Pincer complex-catalyzed allylation of aldehyde and imine substrates via nucleophilic η1-allyl palladium intermediates

Article abstract of DOI:10.1021/ja049357j

Electrophilic allylic substitution of allylstannanes with aldehyde and imine substrates could be achieved by employment of palladium pincer complex catalysts. It was found that the catalytic activity of the pincer complexes is highly dependent on the ligand effects. The best results were obtained by employment of PCP pincer complexes with weakly coordinating counterions. In contrast to previous applications for electrophilic allylic substitutions via bisallylpalladium complexes, the presented reactions involve monoallylpalladium intermediates. Thus, employment of pincer complex catalysts extends the synthetic scope of the palladium-catalyzed allylic substitution reactions. Moreover, use of these catalysts eliminates the side reactions occurring in transformations via bisallylpalladium intermediates. The key intermediate of the electrophilic substitution reaction was observed by ¹H NMR spectroscopy. This intermediate was characterized as an η¹-allyl- coordinated pincer complex. Density functional theory (DFT) modeling shows that the electrophilic attack can be accomplished with a low activation barrier at the γ-position of the η¹-allyl moiety. According to the DFT calculations, this reaction takes place via a six-membered cyclic transition-state (TS) structure, in which the tridentate coordination state of the pincer ligand is preserved. The stereoselectivity of the reaction could be explained on the basis of the six-membered cyclic TS model.

Full text of DOI:10.1021/ja049357j

Published on Web 05/08/2004
Pincer Complex-Catalyzed Allylation of Aldehyde and Imine
Substrates via Nucleophilic η1-Allyl Palladium Intermediates
Niclas Solin, Johan Kjellgren, and Ka´lma´n J. Szabo´*
Contribution from the Arrhenius Laboratory, Department of Organic Chemistry,
Stockholm UniVersity, SE-106 91 Stockholm, Sweden
Received February 5, 2004; E-mail: kalman@organ.su.se
Abstract: Electrophilic allylic substitution of allylstannanes with aldehyde and imine substrates could be
achieved by employment of palladium pincer complex catalysts. It was found that the catalytic activity of
the pincer complexes is highly dependent on the ligand effects. The best results were obtained by
employment of PCP pincer complexes with weakly coordinating counterions. In contrast to previous
applications for electrophilic allylic substitutions via bisallylpalladium complexes, the presented reactions
involve monoallylpalladium intermediates. Thus, employment of pincer complex catalysts extends the
synthetic scope of the palladium-catalyzed allylic substitution reactions. Moreover, use of these catalysts
eliminates the side reactions occurring in transformations via bisallylpalladium intermediates. The key
intermediate of the electrophilic substitution reaction was observed by 1H NMR spectroscopy. This
intermediate was characterized as an η1-allyl-coordinated pincer complex. Density functional theory (DFT)
modeling shows that the electrophilic attack can be accomplished with a low activation barrier at the
γ-position of the η1-allyl moiety. According to the DFT calculations, this reaction takes place via a six-
membered cyclic transition-state (TS) structure, in which the tridentate coordination state of the pincer
ligand is preserved. The stereoselectivity of the reaction could be explained on the basis of the six-membered
cyclic TS model.
1. Introduction
version of these reactions, the key step involves electrophilic
attack on a bisallylpalladium complex (Scheme 1), which can
be generated from a monoallylpalladium complex and allyl-
stannane.3,13 Subsequently, this bisallylpalladium complex reacts
with an electrophile, affording the allylated product and the
regenerated catalyst.3 An important application of this trans-
formation involves enantioselective allyl transfer via chiral
allylpalladium complexes.4,7
Theoretical studies revealed that the electrophilic attack
proceeds via an η1,η3-bisallylpalladium species formed from the
η3,η3-form by ligand coordination (Scheme 2).14 The electro-
philic attack takes place at the γ-position of the η1-moiety with
a remarkably low activation barrier. It was also concluded that
the other η3-coordinated allyl group acts as an electron-supplying
spectator ligand. Accordingly, the spectator ligands (the η3
moiety and L) together occupy three of the four coordination
sites on palladium.
In the traditionally used version of the electrophilic substitu-
tion reactions, formation of various isomeric bisallylpalladium
complexes with different reactivities (Scheme 1) imposes
considerable limitations on the synthetic scope of the catalytic
transformation. Catalytic formation of unsymmetrically substi-
tuted η3,η3-bisallylpalladium complexes (R * H, Scheme 1)
involves generation of various η1,η3-intermediates, and therefore,
either allyl group can be transferred to the electrophile.8,9 Control
Palladium-catalyzed allylic substitution reactions via allylpal-
ladium intermediates represent one of the most important
procedures in modern organic synthesis.1,2 Palladium-catalyzed
allylic displacement of acetate, carbonate, and their congeners
with nucleophilic reagents has became a very important and
versatile synthetic procedure. This nucleophilic substitution
reaction involves oxidative addition of palladium(0) to the allylic
substrate, generating an η3-(mono)allylpalladium intermediate
followed by a nucleophilic attack, providing the allylated product
and regenerating the palladium(0) catalyst.1,2
Possibilities to extend the synthetic scope of the palladium-
catalyzed allylic substitutions to electrophilic substrates have
recently been the subject of a great deal of interest in mechanistic
and in synthetic organic chemistry.3-12 In the commonly used
(1) Tsuji, J. PerspectiVes in Organopalladium Chemistry for the 21st Century;
Elsevier: Amsterdam, 1999.
(2) Tsuji, J. Palladium Reagents and Catalysis: InnoVations in Organic
Synthesis; Wiley: Chichester, U.K., 1995.
(3) Nakamura, H.; Iwama, H.; Yamamoto, Y. J. Am. Chem. Soc. 1996, 118,
6641.
(4) Nakamura, H.; Nakamura, K.; Yamamoto, Y. J. Am. Chem. Soc. 1998,
120, 4242.
(5) Nakamura, H.; Shim, J.-G.; Yamamoto, Y. J. Am. Chem. Soc. 1997, 119,
8113.
(6) Nakamura, H.; Aoyagi, K.; Shim, J.-G.; Yamamoto, Y. J. Am. Chem. Soc.
2001, 123, 372.
(7) Fernandes, R. A.; Stimac, A.; Yamamoto, Y. J. Am. Chem. Soc. 2003,
125, 14133.
(8) Solin, N.; Narayan, S.; Szabo´, K. J. J. Org. Chem. 2001, 66, 1686.
(9) Solin, N.; Narayan, S.; Szabo´, K. J. Org. Lett. 2001, 3, 909.
(10) Wallner, O. A.; Szabo´, K. J. Org. Lett. 2002, 4, 1563.
(11) Wallner, O. A.; Szabo´, K. J. J. Org. Chem. 2003, 68, 2934.
(12) Wallner, O. A.; Szabo´, K. J. Chem. Eur. J. 2003, 9, 4025.
(13) Goliaszewski, A.; Schwartz, J. Tetrahedron 1985, 41, 5779.
(14) Szabo´, K. J. Chem. Eur. J. 2000, 6, 4413.
9
7026
J. AM. CHEM. SOC. 2004, 126, 7026-7033
10.1021/ja049357j CCC: $27.50 © 2004 American Chemical Society

Palladium-Catalyzed Allylic Substitution Reactions
A R T I C L E S
Scheme 1. Formation, Reactivity, and Isomerization Possibilities of Bisallylpalladium Intermediates
Scheme 2. Mechanism of the Electrophilic Attack at the η1-Allyl
Moiety of an η3,η3-Bisallylpalladium Complex
Chart 1. Studied Active Pincer Complex Catalysts
of the regioselectivity is a particularly important issue in
asymmetric applications,4,7 where the electrophilic attack on the
chiral allyl moiety must be avoided to maintain the asymmetric
activity of the allylpalladium catalyst. A further important
problem is that bisallylpalladium complexes may undergo allyl-
allyl (Stille) coupling instead of reaction with electrophiles.13,15-17
This reaction particularly easily occurs in the presence of
excessive amounts of strongly coordinating ligands (L, Scheme
1), which promotes the formation of the η1,η1-bisallylic isomer.
Echavarren and Ca´rdenas and co-workers17 have recently shown
that η1,η1-bisallylpalladium complexes easily undergo reductive
elimination to give 1,5-hexadiene derivatives.
Accordingly, a further innovative development of the pal-
ladium-catalyzed electrophilic substitution reactions requires that
the catalytic transformations do proceed entirely via mono-
allylpalladium intermediates without involvement of bisallyl-
palladium species. However, development of such catalytic
transformations poses a great challenge, since the usual monoallyl-
palladium complexes are considerably more stable as η3-
coordinated species18 (in which the allyl moiety has an elec-
trophilic character) than in their η1-coordinated forms required
in the electrophilic attack (Scheme 2). Therefore, the spectator
ligand in a monoallylpalladium complex reactive toward elec-
trophiles has to fulfill at least four important criteria: (a) The
spectator ligand has to perform a tridentate coordination to
palladium. Thus only one free coordination site will be available
on palladium to ensure η1-coordination of the allyl moiety. (b)
A strong electron donor ability is required to lend a nucleophilic
character to the allyl group. (c) A firm palladium-ligand
bonding is necessary to avoid undesired dynamic processes and
ligand exchange. (d) The electron-supplying part of the spectator
ligand must be forced to a trans position relative to the allyl
ligand in order to avoid reductive coupling (cf. Stille coupling)
between the spectator ligand and the allyl moiety.
As we have shown in a recent paper, the above requirements
can be fulfilled by application of PCP- (1-2) type pincer
complexes in the catalytic electrophilic substitution reactions.19
These type of pincer complexes are characterized by strong
tridentate ligand coordination and direct aryl-metal bonding,
which provides an electron-rich palladium atom (Chart 1).20-26
In this paper we give a full account of our results on pincer
complex-catalyzed electrophilic substitution reactions. In addi-
tion, we present our new studies on some important synthetic
and mechanistic aspects of this reaction: (i) possibilities to
improve the catalytic activity of the applied pincer catalysts (cf.
3); (ii) direct observation of the allylpalladium intermediate of
the reaction to assess the reaction mechanism; and (iii) modeling
the electrophilic attack on the allylic moiety of the pincer
complex intermediate of the reaction by density functional
theory (DFT).
2. Catalytic Electrophilic Substitution by Pincer
Complexes
As we communicated before,19 PCP pincer complexes 1 and
2 are able to catalyze the electrophilic substitution of allylstan-
(19) Solin, N.; Kjellgren, J.; Szabo´, K. J. Angew. Chem., Int. Ed. 2003, 42,
3656.
(20) Albrecht, M.; Koten, G. v. Angew. Chem., Int. Ed. 2001, 3750.
(21) Singleton, J. T. Tetrahedron 2003, 59, 1837.
(22) Rimml, H.; Venanzi, L. M. J. Organomet. Chem. 1983, 259, C6.
(23) Haenel, M. W.; Oevers, S.; Bruckmann, J.; Kuhnigk, J.; Kru¨ger, C. Synlett
1998, 301.
(24) Haenel, M. W.; Jakubik, D.; Kru¨ger, C.; Betz, P. Chem. Ber. 1991, 124,
333.
(25) Beek, J. A. M. v.; Koten, G. v.; Dekker, G. P. C. M.; Wissing, E.; Zoutberg,
M. C.; Stam, C. H. J. Organomet. Chem. 1990, 394, 659.
(26) Slagt, M. Q.; Rodr´ıguez, G.; Grutters, M. M. P.; Gebbink, R. J. M. K.;
Klopper, W.; Jenneskens, L. W.; Lutz, M.; Spek, A. L.; Koten, G. v. Chem.
Eur. J. 2004, 10, 1331.
(15) Goliaszewski, A.; Schwartz, J. J. Am. Chem. Soc. 1984, 106, 5028.
(16) Nakamura, H.; Bao, M.; Yamamoto, Y. Angew. Chem., Int. Ed. 2001, 40,
3208.
(17) Me´ndez, M.; Cuerva, J. M.; Go´mez-Bengoa, E.; Ca´rdenas, D. J.; Echavarren,
A. M. Chem. Eur. J. 2002, 8, 3620.
(18) Solin, N.; Szabo´, K. J. Organometallics 2001, 20, 5464.
9
J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004 7027

A R T I C L E S
Solin et al.
Scheme 3. Pincer Complex-Catalyzed Electrophilic Substitution
Table 1. Pincer Complex-Catalyzeda Electrophilic Substitution
Reactions
Reactions
1
c
entry substrate 1 substrate 2 cat. prod.
Q
R
solvent conditionsb yield
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4a
4b
4b
4b
4b
4cf
4d
4d
4b
5a
5a
5a
5b
5c
5c
5d
5d
5d
5d
5d
5e
5e
5f
1a 7a C6H5
2a 7a C6H5
H
DMF
THF
THF
DMF
DMF
THF
DMF
60/22
60/21
40/18
60/17
60/17
25/15
60/17
81
88
92
88
82
95
82
95
95
95
95
80
80
73
72
66
69
95
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
H
3
7a C6H5
1a 7b 4-CN-C6H5
1a 7c 4-CH3CO-C6H5
3
7c 4-CH3CO-C6H5
1a 7d 4-NO2-C6H5
1b 7d 4-NO2-C6H5
1c 7d 4-NO2-C6H5
2a 7d 4-NO2-C6H5
CHCl3 40/2
CHCl3 20/3
was complete in a couple of hours at room temperature or at
40 °C. However, the acetate and tetrafluoroborate complexes
were rapidly deactivated under the applied catalytic conditions.
Therefore, these catalysts could not be employed for allylation
of nonactivated aldehydes (such as 5a and 5f,g) requiring a
longer reaction time.
THF
THF
THF
THF
THF
THF
DMF
THF
THF
40/21
25/3
3
7d 4-NO2-C6H5
2a 7e cinnamyl
60/21
40/24
60/28
60/26
60/17
60/21
40/16
3
3
3
7e cinnamyl
7f hexyl
7g cyclohexyl
5g
6
6
1a 8a
2a 8a
Increase of the Catalytic Activity by Employment of
Complex 3. Since we found that the catalytic activity of pincer
complexes with weakly coordinating counterions is higher than
with chloride counterion, we considered using trifluoroacetate
complex 3 as a catalyst in the electrophilic allylic substitution
reactions. This pincer complex is readily available by a two-
step synthesis reported by Bedford et al.32 Complex 3 is highly
stable under the applied reaction conditions and its catalytic
activity is preserved at least for 24 h even at 60 °C.
6
3
8a
5a
5d
5d
5g
5d
5d
5a
6
2a 7h C6H5
2a 7i 4-NO2-C6H5
15 7i 4-NO2-C6H5
Ph THF
Ph THF
Ph THF
Ph THF
Me DMF
Me THF
Me THF
Ph THF
60/21 61d
60/21 95d
20/64 85d
60/40 49e
60/14 77g
40/14 65g
60/40 57g
60/21 65h
3
7j cyclohexyl
1a 7k 4-NO2-C6H5
3
3
7k 4-NO2-C6H5
7l C6H5
2a 8b
a Catalyst (5 mol %) was applied. b Temperature (°C)/time (h). c Isolated
yield. d Diastereomer ratio anti/syn28 ) 10:1. e Diasteromer ratio anti/syn29
) 14:1. f Z/E ratio ) 2:1. g Diasteromer ratio syn/anti30 ) 1:1. h Diastere-
omer ratio syn/anti31 ) 12:1.
As expected, trifluoroacetate complex 3 outperforms chloro
complexes 1a and 2a in all catalytic reactions. Thus benzalde-
hyde (5a) was allylated at 40 °C in excellent yield with 3 as a
catalyst (entry 3), while the reaction with complexes 1a and 2a
had to be conducted at 60 °C and for a longer time (entries 1
and 2). Furthermore, allylation of aldehyde 5c was performed
at room temperature with 3 (entry 6), while 60 °C reaction
temperature was required for the same process catalyzed by 1a
(entry 5). By employment of 3, nitrobenzaldehyde (5d) could
be allylated at about the same temperature and in the same
reaction time (entry 11) as with tetrafluoroborate complex 1c.
Not only benzaldehyde derivatives but also cinnamyl aldehyde
(5e) (cf. entries 12 and 13) and benzenesulfonimine (6) (cf.
entries 16-17 and 18) reacted faster with 4a in the presence of
3 than with catalytic amounts of 1a and 2a. Most importantly,
electrophilic substitution of allylstannane 4a with aliphatic
aldehydes 5f,g could be achieved with high yields by employ-
ment of catalyst 3 (entries 14 and 15), while the same catalytic
reaction with 1 and 2 resulted in only traces of products under
similar or even under harsh reaction conditions. Cinnamylstan-
nane 4b could also be reacted with 5g, affording cyclohexyl
derivative 7j in a moderate yield and with excellent stereo- and
regioselectivity (entry 22). Crotylstannanes (4c,d) react much
slower with aldehydes than the parent allylstannane (4a).
Crotyltributylstannane (4d) reacts very slowly with 5d even at
60 °C in the presence of 1a. Therefore, we employed crotyl-
trimethylstannane (4c) to accomplish this reaction (entry 23).
However, in the presence of the highly active catalyst 3,
stannane 4c could be replaced by the less reactive substrate 4d,
and moreover, the reaction temperature was decreased to 40
°C (entry 24). Even catalytic coupling of benzaldehyde (5a)
nanes 4a-c with aldehyde (5a-e) and imine (6) electrophiles.
By use of 5 mol % catalyst 1a and 2a, a typical reaction was
conducted at 60 °C overnight to obtain a full conversion of the
reactants (Scheme 3).
All reactions proceed under neutral conditions, and therefore
many functionalities such as the cyano (entry 4), keto (entry
5), and nitro (entries 7-10) groups are tolerated (Table 1). The
high chemoselectivity of the reaction was demonstrated by
allylation of keto aldehyde 5c (entry 5). In this process the
aldehyde functionality could be allylated, while the keto group
remained unchanged.
We have also studied the regio- and stereoselectivity of the
reaction. It was found that cinnamyl stannane (4b) readily
undergoes palladium-catalyzed allylic substitution with alde-
hydes 5a and 5d, providing exclusively the branched allylic
products (7h and 7i) with high anti diastereoselectivity (entries
19 and 20). Sulfonimine 6 and stannane 4b also react with the
same regioselectivity and with a high stereoselectivity (entry
26). Interestingly, in this reaction the major stereoisomer is the
syn form. The catalytic substitution reaction with crotyl stannane
(4c) also gives the branched allylic isomer (7k); however, the
two diastereomers form in a 1:1 ratio (entry 23).
We attempted to increase the catalytic activity of the pincer
complexes by replacement of the strongly coordinating chloride
counterion with acetate and tetrafluoroborate (e.g., 1b,c).27
Comparison of entry 7 with entries 8 and 9 clearly indicate that
acetate complex 1b and tetrafluoroborate complex 1c are more
reactive catalysts than chloro complex 1a. With catalytic
amounts of 1b and 1c, the reaction with nitrobenzaldehyde (5d)
(28) Takahara, J. P.; Masuyama, Y.; Kurusu, Y. J. Am. Chem. Soc. 1992, 114,
2577.
(29) Loh, T.-P.; Tan, K.-T.; Hu, Q.-Y. Angew. Chem., Int. Ed. 2001, 40, 2921.
(30) Batey, R. A.; Thadani, A. N.; Smil, D. V.; Lough, A. J. Synthesis 2000,
990.
(31) Lu, W.; Chan, T. H. J. Org. Chem. 2000, 65, 8589.
(32) Bedford, R. B.; Draper, S. M.; Scully, P. N.; Welch, S. L. New J. Chem.
2000, 24, 745.
(27) Complexes 1b,c and 2b were generated from 1a and 2a, respectively. The
corresponding chloro complex was dissolved in CHCl3, mixed with AgBF4
(1b and 2b) or AgOAc (1c) at room temperature, and then the mixture
was stirred for 30 min. The precipitate was removed and the resulting clear
solution was used without further purification.
9
7028 J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004

Palladium-Catalyzed Allylic Substitution Reactions
A R T I C L E S
Figure 1. Formation of 7d in the reaction of 4a and 5d (cf. entries 10 and
11) catalyzed by pincer complexes 2a (9) and 3 (b) in CDCl3 at 25 °C.
Scheme 4. Formation of η1-Allylpalladium Complex 9 by
Transmetalation of Allylstannane 4e with Pincer Complex 3
1
Figure 2. (a) H NMR spectrum of the reaction of complex 3 with allyl
stannane 4e. (b) 1H NMR spectrum of the reaction of complex 3 with
allylmagnesium bromide. Both reaction were conducted in THF-d8 (i, THF-
d7; ii, ether; iii, CH2-MgBr). The shift values are given in parts per million.
Chart 2. 1H NMR Shift Values (in ppm) Reported for
η1-Allylpalladium Complexes Analogous to 936,37
with 4d could be performed with catalyst 3 (entry 25), while
this reaction is extremely slow with catalysts 1a and 2a under
the same conditions.
We have monitored the catalytic reaction of 4a with 5d by
1H NMR spectroscopy at 25 °C (Figure 1). The diagram in
Figure 1 clearly indicates that the rate of formation of 7d is
considerably higher with trifluoroacetate complex catalyst 3 than
with chloro complex 2a. This experiment confirms the above
findings that pincer complex 3 is a formidable catalyst for the
electrophilic substitution of allylstannanes.
allylstannanes 4a and 4e even in the absence of aldehydes.19
Monitoring of this reaction with 1H NMR spectroscopy revealed
formation of propene from the allylstannane substrate. We
supposed that the reaction involved transmetalation of the
allylstannanes with palladium; however, the η1-allylpalladium
intermediate formed probably instantaneously reacted with traces
of water present in the reaction mixture to give propene.
However, when a similar reaction was carried out with
complex 3 and allyltrimethylstannane 4e in dry THF-d8, a
number of relevant peaks between 2 and 5 ppm could be
3. Mechanistic Aspects
It is well-known that palladium(II) species readily undergo
transmetalation with allylstannanes to form allylpalladium
complexes. Transmetalation of (mono)allylpalladium complexes
with allylstannane derivatives gives bisallylpalladium com-
plexes, which subsequently react with electrophiles (Scheme
1).3,13 Palladium-pincer complexes have been reported to
undergo transmetalation with organometallic reagents including
organostannanes.33-35 For example, Cotter et al.33 have shown
that 2-furylstannane undergoes tin-to-palladium transmetalation
with the triflate salt of 2. Therefore, it is reasonable to assume
that the first step of the studied reaction (Scheme 3) is
transmetalation of the allylstannane (4) with the pincer complex
catalyst (1-3) to give an η1-allyl-coordinated pincer species
(Scheme 4).
1
observed in the H NMR spectrum of the process (Scheme 4
and Figure 2a). One of these peaks appeared at 2.29 ppm as a
doublet of triplets with coupling constants of 8.8 Hz (3JHH) and
5.1 Hz (JHP). This shift value and coupling pattern are char-
acteristic for the R-protons of the η1-allyl moiety in analogous
pincer complexes (cf. Scheme 4 and Chart 2). For example,
this proton resonates at 2.62 ppm with the same coupling pattern
3
(dt, JHH ) 8.4 Hz, JHP ) 5.3 Hz) in complex 10 reported by
Osborn and co-workers.36 The resonance value for the R-proton
in 9 (2.29 ppm) is also in the range of NMR shifts reported for
similar complexes (Chart 2) such as 1036 (2.62 ppm) and 1137
(1.98 ppm). Two other shift values at 4.15 and 3.84 ppm are
also of great interest, since these shifts can be assigned to the
γ-protons of the η1 moiety of 9. The shift values are again very
close to the γ-proton shifts reported for complex 10 (Chart 2).
Reactive Intermediate of the Reaction. Our previous studies
have shown that pincer complexes 1c and 2b react with
1
To confirm the above H NMR shift assignments, we have
carried out a control experiment. This experiment involved
(33) Cotter, W. D.; Barbour, L.; McNamara, K. L.; Hechter, R.; Lachicotte, R.
J. J. Am. Chem. Soc. 1998, 120, 11016.
(34) Ohff, M.; Ohff, A.; Boom, M. E. v. d.; Milstein, D. J. Am. Chem. Soc.
1997, 119, 11687.
(36) Barloy, L.; Ramdeehul, S.; Osborn, J. A.; Carlotti, C.; Taulelle, F.; Cian,
A. D.; Fischer, J. Eur. J. Inorg. Chem. 2000, 2523.
(37) Kuhn, O.; Mayr, H. Angew. Chem., Int. Ed. 1999, 38, 343.
(35) Rimml, H. Thesis, ETH, Zu¨rich, Switzerland, 1984.
9
J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004 7029

A R T I C L E S
Solin et al.
Scheme 5. Allyl Group Transfer from Complex 9 to the Aldehyde
Electrophile
Species 12a-c involve simplified models for the pincer ligand
of 3 in which the phenyl groups are replaced by hydrogen atoms.
In these calculations the benzaldehyde electrophile (5a) was
approximated by an acetaldehyde molecule. The η1-allyl species
12a represents a minimum on the potential energy surface (PES)
according to the vibrational analysis. The ligating atoms on
palladium are in a square planar arrangement with a strong
Pd-P coordination.
Complex 12b was characterized as a TS structure, as its
vibrational analysis gave only one imaginary frequency. The
transition vector corresponds to the bond formation between
the γ-terminus of the η1-allyl moiety and the carbonyl carbon
of acetaldehyde. The activation energy is only 12.2 kcal mol-1
(the zero-point vibration corrected value is 14.0 kcal mol-1),
which is similar to the activation barrier of these types of
electrophilic attacks on bisallylpalladium complexes (10-15
kcal mol-1).8,12,14
In 12b the η1-allyl moiety, the carbonyl group, and the
palladium atom form a cyclic six-membered ring structure with
distinct equatorial and axial positions. This also requires that
the R-carbon atom of the allyl group and the oxygen of the
aldehyde share a single coordination site on palladium. Thus
the tridentate pincer-type ligation remains largely unaffected
in the TS structure. This is also indicated by the fact that the
strong Pd-P coordination is still maintained. As one goes from
GS 12a to TS structure 12b, the Pd-P bond is changed by only
0.02 Å. The vibrational analysis of the product complex 12c
gave only real frequencies, indicating that it is a minimum on
the PES. Formation of 12c from 12a is exoenergetic with 11.3
kcal mol-1 (7.9 kcal mol-1 with ZPV correction). In this
complex the oxygen of the homoallyl product is coordinated to
palladium, while the carbon-carbon double bond is uncoordi-
nated.
The above DFT calculations with model systems 12a-c
indicate that the electrophilic substitution of the η1-allyl moiety
occurs readily with a low activation barrier. However, bulky
phenyl substituents on the phosphorus atoms in the rigid firmly
tricoordinated pincer ligand can be involved in destabilizing
steric interactions with the substituents of the aldehyde com-
ponent. These interactions can considerably increase the activa-
tion energy of the reaction and change the TS geometry.
Therefore, we undertook theoretical studies for the realistic
species 13a-c, including phenyl groups on the pincer ligands
and on the aldehyde substrate. As one goes from model complex
12a to complex 13a, the allyl group is rotated by about 90° to
avoid the steric interactions with the phenyl groups of the pincer
ligand. Otherwise, the geometrical parameters describing the
allyl-metal and the Pd-P bonding are almost identical. The
activation energy of the electrophilic attack by benzaldehyde
on 13a is 13.1 kcal mol-1 (13b), which is only 0.9 kcal mol-1
higher than the activation energy obtained for the model systems
reaction of allylmagnesium bromide in place of 4e with pincer
complex 3. It is well-known that the allyl Grignard reagents
are much more reactive organometallic substrates than allyl-
stannanes, and therefore these reagents instantaneously trans-
metalate with palladium(II) species including allylpalladium
complexes.13,15,38 Indeed, the same characteristic peaks at 2.29,
1
3.84, and 4.15 ppm appeared in the H NMR spectrum of the
reaction of allylmagnesium bromide with 3 (Figure 2b) as for
the stoichiometric process of allylstannane 4e with complex 3
(Figure 2a). Accordingly, we conclude that the stoichiometric
reaction of 4e with pincer complex 3 leads to η1-allylpalladium
pincer complex 9, which we consider as the reactive intermediate
of the studied catalytic electrophilic substitution (Scheme 3).
The above stoichiometric reaction (Scheme 4) can also be
performed with allyltributylstannane 4a. However, in the
presence of a large excess of allylstannane 4a the 9:3 ratio was
only 0.18, which was remained unchanged in a longer period
of time. On the other hand, when allyltrimethylstannane 4e was
employed (Figure 2a), the 9:3 ratio became considerably higher,
up to 0.56. These observations indicate that under catalytic
conditions the active catalyst 9 is generated in a lower
concentration from allyltributylstannane derivatives than from
allyltrimethylstannane substrates. This also explains the lower
reactivity of allyltributylstannanes compared to allyltrimethyl-
stannanes in the catalytic reactions.
DFT Modeling of the Electrophilic Attack. The key step
of the catalytic reaction is obviously the coupling between the
allyl moiety of the η1-allyl complex (such as 9) with aldehydes
and imine electrophiles (Scheme 5). A peculiar feature of this
reaction step is that only a single coordination site on palladium
is available for the entire process. Thus, rationalization of the
mechanism of the electrophilic attack on the η1-allyl moiety of
9 (and its analogues) requires an explicit knowledge about the
transition-state (TS) structure and the activation energy of the
reaction. Therefore, we carried out density functional theory
(DFT) calculations for the most important ground-state (GS)
and TS structures of the process.
All geometries (12 and 13) were fully optimized by employ-
ing a Becke-type39 three-parameter density functional model
B3PW91 (Figure 3) with a double-ú(DZ)+P basis constructed
from the LANL2DZ basis40-42 by adding one set of d-
polarization functions to the heavy atoms (exponents: C, 0.63;
O, 1.154; P, 0.34) and one set of diffuse d-functions on
palladium (exponent 0.0628). All calculations have been carried
out by employing the Gaussian98 program package.43
(43) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M.
A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A.; Stratmann,
R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels, A. D.; Kudin,
K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone, V.; Cossi, M.; Cammi,
R.; Mennucci, B.; Pomelli, C.; Adamo, C.; Clifford, S.; Ochterski, J.;
Petersson, G. A.; Ayala, P. Y.; Q. Cui; Morokuma, K.; Malick, D. K.;
Rabuck, A. D.; Raghavachari, K.; Foresman, J. B.; Cioslowski, J.; Ortiz,
J. V.; Stefanov, B. B.; Liu, G.; Liashenko, A.; Piskorz, P.; Komaromi, I.;
Gomperts, R.; Martin, R. L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng,
C. Y.; A. Nanayakkara; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.;
Johnson, B.; Chen, W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-
Gordon, M.; Replogle, E. S.; Pople, J. A. Gaussian 98; Gaussian, Inc.:
Pittsburgh, PA, 1998.
(38) Henc, B.; Jolly, P. W.; Salz, R.; Wilke, G.; Benn, R.; Hoffmann, E. G.;
Mynott, R.; Schroth, G.; Seevolgel, K.; Sekutowski, J. C.; Kru¨ger, C. J.
Organomet. Chem. 1980, 191, 425.
(39) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(40) Dunning, T. H.; Hay, P. J. Modern Theoretical Chemistry; Plenum: New
York, 1977; Vol. 3.
(41) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270.
(42) Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 299.
9
7030 J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004

Palladium-Catalyzed Allylic Substitution Reactions
A R T I C L E S
Figure 3. B3PW91/LANL2DZ+P optimized structures occurring in the electrophilic attack on the allyl moiety. The bond lengths are given in angstroms.
The zero-point vibration corrected energies are shown in italic type; all energies are given in kilocalories per mole.
(12a f 12b). Furthermore, the reaction also proceeds through
a six-membered cyclic transition state (13b) and the geometry
of the TS structures 13b and 12b is very similar. The reaction
is exoenergetic with 13.0 kcal mol-1 (13c), which compares
well to the energetic features of the model reaction (12.2 kcal
mol-1). Because of the large size of 13a-c, the vibrational
analysis is computationally prohibitive. However, the close
resemblance between the corresponding species of 12a-c and
13a-c indicates that 13a and 13c are GS structures and 13b is
a TS structure. Accordingly, it can be concluded that the steric
effects of the phenyl substituents of the pincer ligand and the
substituents of the aldehyde component has a relatively weak
effect on the energetic features and on the TS geometry of the
electrophilic attack.
Catalytic Cycle of the Reaction. On the basis of the results
of the stoichiometric reaction of 3 with allylstannane 4e (Scheme
4), the introducing step of the catalytic reaction is the trans-
metalation of the allylstannane substrate with the pincer complex
catalyst (Scheme 6). The stoichiometric reactions indicate that
this transmetalation is an equilibrium process, since only a
certain portion of the pincer complex 3 is converted to 9 even
in the presence of a large excess of allylstannane (4a or 4e).
Therefore, electronic and steric effects shifting the equilibrium
to formation of the η1-allyl complex 9 are particularly important
9
J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004 7031

A R T I C L E S
Solin et al.
Scheme 6. Catalytic Cycle of the Pincer Complex-Catalyzed
Electrophilic Substitution Reaction
Scheme 7. Rationalization of the Stereoselectivity of the Catalytic
Transformation
the other hand, formation of complex 9 is facilitated by a low
electron density on palladium (vide supra). Therefore, a careful
consideration of the electronic effects of the pincer ligands is
necessary to obtain a catalyst that shows high reactivity in both
the transmetalation step and the electrophilic attack of the
catalytic cycle. Pincer complexes 1-3 represent electronically
well-balanced catalysts working equally well in both reaction
steps. On the other hand, pincer complexes 15-17 (Chart 3)
bearing an η1-allyl ligand are expected to undergo a facile
electrophilic attack; however, formation of these species by
transmetalation from allylstannane is probably very slow.
The regioselectivity of the reaction (entries 19-26) can be
explained by the substituent effects of the alkyl and aryl groups
on the stability of η1-allyl moieties. Our previous studies have
clearly shown that the terminally substituted η1-allyl complexes
are more stable when the alkyl or aryl substituents are attached
at the γ-terminus of the allyl moiety.9-12,14,18,19 Since the
electrophilic attack also occurs at the γ-position of the η1-allyl
moiety, the catalytic reaction affords the branched allyl isomer.
Cinnamylstannane 4b, comprising a bulky phenyl substituent,
reacted with high stereoselectivity (entries 19-22 and 26).
Interestingly, the major stereoisomer with aldehydes 5a, 5d, and
5g is the anti isomer, while the reaction with sulfonimine 6
affords mainly the syn diastereomer. The DFT calculations
clearly show that the electrophilic attack proceeds through a
six-membered cyclic TS, in which the allylic terminal substit-
uents and the substituent of the aldehyde component occupy
distinct axial and equatorial positions (12b and 13b). In the
calculated structures the orientation of the aldehyde molecules
renders their substituents (Me in 12b and Ph in 13b) to an
equatorial position. Considering that the trans geometry repre-
sents the most stable form for the cinnamyl moiety (18, Scheme
7), the phenyl substituent will also occupy an equatorial position
in the TS structure (19). Thus in the most stable TS structure
the bulky substituents are in a trans-diequatorial arrangement
over the developing carbon-carbon bond, which leads to trans-
diastereomer 20. A similar model was employed to explain the
stereoselectivity of the electrophilic attack via bisallylpalladium
intermediates.12 When crotylstannane 4c,d is used, the level of
the stereoselectivity is determined by the steric effects of the
methyl substituent in the allyl moiety. Equal formation of syn
and anti diastereomers (entries 23 and 24) can be ascribed to
the small size of the allylic methyl substituent.
Chart 3. Pincer Complexes with Low Catalytic Activity in the
Electrophilic Substitution Reactions
factors to increase the rate of the catalytic transformation.
Weakly coordinating counterions, such as trifluoroacetate (3),
acetate (1b), and tetrafluoroborate (1c and 2b), can be much
more easily replaced with the allyl group than the chloride ion
(1a and 2a), which explains the high catalytic activity of 3 (1b,c)
compared to 1a and 2a. In the transmetalation step, formally
an allyl anion is transferred from 4a to the pincer complex, and
thus formation of 9 is thermodynamically not preferred when
the electron density on palladium is too high. A high electron
density on palladium can be generated by employment of
σ-donor pincer ligands. Therefore, we have also investigated
the catalytic activity of NCN complexes 1525 and 1644 and
carbene complex 17 (Chart 3).45 All three complexes exhibited
low catalytic activity in the electrophilic substitution reactions,
which probably can be due to the slow transmetalation step.
As mentioned above, the steric and electronic features of the
allylstannane substrate also influence the transmetalation. For
example, allyltrimethylstannanes (4b and 4c) are more reactive
than their tributyl-substituted counterparts (e.g., 4a and 4d) in
the catalytic reaction.
The electrophilic attack is probably the most important
reaction step since it determines the regio- and stereoselectivity
of the catalytic process. According to the above DFT studies,
PCP complex 9 readily undergoes electrophilic attack with
aldehydes (Scheme 5 and Figure 3). However, the electrophilic
attack and the transmetalation step have opposite electron
demands. In intermediate 9 a high electron density on palladium
increases the nucleophilic character of the η1-allyl moiety. On
(44) Denmark, S. E.; Stavenger, R. A.; Faucher, A.-M.; Edwards, J. P. J. Org.
Chem. 1997, 62, 3375.
(45) Gru¨ndemann, S.; Albrecht, M.; Loch, J. A.; Faller, J. W.; Crabtree, R. H.
Organometallics 2001, 20, 5485.
9
7032 J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004

Palladium-Catalyzed Allylic Substitution Reactions
A R T I C L E S
5. Experimental Section
The stereoselectivity in the reaction of 4b with 6 (entry 26)
is strongly influenced by the trans geometry of the phenylsul-
fonyl and phenyl groups across the carbon-nitrogen double
bond in 6. Because of this geometry, the lone pair on nitrogen
(interacting with palladium) and the phenyl group are in a cis
arrangement. Furthermore, it is reasonable to assume that in
the TS structure 21 the close vicinity between the palladium
atom and the sulfonyl group will be avoided. As a consequence,
the preferred orientation of 6 in TS structure 21 renders the
phenyl group to an axial position. Since the allylic phenyl group
is in an equatorial position, this model predicts formation of
the syn diastereomer 22. It is interesting to note that a similar
syn diastereoselectivity was reported for the formation of 8b
from 6 and cinnamyl bromide in indium- and zinc-mediated
reactions.31
The theoretical calculations have shown that the electrophilic
attack results in an alkoxide-coordinated homoallylic alcohol
(12c and 13c) as product. Under catalytic conditions these
complexes (such as 14) may undergo ligand exchange with the
organotin salts formed in the transmetalation to give the final
product and to regenerate the catalyst (Scheme 6). This reaction
probably easily occurs due to the well-known low coordination
ability of the alkoxides to palladium.
The NMR spectra were recorded on a Varian 400 MHz NMR
spectrometer. The chemical shifts (ppm) are obtained by use of residual
solvent as internal standard in 1H NMR (THF-d8, 3.58 ppm) and
phosphoric acid in 31P NMR measurements. Merck silica gel 60 (230-
400 mesh) was used for chromatography. All reactions were performed
under an atmosphere of argon by standard manifold techniques. The
solvents were carefully dried and distilled prior to use.
Palladium-Catalyzed Allylation of Aldehydes by Pincer Complex
Catalysts. A mixture of the appropriate catalyst (0.007 mmol, 5 mol
%) and aldehyde 5 (0.15 mmol) was dissolved in the corresponding
solvent: in DMF (500 µL) for catalyst 1a and in THF for 2a (500 µL)
and 3 (250 µL). Subsequently, allylstannane 4 (0.18 mmol) was added
to this mixture at room temperature. The reaction mixture was then
stirred at the allotted temperatures and times listed in Table 1. Thereafter
the reaction mixture with DMF as solvent (1a) was diluted with water
and extracted with ether, while the reaction mixtures with THF as
solvent (2a and 3) were evaporated. The crude products were purified
by silica gel column chromatography with pentane/ethyl acetate as
eluent. The NMR data obtained for products 7 and 8 are in agreement
with the corresponding literature values.28,30,31,46
Comparison of the Catalytic Activity of 2a and 3 (Figure 1). The
appropriate pincer complex (0.0075 mmol) and 5d (22.7 mg, 0.15
mmol) were dissolved in CDCl3 (500 µL) and transferred to an NMR
tube. To this mixture was added allyltributylstannane 4a (56 µL, 0.18
mmol) at 0 °C. Then the reaction mixture was heated to 25 °C and the
4. Conclusions
In this study we have shown that palladium pincer complexes
are formidable catalysts for electrophilic substitution of allyl-
stannanes. Application of this procedure eliminates the side
reactions occurring in the same catalytic transformations via
bisallylpalladium intermediates. It was shown that complex 3
with weakly coordinating trifluoroacetate counterion has a
remarkably high catalytic reactivity and stability under the
applied catalytic conditions, and therefore it can be used for
allylation of a broad variety of electrophiles. Complexes 1a and
2a are also robust catalysts; however, they are inferior to 3 in
the studied electrophilic substitution reactions.
We have identified the active catalytic intermediate of the
reaction, which is an η1-allyl-coordinated pincer complex (9)
formed by transmetalation from allylstannane and the pincer
catalyst. Theoretical calculations indicate that the electrophilic
attack proceeds with a low activation barrier on the γ-position
of the η1-allyl moiety of the active intermediate (9). The
electrophilic attack takes place via a six-membered cyclic TS
structure. The activation energy and the TS geometry are only
slightly affected by the steric interactions between the bulky
phenyl groups on the pincer ligand and the aldehyde substrate.
The regio- and stereoselectivity of the reaction can be rational-
ized on the basis of the calculated TS geometry of the reaction.
It was found that the catalyst activity is determined by the
ligand effects on the transmetalation step and on the electrophilic
substitution step of the catalytic process. Since these steps have
an opposite electron demand, PCP-type complexes are expected
to show the highest activity in electrophilic substitution reac-
tions. Furthermore, weakly coordinating counterions (such as
trifluoroacetate, acetate, and borontetrafluoride) on palladium
facilitate the transmetalation step, and therefore increase the
catalytic activity of the applied pincer complex.
1
progress of the reaction was followed by H NMR spectroscopy.
Stoichiometric Reaction of Pincer Complex 3 with Allyltrimeth-
ylstannane 4e. Pincer complex 3 (0.008 mmol) dissolved in dry THF-
d8 (600 µL) was transferred into a carefully dried NMR tube. To this
solution was added allyltrimethylstannane 4e (0.2 mmol) at -10 °C.
After homogenization, a clear yellow reaction mixture was obtained,
1
which was studied by H and 31P NMR spectroscopy at -10 °C. The
shift values for complex 9 were determined from the NMR spectra of
this reaction mixture. The only exception is the 1H NMR shift at 5.91
ppm, which overlaps with the H2 resonance of 4e. This shift value
was determined from the reaction mixture of 3 with allylmagnesium-
bromide (see below). 1H NMR (see also Figure 2a) in THF-d8: δ 7.76
(m, 8H), 7.52 (m, 12H), 6.97 (t, J ) 8.0 Hz, 1H), 6.66 (d, J ) 8.0 Hz,
2H), 5.91 (m, 1H), 4.15 (d, J ) 16.5 Hz, 1H), 3.84 (dd, J ) 10.0 and
2.2 Hz, 1H) 2.29 (dt, J ) 8.8 and 5.1 Hz, 1H). 31P NMR: δ 147.1
Stoichiometric Reaction of Pincer Complex 3 with Allylmagne-
sium Bromide. The NMR tube charged with 3 (0.01 mmol) dissolved
in THF-d8 (500 µL) was cooled to -10 °C. Thereafter, allylmagnesium
bromide in THF-d8 (200 µL, 0.25 M, 0.05 mmol) was added and the
reaction mixture was homogenized to give a yellow solution. The NMR
shift values obtained for 9 in this experiment (Figure 2b) and in the
experiment given above (Figure 2a) are identical.
Acknowledgment. This work was supported by the Swedish
Natural Science Research Council (NFR). The calculations were
done at the IBM SP2 parallel computer facility of the Paral-
lelldatorcentrum (PDC) at the Royal Institute of Technology,
Sweden. We thank the PDC for a generous allotment of
computer time.
JA049357J
(46) Jones, P.; Knochel, P. J. Org. Chem. 1999, 64, 186.
9
J. AM. CHEM. SOC. VOL. 126, NO. 22, 2004 7033