Chiral Bis-π-allylpalladium Complex Catalyzed Asymmetric Allylation of Imines: Enhancement of the Enantioselectivity and Chemical Yield in the Presence of Water

Article abstract of DOI:10.1021/ja037272x

The chiral π-allylpalladium complex 2a, prepared from exoethylidenenorpinane 7, catalyzed the allylation of diverse imines with allyltributylstannane in the presence of 1 equiv of water in good to high enantioselectivities. The catalyst prepared from a 1:1 mixture of (E)- and (Z) -7 was found to be consisting of two stereoisomers 2a and 2b in 1.3:1 ratio. On separation, 2a catalyzed the allylation of imines in much higher enantioselectivities than 2b, giving the same major enantiomer and thereby justifying the need to separate 2a free of 2b. We have achieved the highest separation ratio of >400:1 for 2a:2b by repeated recrystallizations. Isomerization of 2b to 2a during recovery of 2a from the filtrates was observed as more of 2a was recovered each time during recrystallization. Although dry THF was the best solvent, we tried various additives and found that addition of one equivalent of water gave the best results with respect to shorter reaction time, higher yields and enantioselectivities. Thus, we have developed a more general, reproducible, robust and a non-Lewis acid catalyzed procedure for catalytic asymmetric allylation of imines under essentially neutral conditions.

Full text of DOI:10.1021/ja037272x

Published on Web 10/22/2003
Chiral Bis-π-allylpalladium Complex Catalyzed Asymmetric
Allylation of Imines: Enhancement of the Enantioselectivity
and Chemical Yield in the Presence of Water
Rodney A. Fernandes, Anton Stimac, and Yoshinori Yamamoto*
Contribution from the Department of Chemistry, Graduate School of Science, Tohoku UniVersity,
Sendai 980-8578, Japan
Received July 15, 2003; E-mail: yoshi@yamamoto1.chem.tohoku.ac.jp
Abstract: The chiral π-allylpalladium complex 2a, prepared from exoethylidenenorpinane 7, catalyzed the
allylation of diverse imines with allyltributylstannane in the presence of 1 equiv of water in good to high
enantioselectivities. The catalyst prepared from a 1:1 mixture of (E)- and (Z)-7 was found to be consisting
of two stereoisomers 2a and 2b in 1.3:1 ratio. On separation, 2a catalyzed the allylation of imines in much
higher enantioselectivities than 2b, giving the same major enantiomer and thereby justifying the need to
separate 2a free of 2b. We have achieved the highest separation ratio of >400:1 for 2a:2b by repeated
recrystallizations. Isomerization of 2b to 2a during recovery of 2a from the filtrates was observed as more
of 2a was recovered each time during recrystallization. Although dry THF was the best solvent, we tried
various additives and found that addition of one equivalent of water gave the best results with respect to
shorter reaction time, higher yields and enantioselectivities. Thus, we have developed a more general,
reproducible, robust and a non-Lewis acid catalyzed procedure for catalytic asymmetric allylation of imines
under essentially neutral conditions.
reactions with aldehydes have been well investigated,4 very few
reports have appeared on the asymmetric allylation of imines.5
Introduction
The allylation of carbonyl compounds such as aldehydes,
ketones, and their derivatives is one of the most important C-C
bond forming reactions because of the versatility of homoallylic
alcohols and amines as useful synthetic intermediates.1 Although
many methods have been developed, research in this field
seeking new and more efficient methods are unabated. Among
the various allylmetal reagents, allylstannanes and allylsilanes
are very useful because of their modest reactivity, which in turn
can be increased by catalyst activation and thus allow for
application to catalytic enantioselective reactions. Imine activa-
tion methods using Lewis or Bronstead acids have also been
developed.1 However, these acids coordinate strongly with
amine products, making the catalyst inactive. It is also reported
that fluoride anions or alkoxides also activate the C-Si bond
of allylsilanes by coordinating to the silicon atom.2 Recently,
chiral sulfoxides as neutral coordinate organocatalyst (NCOs)
were used in the allylation of N-acylhydrazones.3 Although the
Our group6 reported the first catalytic asymmetric allylation of
imines in 1998 using allyltributylstannane in the presence of a
chiral π-allylpalladium complex.
In contrast to the Lewis acid mediated allylations, known from
previous studies that require at least a stoichiometric amount
of promoter, these reactions were shown to occur readily in the
presence of catalytic amounts of Pd (II) or Pt (II) complexes.
Mechanistic studies of the Pd (II)-catalyzed reaction revealed
that bis-π-allylpalladium complex (e.g., 3) is a key intermediate7
for the catalytic cycle and reacts with imines as a nucleophile,
although the ordinary π-allylpalladium complexes such as
π-allylPdX (X ) OAc and halides, e.g., 1) act as electrophiles8
(Scheme 1). We have also shown that bis-π-allylpalladium
complex has an amphiphilic character.9 It was our earlier
(3) Kobayashi, S.; Ogawa, C.; Konishi, H.; Sugiura, M. J. Am. Chem. Soc.
2003, 125, 6610.
(4) (a) Oi, S.; Moro, M.; Fukuhara, H.; Kawanishi, T.; Inoue, Y. Tetrahedron
2003, 59, 4351. (b) Yanagisawa, A.; Kageyama, H.; Nakatsuka, Y.;
Asakawa, K.; Matsumoto, Y.; Yamamoto, H. Angew. Chem. Int. Ed. 1999,
38, 3701 and references therein. Also see ref. (2a).
(5) For leading references on enantioselective allylation of imines, see: (a)
Gastner, T.; Ishitani, H.; Akiyama, R.; Kobayashi, S. Angew. Chem. Int.
Ed. 2001, 40, 1896. (b) Fang, X.; Johannsen, M.; Yao, S.; Gathergood, N.;
Hazell, R. G.; Jorgensen, K. A. J. Org. Chem. 1999, 64, 4844. (c) Ferraris,
D.; Dudding, T.; Young, B.; Drury, W. J., III.; Lectka, T. J. Org. Chem.
1999, 64, 2168.
(6) (a) Bao, M.; Nakamura, H.; Yamamoto, Y. Tetrahedron Lett. 2000, 41,
131. (b) Nakamura, H.; Nakamura, K.; Yamamoto, Y. J. Am. Chem. Soc.
1998, 120, 4242. For use of allyltrimethylsilane instead of allyltributyl-
stannane, see reference (2b).
(7) Nakamura, H.; Iwama, H.; Yamamoto, Y. J. Am. Chem. Soc. 1996, 118,
6641. (b) Nakamura, H.; Asao, N.; Yamamoto, Y. J. Chem. Soc., Chem.
Commun. 1995, 1273.
(1) For recent reviews on allylmetal additions, see: (a) Puentes, C. O.;
Kouznetsov, V. J. Heterocycl. Chem. 2002, 39, 595. (b) Denmark, S. E.;
Almstead, N. G. In Modern Carbonyl Chemistry; Otera, J., Ed.; Wiley-
VHC: Weinheim, 2000; Chapter 10. (c) Chemler, S. R.; Roush, R. W. In
Modern Carbonyl Chemistry; Otera, J., Ed.; Wiley-VHC: Weinheim, 2000;
Chapter 11. (d) Kobayashi, S.; Ishitani, H. Chem. ReV. 1999, 99, 1069. (e)
Bloch, R. Chem. ReV. 1998, 98, 1407. (f) Enders, D.; Reinhold, U.
Tetrahedron: Asymmetry 1997, 8, 1895. (f) Yamamoto, Y.; Asano, N.
Chem. ReV. 1993, 93, 2207.
(2) For leading references on fluoride- or alkoxide-promoted allylation of imines
with allylsilanes, see: (a) Yamasaki, S.; Fujii, K.; Wada, R.; Kanai, M.;
Shibasaki, M. J. Am. Chem. Soc. 2002, 124, 6536. (b) Nakamura, K.;
Nakamura, H.; Yamamoto, Y. J. Org. Chem. 1999, 64, 2614. (c) Wang,
D.-K.; Zhou, Y.-G.; Tang, Y.; Hou, X.-L.; Dai, L.-X. J. Org. Chem. 1999,
64, 4233. (d) Pilcher, A. S.; DeShong, P. J. Org. Chem. 1996, 61, 6901.
(e) Sakurai, H. Synlett 1989, 1.
9
10.1021/ja037272x CCC: $25.00 © 2003 American Chemical Society
J. AM. CHEM. SOC. 2003, 125, 14133-14139
14133

A R T I C L E S
Fernandes et al.
Scheme 2. Catalyst Preparation for Asymmetric Allylation (for
Scheme 1. Concept of Chiral Allylation with Bis-π-Allylpalladium
Complexes
separation of 2a and 2b See Supporting Information).
tion using 1H NMR, we found that the π-allylpalladium
complex, obtained from a 1:1 mixture of (E)- and (Z)-7,11
consisted of a 1.3:1 mixture of two stereoisomers 2a and 2b
(Scheme 2). We first tried to separate 2a and 2b by column
chromatography. However, although we could get a small
amount of 2b separated in pure form, 2a always contained 2b.
We opted then for repeated recrystallization. On two recrystal-
lizations in CH2Cl2/hexane, we got 2a:2b ) 12.3:1 (see the
Supporting Information). Although these results could not be
exactly reproduced, we later concluded through several trials
that any material having ratio of 2a:2b > 7 could best be
upgraded by propionitrile recrystallization.12 Thus, two succes-
sive propionitrile recrystallizations gave 2a:2b ) 100:1, having
investigation that, with the proper choice of the two allyl ligands
of bis-π-allylpalladium complexes, catalytic asymmetric ally-
lation was also possible. This is due to the fact that a sterically
bulky π-allyl group acts as a nontransferable π-allyl ligand,
whereas the other reacts with imines as a nucleophile in a chiral
way.6b Among the various π-allylpalladium chloride complexes
prepared from (1R)-(+)-camphor, (1S)-â-(-)-pinene, and (1S)-
(+)-3-carene, the one based on (1S)-â-(-)-pinene gave around
50% enantiomeric excess, whereas the others were unsuccessful.6b
The exomethylene of (1S)-â-(-)-pinene was converted to
exoethylidene, and the corresponding π-allylpalladium chloride
2a obtained was reacted with allyltributylstannane to give the
bis-π-allylpalladium complex 4, which transferred the allyl
group to the imines 5 to furnish homoallylamines 6 in good
enantioselectivities (up to 82% ee).6b Later, when we applied
this catalytic asymmetric allylation to several imines under the
conditions reported in the communication,6b nonreproducible
results were obtained; often, lower enantioselectivities and
chemical yields were obtained, and the reason for this fluctuation
was unclear. This situation left sufficient room for improvement
of the reaction rate, yield, and enantioselectivity.
[R]22 ) -19.4 (c 0.4, CHCl3). Further recrystallization in
D
propionitrile gave the catalyst with 2a:2b ) >400:1,13 having
[R]22D ) -19.9 (c 0.4, CHCl3)14 which was almost constant on
further recrystallization. The filtrates were combined to recover15
the catalyst and more of 2a was always obtained, indicating
isomerization of 2b to 2a during recrystallization. Thus, with
the upgraded catalyst in hand we next planned to examine the
catalytic asymmetric allylation reaction.
Catalytic Asymmetric Allylation: Unprecedented Influ-
ence of Water. We first examined the allylation reaction of 8a
in the presence of 2a (5 mol %) in THF solvent at 0 °C under
anhydrous conditions (eq 2). Although we carried out the
reaction several times under seemingly identical conditions,
nonreproducible results and lower enantioselectivities were
obtained. Also, under anhydrous conditions, catalyst decomposi-
tion was observed. Among many trials, the best results under
the standard conditions using dry THF (dried over Na or LiAlH4)
are shown in entries 1 and 2 of Table 1. We did not understand
why the data on ee and the chemical yield of eq 2 were
fluctuating. However, addition of 1 equiv of water resulted in
an improved yield and enantioselectivity (entry 3). More
importantly, in the presence of water, the data on ee and
Now we wish to report in detail that the asymmetric allylation
of imines with allyltributylstannane in the presence of a catalytic
amount of chiral π-allylpalladium complex 2a and 1 equiV of
water gives more general, reproducible, and robust results over
a wide range of imines (eq 1).
(11) Olefin 7 (E:Z ) 1:1) was prepared starting from (1S)-(-)-â-pinene available
in Aldrich Chemical Co. in 97% ee. Though (1R)-(+)-nopinone (90% ee)
is also available, we preferred to prepare it from (1S)-(-)-â-pinene by
ozonolysis of exocyclic double bond. Further Wittig reaction with ethyl-
triphenylphosphonium bromide and tBuOK gave 7 (E:Z ) 1:1) [see: Brown,
H. C.; Weissman, S. A.; Perumal, P. T.; Dhokte, U. P. J. Org. Chem. 1990,
55, 1217].
Results and Discussion
Preparation of Catalyst. Trost et al.10 reported that the
π-allylpalladium complex obtained from 7 was a single complex
and assigned the syn configuration (Scheme 2). They also stated
that the stereochemistry of the olefin did not determine the
stereochemistry of the product. However, on careful investiga-
(12) The solvents tried were 2-butanone, propionitrile, 2-propanol, and diiso-
propyl ether. Among all, propionitrile gave best upgradation ratio of 2a:
2b and recovery of catalyst during recrystalization. Though the catalyst is
not freely soluble in propionitrile, warming to 90 °C gave clear solution.
Higher temperatures or boiling should be avoided to prevent catalyst
decomposition.
(8) (a) Tsuji, J. In Palladium Reagents and Catalysis; John Wiley and Son:
Chichester, 1995; p 61. (b) Codleski, S. A. In ComprehensiVe Organic
Synthesis; Semmelhack, M. F., Ed.; Pergamon Press: Oxford, 1991; Vol.
4, p 585. (c) Collmann, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G.
In Principles and Applications of Organotransition Metal Chemistry; Mill
Valley, 1987; p 417.
(13) See the Supporting Information for 1H NMR of 2a:2b ) 1.3:1, 100:1, and
>400:1.
(14) There is no literature value of [R]D of 2a available. In our earlier
communication (ref 6b), the catalyst used had [R]22D -13.3 (c 0.4, CHCl3).
1
(9) Nakamura, H.; Shim, J.-G.; Yamamoto, Y. J. Am. Chem. Soc. 1997, 119,
8113.
From the H NMR spectrum available the ratio of the catalyst 2a:2b was
20:1. To the best of our knowledge, this is the first report of preparation of
the catalyst 2a and separated from its undesired stereoisomer with a ratio
of 2a:2b >400:1.
(10) (a) Trost, B. M.; Strege, P. E.; Weber, L.; Fullerton, T. J.; Dietsche, T. J.
J. Am. Chem. Soc. 1978, 100, 3407. Other references for synthesis of chiral-
π-allylpalladium complex complexes, see: (b) Rosset, J.-M.; Glenn, M.
P.; Cotton, J. D.; Willis, A. C.; Kennard, C. H. L.; Byriel, K. A.; Riches,
B. H.; Kitching, W. Organometallics 1998, 17, 1968. (c) Heck, R. F. In
Palladium Reagents in Organic Syntheses; Academic Press: London, 1985;
pp 1-18. (d) Trost, B. M.; Metzner, P. J. J. Am. Chem. Soc. 1980, 102,
3572.
(15) For recovery of catalyst 2a, the later filtrates (see the Supporting
Information) were combined and concentrated to yellow powder. This was
dissolved in propionitrile and crystallized out. The successive three
recrystallizations gave 2a:2b ) 100:1, having [R]22D -17.2 (c 0.4, CHCl3).
This was further upgraded to 2a (2a:2b >400:1), having [R]22 -19.8 (c
0.4, CHCl3). See the Supporting Information for more details.D
9
14134 J. AM. CHEM. SOC. VOL. 125, NO. 46, 2003

Chiral Bis-π-allylpalladium
A R T I C L E S
Table 1. Optimization of Reaction Conditionsa
sieves and MeOH. We do not understand at present the almost
comparable behavior of molecular sieves with the results
obtained with water though they are contradictory in activity.
TBAF, i-PrOH, and KOAc gave slightly lower enantioselec-
tivity, whereas the latter gave lower yield. Na2CO3 proved
unsatisfactory. The reaction required longer time for completion
and resulted in lower yield and enantioselectivity. AcOH gave
comparable results with water, but required longer reaction time
and gave lower yields. Thus, we found that addition of 1 equiv
of water furnished the best results.
The general procedure is as follows: Imine (0.5 mmol) was
taken in a Wheaton microreactor (5 mL capacity), and the
reactor was kept under Ar atmosphere. Dry THF (1.25 mL),
degassed water (9 µL, 0.5 mmol, 1 equiv) and allyltributyl-
stannane (194 µL, 0.625 mmol, 1.25 equiv) were added
sequentially. The mixture was cooled to 0 °C and the chiral
palladium chloride complex 2a (14.56 mg, 0.025 mmol, 5 mol
%) was added under argon. The reaction mixture was flushed
with argon and stirred at 0 °C for specified time. The reaction
progress was monitored with GC-MS or TLC. After comple-
tion, the turbid reaction mixture was quenched with 1N HCl
(2.5 mL). CH3CN16 (1 mL) was added and the reaction mixture
was stirred at room temperature for 10 min. The two layered
solution was extracted with hexane (2 × 3 mL). The hexane
layer was discarded, the aqueous layer was basified with 10%
NaOH aq (1.25 mL), and the resulting solution was stirred for
5 min. The solution was extracted with EtOAc (2 × 5 mL),
dried (Na2SO4) and concentrated. Purification by silica gel
column chromatography (hexane/EtOAc ) 5:1) gave the
corresponding homoallylamines. The enantiomeric excesses
were determined by HPLC.17
Under the above procedure, most imines reacted in shorter
time, with high yields and moderate to good enantioselectivities
(Table 3, entries 1-5). N-Benzylidenebenzylamine 12a reacted
in 48 h to give 13a in 85% ee. The imines 12b and 12e furnished
13b and 13e in 85% and 70% ee, respectively, though a longer
reaction time was needed for completion. The imine 12c with
para-methyl group gave 13c in high yield and 90% ee. We
anticipated that an ortho-chelating group such as methoxy might
help in rigidifying the transition state through chelation and
higher ee would result. The ortho-methoxy substituted imine
12d produced 13d in very good enantioselectivity of 88%, but
a marked contrast between 12d and 12b (para-methoxy
substituted) was not observed.
N-Cyclohexylidenebenzylamine 14 furnished the homoallyl-
amine 15 in 40% ee requiring 111 h for completion of reaction
by our earlier procedure;6b however, it gave 15 in 50% ee and
reacted in 74 h by our above new procedure (entry 6). Similarly,
the imine 16 produced 17 in high yield and good enantioselec-
tivity of 69% (entry 7). The heterocyclic imines 18a and 18b
reacted well to give 19a (67% ee) and 19b (53% ee),
respectively, in high yields and moderate ees (entries 8-9)
without polymerization, which is sometimes noticed under Lewis
acid-catalyzed conditions. Next, we examined the imines with
chelating methoxy substituent either on the aryl group of
aldehyde portion or on benzyl group on imine nitrogen. The
imines 10a-10c with para-methoxybenzyl group gave high
yields of homoallylamines 11a-11c with excellent enantio-
equiv of
entry allylSnBu3
%yield
temp °C time (h) of 9a %ee
solvent
1
2
1
1
dry THFb
dry THFc
0
0
120
120
120
84
96
135
96
192
120
168
51
64
73
76
72
64
68
50
70
42
84
82d
88
90
86d
86
83
83
63
17e
3
4
5
6
7
8
9
10
1
dry THF + 1 equiv H2O
dry THF + 1 equiv H2O
dry THF + 1 equiv H2O
dry THF + 5 equiv H2O
dry THF + 0.5 equiv H2O
dry THF + 1 equiv H2O
dry THF + 1 equiv H2O
dry THF + 1 equiv H2O
0
0
0
0
0
1.25
1.25
1.25
1.25
1.25
1.25
1.25
-20
rt
0
a Standard reaction conditions: To a solution of 8a (0.5 mmol) in dry
THF (1.25 mL, commercial dehydrated THF) was added allylSnBu3 (0.5
mmol) and the mixture was cooled to 0 °C. Catalyst 2a (5 mol %) was
added and the reaction mixture stirred for specified time (GC-MS
monitored). b Dried on Na. c Dried on LiAlH4. d Stannane added last. e Using
the catalyst 2b.
Table 2. Effect of an Additive on the Reaction Time, Chemical
Yield, and Enantioselectivity
%yield of
11a
additive
equiv
time (h)
%ee
no additive
H2O
TBAF
Na2CO3
KOAc
78
73
102
171
91
76
89
81
51
54
81
76
75
71
83
90
86
57
86
87
87
86
87
1
1
2
2
4 A° MS
MeOH
i-PrOH
AcOH
100 mg/mmol
72
89
90
147
1
1
1
chemical yield were not fluctuating, and a reproducible result
was obtained. The use of 1.25 equiv of stannane shortened the
reaction time and gave higher yield and enantioselectivity (entry
4). When the order of addition was changed, with catalyst
addition first and then stannane (entries 2 and 5), it required a
longer reaction time and gave slightly lower enantioselectivity.
Excess of water (5 equiv, entry 6) or less than 1 equiv (0.5
equiv, entry 7) gave inferior results in comparison to entry 4.
The reaction at -20 °C required longer reaction time for
completion and resulted in poor yields (entry 8). Similarly poor
results were obtained at room temperature (entry 9). When the
catalyst 2b was employed, the imine 8a was not completely
consumed even after 168 h and a poor yield and unsatisfactory
enantioselectivity were obtained (entry 10). Although catalyst
2b produces the same enantiomer, it is necessary to separate
2b from 2a.
Apart from water, we were interested in trying out other
additives. The reaction of 10a with allyltributylstannane in the
presence of 2a (5 mol %) and additives in dry THF at 0 °C
gave 11a (eq 3). It can be seen in Table 2, that addition of 1
equiv of water gave the best results concerning both chemical
yield and ee. Comparable results were obtained with molecular
(16) CH3CN was added for better solubilization of stannane byproducts.
(17) Details for each compound are giving in the Supporting Information.
9
J. AM. CHEM. SOC. VOL. 125, NO. 46, 2003 14135

A R T I C L E S
Fernandes et al.
Table 3. Catalytic Asymmetric Allylation of Imines with Allyltribuylstannane in the Presence of the Catalyst 2aa
9
14136 J. AM. CHEM. SOC. VOL. 125, NO. 46, 2003

Chiral Bis-π-allylpalladium
Table 3. (Continued)
A R T I C L E S
a All reactions were carried out with 1.25 equivalent of allyltributylstannane and one equivalent of water in the presence of 5 mol % of 2a and at 0 °C
for the specified time. In some cases additional 0.5 equivalent of allyltributylstannane was added if it was completely consumed and the reaction was not
complete (checked by GC-MS); for determination of absolute configuration, see ref 6b. b Determined by 1H NMR of the crude reaction mixture with
CH3NO2 as internal standard. c Using catalyst 2b. d ee based on known [R]D values (see the Supporting Information)
selectivities of 90%, 85%, and 86%, respectively (entries 10-
12). The imine 10d with nitro group in para-position of the
aldehyde portion gave 11d in excellent yield, but an unsatisfac-
tory enantioselectivity was obtained. The reason for this decrease
of enantioselectivity is not clear at present. The imine 10e with
ortho-methoxybenzyl group gave 11e with 89% ee (entry 14).
N-2-Naphthylidenebenzylamine 20a produced 21a in excellent
enantioselectivity (91% ee) under our new procedure (entry 15).
However, 20b having a para-methoxybenzyl group afforded 21b
with lower ee in comparison to 20a (entry 16). We have also
tried the reaction of imines bearing a group, other than benzyl,
on the imine nitrogen. Thus, the N-allyl imine 22a furnished
the allylated homoallylamine 23a in good yield and 78% ee
(entry 17). The reaction of the imine 22b was very sluggish to
give the product 23b in less than 10% yield even after 144 h,
though the imine was completely consumed in the reaction. The
probable reason for this sluggish reaction is explained in the
mechanistic section. N-3-Pyridylidenemethylamine 24 required
longer reaction time and gave 25 in moderate ee. The reaction
of the bulky imine 26 with three methoxy groups was slow to
furnish 27 in unsatisfactory 34% ee.
We were more interested in the amine 9a obtained by
allylation of the imine 8a. Complete reduction of the double
bond and removal of benzyl protection in 9a would give (R)-
R-propylpiperonylamine. This chiral butylamine is an important
building block of human leukocyte elastase inhibitor L-694,458.
9
J. AM. CHEM. SOC. VOL. 125, NO. 46, 2003 14137

A R T I C L E S
Fernandes et al.
Scheme 3. Plausible Mechanism of Allylation
Scheme 4. Transition State Models
A few syntheses of this amine have been reported earlier.18
Allylation of the imine 8a gave 9a in 76% yield with high
enantioselectivity of 90% (entry 21). This could be easily
converted into (R)-R-propylpiperonylamine by hydrogenation
of the double bond and benzyl deprotection. We examined the
allylation of the arylimines bearing other substituents on the
aryl portion or having different substituents on the benzyl group
with a hope to enhance the ee of the amine. Thus, 8b with para-
methoxybenzyl group and 8c with ortho-methoxybenzyl group
gave 9b and 9c, respectively, in comparable results with 89%
ee (entries 22-23). Introduction of either Cl- or Br- substituent
at the ortho-position of the aryl group could show no difference
(entries 24-25). Introduction of a bulky group such as diphen-
ylmethylene on the imine nitrogen could hardly show any
improvement. Thus, the imine 8f gave 9f in low yield of 30%
with 83% ee. Finally, we tried the allylation of the cis-cyclic
imines 28 and 31 (entries 27-28). Interestingly, the cis-imines
proved to be poorer substrates in the chiral allylation by this
procedure. The reason for this dramatic difference between the
trans- and cis-imines is discussed in the mechanistic section.
Mechanistic Explanations. A plausible mechanism for
allylation is shown in Scheme 3. As shown earlier7 the bis-π-
allylpalladium complex 4 is the reactive intermediate, in which
an allyl ligand acts as a transferable group and the other
nontransferable allyl group determines the stereocontrol of
allylation. Thus, the transmetalation between 2a and allyltribu-
tylstannane would produce tributylstannyl chloride and the bis-
π-allylpalladium complex 4, which would react with the imine
5 to give the π-allylpalladium amide 35 via complex 34. The
key step for chiral induction could be the coordination stage of
imine 5 to the bis-π-allylpalladium complex 4 to give 34 and
the subsequent allylation would proceed in a six-membered chair
like transition state to give 35. The subsequent transmetalation
of allyltributylstannane to palladium would produce the corre-
sponding stannyl homoallylamide 36 and regenerate 4 to
complete the catalytic cycle. Hydrolysis of 36 would give the
product homoallylamine 6. The role of water in this reaction
could be to form the pentacoordinate allylstannate, in which
water coordinates to the tetravalent stannane, and thereby
facilitates the C-Sn bond cleavage and enhances the trans-
metalation step in giving 4 and 36. This contributes to a faster
reaction rate and higher yields.19 The sulfonyl imine 22b (entry
18) failed to give the product amine in reasonable yield, though
it was consumed in the reaction. This probably could be due to
stabilization and decrease in nucleophilicity of nitrogen atom
by the electron withdrawing tosyl group20 in the intermediate
corresponding to 35. This intermediate could not undergo
transmetalation with allyltributylstannane to give the product
amine and reproduce 4. Thus, the catalytic cycle must be
blocked.
For the trans-imines, the probable transition state models are
shown in Scheme 4a. The front side of η3-10-methylpinene
group of the palladium catalyst is highly crowded by the methyl
at C-10 position, and thus an imine is forced to approach from
the less hindered rear side. The nitrogen of imine coordinates
to palladium atom and the C-C bond formation occurs through
the six-membered cyclic chair like transition state. In the case
of 38, there is severe steric repulsion between R group of the
imine and C-7 methylene group. Accordingly, the reaction
proceeds through a transition state model 37 to give the (R)-
homoallylamine predominantly. In the case of the cis-imines
(Scheme 4b), there is not much repulsion between the H of the
imine (model 40) and the C-7 methylene group, H being a small
atom. Hence, cyclic cis-imines without R-substituent to nitrogen
should not undergo severe steric interactions with the allylic
ligand within any of the proposed activated complexes. Thus,
there is not much energy difference between the transition state
models 39 and 40 resulting in poor enantioselectivity.
Conclusion
We have now improved our earlier allylation procedure and
most trans-imines reacted in shorter time, with good to high
(18) (a) van der Sluis, M.; Dalmolen, J.; de Lange, B.; Kaptein, B.; Kellogg, R.
M.; Broxterman, Q. B. Org. Lett. 2001, 3, 3943. (b) Cvetovich, R. J.;
Chartrain, M.; Hartner, F. W., Jr.; Roberge, C.; Amato, J. S.; Grabowski,
E. J. J. J. Org. Chem. 1996, 61, 6575.
(19) However, the enhancement of enantioselectivity by water is not clear.
Formation of pentacoordinate Si is known, see reference (2), probably
similar is the case with stannane.
(20) A similar explanation is given by Hou, see: ref (2c).
9
14138 J. AM. CHEM. SOC. VOL. 125, NO. 46, 2003

Chiral Bis-π-allylpalladium
A R T I C L E S
yields and moderate to high enantioselectivities. The catalyst
prepared is refined and separated from the undesired stereo-
isomer by an easy recrystallization procedure. We have dis-
covered that addition of 1 equiv of water shortens the reaction
time, enhances the yields, and enantioselectivities. We are now
in a position to synthesize diverse chiral homoallylamines from
achiral imines in catalytic manner with a more general,
reproducible, and robust procedure. The Pd-catalyzed asym-
metric allylation proceeds under essentially neutral conditions
in contrast to the Lewis acid catalyzed reaction, making it
feasible to use the substrates with labile functional groups (even
the halogens) and as well to apply the allylation to heterocyclic
substrates which otherwise sometimes are prone to polymeri-
zation. Although at this stage we could not justify the exact
role of water in the ee enhancement, further investigations are
in progress.
Acknowledgment. A.S. thanks JSPS for a research fellow-
ship.
Supporting Information Available: Experimental details for
catalyst preparation and separation and characterization data of
the products. This material is available free of charge via the
Internet at http://pubs.acs.org.
JA037272X
9
J. AM. CHEM. SOC. VOL. 125, NO. 46, 2003 14139