Molecular mechanics predictions and experimental testing of asymmetric palladium-catalyzed allylation reactions using new chiral phenanthroline ligands

Article abstract of DOI:10.1021/ja950860t

Molecular mechanics calculations were used to probe the conformational properties of a number of substituted phenanthrolines and their η³- allylpalladium complexes. Special attention was focused on phenanthrolines bearing chiral, terpene-derived, alkyl and alkenyl groups at C(2). Based upon these calculations, predictions could then be made regarding the suitability of the several ligands for use in asymmetric palladium-catalyzed substitution reactions of allylic acetates. Each of the substituted phenanthrolines was prepared by straightforward means. Use of these ligands in catalytic allylations gave results which were in good agreement with the calculation- based predictions. The highest levels of asymmetric induction were predicted and were obtained with a readily available 2-(2-bornyl)phenanthroline ligand 13. The results were compared with previously reported data obtained using other ligands. Overall, this work provides further indication of the potential utility of a combined calculational/experimental approach for the design of chiral catalysts.

Full text of DOI:10.1021/ja950860t

J. Am. Chem. Soc. 1996, 118, 4299-4313
4299
Molecular Mechanics Predictions and Experimental Testing of
Asymmetric Palladium-Catalyzed Allylation Reactions Using
New Chiral Phenanthroline Ligands
Eduardo Pen˜a-Cabrera,1a Per-Ola Norrby,*,1b Magnus Sjo1gren,1c Aldo Vitagliano,1d
Vincenzo De Felice,1e Johan Oslob,1c Shingo Ishii,1a David O’Neill,1a
Bjo1rn Åkermark,*,1c and Paul Helquist*,1a
Contribution from the Department of Chemistry and Biochemistry, UniVersity of Notre Dame,
Notre Dame, Indiana 46556, Department of Medicinal Chemistry, Royal Danish School of
Pharmacy, UniVersitetsparken 2, DK-2100 Copenhagen, Denmark, Department of Organic
Chemistry, Royal Institute of Technology, S-100 44 Stockholm, Sweden, Dipartimento di Chimica,
UniVersita` di Napoli “Federico II”, Via Mezzocannone 4, I-80134 Napoli, Italy, and Facolta` di
Agraria, UniVersita` del Molise, Via Tiberio 21/A, I-86100 Campobasso, Italy
ReceiVed March 15, 1995. ReVised Manuscript ReceiVed February 18, 1996X
Abstract: Molecular mechanics calculations were used to probe the conformational properties of a number of
substituted phenanthrolines and their η3-allylpalladium complexes. Special attention was focused on phenanthrolines
bearing chiral, terpene-derived, alkyl and alkenyl groups at C(2). Based upon these calculations, predictions could
then be made regarding the suitability of the several ligands for use in asymmetric palladium-catalyzed substitution
reactions of allylic acetates. Each of the substituted phenanthrolines was prepared by straightforward means. Use
of these ligands in catalytic allylations gave results which were in good agreement with the calculation-based
predictions. The highest levels of asymmetric induction were predicted and were obtained with a readily available
2-(2-bornyl)phenanthroline ligand 13. The results were compared with previously reported data obtained using other
ligands. Overall, this work provides further indication of the potential utility of a combined calculational/experimental
approach for the design of chiral catalysts.
Introduction
test the usefulness of our parameter sets in predicting the
stereochemical outcomes for the formation and reactions of η3-
Several approaches have been developed for the asymmetric
synthesis of nonracemic organic compounds. Among the most
common strategies is the use of reactions promoted by metal
catalysts bearing chiral ligands.2 Many of the highly successful
ligand systems that have been developed have arisen largely
by a trial-and-error approach. Our laboratories wished to
explore the feasibility of using molecular mechanics calcula-
tions3 as a basis for the rational design of new chiral ligands.
These calculations would be used to predict the stereoselectivity
of metal complex formation and the stereoselectivity of the
subsequent reactions of the complexes. If this approach were
to prove feasible, then actual laboratory efforts could be invested
in the synthesis of only the most promising candidates for use
as ligands in asymmetric synthesis.
We have previously published some of our studies directed
toward developing molecular mechanics parameter sets for η3-
allylpalladium complexes.4 Likewise, the Pregosin and Albinati
laboratories have employed molecular mechanics calculations
in related studies of chiral palladium complexes.3u,w However,
to meet the longer-term goals stated above, we next needed to
(3) For general reviews of molecular mechanics, see: (a) Burkert, U.;
Allinger, N. L. Molecular Mechanics; American Chemical Society: Wash-
ington, 1982. (b) Brubaker, G. R.; Johnson, D. W. Coord. Chem. ReV. 1984,
53, 1. (c) Hancock, R. D. Prog. Inorg. Chem. 1989, 37, 187. For some
representative examples of molecular mechanics calculations on metal-
containing systems, see: (d) Laier, T.; Larsen, E. Acta Chem. Scand. A
1979, 33, 257. (e) Lauher, J. W. J. Am. Chem. Soc. 1986, 108, 1521. (f)
Brown, J. M.; Evans, P. L.; Lucy, A. R. J. Chem. Soc., Perkin Trans. II
1987, 1589. (g) Brown, J. M.; Evans, P. L. Tetrahedron 1988, 44, 4905.
(h) Bogdan, P. L.; Irwin, J. J.; Bosnich, B. Organometallics 1989, 8, 1450.
(i) Vedani, A.; Huhta, D. W. J. Am. Chem. Soc. 1990, 112, 4759. (j) Lin,
Z.; Marks, T. J. J. Am. Chem. Soc. 1990, 112, 5515. (k) Mayo, S. L.;
Olafson, B. D.; Goddard, W. A., III J. Phys. Chem. 1990, 94, 8897. (l)
Bly, R. S.; Raja, M. Organometallics 1990, 9, 1500. (m) Allured, V. S.;
Kelly, C. M.; Landis, C. R. J. Am. Chem. Soc. 1991, 113, 1. (n) Adam, K.
R.; Antolovich, M.; Brigden, L. G.; Lindoy, L. F. J. Am. Chem. Soc. 1991,
113, 3346. (o) Lee, K. J.; Brown, T. L. Inorg. Chem. 1992, 31, 289. (p)
Bernhardt, P. V.; Comba, P. Inorg. Chem. 1992, 31, 2638. (q) Rappe´, A.
K.; Casewit, C. J.; Colwell, K. S.; Goddard, W. A., III; Skiff, W. M. J.
Am. Chem. Soc. 1992, 114, 10024. (r) Mackie, S. C.; Baird, M. C.
Organometallics 1992, 11, 3712. (s) Polowin, J.; Mackie, S. C.; Baird, M.
C. Organometallics 1992, 11, 3724. (t) Giovannetti, J. S.; Kelly, C. M.;
Landis, C. R. J. Am. Chem. Soc. 1993, 115, 4040. For recent applications
of molecular mechanics calculations to η3-allylpalladium complexes, see:
(u) Pregosin, P. S.; Ru¨egger, H.; Salzmann, R.; Albinati, A.; Lianza, F.;
Kunz, R. W. Organometallics 1994, 13, 83. (v) Gugelchuk, M. M.; Houk,
K. N. J. Am. Chem. Soc. 1994, 116, 330. (w) Trost, B. M.; Breit, B.; Organ,
M. G. Tetrahedron Lett. 1994, 35, 5817. (x) Barbaro, P.; Pregosin, P. S.;
Salzmann, R.; Albinati, A.; Kunz, R. W. Organometallics 1995, 14, 5160.
(y) Taber, D. F.; You, K. K.; Rheingold, A. L. J. Am. Chem. Soc. 1996,
118, 547. For another extensive bibliography, see ref 3r.
X Abstract published in AdVance ACS Abstracts, April 1, 1996.
(1) (a) University of Notre Dame. (b) Royal Danish School of Pharmacy.
(c) Royal Institute of Technology. (d) Universita` di Napoli. (e) Universita`
del Molise.
(2) (a) Kagan, H. B. In ComprehensiVe Organometallic Chemistry;
Wilkinson, G., Stone F. G. A., Abel, E. W., Eds.; Pergamon: Oxford, 1982;
Vol. 8, pp 463-498. (b) Asymmetric Synthesis; Morrison, J. D., Ed.;
Academic Press: New York, 1985; Vol. 5. (c) Asymmetric Catalysis;
Bosnich, B., Ed.; M. Nijhoff: Dordrecht, 1986. (d) Blystone, S. L. Chem.
ReV. 1989, 89, 1663. (e) Scott, J. W. Top. Stereochem. 1989, 19, 209. (f)
Brunner, H.; Zettlmeier, W. Handbook of EnantioselectiVe Catalysis with
Transition Metal Compounds; VCH: Weinheim, 1993. (g) Catalytic
Asymmetric Synthesis; Ojima, I., Ed.; VCH: New York, 1993. (h) Noyori,
R. Asymmetric Catalysis in Organic Synthesis; Wiley: New York, 1994.
(4) (a) Norrby, P.-O.; Åkermark, B.; Hæffner, F.; Hansson, S.; Blomberg,
M. J. Am. Chem. Soc. 1993, 115, 4859. (b) Åkermark, B.; Oslob, J. D.;
Norrby, P.-O. Organometallics 1995, 14, 1688. (c) Andersson, P. G.;
Harden, A.; Tanner, D.; Norrby, P.-O. Chem. Eur. J., 1995, 1, 12. (d)
Tentative parameters were utilized for phenyl substituents on the allyl
moiety: Norrby, P.-O., unpublished data. The complete set of parameters
utilized in this work is available upon request from P.O.N. via e-mail:
peon@medchem.dfh.dk.
S0002-7863(95)00860-2 CCC: $12.00 © 1996 American Chemical Society

4300 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Pen˜a-Cabrera et al.
allylpalladium complexes.5 In this paper, we report studies in
which we have made these predictions for use of several new
ligands, and we report the results of testing these predictions
experimentally.
ment. The specific η3-allyl palladium force field employed in
the present study has been validated both for structures and
energies,4 but for most other synthetically interesting metal
complexes, only generalized parameter sets are available.
Attempts to use some of the general parameter sets that are now
becoming available in commercial programs indicated that these
nonspecific force fields would not yield the accuracy necessary
for our purposes. The present study serves as an additional
test of the usefulness of the parameters that we have developed
to date.4
Results
Molecular Mechanics Studies of the Proposed Ligands and
Their Palladium Complexes: Basic Approach. We begin
this discussion by describing our general molecular mechanics-
based approach and its inherent limitations. As a starting point
for our calculational approach to catalyst design, we chose to
explore the possible conformations of η3-allylpalladium com-
plexes containing various proposed chiral ligands. The molec-
ular mechanics calculations give us two types of relevant
information for each conformation of the complexes, namely
the steric energy and the geometry, both of which have been
shown to be well represented in the case of η3-allylpalladium
complexes with substituted phenanthroline ligands.4 When all
conformations on the reaction pathway have been found, the
product distribution is determined by two factors: the relative
populations and the intrinsic reactivity of the respective
intermediates. Under Curtin-Hammett conditions,6 the popula-
tion is simply scaled by the intrinsic reactivity in order to obtain
the product isomeric ratio. When conformational equilibration
in the intermediate is slower, the intrinsic reactivity is mainly
important in determining which of the allyl termini is attacked
by the nucleophile. Thus, as a very tentative first approximation,
the product distribution can be predicted from the populations
of the respective intermediates, but, in most cases, at least a
qualitative assessment of the reactivity is necessary. Several
structural factors influencing the reactivity have been identified
previously, and therefore a reasonable initial prediction of the
products can usually be made if not only the energies but also
the structures of the intermediates are considered. This is
especially important in locally symmetric cases when attack at
the two different termini on one single conformation will lead
to opposite enantiomers (e.g., in cyclohexenyl complexes). Here
any selectivity can arise only from reactivity differences within
conformations. These points will be discussed further (Vide
infra).
Secondly, and of even more fundamental concern, when
Curtin-Hammett considerations are taken into account, there is
no firm basis for believing that the energy-minimized complexes
that we may identify will be the actual reactive species in the
allylic substitution reactions of interest. Instead, higher energy
complexes that would not be considered in our approach may
be the kinetically more important intermediates in these reac-
tions.6 This Curtin-Hammett factor is, in general, a potential
pitfall in studies based upon the experimental detection or
calculational prediction of reaction intermediates. Our tentative
solution to this problem is to retain high energy species in the
conformational search of the intermediates and to try to identify
any that might have exceptional reactivity. Thorough kinetic
studies coupled with identification of intermediates, their rates
of interconversion, and their involvement in actual reaction
pathways would be required to address this point in greater
detail. However, these further, more elaborate studies lie beyond
the scope of our initial efforts aimed at developing a molecular
mechanics-based approach for the identification of possible
candidates to serve as chiral ligands. Caution must therefore
be exercised in drawing conclusions from our initial results.
Thirdly, a number of reaction parameters can greatly affect
the stereochemical outcome of a given reaction. Among these
factors are the properties of the other reactants such as the
nucleophiles present in a given reaction, the nature of coreagents
such as the bases used to generate reactive intermediates, the
reactivity of leaving groups, the properties of the counterions
present with the reagents and intermediates, and variations in
solvent. Although these variables can have a great bearing on
enantioselectivities of reactions, these factors are not addressed
by the molecular mechanics approach employed in the present
work.
In full recognition of the need to go beyond the present level
of analysis, we have also begun, in parallel work, to address
some of the required reactivity considerations in considerable
detail. For example, in studies to be reported separately,7 we
have developed an approach for quantifying the intrinsic
reactivity based on the structure of several complexes of the
types discussed in the present paper and also based on the steric
effects of the nucleophile itself as it approaches the allyl ligand.
Selection and Study of Ligand Systems. Some of our
earlier work indicated the special role that substituted phenan-
throline ligands play in controlling the stereochemistry of
formation and reactions of η3-allylpalladium complexes.8 For
example, the use of phenanthrolines bearing various substituents
at C-2 and C-9 showed an altered syn,anti-stereoselectivity in
allyl complex formation and important differences in the
reactivity of anti- and syn-substituted termini. This earlier work
There are some readily recognized limitations in this ap-
proach, especially if a generalization to other catalytic systems
is sought. First of all, the required molecular mechanics
parameter sets for metals are still in an early stage of develop-
(5) (a) Maitlis, P. M.; Espinet, P.; Russell, M. J. H. In ComprehensiVe
Organometallic Chemistry; Wilkinson, G., Stone F. G. A., Abel, E. W.,
Eds.; Pergamon: Oxford, 1982; Vol. 6, p 385. (b) Trost, B. M.; Verhoeven,
T. R. In ComprehensiVe Organometallic Chemistry; Wilkinson, G., Stone
F. G. A., Abel, E. W., Eds.; Pergamon: Oxford, 1982; Vol. 8, p 799. (c)
Heck, R. F. Palladium Reagents in Organic Synthesis; Academic Press:
New York, 1985. (d) Trost, B. M. Angew. Chem., Int. Ed. Engl. 1986, 25,
1. (e) Trost, B. M. J. Organomet. Chem. 1986, 300, 263. (f) Tsuji, J. J.
Organomet. Chem. 1986, 300, 281. (g) Tsuji, J. Tetrahedron 1986, 42, 4361.
(h) Consiglio, G.; Waymouth, R. M. Chem. ReV. 1989, 89, 257. (i) Godleski,
S. A. In ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I.,
Semmelhack, M. F., Eds.; Pergamon: Oxford, 1991; Vol. 4, pp 585-661.
(j) Mackenzie, P. B.; Whelan, J.; Bosnich, B. J. Am. Chem. Soc. 1985, 107
, 2046. (k) Hayashi, T.; Yamamoto, A.; Hagihara, T. J. Org. Chem. 1986,
51, 723. (l) Trost, B. M.; Sudhakar, A. R. J. Am. Chem. Soc. 1988, 110,
7933. (m) Cesarotti, E.; Grassi, M.; Prati, L.; Demartin, F. J. Organomet.
Chem. 1989, 370, 407. (n) Hayashi, T.; Kishi, K.; Yamamoto, A.; Ito, Y.
Tetrahedron Lett. 1990, 31, 1743. (o) Fiaud, J.-C.; Legros, J.-Y. J. Org.
Chem. 1990, 55, 4840. (p) Granberg, K. L.; Ba¨ckvall, J.-E. J. Am. Chem.
Soc. 1992, 114, 6858. (q) Dawson, G. J.; Frost, C. G.; Martin, C. J.;
Williams, J. M. J.; Coote, S. J. Tetrahedron Lett. 1993, 34, 7793. (r) Nilsson,
Y. I. M.; Andersson, P. G.; Ba¨ckvall, J.-E. J. Am. Chem. Soc. 1993, 115,
6609. (s) Breutel, C.; Pregosin, P. S.; Salzmann, R.; Togni, A. J. Am. Chem.
Soc. 1994, 116, 4067. (t) Trost, B. M.; Bunt, R. C. J. Am. Chem. Soc. 1996,
118, 235.
(7) Norrby, P.-O.; Helquist, P.; Oslob, J. D.; Åkermark, B. Submitted
for publication.
(8) (a) Åkermark, B.; Hansson, S.; Vitagliano, A. J. Am. Chem. Soc.
1990, 112, 4587. (b) Sjo¨gren, M.; Hansson, S.; Norrby, P.-O.; Åkermark,
B.; Cucciolito, M. E.; Vitagliano, A. Organometallics 1992, 11, 3954. (c)
Hansson, S.; Norrby, P.-O.; Sjo¨gren, M. P. T.; Åkermark, B.; Cucciolito,
M. E.; Giordano, F.; Vitagliano, A. Organometallics 1993, 12, 4940. (d)
Sjo¨gren, M. P. T.; Hansson, S.; Åkermark, B.; Vitagliano, A. Organome-
tallics 1994, 13, 1963.
(6) Curtin, D. Y. Record Chem. Progr. 1954, 15, 111.

Asymmetric Palladium-Catalyzed Allylation Reactions
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4301
Figure 2. Proposed structures of the (η3-allyl)Pd complexes 3 formed
from the ligand rac-2.
Figure 1. Two views of a minimized structure of anti,syn-3.
prompted our initial development of new parameter sets for
performing molecular mechanics calculations on these and other
related palladium-containing systems.4 One of our earlier, key
results was that 2-(8-methoxy-1-naphthyl)phenanthroline forms
(η3-1,3-dimethylallyl)palladium complexes for which the anti,-
syn-1 and syn,syn-1 isomers exist in a ratio of 80:20. An initial
consideration was that an increase in the steric demands of the
phenanthroline substituent should lead to an enhanced anti,-
syn:syn,syn ratio. Furthermore, these types of ligands should
be resolvable due to greatly restricted rotation about the bond
joining the phenanthroline to the methoxy-substituted naphthyl
group, in which case these ligands should induce asymmetry
in the resulting allyl complexes.
After brief molecular mechanics analysis of various deriva-
tives of this system, the benzobicyclo[2.2.2]octane derivative
2 was chosen for more detailed studies. This choice was based
upon (1) prediction of high anti,syn:syn,syn ratios in the
formation of allyl complexes and (2) the potential ease of
preparation of this ligand compared to other possible choices.
Figure 3. Two diastereomeric forms each of 9 and 10.
the fact that a mixture of diastereomeric complexes is present
is likely to lead to a low degree of asymmetric induction. These
results clearly indicated the need to probe the question of
diastereofacial selectivity in the more detailed molecular
mechanics calculations to be performed on each of the further
proposed allyl complexes.
We therefore chose to investigate a series of additional chiral
phenanthroline ligands. Ideally, the ligands would be readily
available in nonracemic form without a need for resolution.
Consideration was given to the substitution of phenanthroline
with groups directly obtainable from naturally occurring terpene
hydrocarbons. After preliminary consideration of several pos-
sible systems, we focused on the use of readily available (1R)-
camphor (5) and (1R)-nopinone (6).
In particular, molecular mechanics calculations were per-
formed on the (η3-1,3-dimethylallyl)palladium complexes 3
containing ligand 2. Figure 1 shows a minimized conformation
of anti,syn-3. Formation of the syn,syn isomer would be
expected to be unfavorable due to interaction of a syn-methyl
group with the benzobicyclooctyl substituent.
High anti,syn preference is not a sufficient condition for a
high level of asymmetric induction in subsequent additions to
the allyl ligand. A brief experimental examination of the
reaction of the racemic ligand 2 with a preformed (η3-1,3-
dimethylallyl)palladium trifluoroacetate complex 4 showed that
a 20:1 mixture of anti,syn and syn,syn complexes 3 is formed
We initially chose to consider the conversion of ketones 5
and 6 into alkenyl derivatives for attachment to phenanthroline
to give ligands 7 and 8, respectively. With our previously
developed molecular mechanics parameter set, the structures
of the anti,syn- and syn,syn-1,3-dimethylallylpalladium com-
plexes 9 and 10 were subjected to an initial minimization.9 The
minimized structures for the anti,syn complexes are shown in
Figure 3.
1
based upon H NMR analysis of the product mixture (eq 1).
However, the anti,syn product consists of a 3:1 mixture of two
diastereomers which can be explained by nondiastereofacial
selective coordination of the 1,3-dimethylallyl ligand to pal-
ladium (Figure 2). The fact that the two anti,syn isomers are
formed suggests that the benzobicyclooctyl group can rotate
away from the anti-methyl group in order to minimize steric
interactions in the second anti,syn isomer depicted in Figure 2.
Even if a nucleophile were to attack selectively at either an anti-
substituted or a syn-substituted terminus of the allyl system,
From Figure 3, we see that the exo face of the allyl group in
both isomers is rather distant from the terpene-derived group.
(9) All the minimizations were performed using the MM2 program,
utilizing the parameter set described in ref 6. MM2(91), version for
Macintosh: MacMimic/MM2(91), InStar Software AB, IDEON Research
Park, S-223 70 Lund, Sweden. MM2 versions for platforms other than
Macintosh are available from the Quantum Chemistry Program Exchange,
Indiana University, Bloomington, IN 47405.

4302 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Pen˜a-Cabrera et al.
Consequently, the achievement of high asymmetric induction
may be thought to be unlikely since the chiral center of the
ligand is on the face of the allyl ligand opposite the exo face
which an incoming nucleophile would attack.5 However, an
interesting feature of the minimized structures is that the allylic
plane does not lie perpendicular to the phenanthroline plane,
but it is rotated and/or pushed away from the coordination plane.
This feature has been observed experimentally by others when
bulky ligands are employed.10 Differences in reactivities of the
allylic termini may therefore result from a more pronounced
distortion caused by the terpene-derived group upon one of the
allylic termini preferentially or from inherent differences in
reactivity of the anti and syn termini (see above),8 provided
that there is good diastereofacial selectivity in the coordination
of the allylic group to the metal.
Initial MM2 calculations on the diastereomeric forms of anti,
syn-9 depicted in Figure 3 show that their energy difference is
very small (0.1 kcal/mol). Also, the calculations indicate that
there is a syn,syn isomer with a steric energy slightly lower
(ca. 0.6 kcal/mol) than that of the anti,syn isomers. A total of
nine different conformers were found within 2 kcal/mol of the
global minimum (a syn,syn isomer). With these small energy
differences, we would predict poor asymmetric induction in
subsequent reactions of the dimethylallyl ligand with nucleo-
philes.
Figure 4. Two diastereomeric forms of complex 12. Calculated bond
distances shown for form 12A. Distance differences are smaller in 12B,
the global minimum energy conformation.
A similar situation arises for anti,syn-10. Our initial calcula-
tions indicate an energy difference of only 0.2 kcal/mol between
the two diastereomeric forms depicted in Figure 3. The
calculations also indicate the existence of two syn,syn diaster-
eomers with steric energies ca. 1 kcal/mol less than their anti,
syn counterparts. The calculated energy difference between the
two syn,syn isomers is only 0.3 kcal/mol. Here we find a total
of six conformers within 2 kcal/mol of the lowest syn,syn isomer.
Therefore, we would again predict poor asymmetric induction
in reactions of 10.
In attempting to reduce the number of factors that affect the
enantioselectivity of the reactions of the allyl complexes, we
decided to consider the use of rac-1-acetoxy-2-cyclohexene (11)
since the question of syn,anti isomerism would not occur in
this cyclic system. The key factor that would govern enantio-
selectivity in this case would be the regiochemistry of nucleo-
philic attack on the coordinated cyclic allyl system.
Figure 5. Minimized structures of ligands 7 and 8 showing coplanarity
between the phenanthroline ring and the olefinic bond of the terpene
unit.
bond distances are unequal, and the two terminal C-Pd
distances are also unequal. These calculated distances are shown
in Figure 4. The slightly greater Pd-N1 distance is likely due
to the closeness of the bulky terpene-derived group at C2. The
somewhat closer interaction between N10 and Pd (increased
electron-donation) may tend to shorten the bond between
palladium and the allylic carbon atom trans to N10 (assuming
the allyl fragment to be an electron-accepting species). If this
perturbation were sufficiently great, it would create greater
carbocation character at the other allylic carbon atom trans to
N1. Therefore, the incoming nucleophile may be expected to
attack preferentially at this latter terminus. Unfortunately, the
difference between the termini in 12B is much smaller, and the
high energy of 12A precludes any dominating influence of this
conformer unless the reactivity of one terminus is extraordinarily
large. Thus, only low asymmetric induction is predicted for
this system.
Further calculations on each of the above allyl systems leads
to additional minimized conformations in addition to those
already discussed, but in no case could a prediction of good
asymmetric induction be made using ligands 7 and 8. These
additional calculations did lead to another important point,
however. Minimized structures of just the noncoordinated
ligands 7 and 8 by themselves are depicted in Figure 5. In
these structures, the olefinic bond of each terpene residue is
coplanar with the phenanthroline ring. The barrier to rotation
of the terpene moiety (calculated to be ca. 7 kcal/mol) would
Molecular mechanics calculations on the corresponding
allylpalladium complex of the camphor-derived ligand leads to
energy minima corresponding to 12A and 12B (Figure 4) having
an energy difference of 1.3 kcal/mol. In order for nucleophilic
addition to take place with significant asymmetric induction,
there must be some factor, either structural or electronic, that
sufficiently differentiates the two allylic termini. As an ad-
ditional complication, complexes of symmetric allyls (e.g., 12)
are expected to show very rapid equilibration between all
possible forms due to fast syn,syn-anti,anti-isomerization (ap-
parent rotation of the allyl moiety).8c According to the Curtin-
Hammett principle,9 the rate of addition to any allylic carbon
is a product of the Boltzmann population of that conformer and
the inherent reactivity of the specific allyl terminus.
Analysis of these structures indicates, as in the acyclic cases,
that the terpene-derived substituent would not be expected to
interact sterically with an incoming nucleophile. Also, the
allylic moiety is seen to lie considerably off of the coordination
plane formed by the palladium atom and the phenanthroline
nitrogen atoms. Of further interest in 12A is that the two Pd-N

Asymmetric Palladium-Catalyzed Allylation Reactions
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4303
Figure 6. Two conformers of ligand 13. The ground state (left) is
hindered at the coordination site, but the low energy conformer to the
right is relatively unhindered.
Figure 8. Steric interactions produced by nucleophilic attack at the
allylic center proximal to the substituted phenanthroline ring of syn,
syn-14.
Figure 7. Minimized diastereomeric forms of complex 14.
necessitate a high energy expenditure upon formation of a metal
complex of either of these ligands. This barrier may reduce
the rate of complexation, or it may destabilize the intermediate
metal complexes and lead to early decomposition of the catalyst.
In either case, these effects may be reflected in low overall rates
of reaction and/or a low yields of the final allylic substitution
products.
Figure 9. Steric interactions that arise from nucleophilic attack at the
syn-substituted allylic C atom of anti,syn-14.
These results indicated the need to modify the structure of
the chiral substituent on the phenanthroline so as to allow free
interaction of the nitrogen atoms with the palladium. Examina-
tion of various modified ligands led us to the dihydro, or bornyl,
analogue 13 of the previously studied bornenylphenanthroline
7. The minimum energy structure of 13 (Figure 6) still shows
severe interaction with the coordination site, but formation of
rotamers with less crowding is much less costly for 13 than for
the previously studied analogue 7.
The first complexes of 13 to be subjected to molecular
mechanics calculations were those of the 1,3-dimethylallyl
system 14. Energy minimization leads to two isomers, namely
syn,syn-14 and anti,syn-14 as depicted in Figure 7, having an
energy difference of only 0.03 kcal/mol. Among the key
structural features of these complexes is the distortion of the
allyl group coordination caused by the terpenyl group. The net
result of this interaction is a displacement of the allyl group
out of the N-Pd-N coordination plane as well as a slight
rotation about the centroid of the allyl group. Another important
feature is the placement of the anti-methyl group in anti,syn-
14 below the unsubstituted carbon ring of the phenanthroline
ligand. Presumably, the methyl group at C-1 of the terpenyl
group prevents placement of the anti-methyl group below the
substituted A ring of the phenanthroline ligand.
Upon addition of nucleophiles to either of these isomers of
14, the initial steric interaction between the incoming nucleo-
phile and the dimethylallyl ligand would not appear to play an
important role in determining enantioselectivity. However, other
important differences may be expected to arise at a later stage
of bond formation. Once nucleophilic addition occurs, the η3-
complex becomes a transient η2-complex by rotation of the
allylic fragment.11 A search of the Cambridge database reveals
that η2-olefin complexes of Pd(0) strongly prefer a trigonal
planar coordination with the olefin parallel to the coordination
plane.12 Thus, if nucleophilic attack were to occur at the allylic
terminus of syn,syn-14 proximal to the substituted phenanthro-
Figure 10. Bottom view of two diastereomeric forms of 15.
line ring, the allylic fragment would rotate counterclockwise in
order to align the newly formed olefin with the coordination
plane, as depicted in Figure 8. This rotation would produce a
strong steric interaction between the methyl group at this
terminus and the terpenyl group. On the other hand, if
nucleophilic attack were to occur on the allylic terminus
proximal to the unsubstituted phenanthroline ring, the subsequent
clockwise rotation would not lead to any significant steric
interactions. Therefore, attack at this latter position and
subsequent formation of the R enantiomer (assuming highest
Cahn-Ingold-Prelog priority for the nucleophile moiety) of
the allylation product would be favored.
Similar reasoning in the case of anti,syn-14, would predict
formation of the S enantiomer of the allylation product due to
unfavorable interactions in formation of the R product (Figure
9). However, because syn,syn-14 would preferentially afford
the R enantiomer, and because the energy difference between
syn,syn-14 and anti,syn-14 is estimated to be only 0.03 kcal/
mol, poor overall asymmetric induction would be predicted for
this dimethylallyl system.
We next studied the cyclic allyl complexes 15 to be derived
from 1-acetoxy-2-cyclohexene (11). Energy minimization leads
to the two structures 15A and 15B (Figure 10), with the latter
being only 0.2 kcal/mol lower in energy than the former. No
significant differences in N-Pd or terminal allylic C-Pd
distances are observed in either structure. However, a key
difference in the two structures is that in 15A the cyclic allyl
group and the terpenyl group lie on opposite sides of the
phenanthroline plane, whereas in 15B, the cyclic allyl group is
much closer to the bulky terpenyl group.
Nucleophilic attack on the allylic terminus closer to the
terpenyl group of 15A would induce clockwise rotation of the
(10) Experimentally, a similar phenomenon has been observed in a related
compound. See: Deeming, A. J.; Rothwell, I. P.; Hursthouse, M. B.; Baker-
Dirks, J. D. J. J. Chem. Soc., Chem. Commun. 1979, 670.
(11) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles
and Applications of Organometallic Chemistry; University Science Books:
Mill Valley, California, 1987; p 417.
(12) Cambridge Structural Database codes for η2-olefin-Pd(0) com-
plexes: BPBZAP, BZYLPD, CARJOU, CUTZAS, DBZACP10, FICBIC,
FMEACA10, GENJAK, HADSIO, SOLTAO, VUFNUF. Only in the case
of the geometrically constrained Pd2(dba)3 (DBZACP10) does any olefin
deviate significantly from the coordination plane of palladium.

4304 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Pen˜a-Cabrera et al.
Figure 11. Effect of rotation of the six-membered ring upon nucleo-
philic attack at either end of the allylic fragment of 15A.
Figure 13. Two diastereomeric forms of syn,syn-16.
Figure 12. Modes of rotation of the six-membered ring upon
nucleophilic attack on 15B.
cyclic allyl unit, whereas attack at the other terminus would
induce counterclockwise rotation (Figure 11). However, in
neither case would significant steric interactions arise between
the allyl residue and the substituted phenanthroline ligand, and
therefore poor asymmetric induction would be predicted for
reaction of 15A.
Figure 14. Predicted regiochemistry for nucleophilic attack on syn,
syn-16B.
On the contrary, attack on the allylic terminus proximal to
the terpenyl group of 15B and subsequent counterclockwise
rotation would reposition the cyclohexenyl residue away from
the terpenyl group, whereas attack at the other terminus
accompanied by clockwise rotation would move the cyclohex-
enyl group toward the bulky substituent (Figure 12). To the
extent that isomer 15B undergoes reaction, good asymmetric
induction would thus be predicted, but this outcome would be
attenuated by competing reaction of 15A. Therefore, at most,
only modest overall asymmetric induction would be predicted
for this cyclic allyl system.
For the purpose of studying a system that would perhaps
exhibit greater interactions between the allylic substituents and
the substituted phenanthroline ligand, we next examined com-
plexes containing the 1,3-diphenylallyl ligand. This allyl system
has been studied extensively in asymmetric allylations by several
other investigators (Vide infra). We considered only syn,syn
isomers since this geometry is greatly favored in this case (as
also confirmed in NMR studies below). Molecular mechanics
calculations lead to two minimized forms, 16A and 16B (Figure
13), having an energy difference of 1.8 kcal/mol, with 16B
having the lower energy. With this energy difference, 16B may
therefore be favored by as much as 95% over 16A at equilib-
rium, although Curtin-Hammett considerations6 must still be
borne in mind. Among the complexes that we have studied,
this system also exhibits the greatest distortion of the allylic
group away from the N-Pd-N coordination plane.
Figure 15. Two diastereomeric forms of 17.
structures anti,syn-17 and syn,syn-17 were found with an energy
difference of only 0.1 kcal/mol slightly favoring the anti,syn
isomer (Figure 15). The forms having the phenyl group
proximal to the terpenyl-substituted phenanthroline ring were
unexpectedly found to be approximately 2 kcal/mol lower in
energy than the forms having the phenyl group proximal to the
unsubstituted phenanthroline ring. A detailed examination of
the possible conformers reveals that the phenyl and methyl
groups experience a comparable number of repulsive van der
Waals interactions (e.g., C-H distances < 2.9 A in the MM2
force field) in the proximal position, whereas the phenyl group
in this position is also stabilized by a large number of attractive
van der Waals interactions (e.g., C-H distances > 3.0 A). Slight
rotation of the allyl causes increased crowding between the syn
position and H9 on the phenanthroline, thus favoring the anti-
methyl isomer. In unsubstituted phenanthroline, where there
is very little rotation of the allyl, the syn/anti ratio is 9/1.8
For either form, we would anticipate at least some preference
for nucleophilic attack at the methyl-substituted terminus of the
allylic system and subsequent rotation of the phenyl-substituted
terminus away from the terpenyl group. Because the two forms
have opposite syn,anti configurations at this site of attack, they
would give opposite configurations of the final allylation
product. However, our prior studies8 indicate the greater
reactivity of anti-substituted allylic termini and, therefore, the
anti,syn isomer may react at the methyl terminus more rapidly
than the syn,syn isomer, thus leading to a modest asymmetric
induction for this site of attack. On the other hand, both forms
have the same syn configuration at the phenyl-substituted
terminus and, therefore, to the extent that reaction may occur
at this position, both forms would give the same configuration
For either of these forms, nucleophilic attack would be
expected to occur at the allylic terminus proximal to the
unsubstituted phenanthroline ring because of the subsequent
rotation of the diphenylallyl group away from the bulky terpenyl
substituent. If the assumption were valid that the more stable
form 16B would also be the principal participant in this reaction,
then we could predict good enantioselectivity to give the R
enantiomer of the allylation product (Figure 14).
As a final case, we chose to study the unsymmetrically
substituted 1-methyl-3-phenylallyl system. The minimized

Asymmetric Palladium-Catalyzed Allylation Reactions
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4305
Table 1. Experimental and Calculated Diastereomeric Ratios of
Some Pd-Allyl Complexes with Chiral Phenanthroline Ligands
Scheme 2
intermediate 22 followed by reaction with phenanthroline.
Rearomatization with dichlorodicyanoquinone (DDQ) then gives
the desired ligand 2 in racemic form.
Several approaches were studied for the preparation of the
chiral terpene-based phenanthroline derivatives, but only the
most successful routes are summarized here. A route quite
different from ours was reported previously by Gladiali et al.,
but in an approach more closely related to ours, the coupling
of a camphor sultam with halophenanthroline derivatives was
reported by Steckhan et al.15
Scheme 1
Based upon the work of Bond,16 the hydrazone 23a obtained
from (1R)-camphor and 2,4,6-triisopropylbenzenesulfonyl-
hydrazide17 undergoes a Shapiro reaction upon treatment with
sec-butyllithium to generate the alkenyllithium derivative 24
(Scheme 2); the simpler p-toluenesulfonylhydrazone 23b was
shown to be useful in subsequent work described below. Direct
reaction18 of 24 with phenanthroline at -78 °C followed by
DDQ-induced rearomatization at -50 °C gives the desired
ligand 7 in 52% yield. Careful temperature control is necessary
during the addition reaction in order to prevent alkene double
bond migration/rearomatization of the intermediate adduct,
which leads to varying amounts of the dihydro, or bornyl,
derivative 13 (Vide infra).
As an alternative approach, we studied the use of the
corresponding alkenyl iodide. Based upon a literature proce-
dure,19 the parent hydrazone derivative 2520 of (1R)-camphor
(5) undergoes reaction with iodine in the presence of triethyl-
amine to give a 1:1 mixture of the desired iodide 26 and the
regioisomeric product 27 (Scheme 3). The literature procedure19
describes the separation of these two products through use of
spinning-band distillation, but we have found the use of column
chromatography employing AgNO3-impregnated silica gel21 to
be a practical option.
of the final product. The overall prediction would therefore be
modest asymmetric induction for the major regioisomer but
much higher asymmetric induction for the minor regioisomer.
Our predictions for each of the preceding systems are
summarized later along with experimental results of actual
catalytic allylation reactions (Table 2).
Synthesis of the Chiral Ligands. The synthesis of 2
proceeds relatively straightforwardly (Scheme 1). The benzo-
bicyclo[2.2.2]octene unit is prepared according to a modification
of the route previously reported by Schmid and Rabai.13 Diels-
Alder reaction of 1,4-benzoquinone with 1-methoxy-1,3-cyclo-
hexadiene affords the tricyclic adduct 18. Reduction with
diisobutylaluminum hydride (DIBALH) produces the allylic diol
19. Aromatization with phosphoryl chloride and pyridine gives
the benzo-fused derivative 20. Hydrogenation of the alkene
bridge provides the benzobicyclo[2.2.2]octene 21. Linking of
21 to phenanthroline is accomplished by methoxy-directed
lithiation14 of 18 with tert-butyllithium to generate the presumed
Lithium/iodide exchange22 of 26 with tert-butyllithium fol-
lowed by addition of the resulting alkenyllithium 24 to phenan-
(15) (a) Gladiali, S.; Chelucci, G.; Soccolini, F.; Delogu, G.; Chessa, G.
J. Organomet. Chem. 1989, 370, 285. (b) Kandzia, C.; Steckhan, E.; Knoch,
F. Tetrahedron: Asymm. 1993, 4, 39.
(16) Chamberlin, A. R.; Stemke, J. E.; Bond, F. T. J. Org. Chem. 1978,
43, 147.
(17) Cusack, N. J.; Reese, C. B.; Risius, A. C.; Roozpeikar, B.
Tetrahedron 1976, 32, 2157.
(18) For related reports of organolithium addition to phenanthroline and
rearomatization with MnO2, see: (a) Dietrich-Buchecker, C. O.; Marnot,
P. A.; Sauvage, J. P. Tetrahedron Lett. 1982, 23, 5291. (b) Lam, F.; Chan,
K. S.; Liu, B.-J. Tetrahedron Lett. 1995, 36, 6261.
(19) Pross, A.; Sternhell, S. Aust. J. Chem. 1970, 23, 989.
(20) Reush, W.; DiCarlo, M. W.; Traynor, L. J. Org. Chem. 1961, 26,
1711.
(13) Schmid, G. H.; Rabai, J. Synthesis 1988, 332.
(14) Shirley, D. A.; Cheng, C. F. J. Organomet. Chem. 1969, 20, 251.
(21) Loev, B.; Bender, P. E.; Smith, R. Synthesis 1973, 6, 362.

4306 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Pen˜a-Cabrera et al.
Table 2. Summary of Predicted Enantioselectivities and Experimental Results of Allylation Reactions
a Unless otherwise noted, the nucleophile was generated from the malonate + NaH. b Obtained using the BSA procedure.
Scheme 3
Scheme 4
throline and treatment with DDQ gives the desired ligand 7 in
30% yield (eq 2). Although the overall yield of 7 is much lower
by this route, this preparation is considerably less expensive
than the first one employing 2,4,6-triisopropylbenzenesulfonyl-
hydrazide (Scheme 2).
8 is obtained in 57% yield (Scheme 4), but if low temperature
is not properly maintained, this compound is again contaminated
by a side product 2915 resulting from alkene double bond
migration/rearomatization.
The spontaneous alkene double bond migration/rearomati-
zation could be used to advantage in the synthesis of the bornyl-
substituted ligand 13 by simply changing the reaction temper-
ature for the addition of the alkenyllithium 24 to phenanthroline.
For incorporation of (1R)-nopinone (6) into the corresponding
ligand 8, we again proceeded via the Shapiro reaction of the
corresponding arenesulfonylhydrazone 28. The desired ligand
(22) Seebach, D.; Neumann, H. Chem. Ber. 1974, 107, 847.

Asymmetric Palladium-Catalyzed Allylation Reactions
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4307
Scheme 5
Figure 16. X-ray structure of 31A. Non-carbon atoms are labeled.
Hydrogens are not shown.
Use of an arenesulfonylhydrazone 23 or the alkenyl iodide 26
as a precursor gives the ligand 13 in overall yields of 50 or
80%, respectively (eqs 3 and 4). In this case, we investigated
the use of the triisopropylbenzenesulfonylhydrazone 23a and
the less expensive p-toluenesulfonylhydrazone 23b, both of
which give 13 in yields of approximately 50%. Furthermore,
p-toluenehydrazone 23b is commercially available as either the
(R)- or the (S)-enantiomer, thus permitting facile synthesis of
either enantiomeric series of ligands. Detailed 1H NMR analysis
of the product 13 reveals the presence of only the endo-bornyl
isomer, which indicates that the alkene double bond migration/
rearomatization is highly stereoselective.
NMR studies of a complex 3 containing the 2-benzobicyclo-
[2.2.2]octenylphenanthroline ligand 2 (Vide supra). The low
3:1 diastereofacial selectivity (Vide supra) prompted our study
of the terpene-derived ligands.
For 2-bornenylphenanthroline (7), the molecular mechanics
calculations predict poor potential utility in asymmetric allylation
reactions. The first experimental examination of this system
was the preparation of η3-1,1-dimethylallyl and η3-1,1-diphen-
ylallyl) complexes 32a and 32b (eq 5). 1H NMR showed them
to exist as 1:1 and 2:1 diastereomer mixtures, respectively, most
likely having their substituted allylic termini remote from the
bornenyl substituent but differing in diastereofacial coordination
of their allyl groups. No attempts were made to distinguish
these isomers. Also, due to their substitution patterns, the
terminal positions of these complexes cannot function as
prochiral centers for nucleophilic addition.
These results together with the NMR studies using the other
chiral ligands are shown in Table 1. The NMR experimental
values agree well with the calculated ratios. The largest error
(19:1 vs. 4:1 for the fourth entry) corresponds to an energy
difference of less than 1 kcal/mol. In other studies, it has
recently been shown that the average error for simple, non-
organometallic, organic compounds with some of the best force
fields in use today is approximately 0.5 kcal/mol.23 A maximum
error of 1 kcal/mol in less fully developed organometallic
systems represents a relatively good level of agreement.
From the perspective of asymmetric allylation reactions, the
results with the 1,3-diphenylallyl, 1-methyl-3-phenyl, and cy-
clohexenyl systems are the most interesting. The results of
studying these allyl systems with the chiral ligand 13 are
especially promising for asymmetric addition reactions. Also,
from the measurements of the diastereomeric ratios of the
complexes, the apparent rotation of the allyl ligands is slow on
The synthesis of 2,9-disubstituted phenanthrolines was briefly
investigated. Addition of bornenyllithium 24 to the monosub-
stituted ligand 7 followed by rearomatization with DDQ gives
the bis(bornenyl)phenanthroline but in a yield of only 15%.
Likewise, the nopinone-derived ligand 8 may be converted into
the corresponding disubstituted phenanthroline but in a yield
of only 6%.
Structural Studies of Palladium Complexes. The feasibility
of forming characterizable η3-allylpalladium complexes contain-
ing phenanthroline ligands described in this paper was first tested
with the benzobicyclo[2.2.2]octenyl derivative 2. Reaction of
bis[η3-allyl(chloro)palladium(II)] 30 with silver(I) tetrafluo-
roborate followed by ligand 2 gives the desired complex 31 in
68% yield as a 1:1 mixture (NMR determination) of diastere-
omeric forms 31A and 31B (Scheme 5). Crystallization from
methylene chloride affords material enriched in 31A as crystals
suitable for X-ray diffraction. Figure 16 depicts the structure
deduced from the X-ray data, because this complex does not
bear substituents at the terminal positions of the allyl group;
however, it was not intended for use in our studies of
asymmetric allylation reactions. Rather, it served as an initial
test case for formation and structural studies of our series of
substituted phenanthroline complexes.
-
-
the NMR time scale when PF6 or BF4 is used as the
counterion.
Catalytic Allylation Reactions. All of the chiral ligands
and allyl systems that had been the subject of calculations were
employed in catalytic allylation reactions using malonate
nucleophiles. The systems that were studied calculationally
provided a wide range of predicted outcomes, varying from very
low levels up to relatively high levels of asymmetric induction.
If these wide-ranging predictions could be confirmed experi-
mentally, then some assurance would be provided for the
1H NMR measurements of the diastereomeric ratios of some
palladium allyl complexes containing our substituted phenan-
throline ligands were performed and compared to the molecular
mechanics-based predictions. We have already described the
(23) Gundertofte, K.; Liljefors, T.; Norrby, P.-O.; Pettersson, I. J. Comput.
Chem. In press.

4308 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Pen˜a-Cabrera et al.
viability of employing this molecular mechanics approach in
additional future efforts directed toward chiral catalyst design.
The conditions of the catalytic allylations are typified by a
control reaction of rac-(E)-4-acetoxy-2-pentene, diethyl sodi-
omethylmalonate, (η3-1,3-dimethylallyl)palladium trifluoroac-
etate 4, and unsubstituted phenanthroline. The product 33 is
obtained, obviously in racemic form, in 84% yield and 100%
consumption of the starting acetate after 14 h at 0 °C (eq 6).
Alternative conditions for generation of the nucleophile entail
in situ treatment of the malonate ester with O,N-bis(trimethyl-
silyl) acetamide (BSA) and potassium acetate or potassium
fluoride, either in the absence or in the presence of 18-crown-
6.24-32
allylation reactions employing our chiral phenanthrolene ligands.
Those systems that had been predicted to give poor asymmetric
induction do, in fact, give zero or very small ee values. Most
gratifying are the two systems that had been predicted to give
significantly higher ee values. Use of the bornylphenanthroline
ligand 13 in the reaction of dimethyl malonate with the 1,3-
diphenylallyl system gives an ee of up to 92% (entry 8). The
(R)-enantiomer is the predominant product as predicted (see
Figure 14). Also of significance is that different levels of
asymmetric induction had been predicted for the cyclohexenyl
system, depending upon the choice of ligand 7 or 13, and again
these predictions are consistent with the experiments (entries
5-7). Very important are the results with the 1-methyl-3-
phenylallyl system for which significantly different outcomes
had been predicted for reactions occurring at the two termini.
Reaction at the methyl-substituted terminus, which is the major
pathway, gives an ee of 33%, whereas reaction at the phenyl-
substituted terminus gives an ee of up to 96% (entry 10).
Although these results are gratifying, an explanation for the
asymmetric induction in this latter case is far from complete.
The substrate for this reaction was the corresponding racemic
allylic acetate. Therefore, if a simple, direct, two-step mech-
anism were invoked, consisting of π-allyl complex formation
and nucleophilic attack with inversion of configuration of the
allylic positions in both steps, the products would be racemic.
However, the two enantiomers of the starting allylic acetate
would afford diastereomeric π-allyl complexes of different
energies, and if stereochemical interconversion were to occur
in this system, net asymmetric induction could result. Various
mechanisms, including π-σ-π rearrangement, anti displacement
of coordinated palladium by external palladium, and reversible
nucleophilic attack, have been proposed previously for stereo-
chemical interconversion of allyl systems.5j-t,8c,41 Although our
results are consistent with a palladium exchange pathway as
proposed in very similar work of others,41 we have insufficient
data to claim a specific stereochemical interconversion pathway
in the present case.
When the experiment is repeated with bornenylphenanthroline
7 as the ligand, one difference is a much slower rate of reaction
which may be due to the steric interaction of the bornenyl
substituent with the metal moiety as discussed above. After
29 h at 0 °C, only 17% of the allylic acetate is consumed, and
the yield of 33 is only 10%. Also, the enantiomeric excess
1
(ee) of the product is 0%, as measured by H NMR using the
chiral shift reagent, tris{3-[heptafluoropropyl)hydroxymeth-
ylene]-(+)-camphorato]europium, Eu(hfc)3.33 This outcome is
also consistent with the very low level of asymmetric induction
predicted by the calculations. A further decrease in reaction
rate is seen upon use of the 2,9-bis(bornenyl)phenanthroline
ligand whereby only 8% of the allylic acetate is consumed after
29 h at 0 °C to give racemic product in a yield of 4%.
Therefore, use of disubstituted ligands was not pursued further.
Emphasis was instead placed upon use of the bornylphenan-
throline ligand 13 for which the previous molecular mechanics
studies had indicated less hindrance toward palladium coordina-
tion. The overall results obtained with the monosubstituted,
terpene-derived phenanthroline ligands are summarized in Table
2 along with the prior predictions of asymmetric induction.
Our results must also be considered in a broader context. Of
the several metal-promoted reactions that have been the subject
of investigations in asymmetric synthesis,2 palladium-catalyzed
allylations5,24-32,35-41 have been under especially intense scru-
tiny. A large number of chiral ligands have been studied with
great success in these reactions (Figure 17). In order to place
our work in proper perspective, we have compiled representative
results in Table 3 for several of these chiral ligands in
comparison with the use of bornylphenanthroline 13.
Our ligands have functioned well for the originally stated
purpose of studying the correlation of predicted and experi-
mental results. This key, initial goal has been met, but we do
not lay claim to superiority of our ligands compared to the best
of the ligands developed previously. However, our results are
at least comparable in some selected cases, and our ligands are
relatively easily prepared in enantiomerically pure form. Also,
in comparison of different conditions for employing malonate
derivatives, our results are consistent with the prior finding of
Discussion
The predicted and the experimentally observed ee values are
in close agreement for the products of the palladium-catalyzed
(24) von Matt, P.; Pfaltz, A. Angew. Chem., Int. Ed. Engl. 1993, 32,
566.
(25) Trost, B. M.; Murphy, D. J. Organometallics 1985, 4, 1143.
(26) Sprinz, J.; Helmchen, G. Tetrahedron Lett. 1993, 34, 1769.
(27) Bovens, M.; Togni, A.; Venazi, L. M. J. Organomet. Chem. 1993,
451, C28.
(28) Tanner, D.; Andersson, P. G.; Harden, A.; Somfai, P. Tetrahedron
Lett. 1994, 35, 4631.
(29) Kubota, H.; Nakajima, M.; Koga, K. Tetrahedron Lett. 1993, 34,
8135.
(30) Kubota, H.; Koga, K. Tetrahedron Lett. 1994, 35, 6689.
(31) Brown, J. M.; Hulmes, D. I.; Guiry, P. J. Tetrahedron 1994, 50,
4493.
(34) Trost, B. M.; Dietsche, T. J. J. Am. Chem. Soc. 1973, 95, 8200.
(35) Okada, Y.; Minami, T.; Sasaki, Y.; Umezu, Y.; Yamaguchi, M.
Tetrahedron Lett. 1990, 31, 3905.
(36) Hayashi, T.; Yamamoto, A.; Hagihara, T.; Ito, Y. Tetrahedron Lett.
1986, 27, 191.
(37) Kang, J.; Cho, W. O.; Cho, H. G. Tetrahedron: Asymm. 1994, 5,
1347.
(32) Abbenhuis, H. C. L.; Burckhardt, U.; Gramlich, V.; Ko¨llner, C.;
Pregosin, P. S.; Salzmann, R.; Togni, A. Organometallics 1995, 14, 759.
(33) (a) Goering, H. L.; Eikenberry, J. N.; Koermer, G. S. J. Am. Chem.
Soc. 1971, 93, 5913. (b) Fraser, R. R.; Petit, M. A.; Saunders, J. K. J. Chem.
Soc., Chem. Commun. 1971, 1450. (c) Wenzel, T. J. NMR Shift Reagents;
CRC Press: Boca Raton, Florida, 1987; pp 127-131.
(38) Yamazaki, A.; Morimoto, T.; Achiwa, K. Tetrahedron: Asymm.
1993, 4, 2287.
(39) Togni, A. Tetrahedron: Asymm. 1991, 2, 683.
(40) Trost, B. M.; Bunt, R. C. J. Am. Chem. Soc. 1994, 116, 4089.
(41) Dawson, G. J.; Williams, J. M. J.; Coote, S. J. Tetrahedron Lett.
1995, 36, 461.

Asymmetric Palladium-Catalyzed Allylation Reactions
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4309
Figure 17. Examples of previously reported chiral ligands for palladium-catalyzed asymmetric allylation reactions.
Table 3. Comparison of Ligand 13 with Other Previously
Reported Chiral Ligands for Palladium-Catalyzed Asymmetric
Allylation Reactions
potential for further development of nitrogen donor ligands in
these reactions.
Conclusion
The studies described in this paper lend support to the notion
that molecular mechanics may be one of various useful tools
for assisting the rational design of asymmetric catalysts. Clearly,
much remains to be done to improve this approach. More
complete development of parameter sets for metal complexes
is an obvious need. At an even more fundamental level, we
must return to key comments in the introduction. The calcu-
lational prediction and experimental detection of relatively stable
intermediates are not sufficient indicators for successful catalyst
design. The reactivity characteristics and the equilibria between
several species must also be taken into account.6 Other concerns
may again be raised about the effects of other reaction
parameters such as solvents, counterions, and specific selections
of bases, nucleophiles, and leaving groups employed in these
reactions.
With the realization that our present results are only a first
step in this direction, we are actively pursuing some of these
further considerations. We are reporting separately our next
step which takes into account the intrinsic reactivity of distorted
allyls and the steric demands of an approaching nucleophile.7
However, all of the above concerns must be addressed if the
molecular mechanics approach is to be developed into a
generally useful tool for catalyst design. Chiral ligands are of
importance in many other synthetically useful, palladium-
catalyzed reactions including couplings (e.g., the Stille42 and
the Suzuki43 reactions), vinylations (e.g., the Heck reaction44 ),
heteronucleophilic additions to alkenes,45 3 + 2 cycloadditions,46
and alkene polymerizations.47 Also, nitrogen donor ligands are
of use in complexes involved in other reactions such as the well-
established alkene oxidations promoted by osmium48 and
(43) (a) Suzuki, A. Pure Appl. Chem. 1991, 63, 419. (b) Martin, A. R.;
Yang, Y. Acta Chem. Scand. 1993, 47, 221.
(44) (a) Heck, R. F. Org. React. 1982, 27, 345. (b) Heck, R. F. In
ComprehensiVe Organic Synthesis; Trost, B. M., Fleming, I., Semmelhack,
M. F., Eds.; Pergamon: Oxford, 1991; Vol. 4, pp 833-863. (c) Cabri, W.;
Candiani, I. Acc. Chem. Res. 1995, 28, 2.
(45) (a) Gasc, M. B.; Lattes, A.; Perie, J. J. Tetrahedron 1983, 39, 703.
(b) Hegedus, L. S. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Fleming, I., Semmelhack, M. F., Eds.; Pergamon: Oxford, 1991; Vol. 4,
pp 551-569.
(46) (a) Trost, B. M. Angew. Chem., Int. Ed. Engl. 1986, 25, 1. (b) Trost,
B. M.; Grese, T. A.; Chan, D. M. T. J. Am. Chem. Soc. 1991, 113, 7350.
(c) Lautens, M.; Ren, Y.; Delanghe, P. H. M. J. Am. Chem. Soc. 1994,
116, 8821.
Trost and others that the use of the alternative BSA conditions
gives superior ee values compared to more traditional con-
ditions.24-32 We have not employed the tetraalkylammonium
malonate derivatives that have also been reported to give
improved results.40
Another point is that in the past, most ligands used in these
reactions were phosphine or phosphite derivatives, but several
nitrogen-based ligands 38-48 have been developed more
recently. Our results give an additional indication of the
(42) (a) Stille, J. K. Pure Appl. Chem. 1985, 57, 1771. (b) Stille, J. K.
Angew. Chem., Int. Ed. En gl. 1986, 25, 508. (c) Mitchell, T. N. Synthesis
1992, 803.
(47) Brookhart, M.; Wagner, M. I.; Balavoine, G. G. A.; Haddou, H. A.
J. Am. Chem. Soc. 1994, 116, 3641.

4310 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Pen˜a-Cabrera et al.
manganese compounds,49 the rich organometallic chemistry of
copper,50 cobalt-promoted alkylations,51 iron-mediated triene
carbocyclizations,52 and nickel-catalyzed conjugate additions to
enones.53 These and many other reactions will continue to
attract the attention of investigators concerned with the rational
design of methods for asymmetric synthesis.
ferrous sulfate. Ethyl acetate was distilled from calcium chloride.
Tetramethylethylenediamine (TMEDA) was freshly distilled under
nitrogen from sodium. Hexane was distilled under nitrogen from
calcium hydride. Triethylamine, pyridine, and diisopropylamine were
distilled under nitrogen from sodium hydroxide. 1,10-Phenanthroline
was heated at 60 °C under vacuum (0.01 mm Hg) for 12 h and was
stored under nitrogen. The racemic allylic acetates were obtained by
acetylating the alcohols with acetic anhydride under standard conditions.
All other chemicals were purchased from Aldrich Chemical Co. unless
otherwise stated. 1-Acetoxy-2-cyclohexene was kindly donated by Dr.
Sverker Hansson (Royal Institute of Technology). (1,2,3-η)-Prope-
nylpalladium chloride dimer,55 bis(µ-trifluoroactetato)bis[(2,3,4-η)-3-
pentenyl]dipalladium,56 and Pd(dba)257 were synthesized using literature
procedures.
Experimental Section
General Methods. All manipulations of air-sensitive compounds
were performed in oven-dried glassware under a nitrogen atmosphere.
Solutions were transferred with hypodermic syringes or with double-
ended needles (cannulas). GC analyses were performed on a Hewlett-
Packard 5890A chromatograph equipped with a capillary column
(methyl silicone gum, 25 m × 0.31 mm). HPLC analyses were
performed using an ISCO Model 2360 gradient programmer, a Model
2350 pump, a V4 variable wavelength UV detector, and a Brownlee
silica Spheri-5 HPLC analytical 4.6 mm x 22 cm cartridge column.
Flash column chromatography and MPLC were performed using EM
Science silica gel 60 (230-400 mesh) or Aldrich neutral aluminum
oxide, Brockmann I (150 mesh). MPLC purifications were performed
as described by Baeckstro¨m et al.54 Analytical TLC was performed
using Whatman AL SIL G/UV(254) silica gel plates (250 µm thick)
or E. Merck 60 F254 neutral (type E) aluminum oxide sheets (0.2 mm
thick). Preparative TLC was performed using Analtech silica gel thin-
layer chromatography plates (250 µm) or E. Merck 60 F254 neutral (type
E) aluminum oxide sheets (0.2 mm thick).
1-Methoxy-1,2,3,4-tetrahydro-1,4-ethanonaphthalene (21). 1,4-
Benzoquinone (2.0 g, 18,5 mmol) and 1-methoxy-1,3-cyclohexadiene
(1.9 g, 17.6 mmol) in CH2Cl2 (50 mL) were stirred for 12 h at 25 °C.
The solvent was then evaporated to give a dark solid which was stirred
in CCl
4. The mixture was passed through a sintered glass funnel, the
solvent was removed under reduced pressure, and the yellow solid was
recrystallized from EtOAc/pentane to give 18 (2.3 g, 60.2%) as pale
yellow needles: mp 118-119 °C. Anal. Calcd for C13H14O3: C, 71.53;
H, 6.47. Found: C, 71.67; H, 6.58. To a solution of 18 (5.0 g, 22.9
mmol) in CH2Cl2 (50 mL) at 0 °C was added a solution of DIBAL
(48.16 mmol) via syringe. The mixture was stirred at 25 °C for 1 h
and was quenched with 1:1 H2O/MeOH and aqueous sodium potassium
tartrate. The mixture was extracted with ether, and the extracts were
dried, treated with activated charcoal, and concentrated. Recrystalli-
zation from EtOAc gave 19 (5.03 g, 99%) as colorless prisms: mp
128-129 °C. A solution of 19 (3.68 g, 16.6 mmol), pyridine (40 mL),
and phosphoryl chloride (6.8 mL, 73.2 mmol) at 0 °C was warmed to
25 °C. After completion of the reaction (TLC), the mixture was poured
into ice-water and extracted with ether. The extracts were washed
with 5% HCl and water, treated with activated charcoal, dried,
concentrated, and purified by column chromatography (silica gel, 2%
EtOAc/pentane) to give 20 (2.76 g, 90%) as a colorless oil. HRMS
m/e calcd for C11H10O (M+ - C2H4) 158.0732, found 158.0736 (M+
- C2H4). Anal. Calcd for C13H14O: C, 83.82; H, 7.58. Found: C,
84.00; H, 7.74. 20 (0.173 g, 0.93 mmol) and 10% Pd/C (0.1 g) in
EtOAc were stirred under hydrogen (1 atm) at 25 °C and monitored
by GC. The mixture was passed through a cotton plug. The filtrate
was concentrated to give 21 (0.174 g, 99%) as a colorless oil: 1H NMR
(300 MHz) δ 7.5-7.0 (m, ArH), 3.53 (s, CH3O), 2.93 (m, bridgehead
Proton (1H) nuclear magnetic resonance (NMR) spectra were
obtained with a Magnachem A-200 (200 MHz), a Bruker ACF-250
(250 MHz), a General Electric GN-300 (300 MHz), or a Bruker AM-
400 (400 MHz) spectrometer. Carbon (13C) NMR spectra were
recorded on the Bruker or General Electric spectrometers at 62.5, 75,
or 100 MHz. CDCl3 containing 5% (v/v) TMS was the solvent unless
otherwise indicated. NMR assignments were made with both COSY
and homodecoupling programs. The subscript P is used to identify the
protons in the phenanthroline ligands according to following numbering
scheme:
H), 2.0-1.8 (m, CH3OC(CH2)2), 1.6-1.4 (m, CH3OC(CH2CH2)2); 13
C
NMR (75 MHz) δ 143.38, 142.50, 125.90, 125.77, 123.21, 120.15 (Ar),
76.58 (COCH3), 51.17 (COCH3), 33.71 (CH), 29.25 (CH3OCCH2),
26.64 (CH3OCCH2CH2); IR (neat) 3050 (Ar), 2950, 2850 (CH), 1330
. HRMS m/e calcd for C13H16O 188.1201, found
188.1187. Anal. Calcd for C13H16O: C, 82.93; H, 8.57. Found: C,
82.79; H, 8.56.
Infrared spectra were recorded on a Perkin-Elmer Model 1420
spectrophotometer. Mass spectral data were obtained on a Finnigan
MAT 8430 spectrometer. HRMS data were obtained by electron
impact, and low resolution data were obtained by isobutane chemical
ionization. Optical rotations were determined with a Rudolph Research
Autopol III polarimeter at 546 nm. Elemental analyses were performed
by M-H-W laboratories (Phoenix, AZ) and are reported when they agree
with the calculated values within (0.4%. Melting points were
determined in Pyrex capillaries with a Thomas Hoover Unimelt
apparatus and are corrected.
Anhydrous THF, ethyl ether, and dimethoxyethane (DME) were
freshly distilled under nitrogen from dark blue or purple solutions of
sodium-benzophenone ketyl/dianion. Hexane and CH2Cl2 used for
chromatography and workup were distilled form calcium hydride.
Pentane (reagent grade) utilized for chromatography was used directly
without purification. Ethyl ether used for workup was distilled from
(C-O-C) cm-1
2-[(1-Methoxy-1,2,3,4-tetrahydro-1,4-ethanonaphthalen-8-yl)]-
1,10-phenanthroline (rac-2). To a solution of t-BuLi (5.3 mmol) and
TMEDA (0.62 g, 5.3 mmol) in hexane (1.0 mL) at 25 °C under nitrogen
was added 21 (0.5 g, 2.7 mmol) over 30 min. The lithiation stopped
at 56% conversion (GC analysis of an aliquot methylated with CH3I).
The solvent was then evaporated to give a dark oil which was dissolved
in THF (2.0 mL) and added to 1,10-phenanthroline (0.96 g, 5.3 mmol)
in 1:1 ether/THF (10.0 mL) at 25 °C. After 12 h, DDQ (1.2 g, 5.3
mmol) in DME (5.0 mL) was added at 25 °C. After 12 h, the reaction
was quenched with aqueous NH4Cl. The CH2Cl2 extracts were passed
Celite, dried, concentrated, and purified by MPLC (neutral alumina,
EtOAc/hexane gradient) to give rac-2 (0.36 g, 37%) which was
(48) (a) Kolb, H. C.; Andersson, P. G.; Sharpless, K. B. J. Am. Chem.
Soc. 1994, 116, 1278. (b) Norrby, P.-O.; Kolb, H. C.; Sharpless, K. B. J.
Am. Chem. Soc. 1994, 116, 8470.
(49) (a) Palucki, M.; Pospisil, P. J.; Zhang, W.; Jacobsen, E. N. J. Am.
Chem. Soc. 1994, 116, 9333. (b) Linde, C.; Åkermark, B.; Norrby, P.-O. J.
Am. Chem. Soc. In press.
(50) Lipshutz, B. H.; Sengupta, S. Org. React. 1992, 41, 135-631.
(51) Bhatia, B.; Reddy, M. M.; Iqbal, J. Tetrahedron Lett. 1993, 34, 6301.
(52) Takacs, J. M.; Myoung, Y.-C.; Anderson, L. G. J. Org. Chem. 1994,
59, 6928.
(53) Bolm, C.; Ewald, M. Tetrahedron Lett. 1990, 31, 5011.
(54) Baeckstro¨m, P.; Stridh, K.; Li, L.; Norin, T. Acta Chem. Scand.
1987, B41, 442.
1
recrystallized from EtOAc/pentane: mp 219-222 °C; H NMR (300
MHz) δ 9.2 (dd, J ) 4.3, 1.8 Hz, HP9), 8.25 (dd, J ) 8.1, 1.8 Hz, HP7),
8.1 (d, J ) 8.3 Hz, HP4), 7.9-7.7 (two d, J P5,6 ) J P6,5 ) 8.8 Hz, HP5
and HP6), 7.8 (d, J ) 8.3 Hz, HP3), 7.6 (dd, J ) 8.1, 4.3 Hz, HP8), 7.5
(dd, J ) 7.6, 1.5 Hz, Hp), 7.25 (dd, J ) 7.5, 7.3 Hz, Hm), 7.2 (dd, J )
(55) (a) Åkermark, B.; Åkermark, G.; Hegedus, L. S.; Zetterberg, K. J.
Am. Chem. Soc. 1981, 103, 3037. (b) Auburn, P. R.; Mackenzie, P. B.;
Bosnich, B. J. Am. Chem. Soc. 1985, 107. 2033.
(56) Powell, J. J. Am. Chem. Soc. 1969, 91, 4311.
(57) Rettig, M. F.; Maitlis, P. M. Inorg. Synth. 1977, 17, 134.

Asymmetric Palladium-Catalyzed Allylation Reactions
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4311
7.4, 1.5 Hz, Ho), 3.05 (bs, bridgehead H), 2.5 (s, CH3O), 2-1.5 (broad
shoulder, 4 CH2); 13C NMR (75 MHz) δ 162.02, 150.23, 146.49, 145.02,
143.43, 140.10, 136.66, 135.84, 132.77, 130.33, 128.66, 127.12, 126.86,
126.48, 125.80, 125.78, 123.89, 122.47 (phen and Ar), 78.45 (COCH3),
50.65 (COCH3), 34.59 (CH), 32.93, 28.42, 26.81, 25.56 (bs CH2); IR
150.08, 146.36, 145.00, 135.60, 134.90, 128.57, 126.70, 126.20, 125.26,
122.64, 122.24 (phen), 52.76, 50.84 (C-2 and C-3), 50.29 (C-7), 45.60
(C-4), 33.74 (C-1), 29.24, 28.27 (C-5 and C-6), 19.47, 18.93, 14.57
(methyl groups); IR (CH2Cl2) 3060 (Ar), 2900 (CH), 900 (C-H Ar)
cm-1. HRMS m/e calcd for C22H24N2 316.1940, found 316.1954. Anal.
Calcd for C22H24N2: C, 83.49; H, 7.65. Found: C, 83.26; H, 7.85.
Method II. 26 (0.5 g, 1.9 mmol) in 1:1 THF/ether (10 mL), t-BuLi
(4.2 mmol), and phenanthroline (0.17 g, 0.95 mmol) in hexane (10
mL) gave 13 (0.24 g, 80%) as a yellow crystalline solid.
(CH2Cl2) 3060 (Ar), 2900 (CH), 1250 (C-O-C), 900 (C-H Ar) cm-1
.
HRMS m/e calcd for C25H22N2O 366.1732, found 366.1703. Anal.
Calcd for C25H22N2O: C, 81.83; H, 6.06. Found: C, 81.81; H, 6.16.
(1R)-2-Iodo-1,7,7-trimethylbicyclo[2.2.1]hept-2-ene (26) and 2,2-
Dimethyl-3-methylene-4-iodobicyclo[2.2.1]heptene (27).21 To a (1R)-
1,7,7-trimethylbicyclo[2.2.1]heptan-2-one hydrazone (25)20 (7.0 g, 42.1
mmol), Et3N (10.0 mL), and ether (50.0 mL) was added dropwise iodine
in ether until the yellow color persisted for about 30 s. The crude
product consisted of a 1:1 mixture (8.0 g, 72% combined yield) of 26
and 27 which was separated by column chromatography (silica gel/
10% silver nitrate,21 pentane) to give 26 (3.2 g, 29%) as a pale yellow
oil: 1H NMR (300 MHz) δ 6.4 (d, J ) 3.4 Hz, CH)C), 2.36 (t, J )
3.4 Hz, H-4), 1-2 (m, 4 H, ring protons), 0.96 (s, CH3 at C-1), 0.85
(s, CH3 at C-7), 0.84 (s, CH3 at C-7) (lit.21 1H NMR); 27 (3.5 g, 43%)
as a pale yellow oil: 1H NMR (300 MHz) δ 5.17 (s, CdCHH), 4.82
(s, CdCHH), 2.5-1.5 (m, 7 H, ring protons), 1.13 (s, CH3), 1.08 (s,
CH3) (lit.21 1H NMR).
2-[(1R)-1,7,7-Trimethylbicyclo[2.2.1]hept-2-en-2-yl]-1,10-phenan-
throline (7). Method I.16 To a suspension of triisopropylbenzene-
sulfonylhydrazone 23a (1.0 g, 2.3 mmol) in 1:1 hexane/TMEDA (20.0
mL) at -55 °C was added a solution of s-BuLi (5.1 mmol) dropwise.
The solution was stirred at -55 °C for 1.5 h and at 0 °C for 30 min
and was then added to a suspension of 1,10-phenanthroline (0.19 g,
1.04 mmol) in hexane (5.0 mL) at -78 °C. After 12 h, DDQ (0.52 g,
2.3 mmol) in DME (5.0 mL) was added at -50 °C. After 6 h at 25
°C, the reaction was quenched with aqueous NH4Cl solution and the
CH2Cl2 extracts were passed through Celite, dried, concentrated, and
purified by MPLC (neutral alumina, EtOAc/hexane gradient) to give 7
(0.17 g, 52%) as a yellow crystalline solid which was recrystallized
from EtOAc/pentane: mp 182-184 °C; [R]25D -129.4° (c 0.05, CH2-
Cl2); 1H NMR (300 MHz) δ 9.2 (dd, J ) 4.3, 1.7 Hz, HP9), 8.2 (dd, J
) 8.1, 1.7 Hz, HP7), 8.1 (d, J ) 8.4 Hz, HP4), 7.75-7.70 (two d, J )
8.8 Hz, HP5 and HP6), 7.72 (d, J ) 8.4 Hz, HP3), 7.6 (dd, J ) 8.0, 4.3
Hz, HP8), 6.8 (d, J ) 3.4 Hz, H-3), 2.51 (app. t, J ) 3.4 Hz, H-4), 2.0
(m, H-5exo), 1.8 (m, H-6endo), 1.6 (m, H-5endo), 1.56 (s, CH3 at C-1),
1.15 (m, H-6exo), 0.96 (s, CH3 at C-7), 0.90 (s, CH3 at C-7); 13C NMR
(75 MHz) δ 157.08, 150.28, 149.41, 146.69, 145.93, 137.66, 126.96,
126.91, 126.42, 125.42, 122.49, 121.01 (phen), 135.83, 135.49 (C-2
and C-3), 57.00, 55.12, 52.06 (C-1, C-4, and C-7), 31.89, 25.76 (C-5
and C-6), 19.92 (CH3 at C-7), 19.65 (CH3 at C-7), 13.02 (CH3 at C-1);
IR (CH2Cl2) 3030 (Ar, CddC), 2980 (C-H), 1420 (CH) cm-1. HRMS
m/e calcd for C22H22N2 314.1783, found 314.1775. Anal. Calcd for
C22H22N2: C, 84.03; H, 7.06. Found: C, 84.31; H, 7.13.
2-{(1R)-6,6-Dimethylbicyclo[3.1.1]hept-2-en-2-yl}-1,10-phenan-
throline (8). Hydrazone 28 was obtained from reaction of 2,4,6-
triisopropylbenzenesulfonylhydrazide (5.2 g, 17.4 mmol), (1R)-6,6-
dimethylbicyclo-[3.1.1]heptan-2-one (2.2 g, 15.8 mmol), and concentrated
aqueous HCl (1.5 mL) in 3:1 CH3CN/THF (40.0 mL) at 25 °C for 14
h to give 4.0 g (61%) of colorless crystals: mp 162-163 °C (dec.).
HRMS m/e calcd for C24H38N2O2S 419.2748, found 419.2740. Anal.
Calcd for C24H38N2O2S: C, 68.86; H, 9.16. Found: C, 68.59; H, 8.92.
By use of Method I, 28 (0.5 g, 1.2 mmol) in 1:1 hexane/TMEDA (10.0
mL), s-BuLi (2.6 mmol), phenanthroline (0.14 g, 0.79 mmol) in hexane
(5.0 mL), and DDQ (0.27 g, 1.2 mmol) in DME (3.0 mL) and
purification with MPLC (neutral alumina, EtOAc/hexane gradient) gave
8 (0.14 g, 57%) as a yellow crystalline solid which was recrystallized
from EtOAc/pentane: mp 169-171 °C; [R]25D +25° (c 0.04, CH2Cl2);
1H NMR (300 MHz) δ 9.2 (dd, J ) 4.4, 1.8 Hz, HP9), 8.2 (dd, J ) 7.8,
1.8 Hz, HP7), 8.1 (d, J ) 8.3 Hz, HP4), 7.75 (d, J ) 8.6 Hz, HP3), 7.7
(two d, J ) 8.7 Hz, HP5 and HP6), 7.55 (dd, J ) 8.4, 4.4 Hz, HP8), 6.91
(m, H-3), 3.5 (dt, J ) 5.6, 1.6 Hz, H-1), 2.65-2.5 (m, H-7exo, H-4exo
,
H-4endo), 1.45 (s, CH3 at C-6), 1.42 (d, J ) 8.6 Hz, H-7endo), 0.92 (s,
CH3 at C-6); 13C NMR (75 MHz) δ 157.33, 150.16, 147.28, 146.34,
135.97, 135.91, 129.00, 126.89, 126.28, 125.43, 122.47 (phen), 145.60
and 118.91 (C-2 and C-3), 42.47 (C-1), 40.62 (C-5), 38.04 (C-6), 32.25,
31.68 (C-7 and C-4), 26.31 (CH3 at C-6), 21.02 (CH3 at C-6); IR (CH2-
Cl2) 3030 (Ar and CdC), 2980 (CH), 1420 (CH) cm-1. HRMS m/e
calcd for C21H20N2 301.1705, found 301.1730. Anal. Calcd for
C21H20N2: C, 83.95; H, 6.72. Found: C, 83.76; H, 6.77.
2,9-bis-[(1R)-1,7,7-Trimethylbicyclo[2.2.1]hept-2-en-2-yl]-1,10-
phenanthroline. Method I. The alkenyllithium from 23a (0.21g, 0.48
mmol) in 1:1 hexane/TMEDA (8.0 mL) and s-BuLi (1.0 mmol) was
stirred with 7 (0.1 g, 0.32 mmol) in hexane (5.0 mL) at -78 °C for 16
h followed by DDQ (0.11 g, 0.48 mmol) in DME (3.0 mL) at 25 °C
for 4 h. Purification with MPLC (neutral alumina, ethyl acetate/hexane
gradient) gave the disubstituted phenanthroline (0.017 g, 17%) as a
pale yellow crystalline solid which was recrystallized from ethyl acetate/
pentane: mp 169-171 °C; [R]25D -333.3° (c 0.03, CH2Cl2); 1H NMR
δ 8.03 (d, J ) 8.4 Hz, HP3 and HP8), 7.71 (d, J ) 8.4 Hz, HP4 and HP7),
7.61 (s, HP5 and HP6), 6.56 (d, J ) 3.3 Hz, H-3), 2.52 (app. t, J ) 3.4
Hz, H-4), 2.00 (m, H-5exo), 1.9-1.5 (m, H-6endo, H-5endo), 1.75 (s, CH3
at C-1), 1.3-1.1 (m, H-6exo), 0.96 (s, CH3 at C-7), 0.93 (s, CH3 at
C-7); 13C NMR (75 MHz) δ 157.03, 150.22, 145.71, 126.91, 125.18,
120.56 (phen), 137.29 and 135.32 (C-2 and C-3), 56.21 (C-7), 55.52
(C-1), 52.18 (C-4), 31.51 (C-6), 26.08 (C-5), 19.77 (CH3 at C-7), 19.58
(CH3 at C-7), 12.53 (CH3 at C-1); IR (CH2Cl2) 3030 (Ar, CdC), 2980
(C-H), 1420 (CH) cm-1. HRMS m/e calcd for C32H36N2 448.2878,
found 448.2873.
Method II.22 A solution of 26 (0.5 g, 1.9 mmol), 4:1 THF/ether
(10 mL), and t-BuLi (4.2 mmol) was stirred at -78 °C for 1 h and
was added to phenanthroline (0.17 g, 0.95 mmol) in 4:1 THF/ether
(10 mL) at -78 °C . After 7 h, DDQ (0.43 g, 1.9 mmol) in DME (2.0
mL) was added at -50 °C. After 6 h at 25 °C, product isolation as in
Method I gave 7 (0.09 g, 30%) as a yellow crystalline solid.
Use of Method II with 26, t-BuLi, 7, and DDQ gave the disubstituted
phenanthroline (14%) as a yellow crystalline solid.
2-[(1R)-2-endo-1,7,7-Trimethylbicyclo[2.2.1]hept-2-yl]-1,10-phenan-
throline (13). Using Method I as above, the alkenyllithium from 23a
(1.0 g, 2.3 mmol) s-BuLi (5.1 mmol) was stirred with 1,10-phenan-
throline (0.210 g, 1.15 mmol) in hexane (20.0 mL) at 25 °C for 60 h.
Purification with MPLC (neutral alumina, EtOAc/hexane gradient) gave
13 (0.19 g, 52%) as a yellow crystalline solid which was recrystallized
from EtOAc/pentane. Alternatively, (R)-camphor p-toluenesulfonyl-
hydrazone (1.25 g, 3.90 mmol) in 1:1 hexane/TMEDA (10 mL) and a
cyclohexane solution of s-BuLi (7.8 mmol) followed by reaction with
phenanthroline (0.345 g, 1.91 mmol) in hexane (10 mL) gave 13 (0.29
g, 48%) as a yellow crystalline solid: mp 157-159 °C; [R]25D -64.5°
Simple Bis(µ-chloro)bis[η3-allyl]dipalladium Complexes.55 Bis-
(µ-chloro)bis[(1,2,3-η)-1-methyl-3-phenyl-2-propenyl]dipalla-
dium. Trifluoroacetic anhydride (920 mg, 4.38 mmol) was added to
trans-4-phenyl-3-buten-2-ol58 (500 mg, 3.37 mmol) and ether (35 mL)
at 0 °C. The mixture was stirred at 25 °C for 12 h, and the ether was
removed by rotary evaporation. Pd(dibenzylideneacetone)2 (1.4 g, 2.55
mmol), acetonitrile (9 mL), and THF (36 mL) were added, and the
mixture was stirred at 25 °C for 15 min. The mixture was extracted
with 90:10 H2O:CH3CN followed by filtration and rotary evaporation.
The residue was dissolved in anhyd acetone (50 mL), and LiCl (340
mg, 8 mmol) was added. The reaction mixture was stirred at 25 °C
for 30 min and was then filtered though silica gel and evaporated to
give 348 mg (50%) of yellow solid: 1H NMR (400 MHz, CDCl3): δ
7.49 (d, J ) 7.3 Hz, 2 H), 7.31 (m, 1 H), 7.25 (m, 2 H), 5.64 (app t,
J ) 10.7 Hz, 1 H), 4.41 (d, J ) 10.7 Hz, 1 H), 3.94 (m, 1 H), 1.29 (d,
J ) 6.2 Hz, 3 H).
1
(c 0.03, CH2Cl2); H NMR (300 MHz) δ 9.2 (dd, J ) 4.3, 1.8 Hz,
HP9), 8.15 (dd, J ) 8.1, 1.8 Hz, HP7), 8.1 (d, J ) 8.4 Hz, HP4), 7.70-
7.65 (two d, J ) 8.8 Hz, HP5 and HP6), 7.6 (d, J ) 8.4 Hz, HP3), 7.55
(dd, J ) 8.0, 4.4 Hz, HP8), 3.8 (ddd, J ) 11.5, 5.3, 2.4 Hz, H-2exo), 2.3
(m, H-3exo), 2.05 (dd, J ) 13.1, 5.3 Hz, H-3endo) 1.85 (m, H-4 and
H-5endo), 1.60 (m, H-5exo), 1.35 (m, H-6endo), 1.25 (m, H-6exo), 1.12 (s,
CH3 at C-1), 0.95 (s, 2 CH3 at C-7); 13C NMR (75 MHz) δ 163.81,

4312 J. Am. Chem. Soc., Vol. 118, No. 18, 1996
Pen˜a-Cabrera et al.
Bis(µ-chloro)bis[(1,2,3-η)-cyclohexenyl]dipalladium. By the above
procedure, 2-cyclohexen-1-ol gave a 20% yield: 1H NMR (250 MHz,
CDCl3): δ 5.48 (app t, J ) 6.0 Hz, 1 H), 5.18 (app t, J ) 6.0 Hz, 2
H), 1.90-1.68 (br m, 4 H), 1.11-1.00 (br m, 2 H).
Bis(µ-chloro)bis[(1,2,3-η)-1,3-diphenyl-2-propenyl]dipalla-
dium. By the above procedure, 1,3-diphenyl-2-propen-1-ol gave a 60%
yield as a yellow solid: 1H NMR (400 MHz, CDCl3): δ 7.64-7.26
(br m, 10 H), 4.67 (br s, 1 H), 4.15 (br s, 2 H).
J ) 8.6 Hz, 1 H, HP4), 7.96 (d, J ) 8.8 Hz, 1 HP5 or 6), 7.93 (d, J ) 8.7
Hz, HP5 or 6), 7.79 (d, J ) 8.6 Hz, HP3), 7.54 (d, J ) 7.9 Hz, 1 H, HP),
6.65 (d, J ) 3.3 Hz, 1 H), 6.22 (dd, J2,1a ) 12.7 Hz, J2,1s ) 7.6 Hz, 1
H, H2), 3.54 (d, 1 H, H1a), 2.53 (t, J ) 3.5 Hz, 1 H), 1.79 (m, 1 H),
1.43 (m, 1 H). Unassigned signals: 7.51-7.45 (br m), 7.38-7.24 (br
m), 7.16 (t, J ) 7.8 Hz), 7.08 (m), 3.62 (d, J ) 5.9 Hz), 2.04 (m), 1.11
(m), 0.90-0.70 (br m).
{2-[(1R)-1,7,7-Trimethylbicyclo[2.2.1]hept-2-en-2-yl]-1,10-phenan-
throline}[1,1-dimethyl-(1,2,3-η)-2-propenyl]palladium Tetrafluo-
roborate. Obtained as a 1:1 mixture of diastereomers. Major
isomer: 1H NMR (400 MHz, CDCl3): δ 9.11 (dd, JP9,8 ) 5.0 Hz, 1 H,
JP9,7 ) 1.4 Hz, HP9), 8.68 (dd, JP7,8 ) 8.3 Hz, 1 H, HP7), 8.55 (d, JP4,3
) 8.6 Hz, 1 H, HP4), 8.19 (dd, 1 H, HP8), 8.02 (br s, 2 H, HP5,6), 7.87
(d, 1 H, HP3), 6.93 (d, J ) 3.3 Hz, 1 H), 5.26 (dd, J2,1a ) 13.0 Hz, J2,1s
) 7.6 Hz, 1 H, H2), 4.19 (dd, J1s,2 ) 7.6 Hz, J1s,1a ) 1.1 Hz, 1 H, H1s),
3.46 (dd, 1 H, H1a), 2.75 (t, J ) 3.3 Hz, 1 H). Minor isomer: 9.08
(dd, JP9,8 ) 5.0 Hz, JP9,7 ) 1.4 Hz, 1 H, HP9), 8.71 (dd, J P7,8 ) 8.2, 1
{2-[(1-Methoxy-1,2,3,4-tetrahydro-1,4-ethanonaphthalen-8-yl)]-
1,10-phenanthroline}[(1,2,3-η)-propenyl]palladium(II) Tetrafluo-
roborate (31). (1,2,3-η)-Propenylpalladium58 chloride dimer (0.006
g, 0.026 mmol) in CH2Cl2 (0.5 mL) was added to a suspension of AgBF4
(0.0062 g, 0.032 mmol) in CH2Cl2 (0.5 mL) at 0 °C. After 20 min,
rac-2 (0.012 g, 0.032 mmol) in CH2Cl2 (0.5 mL) was added. After 30
min at 25 °C, the mixture was passed through a cotton plug and
magnesium sulfate. Evaporation gave 0.014 g (68%) of off-white solid
whose 1H NMR spectrum showed a 1:1 mixture of diastereomers.
Recrystallization from CH2Cl2/hexane gave 31 as colorless prisms: mp
H, HP7), 8.56 (d, J
)8.5 Hz, 1 H, HP4), 8.18 (dd, 1 H, HP8), 8.04
P4,3
1
> 230 °C; H NMR (300 MHz, CD2Cl2) δ 9.15 (dd, J ) 5.0, 1.5 Hz,
(br s, 2 H, HP5,6), 7.86 (d, J ) 8.5 Hz, 1 H, HP3), 6.86 (d, J ) 3.4 Hz,
1 H), 5.47 (dd, J ) 13.4 Hz, J ) 7.7 Hz, 1 H, H2), 4.32 (dd, J
1 H, HP9), 8.72 (dd, J ) 8.3, 1.4 Hz, 1 H, HP7), 8.5 (app. t, J ) 8.2 Hz,
1 H, HP4), 8.2-8.0 (two d, J ) 8.8 Hz, 2 H, HP5 and HP6), 8.0-7.9 (m,
2 H, HP8 and HP3), 7.5-7.0 (m, 3 H, Ar-H), 5.62 (m, 1 H, central
allylic H, one diastereomer), 5.49 (m, 1 H, central allylic H, other
diastereomer), 4.3 (bd, J ) 7.0 Hz, 2 H, Hsyn), 3.55 (d, J ) 12.1 Hz,
2 H, Hanti, one diastereomer), 3.45 (d, J ) 12.3 Hz, 2 H, Hanti, other
diastereomer), 3.15 (m, 1 H, bridgehead H), 2.78 (s, 3 H, CH3O, one
diastereomer), 2.62 (s, 3 H, CH3O, other diastereomer), 2.3-1.75 (m,
4 H, CH3OC(CH2)2), 1.75-1.2 (m, 4 H, CH3OC(CH2CH2)2); IR (CH2-
Cl2) 3060 (ArH), 2900 (CH), 1420 (CdC allylic), 1250 (C-O-C), 1050
(C)C allylic), 900 (CH Ar) cm-1. HRMS (FAB)m/e calcd for the
cation C28H27N2OPd 513.1169, found 513.1167.
2,1a
2,1s
1s,1a ) 1.0 Hz, 1 H, H1s), 3.46 (dd, 1 H, H1a), 2.75 (t, J ) 3.4 Hz, 1 H).
Unassigned signals: 2.13 (m), 1.85 (m), 1.82 (s), 1.74 (s), 1.53 (m),
1.50 (s), 1.42 (m), 1.27 (s), 1.26 (s), 1.24 (s), 1.17 (s), 1.06 (s), 1.03
(s), 0.94 (d, J ) 5.1 Hz), 0.87 (m).
{2-[(1R)-6,6-Dimethylbicyclo[3.1.1]hept-2-en-2-yl]-1,10-phenan-
throline}[1,1-dimethyl-(1,2,3-η)-2-propenyl]palladium Hexafluoro-
phosphate. Obtained as a 2:1 mixture of diastereomers. Major
isomer: 1H NMR (400 MHz, CDCl3): δ 9.03 (d, J P9,8 ) 4.3 Hz, 1 H,
HP9), 8.64 (d, J P7,8 ) 8.2 Hz, 1 H, HP7), 8.56 (d, J P4,3 ) 8.5 Hz, 1 H,
HP4), 8.10 (dd, 1 H, HP8), 7.98 (m, 2 H, HP5,6), 7.65 (d, 1 H, HP3), 5.23
(dd, J2,1a ) 12.9 Hz, J2,1s ) 7.6 Hz, 1 H, H2), 4.06 (dd, 1 H, J1s,1a
)
X-ray Structure Determination of {2-[(Methoxy-1,2,3,4-tetrahy-
dro-1,4-ethanonaphthalen-8-yl)]-1,10-phenanthroline}[(1,2,3-η)-pro-
penyl]palladium(II) Tetrafluoroborate (31A). The structure shown
in Figure 16 was solved by standard methods.59-63 Full details and
tables of data are available as supporting information.
Synthesis of Complexes with Other Phenanthroline Ligands.
General Procedure. AgBF4(s) or AgPF6(s) (0.06 mmol) was added to
[(allyl)PdCl]2 (0.06 mmol) in CH2Cl2 (3 mL) at 0 °C. After 20 min,
the substituted phenanthroline (0.06 mmol) was added. After 5 min at
20 °C, the mixture was filtered through cotton and Celite. The solvent
was evaporated, and the product was recrystallized twice from CH2-
Cl2/hexane.
1.0 Hz, H1a), 3.36 (dd, 1 H, H1s). Minor isomer: 9.00 (d, JP9,8 ) 4.4
Hz, 1 H, HP9), 8.66 (d, JP7,8 ) 8.4 Hz, 1 H, HP7), 8.57 (d, JP4,3 ) 8.5
Hz, 1 H, HP4), 8.10 (dd, 1 H, HP8), 7.99 (m, 2 H, HP5,6), 7.67 (d, 1 H,
HP3), 5.58 (dd, J2,1a ) 13.3 Hz, J2,1s ) 7.6 Hz, 1 H, H2), 4.36 (d, 1 H,
H1a), 3.62 (d, 1 H, H1s). Unassigned signals: 1H NMR(400 MHz,
CDCl3): δ 2.97 (m), 2.72 (m), 2.61 (m), 2.30 (s), 1.83 (m), 1.71 (s),
1.61 (s), 1.59 (s), 1.56 (s), 1.51 (s), 1.29-1.12 (br m), 0.90 (m).
{2-[(1R)-2-endo-1,7,7-Trimethylbicyclo[2.2.1]hept-2-yl]-1,10-
phenanthroline}[1,1-dimethyl-(1,2,3-η)-2-propenyl]palladium Hexaflu-
orophosphate. Obtained as a 3:1 mixture of diastereomers. Major
isomer: 1H NMR (400 MHz, CDCl3): δ 9.05 (d, JP9,8 ) 4.0 Hz, 1 H,
HP9), 8.65 (d, JP4,3 ) 8.2 Hz, 1 H, HP4), 8.64 (d, JP7,8 ) 8.5 Hz, 1 H,
HP7), 8.11 (dd, 1 H, HP8), 8.06 (d, JP5,6 ) 8.8 Hz, 1 H), 8.01 (d, 1 H,
HP3), 8.00 (d, 1 H), 5.32 (dd, J2,1a ) 12.7 Hz, J2,1s ) 7.6 Hz, 1 H, H2),
4.13 (d, 1 H, H1a), 3.47 (d, 1 H, H1s). Minor isomer: 9.03 (d, J ) 5.5
Hz, 1 H, HP), 8.70 (d, J ) 8.2 Hz, 1 H, HP), 8.17 (d, J ) 5.5 Hz, 1 H,
NMR Determination of Diastereomeric Ratios. Equilibrium
between the diastereomers of the complexes was ensured by 1H NMR
integration of the isomers at regular intervals (ca. 2 days) and by adding
a catalytic amount of CF3CO2H 24 h before the last measurement. The
-
-
ratios were independent of the use of the BF4 or PF6 counterions.
{2-[(1R)-1,7,7-Trimethylbicyclo[2.2.1]hept-2-en-2-yl]-1,10-phenan-
throline}[1,1-diphenyl-(1,2,3-η)-2-propenyl]palladium Tetrafluo-
roborate. Obtained as a 2:1 mixture of diastereomers. Major
isomer: 1H NMR (400 MHz, CDCl3): δ 8.51 (dd, J ) 8.0 Hz, J ) 1.4
Hz, 1 H, HP9), 8.46 (d, J ) 8.6 Hz, 1 H, HP4), 7.95 (d, J ) 9.0 Hz, 1
H, HP5 or 6), 7.91 (d, J ) 8.8 Hz, 1 H, HP5 or 6), 7.79 (d, J ) 8.6 Hz, 1
H, HP3), 7.53 (d, J ) 8.0 Hz, 1 H, HP8), 6.48 (d, J ) 3.2 Hz, 1 H), 6.05
(dd, J2,1a ) 12.7 Hz, J2,1s ) 7.6 Hz, 1 H, H2), 4.47 (d, 1 H, H1s), 3.48
(d, 1 H, H1a), 2.36 (t, J ) 3.4 Hz, 1 H), 1.90 (m, 1 H), 1.58 (m, 1 H).
Minor isomer: 8.55 (dd, J ) 7.9 Hz, J ) 1.5 Hz, 1 H, HP9), 8.46 (d,
HP), 8.03 (d, J ) 8.2 Hz, 1 H, HP), 5.55 (dd, J2,1a ) 13.4 Hz, J2,1s
)
7.7 Hz, 1 H, H2), 4.54 (d, 1 H, H1a), 3.61 (d, 1 H, H1s). Unassigned
signals: 8.58 (m), 5.09 (m), 3.80 (m), 2.49 (m), 2.00-1.89 (br m),
1.85-1.78 (br m), 1.72 (s), 1.58 (s), 1.56-1.47 (br m), 1.46-0.68 (br
m).
{2-[(1R)-2-endo-1,7,7-Trimethylbicyclo[2.2.1]hept-2-yl]-1,10-
phenanthroline}[1,1-diphenyl-(1,2,3-η)-2-propenyl]palladium Hexaflu-
orophosphate. Obtained as an 11:1 mixture of diastereomers. Major
isomer: 1H NMR (400 MHz, CDCl3): δ 8.64 (d, JP4,3 ) 8.8 Hz, 1 H,
HP4), 8.62 (dd, JP9,8 ) 8.2 Hz, JP9,7 ) 1.4 Hz, 1 H, HP9), 8.03 (m, 2 H,
HP5,6), 7.98 (d, 1 H, HP3), 6.22 (dd, J2,1a ) 12.4 Hz, J2,1s ) 7.5 Hz, 1
H H2), 4.38 (d, 1 H, H1s), 3.64 (d, 1 H, H1a). Minor isomer: 6.40 (dd,
J2,1a ) 13.4 Hz, J2,1s ) 7.6 Hz, 1 H H2), 4.95 (d, 1 H, H1s), 3.74 (d, 1
H, H1a). Unassigned signals: 7.59-7.51 (m), 7.50-7.46 (m), 7.42 (m),
7.36 (m), 2.40 (m), 2.10 (m), 1.76 (m), 1.45 (m), 1.25 (s), 1.12 (m),
1.01 (s), 0.95 (m), 0.87 (s), 0.83 (s), 0.27 (s), 0.07 (s).
{2-[(1R)-2-endo-1,7,7-Trimethylbicyclo[2.2.1]hept-2-yl]-1,10-
phenanthroline}[(1,2,3-η)-2-cyclohexenyl]palladium Hexafluoro-
phosphate. Obtained as a 4:1 mixture of diastereomers. Major
isomer: 1H NMR (400 MHz, CDCl3): δ 9.39 (dd, J1 ) 4.9 Hz, J2 )
1.3 Hz, 1 H, HP), 8.65 (app. t, J ) 8.3 Hz, 2 H, HP), 8.10 (d, J ) 5.0
Hz, 1 H, HP), 8.08 (d, J ) 5.0 Hz, 1 H, HP), 8.03 (d, J ) 8.8 Hz, 1 H,
HP), 8.02 (d, J ) 8.8 Hz, 1 H, HP), 6.04 (t, J ) 6.7 Hz, 1 H, H2), 5.75
(quintet, J1 ) 11.5 Hz, J2 ) 6.2 Hz, 2 H, H1,3), 4.00 (dd, J1 ) 11.4 Hz,
J2 ) 3.4 Hz, 1 H, HT), 2.55 (m, 1 H, HT), 1.23 (s, 3 H, Me), 1.02 (s,
(58) Nutaitis, C. F.; Bernardo, J. E. J. Org. Chem. 1989, 54, 5629.
(59) Main, P.; Fiske, S. J.; Hull, S. E.; Lessinger, L.; Germain, G.;
DeClercq, J.-P.; Woolfson, M. M. Multan 11/82; University of York: York,
England, 1992.
(60) Cromer, D. T.; Waber, J. T. International Tables for X-Ray
Crystallography; The Kynoch Press: Birmingham, England, 1974; Vol.
IV, Table 2.2B.
(61) Cromer, D. T.; Waber, J. T. International Tables for X-Ray
Crystallography; The Kynoch Press: Birmingham, England, 1974; Vol.
IV, Table 2.3.1.
(62) Cruickshank, D. W. J. Acta Crystallogr. 1949, 2, 154.
(63) Frenz, B. A. The Enraf-Nonius CAD4 SDP -- A Real-Time System
for Concurrent X-Ray Data Collection and Crystal Structure Solution In
Computing in Crystallography; Schenk, H.; Olthof-Hazelkamp, R., van
Konigsveld, H., Bassi, G. C., Eds.; Delft University Press: Delft, The
Netherlands, 1978; pp 64-71.

Asymmetric Palladium-Catalyzed Allylation Reactions
J. Am. Chem. Soc., Vol. 118, No. 18, 1996 4313
56
3 H, Me), 0.78 (s, 3 H, Me). Minor isomer: 9.32 (d, J ) 4.5 Hz, 1 H,
HP), 8.58 (m, 2 H, HP), 5.90 (d, J ) 6.8 Hz, 1 H, H2), 5.54 (m, 2 H,
Pd(O2CCF3)]2 (0.0023 g, 0.0004 mmol), rac-(E)-1-acetoxy-1,3-
diphenyl-2-propene (0.1 g, 0.4 mmol), dimethyl malonate (0.078 g,
0.6 mmol), and NaH (0.0015 g, 0.6 mmol) in THF (10 mL) (or use of
BSA as in the alternative procedure) gave 0.1 g (80% yield) of colorless
H1,3), 3.87 (m, HT). Unassigned signals: 2.15 (m), 1.70 (m), 1.58 (m),
1.28 (m), 0.83 (m).
{2-[(1R)-2-endo-1,7,7-Trimethylbicyclo[2.2.1]hept-2-yl]-1,10-
phenanthroline}[1-phenyl-3-methyl-(1,2,3-η)-2-propenyl]palladi-
um Hexafluorophosphate. Obtained as a 5:2:1 mixture of diastere-
omers. Major isomer: 1H NMR (400 MHz, CDCl3): δ 8.65 (d, J )
8.5 Hz, 1 H, HP), 8.05 (m, 2 H, HP), 7.95 (d, J ) 8.8 Hz, 1 H, HP),
6.02 (t, J ) 11.3 Hz, 1 H, H2), 5.23 (d, J ) 12.5 Hz, 1 H, H1), 4.23
(m, 1 H, H3), 3.73 (dd, J1 ) 10.9 Hz, J2 ) 3.3 Hz, 2 H, HT), 2.51 (m,
1 H, HT), 1.55 (s, 3 H, MeT), 1.21 (s, 3 H, MeT), 1.02 (s, 3 H, MeT).
Minor isomer: 8.47 (dd, J1 ) 8.2 Hz, J2 ) 1.3 Hz, 1 H, HP), 5.60 (t,
J ) 7.7 Hz, 1 H, H2), 4.67 (m, 2 H, H1,3), 3.23 (m, 1 H, HT). Minor
isomer: 5.60 (dd, J2,1 ) 12.8 Hz, J2,3 ) 9.2 Hz, 1 H, H2), 5.20 (d, J )
12.8 Hz, 1 H, H1), 4.35 (m, 1 H, H3), 3.85 (m, 1 H, HT). Unassigned
signals: 8.56 (m), 7.17-7.89 (br m), 6.95 (d), 2.05 (t), 1.95 (d), 1.92
(d), 0.95-1.25 (br m).
oil: [R]25D +14.6° (c 0.01, CHCl3) [lit.26 calcd for the (S)-enantiomer
of 100% ee, [R]25 -22.4° (c 1.80, CHCl3)]; H NMR (300 MHz) δ
1
D
6.5 (d, J ) 15.7 Hz, PhCHdCH), 6.3 (dd, J ) 15.7, 8.6 Hz,
PhCHdCH), 4.25 (dd, J ) 11.0, 8.8 Hz, PhCHdCHCH), 3.95 (d, J )
11.0 Hz, (CH3OC(CdO))2CH), 3.7 and 3.52 (two s, 2 CH3O(CdO)).
With Eu(hfc)3, the signal at 3.7 ppm was split into two singlets (80%
ee with the NaH procedure; up to 92% ee with the BSA procedure).
The absolute configuration was assigned as R by comparing the sign
of its optical rotation with that of known product.26
Reaction of rac-(E)-2-Acetoxy-4-phenyl-3-butene with Dimethyl
Malonate. 13 (0.010 g, 0.03 mmol), [(MeCHCHCHMe)Pd(O2CCF3)]256
(0.009 g, 0.016 mmol), rac-(E)-2-acetoxy-4-phenyl-3-butene (0.30 g,
1.6 mmol), dimethyl malonate (0.30 g, 2.4 mmol), and NaH (0.062 g,
2.6 mmol) in THF (10 mL) gave 0.31 g (75%) of colorless oil: 1H
NMR showed an 80:20 inseparable mixture of A from attack at C-2
(33% ee) and B from attack at C-4 (95% ee with NaH; up to 96%
using the BSA procedure), respectively: 1H NMR (300 MHz) δ (A)
7.5-7.0 (m, 5 H, ArH), 6.48 (d, J ) 15.0 Hz, PhCHdCH), 6.2 (dd, J
) 15.1, 9.0 Hz, PhCHdCH), 3.75 and 3.67 (two s, 2 CH3OC(dO)),
3.4 (d, J ) 9.0 Hz, (CH3OC(dO))2CH), 3.1 (m, PhCHdCHCH), 1.2
(d, J ) 6.8 Hz, CH3CH); δ (B) 7.5-7.0 (m, 5 H, ArH), 5.7-5.4 (m,
CHdCH), 4.1 (m, PhCH), 3.8 (d, J ) 11.0 Hz, (CH3OC(dO))2CH),
3.73 and 3.48 (two s, 2 CH3OC(dO)), 1.63 (bd, J ) 4.5 Hz, CH3-
CHdCH) (lit.5k 1H NMR). With Eu(hfc)3, the signal of A at 1.2 ppm
was split into two doublets, and the signal of B at 1.63 ppm was split
into two doublets.
{2-[(1R)-2-endo-1,7,7-Trimethylbicyclo[2.2.1]hept-2-yl]-1,10-
phenanthroline}[1,3-diphenyl-(1,2,3-η)-2-propenyl]palladium Hexaflu-
orophosphate. Obtained as a 4:1 mixture of diastereomers. Major
isomer: 1H NMR (400 MHz, CDCl3): δ 8.98 (d, J ) 3.2 Hz, 2 H,
HP), 8.63 (dd, J1 ) 8.1 Hz, J2 ) 1.4 Hz, 2 H, HP), 8.55 (d, J ) 8.4 Hz,
2 H, HP), 7.92 (dd, J1 ) 8.1 Hz, J2 ) 4.5 Hz, 2 H, Ph), 6.12 (t, J )
10.8 Hz, 1 H, H2), 5.23 (d, J ) 12.1 Hz, 1 H, H1), 5.01 (d, J ) 10.6
Hz, 1 H, H3), 3.25 (m, 1 H, HT), 2.94 (m, 3 H, HT), 1.14 (s, 3 H, MeT),
0.98 (s, 3 H, MeT), 0.92 (s, 3 H, MeT). Minor isomer: 1H NMR (400
MHz, CDCl3): δ 8.57 (d, J ) 8.5, 2 H, HP), 6.23 (t, J ) 11.9 Hz, 1 H,
H2), 5.37 (d, 1 H, H1), 5.08 (d, J ) 10.4 Hz, 1 H, H3), 3.50 (m, 1 H,
HT), 3.03 (t, J ) 8.0 Hz, 2 H, HT). Unassigned signals: 1H NMR(400
MHz, CDCl3): δ 7.20-8.10 (m), 1.20-1.85 (m), 0.52 (s), 0.04 (s).
General Procedure for Allylations. The allylic acetate (0.10 mmol)
was added to the ligand (typically 0.005 mmol), Pd(0) complex (0.001
mmol in Pd), and THF (5.0 mL) at 25 °C. After 15 min, the suspension
from reaction of the dialkyl malonate (0.15 mmol) and NaH (0.17
mmol) in THF (5.0 mL), or alternatively, the malonate (0.15 mmol),
N,N-bis(trimethylsilyl) acetamide (BSA, 0.15 mmol), anhydrous KOAc
(0.01 mmol) and, in some cases, 18-crown-6 (0.01 mmol) were added
at 0 °C. The mixture was stirred at 25 °C and monitored by GC or
TLC. After completion, satd aq NH4Cl was added, the ether extracts
were dried and concentrated, and the product was purified by MPLC
Computational Details. Energy minimizations were performed
using the MacMimic/MM2(91) package,9 augmented by a parameter
set for (η3-allyl)palladium complexes.4a,c Calculations were done on a
Macintosh Quadra 800. Analysis of structures and drawings in wire
frame format were performed with MacMimic.
Acknowledgment. We thank Dr. Sverker Hansson for
valuable discussions, Mr. Donald Schifferl for NMR assistance,
Dr. Kenneth Haller for X-ray crystallography, and Dr. Bruce
Plashko and Mr. David Griffith for mass spectrometry. We
are grateful for financial support from the National Science
Foundation (U.S.A.), the National Institutes of Health (U.S.A.),
the Natural Science Research Council (Sweden), the Technical
Research Council (Sweden), and the National Research Council
(CNR, Italy). E.P.-C. is grateful for fellowships from the
Lubrizol Foundation (U.S.A.) and the National Council for
Science and Technology (Mexico). P.-O.N. thanks the Danish
Medical Research Council for financial support. J.O. thanks
the Royal Institute of Technology for a research scholarship.
We are grateful to Johnson Matthey Ltd. for a loan of palladium
compounds.
1
(silica gel, ether/pentane gradient). The % ee was determined by H
NMR using 8-10% tris[3-heptafluoropropyl)hydroxymethylene]-(+)-
camphorato]europium [Eu(hfc)3] in CDCl3.33 Absolute configurations
were not assigned unless otherwise indicated.
Reaction of rac-(E)-2-Acetoxy-3-pentene with Dimethyl Mal-
onate. Ligand 13 (0.0056 g, 0.018 mmol), [(MeCHCHCHMe)Pd(O2-
56
CCF3)]2 (0.0026 g, 0.005 mmol), rac-(E)-2-acetoxy-3-pentene (0.1
g, 0.9 mmol), dimethyl malonate (0.17 g, 1.3 mmol), and NaH (0.035
g, 1.4 mmol) in THF (10 mL) gave 0.13 g (71% yield) of colorless
oil: 1H NMR (300 MHz) δ 5.56 (ddq, J ) 15.2, 6.2, 0.8 Hz,
CH3CHdH), 5.34 (ddq, J ) 15.2, 8.1, 1.4 Hz, CH3CHdCH), 3.70 and
3.65 (two s, (CH3OC(dO))2C), 3.25 (d, J ) 9.1 Hz, (CH3OC-
(dO))2CH), 2.9 (m, CHdCHCH), 1.63 (dd, J ) 6.4, 1.5 Hz, CH3-
CHdCH), 1.06 (d, J ) 6.8 Hz, CHdCHCHCH3). With Eu(hfc)3, the
signal at 1.06 ppm was split into two doublets (24% ee).
Supporting Information Available: Full details of the
crystal structure determination for 31A and five tables of crystal
data and intensity data collection parameters, interatomic
distances, interatomic angles, torsional angles, fractional triclinic
coordinates, and isotropic displacement parameters for the non-
hydrogen atoms of 31A (10 pages). This material is contained
in many libraries on microfiche, immediately follows this article
in the microfilm version of the journal, can be ordered from
the ACS, and can be downloaded from the Internet; see any
current masthead page for ordering information and Internet
access instructions.
Reaction of rac-1-Acetoxy-2-cyclohexene with Diethyl Methyl-
malonate. 13 (0.0090 g, 0.028 mmol), [(MeCHCHCHMe)Pd(O2-
56
CCF3)]2 (0.004 g, 0.007 mmol), rac-1-acetoxy-2-cyclohexene (0.20
g, 1.4 mmol), diethyl methylmalonate (0.36 g, 2.1 mmol), and NaH
(0.054 g, 2.2 mmol) in THF (10 mL) gave 0.28 g (80% yield) of
colorless oil: 1H NMR (300 MHz) δ 5.75 (m, CHdCH), 5.5 (m,
CHdCH), 4.2 (m, 2 CH3CH2), 3.05 (m, CHdCHCH), 2.0-1.4 (m,
-(CH2)3-), 1.32 (s, CH3C), 1.24 (two overlapping t, J ) 8.0 Hz, 2CH3-
CH2). With Eu(hfc)3, the signal at 1.32 ppm was split into two singlets
(55% ee).
Reaction of rac-(E)-1-Acetoxy-1,3-diphenyl-2-propene with Di-
methyl Malonate. 13 (0.0025 g, 0.008 mmol), [(MeCHCHCHMe)-
JA950860T