On the mechanism of the palladium catalyzed intramolecular Pauson-Khand-Type reaction

Article abstract of DOI:10.1021/jo900919v

(Chemical Equation Presented) Density functional theory calculations and experimental studies have been carried out on the intramolecular Pauson-Khand-Type reaction mediated by a PdCl₂-thiourea catalyst, which proceeds under mild reaction condi

Full text of DOI:10.1021/jo900919v

On the Mechanism of the Palladium Catalyzed Intramolecular
Pauson-Khand-Type Reaction
Yu Lan,† Lujiang Deng,† Jing Liu,† Can Wang,† Olaf Wiest,*,‡ Zhen Yang,*,† and
Yun-Dong Wu*,†,§
College of Chemistry, Peking UniVersity, Beijing, 100871, China, Department of Chemistry, The Hong
Kong UniVersity of Science and Technology, Clear Water Bay, Kowloon, Hong Kong, China, and
Department of Chemistry and Biochemistry, UniVersity of Notre Dame, Notre Dame, Indiana 46556
chydwu@ust.hk; zyang@pku.edu.cn; owiest@nd.edu
ReceiVed May 3, 2009
Density functional theory calculations and experimental studies have been carried out on the intramolecular
Pauson-Khand-Type reaction mediated by a PdCl2-thiourea catalyst, which proceeds under mild reaction
conditions and provides a useful alternative to traditional Pauson-Khand reactions. The classical
mechanism of the Pauson-Khand reaction involving the alkyne/alkene C-C bond formation as the key
step has been found to be energetically unfavorable and is not in line with the experimental observations.
A novel reaction mechanism has been proposed for the reaction. The first step involves the
cis-halometalation of the alkyne, followed by sequential alkene and carbonyl insertion. The rate-determining
fourth step is an intramolecular C-Cl oxidative addition, leading to a PdIV intermediate. A C-C bond
formation by reductive elimination completes the reaction. The mechanism is in agreement with the key
experimental observations including (1) the need of a chloride for catalytic activity and the absence of
catalysis with Pd(OAc)2 alone; (2) the rate acceleration by the addition of LiCl; both with PdCl2 and
Pd(OAc)2 catalysts; and (3) the preferred formation of the trans diastereomer in substituted cases. The
cis halometalation and the formation and stability of the PdIV intermediate is studied in detail and provides
general insights into these novel steps.
Introduction
used for the effective synthesis of cyclopentenone frameworks,2
many of which appear in natural products.3 Although the
The Pauson-Khand reaction (PKR) is formally a [2 + 2 +
1] cycloaddition in which an alkyne, an alkene and a carbon
monoxide couple to form a cyclopentenone.1 The PKR is widely
(2) (a) Shibata, T. AdV. Synth. Catal. 2006, 348, 2328. (b) Gibson, S. E.;
Mainolfi, N. Angew. Chem., Int. Ed. 2005, 44, 3022. (c) Jaime, B. U.; Loreto,
A.; Leticia, P. S.; Gema, D.; Javier, P. C. Chem. Soc. ReV. 2004, 33, 32. (d)
Pauson, P. L. In Organometallics in Organic Synthesis. Aspects of a Modern
Interdisciplinary Field; de Meijere, A., tom Dieck, H., Eds.; Springer: Berlin,
1988; p 233.
(3) (a) Gao, P.; Xu, P. F.; Zhai, H. J. Org. Chem. 2009, 74, 2592. (b) Eduardo,
A.; Jaime, B. U.; Delbrin, A.; Gema, D.; Javier, P. C. J. Organometall. Chem.
2008, 693, 2431. (c) Kaneda, K.; Honda, T. Tetrahedron 2008, 64, 11589. (d)
Nadu, C. E.; Lovely, C. J. Org. Lett. 2007, 9, 4697. (e) Kubota, H.; Lim, J.;
Depew, K. M.; Schreiber, S. L. Chem. Biol. 2002, 9, 265.
† Peking University.
‡ University of Notre Dame.
§ The Hong Kong University of Science and Technology.
(1) (a) Khand, I. U.; Knox, G. R.; Pauson, P. L.; Watts, W. E. J. Chem.
Soc., Chem. Commun. 1997, 36. (b) Khand, I. U.; Knox, G. R.; Pauson, P. L.;
Watts, W. E.; Foreman, M. I. J. Chem. Soc., Perkin Trans. 1 1973, 975. (c)
Khand, I. U.; Knox, G. R.; Pauson, P. L.; Watts, W. E.; Foreman, M. I. J. Chem.
Soc., Perkin Trans. 1 1973, 977.
10.1021/jo900919v CCC: $40.75  2009 American Chemical Society
Published on Web 05/26/2009
J. Org. Chem. 2009, 74, 5049–5058 5049

Lan et al.
SCHEME 1. PdII Catalyzed Pauson-Khand-Type Reaction
SCHEME 2. Two Possible Competitive Mechanisms of
Palladium Catalyzed 1,6-Enyne PKR
reaction is most commonly catalyzed by Co,4 the use of several
other transition metals such as Ti,5 Zr,5b,6 Ni7, Mo,8 Ru,9 Rh,9d,10
and Ir11 in Pauson-Khand-like reactions (for convenience it is
still referred to as PKR hereafter) has been described. Recently,
one of us found that PdCl2 coordinated to a thiourea ligand could
also catalyze an intramolecular PKR, as shown in Scheme 1.12
With reaction temperatures of 50 °C and balloon pressure carbon
monoxide, the reaction conditions are very mild. Several
interesting features of the novel reaction were observed,12
namely that; (a) the reaction could be catalyzed by PdCl2 alone,
but the yield is low; (b) addition of thiourea, especially
tetramethyl thiourea (tmtu) greatly increases the reaction yield;
(c) the presence of chloride is essential, and the reaction could
not be catalyzed by Pd(OAc)2 or Pd(dba)2; (d) added Lewis
acids such as LiCl can increase the reaction rate and ultimately
the yield; (e) the stereochemistry of PdCl2 catalyzed PKR is
different with other metal-catalyzed PKR in that substituted
substrates selectively form the trans diastereomer, thus providing
access to different stereoisomers. These observations are intrigu-
ing and cannot be explained by the widely accepted mechanism
of the PKR, but an explanation of the experimental findings is
not obvious. We therefore decided to pursue a detailed
computational study of the reaction mechanism.
14
somerizations where the catalyst is Pd2(dba)3 or palladacy-
clopentadiene15 are reported to occur via this route. After the
oxidative cyclization, which is thought to be the rate-determining
step, CO inserts to the Pd-C bond. After the reductive
elimination, cyclopentenone is formed. This pathway is in
analogy to the catalytic cycle proposed by Magnus et al.13 for
the Co2(CO)8 catalyzed PKR16,17 that forms the basis for the
proposed mechanisms of almost all the transition metal-catalyzed
PKRs. However, this pathway does not explain the current
experimental observations that are described above. Specifically,
it does not explain the essential role of the chloride anion,
observed stereochemistry, and the acceleration of the reaction
by added LiCl.12
In the alternative pathway B, the reaction is initiated by a
cis-addition of palladium and chloride to the alkyne in a
regioselective fashion where the chloride is transferred to the
terminal position of alkyne. This initiation step has been shown
In principle, the palladium catalyzed 1,n-enyne cycloisomer-
ization can proceed via two possible pathways shown in Scheme
2.13 In the first one (pathway A), a palladacyclopentene is
formed via an oxidative cyclization. Some 1,n-enyne cycloi-
18
19
to be also possible for the cases of PdII or hydrogenated Pd0
catalyzed cycloisomerizations. After insertions of alkene and
CO, Pd is oxidatively inserted into the vinyl-chloride bond,
leading to the formation of the same intermediate as in pathway
A. In both cases, the cyclopentenone product is formed after
the reductive elimination and closing of the catalytic cycle.
Although Pathway B appears reasonable and in agreement
with the experimental observations, as will be discussed in more
detail below, there is to the best of our knowledge no precedence
for such a mechanism. Here, we present a study of the two
pathways shown in Scheme 2. Starting from Nakamura’s work16
for the Co-catalyzed PKR, we investigated the cyclizations of
1-allyloxy-2-butyne as a typical 1,6-enyne by Pd(Cl)2 and
tetramethyl thiourea (tmtu) as the ligand to model the Pd-
catalyzed PKR. We will start by investigating the energetics
(4) (a) Muller, J. L.; Rickers, A.; Leitner, W. AdV. Synth. Catal. 2007, 349,
287. (b) Gibson, S. E.; Stevenazzi, A. Angew. Chem., Int. Ed. 2003, 42, 1800.
(c) Fletcher, A. J.; Christie, S. D. R. J. Chem. Soc., Perkin Trans. 1 2000, 1657.
(5) (a) Hicks, F. A.; Kablaoui, N. M.; Buchwald, S. L. J. Am. Chem. Soc.
1996, 118, 9450; 1994, 116, 8593. (b) Hicks, F. A.; Berk, S. C.; Buchwald,
S. L. J. Org. Chem. 1996, 61, 2713. (c) Hicks, F. A.; Buchwald, S. L. J. Am.
Chem. Soc. 1996, 118, 11688. (d) Hicks, F. A.; Buchwald, S. L. J. Am. Chem.
Soc. 1999, 121, 7026. (e) Hicks, F. A.; Kablaoui, N. M.; Buchwald, S. L. J. Am.
Chem. Soc. 1999, 121, 5881. (f) Zhao, Z. B.; Ding, Y.; Zhao, G. J. Org. Chem.
1998, 63, 9285.
(6) (a) Negishi, E.-I.; Holmes, S. J.; Tour, J. M.; Miller, J. A. J. Am. Chem.
Soc. 1985, 107, 1568. (b) Agnel, G.; Negishi, E.-I. J. Am. Chem. Soc. 1991,
113, 7424.
(7) (a) Tamao, K.; Kobayashi, K.; Ito, Y. J. Am. Chem. Soc. 1988, 110, 1286.
(b) Zhang, M.; Buchwald, S. L. J. Org. Chem. 1996, 61, 4498.
(8) (a) Mukai, C.; Uchiyama, M.; Hanaoka, M. J. Chem. Soc. Chem. Commun.
1992, 1014. (b) Jeong, N.; Lee, S. L.; Lee, B. Y.; Chung, Y. K. Tetrahedron
Lett. 1993, 34, 4027. (c) Kent, J. L.; Wan, H. H.; Brummond, K. M. Tetrahedron
Lett. 1995, 36, 2407. (d) Adrio, J.; Rivero, M. R.; Carretero, J. C. Org. Lett.
2005, 7, 431. (e) Adrio, J.; Carretero, J. C. J. Am. Chem. Soc. 2007, 129, 778.
(9) (a) Morimoto, T.; Chatani, N.; Fukumoro, Y.; Murai, S. J. Org. Chem.
1997, 62, 3762. (b) Kondo, T.; Suzuki, N.; Okada, T.; Mitsudo, T. J. Am. Chem.
Soc. 1997, 119, 6187. (c) Koga, Y.; Kobayashi, T.; Narasaka, K. Chem. Lett.
1998, 249. (d) Kobayashi, T.; Koga, Y.; Narasaka, K. J. Organomet. Chem.
2001, 624, 73. (e) Kondo, T.; Nomura, M.; Ura, Y.; Wada, K.; Mitsudo, T.-A.
J. Am. Chem. Soc. 2006, 128, 14816.
(14) (a) Trost, B. M.; Romero, D. L.; Rise, F. J. Am. Chem. Soc. 1994, 116,
4268. (b) Yamamoto, Y.; Nagata, A.; Itoh, K. Tetrahedron Lett. 1999, 40, 5035.
(15) (a) Trost, B. M.; Tanoury, G. J. J. Am. Chem. Soc. 1988, 110, 1636. (b)
Trost, B. M. Acc. Chem. Res. 1990, 23, 34. (c) Trost, B. M.; Stephen, A.; Hashmi,
K. J. Am. Chem. Soc. 1994, 116, 2183. (d) Trost, B. M.; Yanai, M.; Hoogsteen,
K. J. Am. Chem. Soc. 1993, 115, 5294.
(10) (a) Koga, Y.; Kobayashi, T.; Narasaka, K. Chem. Lett. 1998, 249. (b)
Jeong, N.; Lee, S.; Sung, B. K. Organometallics 1998, 17, 3642. (c) Jeong, N.;
Sung, B. K.; Choi, Y. K. J. Am. Chem. Soc. 2000, 122, 6771. (d) Fan, B. M.;
Xie, J. H.; Li, S.; Tu, Y. Q.; Zhou, Q. L. AdV. Syn. Cat. 2005, 347, 759.
(11) (a) Shibata, T.; Takagi, K. J. Am. Chem. Soc. 2000, 122, 9852. (b)
Shibata, T.; Toshida, N.; Yamazaki, M.; Maekawa, S.; Takagi, K. Tetrahedron
2005, 61, 9974.
(12) (a) Tang, Y.; Deng, L.; Zhang, Y.; Dong, G.; Chen, J.; Yang, Z. Org.
Lett. 2005, 7, 1657. (b) Deng, L.; Liu, J.; Huang, J.; Hu, Y.; Chen, M.; Lan, Y.;
Chen, J.; Lei, A.; Yang, Z. Synthesis 2007, 2565.
(16) Yamanaka, M.; Nakamura, E. J. Am. Chem. Soc. 2001, 123, 1703.
(17) (a) Bruin, T. J. M. de.; Milet, A.; Robert, F.; Gimbert, Y.; Green, A. E.
J. Am. Chem. Soc. 2001, 123, 7184. (b) de Bruin, T. J. M.; Milet, A.; Greene,
A. E.; Gimbert, Y. J. Org. Chem. 2004, 69, 1075. (c) Del Valle, C. P.; Milet,
A.; Gimbert, Y.; Greene, A. E. Angew. Chem., Int. Ed. 2005, 44, 5717.
(18) (a) Ji, J.; Wang, Z.; Lu, X. Organometallics 1996, 15, 2821. (b) Li, J.;
Jiang, H.; Chen, M. J. Org. Chem. 2001, 66, 3627. (c) Wang, Z.; Zhang, Z.; Lu,
X. Organometallics 2002, 19, 775. (d) Lu, X. Top. Catal. 2005, 35, 73. (e) Zhu,
G.; Zhang, Z. J. Org. Chem. 2005, 70, 3339.
(19) (a) Yamamoto, Y.; Nagata, A.; Nagata, H.; Ando, Y.; Arikawa, Y.;
Tatsumi, K.; Itoh, K. Chem.sEur. J. 2003, 9, 2469. (b) Yamamoto, Y.;
Kuwabara, S.; Ando, Y.; Nagata, H.; Nishiyama, H.; Itoh, K. J. Org. Chem.
2004, 69, 6697.
(13) (a) Magnus, P.; Pr´ıncipce, L. M. Tetrahedron Lett. 1985, 26, 4851. (b)
Gimbert, Y.; Lesage, D.; Milet, A.; Fournier, F.; Greene, A. E.; Tabet, J.-C.
Org. Lett. 2003, 5, 4073.
5050 J. Org. Chem. Vol. 74, No. 14, 2009

Intramolecular Pauson-Khand-Type Reaction
FIGURE 1. Relative Gibbs free energies (in THF) of the initial
oxidative coupling step with gas phase Gibbs free energies in
parentheses.
and selectivity of the Magnus-type mechanism (pathway A) for
the case of the PdII-tmtu system. We will then study the
individual steps of the proposed mechanism (Pathway B).
Finally, we will discuss the relationship of the results with the
experimental data.
FIGURE 2. Relative Gibbs free energies (in THF) of the two possible
initial cis-insertion steps with gas phase Gibbs free energies in
parentheses.
Computational Methods
All calculations were carried out with the GAUSSIAN 0320 series
of programs using the BP86 density functional.21 Each structure
was fully optimized using the SDD basis set for Pd and the
6-31+G(d) basis set for all other atoms. Harmonic frequency
calculations were performed for all structures to confirm them as a
local minima or transition structures and to derive the thermo-
chemical corrections for the enthalpies and free energies. Solvent
effects were evaluated using the Polarizable Continuum Model
(PCM)22 using simple united atom topological model radii with
the default parameters for THF. All enthalpies and the Gibbs free
energies in the text are in kcal/mol and are calculated for standard
conditions (298 K, atm) and are corrected for solvation effects as
described above. All distances are given in Å.
d-orbitals of late transition metals are much lower than those
of early ones, leading to much smaller d-π* back-donation,
which means neither the acetylene nor the ethylene are ef-
fectively activated to give an accessible barrier.34
The C-C coupling reaction of the Pd(II) to Pd(IV) mecha-
nism occurs by the alkene rotating to the top of the Pd-alkyne
plane, and then attacking the alkyne through transition structure
12-ts. The free energy of activation for this step is 30.6 kcal/
mol. Compared with the initial step of mechanism B, this
mechanism is not reasonable as will be discussed in more detail
below.
Halometalation Reaction and Regiochemistry. The alterna-
tive mechanism (Pathway B) is initiated by the cis-insertion of
the Pd-Cl bond to the alkyne. There are two possible regio-
chemistries for the insertion with either the palladium or the
chloride adding to the terminal position of the alkyne. The results
for the two possible attacks for the methyl and phenyl substituted
substrates 11 and 17, coordinated to tmtu-PdCl2 are shown in
Figure 2.
Both, the Me and Ph substituted, 1,6-enynes show the same
preferred regiochemistry for the insertion. The relative Gibbs
free energy of activation through 18-ts is 4.5 kcal/mol lower
than 20-ts, and the relative Gibbs free energy of 22-ts is 9.2
kcal/mol lower than 24-ts in THF. The values show that the
formation of 19 and 23 is more facile than the formation of the
regioisomers 21 and 25. The insertion is a reversible, endo-
thermic reaction due to the generation of an open coordination
site on the palladium. Thus a follow-up reaction, for example
the coordination of another ligand such as carbon monoxide or
the alkene moiety of the enyne to the vacant site of Pd is
necessary to decrease the relative free energy and drive the
reaction to completion. The results for these follow-up steps
leading to the formation of the final product are shown in Figure
3 for the case of the methyl-substituted reactant. The cis- and
trans- alkyne insertion step was investigated previously by Lu
and co-workers.18a,d,24,25 These results show that when no or
only a small quantity (1 equiv or less) of LiCl is added, the
cis-insertion product is the major stereoisomer formed. Under
Results and Discussion
C-C Coupling Pathway. Earlier experimental and compu-
tational studies of the Magnus mechanism for the PKR identified
the oxidative addition as the rate determining step of the
reaction.16,17 We therefore started the investigation by calculat-
ing the free energies of activation for the enyne oxidative
coupling step. Both Pd(0) to Pd(II) and Pd(II) to Pd(IV)
mechanism were studied. The results for this mechanism are
summarized in Figure 1. Only the alkyne part of enyne is
coordinated to PdCl2-tmtu to form the intermediate 11. The
pentacoordinated species where PdII is coordinated with both
the double and triple bond is subject to a Jahn-Teller effect23
and could thus not be located. For the Pd(0) to Pd(II) process,
the tricoordinated species 14 leads to 16 through transition
structure 15-ts which has an extraordinarily high Gibbs free
energy of activation of 37.2 kcal/mol. The reason is that the
(20) Frisch, M. J.; et al. Gaussian 03, revision C.02; Gaussian, Inc.:
Wallingford, CT, 2004.
(21) (a) Becke, A. D. J. Chem. Phys. 1993, 98, 1372. (b) Becke, A. D.
J. Chem. Phys. 1993, 98, 5648. (c) Perdew, J. P.; Chevary, J. A.; Vosko, S. H.;
Jackson, K. A.; Pederson, M. R.; Singh, D. J.; Fiolhais, C. Phys. ReV. B 1993,
48, 4978. (d) Perdew, J. P. Phys. ReV. B 1986, 33, 8822.
(22) (a) Barone, V.; Cossi, M. J. Phys. Chem. A 1998, 102, 1995. (b) Barone,
B.; Cossi, M.; Tomasi, J. J. Comput. Chem. 1998, 19, 404.
(23) (a) Englman, R. The Jahn-Teller Effect; Wiley: New York, 1972. (b)
Harvey, P. D.; Reber, C. Can. J. Chem. 1999, 77, 16.
J. Org. Chem. Vol. 74, No. 14, 2009 5051

Lan et al.
FIGURE 3. Relative Gibbs free energies (in THF) for pathway B steps with gas phase Gibbs free energies in parentheses.
the reaction conditions studied here, the cis-alkyne insertion
occurs easily due to the preorientation of the Pd-Cl bond. It
should be mentioned that the trans-insertion may also be
possible in the case where more external chloride ions, for
example from the lithium chloride addition, are present. Based
on the observation that PdCl2 alone is sufficient to catalyze the
reaction, this possibility is not further pursued at this point,
although the effect of the LiCl amount on the reaction yield
will be discussed later.
Pathway B. Figure 3 summarizes the results for the alterna-
tive pathway B, which has four major steps. The catalytic cycle
begins from the alkyne coordinated intermediate 11, which is
the same initiation step as in the oxidative coupling mechanism,
and is set to be the reference energy to provide a direct
comparison between the two pathways. After the formation of
19 through the pathway discussed above, the alkene moiety of
enyne rapidly occupies the open coordination site to form
intermediate 26 in a step that is exergonic by 16.2 kcal/mol.
The coordinated alkene is then inserted into the Pd-C bond
via the transition structure 27-ts with a barrier of 15.2 kcal/
mol. The alkene insertion product 28 has formally an open
coordination site, but is 1.5 kcal/mol more stable than inter-
mediate 26. This suggests an additional stabilization by in-
tramolecular coordination to the vinyl chloride, which in turn
is rapidly displaced by carbon monoxide to form intermediate
29 with a 16.5 kcal/mol Gibbs free energy decrease. It should
be pointed out that this exergonicity is overestimated by ∼2-3
kcal/mol because the loss of entropy from translational and
rotational degrees of freedom in the bimolecular reaction are
not considered in the calculations. The first of the two
carbon-carbon bond formation where the coordinated carbon
monoxide inserts into the Pd-C bond occurs via transition
structure 30-ts. The barrier of the carbon monoxide insertion is
about 13.8 kcal/mol, being relatively low. The stable alkene
coordinated intermediate 31 is formed with 3.5 kcal/mol Gibbs
free energy decrease.
The key step of the new mechanism proposed here is the
oxidative addition of the palladium into the vinylic carbon-
chloride bond via transition structure 32-ts, forming the PdIV
intermediate 33. Although the PdII to PdIV transformation with
oxidative additions to carbon-bromide or carbon-iodide has been
described in both experimental26 and computational studies,27
there is to the best of our knowledge little precedence for the
addition into a carbon-chloride bond in the literature. This step
will therefore be discussed in more detail later. Nevertheless,
the results clearly indicate that the Gibbs free energy of
activation of this step is with 23.9 kcal/mol in THF (23.9 kcal
in the gas phase) almost 7 kcal/mol lower than the rate
determining step (via 12-ts) of the oxidative coupling mecha-
nism, which was calculated to be 30.6 kcal/mol at the same
level of theory. The product of this reaction, the palladium-
containing six-member ring 33 is formed with a 2.3 kcal/mol
increase in Gibbs free energy. The reductive elimination via
34-ts has a barrier of only 13.7 kcal/mol, leading to the
(26) (a) Canty, A. J. Acc. Chem. Res. 1992, 25, 83, and references cited
therein. For a recent X-ray structure of a Pd(IV) species, see: (b) Guo, R. Y.;
Portscheller, J. L.; Day, V. W.; Malinakova, H. C. Organometallics 2007, 26,
3874. (c) Bressy, C.; Alberico, D.; Lautens, M. J. Am. Chem. Soc. 2005, 127,
13148. (d) Faccini, F.; Motti, E.; Catellani, M. J. Am. Chem. Soc. 2004, 126,
78. (e) Whitfield, S. R.; Sanford, M. S. J. Am. Chem. Soc. 2007, 129, 15412. (f)
Furuya, T.; Ritter, T. J. Am. Chem. Soc. 2008, 120, 10060.
(24) (a) Ma, S.; Lu, X. J. Org. Chem. 1991, 56, 5120. (b) Ma, S.; Zhu, G.;
Lu, X. J. Org. Chem. 1993, 58, 3692.
(25) (a) Wang, Z.; Zhang, Z.; Lu, X. Organometallics 2000, 19, 775. (b)
Zhang, Z.; Lu, X.; Xu, Z.; Zhang, Q.; Han, X. Organometallics 2001, 20, 3724.
(27) Sundermann, A.; Uzan, O.; Martin, J. M. L. Chem.sEur. J. 2001, 7,
1703.
5052 J. Org. Chem. Vol. 74, No. 14, 2009

Intramolecular Pauson-Khand-Type Reaction
FIGURE 4. Calculated geometries of key intermediates and transition structures of the two proposed pathways. Bond lengths are in Å.
cyclopentenone coordinated to PdII complex 35. This rapid
follow-up reaction is exergonic by 21.6 kcal/mol, rendering the
overall reaction effectively irreversible. As will be discussed
in more detail below, this pathway is not only energetically more
favorable, but also in better agreement with the other experi-
mental findings.
The calculated geometries of intermediates 31 and 33 as well
as the rate-determining transition structure 32-ts are shown in
Figure 4. In 31, the alkene coordination discussed above is
reflected in the short bond lengths of the Pd-C(1) bond and
Pd-C(2) bond if 2.24 Å and 2.20 Å, respectively. The
carbon-carbon bond length of the olefin is 1.41 Å, much longer
than a typical CdC double bond. This observation can be
rationalized by a combination of donation of π-electrons from
the olefin to the coordinated PdII and the back-donatation from
the d-orbitals of PdII to the antibonding π* orbital. Both effects
lead to the increase of CdC bond length. In contrast, the bond
length of C(1)-Cl(1) is 1.81 Å, close to a normal carbon-
FIGURE 5. Relative Gibbs free energies (in THF) of the competitive
chloride single bond. In 32-ts, Cl(1) is with a distance of 2.44
carbonyl insertion reaction with gas phase Gibbs free energies in
parentheses.
Å coordinated but not completely bonded to the palladium, as
can be seen by comparison with the Pd-Cl bond of 2.35 Å in
33. The C(1)-Cl(1) bond is weakened by this coordination to
a bond length of 2.20 Å, much longer than the same bond in
31. C(2) is completely dissociated from the metal, the distance
between Pd and C(2) is 3.00 Å. The structure of 32-ts shows
that the coordination of alkene is weakened and the C(1)-Cl(1)
bond is broken, but the Pd-Cl(1) and Pd-C(1) bonds have
not been fully formed yet. Consistent with that, the energy of
32-ts is relatively high. The key intermediate 33 is then formed
via oxidative addition. The palladium adopts the characteristic
square pyramidal coordination geometry described previously
for PdIV,27 although the vacant site of the octahedron could also
be occupied by a THF solvent molecule to complete the
coordination sphere.
CO Insertion and Olefin Insertion Sequence. A key feature
of our proposed mechanism is the order in which insertions
occur. After the alkyne inserted into the Pd-C(l) bond, carbon
monoxide insertion is a competitive reaction. Figure 5 compares
the Gibbs free energies for the alkene and the carbon monoxide
insertions. Although carbon monoxide is in most cases a better
ligand than an alkene, in the case of an intra- vs. intermolecular
coordination, the relative free energy of carbon monoxide
coordinated intermediate 36 and the alkene coordinated inter-
mediate 26 are nearly the same because of the entropy loss when
carbon monoxide coordinated to Pd. The barrier for carbon
monoxide insertion via transition structure 37-ts is only about
9.3 kcal/mol, but the insertion product 38 is only 2.8 kcal/mol
more stable than 36. In comparison, the product of the alkene
coordination to palladium, intermediate 39, is 3.6 kcal/mol more
stable than 38. Thus, it can be expected that intermediates 26,
36, 38 and 39 will be in rapid equilibrium. When intermediate
39 is formed, the alkene is likely to insert into the Pd-C bond.
However, the barrier of the alkene insertion step is 29.8 kcal/
mol. The Gibbs free energy of the transition structure 40-ts is
9.9 and 14.1 kcal/mol higher than 27-ts and 37-ts, respectively.
Therefore, it is unlikely that 41 can be formed through this
pathway and 39 will reverse to 26, and further react to 29. The
barrier of complete pathway from 39 to 29 is only 20.9 kcal/
mol, compared to a barrier for the reaction of 39 to 41 of 29.8
kcal/mol.
Generation of the Pd(IV) Intermediate. The rate determin-
ing key step of pathway B is the oxidative addition to form a
PdIV species. Although there is some precedence for the
formation of PdIV through oxidative addition to carbon-iodide
and carbon-bromide bonds,26,27 some of this work is contro-
versial35 and may only apply to the sp3 carbons studies there.
In comparison, there are only few, intramolecular examples for
the addition to crbon-chloride bonds.36 This limited precedence
prompted an investigation of the different effects that might be
able to stabilize the PdIV.
We started our investigation by considering the inter- vs.
intramolecular reaction, shown in Figure 6. The intramolecular
binding of olefin leads to about 10.5 kcal/mol enthalpy release
in THF (from 31a to 31). Taking the entropy loss into account,
the relative free energy of 31 is about 5.3 kcal/mol lower than
31a in THF. Thus, the overall activation free energy for the
formation of Pd(IV) intermediate 33 is about 24.0 kcal/mol
(from 31 to 32-ts). The intermediate 33 is only about 2.3 kcal/
mol endogonic relative to 31 in terms of free energy. A similar
J. Org. Chem. Vol. 74, No. 14, 2009 5053

Lan et al.
TABLE 1. Free Energy Barriers for the Oxidative Addition in
Different Model Systems
FIGURE 6. Comparison of Gibbs free energies (in THF) for the inter-
and intramolecular oxidative addition with enthalpy values (in THF)
in parentheses.
model of 42a and 2-chloride propene 42b are used to study the
corresponding intermolecular oxidative addition. The binding
enthalpy of 42b is about 10.0 kcal/mol in THF, just the same
as the intramolecular result. Yet the entropy loss causes 42 to
be about 7.2 kcal/mol higher than 42a and 42b in terms of free
energy in THF. The activation free energy of the oxidative
addition for intermolecular C-Cl bond is 27.8 kcal/mol in THF
(from 42 to 43-ts). Thus, the overall barrier for the formation
of Pd(IV) intermediate 43a is 35.0 kcal/mol (from 42a+42b to
43-ts). And the intermediate 43a is about 18.0 kcal/mol
endogonic (from 42a+42b to 43a). Thus the generation of
intermolecular oxidative addition Pd(IV) intermediate 43a is
not only kinetically, but also thermodynamically unfavorable.
To probe the electronic influence of the ligands coordinated
to the palladium, we studied additional model systems for the
SCHEME 3. Experimentally Observed
Diastereoselectivities12
(28) There are no directly comparable examples of 3-methyl-1,6-enyne in
Co catalyzed PKR, but 3-TBSO-1,6-enyne in the references below show the
stereochemistry of Co catalyzed PKR. (a) Krafft, M. E.; Bonaga, L. V. R.;
Hirosawa, C. Tetrahedron Lett. 1999, 40, 9171, Entry 12 and 13. (b) Breczinski,
P. M.; Stumpf, A.; Hope, H.; Krafft, M. E.; Casalnuovo, J. A.; Schore, N. E.
Tetrahedron 1999, 55, 6797, Table 2. (c) Moriarty, R. M.; Rani, N.; Enache,
L. A.; Rao, M. S.; Batra, H.; Guo, L.; Penmasta, R. A.; Staszewski, J. P.;
Tuladhar, S. M.; Prakash, O.; Crich, D.; Hirtopeanu, A.; Gilardi, R. J. Org.
Chem. 2004, 66, 1890.
(29) A direct comparison for 42c is not possible because this substrate
undergoes a series of side reactions leading to undefined product mixtures.
(30) (a) Tietze, L. F.; Schulz, G. Liebigs Ann. 1996, 1575. Compare also: (b)
Allinger, N. L.; Hirsch, J. A.; Miller, M. A.; Tyminski, J. J. Am. Chem. Soc.
1968, 90, 5773.
(31) (a) Neumann, F.; Hampel, F.; Schleyer, P. V. Inorg. Chem. 1995, 34,
6553. (b) Jiao, H. J.; Schleyer, P. V. J. Am. Chem. Soc. 1995, 117, 11529.
(32) (a) Holzapfel, C. W.; Marais, L. Tetrahedron Lett. 1997, 38, 8585. (b)
Lu, X. Top. Catal. 2005, 35, 73. (c) Wang, Z.; Zhang, Z.; Lu, X. Organometallics
2000, 19, 775. (d) Wang, Z.; Lu, X. Tetrahedron Lett. 1997, 38, 5213. (e) Zhang,
Z.; Lu, X.; Xu, Z.; Zhang, Q.; Han, X. Organometallic 2001, 20, 3724.
(33) Ba¨ckvall, J. E.; Nilsson, Y. I. M.; Gatti, R. G. P. Organometallics 1995,
14, 4242.
(34) Imabayashi, T.; Fujiwara, Y.; Nakao, Y.; Sato, H.; Sakaki, S. Organo-
metallics 2005, 24, 2129.
(35) (a) Consorti, C. S.; Flores, F. R.; Dupont, J. J. Am. Chem. Soc. 2005,
127, 12054. (b) Bergbreiter, D. E.; Osburn, P. L.; Frels, J. D. AdV. Synth. Cat.
2005, 347, 172.
(36) Liebeskind, L. S.; Baysdon, S. L.; South, M. S.; Iyer, S.; Leeds, J. P.
Tetrahedron 1985, 41, 5839.
(37) For example: (a) Tanaka, K.; Fu, G. C. J. Am. Chem. Soc. 2001, 123,
11492. (b) Onitsuka, K.; Segawa, M.; Takahashi, S. Organometallics 1998, 17,
4335.
oxidative addition. As can be seen from the results summarized
in Table 1, replacement of the tmtu ligand with a THF solvent
increases the barrier for the intramolecular oxidative addition
to 27.6 kcal/mol (entry 3), close to the value for the intermo-
lecular reaction. This is in agreement with the experimental
finding that the tmtu ligand increases the reaction rate.12
A
similar stabilization can be achieved using the electron rich PMe3
ligand (entry 4). Conversely, replacement of the electron
withdrawing acyl group by a methylene lowers the free energy
of activation to 21.2 kcal/mol (entry 5). These findings are
consistent with the idea that a polarizable, Lewis-basic ligand
stabilizes the transition state for the formation of the PdIV
intermediate whereas electron withdrawing groups or less Lewis
basic ligands increase the barrier for the formation of the high
oxidation state of the palladium. The stabilization by tmtu is
also the reason why the novel mechanism involving the PdIV
intermediate is energetically more favorable than the traditional
5054 J. Org. Chem. Vol. 74, No. 14, 2009

Intramolecular Pauson-Khand-Type Reaction
SCHEME 4. Transition States for the Formation of the Diastereomeric Products 51 and 52
Magnus mechanism, which is energetically prohibitive in the
case of late transition metals such as palladium.
to the cis- arrangement of the substituents R1 and R2 in the
allyl system. This repulsive interaction favors the trans
arrangement of the substituents in 55b-ts and 55c-ts by 3.2
and 3.0 kcal/mol, respectively. Although the energy differ-
ences are overestimated as compared to the experimentally
observed selectivities of approximately 9:1 for 50b,29 they
are consistent with other computational studies of 1,3-allyl
strain.30 This overestimation is common in such studies where
only a single conformer is considered because the reaction
will also proceed through other, slightly higher energy
transition state conformations not considered here. More
importantly, these results rationalize the experimentally
observed diastereoselectivity based on the stereoselecting
transition structures of the novel mechanism and allow the
semiquantitative prediction of the diastereomer formed.
Effect of LiCl. Based on the earlier observation that chloride
is necessary for the reaction to occur and the computational
results discussed above, we reinvestigated the reaction and found
that the presence of LiCl speeds up the reaction and ultimately
increases the isolated yield. Interestingly, neither LiBr nor NaCl
show similar enhancements.12b We therefore calculated a model
system of the palladium catalyzed PKR containing LiCl. The
resulting energy profile is shown in Figure 8. In solution, LiCl
will be solvated by THF,31 leading to a tetrahedral lithium
complex modeled here by LiCl(Me2O)3. LiCl(Me2O)3 is bonding
to 31 and dissociates two Me2O molecules to form intermediate
56. The calculated free energy change of this step is 7.8 kcal/
mol. While this value might not be accurate due to increased
charge delocalization in the small model system, it is clear that
31 and 56 will be in equilibrium favoring the later. Coordination
of the lithium species stabilizes transition structure 57-ts and
the activation free energy from 56 to 57-ts is reduced to 22.0
kcal/mol. For comparison, the free energy of activation from
31 to 32-ts is 23.9 kcal/mol, consistent with the experimentally
observed rate increase. The overall reaction energy to form
intermediate 58 via 57-ts is also 7.6 kcal/mol less endergonic
than in the case of 33. This is in agreement with the earlier
observation that the yield increases from only a trace amount
Diastereoselectivity of the Reaction. As mentioned earlier,
the new mechanism presented here is not only found to be
energetically more favorable for the palladium catalyzed
PKR, but is also in better agreement with the experimental
observations. As shown in Scheme 3, we found that the
diastereoselectivity of allylic alkane substituted 1,6-enyne is
opposite to the one described obtained with other transition
metal catalysts.5b,28 If the racemic substrates 50a,b are used
in the palladium catalyzed PRK, the major products are 51
as is unambiguously demonstrated by the X-ray structure of
product 51b (See Supporting Information). In contrast, the
TiCp25b and Co2(CO)828 catalyzed PKR give the products 52
as the major products. Similar results were obtained for diyne
50b, which gave under Pd catalyzed PKR condition the
product 51b as the major product with a d:r of 9:1.
The diastereoselectivity of the reaction is determined by
the position of the alkene moiety of the enyne in the insertion
step. As shown in Figure 3, this step is irreversible and sets
the stereochemistry based on the relative energies of the
diastereomeric transition states. For the substrates shown in
Scheme 4, 50b and 50c, there are two possible alkene
insertion transition structures for the formation of the
diastereomers from the complex 53. The product 51, in which
the bridgehead hydrogen and the substituents R2 are trans to
each other, is produced via 54-ts, and product 52 with the
corresponding cis stereochemistry is produced via 55-ts. We
have calculated the relative energies of the different possible
transition structures 54-ts and 55-ts for the two model
substrates 50b and 50c to elucidate the origin of the
experimentally observed stereochemistry in the Pd catalyzed
PKR. The results are summarized in Figure 7.
It can be seen that the relative energies of 54-ts and 54c-
ts, leading to the experimentally observed products, are
significantly lower than that for the diastereomeric transition
structures 55b-ts and 55ts. In both cases, the transition
structures 55 are subject to a significant 1,3-allylic strain due
FIGURE 7. Geometries and free energies of activation of stereoselecting transition structures in pathway B.
J. Org. Chem. Vol. 74, No. 14, 2009 5055

Lan et al.
effect, the bond length of Pd-Cl(3) is significantly longer than
the ones of Pd-Cl(1) and Pd-Cl(2). The octahedral environ-
ment is further stabilized by the chelation by the lithium which
would be much weaker in the case of sodium or potassium.
Trans Halometalation of Alkyne. Although the novel
mechanism presented here is consistent with the experimental
observations, alternatives to the oxidative addition step need to
be considered. Specifically, a direct trans-halometalation of the
alkyne followed by an insertion of the vinyl chloride moiety
into the acylpalladium species and chloride elimination could
rationalize the formation of 35 without the involvement of the
PdIV species. Previously, the trans-insertion was found to be
competitive to the cis-insertion when large concentrations of
LiCl, which could presumably facilitate the trans-halometalation
by attack of the chloride trans to the palladium, are used.32
Insertions of haloalkenes in acyl-palladium intermediates have
also been reported.37 We therefore decided to investigate this
alternative mechanism.
The Gibbs free energy surface for this alternative mech-
anism is shown in Figure 10. We were unable to locate the
transition structure of the trans-alkyne insertion, but the
insertion intermediate 59 is found to be more stable than cis-
insertion intermediate 26. Furthermore, all intermediates and
transition structures of trans-insertion mechanism (59-64)
are 1-2 kcal/mol more stable than the corresponding
structures in the case of the cis-insertion mechanism (26-31).
After insertion of the olefin and the carbonyl, the fairly stable
intermediate 64 is formed. However, the key insertion step
via 65-ts has a very large free energy of activation of 38.5
kcal/mol, rendering this step prohibitive. Therefore, if the
trans-insertion were to take place, it would lead to the stable
intermediates 62 and 64 that cannot react further.
The geometry of transition structure 65-ts, shown in Figure
11, provides some insight into the high barrier of this step.
Comparison with the alternative transition structure 32-ts reveals
significant strain is present in the [2.1.1] bicyclic system formed
by Pd, C(2), C(1) and C(5) and the five-membered ring. This
leads for example to a compression of the C(2)-C(1)-C(5)
angle to 53° and to several eclipsed interactions that are not
present in 32-ts.
The results shown in Figure 10 are also consistent with the
fact that, because they are easily formed but a further reaction
via transition structures such as 65-ts is energetically prohibitive,
the products of the trans-insertion can be trapped in certain
cases.33 In the present example, 62 or 64 cannot revert to the
starting materials based on the overall exergonicity of the
reaction. It is most likely that these intermediates will decom-
pose, thus lowering the overall yield of the reaction. To test
FIGURE 8. Relative Gibbs free energies (in THF) of the oxidative
addition step in the presence of LiCl with gas phase Gibbs free energies
in parentheses.
of the product present to 82%, using only 1 mol % of LiCl as
a cocatalyst.12b
The origin of these changes in the relative free energies can
be rationalized by analysis of the geometries of 56, 57-ts and
58, shown in Figure 9. For all structures, it appears that one
reason for the stabilization is that the chloride coordinates to
the open site on the palladium in the different species along
the reaction pathway while the lithium coordinates in a chelate
fashion to two chlorides and the carbonyl oxygen. The bond
lengths of Pd-C(1) and Pd-C(2) in 56 are 2.16 Å and 2.17 Å,
much shorter than the same bonds in 31. Conversely, the bond
length of C(1)-C(2) in 56 is 1.43 Å, longer than the same bond
in 31. This is consistent with a stronger coordination of
palladium to the olefin up to the point were the Pd-C(1) and
Pd-C(2) bonds have some single bond character. In this
coordination sphere, the palladium with be electronically closer
to a +IV oxidation state character that is stabilized by the
chloride as well as by a stronger coordination of the olefin. A
d-π* back-donation into the double bond would also decrease
the electron density of palladium, leading to a stronger
coordination of the vinylic chloride. In agreement with this
interpretation, the distance between Pd and Cl(3) in 56 is 2.63
Å. Though longer than a typical Pd-Cl single bond, it clearly
shows a strong interaction. The distance between Pd and Cl3
in 57-ts is decreased to 2.58 Å and the Li-Cl(3) distance is
increased to 2.35 Å, indicating that the interaction Pd-Cl(3)
interaction increases as the Pd-C(2) interaction decreases. In
this approximately octrahedral geometry in 57-ts, the forming
Pd-C(1) bond length is 2.18 Å with Cl(3) occupying the trans-
position relative to the forming bond. The coordination in
intermediate 58 is also octahedral. Because of the strong trans-
FIGURE 9. BP86/6-31+G(d)SDD structures of important intermediates and transition structures of the LiCl-accelerated pathway.
5056 J. Org. Chem. Vol. 74, No. 14, 2009

Intramolecular Pauson-Khand-Type Reaction
FIGURE 10. Relative Gibbs free energies (in THF) for the alternative trans- alkyne insertion mechanism with gas phase Gibbs free energies in
parentheses.
product yield. Unfortunately, we were unable to isolate other
products in this case and no remaining starting material was
observed.
Conclusions
In conclusion, we present a novel mechanism for the
palladium catalyzed Pauson-Khand-type reaction that in-
volves an oxidative addition to a C-Cl bond resulting in
the formation of a PdIV species, which has to the best of our
knowledge not been previously described. The free energies
of activation for this pathway were calculated to be ∼4 kcal/
mol more favorable than the ones for the classical Magnus
FIGURE 11. Structure of 65-ts.
mechanism initiated by oxidative coupling of the olefin. An
alternative mechanism involving a trans-hydrometalation was
TABLE 2. Dependence of Product Yield on the Amount of LiCl
also found to be energetically prohibitive. More importantly,
the new mechanism accounts for a series of experimental
observations that would be hard to explain in the classical
mechanism. Specifically, the effect of Cl- can be explained
by the necessity of a cis-addition to the alkyne for the
equivalent of LiCl
yield of product
following key oxidative addition into the vinyl chlorine bond,
which also rationalizes the previous observations by Ba¨ckvall
et al.33 The observed diastereoselectivity, which is opposite
from other PKRs and therefore synthetically useful, can be
explained by the a 1,3-allylic strain, leading to a repulsive
interaction of the substituents in the R-position and the
terminal position of the alkyne in the stereoselecting transition
structures. Finally, the effect of experimentally observed rate
acceleration by LiCl addition can be rationalized by stabiliza-
tion of the PdIV obtained in the oxidative addition step by a
coordinated chloride and the stabilization of the octahedral
coordination26b of the PdIV species by a chelating lithium.
Further increase of the amount of LiCl was found to decrease
0.5
1.0
2.0
5.0
10.0
29%
36%
42%
25%
22%
this hypothesis, we studied the effect of the amount of LiCl on
the reaction yield and the results are summarized in Table 2.
When 2 eq of LiCl were used in the reaction, the cis- insertion
is the major pathway and the usual product is formed. An
increase of the amount of LiCl to 5 or more equivalents would
lead, in agreement with previous experimental results,33 to
an increased trans-insertion and ultimately to a decreased
J. Org. Chem. Vol. 74, No. 14, 2009 5057

Lan et al.
Compound 51b and 52b.
the product yield, presumably by favoring the trans-insertion
that does not lead to product formation.
Further studies of the scope and limitations of the novel
oxidative addition step as well as the LiCl catalysis of
metalorganic reactions, for which the detailed knowledge of the
geometric and electronic structure of the transition structures
and intermediates involved in the Pd-catalyzed PRK should be
useful in expanding the application of this synthetically useful
reaction. Specifically, the dependence of the stability of the PdIV
intermediate, which has found significant interest in the recent
literature,26 should also be transferable to other cases. The
unique role of the thiourea ligand, which was previously
postulated to stabilize the high oxidation state of the palladium
in a carbon monoxide environment,12 could lead to the develop-
ment of an enantioselective version of this reaction. This and
additional studies of intermolecular examples of the reaction
are currently in progress and will be reported in upcoming
publications.
Typical Procedure for the Compound 51b and 52b. Procedure
A: To a mixture of PdCl2 (8.9 mg, 0.05 mmol), TMTU (6.6 mg,
0.05 mmol) and LiCl (21 mg, 0.5 mmol) in dry THF charged with
CO balloon was added 50b (93 mg, 0.5 mmol) in 15 mL THF.
The reaction mixture was stirred at 60 °C for 48 h. Then, the solvent
was removed under reduced pressure, and the residue was purified
by flash chromatograph (petrol ether/ethyl acetate ) 6:1), to afford
51b and 52b (75 mg, 70%) as colorless oil. Procedure B: To a
mixture of PdCl2 (8.9 mg, 0.05 mmol) and TMTU (6.6 mg, 0.05
mmol) in dry THF charged with CO balloon was added 50b (93
mg, 0.5 mmol) in 15 mL THF. The reaction mixture was stirred at
60 °C for 48 h. Then, the solvent was removed under reduced
pressure, and the residue was purified by flash chromatograph
(petrol ether/ethyl acetate ) 6:1), to afford 51b and 52b (48 mg,
45%) as colorless oil.
Spectroscopic Data for 52b. 1H NMR (200 MHz, CDCl3): 1.15
(d, J ) 6.4 Hz, 3 H), 2.35 (dd, J ) 3 Hz, J ) 17.8 Hz, 1 H), 2.81
(dd, J ) 5.4 Hz, J ) 18.2 Hz, 1 H), 3.36-3.39 (m, 2 H),4.29-4.30
(m, 1 H), 5.23 (q, J ) 6.4 Hz, 1H), 7.26-7.40 (m, 5H); 13C NMR
(50 MHz, CDCl3): δ 17.2, 40.0, 45.0, 69.4, 72.8, 128.3, 128.4,
129.0, 130.0, 154.9, 182.1, 207.1; LRMS (EI): 214 (M+).
Experimental Section
Flash chromatographic separations were carried out on Silica
gel (200-300 mesh).1H NMR and 13C NMR were recorded on
a 300 MHz spectrometer or a 400 MHz spectrometer. LiCl · H2O
was flame-dried before being evacuated by oil pump, and was
stored in desiccators. All the weighing manipulations were
conducted under aerobic conditions. THF was distilled from
sodium under nitrogen atmosphere. 50b was prepared following
literature methods.12a,38
Acknowledgment. We are grateful to the National Natural
Science Foundation of China (20225312, 20325208, and
20832003) and the Research Grants Council of Hong Kong for
financial support of the research.
Supporting Information Available: Cartesian coordinates
of all structures discussed, computational details as well as X-ray
structure determination of 51b. This material is available free
of charge via the Internet at http://pubs.acs.org.
(38) (a) Kobayashi, T.; Koga, Y.; Narasaka, K. J. Organomet. Chem. 2001,
624, 73. (b) Krafft, M. E.; Romreo, R. H.; Scott, I. L. J. Org. Chem. 1992, 57,
5277.
JO900919V
5058 J. Org. Chem. Vol. 74, No. 14, 2009