater remote
substituent effect on the regioselectivity of the cycloaddition
than the corresponding endo-2-substituents (Table 1 vs Table
2). The regioselectivity of the cycloadditions with the C-2
substituents containing an exocyclic double bond (sp2-hybridized
carbon at C2) is much higher than that of the cycloadditions
with the exo- and endo 2-substituted norbornenes (Table 3 and
Scheme 4 vs Tables 1 and 2). These studies on the remote
3802 J. Org. Chem., Vol. 71, No. 10, 2006

Ruthenium-Catalyzed [2+2] Cycloadditions
162.7, 154.0, 132.1, 130.1, 128.8(2), 128.3, 76.2, 60.0, 44.6, 41.3,
40.1, 38.2, 33.7, 27.7, 21.2, 14.3. Visible peaks for minor isomer
7a: δ 170.6, 155.6, 132.2, 127.0, 72.3, 45.0, 41.3, 40.8, 39.7, 38.6,
34.0, 27.8.
Computational Details
All computations in this study were carried out with the Gaussian
9823 suite of programs. The Becke three-parameter hybrid func-
tional24 combined with the Lee, Yang, and Parr (LYP) correlation
functional,25 B3LYP, was used. The Los Alamos effective core
potential plus double-ú basis set LANL2DZ26 was employed on
ruthenium, and the 6-31G(d) basis set27 was used for the rest of
the atoms. All electronic structures were closed shells. Harmonic
vibrational frequencies were computed to verify the nature of the
stationary points. Transition-state structures were characterized by
one imaginary frequency and are first-order saddle points. To ensure
that the transition states link the products to the expected reactants,
the normal modes corresponding to the imaginary frequencies were
animated. All reactant and product structures exhibited zero
imaginary frequencies and are minima. The relative Gibbs free
energies (∆G values) were obtained by taking into account zero-
point energies, thermal motion, and entropy contribution at standard
conditions (temperature of 298 K, pressure of 1 atm). Natural bond
orbital (NBO) analyses were performed with the NBO program19
in Gaussian 98. The resulting natural population analysis (NPA)
charge,19b,c Wiberg indices,21 and orbital interactions22 were used
for a detailed study of the electronic structure and bond interactions.
The solvation energies, G298(THF), were computed using the polarized
continuum model of Tomasi and co-workers18 as implemented in
Gaussian 98. THF (ꢀ ) 7.58), which was used in the experiments,
is specified as the solvent in the solvation model.
Fractional recrystallization of the above mixture in EtOAc/
hexanes (1:9) afforded a pure sample of 6a as white crystals. The
regiochemistry of 6a was determined by NMR experiments (1H
NMR, APT, HCOSY, HSQC, HMBC, and NOESY or GOESY
experiments) and confirmed by X-ray crystallography. Rf ) 0.39
(EtOAc/hexanes ) 1:9). Mp 71 °C. GC (HP-1 column): retention
time for major isomer 6a ) 30.571 min. IR (CH2Cl2) νmax (cm-1
)
3066, 2969, 2937, 1734, 1704, 1617, 1492, 1448, 1374, 1297, 1244,
1
1217, 1204, 1136, 1109. H NMR (CDCl3, 400 MHz) δ 8.02 (m,
2H), 7.31 (m, 3H), 4.63 (dm, 1H, J ) 5.3 Hz), 4.22 (q, 2H, J )
7.1 Hz), 2.79 (d, 1H, J ) 3.3 Hz), 2.64 (d, 1H, J ) 3.3 Hz), 2.35
(br. s, 1H), 2.33 (d, 1H, J ) 4.0 Hz), 2.00 (s, 3H), 1.73 (dd, 1H,
J ) 13.3, 7.3 Hz), 1.58 (ddd, 1H, J ) 13.3, 3.9, 2.7 Hz), 1.37 (br.
s, 2H), 1.31 (t, 3H, J ) 7.1 Hz). 13C NMR (APT, CDCl3, 100 MHz)
δ 170.5, 162.5, 153.9, 132.0, 130.0, 128.7, 128.6, 128.2, 76.1, 59.9,
44.5, 41.2, 40.0, 38.1, 33.6, 27.6, 21.1, 14.2. Anal. Calcd for
C20H22O4: C, 73.60; H, 6.79. Found C, 73.56; H, 6.77.
Cycloadducts 11m and 12m (Scheme 4). Yield: 88% (white
solid, 11m/12m ) 15:1 measured by GC and 1H NMR). Rf ) 0.62
(EtOAc/hexanes ) 2:3). GC (HP-1 column): retention time for
major isomer 11m ) 20.016 min and for minor isomer 12m )
20.533 min. IR (CH2Cl2) νmax (cm-1) 2985, 2929, 2853, 2233, 1717,
1705, 1699, 1650, 1615, 1450, 1371, 1336, 1267, 1249, 1206, 1185,
Acknowledgment. This work was supported by NSERC
(Canada) and Boehringer Ingelheim (Canada). W.T. thanks
Boehringer Ingelheim (Canada) Ltd. for a Young Investigator
Award, and R.W.J. thanks the Ontario government and NSERC
(Canada) for postgraduate scholarships (OGS and NSERC PGS
B).
1
1121, 1035. H NMR (CDCl3, 400 MHz) δ 4.19 (q, 2H, J ) 7.1
Hz), 3.42 (s, 0.06H), 3.33 (s, 0.94H), 2.79 (tt, 1H, J ) 11.4, 3.3
Hz), 2.71 (d, 0.94H, J ) 3.2 Hz), 2.65 (dd, 1H, J ) 19.2, 4.4 Hz),
2.63-2.66 (m, 0.12H), 2.55 (m, 1.88H), 2.47 (d, 0.06H), 2.27 (dd,
0.94H, J ) 19.2, 4.2 Hz), 2.25 (dd, 0.06H, J ) 19.2, 4.2 Hz), 1.76
(m, 5H), 1.69 (dm, 1H, J ) 11.2 Hz), 1.42 (dm, 1H, J ) 11.2 Hz),
1.30 (t, 3H, J ) 7.1 Hz), 1.14-1.35 (m, 5H). 13C NMR (APT,
CDCl3, 100 MHz) major isomer 11m: δ 190.1, 163.8, 162.1, 129.8,
111.5, 111.4, 80.3, 60.1, 45.6, 43.2, 42.5, 39.1, 38.8, 33.7, 31.7,
30.7, 30.3, 25.8, 25.60, 25.57, 14.3. Visible peaks for minor isomer
12m: δ 45.4, 44.6, 41.5, 39.3, 34.1.
Supporting Information Available: Detailed experimental
procedures, full characterization of new compounds, and listing of
Cartesian coordinates and total energies for the optimized geom-
etries of calculated species. This material is available free of charge
via the Internet at http://pubs.acs.org.
JO060103L
The major isomer 11m was isolated from the above mixture using
a Chromatotron (EtOAc/hexanes ) 1:9, 1:4) as a white solid. The
regiochemistry of 11m was characterized through the use of NMR
experiments (1H NMR, APT, HSQC, and NOESY) and confirmed
by X-ray crystallography. Mp 88-89 °C. GC (HP-1 column):
retention time ) 19.921 min. 1H NMR (CDCl3, 400 MHz) δ 4.19
(q, 2H, J ) 7.1 Hz), 3.33 (s, 1H), 2.79 (tm, 1H, J ) 11.2 Hz), 2.71
(d, 1H, J ) 3.1 Hz), 2.65 (dd, 1H, J ) 19.2, 4.4 Hz), 2.55 (m,
2H), 2.27 (dd, 1H, J ) 19.2, 4.2 Hz), 1.68-1.78 (m, 6H), 1.42
(dm, 1H, J ) 11.2 Hz), 1.30 (t, 3H, J ) 7.1 Hz), 1.19-1.32 (m,
5H). 1H NMR (C6D6, 400 MHz) δ 3.96 (q, 2H, J ) 7.1 Hz), 2.83
(s, 1H), 2.73 (tt, 1H, J ) 11.6, 3.2 Hz), 2.16 (d, 1H, J ) 3.1 Hz),
1.96 (d, 1H, J ) 4.0 Hz), 1.84 (d, 1H, J ) 3.1 Hz), 1.72 (ddd, 1H,
J ) 19.1, 4.0, 0.9 Hz), 1.50-1.61 (m, 5H), 1.47 (dd, 1H, J ) 19.1,
4.1 Hz), 1.22 (ddm, 1H, J ) 11.1, 4.1 Hz), 1.03-1.13 (m, 3H),
0.91-1.01 (m, 2H), 0.94 (t, 3H, J ) 7.1 Hz), 0.53 (dd, 1H, J )
11.1, 0.9 Hz). 13C NMR (APT, CDCl3, 100 MHz) δ 190.1, 163.9,
162.1, 129.8, 111.53, 111.45, 80.4, 60.1, 45.6, 43.3, 42.5, 39.1,
38.8, 33.7, 31.7, 30.7, 30.3, 25.8, 25.61, 25.58, 14.3. Anal. Calcd
for C21H24O2N2: C, 74.97; H, 7.19. Found C, 74.87; H, 7.15.
(23) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A.
D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi,
M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.;
Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.; Morokuma, K.;
Salvador, P.; Dannenberg, J. J.; Malick, D. K.; Rabuck, A. D.; Raghavachari,
K.; Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Baboul, A. G.; Stefanov,
B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.;
Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.;
Nanayakkara, A.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.; Replogle,
E. S.; Pople, J. A. Gaussian 98, revision A.11; Gaussian, Inc.: Pittsburgh,
PA, 2001.
(24) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(25) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(26) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
(27) (a) Ditchfield, R.; Hehre, W. J.; Pople, J. A. J. Chem. Phys. 1971,
54, 724. (b) Hehre, W. J.; Ditchfield, R.; Pople, J. A. J. Chem. Phys. 1972,
56, 2257. (c) Hariharan, P. C.; Pople, J. A. Theor. Chim. Acta 1973, 28,
213. (d) Hariharan, P. C.; Pople, J. A. Mol. Phys. 1974, 27, 209. (e) Gordon,
M. S. Chem. Phys. Lett. 1980, 76, 163.
J. Org. Chem, Vol. 71, No. 10, 2006 3803