Substituent effects of ligands on asymmetric induction in a prototypical palladium-catalyzed allylation reaction: Making both enantiomers of a product in high optical purity using the same source of chirality

Article abstract of DOI:10.1021/jo9911387

The substituent effects of the ligands 1 through 8 play a crucial role in determining the enantioselectivity in the palladium-catalyzed asymmetric allylation reaction between 1,3-diphenylprop-2-en-1-yl acetate and the sodium salt of diethyl malonate. For

Full text of DOI:10.1021/jo9911387

J . Org. Chem. 1999, 64, 7601-7611
7601
Su bstitu en t Effects of Liga n d s on Asym m etr ic In d u ction in a
P r ototyp ica l P a lla d iu m -Ca ta lyzed Allyla tion Rea ction : Ma k in g
Both En a n tiom er s of a P r od u ct in High Op tica l P u r ity Usin g th e
Sa m e Sou r ce of Ch ir a lity
Dean S. Clyne, Yvan C. Mermet-Bouvier, Nobuyoshi Nomura, and T. V. RajanBabu*
Department of Chemistry, The Ohio State University, 100 West 18th Avenue, Columbus, Ohio 43210
Received J uly 16, 1999
The substituent effects of the ligands 1 through 8 play a crucial role in determining the
enantioselectivity in the palladium-catalyzed asymmetric allylation reaction between 1,3-diphe-
nylprop-2-en-1-yl acetate and the sodium salt of diethyl malonate. For a given chirality of the
backbone, electron-deficient and electron-rich ligands generally gave opposite enantiomers, while
sterically hindered ligands had the same enantioselectivity as electron-rich ligands. In the case of
flexible backbones with ligands of comparable size, a variation of the enantioselectivity with the
electronic properties of the ligand is predictable. In ligands with rigid backbones, the steric effects
of the substituents appear to play a more decisive role, and caution should be exercised in
interpreting the role of electronic effects in such cases. Examples are provided for maximizing both
the chemical yield and the enantioselectivity of the allylation reaction through the tuning of the
electronic properties of the ligands. In selected cases, the major sense of the asymmetric induction
could be reversed solely by changing the electronic properties of the ligands. Bisphosphinites from
(R)-(+)-1,1′-bi-2-naphthol (BINOL) can be tuned to produce both (R)- and (S)-products in 80 and
87% ee, respectively. Stoichiometric reaction of complexes 10e* and 10j* with the sodium salt of
diethyl malonate gave malonate adducts with enantiomeric excesses in agreement with those
obtained under the catalytic conditions. Also reported are the details of an NMR study of the Pd-
(η3-1,3-diphenylallyl)bis-diphenylphosphinite and the corresponding bis-dicyclohexylphosphinite
complexes 10e* and 10j*.
In tr od u ction
reactions including hydrocyanation,2d,3 hydrogenation,4
and hydroformylation5 reactions. The origin of the en-
hanced enantioselectivity remains uncertain in these
cases due to the problems associated with studying the
catalytically relevant intermediates.6 In contrast, for the
palladium-catalyzed asymmetric allylation reaction (eq
1), a direct correlation between well-characterized inter-
mediates and the chirality of the products has been
established.7 With the hope that a study of the role of
Until recently, strategies for controlling enantioselec-
tivity in metal-catalyzed asymmetric reactions have
depended largely on the design and application of chiral
ligands that would provide optimum steric matching
between the catalyst and the substrate(s).1 Increasingly,
new examples of asymmetric reactions which employ
electronic tuning of ligands as a control element are
beginning to appear in the literature.2-5 We have suc-
cessfully used this approach in several key asymmetric
(1) (a) Ojima, I. Catalytic Asymmetric Synthesis; VCH: Weinheim,
1993. (b) Noyori, R. Asymmetric Catalysis in Organic Synthesis; J ohn
Wiley: New York, 1994. (c) Brunner, H.; Zettlmeier, W. Handbook of
Enantioselective Catalysis; VCH: Weinheim, 1993.
(2) (a) Inoguchi, K.; Sakuraba, S.; Achiwa, K. Synlett 1992, 169 and
references therein. (b) J acobsen, E. N.; Zhang, W.; Gu¨ler, M. L. J . Am.
Chem. Soc. 1991, 113, 6703. (c) Nishiyama, H.; Yamaguchi, S.; Kondo,
M.; Itoh, K. J . Org. Chem. 1992, 57, 4306. (d) RajanBabu, T. V.;
Casalnuovo, A. L. J . Am. Chem. Soc. 1992, 114, 6265. (e) RajanBabu,
T. V.; Casalnuovo, A. L.; Ayers, T. A. In Advances in Catalytic
Processes; Doyle, M., Ed.; J AI Press: Greenwich, CT, 1998; Vol. 2, p
1. (f) Williams, J . M. J . Synlett 1996, 705. (g) Schnyder, A.; Hinter-
mann, L.; Togni, A. Angew. Chem., Int. Ed. Engl. 1995, 34, 931 and
references therein.
(3) (a) Casalnuovo, A. L.; RajanBabu, T. V.; Ayers, T. A.; Warren,
T. H. J . Am. Chem. Soc. 1994, 116, 9869. (b) RajanBabu, T. V.;
Casalnuovo, A. L. J . Am. Chem. Soc. 1994, 118, 6325. (c) Casalnuovo,
A. L.; RajanBabu, T. V. In Chirality in Industry II; Collins, A. N.,
Sheldrake, G. N., Crosby, J ., Eds.; J ohn Wiley: Chichester, 1996; p
309.
(4) (a) RajanBabu, T. V.; Ayers, T. A.; Casalnuovo, A. L. J . Am.
Chem. Soc. 1994, 116, 4101. (b) RajanBabu, T. V.; Ayers, T. A.;
Halliday, G. A.; You, K. K.; Calabrese, J . C. J . Org. Chem. 1997, 62,
6012. (c) RajanBabu, T. V.; Radetich, B.; You, K. K.; Ayers, T. A.;
Casalnuovo, A. L.; Calabrese, J . C. J . Org. Chem. 1999, 64, 3429.
(5) RajanBabu, T. V.; Ayers, T. A. Tetrahedron Lett. 1994, 35, 4295.
ligand electronic effects in this reasonably well-delineated
system could have broader implications for the design of
asymmetric catalysts, we undertook a systematic study
of a variety of chiral backbones which could be modified
with chelating atoms of varying electron density.
The palladium-catalyzed asymmetric allylation has
been employed for a variety of C-C and C-N bond-
(6) (a) Landis, C. R.; Halpern, J . J . Am. Chem. Soc. 1987, 109, 1746.
(b) Brown, J . M. Chem. Soc. Rev. 1993, 25. (c) Bircher, H.; Bender, B.
R.; von Philipsborn, W. Magn. Reson. Chem. 1993, 31, 293. (d)
Giovannetti, J . S.; Kelly, C. M.; Landis, C. R. J . Am. Chem. Soc. 1993,
115, 4040.
(7) (a) Trost, B. M.; Van Vranken, D. L. Chem. Rev. 1996, 96, 395.
(b) Mackenzie, P. B.; Whelan, J .; Bosnich, B. J . Am. Chem. Soc. 1985,
107, 2046. (c) von Matt, P.; Lloyd-J ones, G. C.; Minidis, A. B. E.; Pfaltz,
A.; Macko, L.; Neuberger, M.; Zehnder, M.; Ru¨egger, H.; Pregosin, P.
S. Helv. Chim. Acta 1995, 78, 265. (d) Steinhagen, H.; Reggelin, M.;
Helmchen, G. Angew. Chem., Int. Ed. Engl. 1997, 36, 2108. (e) Evans,
D. A.; Campos, K. R.; Tedrow, J . S.; Michael, F. E.; Gagne´, M. R. J .
Org. Chem. 1999, 64, 2994.
10.1021/jo9911387 CCC: $18.00 © 1999 American Chemical Society
Published on Web 09/16/1999

7602 J . Org. Chem., Vol. 64, No. 20, 1999
Clyne et al.
forming reactions. Using carefully designed chiral ligands,
high enantioselectivities have been achieved for the
reactions with a prototypical substrate, 1,3-diphenylallyl
acetate, and a number of soft nucleophiles.7-18 Even
though chelating ligands with widely different donor-
acceptor properties (e.g., P/P, N/N, P/S, P/N) have been
successfully employed for this reaction, a systematic
study of the variation of electronic effects in truly C2-
symmetric ligands has not been carried out before. In this
context, note that our goal was to keep the two ligating
atoms in strictly identical steric environments, so that
the variations in enantioselectivity could be reliably
attributed to electronic effects. We have examined the
electronic effects for a variety of chiral backbones with
the goal of achieving practical levels of asymmetric
induction by electronic tuning of readily available ligands,
prototypes of which are diethyl tartrate and binaphthol.
We report here the first examples of the predictable
dependence of enantioselectivities on ligand electronic
effects in Pd-catalyzed allylation reactions. Also reported
is a rare example of the switching of the sense of chiral
induction by changes in the nonchiral substituents of a
C2-symmetric ligand. A careful structural investigation
of the intermediates by NMR rules out at least one
principal mechanism, viz., the intermediacy of geometric
isomers arising via π-σ-π interconversion, for this di-
chotomy. The finding that the π-allyl Pd intermediates
produced under stoichiometric or catalytic conditions
react with the malonate nucleophile to give essentially
identical enantioselectivity further supports the rel-
evance of such a mechanistic investigation.
F igu r e 1. Chiral phosphinites.
F igu r e 2. Substituents on phosphorus.
Ta ble 1. Electr on ic a n d Ster ic Effects on
En a n tioselectivity u sin g Va r iou s Su bstitu en ts on
P h osp h or u s w ith Liga n d 1
Resu lts a n d Discu ssion
We examined several C2- or pseudo-C2-symmetric diols
as precursors of the chiral backbones (Figure 1). These
were easily converted into bisphosphinites by reaction
with a chlorophosphine (eq 2). Reactions of (E)-1,3-
entry
Ra
ee (%)b
config.c
1
2
3
4
5
6
7
8
9
3,5-(CF3)2C6H3
3,5-F2C6H3
4-(CF3)C6H4
4-FC6H4
39
41
55
17
∼0
∼0
16
25
18
59
(R)
(S)
(S)
(S)
Ph
3,5-Me2C6H3
3,5-(TMS)2C6H3
3,5-(t-Bu)2-4-(MeO)C6H2
Et
(R)
(R)
(R)
(R)
diphenylprop-2-en-1-yl acetate with sodium diethyl ma-
lonate were carried out in THF at 25 °C in the presence
10
Cy
a
Ar of 1 was phenyl except entry 1 in which 2-naphthyl was
b
(8) Hayashi, T. In ref 1a, p 325.
applied. Determined by HPLC analysis with a chiral column
(Daicel OJ ). c Determined by comparison with an authentic (R)
sample prepared by a reaction with (S,S)-CHIRAPHOS.
(9) (a) Yamaguchi, M.; Shima, T.; Yamagishi, T.; Hida, M. Tetra-
hedron: Asymmetry 1991, 2, 663. (b) Yamaguchi, M.; Yabuki, M.;
Yamagishi, T.; Sakai, K.; Tsubomura, T. Chem. Lett. 1996, 241.
(10) (a) Frost, C. G.; Howarth, J .; Williams, J . M. J . Tetrahedron:
Asymmetry 1992, 3, 1089. (b) Allen, J . V.; Coote, S. J .; Dawson, G. J .;
Frost, C. G.; Martin, C. J .; Williams, J . M. J . J . Chem. Soc., Perkin
Trans. 1 1994, 2065.
(11) Sprinz, J .; Kiefer, M.; Helmchen, G.; Reggelin, M.; Huttner, G.;
Walter, O.; Zsolnai, L. Tetrahedron Lett. 1994, 35, 1523.
(12) Togni, A.; Breutel, C.; Schnyder, A.; Spindler, F.; Landert, H.;
Tijani, A. J . Am. Chem. Soc. 1994, 116, 4062.
of a mixture of 0.5 mol % of [Pd(η3-C3H5)(µ-Cl)]2 and 1.2-
1.4 mol % of the corresponding bisphosphinite (eq 1), and
the products were analyzed by HPLC. The bisphosphin-
ites are in general excellent ligands for the allylation
reaction. In many cases the yields of the reaction are
nearly quantitative.
(13) Brown, J . M.; Hulmes, D. I.; Guiry, P. J . Tetrahedron 1994,
50, 4493.
(14) Andersson, P. G.; Harden, A.; Tanner, D.; Norrby, P.-O. Chem.
Eur. J . 1995, 1, 12.
(15) Seebach, D.; Devaquet, E.; Ernst, A.; Hayakawa, M.; Ku¨hnle,
F. N. M.; Schweizer, W. B.; Weber, B. Helv. Chim. Acta 1995, 78, 1636.
(16) Pen˜a-Cabrera, E.; Norrby, P.-O.; Sjo¨egren, M.; Vitagliano, A.;
De Felice, V.; Oslob, J .; Ishii, S.; O’Neill, D.; Åkermark, B.; Helquist,
P. J . Am. Chem. Soc. 1996, 118, 4299.
(17) For a preliminary account, see: Nomura, N.; Mermet-Bouvier,
Y. C.; RajanBabu, T. V. Synlett 1996, 745. While this manuscript was
being reviewed another report dealing with electronic effects in Pd-
catalyzed allylation appeared. Saitoh, A.; Misawa, M.; Morimoto, T.
Synlett 1999, 483.
First, we examined the role played by electronic effects
on the enantioselectivity of the reaction using rigid
pseudo-C2-symmetric, glucose-derived ligands repre-
sented by the generic structure 1 (Figure 1).17 In previous
work, this ligand backbone had been successfully used
in asymmetric hydrocyanation and hydrogenation reac-
tions.3,4 Systematic changes of the nonchiral substituents
of 1 at the phosphorus atoms (Figure 2) leads to both
(R)- and (S)-enantioselectivities, although the enantiose-
lectivity remains modest (Table 1). The unsubstituted
diphenylphosphinite and the corresponding 3,5-Me2C6H3-
substituted phosphinite gave almost racemic product
(18) Trabesinger, G.; Albinati, A.; Feiken, N.; Kunz, R. W.; Pregosin,
P. S.; Tschoerner, M. J . Am. Chem. Soc. 1997, 119, 6315.

Substituent Effects in Pd-Catalyzed Allylation
J . Org. Chem., Vol. 64, No. 20, 1999 7603
Ta ble 2. Electr on ic a n d Ster ic Effects for En a n tioselectivity Usin g 2-8 w ith Va r iou s Su bstitu en ts on P h osp h or u sa
substituent
Rb
backbone
entry
2
3
4
5
6
7
8
1
2
3
4
3,5-(CF3)2C6H3
4-(CF3)C6H4
Ph
Cy
21 (S)
25 (R)
7 (S)
19 (S)
8 (R)
73 (R)
51 (R)
13 (S)
41 (S)
∼0
77 (R)
45 (R)
48 (S)
23 (R)
11 (S)
47 (R)
63 (R)
66 (S)
31 (S)
18 (R)
15 (R)
∼0
54 (R)
∼0
a
Each result contains enantioselectivity of the product (absolute configuration) as determined by HPLC analysis with a chiral column
b
(Daicel OJ ). The R substituents are placed in the increasing order of ligand electron density.
Ta ble 3. Electr on ic a n d Ster ic Effects for
En a n tioselectivity Usin g 4 a n d 8 w ith Va r iou s
Su bstitu en ts on P h osp h or u sa ,b
(∼0% ee, Table 1, entries 5 and 6), while the ligands with
aromatic rings having electron-withdrawing groups gen-
erally yielded (S)-enantioselectivity (Table 1, entries
2-4), with an exception of the 3,5-(CF3)2C6H3-substituted
phosphinite (Table 1, entry 1). Since there is good
correlation between electron-deficiency of the ligands and
the (S)-enantioselectivity, we suspect that in the case of
the 3,5-(CF3)2-substituted aromatic rings, the electronic
effects could play a secondary role to steric effects. A
trifluoromethyl group is reasonably large, having roughly
the same size as an isopropyl group.19 To probe the effect
of bulky meta-substituents we prepared the 3,5-
(TMS)2C6H3 and 3,5-(t-Bu)2-4-(MeO)C6H2 ligands which
are electronically different from the CF3-substituted
ligand. As expected, these ligands showed the same sense
of asymmetric induction (Table 1, entries 7 and 8) as the
3,5-(CF3)2C6H3 ligand. In light of these results, we believe
the steric bulkiness of the trifluoromethyl group led to
(R)-enantioselectivity rather than the (S)-enantioselec-
tivity predicted for an electron-deficient ligand. We have
previously observed such abnormal behavior for the bis-
3,5-(trifluoromethyl)phenyl substituents in Rh-catalyzed
hydrogenation reactions.4 A clear indication that elec-
tronic effects can be used to enhance the selectivity of
the (R)-enantiomer is shown in entries 9 and 10 in Table
1. Diethylphosphinite (Table 1, entry 9), which is pre-
sumably smaller than diphenylphosphinite (A values Et
) 1.8; Ph ) 2.8), yet more electron-rich, gave a modest
(R)-enantioselectivity. Making the ligand electron-rich
and bulkier leads to better enantioselectivity, for ex-
ample, in the dicyclohexylphosphinite (A value 2.2, Table
1, entry 10). These initial results prompted us to examine
the generality of these electronic effects using a number
of well-known and readily available ligand scaffolds. The
results are shown in Table 2.
As shown in Table 2, in each case, the enantioselec-
tivity is affected by the substituents on phosphorus. The
phosphinites 2 derived from (1R,2R)-cyclohexane-1,2-diol
were examined as a simpler analogue of 1. Considering
this backbone has the opposite configurations at its chiral
centers in comparison to 1, we obtained selectivities
similar to 1, except for the dicyclohexyl phosphinite. Even
though the selectivities are poor in the case of 2, since
both (R)- and (S)-products are obtained, these results
confirm that the unexpected electronic effects are not
simply caused by the pseudo-C2-symmetric nature of the
sugar backbone. On the basis of our previous studies on
asymmetric hydrogenation of dehydroamino acids we had
indeed expected, lower enantioselectivities with the cy-
clohexyl-1,2-phosphinites (2) vis-a`-vis the sugar phos-
phinites (1).4c The selectivity observed with the 3,5-
(CF3)2C6H3-substituted bisphosphinite ligand (2a ) was
opposite to what was seen for the other electron-deficient
substituent
Rd
backbone
entry labelc
4
8
1
2
3
4
5
6f
7
9
a
b
c
d
e
f
3,5-(CF3)2C6H3
96%
6%
73% ee (R) 11% ee (S)
90% 35%
71% ee (R) 52% ee (R)
38% >99%
51% ee (R) 47% ee (R)
3,5-F2C6H3
4-(CF3)C6H4
4-FC6H4
Ph
22%
>99%
48% ee (R)
92%
5% ee (S)
46%
13% ee (S) 63 (80) % ee (R)e
Ph
81%
64% ee (S)f,g
g
h
i
3,5-(TMS)2C6H3 96%
58% ee (R) 34% ee (S)
49% 95%
53% ee (S) 30% ee (S)
>99%
Et
10
Cy
Cy
94%
>99%
41% ee (S) 66 (87) % ee (S)e
11f
j
82%
73% ee (R)f,g
a
Each result contains enantioselectivity of the product (absolute
configuration) and the isolated yield of the product 1. The
b
enantioselectivity was determined by HPLC analysis using a
CHIRALCEL OJ column (Daicel). c See Figure 2 for substituents.
d
The R substituents are placed in the increasing order of ligand
electron density. e The numbers in the parentheses were obtained
at -30 °C. f The stoichiometric reactions were performed with (S)-
g
BINAPO as the backbone. The reaction was performed using a
stoichiometric amount of [Pd(η3-PhCHCHCHPh)(L*)]SbF6 com-
plex.
ligands, indicative of the dominance of steric effects over
electronic effects for the 3,5-(CF3)2C6H3 substituents.
Of the other readily available ligands, L-(+)-diethyl
tartrate based phosphinites 4 and (R)-(+)-binaphthol
based phosphinites 8 gave the most promising results and
these ligands were chosen for a detailed study of elec-
tronic and steric effects in the allylic alkylation reaction.
Diethyl tartrate is one of the most abundantly avail-
able chiral precursors. Both enantiomers have been used
extensively in asymmetric synthesis. It is more flexible
than backbones 1 and 2 (and possibly 8), and the
electronic effects appear to be more pronounced and
predictable in this system.20 The diphenylphosphinite
substituted ligand gave 13% ee (S) (Table 3, ligand 4,
entry 5). In comparison with this result, electron-
withdrawing substituents on aromatic rings consistently
increased the proportion of the (R)-enantiomer (Table 3,
entries 1-5) in a predictable fashion as the electron
density3a at the phosphorus decreased. The 3,5-(CF3)2
phenyl phosphinite (Table 3, entry 1) gave the highest
(20) We have previously seen that in Rh-catalyzed asymmetric
hydrogenation of dehydroamino acids, conformationally flexible ligands
show more pronounced, and often predictable, electronic effects. See
refs 4b and 4c.
(19) (a) Dutta, D.; Majumdar, D. J . Phys. Org. Chem. 1991, 4, 611.
(b) White, D.; Coville, N. J . Adv. Organomet. Chem. 1994, 36, 95.

7604 J . Org. Chem., Vol. 64, No. 20, 1999
Clyne et al.
Sch em e 1. Mech a n ism s for Ch ir a l In d u ction in
P d -Ca ta lyzed Allyla tion
ee. Note the switch of the sense of chiral induction and
enhancement of the enantioselectivity in going from the
diphenyl derivative (13% ee S) to the 3,5-(CF3)2C6H3
derivative (73% ee R). Thus by tuning the electronic
properties of this readily available ligand system, one is
able to get better yields and nearly the same selectivity
achieved by the unstable and synthetically more complex
tartrate-derived TADDOP ligand.15 Note that the 3,5-
(CF3)2-substituted aromatic rings did not appear to cause
a switching of the sense of enantioselectivity with this
backbone compared to the electronically similar 3,5-F2-
substituted ligand, as was seen with ligands 1 and 2. The
trifluoromethyl groups therefore function as true electron-
withdrawing groups with much less steric interference
in the case of this ligand with a flexible backbone. In
contrast, electron-rich dialkylphosphinites (Table 3, en-
tries 9 and 10) gave increased (S)-enantioselectivity, up
to 53% ee. The marginal difference between the bulky
cyclohexyl-substituted ligand as compared to that of the
ethyl derivative is more difficult to interpret. These
results clearly indicate that by tuning the electronic
properties of the ligands alone, the enantioselectivity of
a reaction can be switched. However, the unexpected (R)-
selectivity shown by bulky 3,5-(TMS)2-substituted aro-
matic ligands demonstrate that caution should be exer-
cised in exceptional cases.
unpredictable results. (b) In ligands with rigid backbones,
caution should be exercised in predicting the electronic
effects. In these instances, bulky substituents appear to
have the determining role. (c) In a given series, catalysts
with electron-deficient ligands give the opposite enanti-
omer in comparison to catalysts with electron-rich ligands.
Thus enantioselectivity can be switched by changing the
electronic properties of the chelating atoms, even when
the chirality of the backbone remains the same.
Stoich iom etr ic Allyla tion Rea ction s w ith Com -
p lexes 10e* a n d 10j*. Studying the ground-state struc-
tures of reaction intermediates would be irrelevant if it
cannot also be shown that the same intermediates are
involved under catalytic and stoichiometric conditions.
Therefore we undertook a study of the stoichiometric
allylation of complexes 10e* and 10j* (vide infra, eq 4).
[The asterisk indicate that the (S)-binaphthol-derived
phosphinites were used in these reactions]. The com-
plexes 10e* and 10j* were prepared according to eq 4
and then reacted with an equivalent amount of sodium
diethyl malonate under conditions similar to those used
in the catalytic studies. The reaction with the Pd-allyl
complex 10e* proceeded to give predominantly the (S)-
enantiomer of the malonate adduct in 64% ee (Table 3,
entry 6) at room temperature. The reaction of the allyl
complex 10j* with sodium diethyl malonate gave the (R)-
enantiomer in 73 % ee (Table 3, entry 11). The catalytic
and the stoichiometric results are consistent, yielding the
same products in approximately the same enantioselec-
tivity, thus giving further credence to the structural
studies reported below.
Asym m etr ic In d u ction Mech a n ism s. It is believed
that soft nucleophiles, such as the malonate anion,
directly attack one of the terminal carbons of the allyl
moiety from the face opposite to the coordinated pal-
ladium.7 Thus, the chiral ligands are remote from the
reacting center of the molecule and, a priori, difficulties
in obtaining high enantioselectivity should be antici-
pated. Nevertheless, remarkable enantioselectivity has
been achieved in this reaction. Mechanistic rationale for
the observed enantioselectivity can be summarized as
follows (Scheme 1): (i) an interaction of a ligand pendant
group with the incoming nucleophile which results in
preferential attack at one of the allyl terminal carbon
atoms,8 (ii) a distortion of the allylic unit caused by a
steric repulsion of a ligand substituent with the allylic
system,22 (iii) an electronic differentiation of the terminal
allyl carbon atoms by the trans effect resulting from
The (R)-binaphthol-derived phosphinites21 are excellent
ligands for the catalytic allylation reaction, often giving
quantitative yields of the products, except for the very
electron-deficient ligands (entries 1 and 2 under 8, Table
3). As in the case of ligand 1, electron-deficient and
electron-rich ligands gave opposite enantioselectivities
(Tables 1 and 3). Some of the electron deficient ligands
gave lower selectivities, (8a -d ), and in some cases lower
yields (8a and 8b) as compared to the diphenylphosphin-
ite (8e). The binaphthyl backbone is itself electron-
withdrawing, and in the presence of additional electron-
withdrawing substituents on phosphorus, coordination
ability of the ligand may have been reduced resulting in
poor reactions. This suggests that the appropriate elec-
tron density at phosphorus is a requirement for this
reaction. Dicyclohexylphosphinite gave an ee of 66% (S),
the highest of all electron-rich phosphinites. Note that
the sense of asymmetric induction was opposite to what
was observed with the corresponding diphenylphosphin-
ite. Since these ligands impart high activity for the
catalysts, we can carry out the reaction at temperatures
as low as -30 °C to achieve one of the highest levels of
enantioselectivity for C2-symmetric chelating phosphine
ligands (Table 3, entries 5 and 10).9 Note that both
enantiomers of the product were produced by a single
chiral backbone in ee’s of 80% and 87%, respectively
(Table 3, ligand 8, entries 5 and 10).
These results are summarized as follows: (a) In using
ligands with flexible backbones, electronic effects can be
used to enhance the enantioselectivity of an inherently
poor ligand. The sense, and the extent of enantioselec-
tivity are predictable, especially with ligands of similar
steric bulk. Ligands having bulky substituents can give
(21) Indeed diphenylphosphinite from binaphthol was one of the first
chiral ligands used for enantioselective allylation of malonates. Trost,
B. M.; Murphy, D. J . Organometallics 1985, 4, 1143. These workers
were also among the first to prepare the 3,5-bis-TMS-aryl derivative.
It is not apparent whether the switching of the enantioselection
occurred with the allylation of the cyclic substrate that was studied,
the increase in selectivity notwithstanding.
(22) Pfaltz, A. Acc. Chem. Res. 1993, 26, 341.

Substituent Effects in Pd-Catalyzed Allylation
J . Org. Chem., Vol. 64, No. 20, 1999 7605
different donor-acceptor properties of the chelating atom
in the square planar Pd-complex,10,11,23 and (iv) the
relative stabilities of the intermediate Pd(0)-olefin π-com-
plexes resulting from the addition of the nucleophile to
the π-allyl complexes.13
Mechanism i applies to only specially designed ligands,
whereas mechanism ii depends on intramolecular steric
effects and the attendant charge separation in the Pd-C
bonds of the π-allyl Pd intermediate(s) to differentiate
the π-allyl termini. In the case of mechanism iii steric
as well as electronic effects are important. A lack of C2-
symmetry in the ligands can give rise to more than one
intermediate complex whose reactivities determine the
enantioselection. It should be noted that rationalizations
ii and iii rely on typical ground-state arguments and are
valid only for nucleophilic attack proceeding through an
early-transition state. On the other hand, mechanism iv
is based on the assumption that the reaction has a late-
transition state. Although the asymmetric induction
explanations conclude that the enantioselectivity is re-
flected in the site-selectivity of the allyl terminal carbons
of thermodynamically stable (syn,syn)-allyl complexes,
some uncertainty remains as to which diastereomers of
the π-allyl complexes are catalytically relevant, especially
when the enantioselectivity is low. For example, NMR
studies have revealed that in some cases diastereomeric
syn/syn and syn/anti isomers of the π-allyl intermediates
can coexist in solution and these can be attacked at
different rates (vide infra).18,24,27
After careful inspection of a series of the malonate
alkylation reactions with various Pd(0)L2 complexes and
the X-ray structures of the putative intermediates, See-
bach15 proposed the following generalization: for a reac-
tion that proceeded through the λ-conformation26 for the
allyl intermediate, the (S)-isomer of the product was
likely to be the major one, whereas for the reaction
proceeding through the δ-conformation, the (R)-enanti-
omer (Figure 3) was likely to predominate. In a recently
completed study we have shown that seven-membered
1,2-bis-diarylphosphinite ligands form C2-symmetric Ni-
(II) or Rh(I)-complexes whose conformations can be
clearly defined by the λ/δ convention (Figure 4).4c We
expect the Pd-complexes presented here also to behave
accordingly, and the λ/δ models for the intermediates
should provide a basis for the discussion of the origin of
enantioselectivity in reactions of these complexes. The
bis-diarylphosphinite from (R)-binaphthol should form a
well-defined δ-chelate (Figure 3),28 and in accordance
with the empirical generalizations, the major allylation
product should be the (R)-isomer, as has been observed.
However, difficulties arise in providing a satisfactory
explanation as to how, for a given chiral backbone, a
ligand substituent quite removed from the coordination
F igu r e 3. A and
B
are conformations of η3-allylPd(II)
complexes.15 C and D are λ- and δ-conformations of C2-
symmetric [diarylphosphinite]Pd(II) complexes leading to (S)
and (R) adducts.
sphere (for example, a 4-fluoro- or 3,5-difluoro-substitu-
ent on a P-aromatic ring), or a change from P-phenyl (A
value 2.8) to P-cyclohexyl (A value 2.2) can bring about
conformational change of a reactive diastereomer. Thus
for the dramatic reversal of the selectivity, for example,
in going from complex 10e (80% (R)) to complex 10j (87%
(S)), one should look elsewhere for an explanation. Two
alternate explanations can be offered for this observation.
(i) The reactivities of the two allyl termini (Scheme 2,
see the syn,syn complex) are different for the two
different ligands. (ii) The (syn,syn)/(syn,anti) equilibrium
could be different for the phenyl and cyclohexyl phos-
phinites (Scheme 2).18,24,27 Nucleophilic addition at the
carbon carrying an anti-phenyl group in the (syn,anti)
complex would lead to the opposite enantiomer to that
obtained from attack of the same carbon atom in the
(23) (a) Åkermark, B.; Hansson, S.; Krakenberger, B.; Vitagliano,
A.; Zetterberg, K. Organometallics 1984, 3, 679. (b) Åkermark, B.;
Krakenberger, B.; Hansson, S.; Vitagliano, A. Organometallics 1987,
6, 620.
(24) Barbaro, P.; Pregosin, P. S.; Salzmann, R.; Albinati, A.; Kunz,
R. W. Organometallics 1995, 14, 5160.
(25) Togni, A.; Burckhurdt, U.; Gramlich, V.; Pregosin, P. S.;
Salzmann, R. J . Am. Chem. Soc. 1996, 118, 1031.
(26) For definitions of λ- and δ-conformations, see: (a) Kagan, H.
B. Asymmetric Synthesis Using Organometallic Catalysts. In Com-
prehensive Organometallic Chemistry; Wilkinson, G., Stone, F. G. A.,
Abel, E. W., Eds.; Pergamon Press: Oxford, 1982; Vol. 8, p 463. (b)
Pavlov, V. A.; Klaunovskii, E. I.; Struchkov, Y. T.; Voloboev, A. A.;
Yanovsky, A. I. J . Mol. Catal. 1988, 44, 217.
(28) Note that for a given backbone chirality, the δ/λ-conformations
alternate in going from a seven-membered to a nine-membered chelate.
Thus (R)-BINAP-metal chelates exhibit a λ-conformation and (R)-
BINAPO-phosphinite complexes a δ-conformation. The same has been
seen before in going from five- to seven-membered chelates. For a
compilation, see ref 15. Seebach, D.; Plattner, D. A.; Beck, A. K.; Wang,
Y. M.; Hunziker, D.; Petter, W. Helv. Chim. Acta 1992, 75, 2171.
(27) Herrmann, J .; Pregosin, P. S.; Salzmann, R.; Albinati, A.
Organometallics 1995, 14, 3311.

7606 J . Org. Chem., Vol. 64, No. 20, 1999
Clyne et al.
ligand, a mixture of two isomers was observed in solution.
The minor isomer (29%) was determined to be a (syn,-
syn) allyl-complex while the major isomer (71%) was a
(syn,anti) allyl-complex. The ratio of the two configura-
tional isomers (29:71) reflected the observed ratio of the
enantiomers for the product of the catalytic amination
(30:70). It was assumed that the nucleophile attacks the
same terminal allyl carbon trans to the phosphorus in
both cases. Since there was no equilibration of diaster-
eomeric complexes on the NMR time-scale, the conclusion
was reached that the ratio of the intermediate complexes
did reflect the enantioselectivity of the reaction. Two
other cases of the existence of syn,syn/syn,anti isomerism
have also been reported.18,24 In these cases the reactions
appeared to proceed under Curtin-Hammett conditions,
and the observed ratio of the diastereomers as deter-
mined by NMR (∼4:1) did not correspond to the enanti-
oselectivity (∼95:5) of the reaction.
NMR Stu d y of Allyl P a lla d iu m Bis-p h osp h in e a n d
Bis-p h osp h in ite Com p lexes. Since the catalytic and
stoichiometric reactions lead to the same level of asym-
metric induction (but in opposite absolute sense) for the
complexes 10e and 10j, it should be possible to explore,
by NMR, whether differing syn,syn/syn,anti ratios of
intermediates is responsible for this unusual mechanistic
dichotomy and the following studies address this issue.
F igu r e 4. Crystal structure of a prototypical 1,2-bis-dia-
rylphosphinite complex.4c
Sch em e 2. Role of syn ,syn /syn ,a n ti Isom er iza tion
in Sw itch in g of th e En a n tioselectivity
In enantioselective Pd-catalyzed allylation reactions,
analysis of ground-state properties of intermediates has
been used successfully to predict the sense, if not the
absolute value, of the enantioselectivity.7 In this connec-
tion, NMR methods have been used extensively. The
analysis of the data from a variety of 1- and 2-D NMR
experiments giving both scalar (J , in hertz) and through-
space coupling (NOE) information can provide a remark-
ably clear picture of the solution structure of the inter-
mediate Pd-complex(es).7d,29 We undertook a detailed
NMR study of the solution structures of the bis-phos-
phinite complexes 10e* ([Pd(η3-PhCHCHCHPh)((S)-8e)]-
SbF6) and 10j* ([Pd(η3-PhCHCHCHPh)((S)-8j)]SbF6) in
order to probe the origin of the unprecedented switching
in the alkylation selectivity when the P-substituent of
the BINAPO ligand was changed from phenyl to cyclo-
hexyl [(Table 3, ligand 8, entries 6 and 11); note that the
stoichiometric reactions and the NMR studies were
carried out with the (S)-BINAPO series of ligands, and
that these ligands and their complexes are indicated by
an asterisk (for example, 8e*, 8j* and 10e*, 10j*) next
to the number.] For comparison purposes, and to ascer-
tain our ability to detect all diasteromeric Pd(II)-
complexes (including any syn,anti isomers of the η3-1,3-
diphenylallyl complex) involved in the reaction, we also
studied the intermediate(s) from (S,S)-CHIRAPHOS, a
ligand that provided ee’s comparable to our system.30
When we started this work, there was only one report of
a [Pd(η3-PhCHCHCHPh)(L*)]X complex with a C2-sym-
metric chiral phosphine ligand.24 Several systems de-
scribed previously had an asymmetric allyl component
Sch em e 3. F or m a tion of a syn ,a n ti P d (a llyl)
Com p lex
(syn,syn) complex. Each diastereomeric intermediate
could react at a different rate. As long as the rate of
nucleophilic addition is slower than syn,syn/syn,anti
equilibration, the ratio of diastereomers need not reflect
the enantiomeric ratio of the products. In one case
reported by Togni (Scheme 3), a series of substituted
ferrocenyl pyrazole ligands were used in a palladium-
catalyzed asymmetric allylic amination reaction.25 The
reaction with a bulky 9-anthryl ligand gave 40% ee for
the (S)-product, a result opposite to that of the other
ligands substituted with methyl, phenyl, cyclohexyl,
1-naphthyl, or 2-naphthyl groups. They attributed this
inversion of enantioselectivity for the 9-anthryl-substi-
tuted case to a major steric influence of the bulky
substituent on the configuration of the allyl-palladium
intermediate. In an NMR study of the [Pd(η3-PhCHCH-
CHPh)(L*)]PF6 complex with the 9-anthryl-substituted
(29) (a) Pregosin, P. S.; Salzmann, R. Coord. Chem. Rev. 1996, 115,
35. (b) Pregosin, P. S.; Trabesinger, G. J . Chem. Soc., Dalton Trans.
1998, 727.
(30) Since there was a discrepancy in the ee values reported for the
allylation of dimethyl malonate (22%7b vs 90%9) using (S,S)-CHIRA-
PHOS as a ligand, we repeated the reaction (with diethyl malonate,
the substrate under our study) and observed ee’s of 75% (R) in THF
at room temperature.

Substituent Effects in Pd-Catalyzed Allylation
J . Org. Chem., Vol. 64, No. 20, 1999 7607
like η3-â-pinene,31-34 η3-methylallyl,35 or an η3-exometh-
ylenecyclopentene allyl moiety.32 Alternatively, the re-
ported complexes contained an asymmetric, mixed che-
late ligand with P/S,36 N/S,37 or P/N donor atoms.7d,11
attack takes place at the C-terminal carrying the stacked
Ph group.29
The [Pd(η3-PhCHCHCHPh)(L*)]SbF6 bisphosphinite
complexes 10e* and 10j* were prepared by the reaction
of [Pd(η3-PhCHCHCHPh)(µ-Cl)]2 with AgSbF6 and the
bisphosphinite ligands 8e* and 8j* (eq 4). Purification
The complex 9 ([Pd(η3-PhCHCHCHPh)[(S,S)-CHIRA-
PHOS]SbF6), a red solid, was prepared according to eq
1
3. The H, 13C, and 31P NMR spectra are included in the
Supporting Information. Surprisingly, the coupling pat-
tern, and to some extent, the chemical shifts in both the
1
31P and H spectra of this complex are highly dependent
on the solvent and the concentration of the complex in
solution. For example, the 31P{1H} NMR spectrum of the
crude material (32 mg) in CDCl3 solution (ca. 1 mL) at
30 °C contained a singlet at 49.14 ppm. The crude
complex 9 was purified by reprecipitation from dichlo-
romethane. Visible in the 31P{1H} NMR spectrum of this
pure complex (ca. 10 mg in 1 mL) was an AB quartet
(∆ν ) 0.58, J PP ) 67.2 Hz) with νA ) 49.26 and νB ) 48.68
ppm. It was determined that a low concentration of
complex 9 was essential in obtaining well-resolved 31P-
{1H} NMR spectra. Curiously a larger ∆ν (1.32 ppm) was
observed for the two 31P chemical shifts at 30 °C in C6D6
(νA ) 49.10 and νB ) 47.78, J PP ) 66.1 Hz) as compared
to the value in CDCl3, while under similar conditions of
concentration and temperature, in THF-d8, a singlet was
by precipitation from a benzene solution with hexane
gave the desired complexes as orange solids.
The aromatic region (6.64-8.06 ppm) of the H NMR
1
spectrum of 10e* in THF-d8 contained signals for the
protons of the binaphthalene backbone, the phenyl sub-
stituents of the allyl moiety and the phenyl substituents
on phosphorus (see Supporting Information for spectra).
Interestingly, one pair of the aromatic protons was
shielded and visible at δ 5.91 (dd, J PH ) 11.5, J HH ) 7.5).
The allyl methine proton signals were visible at δ 6.29
(dd, J ) ∼13, 13, 1H), 5.85 (dd, br, 1H), and 4.96 (dd, J )
∼12, 12, 1H) ppm. The 31P{1H} spectrum of 10e* con-
tained two doublets at 142.64 and 126.99 ppm with J PP
of 79.5 Hz and a ∆δ of 15.65 ppm (Table 4). Only one set
of phosphorus signals were observed at room temperature
indicating that either several rapidly equilibrating com-
plexes were present, or alternatively only a single isomer
of the complex of 10e* was present. Attempts to locate
minor isomers by a 2-D NOESY spectrum27 were not
successful. The topology of the complex is determined by
the chirality of the binaphthyl backbone and the attached
phenyl groups, with different diastereomeric complexes
1
observed at 49.73 ppm.38 The H [CDCl3 δ 0.91-1.15 (m,
6H), 2.09-2.27 (m, 1H), 2.42-2.60 (m, 1H), 4.79-4.92
(ddd, J ) 13.0, 9.1, 3.9, 1H), 5.07-5.23 (ddd, J ) 13.0,
8.8, 4.1, 1H), 6.49 (t, J ) 13.0, 1H), 6.55-6.75 (m, 2H,
aromatic), 6.85-7.75 (m, aromatic)] and 13C NMR spec-
trum [CDCl3 inter alia 112.8 (dm, J CP ) 12.0, C2), 91.21
(dd, J CP ) 22.6, 7.8, C1 or C3), 89.14 (dd, J CP ) 24.0, 11.0
C3 or C1, allylic fragment)] was also consistent with the
proposed structure. These data is in agreement with a
syn,syn configuration for the allyl unit, and there is no
indication of a syn,anti isomer in any of the solvents
examined. On the basis of previous results,24 we expected
to see an additional set of 31P signals if any of the syn,-
anti isomer was present in solution. From these studies
we conclude that in all probability the two enantiomers
arise via alkylation at the different allyl termini of the
coordinated unit, and not as a result of syn,syn to syn,-
anti isomerization followed by reactions of each of these
isomers. The major product arises via a δ-conformation15
of the π-allyl complex (Figure 3) where the nucleophilic
(38) One intriguing possibility is that THF is involved in a π-σ-π
equilibration process which renders the two phosphorus nuclei identi-
cal. Initial attack by THF will be followed by a rotation around a Pd-
olefin complex, which has been identified as a viable intermediate by
Reggellin and Helmchen in reactions of Pd-allyl complexes with
nucleophiles. See ref 7d.
(31) Ru¨egger, H.; Kunz, R. W.; Ammann, C. J .; Pregosin, P. S. Magn.
Reson. Chem. 1991, 29, 197.
(32) Pregosin, P. S.; Ru¨egger, H.; Salzmann, R.; Albinati, A.; Lianza,
F.; Kunz, R. W. Organometallics 1994, 13, 5040.
(33) Pregosin, P. S.; Ru¨egger, H.; Salzmann, R.; Albinati, A.; Lianza,
F.; Kunz, R. W. Organometallics 1994, 13, 83.
(34) Pregosin, P. S.; Salzmann, R.; Togni, A. Organometallics 1995,
14, 842.
(35) Pregosin, P. S.; Salzmann, R. Magn. Reson. Chem. 1994, 32,
128.
(36) Barbaro, P.; Currao, A.; Herrmann, J .; Nesper, R.; Pregosin,
P. S.; Salzmann, R. Organometallics 1996, 15, 1879.
(37) Boog-Wick, K.; Pregosin, P. S.; Trabesinger, G. Organometallics
1998, 17, 3254.

7608 J . Org. Chem., Vol. 64, No. 20, 1999
Clyne et al.
Ta ble 4. Selected 1H a n d 31P Ch em ica l Sh ifts for
a
Com p lexes 10e a n d 10j in THF -d 8
nucleus
10e
10j
PA
PB
H1
H2
H3
126.99
142.64
4.96
6.29
5.85
178.99
173.86
4.85
6.62
5.61
HAax (on PAax
HAeq (on PAeq
HBax (on PBax
)
)
)
6.77
7.15
5.91
7.84
2.05-2.30
1.25-1.96
HBeq (on PBeq
)
C1
C2
C3
89.95
112.92
103.03
81.46
111.78
103.48
a
Chemical shifts in ppm relative to TMS [1H] and H3PO4 [31P].
See text.
resulting from geometric isomers of the coordinated allyl
moiety (Scheme 2). The (syn,syn) configurations of the
allyl moiety are generally preferred, unless ligands with
exceptional steric demands are involved.25,32 In a series
of 1H and 31P DNMR experiments with 10e* in THF-d8,
no evidence for the presence of more than one complex
was observed at temperatures as low as 210 K. Thus, the
possibility of the rapid equilibration of diastereomeric
complexes was eliminated and we believe that a single
complex of 10e* is present in solution.39
F igu r e 5. 1H, 13C, and 31P NMR data for 10e*.
The orientation of the allyl moiety relative to the
bisphosphinite ligand was determined through H NOE
1
experiments, where through space relationships between
the allyl and the phosphorus phenyl ortho protons were
established. Irradiation at 5.85 ppm (H3) gave an NOE
of the HBeq ortho hydrogen signal, while irradiation of
H1 (δ 4.96) gave an NOE of the HAax signal at δ 6.77.
This NMR data is consistent with the (syn,syn) config-
uration of the allyl complex shown in Figure 5. Examina-
tion of this structure shows stacking of the PA equatorial
phenyl group and the phenyl group attached to C1 of the
allyl moiety. It is likely that the proximate position of
the C3-allyl phenyl group results in the shielding of the
ortho-hydrogens of the axial PB-phenyl group resulting
in an upfield shift (5.91 ppm). Further verification of the
relative orientation of the PB-phenyl groups come from
the strong NOE observed for the ortho hydrogen at δ
7.84, when H3 at δ 5.85 is irradiated. The shielding of
H1 relative to H3 (δ 4.96 and 5.85 respectively, see Figure
5) also confirms these assignments.
The allyl methine proton signals of 10e* were assigned
1
on the basis of H decoupling and COSY NMR experi-
ments. Based on a series of experiments (vide infra) the
signal at 6.29 ppm was assigned to the central allyl
proton H2 and the signals at 4.96 and 5.85 ppm to the
terminal protons H1 and H3. This chemical shift trend
has been observed in other complexes of this type.36,37 The
3
H2 signal was a dd with J HH ) ∼13 Hz which is typical
1
for a trans H-H arrangement. In a series of H NOE
experiments, NOEs were observed between protons H1
and H3, but not between H2 and either H1 or H3. These
data is consistent with the 1,3-diphenylallyl moiety
having the (syn,syn) configuration, eliminating the (syn,-
anti) complexes as likely structures for complex 10e*.
It is known that for a metal-coordinated allyl group,
coupling occurs through the metal between the phospho-
rus atom and a trans allyl proton.40 This trans P-H
coupling is large while the corresponding cis coupling is
small (ca. 0 Hz). The ortho protons of the phosphorus
phenyl substituents are also coupled to the phosphorus.
These “reporter” proton signals29 are extremely useful in
determining the overall 3-D structures of these complexes
Through selective proton decoupling in the 13C NMR
of 10e* at -40 °C, the chemical shifts for the allylic
carbon resonances were established at δ 89.95 (C1,
coupled to H1 at δ 4.96), 112.92 (C2, coupled to H2 at δ
6.29), and 103.03 (C3 coupled to H3 at δ 5.85).
1
and can be identified with H{31P} experiments.
The major product observed, i.e., the (S)-isomer results
from attack of the nucleophile at C1 (Figure 3). Note that
this carbon carries the phenyl group that encounters
maximum steric repulsion from one of the P-aryl groups,
in this case, the PA-equatorial phenyl group.41
Irradiation at 126.99 (PA) allowed for the assignment
of the trans H3 at 5.85 ppm and HAax and HAeq, the ortho
hydrogen atoms of the phenyl substituents on PA at 6.77
and 7.15 ppm, respectively. Similarly, irradiation of the
142.64 ppm phosphorus signal (PB) identified PB as trans
to H1 at δ 4.96. Also assigned were signals at δ 5.91 and
7.84 for HBax and HBeq the ortho hydrogen signals of the
phenyl substituents on PB. The ortho hydrogen atoms of
the axial phenyl substituents are typically shielded
relative to the equatorial substituent.31,32,35,40 Thus the
6.77 ppm and the 5.91 ppm signals were assigned to the
ortho hydrogen atoms HAax and HBax, respectively.
A similar series of NMR experiments were performed
with 10j* to determine the solution structure of this
1
complex. The aromatic region of the H NMR spectrum
of 10j* in THF-d8 contains signals for the protons of the
binaphthalene backbone and the phenyl substituents of
the allyl moiety between 7.06 and 8.22 ppm. The allyl
methine proton signals, all of which appear to be much
broader than those of 10e*, are seen at δ 6.62, 5.61, and
4.85 ppm. The highest field signal (δ 4.85) appears as a
broad multiplet. As the solution is cooled evolution of
some fine structure is seen, and at 270 K the signal at δ
4.85 is a broad doublet of doublets (J ) 10.4, 10.4 Hz).
(39) Of course, because of the C2-symmetric nature of the ligand,
the two (syn,syn) complexes derived from the two enantiomeric allyl
acetates are identical.
(40) Ammann, C. J .; Pregosin, P. S.; Ru¨egger, H.; Albinati, A.;
Lianza, F.; Kunz, R. W. J . Organomet. Chem. 1992, 423, 415.

Substituent Effects in Pd-Catalyzed Allylation
J . Org. Chem., Vol. 64, No. 20, 1999 7609
Clearly the dicyclohexylphosphinites appear to be much
more fluxional than the diphenylphosphinites. The ali-
phatic region of the spectrum contains broad signals for
the cyclohexyl substituents on phosphorus between 0 and
2.65 ppm. The 31P{1H} spectrum of 10j* contains two
doublets at δ 178.99 and 173.86 ppm with J PP of 53.5 Hz
and a ∆δ of 5.13 ppm (Table 4). Only one set of
phosphorus signals were observed for 10j* in THF-d8 at
temperatures as low as 210 K. As in the case of 10e*,
this is consistent with a single diastereomeric complex
of 10j* present in solution.39
different allylic terminal carbons have to be invoked in
order to explain the opposite selectivities observed. How
this depends on whether the phosphorus atoms carry a
phenyl or a cyclohexyl substituent is not entirely clear.
One possible explanation is that the stacking effects that
have been proposed for the P-phenyl ligands are not
possible with cyclohexyl-substituted phosphinites. In
addition, cyclohexyl derivatives are sterically bulkier (for
example, the cone angle for tricyclohexylphosphine is
larger than that of triphenylphosphine: 185° vs 145°).
These effects may have attendant consequences such as
differing Pd-Callyl bond lengths, and hence differing
electrophilicities for the associated carbons in the rel-
evant diasteromers.42 One measure of such differences
is the relative chemical shifts of the phosphorus atoms
and the associated trans allylic carbon atoms. Curiously,
in the case of the diphenylphosphinites, the low-field
carbon is trans to the high-field phosphorus and in the
case of the dicyclohexylphosphinites the high field carbon
is trans to the high-field phosphorus. The relevance of
this observation remains to be clarified. For example, the
difference in donor-acceptor properties of the two chelat-
ing phosphorus atoms of a C2-symmetric ligand could be
responsible for the preferred site of attack by a nucleo-
phile. A more likely scenario is that a synergistic effect
of the electrophilicities of the terminal allyl carbons and
the electronic properties of the relevant trans-chelating
atom is responsible for the observed selectivity. In the
BINAPO system, assuming the same conformation for
both the diphenylphosphinite and the dicyclohexylphos-
phinite complexes, the major product in each case arises
from the attack of the nucleophile trans to the more
deshielded phosphorus. Our IR studies with [1,2-bis-
(diarylphosphinite)Ni(CO)2 complexes3a support the sug-
gestion that the more electron-withdrawing the phos-
phorus, the better the back-bonding from the metal as
judged by the carbonyl (νA1 and νB1) frequencies. In Pd-
catalyzed allylation reactions, a number of highly suc-
cessful chelating ligands have been designed on the basis
of preferential nucleophilic attack on the allylic carbon
trans to the better π-acceptor.7
The allyl methine proton signals of 10j* were assigned
1
on the basis of H decoupling and COSY NMR experi-
ments. The signal at δ 6.62 ppm was assigned to allyl
proton H2, and the signals at δ 4.85 and 5.61 ppm were
assigned to the terminal hydrogens H1 and H3. The H2
3
signal was a dd with a J HH of 11.4 Hz which is typical
for a trans H-H arrangement. An NOE was observed
between the H1 and H3 protons but not between H2 and
the other allyl protons. These data is consistent with the
1,3-diphenylallyl moiety having a (syn,syn) configuration.
Irradiation of the 178.99 ppm phosphorus signal (PA)
1
in a H{31P} experiment at -20 °C identified PA as trans
to H3 at 5.61 ppm. Also identified were signals at 2.30
and 2.05 ppm (all greater than δ 2.00) in the aliphatic
region for protons of the PA-cyclohexyl substituents. On
account of the small chemical shift difference between
PA and PB, the decoupler power in these experiments had
to be very carefully set in order to irradiate each of the
31P nuclei selectively. Irradiation of PB at 173.86 ppm
allowed for the assignment of the trans H1 at 4.85 ppm,
and also signals between 1.25 and 1.96 ppm (most
notably none greater than δ 2.00) for protons on the
cyclohexyl substituents. The line shape of the aliphatic
1
region of the H spectrum of 10j* was very complex and
overlapping, and the unambiguous assignment of signals
in this region was not possible.
Irradiation of the allyl protons of 10j* in NOE experi-
ments gave data that was consistent with the relative
orientation of the allyl and the bisphosphinite ligand
1
moieties determined by H{31P} measurements. Irradia-
tion of H1 which is cis to PA, gave an enhancement of the
signals at 2.30 and 2.05 ppm for protons on the PA-
cyclohexyl substituents. Irradiation of H3 gave an NOE
to signals below 2.00 ppm.
Due to the high degree of signal overlap in the aliphatic
region of the 1H NMR spectrum, an unambiguous as-
signment of signals from each of the four cyclohexyl
substituents, and hence the identification of reporter
atoms in each quadrant of 10j* was not possible. This is
in contrast to complex 10e* where the phenyl ortho
hydrogen atom signals were used to determine the overall
3-D structure of the complex.
Con clu sion s
Our results can be summarized as follows. Even when
the bisphosphinite ligands are C2-symmetric, in the
formation of complexes (e.g., 10e* or 10j*), this symmetry
is broken, as indicated by the AB-nature of the 31P NMR.
On the basis of the NMR experiments, we propose a
λ-conformation for the complex 10e*. The 1,3-diphenyl-
allyl unit exists exclusively, as the (syn,syn) isomer.
Attack by a malonate anion preferentially occurs at the
carbon that carries the phenyl group which is eclipsed
(stacked) with respect to one of the P-aryl groups. In the
case of the cyclohexyl substituted phosphinite complex
10j*, which gave the opposite enantiomer as the major
product, the conformation of the chelate could not be
determined with certainty. The results of these NMR
studies clearly rule out one of the possible mechanisms
for this unusual stereochemical result: the one involving
Through selective proton decoupling in the 13C NMR
of 10j* at -40 °C, the chemical shifts for the allylic
carbon resonances were established at δ 81.46 (C1,
coupled to H1 at δ 4.85), 111.78 (C2, coupled to H2 at δ
6.62), and 102.48 (C3 coupled to H3 at δ 5.61). An HMQC
experiment at -20 °C confirmed these assignments.
The structure of 10j* could not be rigorously estab-
lished on the basis of these NMR experiments. However,
with the syn,syn configuration of the allyl moiety con-
firmed, these NMR results are consistent with a confor-
mation of 10j* which is similar to that of 10e*. If the
basic assumption is made that both 10e* and 10j* have
the same λ-conformation, nucleophilic attacks at two
(41) It should be noted this is counter to the notion that the most
deshielded carbon of the allyl unit is the one that undergoes the
nucleophilic attack. However, the 13C chemical shift criterion for
reactivity24 as well as Pd-C bond strength42 has been labeled “suspect”.
(42) For an example of what appears to be a steric acceleration of a
Pd-C bond cleavage in an allyl complex, see: Breutel, C.; Pregosin,
P. S.; Salzmann, R.; Togni, A. J . Am. Chem. Soc. 1994, 116, 4067.

7610 J . Org. Chem., Vol. 64, No. 20, 1999
Clyne et al.
Dieth yl (2R,3R)-2,3-Bis[(d ip h en yp h osp h in o)oxy]su c-
cin a te (4e), Meth od B. To a solution of diphenylchlorophos-
phine (196 mg, 0.888 mmol) in toluene (3 mL) was added
DMAP (112 mg, 0.917 mmol) in toluene (1 mL) and L-diethyl
tartrate (73 mg, 0.35 mmol). The reaction mixture was stirred
at room temperature for 3 h and filtered through Celite.
Concentration of the filtrate and removal of excess reagents
under vacuum overnight gave the bisphosphinite 4e (248 mg,
a syn,syn to syn,anti interconversion. Nucleophilic attack
at the allylic carbon which is trans to the more deshielded
and π-acceptor phosphorus atom provides a satisfactory
explanation for the observed reversal of selectivity in
going from the diphenylphosphinite to the dicyclohexyl-
phosphinite BINAPO complex. Whatever be the origin
of the reversal of stereoselectivity, in selected cases it has
been shown to be possible to reverse the outcome of an
enantioselective reaction simply by changing the non-
chiral substituents on a ligand.
1
0.432 mmol, crude). H NMR: δ 0.91 (t, J ) 7.1 Hz, 6H), 3.78
(dq, J ) 7.1, 10.8 Hz, 2H), 3.96 (dq, J ) 7.1, 10.8 Hz, 2H),
5.20 (d, J P-H ) 9.0 Hz, 2H), 7.13-7.66 (m, 20H). 31P{1H}
NMR: δ 119.5 (s).
(1R,2R)-1,2-Cycloh exylen e Bis(d ip h en ylp h osp h in ite)
(2e), Meth od C. To a solution of (1R,2R)-trans-cyclohexane-
1,2-diol (42 mg, 0.37 mmol) in THF (5 mL) at -30 °C was
added a solution of n-butyllithium in hexane (280 µL, 0.730
mmol). The reaction mixture was stirred with warming to room
temperature over 20 min, and diphenylchlorophosphine (160
µL, 0.730 mmol) was added. The reaction mixture was stirred
at room temperature for 9 h. The solvent was evaporated in
vacuo, and the residue was extracted with diethyl ether and
filtered through Celite. Concentration of the filtrate and
removal of excess reagents under vacuum overnight gave the
bisphosphinite 2e (195 mg, 0.401 mmol, crude). 1H NMR: δ
1.20-1.30 (m, 2H), 1.47-1.51 (m, 2H), 1.61-1.67 (m, 2H),
2.03-2.08 (m, 2H), 3.99-4.04 (m, 2H), 7.17-7.53 (m, 20H).
31P{1H} NMR: δ 106.4 (s).
Ba ck bon e 6. To L-threitol (56.9 mg, 0.466 mmol) in dichlo-
romethane (10 mL) were added triphenylmethyl chloride
(300.0 mg, 1.076 mmol), triethylamine (2 mL), and a catalytic
amount of DMAP. The reaction mixture was stirred at room
temperature for 16 h, poured into crushed ice, and extracted
with dichloromethane. The organic extracts were sequentially
washed with saturated ammonium chloride solution, water,
dried over anhydrous Na2SO4, and concentrated in vacuo.
Column chromatography of the residue with 6% ethyl acetate
in hexane as eluant gave backbone 6 (149 mg, 52%). 1H NMR:
δ 2.62-2.67 (m, 2H), 3.11 (dd, J ) 5.6, 9.6 Hz, 2H), 3.35 (dd,
J ) 4.5, 9.6 Hz, 2H), 3.74-3.85 (m, 2H), 7.18-7.41 (m, 30H).
Exp er im en ta l Section
Gen er a l Meth od s. Unless otherwise stated, all reactions
were carried out under a nitrogen atmosphere inside a Vacuum
Atmospheres drybox. Analytical TLC were performed on E.
Merck precoated (0.25 mm) silica gel 60 F254 plates. Column
chromatography was conducted with silica gel 40 (Scientific
Adsorbents Incorporated, 40 Microns Flash).43 NMR spectra
were obtained on CDCl3 solutions unless otherwise stated.
Splitting patterns are indicated as the following: s, singlet;
d, doublet; t, triplet; q, quartet; m, multiplet; br, broad. Low-
resolution mass spectra were recorded in fast atom bombard-
ment mode (MSFAB). Enantiomeric excesses were determined
by HPLC using a Daicel Chiralcel OJ column and eluting with
5% 2-propanol in hexane as eluant. Tetrahydrofuran, diethyl
ether, and hexane were distilled from sodium benzophenone
ketyl. Dichloromethane was distilled from CaH2. The solvents
in the drybox were stored over activated molecular sieves after
distillation as described above. The bisphosphinites 1-8
(Figures 1 and 2) were synthesized according to the methods
A-C. The cationic (phosphine)Pd(allyl) complexes were syn-
thesized by the method D. Representative examples are given
below. Characterization data for other bisphosphinites are
included in the Supporting Information. Air-sensitive trivalent
phosphorus containing compounds give poor microanalyses.
The purity of the ligands were ascertained by H and 31P{1H}
NMR. In most instances the ligands were >95% pure.
1
P r ep a r a tion of P d -Allyl Com p lexes. [P d (η3-P h CHCH-
CHP h )((S,S)-CHIR AP HOS)]Sb F 6 (9), Met h od D. [Pd(η3-
PhCHCHCHPh)(µ-Cl)]2 complex (11.5 mg, 0.0172 mmol) and
silver antimony hexafluoride (17.6 mg, 0.0512 mmol) were
stirred in CH2Cl2 (5 mL) for 15 min. The mixture was filtered
through Celite into a solution of (S,S)-CHIRAPHOS (14.8 mg,
0.0347 mmol) and stirred for an additional 45 min. The
reaction was concentrated and the resultant solids were
dissolved in benzene, transferred, and precipitated with hex-
ane. The precipitate was filtered with Celite and washed with
hexane. The precipitate was dissolved with CH2Cl2, eluted
through Celite and concentrated to give complex 9 (19.1 mg,
58%) as a red solid. 1H NMR: δ 0.91-1.15 (m, 6H), 2.09-2.27
(m, 1H), 2.42-2.60 (m, 1H), 4.79-4.92 (ddd, J ) 13.0, 9.1, 3.9
Hz, 1H), 5.07-5.23 (ddd, J ) 13.0, 8.8, 4.1 Hz, 1H), 6.49 (t, J
) 13 Hz, 1H), 6.55-6.75 (m, 2H, aromatic), 6.85-7.75 (m,
aromatic). 13C NMR (75 MHz): allylic carbons 89.14, 91.21,
112.8. 31P{1H} NMR (101 MHz): δ 48.68 (d, J P-P ) 67.2 Hz),
49.26 (d, J P-P ) 67.2 Hz). MSFAB (SIMS(+)) m/z (rel inten-
sity): 726 (C43H41P2Pd, M+, 35), 725 (M+ - 1, 100).
The precursor diol for bisphosphinite 1 and the O-substi-
tuted pyranose bisphosphinites 1a -g were prepared by the
method of RajanBabu and co-workers.3a The precursor diol for
bisphosphinite 2 was purchased from Fluka. The precursor
diols for bisphosphinites 3, 4, 7, and 8 and (L)-threitol used in
the preparation of bisphosphinites 5 and 6 were purchased
from Aldrich. Backbone 5 was prepared according to the
method of Mash and co-workers.44 The chlorophosphines were
prepared according to known procedures.45,46 The [Pd(η3-
PhCHCHCHPh)(µ-Cl)]2 complex was prepared according to the
method of Bosnich and co-workers.47
Ch em ica l Meth od s. P r ep a r a tion of th e P h osp h in ites.
(1R,2R)-1,2-Cycloh exylen e Bis[3,5-bis(tr iflu or om eth yl)-
ph en ylph osph in ite] (2a), Meth od A. To 1.0 equiv of (1R,2R)-
trans-cyclohexane-1,2-diol (22.4 mg, 0.190 mmol) in dichlo-
romethane (1 mL) were added pyridine (0.5 mL), a catalytic
amount of DMAP (4-(dimethylamino)pyridine), and chlorobis-
[bis-3,5-(trifluoromethyl)phenyl]phosphine (209 mg, 0.425 mmol)
in dichloromethane (1 mL). The reaction mixture was stirred
at room temperature for 10 h, and the solvent was evaporated
in vacuo. The residue was extracted with dry diethyl ether
and filtered through Celite. Concentration of the filtrate and
removal of excess reagents under vacuum overnight gave the
bisphosphinite 2a (207 mg, 0.202 mmol, crude). 1H NMR: δ
1.20-1.45 (m, 2H), 1.46-1.65 (m, 2H), 1.66-1.85 (m, 2H),
1.86-2.10 (m, 2H), 4.15-4.28 (m, 2H), 7.58-8.26 (m, 16H).
31P{1H} NMR: δ 102.1 (s).
[P d (η3-P h CHCHCHP h )(8e*)]SbF 6 (10e*), Meth od D.
The reaction of [Pd(η3-PhCHCHCHPh)(µ-Cl)]2 complex (10.2
mg, 0.0152 mmol), silver antimony hexafluoride (16.2 mg,
0.0472 mmol) and bisphosphinite 8e* (20.2 mg, 0.0309 mmol)
1
gave complex 10e* (28.8 mg, 80%) as an orange-red solid. H
NMR (THF-d8): δ 4.96 (dd, J ) 12.1, 12.1 Hz, 1H), 5.85 (br
dd, 1H), 5.91 (dd, J ) 7.5, 11.5 Hz, 2H), 6.29 (dd, J ) 12.7,
12.7 Hz, 1H), 6.64-6.85 (m, 12H), 6.97-7.48 (m, 20H), 7.64-
7.73 (m, 4H), 7.83 (d, J ) 11.1 Hz, 1H), 7.85 (d, J ) 11.7 Hz,
1H), 8.05 (d, J ) 8.5 Hz, 1H), 8.09 (d, J ) 9.1 Hz, 1H). 13C
NMR (125 MHz, THF-d8, -40 °C): allylic carbons 89.95,
(43) Still, W. C.; Kahn, M.; Mitra, A. J . Org. Chem. 1978, 43, 2923.
(44) Mash, E. A.; Nelson, K. A.; Van Deusen, S.; Hemperly, S. B.
Org. Synth. 1989, 68, 92.
(45) Unruh, J . D.; Christenson, J . R. J . Mol. Catal. 1982, 14, 19.
(46) Perich, J . W.; J ohns, R. B. Synthesis 1988, 142.
(47) Auburn, P. R.; Mackenzie, P. B.; Bosnich, B. J . Am. Chem. Soc.
1985, 107, 2033.
103.03, 112.92. 31P{1H} NMR (THF-d8): δ 126.99 (d, J P-P
)
79.5 Hz), 142.64 (d, J P-P ) 79.5 Hz). MSFAB (SIMS(+)) m/z
(rel intensity): 954 (C59H45O2P2Pd, M+, 34), 953 (M+ - 1, 65),
453 (100).

Substituent Effects in Pd-Catalyzed Allylation
J . Org. Chem., Vol. 64, No. 20, 1999 7611
[P d (η3-P h CHCHCHP h )(8j*)]SbF 6 (10j*), Meth od D. The
reaction of [Pd(η3-PhCHCHCHPh)(µ-Cl)]2 complex (8.1 mg,
0.012 mmol), silver antimony hexafluoride (12.7 mg, 0.0370
mmol), and bisphosphinite 8j* (16.8 mg, 0.0247 mmol) gave
4.28 (dd, J ) 8.2, 10.9 Hz, 1H), 6.38 (dd, J ) 8.2, 15.7 Hz,
1H), 6.49 (d, J ) 15.7 Hz, 1H), 7.20-7.32 (m, 10H).
(b) Rea ction of Dieth yl Ma lon a te An ion w ith a Sto-
ich iom etr ic Am ou n t of [P d (η3-P h CHCHCHP h )(8j*)]SbF 6
Com p lex (10j*). A solution of sodium diethyl malonate (3.6
mg, 0.020 mmol) in THF (3 mL) was added to a solution of
10j* (19.0 mg, 0.0156 mmol) in THF (2 mL), and the resulting
mixture was stirred at room temperature for 10 h. The deep-
red reaction solution was removed from the drybox, diluted
with diethyl ether, washed with saturated ammonium chloride
solution and brine, and dried over anhydrous MgSO4. The
reaction was filtered and concentrated in vacuo. Column
chromatography of the residue with 2% ethyl acetate in hexane
as eluant gave adduct 11 as a colorless oil (4.5 mg, 82%).
1
complex 10j* (21.5 mg, 73%) as an orange-red solid. H NMR
(THF-d8): δ 0.00-2.35 (m, 44H), 4.85 (br dd, J ) 10.4, 10.4
Hz, 1H), 5.61 (br dd, J ) 10.9, 10.9 Hz, 1H), 6.62 (br dd, J )
11.4, 11.4 Hz, 1H), 7.04 (d, J ) 8.2 Hz, 1H), 7.09 (d, J ) 8.2
Hz, 1H), 7.30 (t, J ) 8.2 Hz, 1H), 7.31 (t, J ) 8.2 Hz, 1H),
7.36-7.41 (m, 11H), 7.56-7.61 (m, 2H), 7.71 (d, J ) 8.8 Hz,
1H), 7.99 (d, J ) 8.2 Hz, 1H), 8.03 (d, J ) 8.2 Hz, 1H), 8.13 (d,
J ) 9.5 Hz, 1H), 8.22 (d, J ) 8.8 Hz, 1H). 13C NMR (125 MHz,
THF-d8, -40 °C): allylic carbons 81.46, 102.48, 111.78. 31P-
{1H} NMR (THF-d8): δ 173.86 (d, J P-P ) 48.9 Hz), 178.99 (d,
J P-P ) 48.9 Hz). MSFAB (SIMS(+)) m/z (rel intensity): 978
(C59H69O2P2Pd, M+, 50), 977 (M+ - 1, 100).
Ack n ow led gm en t. We are grateful to the National
Science Foundation (CHE-9706766), the EPA/NSF Pro-
gram for Technology for Sustainable Environment (EPA
R826120-01-0), and the Petroleum Research Fund (ACS)
for financial support of this work. We thank The DuPont
Company for a grant from the Aid to Education funds.
Financial support is also acknowledged from The Ohio
State University (postdoctoral fellowship, D.S.C).
Dieth yl 1-(1,3-Dip h en ylp r op -2-en yl)m a lon a te (11). (a )
Rea ction of Dieth yl Ma lon a te An ion Med ia ted by a
Ca ta lytic Am ou n t of a P a lla d iu m Bisp h osp h in ite Com -
p lex. To a solution of [Pd(η3-C3H5)(µ-Cl)]2 (1.0 mg, 2.7 mmol;
0.55 mol %) in THF (1 mL) was added a solution of the
corresponding bisphosphinite (6.5 mmol, 1.3 mol %) in THF
(3 mL). A solution of 1,3-diphenylprop-2-en-1-yl acetate (125
mg, 0.495 mmol) in THF (2 mL) was treated successively with
this catalyst solution and a solution of diethyl malonate sodium
salt (109 mg, 0.598 mmol) in THF (4 mL). After the conversion
was complete according to TLC analysis, the reaction mixture
was removed from the drybox, and the reaction was quenched
with saturated ammonium chloride solution (10 mL). The
organic material was extracted with diethyl ether (20 mL),
dried over anhydrous MgSO4, and concentrated in vacuo.
Column chromatography of the residue with 10% ethyl acetate
in hexane as eluant gave malonate adduct 11. 1H NMR: δ 1.02
(t, J ) 7.1 Hz, 3H), 1.22 (t, J ) 7.1 Hz, 3H), 3.98 (d, J ) 10.9
Hz, 1H), 3.99 (q, J ) 7.1 Hz, 2H), 4.19 (q, J ) 7.1 Hz, 2H),
Su p p or tin g In for m a tion Ava ila ble: Experimental de-
tails and characterization data for bisphosphinite ligands,
NMR spectra for complexes 9, 10e, and 10j as well as two
typical HPLC chromatograms used in determining the enan-
tioselectivity of the allylation reaction with bisphosphinites
8e and 8j. This material is available free of charge via the
Internet at http://pubs.acs.org.
J O9911387