Bite angle effect of bidentate P-N ligands in palladium catalysed allylic alkylationf

Article abstract of DOI:10.1039/b001355m

Two series of new bidentate P-N ligands have been synthesized. Application of these ligands in the palladium catalysed allylic alkylation of crotyl chloride and cinnamyl chloride leads to the preferential formation of the branched product. A larger bite angle of the ligand leads to higher regioselectivity. Stoichiometric alkylation of the complex [Pd(C₄H₇){p-MeOC₆H₄C=N(CH ₂)₄OPPh₂}][O₃SCF₃] proceeds with 88% regioselectivity to the branched product. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/b001355m

View Article Online / Journal Homepage / Table of Contents for this issue
Bite angle effect of bidentate P–N ligands in palladium catalysed
allylic alkylation†
Richard J. van Haaren,a Cees J. M. Druijven,a Gino P. F. van Strijdonck,a Henk Oevering,b
Joost N. H. Reek,a Paul C. J. Kamera and Piet W. N. M. van Leeuwen*a
a Institute of Molecular Chemistry, Faculty of Chemistry, University of Amsterdam,
Nieuwe Achtergracht 166, 1018 WV, Amsterdam, The Netherlands
b DSM Research B.V., PO Box 18, 6160 MD, Geleen, The Netherlands
Received 18th February 2000, Accepted 28th March 2000
Published on the Web 28th April 2000
Two series of new bidentate P–N ligands have been synthesized. Application of these ligands in the palladium
catalysed allylic alkylation of crotyl chloride and cinnamyl chloride leads to the preferential formation of the
branched product. A larger bite angle of the ligand leads to higher regioselectivity. Stoichiometric alkylation
of the complex [Pd(C H ){p-MeOC H C᎐N(CH ) OPPh }][O SCF ] proceeds with 88% regioselectivity to the
᎐
4
7
6
4
2
4
2
3
3
branched product.
Introduction
A tremendous research effort has been devoted to the enantio-
selective alkylation of symmetrically disubstituted allylic
substrates, such as 3-acetoxy-1,3-diphenyl-1-propene and cyclo-
hex-2-enyl acetate1–3 using malonate nucleophiles. In the case
of non-symmetrically monosubstituted substrates, e.g. crotyl
acetate (but-2-enyl acetate) or cinnamyl acetate (3-phenyl-
prop-2-enyl acetate), regiocontrol is a prerequisite for enantio-
control.4 Palladium complexes have a preference for the
formation of the linear, achiral product.5 Excellent regio- and
enantio-selectivities6 have been obtained using metals other than
palladium such as iridium,6a rhodium,6b iron6c and tungsten.4a,6d
Recently it was shown that, by careful ligand design, pal-
ladium can also show a preference for the branched, chiral
product. Up to 96% regioselectivity to the formation of the
branched, chiral product was obtained for cinnamyl acetate.4b
Very high enantioselectivities have been found using phos-
phinooxazoline ligands (Pfaltz,4a,b Williams4e,f) and the MAP
ligand (R(Ϫ)-2-(dimethylamino)-2Ј-(diphenylphosphino)-1,1Ј-
binaphthyl, Kocovsky4c,g). The phosphinooxazoline ligands
contain a soft phosphorus donor atom and a relatively hard
nitrogen donor atom. It has been established that in the corre-
sponding (allyl)palladium complex the nucleophilic attack takes
place at the allylic carbon atom trans to phosphorus.1,4,7 In the
case of monosubstituted substrates it is therefore essential that
the substituted allylic carbon atom is bound trans to phos-
phorus. In the above mentioned phosphinooxazoline ligands
the steric bulk of the nitrogen donor is much smaller than that
of the phosphorus donor. A substituent on the allyl moiety will
therefore encounter less steric hindrance when it is positioned
cis to the nitrogen donor ligand.
Fig. 1 Regioselectivity in the palladium catalysed allylic alkylation.
C2–C3)–σ(to C1–C2) type complex. Since phosphorus exerts a
stronger trans influence than nitrogen, the Pd–allyl bond trans
to phosphorus is weakened. As a result, the allylic C–C bond
trans to phosphorus has more double bond character. Thus
σ(C1–C2) will be found cis to phosphorus and π(C2–C3) trans
to phosphorus. The preference for nucleophilic attack on the
allylic carbon atom trans to phosphorus is therefore caused
both by steric and electronic effects (see Fig. 1).
Apart from these steric reasons, the electronic difference
between phosphorus and nitrogen also plays
a
role.1,4,7
Although the allyl moiety remains bonded via all three carbon
atoms in a covalent manner,8 the presence of a substituent on
one of the terminal positions distorts its symmetry.9b The allylic
carbon–carbon bond next to the substituent (C3–C2) will show
more double bond character than the other allylic carbon–
carbon bond (C1–C2). As a consequence, the bonding between
palladium and the substituted allyl will be distorted to a π(to
Part of our own work in the field of allylic alkylation5 has
been concerned with the effect of the bite angle of achiral,
symmetric bidentate phosphine ligands on the regioselectiv-
ity.5a,b It was found that a larger bite angle of the ligand results
in 100% formation of the linear, non-chiral product for trans-
hex-2-enyl acetate. We now report on the remarkable, opposite
effect of the bite angle of bidentate P–N ligands on the regio-
selectivity of the alkylation of non-symmetrically substituted
† Electronic supplementary information (ESI) available: characteris-
ation data. See http://www.rsc.org/suppdata/dt/b0/b001355m/
DOI: 10.1039/b001355m
J. Chem. Soc., Dalton Trans., 2000, 1549–1554
This journal is © The Royal Society of Chemistry 2000
1549

View Article Online
Fig.
3
Possible isomeric structures of cationic (C4H7)Pd(P–N)
complexes.
Fig. 2 Generic structures of two new classes of P–N ligands.
allyl moieties. In this study the (allyl)palladium complexes bear-
ing new P–N ligands are synthesized and characterized. The
exact orientation of the substituent on the allyl group is estab-
lished by means of NMR spectroscopy. Use of the complexes
in the stoichiometric and the catalytic alkylation shows a pro-
nounced effect of the counter ion of the cationic (allyl)-
(ligand)palladium complex.
Results
Ligand synthesis
We have prepared two series of mixed phosphorus–nitrogen
ligands. One series of ligands consists of a Ph2PO unit that is
connected to an ortho substituted pyridine moiety via an alkyl
chain of variable length (P–Py ligands (a–c); the generic struc-
ture is given in Fig. 2). By changing the length of the alkyl chain
the bite angle of the ligand can be tuned. Another class of
ligands (d–g) is based on the same phosphorus unit, with an
alkyl group of variable length linked to an imine moiety (see
Fig. 2). At the para position of the imine moiety a substitu-
ent was introduced. By changing this substituent the elec-
tronic properties of the ligand can be tuned. The P–Py type
ligands (a–c) were conveniently prepared by coupling Ph2PCl to
o-NC5H4(CH2)nOH in the presence of NEt3. The P–Im class of
ligands (d–g) were synthesized in two steps. The imine was syn-
thesized according to a literature procedure10 by condensation
of the para-substituted aldehyde with H2N(CH2)nOH (n = 3
or 4), followed by coupling to Ph2PCl.
Fig. 4 Variable temperature NMR spectra of cationic (C4H7)Pd(c) at
218 (a), 273 (b) and 328 K (c), recorded in CDCl3.
the crotyl and cinnamyl complexes. When the ligand with the
smallest bite angle is used (a), only 81% of the syn-trans-P iso-
mer is formed, with the anti-trans-P complex (14%) as the other
main isomer. The remaining signals (4 and 1%) are ascribed to
the syn-trans-N and the anti-trans-N isomer.
Synthesis and structures of Pd(allyl)(P–N)X complexes
We have prepared cationic crotyl and cinnamyl palladium
complexes of these new ligands by reaction of the appropriate
ligand with the [(C4H7 or C9H9)PdCl]2 dimer,11 followed by
chloride abstraction with silver triflate. Using these P–N lig-
ands, four isomeric complexes can be formed: the substituent
can be oriented either syn or anti with respect to Hd and cis or
trans to the phosphorus atom (see Fig. 3). The value of the
coupling constant (4J(P–CH3)) is diagnostic for the orientation of
the substituent, enabling elucidation of the structure of the
complexes.5b The value of the coupling constant of the syn
oriented CH3 groups with a trans phosphorus atom is around
10–12 Hz (4J(P–CH3)). Both a cis orientation with respect to
phosphorus and an anti orientation with respect to Hd would
result in a lower value of 4J(P–CH3) (around 6 Hz). When ligand a
is used all four isomers of the (C4H7)Pd(a) complex are formed.
The syn-trans-P isomer predominates over the other isomers
(>90%). When the substituent is a phenyl rather than a methyl
group the amount of syn-trans-P isomer exceeds 97%. The
minor isomer (<3%) could not be identified by NMR. Based on
the results of allylic alkylation (see below) we assign the minor
signal to the anti-trans-P isomer.
In general, the signals in the 1H NMR spectra (at 298 K) are
relatively broad, which is indicative of a dynamic exchange
process. As it is crucial to determine the orientation of the sub-
stituent on the allyl moiety, we conducted variable temperature
NMR experiments with the cationic (C4H7)Pd(c) complex (Fig.
4). The fast exchange limit is reached at ϩ55 (c) and at Ϫ55 ЊC
(a) the slow exchange limit is almost reached.12 All signals in the
fast exchange spectrum decoalesce into two signals when going
4
to lower temperatures. At Ϫ55 ЊC the value of J(P–CH3) could
not be determined exactly, but was approximately the same as in
the fast exchange limit.
In order to gain more insight we have prepared the imine
based ligand g, that differs from ligand c only in the nature of
the nitrogen donor group. The crotyl complex of ligand g shows
the same type of fluxional behaviour as the analogous complex
of ligand c. A low temperature (Ϫ25 ЊC) spectrum recorded for
the complex of ligand g shows that in the slow exchange regime
the methyl groups of both isomers have the same coupling
constant (4J(P–CH3)). This indicates that both isomers have the
methyl group in a syn orientation. The spectra of all complexes
ligated with the imine type ligands show the presence of two
structures. All are identified as syn-trans-P complexes.
1
It was concluded from 31P and H NMR spectroscopy that
4
the bite angle of the ligand has an effect on the isomer distribu-
tion. Use of the ligand b or c, having a large bite angle, results in
almost exclusive formation (>97%) of the syn-trans-P isomer of
The value for J(P–CH3) in the fast exchange spectrum of the
complex of ligand c is the same as in the slow exchange spec-
trum of the complex of ligand g. Apart from the value for
1550
J. Chem. Soc., Dalton Trans., 2000, 1549–1554

View Article Online
Table 1 Stoichiometric alkylation of (C4H7)Pd(P–N)OTf complexes
(OTf = O3SCF3)
Table 3 Catalytic alkylation of crotyl chloride
TOF after
5 min/mol
Entry
Ligand
% Branched
% trans
% cis
Entry
Ligand
molϪ1 hϪ1
% Branched
% trans
% cis
1
2
3
4
5
6
a
b
c
d
e
f
37.2
65.6
79.0
84.1
88.0
79.8
56.1
31.0
18.2
14.2
10.9
17.9
6.7
3.4
2.8
1.7
1.1
2.3
1
2
3
4
5
6
a
b
c
d
e
f
1000
200
1000
900
300
200
33.3
41.7
55.1
52.3
50
58.6
49.6
37.8
39.0
41
8.1
8.7
7.1
8.7
9
33
57
10
Table 2 Stoichiometric alkylation of (C9H9)Pd(P–N)OTf complexes
Table 4 Catalytic alkylation of cinnamyl chloride
Entry
Ligand
% Branched
% trans
TOF after
5 min mol
1
2
3
4
5
6
a
b
c
d
e
f
43.7
67.5
77.8
75.3
78.4
81.8
56.3
32.5
22.2
24.7
21.6
18.2
Entry
Ligand
molϪ1 hϪ1
% Branched
% trans
1
2
3
4
5
6
a
b
c
d
e
f
2700
1700
2800
7300
3000
3500
29.4
26.7
56.9
22.5
31.5
13
70.6
73.3
43.1
77.5
68.5
87
4J(P–CH3), the chemical shift of the CH3 group for all com-
plexes > 1.3 ppm, is also indicative of a trans-P orientation. A
trans-N orientation of the methyl group would result in a signal
at higher field (<0.6 ppm).13e,f
Table 5 Regioselectivity in the catalytic alkylation of crotyl acetatea
Apart from the splitting of all signals, it can be seen that at
ϩ55 ЊC the allylic protons Ha and Hb are in a fast exchange. At
low temperature (Ϫ55 ЊC), the one, averaged signal of these two
protons is not split into only two, but into four signals, two for
each proton. The signals of the ortho-pyridine proton are sep-
arated from the rest of the signals and were used to obtain rate
data by simulation of the spectra.14 The Eyring plot shows that
the rate of exchange is linearly dependent on the reciprocal
temperature: Ϫln (k/T) = Ϫ33 ϩ 7 × 103 TϪ1. From this it
follows that ∆H‡ = 60 kJ molϪ1 and ∆S‡ = 34 J KϪ1 molϪ1.
Complex
Entry
(ligand/Pd)b
% Branched
% trans
% cis
1
a (1)
29.6
29.2
24.5
24.5
50.4
62.3
63.1
68.3
64.6
44.0
8.0
7.7
7.2
10.9
5.6
2b
3b
4
a (3)
a (4)
b (1)
5
b (1) 20 equivalents
LiBr/Pd
b (1.5)
b (2)
6b
7b
8
19.3
13.9
29.1
77.5
84.1
59.6
3.2
2.0
11.3
c (1)
Stoichiometric alkylation
a After 24 hours quantitative conversion was reached. Initial reaction
rates were not determined. b Extra ligand was added from a stock solu-
tion to the isolated complex.
The results of the stoichiometric alkylation of these com-
plexes with sodium diethyl 2-methylmalonate are presented in
Tables 1 and 2 (see also Fig. 1). Reaction of the crotyl com-
plex provided mainly the linear trans and the branched prod-
uct. The linear cis product is formed in only minor amounts
(up to 7%). When the bridge length and consequently the bite
angle of the ligand are larger, the regioselectivity to the
branched product of the stoichiometric alkylation increases,
up to 79% for ligand c.
reactions, the regioselectivity for the branched product
increases when the bite angle of the pyridine based ligands is
larger (entries 1–3, Tables 3 and 4). The branched:linear ratio
obtained in the catalytic reactions, however, is lower than that
found for the stoichiometric reactions.
The alkylation of allylic acetates compared to chlorides
resulted in a significant decrease of the regioselectivity towards
the branched product: from 56.9% for crotyl chloride to 24.5%
for crotyl acetate (see Table 5). Addition of extra halide (LiBr)
to the reaction mixture containing crotyl acetate restored the
regioselectivity to a value of 50.4%.
When extra ligand was added to the isolated complexes a
sharp decrease in regioselectivity was observed (Tables 5 and 6).
The addition of two extra equivalents of ligand b resulted in
the formation of only 6% of the branched product (entry 7,
Table 6). The effect is more pronounced for the ligand having
the longer bridge length (ligand b versus a). This product distri-
bution is similar to that obtained with the bidentate phosphine
ligand dppe (entry 9, Table 6).15
Kinetic experiments on the catalytic alkylation of cinnamyl
chloride using the ligand c were performed by variation of the
concentrations of palladium, diethyl 2-methylmalonate and
cinnamyl chloride. The reactions were monitored with GC. The
alkylation reaction proceeds with a zeroth order dependency
on the concentration of cinnamyl chloride, and a first order
dependency on both the malonate anion and the palladium
complex.
If the substituent on the allyl moiety is a bulkier phenyl
rather than a methyl group only two products are observed: the
linear trans and the branched product. Again the ligands with a
larger bite angle direct the regioselectivity towards the preferen-
tial formation of the branched product up to 78% (ligand c).
There is no significant change in regioselectivity when neutral
instead of cationic complexes are used (entry 4).
Tables 1 and 2 show that the regioselectivity of stoichio-
metric alkylation of the complexes ligated with imine based
ligands (d–f) is similar to that obtained using the pyridine based
ligands (a–c). For the crotyl complexes, the highest regio-
selectivity is found for methoxy substituted ligand e, which
gives 88% of the branched product, whereas for cinnamyl com-
plexes the highest regioselectivity (82%) is found using the
fluoro substituted ligand f.
Catalytic alkylation and kinetics
We have studied the catalytic alkylation of crotyl chloride
(trans-but-2-enyl chloride) and cinnamyl chloride (trans-3-
phenylprop-2-enyl chloride) using the corresponding (allyl-
palladium) complex as the catalyst. Also in the catalytic
J. Chem. Soc., Dalton Trans., 2000, 1549–1554
1551

View Article Online
Table 6 Regioselectivity in the catalytic alkylation of cinnamyl
ments show that the cationic complexes do not interconvert
between the different isomers.
acetatea
The regioselectivities found for the stoichiometric reactions
therefore indicate that the nucleophilic attack takes place trans
to the phosphorus atom, having a larger trans influence. This
observation is in agreement with a model, recently presented in
the literature, in which it is predicted that the regioselectivity for
the branched product is higher when the non-symmetry of the
allyl moiety is enhanced.9 Substitution of the imine at the para
position with an electron donating group, such as a methoxy (e)
or a dimethylamino (d) group, will increase the non-symmetry
of the allyl. The results in Tables 1 and 2 show that this is indeed
the case, both for bridge length 3 and 4. The regioselectivity
towards the branched product can be increased to a value of
88% for the methoxy substituted ligand (e).
The regioselectivity for the branched product increases when
going from a small bite angle (ligand a) to a larger bite angle
(ligand c). This is observed both for the crotyl and the cinnamyl
complexes. The regioselectivities obtained using P–N type lig-
ands may therefore be related to the non-symmetry of the allyl
moiety. The substituted allylic carbon atom C3 is more electro-
philic than C1 and becomes even more electrophilic when the
bite angle is larger.
Complex
Entry
(ligand/Pd)b
% Branched
% trans
1
a (1)
a (2)
a (3)
a (4)
b (1)
b (2)
b (3)
c (1)
dppec
33.0
22.4
22.3
5.9
21.5
7.7
5.9
39.1
4.7
67.0
77.6
77.7
94.1
78.5
92.3
94.4
60.9
95.3
2b
3b
4b
5
6b
7b
8
9
a After 24 hours quantitative conversion was reached. Initial reaction
rates were not determined. b Extra ligand was added from a stock
solution to the isolated complex. c 1,2-Bis(diphenylphosphino)ethane.
For cinnamyl complexes, the difference between the regio-
selectivity obtained with the fluoro substituted ligand f and the
methoxy substituted ligand e is smaller than for the correspond-
ing crotyl complexes. The regioselectivity obtained using ligand
f is slightly higher than that obtained using ligand e. Since these
observations cannot be explained in terms of electronic non-
symmetry of the allyl moiety only, it is concluded that in this
case either steric or other secondary interactions, such as π–π
stacking, can play a role.16
Fig. 5 Dynamic exchange of endo and exo isomers in cationic
(C4H7)Pd(P–N) complexes via π–σ–π rearrangement.
Discussion
Structures of the complexes
The 1H NMR spectra showed that all complexes exhibit
dynamic behaviour at room temperature. Variable temperature
NMR spectroscopy of the complex bearing ligand c showed
that the observed exchange process does not involve the syn-
trans-P and one of the other isomers of the types syn-trans-N,
anti-trans-P or anti-trans-N. Surprisingly, the two different iso-
mers that were observed under the slow exchange conditions
are both identified as a syn-trans-P isomer. The two isomers are
proposed to be conformers differing in the orientation of the
ligand backbone, which can be either up (endo) or down (exo)
with respect to the orientation of the allyl moiety.
The protons Ha and Hb (see Fig. 1) are in fast exchange at
ϩ55 ЊC, via the π–σ–π rearrangement13 of the allyl moiety, one
averaged signal being observed. This one signal is split into two
signals for each proton at Ϫ55 ЊC. The fluxional behaviour is
thus caused by a selective π–σ–π isomerization, during which
the Pd–allyl bond trans to phosphorus is broken. The syn-trans-
P structure is retained during this rearrangement, but the orien-
tation has changed from endo to exo or vice versa (see Fig. 5).
Thus, the dynamic exchange between the endo and the exo form
of the complex is a result of the π–σ–π rearrangement of the
allyl moiety and not of other processes such as flipping of the
backbone.
Catalytic alkylation, kinetics
The regioselectivity found for the catalytic reactions is lower
than that found for the stoichiometric reactions. This can be
explained as follows. After oxidative addition of the substrate
two types of isomeric palladium complexes can be formed:
one type having the substituent oriented trans to the phos-
phorus atom and one type having the substituent oriented
trans to the nitrogen atom. The syn-trans-P isomer is thermo-
dynamically favoured and the syn-trans-N will isomerize to the
more stable syn-trans-P isomer. This isomerization is facili-
tated by strongly co-ordinating counter ions, such as chloride,
which is the leaving group in the catalytic reactions (Tables 3
and 4).13
If the subsequent alkylation step is slow relative to isomeriz-
ation (trans-N to trans-P), the product distribution will be
determined by the syn-trans-P isomer only and therefore will be
equal to the regioselectivity in the stoichiometric alkylation. If
the alkylation, however, occurs at a rate similar to or faster than
the rate of isomerization, the observed regioselectivity is deter-
mined by the isomer ratio formed, their rate of isomerization
and their respective rates of alkylation. Since nucleophilic
attack will primarily take place at the carbon atom trans to
phosphorus, alkylation of the syn-trans-N isomer will yield
relatively large amounts of the linear trans product. This
explanation in terms of competition between isomerization and
alkylation is supported by the results of the catalytic alkylation
of both crotyl and cinnamyl acetate. The acetate anion co-
ordinates not as strongly to the palladium centre as the chloride
anion. It was shown that a co-ordinating anion facilitates
dynamic behaviour of the allyl moiety.13d,e It has indeed been
found that the regioselectivity is lower when acetate is the leav-
ing group (and therefore also the counter ion (Table 6)). The
influence of a co-ordinating anion is further shown by addition
of LiBr to the crotyl acetate reaction mixture (BrϪ replaces
OAcϪ: entry 5, Table 5). In this case the regioselectivity is
similar to that obtained using crotyl chloride.
The partial deco-ordination of the allylic moiety during this
process accounts for the observed positive ∆S‡ value. The
values of the activation parameters ∆H‡ and ∆S‡ correspond
to literature values for the π–σ–π rearrangement.13e
Stoichiometric alkylation
As expected, the syn-trans-P isomeric form of the cationic
(C4H7 or C9H9)Pd(P–N) complexes is the main isomer present
in solution. The NMR spectra show that for a small methyl
substituent there is 3% of the anti-trans-P isomer present. The
study of the dynamic behaviour shows that there is no isomeriz-
ation from the syn to the anti isomer on the NMR timescale.
The 3% linear cis product that is found (Table 1) corresponds
with the 3% anti-trans-P isomer that is present. When the sub-
stituent is a larger phenyl group only the syn-trans-P isomer is
observed. This is also reflected in the absence of the linear cis
product in the alkylation reactions (Table 2). VT-NMR experi-
1552
J. Chem. Soc., Dalton Trans., 2000, 1549–1554

View Article Online
General synthetic procedures
Sodium diethyl 2-methylmalonate (0.5 M in THF) was pre-
pared from diethyl 2-methylmalonate and NaH in THF at
273 K. The dimers [Pd(C4H7)(µ-Cl)]2 and [Pd(C9H9)(µ-Cl)]2
were prepared following a literature procedure.11 The synthesis
and characterization of [Pd(dppe)(C4H7)]OTf,5b crotyl acetate5b
and the alkylation products of the coupling of crotyl acetate to
sodium diethyl 2-methylmalonate are described elsewhere.13e
The P–Py type ligands (a–c) were prepared via condensation
of 2-pyridine-methanol (P–Py1 (a)), -ethanol (P–Py2 (b)) and
-propanol (P–Py3 (c)) with Ph2PCl in the presence of an excess
of NEt3. The alcohol and NEt3 were dissolved in diethyl ether
and cooled to 0 ЊC, after which a solution of Ph2PCl in diethyl
ether was added dropwise. A white precipitate was formed
(NEt3ؒHCl). After filtration the reaction volume was concen-
trated under reduced pressure. Flash chromatography (silica)
of this residue yielded the ligand as a colourless oil, in a yield of
around 70%, based on Ph2PCl.
Fig. 6 Proposed formation of bis-phosphine co-ordinated complexes
by addition of extra ligand to cationic (C4H7 or C9H9)Pd(P–N) com-
plexes.
The catalytic intermediates used as precursors in the catalytic
experiments contain only one ligand per palladium atom. In
Tables 5 and 6 it is shown that the addition of extra ligand
to the isolated complex causes a significant change in regio-
selectivity. The resulting regioselectivity is similar to that
obtained when bidentate phosphine ligands are used (entry 9,
Table 6). When extra ligand is present it is plausible that two
ligands are co-ordinated to palladium, both via the phosphorus
atom (Fig. 6). The system then behaves as a bidentate phos-
phine ligand.
The P–Im type ligands (d–g) were synthesized using a similar
procedure, but required as an extra step the synthesis of the
imine alcohol.10 In a typical procedure, the aldehyde was
dissolved in toluene and the α- to ω-aminoalcohol was added in
one portion. The equilibrium of this condensation reaction was
directed towards the imine alcohol by removing the water
formed with dehydrated K2CO3. After stirring for 12 hours the
reaction was complete and after filtration the solvent was
removed by evaporation. The alcohol was then coupled to
Ph2PCl using the same procedure as for the P–Py type ligands.
As a side reaction, cyclization of the imine took place.10 This
could only be prevented by the presence of an excess of the
aldehyde. It should be noted that especially aldehydes bearing
electron withdrawing substituents required this excess. The
aminoalcohol could be quantitatively converted into the corre-
sponding imine. The mixture of imine and aldehyde (colourless
oil) was used as such in the coupling to Ph2PCl. The aldehyde–
ligand mixture (colourless oil) was then used in the synthesis of
the corresponding (allyl)palladium complexes.
The oily ligand (1.0 equivalent) was weighed in a Schlenk
Vessel and dissolved in 10 mL of CH2Cl2. To this mixture a
solution of exactly 0.5 equivalent of the [Pd(C4H7)(µ-Cl)]2 or
[Pd(C9H9)(µ-Cl)]2 dimer was added (dissolved in 10 mL of
CH2Cl2). The solution changed from colourless to bright yellow
instantaneously. After stirring for 15 minutes, exactly 1.0
equivalent of AgOTf was added. Immediately, a white precipi-
tate was formed and the solution changed to light yellow. After
filtration over Celite, the solvent was removed in vacuo. After
this step, the excess of aldehyde could be removed by repetitive
washing with either pentane or benzene. All palladium com-
plexes were isolated as a white microcrystalline powder in ca.
90% yield based on palladium and were used in the alkylation
reaction.
Kinetic experiments on the alkylation of cinnamyl chloride,
using (C9H9)Pd(c)OTf as the catalyst, show that the rate of
reaction is independent of the cinnamyl chloride concentration
and that it is linearly dependent on the concentration of both
the malonate anion and the palladium complex. This indicates
that the nucleophilic attack is the rate determining step of the
reaction.
Although palladium allyl complexes have been studied in
great detail, little is known about the influence of the properties
of ion pairs, which seem to be important for this chemistry. It
has been reported that solvent polarity can have an effect on the
outcome of enantioselective alkylation reactions.2h Currently,
we are studying the effect of polarity effects on the regio-
selectivity of the alkylation reaction.5d
Conclusion
In conclusion, we have shown that mixed bidentate P–N ligands
with a large bite angle direct the regioselectivity to the form-
ation of the branched product. Since the nitrogen donor atom is
incorporated in a small pyridine group the effect is electronic in
nature. This is in contrast with our previous results concerning
the effect of the bite angle of bidentate phosphine ligands,
which could be explained in terms of steric hindrance.5a,b Thus,
the effect of a larger bite angle on the regioselectivity has a
steric component (leading to more linear product) and an
electronic component (leading to more branched product).
Therefore we conclude that for a rational design of ligands
that favour the formation of the branched product the following
parameters are of importance: (1) relative donor–acceptor
strength of the ligand donor atoms; (2) steric hindrance in the
transition state; (3) bite angle of the ligand.
Alkylation reactions
Experimental
1H (300 MHz, TMS, CDCl3), 31P-{1H} (121.5 MHz external
85% H3PO4, CDCl3) and 13C NMR (75.4 MHz, TMS, CDCl3)
were recorded on a Bruker AMX-300 spectrometer. Elemental
analyses were performed on an Elementar Vario EL apparatus
(Foss Electric). The product distribution of the alkylation
experiments was measured on an Interscience Mega2 apparatus,
equipped with a DB1 column, length 30 m, inner diameter 0.32
mm, film thickness 3.0 µm, and a flame ionization detector.
All experiments were carried out using standard Schlenk
techniques. All solvents were freshly distilled prior to use.
Crotyl chloride, cinnamyl chloride, benzaldehyde, 4-fluoro-
benzaldehyde, 4-dimethylaminobenzaldehyde, 4-methoxy-
benzaldehyde, 3-aminopropanol, 4-aminobutanol, diethyl
2-methylmalonate, NaH, AgOTf and PdCl2 were obtained
from Aldrich. Crotyl chloride and the aldehydes were distilled
prior to use.
The stoichiometric alkylation reactions were performed at
room temperature (292 K) by adding an excess of sodium
diethyl 2-methylmalonate (0.1 ml of a 0.5 M solution in THF)
to a solution of 10 mg of the palladium complex in 1 ml of
THF. Reaction was instantaneous and after one minute the
mixture was worked up with water, filtered over silica and
analysed by GC.
The catalytic reactions were performed at room temperature
(292 K) in THF (5 mL), using 0.05 mol% of catalyst (0.25
µmol), 0.5 mmol of substrate and 1.0 mmol of sodium diethyl
2-methylmalonate. The reaction was monitored by taking
samples from the reaction mixture which, after aqueous work-
up, were analysed by GC using decane as the internal standard.
The kinetic experiments were performed using stock solu-
tions of all reaction components. The amount of each reagent
was, one at a time, varied by changing the amount of stock
J. Chem. Soc., Dalton Trans., 2000, 1549–1554
1553

View Article Online
P. S. Pregosin and R. Salzmann, J. Am. Chem. Soc., 1996, 118, 1031;
solution added. The amount of catalyst was varied from
0.000625 to 0.0200 mmol, the amount of cinnamyl chloride
from 0.125 to 1.000 mmol and the amount of malonate from
0.200 to 2.000 mmol.
K. Selvakumar, M. Valentini, M. Wörle, P. S. Pregosin and
A. Albinati, Organometallics, 1999, 18, 1207; G. Helmchen,
J. Organomet. Chem., 1999, 576, 203; A. Albinati, P. S. Pregosin and
K. Wick, Organometallics, 1996, 15, 2419.
NMR data were obtained in CDCl3 (δ in ppm), IR recorded
in CH2Cl2. Complete analytical data of the compounds pre-
sented in this paper are available as supporting information.†
Pd(P–Py1 (a))(C4H7)[SO3CF3]: δH (syn-trans-P isomer) 1.83
(d, J1 = 6.2, J2 = 10.6, 3 H (Me)); 3.21 (hump, 2 H (Ha and
Hb)); 5.05 (m, 1 H (Hc)); 5.16 (d, J = 21.9 Hz, 2 H (POCH2));
5.75 (dt, J1 = J2 = 9.4, J = 13.2, 1 H (Hd)); 7.5 (m, 10 H (Ar));
7.63 (t, J = 7.4, 1 H (m-H of pyridine) (NCHCH)); 7.65 (d,
J = 7.4, 1 H (m-H of pyridine) (NCCH)); 7.95 (t, J = 7.4, 1 H
(p-H of pyridine); and 8.81 (d, J = 5.3 Hz, o-H of pyridine); δC
17.7, 50.6, 73.1, 103.7, 119.0, 121.7, 123.3, 126.8, 127.1, 129.3,
129.5, 131.8, 132.6, 134.3, 140.7, 154.5 and 155.0; δP {1H} (syn-
trans-P isomer) 128.3 (s, 1P); other isomers appear at 132.7 (s,
4 (a) R. Pretot, G. C. Lloyd-Jones and A. Pfaltz, Pure Appl. Chem.,
1998, 70, 1035; (b) R. Pretot and A. Pfaltz, Angew. Chem., Int. Ed.,
1998, 37, 323; (c) P. Kocovsky, S. Vyskocil, I. Cisarova, J. Sejbal,
I. Tislerova, M. Smrcina, G. C. Lloyd-Jones, S. C. Stephen, C. P.
Butts, M. Murray and V. Langer, J. Am. Chem. Soc., 1999, 121,
7714; (d) K. Selvakumar, M. Valentini, P. S. Pregosin and A.
Albinati, Organometallics, 1999, 18, 4591; (e) C. J. Martin, D. J.
Rawson and J. M. J. Williams, Tetrahedron: Asymm., 1998, 9, 3723;
( f ) J. F. Bower, R. Jumnah, A. C. Williams and J. M. J. Williams,
J. Chem. Soc., Perkin Trans. 1, 1997, 1411; (g) S. Vyskocil,
M. Smrcina, V. Hanus, M. Polasek and P. Kocovsky, J. Org. Chem.,
1998, 63, 7738; (h) H. Steinhagen, M. Reggelin and G. Helmchen,
Angew. Chem., Int. Ed. Engl., 1997, 36, 2108.
5 (a) M. Kranenburg, P. C. J. Kamer and P. W. N. M. Van Leeuwen,
Eur. J. Inorg. Chem., 1998, 1, 25; (b) R. J. Van Haaren, H. Oevering,
B. B. Coussens, G. P. F. Strijdonck, J. N. H. Reek, P. C. J. Kamer and
P. W. N. M. Van Leeuwen, Eur. J. Inorg. Chem., 1999, 1237; (c) D. De
Groot, E. G. Eggeling, J. C. De Wilde, H. Kooijman, R. J. Van
Haaren, A. W. Van Der Made, A. L. Spek, D. Vogt, J. N. H. Reek,
P. C. J. Kamer and P. W. N. M. Van Leeuwen, Chem. Commun.,
1999, 1623; (d) G. E. Oosterom, R. J. Van Haaren, J. N. H. Reek,
P. C. J. Kamer and P. W. N. M. Van Leeuewen, Chem. Commun.,
1999, 1119.
0.04P), 130.1 (s, 0.03P) and 129.7 (s, 0.14P); IR (νmax
/cmϪ1)
˜
3058, 2991, 2923, 1607 and 1438; FAB-MS m/z = 454.0550
(C22H23NOPPdϩ requires 454.0552); Found C, 45.35; H, 3.84;
Calc. for C22H23NOPPdϩCF3SO3 ϩ 0.1CH2Cl2 C, 45.31; H,
3.82%.
Acknowledgements
6 (a) R. Takeuchi and M. Kashio, J. Am. Chem. Soc., 1998, 120, 8647
and references therein; (b) P. A. Evans and J. D. Nelson, J. Am.
Chem. Soc., 1998, 120, 5581; (c) Y. Xu and B. Zhou, J. Org. Chem.,
1987, 52, 974; (d) G. C. Lloyd-Jones and A. Pfaltz, Angew. Chem.,
Int. Ed. Engl., 1995, 34, 462.
This research was carried out with financial support from
DSM Research B.V. and with a subsidy from the Ministerie
van Onderwijs, Cultuur en Wetenschappen as part of the
E.E.T. program for clean chemistry.
7 T. Suzuki and H. Fujimoto, Inorg. Chem., 1999, 38, 370; P. E.
Blöchl and A. Togni, Organometallics, 1996, 15, 4125; T. R. Ward,
Organometallics, 1996, 15, 2836.
8 K. J. Szabo, Organometallics, 1996, 15, 1128.
References
9 J. D. Oslob, B. Åkermark, P. Helquist and P.-O. Norrby,
Organometallics, 1997, 16, 3015; V. Branchadell, M. Moreno-Manas,
F. Pajuelo and R. Pleixats, Organometallics, 1999, 18, 4934.
10 E. D. Bergmann and A. Kaluszyner, Recl. Trav. Chim. Pays-Bas,
1959, 78, 315; J. M. Lerestif, J. Perrocheau, F. Tonnard, J. P.
Bazureau and J. Hamelin, Tetrahedron, 1995, 51, 6757.
1 G. Knuhl, P. Sennhenn and G. Helmchen, Chem. Commun., 1995,
1845; S. Kudis and G. Helmchen, Angew. Chem., Int. Ed., 1998,
37, 3047; H. Rieck and G. Helmchen, Angew. Chem., Int. Ed.
Engl., 1995, 34, 2687; G. Helmchen, S. Kudis, P. Sennhenn and
H. Steinhagen, Pure Appl. Chem., 1997, 69, 513; B. Wiese and
G. Helmchen, Tetrahedron Lett., 1998, 39, 5727; M. Gomez,
S. Jansat, G. Muller, D. Panyella, P. W. N. M. Van Leeuwen, P. C. J.
Kamer, K. Goubitz and J. Fraanje, Organometallics, 1999, 18, 4970;
S. R. Gilbertson and D. Xie, Angew. Chem., Int. Ed., 1999, 38, 2750.
2 (a) B. M. Trost and D. L. Van Vranken, Chem. Rev., 1996, 96,
395; (b) J.-C. Galland, S. Roland, J. Malpart, M. Savignac and
J.-P. Genet, Eur. J. Inorg. Chem., 1999, 621; (c) H. Brunner,
I. Deml, W. Dirnberger, K.-P. Ittner, W. Reisser and M.
Zimmermann, Eur. J. Inorg. Chem., 1999, 51; (d) R. Kuwano
and Y. Ito, J. Am. Chem. Soc., 1999, 121, 3236; (e) M. Yamaguchi,
K. Ohba, H. Tomonaga and T. Yamagishi, J. Mol. Catal. A-Chem.,
1999, 140, 255; ( f ) K. Yonehara, T. Hashizume, K. Mori, K. Ohe
and S. Uemara, Chem. Commun., 1999, 415; (g) S.-L. You,
Y.-G. Zhou, X.-L. Hou and L.-X. Da, Chem. Commun., 1998, 2765;
(h) B. M. Trost and R. C. Bunt, J. Am. Chem. Soc., 1998, 120, 70;
(i) B. M. Trost and J. D. Oslob, J. Am. Chem. Soc., 1999, 121, 3057;
(j) J. Tsuji, H. Takahashi and M. Morikawa, Tetrahedron Lett.,
1965, 4387; (k) J. Tsuji, Tetrahedron, 1986, 42, 4361; (l) P. Dierkes,
S. Ramdeehul, L. Barloy, A. De Cian, J. Fischer, P. C. J. Kamer,
P. W. N. M. Van Leeuwen and J. A. Osborn, Angew. Chem., Int. Ed.,
1998, 37, 3116.
11 W. D. Dent, R. Long and A. J. Wilkinson, J. Chem. Soc., 1964, 1585.
12 The use of other solvents with lower boiling point did not improve
the results, either because the energy barrier for the exchange
process was even lower than in CDCl3 or because of very poor
solubility.
13 (a) R. Kuwano and Y. Ito, J. Am. Chem. Soc., 1999, 121, 3236;
(b) B. Åkermark, G. Åkermark, L. Hegedus and K. Zetterberger,
J. Am. Chem. Soc., 1981, 103, 3037; (c) B. M. Trost and J. D. Oslob,
J. Am. Chem. Soc., 1999, 121, 3057; (d) K. Vrieze and P. W. N. M.
Van Leeuwen, Dynamic Nuclear Magnetic Resonance Spectroscopy,
Academic Press, New York, 1975; (e) M. P. T. Sjögren, S. Hansson,
B. Åkermark and A. Vitagliano, Organometallics, 1994, 13, 1963;
( f ) S. Hansson, P.-O. Norrby, M. P. T. Sjögren, B. Åkermark,
M. E. Cucciolito, F. Giordano and A. Vitagliano, Organometallics,
1993, 12, 4940.
14 The NMR data were compared to simulated spectra in 16 steps from
218 to 338 K; coalescence of the ortho-pyridine protons occurred at
251 K, R2 = 0.981. Simulation of the spectra was performed using
software by P. H. M. Budzelaar, gNMR version 3.5 M, Ivorysoft,
Amerbos, Amsterdam, 1995.
15 The alkylation of cinnamylpalladium complexes bearing other
bidentate ligands resulted in similar regioselectivities.
3 J.-M. Brunel, T. Constantieux, A. Labande, F. Lubatti and
G. Buono, Tetrahedron Lett., 1997, 38, 5971; K. Hiroi, Y. Suzuki
and I. Abe, Chem. Lett., 1999, 149; B. Bartels and G. Helmchen,
Chem. Commun., 1999, 741; A. Togni, U. Burckhardt, V. Gramlich,
16 Molecular modelling (Spartan PM3(tm) method) shows that the
phenyl substituent on the allyl moiety is oriented parallel to
the imine functionality.
1554
J. Chem. Soc., Dalton Trans., 2000, 1549–1554