Chiral phosphoramide-catalyzed enantioselective addition of allylic trichlorosilanes to aldehydes. Preparative studies with bidentate phosphorus-based amides

Article abstract of DOI:10.1021/jo052203h

On the basis of the mechanistic insight that more than one Lewis basic moiety (phosphoramide) is involved in the rate- and stereochemistry-determining step of enantioselective allylation, bidentate chiral phosphoramides were developed. Different chiral phosphoramide moieties were connected by tethers of methylene chains of varying length. The rate and enantioselectivity of allylation with allyltrichlorosilane promoted by the bidentate phosphoramides was found to be highly dependent on the tether length. A new phosphoramide based on a 2,2′-bispyrrolidine skeleton has been designed and afforded good yield, efficient turnover, and high enantioselectivity in allylation reactions. The synthesis of enantiopure 2,2′-bispyrrolidine was easily accomplished on large scale by photodimerization of pyrrolidine followed by resolution with L(or D)-tartaric acid. The scope of the allylation reaction was examined with variously substituted allylic trichlorosilanes and unsaturated aldehydes. This method has been applied to the construction of stereogenic, quaternary centers by the addition of unsymmetrically γ-disubstituted allylic trichlorosilanes.

Full text of DOI:10.1021/jo052203h

Chiral Phosphoramide-Catalyzed Enantioselective Addition of
Allylic Trichlorosilanes to Aldehydes. Preparative Studies with
Bidentate Phosphorus-Based Amides
Scott E. Denmark,* Jiping Fu, and Michael J. Lawler
Roger Adams Laboratory, Department of Chemistry, UniVersity of Illinois, Urbana, Illinois 61801
denmark@scs.uiuc.edu
ReceiVed October 21, 2005
On the basis of the mechanistic insight that more than one Lewis basic moiety (phosphoramide) is involved
in the rate- and stereochemistry-determining step of enantioselective allylation, bidentate chiral phosphor-
amides were developed. Different chiral phosphoramide moieties were connected by tethers of methylene
chains of varying length. The rate and enantioselectivity of allylation with allyltrichlorosilane promoted
by the bidentate phosphoramides was found to be highly dependent on the tether length. A new
phosphoramide based on a 2,2′-bispyrrolidine skeleton has been designed and afforded good yield, efficient
turnover, and high enantioselectivity in allylation reactions. The synthesis of enantiopure 2,2′-bispyrrolidine
was easily accomplished on large scale by photodimerization of pyrrolidine followed by resolution with
L(or D)-tartaric acid. The scope of the allylation reaction was examined with variously substituted allylic
trichlorosilanes and unsaturated aldehydes. This method has been applied to the construction of stereogenic,
quaternary centers by the addition of unsymmetrically γ-disubstituted allylic trichlorosilanes.
Introduction and Background
rosilanes 1 to aldehydes is promoted by chiral phosphoramide
4 to give chiral, nonracemic homoallylic alcohols 3 with high
diastereocontrol yet with only modest enantioselectivity. Al-
though the yield and enantioselectivity were modest, the reaction
was still judged to have great potential because (1) the uncata-
lyzed reaction does not proceed in the absence of promoter, (2)
the allylation displays high diastereocontrol suggesting a rigid,
closed transition structure, and (3) the reaction could be
promoted with a substoichiometric amount of Lewis base.
It is axiomatic that the invention of new, catalytic enantio-
selective transformations is inexorably bound to the design and
development of highly selective and efficient catalysts.1 In the
development Lewis base catalyzed, enantioselective allylation2
of aldehydes, ketones, and azomethine derivatives, the design
of chiral Lewis bases has met with modest success. Chirally
modified formamides, aromatic and aliphatic N-oxides, ureas,
diamines, sulfoxides, and phosphine oxides have been described
for this transformation.3 In 1994, we described the first chiral
Lewis base promoted allylation by the use of chiral phosphor-
amides for this transformation.4a Since that time, we have been
engaged in a broadly based program on the design and testing
of different structures of phosphoramides derived primarily from
chiral 1,2-diamines (Scheme 1). The addition of allyltrichlo-
SCHEME 1
(1) ComprehensiVe Asymmetric Catalysis; Jacobsen, E. N., Pfaltz, A.,
Yamamoto, H., Eds.; Springer-Verlag: Heidelberg, 1999; Vols. I-III.
(2) For reviews of catalytic enantioselective allylation, see: (a) Denmark,
S. E.; Fu, J. Chem. Commun. 2003, 167-170. (b) Denmark, S. E.; Fu, J.
Chem. ReV. 2003, 103, 2763-2793. (c) Yanagisawa, A. In ComprehensiVe
Asymmetric Catalysis; Jacobsen, E. N., Pfaltz, A., Yamamoto, H., Eds.;
Springer-Verlag: Heidelberg, 1999; Vol. II, Chapter 27. (d) Kobayashi,
S.; Sugiura, M.; Ogawa, C. AdV. Synth. Catal. 2004, 346, 1023-1034.
Unfortunately, extensive empirical optimization did not
provide either a better catalyst or even a framework for structural
10.1021/jo052203h CCC: $33.50 © 2006 American Chemical Society
Published on Web 01/17/2006
J. Org. Chem. 2006, 71, 1523-1536
1523

Denmark et al.
modification.5 This failure stimulated an in-depth mechanistic
study to probe the origins of activation and stereoselection. It
was demonstrated that the reaction proceeds through simulta-
neously operating pathways that involve either one or two
phosphoramides in the rate- and stereochemistry-determining
steps (Figure 1).4b Although mechanistically intriguing, these
dual mechanistic pathways are disadvantageous in a catalytic
process for both selectivity and reactivity reasons. First, because
the reaction is second order in catalyst, the rate of the reaction
falls off as the square of catalyst concentration. Second, at lower
catalyst loadings, a competing, less selective pathway can com-
promise the overall reaction selectivity. Third, structural modi-
fication of the catalyst often requires increasing steric bulk that
disfavors the binding of two catalyst molecules to the silicon
center.
FIGURE 2. Design of chiral phosphoramides.
Results
1. Bidentate Ligands Derived from (R,R)-1,2-trans-Cyclo-
hexanediamine. 1.1. Bisphosphoramides. Although prior stud-
ies in these laboratories provided a good notion of the geometric
requirements to accommodate cis coordination by phosphora-
mides,8 it was difficult to predict, a priori, the ideal type and
length of linker to connect the two participating functions. Thus,
our initial design of bidentate ligands focused on the utilization
of Type A chiral bisphosphoramides. Specifically, polymeth-
ylene tethers were utilized as linkages that would allow a broad
survey of optimal tether length for chelation. In addition,
(1R,2R)-N,N-dimethyl-1,2-trans-cyclohexanediamine was cho-
sen as the chiral phosphoramide subunit because the corre-
sponding monophosphoramide 4 provided one of highest
enantioselectivities in the allylation reaction promoted by
monophosphoramide.5 These considerations led to the formula-
tion of bisphosphoramides 5. In addition, monophosphoramide
6 was prepared to closely mimic the behavior of dimeric ligands
5 if they bound in only a monodentate fashion (Chart 1).
FIGURE 1. Dual mechanistic pathways for the allylation reaction.
To address these problems, we envisioned the utilization of
bidentate ligands with the expectation of increasing the effective
concentration of the second catalyst molecule through ap-
proximation.6 In designing these bidentate chiral phosphor-
amides, we envisioned two distinct types of bisphosphoramides
(Figure 2), namely (1) Type A, the two chiral phosphoramide
units separated with achiral linker, and (2) Type B, chiral
linkages employed to connect two achiral phosphoramide units.
With the conformational restriction provided by the linkage,
the selectivity and reactivity of the catalyst would depend on
not only the phosphoramide subunit but also the connecting
tether.7 We describe herein studies on the optimization of the
phosphoramide unit as well as the linkage that led to the
development of a highly selective catalyst.4b-d
CHART 1
The bisphosphoramides 5 were prepared by condensation
of (1R,2R)-N,N-dimethyl-1,2-trans-cyclohexanediamine 7 with
phosphorus oxychloride to afford the diaminophosphoryl chlo-
ride 8.9 This chiral diaminophosphoryl chloride was subse-
quently combined with dilithiated N,N-dimethyldiamine linkers
(n ) 2-6) (Scheme 2). The bisphosphoramides with shorter
tethers (5a (n ) 2), 5b (n ) 3), 5c (n ) 4)) were crystalline,
whereas the bisphosphoramides 5d (n ) 5) and 5e (n ) 6) were
(3) (a) Iseki, K.; Mizuno, S.; Kuroki, Y.; Kobayashi, Y. Tetrahedron
1999, 55, 977-988. (b) Hellwig, J.; Belser, T.; Muller, J. F. K. Tetrahedron
Lett. 2001, 42, 5417-5419. (c) Nakajima, M.; Saito, M.; Shiro, M.;
Hashimoto, S. J. Am. Chem. Soc. 1998, 120, 6419-6420. (d) Malkov, A.
V.; Bell, M.; Orsini, M.; Pernazza, D.; Massa, A.; Hermann, P.; Meghani,
P.; Kocovsky, P. J. Org. Chem. 2003, 68, 9659-9668. (e) Shimada, T.;
Kina, A.; Ikeda, S. Hayashi, T. Org. Lett. 2002, 4, 2799-2801. (f)
Chataigner, I.; Piarulli, U.; Gennari, C. Tetrahedron Lett. 1999, 40, 3633-
3634. (g) Angell, R. M.; Barrett, A. G. M.; Braddock, D. C.; Swallow, S.;
Vickery, B. D. Chem. Commun. 1997, 919-920. (h) Massa, A.; Malkov,
A. V.; Kocovsky, P. Scettri, A. Tetrahedron Lett. 2003, 44, 7179-7181.
(i) Traverse, J. F.; Zhao, Y.; Hoveyda, A. H.; Snapper, M. L. Org. Lett.
2005, 7, 3151-3154. (j) Pignataro, L.; Benaglia, M.; Cinquini, M.; Cozzi,
F.; Celentano, G. Chirality 2005, 17, 396-403. (k) Mu¨ller, C. A.; Hoffart,
T.; Holbach, M.; Reggelin, M. Macromolecules 2005, 38, 5375-5380. (l)
Nakajima, M.; Kotani, S.; Ishizuka, T.; Hashimoto, S. Tetrahedron Lett.
2005, 46, 157-159.
SCHEME 2
(4) (a) Denmark, S. E.; Coe, D. M.; Pratt, N. E.; Griedel, B. D. J. Org.
Chem. 1994, 59, 6161-6163. (b) Denmark, S. E.; Fu, J. J. Am. Chem. Soc.
2000, 122, 12021-12022. (c) Denmark, S. E.; Fu, J. J. Am. Chem. Soc.
2001, 123, 9488-9489. (d) Denmark, S. E.; Fu, J. Org. Lett. 2002, 4, 1951-
1953.
(5) Denmark, S. E.; Fu, J.; Coe, D. M.; Su, X.; Pratt, N. E.; Griedel, B.
D. J. Org. Chem. 2006, 71, 1513-1522.
(6) For another study on dimeric catalysts tethered with methylene units,
see: Konsler, R. G.; Karl, J.; Jacobsen, E. N. J. Am. Chem. Soc. 1998,
120, 10780-10781.
(7) For a review on the effect of ligand bite angle effect in phosphine
ligands, see: van Leeuwen, P. W. N. M.; Kamer, P. C. J.; Reek, J. N. H.;
Dierkes, P. Chem. ReV. 2000, 100, 2741-2769.
1524 J. Org. Chem., Vol. 71, No. 4, 2006

Chiral Phosphoramide-Catalyzed EnantioselectiVe Allylation
oils. For convenience of manipulation, stock solutions of
phosphoramides in CH2Cl2 were prepared and used in the
allylation reactions. The synthesis of 6 was accomplished by
direct condensation of the diamine 7 with the phosphoryl
dichloride 9.10
additive in the reaction. The use of i-Pr2NEt in reactions of
allylic trichlorosilanes was initially developed by Nakajima for
reactions catalyzed by bis-N-oxides, wherein it was found that
i-Pr2NEt could enhance the reaction rate without affecting the
enantioselectivity.3c Nakajima proposed that i-Pr2NEt functions
to assist turnover by releasing the catalyst from the product
trichlorosilyl ether. This effect is clearly seen by comparing
entries 7 and 9 (Table 1). In the reaction catalyzed by 0.1 equiv
of 5d, the addition of 5.0 equiv of i-Pr2NEt improved the
reaction yield to 73% without affecting the enantioselectivity.
1.2. Bisphosphonamides. The allylation reaction clearly
manifested the beneficial effect of uniting the two phosphor-
amide units through a diaminoalkyl chain presumably as a
consequence of chelation. Because phosphonic amides such as
10 are also effective promoters for the allylation reaction,5 the
dimeric phosphonic amides 11 tethered with various methylene
units were prepared as well (Chart 2). The bidentate ligands
11b-d were prepared by treatment of the lithiated phosphon-
amide 1211 with 1,n-dihalides (Scheme 3). Attempts to syn-
thesize 11a by adding the anion to 1,2-diiodoethane, however,
did not provide the desired dimeric product. Instead, 11a was
synthesized by condensation of diamine 7 with bisphosphonic
acid chloride 13 in the presence of triethylamine.
With these new phosphoramides in hand, the addition of
allyltrichlorosilane (1a) to benzaldehyde (2) was then carried
out, and the results are collected in Table 1. Because compounds
5 are bisphosphoramides, 0.5 equiv was used for comparison
to the monophosphoramide 6. The dependence of the product
er on the linker length was striking. Whereas the monophos-
phoramide 6 gave an er of 75.5/24.5, nearly racemic product
was obtained with 5a (entries 1 and 2). A dramatic increase in
er was seen in the change from 5c, where the er of the product
was only 58.5/41.5, to 5d, wherein the er was 82.5/17.5 (entries
4 and 6). The er decreased again to 73.0/26.0 when 5e was used
(entry 8). This behavior implies a cooperativity of binding with
bisphosphoramides, because if there were no chelation with
silicon, the er of 5 would be chain-length independent and
similar to that obtained with 6. For reasons that were not at all
obvious, bisphosphoramide 5d tethered with five methylene
units gave the highest enantioselectivity.
TABLE 1. Allylation with 1a Promoted by Bisphosphoramidesa
CHART 2
entry
promoter
tether, n
equiv
erb
yield, %
1
2
3
4
5
6
7
8
9c
6
1.0
0.5
0.5
0.5
0.1
0.5
0.1
0.5
0.1
75.5/24.5
50.0/50.0
62.5/17.5
58.5/41.5
55.0/45.0
82.5/17.5
86.0/14.0
73.0/26.0
86.0/14.0
73
60
72
82
52
78
54
75
73
5a
5b
5c
5c
5d
5d
5e
5d
2
3
4
4
5
5
6
5
SCHEME 3
a Reaction performed at 1.0 M concentration at -78 °C for 6 h using
100% ee promoter. b Determined by CSP-SFC. c 5.0 equiv of i-Pr2NEt was
used as an additive.
Foregoing studies with monophosphoramide 4 showed that
the allylation selectivity could be dependent on the promoter
loading.5 Thus, the dependence of the product enantiomeric
composition on the promoter loading (concentration) was then
studied with bisphosphoramides 5c and 5d, which gave the
lowest and highest enantioselectivities, respectively. Entries 4-5
in Table 1 show that when the loading of 5c was decreased to
0.1 equiv, the er of the product obtained also decreased from
58.5/41.5 to 55.0/45.0. On the other hand, a similar lowering
of the loading of 5d (entries 6-7) brought about an increase in
the er of 3aa from 82.5/17.5 to 86.0/14.0. It was rather surprising
that in the reaction promoted by 5d, even higher ee was achieved
at lower promoter concentration, but this clearly suggested a
different mode of binding of 5d compared to 4 and 5c.
The addition of 1a to 2 was then carried out with 0.1 equiv
of catalyst at -78 °C, and the results are collected in Table 2.
The reactions gave rather low yields of 3aa with a catalytic
amount of promoter. In the reaction catalyzed by bidentate
ligands 11, the enantioselectivity was also found to be dependent
on the tether length. The magnitude of er variation in 11 was
much less than that with chiral bisphosphoramides 5. The
bisphosphonic diamide with four methylene tether 11a gave an
er of 77.0/23.0, and the er increased to 87.0/13.0 with 11b
(entries 2 and 3). The enantioselectivity decreased to 81/19 when
11c was used, and the highest selectivity (89.0/11.0 er) was
observed with 11d tethered by seven methylene units (entires
4 and 5). Again, the effect of tether length on the enantiose-
lectivity and the enhanced enantioselectivity observed with 11d
Finally, to further improve the yield of 3aa obtained with
substoichiometric amounts of 5d, i-Pr2NEt was used as an
(8) For previous studies of monophosphoramide/Lewis acid complexes,
see: Denmark, S. E.; Su, X. Tetrahedron 1999, 55, 8727-8738.
(9) Alexakis, A.; Mutti, S.; Magneney, P. J. Org. Chem. 1992, 57, 1224-
1237.
(10) Denmark, S. E.; Su, X.; Nishigaichi, Y.; Coe, D. M.; Wong, K.-T.;
Winter, S. B. D.; Choi, J. Y. J. Org. Chem. 1999, 64, 1958-1967.
(11) Lawrence, R. M.; Biller, S. A.; Dickson, J. K., Jr.; Logan, J. V. H.;
Magnin, D. R.; Sulsky, R. B.; DiMarco, J. D.; Gougoutas, J. Z.; Beyer, B.
D.; Taylor, S. C.; Lan, S.-J.; Cioseck, C. P., Jr.; Harrity, T. W.; Jolibois,
K. G.; Kunselman, L. K.; Slusarchyk, D. A. J. Am. Chem. Soc. 1996, 118,
11668-11669.
J. Org. Chem, Vol. 71, No. 4, 2006 1525

Denmark et al.
amine-based bisphosphoramides 5.10 Thus, starting from diamine
16, formation of diaminophosphoryl chloride 17 (from POCl3),9
followed by reaction with the dilithiated N,N′-dimethyldiamine
linker gave the bisphosphoramide 14 in 68% yield (Scheme 4).
The synthesis of 15 was achieved in a one-pot procedure, in
which the two phosphorus atoms were linked with the diamine
at the P(III) stage.10
TABLE 2. Addition of Allyltrichlorosilane to Benzaldehyde
Catalyzed by Bidentate Phosphonic Amidesa
entry
cat.
yield, %
erb
1
2
3
4
5
10
27
37
42
40
39
83.0/17.0
77.0/23.0
87.0/13.0
81.0/19.0
89.0/11.0
SCHEME 4
11a
11b
11c
11d
a Reaction performed at -78 °C for 6 h at 1.0 M concentration.
b Determined by CSP-SFC.
again suggested the operation of chelation. However, in this
series, high enantioselectivity was also seen with the monomeric
phosphonamide 10, so the magnitude of the cooperativity is less
compelling than with the phosphoramide series 5.
2. Bidentate Ligands Derived from Stilbene and Binaph-
thyl Diamines. The chain length dependence observed with
bidentate ligands 5 successfully established that the two phos-
phoramide subunits can behave cooperatively to enhance selec-
tivity. Accordingly, we next investigated the development of
Type A bisphosphoramides bearing a pentamethylene tether that
would bring together different phosphoramide units to further
improve the stereoselectivity. In contemporaneous investigations
in these laboratories on the catalytic enantioselective aldol
addition reaction,12 we have documented good success with both
monomeric13 and dimeric14 phosphoramide catalysts derived
from readily available diamines (R,R)-1,2′-stilbenediamine and
(R)-1,1′-binaphthyl-2,2′-diamine (Chart 3). Thus, the bisphos-
phoramides 14 and 15 derived were prepared and tested in the
allylation.
These two new bisphosphoramides, unfortunately, were rather
ineffective in the allylation reaction, giving low enantioselec-
tivities and/or reactivities. In the test reaction with 1a and 2,
catalyst 14 (5 mol %) provided 3aa in 82% yield but only with
60.0/40.0 er. With catalyst 15 (5 mol %), the reaction was rather
sluggish, affording 3aa in only 4% yield.
3. Bisphosphoramides Linked by Chiral Diamines (Type
B). Although seven atom tethers were found to be the ideal
tether length for chelation, the Type B chiral bisphosphor-
amide 20, was first tested in which two achiral subunits are
separated by six atoms comprising four carbon and two nitrogen
atoms. The bisphosphoramide 20 was synthesized from 1,1′-
binaphthyl-2,2′-diamine 18 by coupling the dilithiated diamide
with 19 followed by oxidation with (TMSO)2 and m-CPBA
(Scheme 5).
CHART 3
SCHEME 5
Phosphoramide 14 was synthesized in a straightforward
manner by the route developed for the trans-1,2-cyclohexanedi-
The bisphosphoramide 20 catalyzed the addition of 1a to
2 and provided the adduct 3aa with up to 90/10 er, and
more importantly, the enantioselectivity was independent of
the loading of the catalyst (Table 3). The insensitivity of the
enantioselectivity on the catalyst strongly supported the opera-
tion of chelation. Under the optimized reaction condition with
10 mol % of 20 and 5.0 equiv of i-Pr2NEt as an additive,
the homoallylic alcohol 3aa was obtained in 67% yield and
90.0/10.0 er.
4. Bidentate Ligands Derived from 2,2′-Bispyrrolidine and
2,2′-Bispiperidine. 4.1. Design and Preparation. The failure
to improve the enantioselectivity of allylation by empirical
modification with conventional diamines (in Type A or Type
B ligands) suggested the need to better understand the origin
(12) Denmark, S. E.; Fujimori, S. In Modern Aldol Reactions; Mahrwald,
R., Ed.; Wiley-VCH: Weinheim, 2004; Vol. 2, Chapter 7.
(13) Monophosphoramides derived from 14 and 15 have provided
excellent enantioselectivities in reactions with other chlorosilane species
such as the aldol reaction of enoxytrichlorosilanes and epoxide-opening
reactions (silicon tetrachloride). For aldol reaction, see: (a) Denmark, S.
E.; Stavenger, R. A. Acc. Chem. Res. 2000, 33, 432-440. Epoxide-opening
reaction, see: (b) Denmark, S. E.; Barsanti, P. A.; Wong, K.-T.; Stavenger,
R. A. J. Org. Chem. 1998, 63, 2428-2429.
(14) These bisphosphoramides have been employed with good success
in aldol addition reactions of enoxytrichlorosilanes (14) and Mukaiyama-
type aldol addition with silicon tetrachloride (15). See: (a) Denmark, S.
E.; Pham, S. M. J. Org. Chem. 2003, 68, 5045-5055. (b) Denmark, S. E.;
Beutner, G. L.; Wynn, T.; Eastgate, M. D. J. Am. Chem. Soc. 2005, 127,
3774-3789.
1526 J. Org. Chem., Vol. 71, No. 4, 2006

Chiral Phosphoramide-Catalyzed EnantioselectiVe Allylation
dihydroxylation of olefins17 and also for catalytic enantiose-
lective additions to nitroalkenes.18 Several syntheses of enan-
tiopure 2,2′-bispyrrolidine have been previously reported. Most
of these routes required a multiple-step synthesis from chiral
starting materials. For example, the synthesis developed by
Hirama and co-workers requires 11 steps from D-tartaric acid.17c
Alternatively, Alexakis has developed a synthesis of (R,R)-2,2′-
bispyrrolidine by an organometallic addition to a chiral diimine
that required five steps.18b In a simpler synthesis reported by
Masamune, the dl/meso mixture 2,2′-bispyrrolidine was syn-
thesized in two steps, which was then resolved with tartaric
acid.19 In our initial attempts to prepare this compound, the
Masamune approach was chosen for the initial synthesis of
(R,R)-2,2′-bispyrrolidine because of its apparent simplicity
(Scheme 6). Condensation of pyrrole with 2-pyrrolidinone under
the action of phosphorus oxychloride gave 23, which was
subsequently hydrogenated with rhodium on alumina to provide
a dl/meso mixture of 2,2′-bispyrrolidine 21. The hydrogenation
step, however, was found to be very sluggish and irreproducible.
When carried out on a 12 g scale, the reaction required 5 days
to achieve a 90% conversion. Furthermore, attempts to scale-
up the synthesis to 100 g led to even slower reactions, in which
only 40% conversion was obtained after 30 days.
TABLE 3. Addition of Allyltrichlorosilane Catalyzed by
Bisphosphoramide 20a
entry
loading
additive
yield, %
erb
1
2
3
0.5
0.1
0.1
76
49
67
90.0/10.0
90.0/10.0
90.0/10.0
i-Pr2NEt
a Reaction performed at 1.0 M concentration at -78 °C for 6 h.
b Determined by CSP-SFC.
of asymmetric induction in this process. Thus, to refine our
understanding and assist in the design of more selective catalysts,
we first chose to study the structure of our Group 14 Lewis
acid complexes of the chiral phosphoramides. Because no
suitable crystals could be obtained from silicon-based Lewis
acids,15 we employed tin tetrachloride as a surrogate, fully
cognizant of the difference between silicon and tin.16 For this
study, we selected one of the more selective bisphosphoramide
catalysts, 5d. Examination of the X-ray crystal structure of 5d‚
SnCl4 revealed that the disposition of the internal, N-methyl
substituents was significantly influenced by the cyclohexane-
1,2-diamine skeleton (Figure 3a). We reasoned that connecting
the substituent on the stereogenic center to the nitrogen atom
by enclosure in a ring should enforce a more rigid control of
the orientation of the N-substituents and thus impose a more
highly dissymmetric coordination environment. This notion of
backbone-induced nitrogen distortion is presented in Figure 3b,c.
To accomplish this structural modification, we envisioned the
utilization of 2,2′-bispyrrolidine (R,R)-21 and bispiperidine
(R,R)-22 (Chart 4) as entries into a new class of chiral diamine-
derived phosphoramides.
SCHEME 6
These problems encouraged us to reconsider an alternative
preparation of 21 that was initially dismissed as being less
preparatively useful. In 1989, Crabtree reported an intriguing
oxidative photodimerization of amines.20 It was reported that
2,2-bispyrrolidine could be obtained by heating neat pyrrolidine
to reflux in under ultraviolet irradiation in the presence of a
trace amount of mercury. Surprisingly, despite many efforts on
the synthesis of (R,R)-2,2′-bispyrrolidine, this method had not
been developed. If this reaction were scalable, then given the
direct resolution of the resulting 2,2′-bispyrrolidine established
by Masamune, it would provide a very convenient, practical
method for the synthesis of (R,R)-2,2′-bispyrrolidine.
Ultimately, it was found that the photodimerization reaction
was amenable to scale-up with a rather simple apparatus: a 500-
mL flask equipped with quartz refluxing columns and water
condensers (Scheme 7).21 To this flask were added 200 mL of
FIGURE 3. (a) Chem 3D presentation of 5d‚SnCl4, hydrogens omitted
for clarity. (b) Nitrogen distortion in the ring-fused phosphoramide tin
complex. (c) Up-down geometry of phosphoramide in the hypothetical
phosphoramide tin complex.
CHART 4
(17) (a) Hirama, M.; Oishi, T.; Ito, S. J. Chem. Soc., Chem. Commun.
1989, 665-666. (b) Oishi, T.; Hirama, M. J. Org. Chem. 1989, 54, 5834-
5835. (c) Kotsuki, H.; Kuzume, H.; Ghoda, T.; Fukuhara, M.; Ochi, M.;
Oishi, T.; Hirama, M.; Shiro, M. Tetrahedron: Asymmetry 1995, 6, 2227-
2236.
(18) (a) Andrey, O.; Alexakis, A.; Tomassini, A.; Bernarindelli, G. AdV.
Synth. Catal. 2004, 346, 1147-1168. (b) Alexakis, A.; Tomassini, A.;
Chouillet, C.; Roland, S.; Mangeney, P.; Bernardinelli, G. Angew. Chem.,
Int. Ed. 2000, 39, 4093-4095.
(R,R)-2,2′-Bispyrrolidine 21 was initially developed by
Hirama and has been successfully applied in the asymmetric
(19) Oishi, T.; Hirama, M.; Sita, L. R.; Masamune, S. Synthesis 1991,
789-792.
(15) For a recent X-ray crystal structure of a bis-N-oxide complex of
silicon tetrachloride, see: Denmark, S. E.; Fan, Y.; Eastgate, M. D. J. Org.
Chem. 2005, 70, 5235-5248.
(20) (a) Ferguson, R. R.; Boojamra, C. G.; Brown, S. H.; Crabtree, R.
H. Heterocycles 1989, 28, 121-124. (b) Krajnik, P.; Ferguson, R. R.;
Crabtree, R. H. New J. Chem. 1993, 17, 559-566.
(16) Denmark, S. E.; Fu, J. J. Am. Chem. Soc. 2003, 125, 2208-2216.
(21) Denmark, S. E.; Fu, J.; Lawler, M. J. Org. Synth. 2005, 83, in press.
J. Org. Chem, Vol. 71, No. 4, 2006 1527

Denmark et al.
CHART 5
pyrrolidine, a drop of mercury, and boiling chips. The flask
was placed in a Rayonet photoreactor with 14 × 8-W low-pres-
sure Hg lamps (254 nm). The mixture was then heated to reflux
before the lamps were turned on, after which the mixture was
heated to reflux under irradiation for 10 days. After the unreacted
pyrrolidine was separated by simple distillation, 84.0 g of the
2,2′-bispyrrolidine 21 was obtained by vacuum distillation.
The crude product containing a 1/1 dl/meso mixture of 2,2′-
bispyrrolidine was of sufficient purity to continue to the
resolution step. Addition of L-tartaric acid to an aqueous solu-
tion of 2,2′-bispyrrolidine produced (R,R)-2,2′-bispyrrolidine‚
L-tartaric salt as prismatic crystals, which after neutralization
gave (R,R)-21 in 55% overall yield (based on isomer content).
The mother liquor from the resolution step was also basified to
recover the rest of 2,2′-bispyrrolidine, from which (S,S)-21 was
obtained in 48% yield with the use of D-tartaric acid as the
resolving reagent. Thus, through sequential resolution using L-
or D-tartaric acid, both (R,R)-21 and (S,S)-21 could be obtained
in enantiopure form in two steps.
SCHEME 9
SCHEME 7
By the same method, dl/meso-2,2′-bispiperidine 22 (dl/meso,
1/1) was synthesized on a large scale (Scheme 8). To simplify
the resolution of dl-2,2′-bispiperidine, diastereomerically pure
dl-2,2′-bispiperidine was required.22 Thus, the dl/meso mixture
of 2,2′-bispiperidines was derivatized with 4-nitrobenzaldehyde,
and the chiral diastereomer 24 was obtained in 31% yield after
chromatographic separation.23 Hydrolysis of 24 with an aqueous
solution of HCl provided the d,l-2,2′-bispiperidine (85%), which
was then resolved with L-tartaric acid to give (R,R)-22 in 81%
yield (based on available isomer content).
5, Scheme 9). As shown in Scheme 9, formation of phosphoryl
chlorides 28 and 29 with POCl3 from diamines (R,R)-21/22,
respectively, followed by reaction with dilithiated N,N′-dimethyl-
1,n-diamine linkers provided the bisphosphoramides 25c-e and
27. In the phosphoramides derived from 2,2′-bispyrrolidine, the
monophosphoramide 26 was also prepared for comparison. The
synthesis of 26 was achieved by direct condensation of (R,R)-
21 with piperidinophosphoric dichloride 30.
4.2. Allylations with (R,R)-25, (R,R)-26, and (R,R)-27.
These new phosphoramides were then evaluated in the addition
of allyltrichlorosilane (1a) to benzaldehyde (2) in the presence
of 5.0 equiv of i-Pr2NEt was an additive. With 20 mol % of
monophosphoramide 26 as the catalyst, the adduct was obtained
only in modest yield and enantioselectivity (Table 4, entry 1).
SCHEME 8
TABLE 4. Allylation of Benzaldehyde Catalyzed by
Phosphoramidesa
entry
catalyst
erb (S/R)
yield,c %
1d
2
3
4
5
26
78.0/22.0
59.0/41.0
93.5/6.5
83.5/16.5
79.0/21.0
56
54
85
58
72
25c
25d
25e
27
With enantiopure diamines (R,R)-21 and (R,R)-22 in hand,
the bisphosphoramides 25c-e and 27 were synthesized (Chart
a All reactions run at 1.0 M concentration in CH2Cl2/i-Pr2NEt, 1/1 at
-78 °C for 8 h, using 5 mol % catalyst. b Determined by CSP-SFC. c Yield
of chromatographically homogeneous product. d 20 mol % of catalyst was
used.
(22) Sato, M.; Sato, Y.; Yano, S.; Yoshikawa, S. J. Chem. Soc., Dalton
Trans. 1985, 895-898.
(23) Chivers, P. J.; Crabb, T. A.; Williams, R. O. Tetrahedron 1968,
24, 4411-4421.
1528 J. Org. Chem., Vol. 71, No. 4, 2006

Chiral Phosphoramide-Catalyzed EnantioselectiVe Allylation
In the reaction catalyzed by bisphosphoramides 25, a depen-
dence of product enantioselectivity on tether length was
observed. The enantioselectivity variation on the tether length
followed the same trend as that seen with bisphosphoramides
5. Whereas 25c gave only 59.0/41.0 er in the reaction, 25d,
bearing a five-methylene tether dramatically improved the
enantioselectivity to 93.5/6.5 er (entries 2 and 3). In this case,
the enantioselectivity observed with 25d was significantly higher
than that obtained with 5d and 11d derived from trans-
cyclohexane-1,2-diamine, which clearly illustrated the success
of the design. When 25e was used, the enantioselectivity
decreased slightly to 83.5/16.5 er. In addition, the bisphos-
phoramide 25d proved to be much more reactive than the
bisphosphoramides with other tethers as well as the monophos-
phoramide, giving an 85% yield with only 5 mol % catalyst
loading. Compared to 25d, the bisphosphoramide 27d derived
from 2,2-bispiperidine, however, provided a lower selectivity
and yield (entry 5).
zaldehyde, cinnamaldehyde, and 2-furaldehyde gave slightly
diminished enantioselectivities.
An important demonstration of scope was in the extension
to the reactions of γ-substituted allylic trichlorosilanes. An
excellent correlation of geometrical homogeneity of the silanes
with the diastereomeric composition of the products ((E)-1b f
anti-3; (Z)-1b f syn-3) was seen. The results in Table 5 show
clearly that (Z)-1b leads to much higher enantioselectivity
compared to (E)-1b (compare entries 7-9 and 10-12). Fur-
thermore, γ-disubstituted allylic trichlorosilane 1c also reacted
under these conditions to provide prenylation products with
excellent selectivity (Table 5, entries 17-19). Apparently, the
Z-substituent on the allylic trichlorosilane has a beneficial
effect as evidenced by the highly selective syn-butenylation
and prenylation processes. Further, electron-rich aromatic al-
dehydes seemed to react with higher enantioselectivities com-
pared to electron-deficient aromatic substrates (cf. entries 3 vs
4, 12 vs 13).
5. Scope of the Reaction. The successful design and
implementation of bisphosphoramide 25d derived from (R,R)-
2,2′-bispyrrolidine demonstrated our understanding of the origin
of stereochemical induction in the allylation reactions and
produced high enantioselectivity in the addition of allyltrichlo-
rosilane to benzaldehyde. To establish the generality of the
bisphosphoramide 25d in the catalytic enantioselective allylation
reaction, a survey of aldehydes and allylic trichlorosilanes was
undertaken, and the results are collected in Table 5. Under the
established reaction condition with 5 mol % of 25d, 1a
underwent addition to aromatic (entries 1-4), heteroaromatic
(entry 6), and unsaturated aldehydes (entry 5) to give homoal-
lylic alcohols in good selectivities and yields. 2-Naphthaldehyde
and 4-methoxybenzaldehyde provided enantioselectivities simi-
lar to that with benzaldehyde, whereas 4-trifluoromethylben-
6. Quaternary Stereogenic Centers. The successful and
highly enantioselective addition of 1c promoted by 25d, together
with the strong stereochemical coupling of geometry with
diastereoselectivity (for (E)-1b and (Z)-1b), suggested the
opportunity for enantioselective construction of quaternary
stereogenic centers by the addition of unsymmetrical γ-disub-
stituted allylic trichlorosilanes to aldehydes.24 Trisubstituted
silanes (E)-33 and (Z)-33 were chosen as test substrates and
were readily synthesized from commercial available geraniol
(E)-31 and nerol (Z)-31, respectively (Scheme 10). The geo-
metrically homogeneous alcohols (E)- and (Z)-31 were con-
verted to the corresponding chlorides (E)-32 and (Z)-32 by the
Corey-Kim procedure25a,b or with PPh3/CCl425c which were then
treated with diisopropylethylamine, trichlorosilane, and a cata-
lytic amount of CuCl to provide allylic trichlorosilanes (E)-33
and (Z)-33 also in geometrically homogeneous form.26 The
catalyzed additions of (E)-33 and (Z)-33 to benzaldehyde
provided adducts anti-34 and syn-34 with excellent diastereo-
and enantioselectivities (Scheme 10).
TABLE 5. Addition of Allylic Trichlorosilanes to Aldehydes
Catalyzed by 25da
The configurational assignment in the formation of quaternary
centers was unambiguously confirmed by X-ray crystallographic
analysis of the adduct of (E)-33 with 2-naphthaldehyde. The
addition of (E)-33 to 2-naphthaldehyde catalyzed by (R,R)-25d
provided the homoallylic alcohol 35 with high diastereo- and
enantioselectivity (Scheme 11). Treatment of the 35 with
4-bromobenzoyl chloride and a stoichiometric quantity of
DMAP in the presence of Et3N led to the acylated product 36.
Repeated recrystallization provided enantiomerically pure 36
which was suitable for single-crystal X-ray diffraction.27 The
yield,c
%
entry silanesb
R
product
syn/antid
ere
1
2
3
4
5
6
7
8
9
1a
1a
1a
1a
1a
1a
C6H5
3aa
3ab
3ac
3ad
3ae
3af
3ba
3bb
3be
3ba
3bb
3bc
3bd
3be
85
92
84
79
84
59
82
83
59
89
88
91
85
78
62
82
89
71
71
93.5/6.5
93.4/6.6
94.0/6.0
90.1/9.9
90.7/9.3
90.3/9.7
2-naphthyl
4-CH3OC6H4
4-CF3C6H4
(E)-C6H5CHdCH
2-furyl
(E)-1b C6H5
(E)-1b 2-naphthyl
(E)-1b (E)-C6H5CHdCH
1/99 92.8/7.2
1/99 90.4/9.6
1/99 90.4/9.6
99/1 96.7/3.3
99/1 96.8/3.2
99/1 97.3/2.7
99/1 92.3/7.7
99/1 94.0/6.0
95/5 96.1/3.9
99/1 97.8/2.2
98.0/2.0
(24) For reviews on enantioselective construction of quaternary stereo-
centers, see: (a) Christoffers, J.; Barao, A. AdV. Synth. Catal. 2005, 347,
1473-1482. (b) Douglas, C. J. Overman, L. E. Proc. Nat. Acad. Sci. U.S.A.
2004, 101, 5363-5367. (c) Christoffers, J.; Mann, A. Angew. Chem., Int.
Ed. 2001, 40, 4591-4597. (d) Corey, E. J.; Guzman-Perez, A. Angew.
Chem., Int. Ed. 1998, 37, 388-401. (e) Fuji, K. Chem. ReV. 1993, 93, 2037-
2066.
(25) (a) Corey, E. J.; Kim, C. U.; Takeda, M. Tetrahedron Lett. 1972,
13, 4339-4342. (b) Davisson, V. J.; Woodside, A. B.; Neal, T. R.; Stremler,
K. E.; Muehlbacher, M.; Poulter, C. D. J. Org. Chem. 1986, 51, 4768-
4779. (c) Calzada, J. G.; Hooz, J. Organic Syntheses; Wiley: New York,
1988; Collect. Vol. VI, pp 634-637.
(26) Kobayashi, S.; Nishio, K. J. Org. Chem. 1994, 59, 6620-6628.
(27) The crystallographic coordinates of 36 have been deposited with
the Cambridge Crystallographic Data Centre; deposition no. CCDC 165491.
These data can be obtained free of charge via www.ccdc.cam.ac.uk/conts/
retrieving.html (or from the Cambridge Crystallographic Data Centre, 12
Union Road, Cambridge CB2 1EZ, U.K.; fax: (+44) 1223-336-033; or
deposit@ccdc.cam.ac.uk).
10 (Z)-1b C6H5
11 (Z)-1b 2-naphthyl
12 (Z)-1b 4-CH3OC6H4
13 (Z)-1b 4-CF3C6H4
14 (Z)-1b (E)-C6H5CHdCH
15 (Z)-1b (E)-C6H5CHdC(CH3) 3bg
16 (Z)-1b 2-furyl
17 1c
18 1c
19 1c
3bf
3ca
3ce
3cf
C6H5
(E)-C6H5CHdCH
2-furyl
94.1/5.9
97.3/2.6
a Reaction performed at -78 °C for 8-10 h with 5 mol % of 25d. b (E)-
1b and (Z)-1b both greater than 99/1 isomerically pure by 1H NMR analysis.
c Yield of chromatographically homogeneous product. d Determined by 1H
NMR (400 or 500 MHz) analysis. e Determined by CSP-SFC or CSP-GC.
J. Org. Chem, Vol. 71, No. 4, 2006 1529

Denmark et al.
SCHEME 10
configuration of 36 was determined to be (1S,2S). The sense of
relative stereoinduction clearly supported the intermediacy of
a siliconate complex which reacts through a closed, chair-
like transition structure. In addition, the observation that the
major adduct had the S configuration at the hydroxyl center is
also consistent with the results obtained in the addition of
allylsilanes 1.
12). LY426965 belongs to a family of arylpiperazines that
are effective pharmaceutical agents for the treatment of con-
ditions related to or affected by the serotonin 1A recep-
tor. Preclinical studies indicated that LY426965 is a selective,
full 5-HT1A antagonist that may have clinical use as phar-
macotherapy for smoking cessation and depression-related
disorders.
The key steps in the synthesis of LY426965 are shown in
Scheme 12.4d To generate the S configuration of the quaternary
center required the combination of E-configured allylic silane
40 under the action of chiral catalyst (S)-(l,l)-25. The synthesis
of (E)-39 began by zirconium-catalyzed carbometalation of
phenylacetylene 37 followed by activation and trapping of the
resulting alane with formaldehyde to provide the homoallyl
alcohol 38 as a single geometrical isomer.29 Alcohol 38 was
converted by the Corey-Kim procedure25 to the corresponding
chloride 39, which was then treated with diisopropylethylamine,
trichlorosilane, and a catalytic amount of CuCl to give trichlo-
rosilane 40 in geometrically homogeneous form in 78% yield
for the two steps. The addition of this chlorosilane to benzal-
dehyde was effectively catalyzed by (S)-(l,l)-25d and provided
adduct 41 with excellent diastereoselectivity (anti/syn 99/1) and
enantioselectivity (96.5/3.5). In this case, it was found that the
addition of 0.2 equiv of tetrabutylammonium iodide could
improve the yield without affecting the selectivity.30 Selective
hydroboration/oxidation of the terminal double bond afforded
intermediate 42.
SCHEME 11
Because geometrically defined γ-disubstituted allylic alcohols
are widely accessible, this method represents a versatile route
for the construction of quaternary stereogenic centers. As a
demonstration of such a method in synthesis, it was applied to
the enantioselective synthesis of serotonin antagonist (LY426965)
(46)28 which contains an R-carbonyl quaternary center (Scheme
The final stages of the synthesis of LY426965 required
selective saturation of one of the phenyl rings. This maneuver
could be accomplished by a two-stage hydrogenation, first with
rhodium on alumina31 and then at higher pressure with platinum
on carbon to afford diol 43. Adjustment of the oxidation state
under Swern oxidation conditions afforded keto aldehyde 44,
(28) (a) Rasmussen, K.; Calligaro, D. O.; Czachura, J. F.; Dreshfield-
Ahmad, L. J.; Evans, D. C.; Hemrick-Luecke, S. K.; Kallman, M. J.;
Kendrick, W. T.; Leander, J. D.; Nelson, D. L.; Overshiner, C. D.;
Wainscott, D. B.; Wolff, M. C.; Wong, D. T.; Branchek, T. A.; Zgombick,
J. M.; Xu, Y.-C. J. Pharmacol. Exp. Ther. 2000, 294, 688-700. (b) Godfrey,
A. G.; Kohlman, T. D.; O’Toole, J. C.; Xu, Y.-C.; Zhang, T. Y. U.S. Pat.
Appl. Publ. US2004044009, 2004.
(29) Okukado, N.; Negishi, E.-I. Tetrahedron Lett. 1978, 19, 2357-
2360.
(30) (a) Short, J. D.; Attenoux, S.; Berrisford, D. J. Tetrahedron Lett.
1997, 38, 2351-2354. (b) Denmark, S. E.; Su, X.; Nishigaichi, Y. J. Am.
Chem. Soc. 1998, 120, 12990-12991.
(31) Stocker, J. H. J. Org. Chem. 1962, 27, 2288-2289.
1530 J. Org. Chem., Vol. 71, No. 4, 2006

Chiral Phosphoramide-Catalyzed EnantioselectiVe Allylation
SCHEME 13
which underwent clean reductive amination with commercially
available 4-(2-methoxyphenyl)piperazine (45) to complete the
synthesis of LY426965.
SCHEME 12
be in the form of the R-chloro silyl ether, many stronger
nucleophiles were capable of providing addition products, such
as trichlorosilyl ketene acetals,15 TBS ketene acetals,14b con-
jugated ketene acetals,14b,33 and isocyanides.34 Clearly, the
R-chloro silyl ether iii exists in an equilibrium with the acti-
vated aldehyde ii. Thus, depending upon the position of that
equilibrium and the reactivity of the competing nucleophile,
measurable addition rates can be seen. However, because 1 and
its relatives are on the lower end of the nucleophilicity scale,35
our efforts to promote addition focused on shifting the posi-
tion of the equilibrium toward ii. We envisioned that sequester-
ing the chloride anion would affect the equilibrium to favor
the complexed aldehyde and allow addition of the allyl group.
Strategies such as ion pairing and solvation effects were
considered.
Clearly, a more direct approach to LY426965 would have
been the addition of (E)-39 to cyclohexanecarboxaldehyde.
Unfortunately, when this aldehyde was employed under the
standard allylation conditions, no desired product was formed.
This outcome was not unexpected from previous studies on the
scope of the addition reaction of simpler allylic silanes 1. This
limitation has plagued the Lewis base catalyzed allylation
reactions of allylic trichlorosilanes, and therefore, a systematic
study on the origins and potential solutions was initiated.
7. Allylation of Aliphatic Aldehydes. Under conditions by
which aromatic and unsaturated aldehydes successfully reacted
with allylic trichlorosilanes, aliphatic aldehydes did not under-
go addition. This is surprising as aliphatic aldehydes should be
more reactive substrates. Careful analysis of the reaction
mixtures by 1H NMR spectroscopy revealed that aliphatic
aldehydes are immediately consumed in the formation of an
R-chloro silyl ether iii, which is unreactive toward the allylating
reagents (Scheme 13). In the preceding paper, it was shown
that the mechanism of the allylation involves an ionization of
a chloride ion from 1 to generate a cationic silicon species i.
Binding of i to the aldehyde oxygen constitutes activation of
the carbonyl group in species ii. With aromatic aldehydes, ii
proceeds directly to the product (aromatic aldehydes show no
tendency to form R-chloro silyl ethers by 1H NMR spectroscopic
analysis), whereas with aliphatic aldehydes, collapse of the
zwitterion is much faster and only R-chloro silyl ether iii is
observed.
The chloride anion is known to form ate complexes with
various inorganic salts that have high chloride affinities.36 Thus,
we devised an empirical survey of the effect of various inorganic
salts on the rate, yield, and selectivity of the addition of 1 to an
aliphatic aldehyde. Hydrocinnamaldehyde was chosen as a
representative aliphatic aldehyde, and the allylation was carried
out in CH2Cl2 at 0.2 M concentration at room temperature for
1 h. After quenching the reactions with KF and aqueous workup,
the reaction mixtures were subjected to GC analysis (internal
standard) to determine the conversion (Table 6).
In absence of an inorganic salt, HMPA-promoted the allyl-
ation of hydrocinnamaldehyde at a rather slow rate at room
temperature, reaching only 6% conversion after 2 h (Table 6).
With addition of inorganic salts, the reaction was enhanced
(entries 3-14), particularly with 10 mol % of HgCl2 or SbCl3
for which up to 28% conversion was achieved.
In subsequent studies to establish the optimal loading of the
salt (entries 15-18), it was found that the reaction rate
significantly dropped when the loading of HgCl2 was decreased
to 2 mol %. However, an increase of loading from 10 mol % to
50 mol % only gave a slightly improved reaction rate.
In addition, the effect of phosphoramide stoichiometry and
structure on the reaction rate was also investigated (entries 18-
(33) Denmark, S. E.; Heemstra, J. H., Jr. J. Am. Chem. Soc. 2006, 128,
1038-1039.
(34) (a) Denmark, S. E.; Fan, Y. J. Am. Chem. Soc. 2003, 125, 7825-
7827. (b) Denmark, S. E.; Fan, Y. J. Org. Chem. 2005, 70, 9667-9676.
(c) Denmark, S. E.; Bui, T. J. Org. Chem. 2005, 70, 10190-10193.
(35) Mayr, H.; Kempf, B.; Ofial, A. R. Acc. Chem. Res. 2003, 36, 66-
77.
(36) (a) Larson, J. W.; McMahon, T. B. J. Am. Chem. Soc. 1985, 107,
766-773. (b) Kraus, K. A.; Moore, G. E. J. Am. Chem. Soc. 1953, 75,
1460-1462. (c) Strelow, F. W. E.; Rethemeyer, R.; Bothma, C. J. C. Anal.
Chem. 1965, 37, 106-111. (d) Strelow, F. W. E.; Van Zyl, C. R.; Bothma,
C. J. C. Anal. Chim. Acta 1969, 45, 81-92.
The formation of R-chloro silyl ether has also been observed
in reactions with other trichlorosilyl species.14b,32 Neverthe-
less, even when the resting state of the aldehyde was shown to
(32) (a) Denmark, S. E., Ghosh, S. K. Angew. Chem., Int. Ed. 2001, 40,
4759-4762. (b) Denmark, S. E.; Bui, T. Proc. Nat. Acad. Sci. U.S.A. 2004,
101, 5439-5444.
J. Org. Chem, Vol. 71, No. 4, 2006 1531

Denmark et al.
20). Without HMPA, there was essentially no reaction with
HgCl2 alone. Compared to the reaction with 20 mol % of
HMPA, only a marginal improvement on the conversion was
observed when the loading of the HMPA was increased to 50
mol %. The most promising result was obtained when bispho-
sphoramide 25d was used. With 5 mol % of bisphosphoramide
25d and 10 mol % of HgCl2, the allylation of hydrocinnamal-
dehyde achieved 59% conversion after 1 h.
allyl alcohol adduct was isolated with up to 32% yield.
Unfortunately, the enantioselectivity decreased only 55.5/44.5
er. A similar effect of HgCl2 on the enantioselectivity was
observed in the allylation of benzaldehyde. At -25 °C, the
bisphosphoramide 25d catalyzed the allylation reaction and gave
the homoallylic alcohol in 86% yield and 85.0/15.0 er. In
contrast, nearly racemic product was isolated when HgCl2 was
added.
TABLE 6. Addition of Allyltrichlorosilane to
SCHEME 15
Hydrocinnamaldehyde Catalyzed by HMPA and Inorganic Saltsa
entry
HMPA, equiv
additive, equiv
conversion, %b
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
1.3
6
28
26
9
13
5
9
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
0.2
SbCl3, (0.1)
HgCl2, (0.1)
ZrCl4, (0.1)
CuCl2, (0.1)
PdCl2(0.1)
As an alternative approach to solve the lack of reactivity of
aliphatic aldehydes, it was reasoned that the formation of the
R-chloro silyl ether might be disfavored if the chloride anion
could be exchanged with a less nucleophilic anion such as a
triflate anion. Thus, the equilibrium would favor the complexed
aldehyde, which would allow addition of the allyl group. This
exchange should be promoted by phosphoramides since the
ionization of chloride took place only upon coordination of
phosphoramides to allylic trichlorosilane.
GeCl4, (0.1)
NiCl2, (0.1)
PtCl2, (0.1)
BiCl3, (0.1)
InCl3, (0.1)
PbCl2, (0.1)
HgCl2, (0.1)
HgCl2, (0.02)
HgCl2, (0.5)
HgCl2, (1.0)
HgCl2, (0.1)
HgCl2, (0.1)
HgCl2, (0.1)
9
6
19
13
6
26
8.7
31
30
1.3
31
59
Thus, to a solution of allyltrichlorosilane with a 20 mol %
of HMPA was added TMSOTf as the triflate anion source. Upon
addition, TMSOTf disappeared rapidly with the formation of
TMSCl. The reaction, however proceeded to 20% conver-
sion as judged by 1H NMR analysis. To achieve complete
reaction required the use of 1.0 equivalent of HMPA. The
stoichiometries of the reagents utilized in the reaction suggested
the formation of ionic HMPA/allyldichlorotriflate complex 49
(Scheme 16).
To test the reactivity of complex 49 toward aliphatic
aldehydes, cyclohexanecarboxaldehyde was added to the reac-
tion mixture. After 3 h at 0 °C, the allylation adduct 50 was
obtained only in 8% yield (Scheme 16). The allylation rate could
be significantly accelerated by the addition of another equivalent
of HMPA. With 2 equiv of HMPA, the allylation product 50
was obtained in 56% yield after 3 h at 0 °C.
0.5
25d (0.05)
a Reaction done at 0.20 M concentration at rt for 2 h. b Conversion
determined by GC analysis.
The catalyst system was further applied to the addition of
2-butenyltrichlorosilanes 1b to examine the effect of HgCl2 on
the diastereoselectivity. With 50 mol % of HMPA and 10 mol
% of HgCl2, the silane (Z)-1b successfully added to the
hydrocinnamaldehyde and provided the syn homoallyl alcohol
48 in 44% yield and high selectivity (syn/anti, 98/2) (Scheme
14). The high stereospecificity strongly suggested a closed,
chairlike transition structure.
SCHEME 14
SCHEME 16
Because of the enhanced reaction rate observed with the
addition of HgCl2, the effect of HgCl2 on the enantioselec-
tivity was then investigated. At -78 °C, the bisphosphoramide
25d did not promote the addition of allyltrichlorosilane to
hydrocinnamaldehyde. Increasing the reaction temperature to
-25 °C allowed the addition to take place, and the homoallylic
alcohol adduct could be isolated in 25% yield and er of
66.0/34.0 (Scheme 15). Upon addition of 10 mol % of HgCl2
salt, the reaction rate was significantly enhanced. The homo-
These studies demonstrated that the allylation of aliphatic
aldehydes could be achieved by intercepting the chloride anion.
Particularly, with HgCl2 as an additive, a catalytic amount of
phosphoramide could afford the allylation adduct with high
1532 J. Org. Chem., Vol. 71, No. 4, 2006

Chiral Phosphoramide-Catalyzed EnantioselectiVe Allylation
conversion. Unfortunately, in this case, the enantioselectivity
observed was adversely affected.
Indeed, loading studies with bisphosphoramides showed that
when the loading of 5c was decreased from 1.0 to 0.1 equiv,
the er of product obtained also decreased from 58.5/41.5 to
55/45. On the other hand, a lower of the loading of 5d brought
about an increase in the ee of the product from 82.5/17.5 to
86/14. Thus, if the reaction catalyzed by 5c primarily involved
two separate bisphosphoramides because of the low propensity
of 5c to chelate, then lowering the concentration of 5c would
allow intervention of the unselective nonchelated one-phos-
phoramide pathway (Figure 5c). On the other hand, if 5d reacts
primarily through a chelated, one-bisphosphoramide pathway
(Figure 5a) with some of the two-separate-bisphosphoramide
pathway in competition, then lowering the concentration of 5d
would lead to enhanced selectivity by favoring the more
selective chelated pathway. This is what was observed.
Discussion
1. Reaction Promoted by Bidentate Ligands. The poor
selectivities and reaction rates observed with monodentate
phosphorus-based ligands as catalysts in allylation reactions was
shown to arise from the intervention of two mechanistic
pathways involving either one or two ligands in the rate- and
stereochemistry-determining transition structure. The use of
bidentate ligands addressed this problem by increasing the
effective concentration of the second basic group through
tethering. An important consequence of tethering is that the
arrangement of the Lewis basic subunits in the complex is
dictated by the tether length and flexibility. Not surprisingly,
the selectivities were found to be quite different for two different
phosphoramide subunits and one phosphonamide (Figure 4). The
graphical presentation of the data clearly shows the dramatic
chain length dependence of enantioselectivity, which strongly
supports the hypothesis that the two subunits are cooperative.
Both of the phosphoramide series 5 and 25 show significant
increase in selectivity with the pentamethylene tether. Even the
bisphosphonamide series 11 showed a dependence of the product
er on the tether length albeit with a different pattern, perhaps
reflecting the all-carbon chain that connects the two phosphon-
amide units. Nevertheless, the maximum was seen with a hepta-
methylene tether that corresponds to the same 12-membered
chelate ring for complexation of the siliconium cation.
FIGURE 5. (a) Chelated one-phosphoramide pathway. (b) Nonchelated
two-phosphoramide pathway. (c) Nonchelated one-phosphoramide
pathway.
Structural support for the formation of these various com-
plexes has been provided by solution NMR spectroscopic and
solid X-ray crystallographic studies of the bisphosphoramide‚
SnCl4 complexes.16 From 119Sn and 31P NMR solution studies,
it was found that the catalysts 5c, 25c, 5d, and 25d bearing
four or five methylene units formed exclusively the cis-chelate
complex with SnCl4 at high concentration.
In addition, the NMR spectroscopic studies revealed that the
formation of chelated complexes was not as favorable with
bisphosphoramides 5a, 5b, 5e, and 25e containing other tether
lengths. Instead these potentially chelating bisphosphoramides
functioned as monodentate ligands. In the allylation reaction
the bisphosphoramides, 5b, 5e, and 25e provided rather similar
enantioselectivity compared to the monophosphoramide 6;
therefore, it was most likely that these ligands behave as
monodentate ligands in the reaction as well.
FIGURE 4. Tether length dependence of enantioselectivity for various
dimeric catalysts.
In the case of bisphosphoramide 5a, upon coordination of
one phosphoramide unit, the appended bulky phosphoramide
is in close proximity to the coordinating phosphoramide unit.
Thus, the bisphosphoramide could be viewed as a bulky
monodentate phosphoramide. In the allylation reaction promoted
by 5a, the product was obtained in racemic form. Since the
sterically bulky phosphoramides have been found give poor
enantioselectivity, the reason for the low selectivity was pre-
sumably due to the steric bulk of this monodentate phosphor-
amide.
The dependence of er on the promoter concentration provided
additional support for the operation of chelation. In the allylation
promoted by bisphosphoramides, there are two possible compet-
ing pathways involving two phosphoramide units in the transi-
tion structure: (a) one bisphosphoramide chelated to silicon and
(b) two separate bisphosphoramides (nonchelated) bound in the
transition structure (Figure 5). The rate of the pathway (b) is
second order in phosphoramide concentration, while the rate
of pathway (a) is first order in phosphoramide concentration.
Therefore, the chelated pathways would be more favorable at
lower phosphoramide concentration. Because the enantioselec-
tivities of the chelated and nonchelated pathways were different,
the net reaction selectivity would be concentration dependent.
Clearly, tether length and the resulting chelation are not the
only factors that influenced the stereoselectivity of the allylation.
Obviously, the structure of the phosphorus-containing units that
create the chiral environment are crucial for high facial
J. Org. Chem, Vol. 71, No. 4, 2006 1533

Denmark et al.
X-ray crystal structure with our recently reported computational
analysis of the preferences for coordination geometries in
hypervalent silicon species.15 These calculations revealed a
strong preference for the trans arrangement of chlorines in one
of the two 3-center-4-electron bonds and the placement of the
coordinated aldehyde in the sp hybrid orbital. This geometry
then places the allyl group in the second hypervalent bond,
which is also where the electronic activation should be the
strongest. Two limiting arrangements of the aldehyde and allyl
groups in chairlike conformations (as must be the case because
of the high diastereoselectivity) are shown in Figure 7.37 In one
of the possible chairlike assemblies (Figure 7b), leading to the
Re-face attack, the benzaldehyde ring is located in a quadrant
of the ligand environment that is occupied by a forward-pointing
pyrrolidine ring, thereby creating unfavorable steric interactions.
In the diastereomorphic chairlike arrangement (Figure 7a), the
benzaldehyde ring is placed in a quadrant that is occupied by a
rearward-pointing pyrrolidine ring, and no unfavorable steric
interactions are evident. On the basis of this analysis, we propose
that the reaction proceeds through a transition structure similar
to the arrangement shown in Figure 7a, leading to the observed
homoallyl alcohol of S-configuration at the hydroxyl center.
FIGURE 6. Chem 3D presentation of the X-ray structure of 25d‚
SnCl4.
differentiation. Compare, for example, the phosphoramides 5d,
14, 15, 20, 25d, and 27, all of which can achieve 12-membered
ring chelation. The range of enantioselectivities is striking from
60/40 up to 93.5/6.5. Perhaps most illustrative of the sensitivity
of these reactions to structural details is the drop in selectivity
seen for the bispiperidine catalyst 27. Given the generality and
high selectivity observed with catalyst 25d, a rationalization of
preferred transition-structure arrangements is warranted.
2. Transition Structure for the Allylation Catalyzed by
25d. Among the bisphosphoramide examined, the bisphosphor-
amide 25d derived from 2,2′-bispyrrolidine provided the highest
enantioselectivity. To provide an understanding of the special
features of 25d, the complexation of 25d with SnCl4 as a Lewis
acid has been studied, and the solid-state structure of 25d‚SnCl4
complex is shown in Figure 6.16 In this complex, the tin atom
adopts an octahedral geometry with bisphosphoramide chelated
in a cis manner. Most importantly, the two pyrrolidine rings
adopt a stairlike geometry, creating a highly asymmetric
environment.
The bisphosphoramide 25d possesses two features that
contribute to the high enantioselectivities: (1) a 2,2′-bispyrro-
lidine as chiral diamine backbone and (2) five methylene units
as the tether. Interesting, the bisphosphoramide 27d derived from
2,2′-bispiperidine was much less selective than 25d. It is
conceivable that the longer linkage between the nitrogen and
the stereogenic center (four ring methylenes vs three ring
methylenes) and the attendant increased flexibility of the in the
piperidine attenuates the influence of the backbone on the
geometry of the piperidine ring, which leads to a decreased
enantioselectivity. In addition, in a detailed solution- and solid-
state study on the bisphosphoramide 25c‚SnCl4 and 25d‚SnCl4,
A transition structure for the allylation reaction catalyzed by
25d can be formulated by combining the coordinates from the
FIGURE 7. Hypothetical transition structure for the reaction promoted by 25d.
1534 J. Org. Chem., Vol. 71, No. 4, 2006

Chiral Phosphoramide-Catalyzed EnantioselectiVe Allylation
it was found that a five methylene unit tether was not only
important for chelation but also necessary to bring the chiral
information close to the reaction center, thereby achieving high
asymmetric induction.16
This technology represents the state of the art for catalytic,
enantioselective allylation of aromatic and olefinic aldehydes.
Extremely high enantio- and diastereoselectivity can be obtained
from simple allylic reagents and from geometrically defined
donors, even for the construction of quaternary stereogenic
centers. The singular (albeit significant) shortcoming of this
method is the inability to engage aliphatic aldehydes. Because
this is a mechanism-based problem associated with the necessary
ionization of chloride from the chlorosilyl reagents, future
studies will focus on allylating agents that are capable of
nucleophilic activation without the need for ionization of a
ligand. These efforts along with applications of the chlorosilyl
species in synthesis endeavors will be reported in due course.
3. Reaction with γ-Substituted Allylic Trichlorosilanes. In
the addition of allylic trichlorosilanes to aromatic aldehydes,
the electron-rich aromatic aldehyde provided higher enantiose-
lectivity than the electron-poor aromatic aldehydes. One might
expect that an electron-rich aldehyde would coordinate more
tightly to the silicon center, thus contracting the transition
structure and intensifying steric interaction, leading to the higher
selectivity. Alternatively, it is possible that the less reactive
electron-rich aldehydes react via late transition states that can
also enhance interactions of substrate and chiral catalyst.38
Experimental Section
A more important demonstration of the chiral Lewis base
catalyzed allylation is that uniformly high diastereoselectivities
and enantioselectivities are observed in the addition of (E)- and
(Z)-2-butenyltrichlorosilanes to aldehydes. The high stereospeci-
ficity clearly suggested a closed, chairlike transition structure.
Although asymmetric allylation has seen wide application in
organic synthesis, few applications of this method in the con-
struction of quaternary centers have been reported. The chal-
lenges of performing such a reaction are as follows: (1) selective
synthesis of geometrically pure 3,3-disubstitued allylic orga-
nometallic reagents; (2) correlation of the geometrical composi-
tion of the allylic organometallic reagents to the diastereomeric
composition of the products; (3) control of asymmetric induction
with internal chiral auxiliaries or external chiral catalysts. The
allylation method described here has been proven to satisfy these
requirements. This method provided a very versatile route for
the construction of quaternary centers from the corresponding
γ-disubstituted allylic alcohol. The application of this method
to the synthesis of serotonin antagonists 46 demonstrated not
only the efficiency of this allylation method but also the versatile
functionality provided by the allylation adducts.
General Experimental Procedures. See the Supporting Infor-
mation.
Preparation of (3aR,4aR)-7-Chlorooctahydro-6a,7a-diaza-7-
phosphacyclopenta[a]pentalene 7-Oxide (28). In a 100-mL, two-
neck, round-bottom flask with a nitrogen adapter and septum were
added a solution of (R,R)-21 (2.0 g, 14.3 mmol) in 80 mL of diethyl
ether at 0 °C (ice bath) and triethylamine (4.0 mL, 28.6 mmol, 2.0
equiv) followed by the slow addition of phosphorus oxychloride
(1.33 mL, 14.3 mmol, 1.0 equiv) by syringe. The mixture was stirred
at room temperature for 2 h, whereupon the salt was filtered though
a glass filter frit and the filtrate was concentrated to give 2.1 g
(67%) of 28 as a light yellow oil. This crude material was used
directly in the next step. An analytical sample was obtained by a
chromatographic purification (silica gel, CH2Cl2/acetone, 19/1). Data
for (+)-28: 1H NMR (500 MHz, CDCl3) 3.69-3.57 (m, 3 H,
HC(3a), HC(4a), HâC(6)), 3.44-3.36 (m, 1 H, HRC(6)), 3.06-
2.98 (m, 2 H, H2C(1)), 2.06-1.89 (m, 6 H, H2C(2), H2C(5), H2C-
(3 or 4), 1.75-1.51 (m, 2 H, H2C(3 or 4)); 13C NMR (500 MHz,
CDCl3) 68.7 (d, J ) 14.7, C(3a or 4a)), 66.1 (d, J ) 17.5, C(3a or
4a)), 45.4 (d, J ) 1.9, C(1 or 6)), 45.2 (d, J ) 3.7, C(1 or 6)), 30.7
(d, J ) 6.4, C(3 or 4)), 30.3 (d, J ) 2.7, C(3 or 4)), 27.5 (d, J )
4.6, (C(2 or 5)), 27.1 (d, J ) 5.5, C(2 or 5)); 31P NMR (202 MHz,
CDCl3): 28.35; IR (NaCl): 2969 (m), 2879 (m), 1446 (w), 1278
(s), 1197 (m), 1110 (s), 1066 (m), 1033 (s), 993 (w), 910 (w); MS
(FAB) 224 (3), 223 (29), 222 (10), 221 (100), 141 (33), 132 (16),
Conclusions
118 (23); [R]24 +37.25 (c ) 1.21, benzene); TLC Rf 0.34 (CH2-
A new family of phosphorus-based bidentate ligands has been
developed for catalytic enantioselective allylation reaction. From
a combination of mechanistic insights and X-ray crystallographic
analysis, it was found that linking two phosphonamide units
through a pentamethylene tether leads to optimal cooperativity
in the activation of allylic trichlorosilyl reagents. The bidentate
phosphoramide derived from 2,2-bispyrrolidine provided the
highest yield and enantioselectivity in the addition of various
allylic trichlorosilyl species with aromatic and unsaturated
aldehydes. Aliphatic aldehydes are prone to formation of
unreactive R-chlorosilyl ethers and do not undergo addition in
useful yields. The high stereochemical fidelity between allyl
geometry and product configuration supports the hypothesis of
a highly organized chairlike transition structure and also allowed
for the development of catalytic enantioselective construction
of quaternary stereogenic centers. This application was il-
lustrated in the synthesis of serotonin agonist LY426965.
D
Cl2/acetone, 19/1) [KMnO4]; HRMS calcd for C8H15ClN2OP (M+
+ H) 221.0610, found 221.0611.
Preparation of N,N′-Dimethyl-N,N′-bis((3′aR,4′aR)-7′-oxooc-
tahydro-6′a,7′a-diaza-7′-phosphacyclopenta[a]pentalene-7′-yl)-
pentane-1,5-diamine (25d). In a 50-mL, three-necked, round-
bottom flask fitted with a nitrogen inlet adapter, septum, and
thermocouple was placed a solution of N,N′-dimethyl-1,5-pen-
tanediamine (476 µL, 3.0 mmol, 1.0 equiv) in 28.0 mL of THF,
and the solution was cooled to -78 °C in a dry ice/i-PrOH bath.
To this solution was added dropwise n-BuLi (4.0 mL, 1.55 M, 6.1
mmol, 2.0 equiv). The reaction mixture was warmed to 0 °C and
was stirred at 0 °C in an ice bath for 30 min and then was cooled
to -78 °C. To this solution was added a solution of 28 (1.50 g, 6.8
mmol, 2.3 equiv) in 17.0 mL of THF. The mixture was stirred at
0 °C in an ice bath for 1 h and then at room temperature for 2 h.
After evaporation of the solvent, the product was purified by silica
gel column chromatography (CH2Cl2/i-PrOH, 9/1 followed by
CH2Cl2/i-PrOH, 4/1) to provide 1.41 g (93%) of 25d as a clear,
1
colorless oil. The product was pure by H NMR (S/N > 100/1)
and 31P NMR (S/N > 60/1) analysis.
(37) This analysis is different from that presented in our previous account
(ref 16) wherein the aldehyde was suggested to be coordinated trans to a
chloride in an axial position. Our calculations strongly suggest that this is
not a bound species and would also be deactivated for addition to the
aldehyde group.
(38) For an in-depth mechanistic discussion of the addition of enox-
ytrichlorosilyl species to aldehydes, see: Denmark, S. E.; Bui, T. J. Org.
Chem. 2005, 70, 10393-10399.
Data for (+)-25d: 1H NMR (500 MHz, CDCl3) 3.53-3.46 (m,
6 H, 2 × HC(3′a), 2 × HC(4′a), 2 × HâC(6′)), 3.05-2.92 (m, 8
H, 2 × H2C(1′), H2C(1), H2C(5)), 2.75-2.70 (m, 2 H, 2 × HRC-
(6′)), 2.57 (d, J ) 10.3, 6 H, H3C(Me)), 1.95-1.81 (m, 12 H, 2 ×
H2C(2′), 2 × H2C(5′), 2 × H2C(3′ or 4′)), 1.62-1.47 (m, 8 H,
H2C(2), H2C(4), 2 × H2C(3′ or 4′)), 1.29-1.24 (m, 2 H, H2C(3));
J. Org. Chem, Vol. 71, No. 4, 2006 1535

Denmark et al.
13C NMR (126 MHz, CDCl3) 68.6 (d, J ) 10.1, C(3′a or 4′a), 66.6
(d, J ) 9.4, C(3′a or 4′a), 49.1 (d, J ) 3.6, C(1), C(5)), 45.4 (d, J
) 2.8, C(1′ or 6′)), 43.5 (d, J ) 3.6, C(1′ or 6′)), 33.0 (d, J ) 3.7,
C(Me)), 30.9 (d, J ) 2.8, C(3′ or 4′)), 30.5 (d, J ) 5.5, C(3′ or
4′)), 28.2 (d, J ) 2.7, C(2′ or 5′)), 27.9 (d, J ) 6.5, C(2′ or 5′)),
27.4 (d, J ) 3.7, C(2), C(4)), 24.0 (C(3)); 31P NMR (202
MHz, CDCl3) 25.37; IR (NaCl) 3039 (w), 2958 (s), 2867 (s), 1714
(w), 1454 (m), 1328 (m), 1226 (s), 1207 (s), 1135 (s), 1091 (s),
1070 (s), 1027 (s), 985 (s), 730 (s); MS (FAB) 224 (3), 223 (29),
CH2Cl2/pentane, 3/1 followed by CH2Cl2) to give 254 mg (85%)
of 3aa as a colorless oil. Data for (S)-3aa: 1H NMR (500 MHz,
CDCl3) 7.37-7.26 (m, 5 H, HC(Aryl)), 5.80 (dddd, J ) 17.5, 10.5,
7.7, 6.6, 1 H, HC(3)), 5.19-5.14 (m, 2 H, H2C(4)), 4.75 (dd, J )
7.7, 5.0, 1 H, HC(1)), 2.57-2.46 (m, 1 H, HC(2)), 2.02 (br, 1 H,
OH); 13C NMR (126 MHz, CDCl3) 143.8 (C(1′)), 134.4 (C(3)),
128.4 (C(3′)), 127.6 (C(4′)), 125.8 (C(2′)), 118.5 (C(4)), 73.2 (C(1)),
43.8 (C(2)); IR (NaCl) 3368 (s), 3076 (w), 3030 (w), 2906 (w),
1641 (m), 1493 (w), 1453 (m), 1315 (w), 1198 (w), 1047 (s), 916
222 (10), 221 (100), 141 (33), 132 (16), 118 (23); [R]24 +12.26
(s), 757 (s), 700 (s); [R]24D -48.12 (c ) 1.05, benzene) (lit.39 [R]24
D
D
(c ) 1.11, EtOH); TLC Rf 0.45 (CH2Cl2/EtOH, 9/1) [KMnO4];
HRMS calcd for C23H45N6O2P2 (M+ + H) 499.3079; found
499.3081.
-17.8 (c ) 7.38, benzene) for 30% ee of (S)-3aa); TLC Rf +0.39
(CH2Cl2) [KMnO4]; SFC (R)-3aa, tR 2.97 min (6.5%); (S)-3aa, tR,
3.46 min (93.4%) (Chiralpak OD, 40 °C, 150 bar, 4% MeOH in
CO2, 3.0 mL/min, 258 nm).
Addition of Allylic Trichlorosilanes to Aldehydes. Preparation
of (S)-1-Phenyl-3-buten-1-ol (3aa). In a 10-mL, three-necked,
round-bottom flask fitted with a nitrogen inlet adapter, septum, and
thermocouple was placed a solution of 25d (50 mg, 0.1 mmol, 0.05
equiv) in 1.0 mL of dichloromethane and 1.0 mL of diisopropyl-
ethylamine under N2 at -78 °C, and then allylic trichlorosilane 1a
(580 µL, 4.0 mmol, 2.0 equiv) was added. The solution was stirred
at -78 °C for 10 min before benzaldehyde (200 µL, 2.0 mmol)
was added. The resulting mixture was stirred at that temperature
for 8 h, whereupon the cold solution was poured into a mixture of
10 mL of saturated aqueous NaHCO3 and 10 mL of saturated
aqueous KF solution at 0 °C with vigorous stirring. The mixture
was stirred for 2 h at room temperature, and then it was filtered
through Celite. The layers were then separated, and the aqueous
layer was extracted with CH2Cl2 (3 × 30 mL). The combined
organic extracts were dried (Mg2SO4), filtered, and concentrated.
The oily residue was purified by column chromatography ((SiO2),
Acknowledgment. We are grateful to the National Science
Foundation for support of this research. J.F. thanks the
Boehringer Ingelheim Pharmaceutical Co. for a graduate fel-
lowship. Justin Montgomery is thanked for his help with the
transition-structure modeling.
Supporting Information Available: Preparation and full
characterization ofphosphoramide catalysts, all addition products,
a representative allylation procedure, and crystallographic data for
36. This material is available free of charge via the Internet at
http://pubs.acs.org.
JO052203H
(39) Hoffmann, R. W.; Herold, T. Chem. Ber. 1981, 114, 375-383.
1536 J. Org. Chem., Vol. 71, No. 4, 2006