A comparison of (R,R)-Me-DUPHOS and (R,R)-DUPHOS-iPr ligands in the Pd 0-catalysed asymmetric allylic alkylation reaction: Stereochemical and kinetic considerations

Article abstract of DOI:10.1002/ejoc.200900649

A relatively unexploited commercial ligand, (R,R)-DUPHOSiPr (1b) was tested in the Pd⁰-catalysed asymmetric allylic alkyation reaction using rac-1,3-diphenylpropenyl acetate as substrate, malonate as nucleophile and a variety of Pd precatalysts

Full text of DOI:10.1002/ejoc.200900649

FULL PAPER
DOI: 10.1002/ejoc.200900649
A Comparison of (R,R)-Me-DUPHOS and (R,R)-DUPHOS-iPr Ligands in
0
the Pd -Catalysed Asymmetric Allylic Alkylation Reaction: Stereochemical and
Kinetic Considerations
[a]
[a]
[b]
Vanda Raquel Marinho, J. P. Prates Ramalho, Ana I. Rodrigues, and
Anthony J. Burke*[
a]
Keywords: Asymmetric catalysis / Alkylation / Palladium / P ligands / Fukui function / Kinetics
A relatively unexploited commercial ligand, (R,R)-DUPHOS-
iPr (1b) was tested in the Pd -catalysed asymmetric allylic
emperical computational study was undertaken. DFT calcu-
lations of the Fukui Function (FF) were also conducted to
confirm the most likely site (electrophilic) for nucleophilic at-
tack on the active Pd-allyl complex. In the case of the Pd
catalyst formed with 1b. This result indicated that stereo-
chemical factors were more important than electronic factors
in this reaction with 1b. Kinetic studies were also conducted
to compare the activities of the catalysts 2a and 2b formed
with (R,R)-Me-DUPHOS (1a) and (R,R)-DUPHOS-iPr (1b).
0
alkyation reaction using rac-1,3-diphenylpropenyl acetate as
substrate, malonate as nucleophile and a variety of Pd pre-
catalysts under standard conditions. Excellent ee values (up
to Ն 98%) could be obtained with 1b, but the conversions
were generally inferior to those obtained using (R,R)-Me-
DUPHOS (1a) under similar conditions. It was also observed
that there was a switch in the absolute configuration of the
malonate product to (R) on using 1b to form the catalyst. This
was an indication of the complementary nature of these two
ligands. In order to explain this, a rigorous detailed semi-
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2009)
WalPhos[6] – which has been so successful in catalytic asym-
Introduction
[
7,8]
metric hydrogenation
– shows potential in other catalytic
At the present time asymmetric synthesis finds consider- asymmetric reactions like, the AAA reaction, ketone hydro-
able application in providing useful enantiomerically pure silylation and vinylarene hydroboration.
compounds. Since 1977 the asymmetric allylic alkylation
We became interested in accessing (R,R)-DUPHOS-iPr
AAA) reaction has become a standard approach for the (1b)[ in the AAA reaction of both rac-1,3-diphenylpro-
9]
(
[1,2]
creation of both C–C and C–X bonds.
Many chiral li- penyl acetate and rac-1-acetoxycyclohexene with malonate
gands have been screened for this particular reaction, these nucleophile (Figure 1). Our rationale was based on the sup-
generally contain P or N, possess C or C symmetry and position that the presence of bulky groups around the P
2
1
can be of a homobidentate type (e.g., diphosphanes, bisoxaz- atoms should give high enantioselectivities. It must be noted
[
10]
olines, etc.) or of a heterobidentate nature (e.g. phosphanyl- that, Drago and Pregosin have already demonstrated (as
[
1]
oxazolines etc). Several useful chiral diphosphane ligands a one-off study) an ee as high as 98% for the alkylation of
are known for catalytic asymmetric hydrogenations, like, rac-1,3-diphenylpropenyl acetate with dinuclear [PdCl-
[
3]
[3]
[4]
chiraphos, DIPAMP, DIOP and many more. Unfortu- (PhCHCHCHPh)] , malonate as nucleophile and (R,R)-
2
[
11]
nately these “borrowed” ligands are less successful in other Me-DUPHOS and Malaisé et al. showed a similar result,
catalytic reactions. For this reason we have established a but only when the sodium salt of dimethyl maloante was
research program to screen key phosphane and other li- used [as they claimed that BSA was not an appropriate base
gands (with a proven track record) for effectiveness in other for reactions involving (R,R)-Me-DUPHOS]. In the case of
key asymmetric catalytic processes. For example, we have (R)-Josiphos it was shown by Togni and Spindler that ee
recently demonstrated that Berens’ ligand[
4,5]
and values of up to 93% could be achieved for this reaction.
[12]
[
[
a] Departamento de Química and Centro de Química de Évora,
Universidade de Évora,
Rua Romão Romalho 59, 7000 Évora, Portugal
Fax: +351-266745303
E-mail: ajb@dquim.uevora.pt
b] Departamento de Tecnologia de Indústrias Químicas, Instituto
Nacional de Engenharia, Tecnologia e Inovação,
Estrada do Paço do Lumiar, Edifício F, 1649-038 Lisboa, Por-
tugal
Figure 1. (R,R)-Me-DUPHOS (1a) and (R,R)-DUPHOS-iPr (1b).
Eur. J. Org. Chem. 2009, 6311–6317
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
6311

V. R. Marinho, J. P. P. Ramalho, A. I. Rodrigues, A. J. Burke
Support for this conjecture comes from the observation
FULL PAPER
DUPHOS-iPr (1b) has previously been screened in the
AAA reaction of ethyl 2-oxocyclohexanecarboxylate with that when the reaction was performed according to the con-
[
11]
cinnamyl acetate, but the authors claimed unsatisfactory re- ditions of Malaisé et al. (Table 1, entry 3) a poorer con-
sults for the reaction.[
In this paper we disclose our results on the use of ligand yield reported by these workers using Me-DUPHOS.
b in the AAA reaction, comparative kinetics between the Some interesting temperature effects were observed. In
catalysts bearing 1a and 1b and computational studies to the case of the reactions using [Pd(allyl)Cl] when the tem-
13]
version of 60% was obtained as opposed to the quantitative
1
2
explain the switch in configuration using 1b.
perature was increased to 40 °C there was a substantial re-
duction in the reaction conversion and ee (Table 1, entry 2)
and we assume that this is due to chiral catalyst degradation
at the higher temperature. The same was observed on using
Results and Discussion
Pd(dba) as the conversion dropped to 2% and the ee to
2
(R,R)-DUPHOS-iPr (1b) was tested in the AAA reaction
6
9
2% as opposed to a conversion of 12% and an ee of Ն
8% (Table 1, entry 4). However, in the case of Pd (dba) ·
with 1,3-diphenylpropenyl acetate using dimethyl malonate
with a variety of Pd pre-catalysts (Scheme 1) (the active chi-
ral catalyst was thus prepared in situ) giving the following
results (Table 1).
2
3
CHCl when the temperature was raised, the conversion
3
doubled although the ee remained constant (Table 1 entries
6
and 7). In terms of the effect exerted by the solvents in the
reactions using [Pd(allyl)Cl] and Pd (dba) pre-catalysts, it
2
2
2
would appear that when BSA/KOAc are used, it was the
polar non-coordinating solvents CH Cl and ClCH CH Cl
2
2
2
2
which gave the highest ee values (Table 1, entries 1, 4 and
5) when compared to the polar coordinating solvent, THF
and acetonitrile which gave low conversions and ee values.
However, in one case when THF was used with NaH as
base an ee of Ն 98% was obtained (Table 1, entry 3).
In all cases the (R)-enantiomer of the alkylated malonate
product was the major isomer. A surprising result and con-
Scheme 1. Palladium-catalysed asymmetric allylic alkylation of rac-
1,3-diphenylpropenyl acetate with dimethyl malonate.
Table 1. Asymmetric allylic alkylation of rac-1,3-diphenylpropenyl
acetate using ligand 1b.[
a]
Conv.[b] (%) ee[c] (%)
Entry Pre-catalyst
[Pd(allyl)Cl]
[Pd(allyl)Cl]
[Pd(allyl)Cl]
Pd(dba)
Pd(dba)
Pd (dba)
Pd (dba)
Pd (dba)
Solvent
CH Cl
CH Cl
THF
CH Cl
ClCH
CH Cl
CH
CH
1
2
2
2
25
6
94 (R)
53(R)
Ն 98% (R)
Ն 98% (R)
Ն 98% (R)
81 (R)
[
10,11]
[
d]
trary to the results obtained using 1a
where the (S)-
2
3
4
5
6
7
8
2
2
2
[
d,e]
malonate product was the major enantiomer. Drago and
2
60
12
21
11
23
58
[
10]
2
2
2
Pregosin have previously shown on the basis of extensive
2
2
CH
2
Cl
1
H NMR studies on the isolated Pd-allyl complex with an
2
3
·CHCl
·CHCl
·CHCl
3
2
2
2
2
exo-conformation that the malonate nucleophile attacks
preferentially C-1 as this is the most electrophilic of the two
terminal allyl carbons. This was postulated to be due to the
steric hindrance imparted by the ligand methyl group. Due
to facial selectivity in this case the (S)-malonate enantiomer
[
d]
2
3
3
2
Cl
Cl
82 (R)
[
f]
2
3
3
2
Ն 98% (R)
[
a] Reaction conditions: 2.5 mol-% of ligand and pre-catalyst 1 mol-%
were used, complexation conducted at reflux temperature for 2 h and
then room temp. for 20 h (base: BSA/KOAc, 3 equiv. BSA and 1 mol-
%
KOAc was used). [b] Pertains to the ratio of substrate to product as is preferred. We prepared the triflate salt of the Pd complex
determined by HPLC. [c] Determined by HPLC using a Chiralcel OD-
H column. [d] Conducted at reflux temperature. [e] 1.2 mol-% of ligand
and pre-catalyst 0.5 mol-% were used, complexation conducted at reflux
temperature for 2 h and then room temp. for 20 h, base: NaH. [f]
2
c (Figure 1) by a method similar to that described by Pre-
[10]
gosin and Drago (in fact we had problems making it by
their method), but despite our best efforts we have been
2
.5 mol-% of ligand and pre-catalyst 10 mol-% were used, complexation unable to grow suitable crystals of 2c for single-crystal X-
conducted at reflux temperature for 2 h and then room temp. for 1 h.
ray crystallographic analysis. However, we did indeed ob-
tain a 31P NMR spectrum of 2c which had much the same
characteristics as the triflate complex of 2a reported by Pre-
Despite the generally very good enantioselectivities (with
a maximum of Ն 98%) obtained, in all cases the substrate
conversions were moderate (with a maximum of 60%, see
Table 1). We assume that this may be linked to the consider-
able steric hindrance inherent in the active allylpalladium
complex 2b derived from rac-1,3-diphenylpropenyl acetate,
or in forming the intermediate π-allyl palladium complex
that is attacked by malonate (see Figure 2).
[
10]
gosin and Drago
31 Hz for P1 and another doublet at 62.55 ppm with J =
1 Hz for P2). This thus prompted us to carry out a very
(one doublet at δ = 64.35 ppm with J
=
3
detailed computational study to understand the nature of
the configurational switch of the malonate product on go-
ing from 1a to 1b.
In our hands the results using rac-1-acetoxycyclohexene
as substrate in this reaction were very poor, in fact, moder-
ate conversions and no enantioselectivity was obtained.
Theoretical Studies
Considering the number and the size of these systems,
a semi-empirical method proved to be the best choice for
Figure 2. Pd-allyl complexes 2.
6312
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 6311–6317

0
Pd -Catalysed Asymmetric Allylic Alkylation Reaction
[
14]
geometry optimization. The PM6 Hamiltonian included
Another approach is the method proposed by Contreras
in the recent version of MOPAC 2007[ was employed. The et al.
15]
[22]
in the framework of frontier orbital theory. This
geometries were fully optimized in Cartesian coordinates approach involves a single calculation with the FF at an
and the stationary points were further characterised by fre- atom k being given by
quency calculations as minima. Following the initial geome-
try optimization with the PM6 method, Density Functional fkα
Theory (DFT) calculations were carried out on all com-
plexes. Calculations carried out at the DFT level were per-
fµα
=
͚
෈
k
(3)
(4)
µ
[
16]
having
formed using the hybrid Becke exchange functional and
[17]
the correlation functional B3LYP
Gamess package.
as contained in the
The Lanl2DZ effective core basis set
was employed for the metal and the P atom while the 6-
[
18]
α
= |cµα|2 + cµα
f
µ
͚
να µν
c S
ν϶µ
3
1G* basis set was used for all the other atoms.
There was a significant degree of approximation between
where c is the frontier molecular orbital coefficients and
µν
Sµν is the overlap integral between the atomic basis func-
tions χ (r) and χ (r). Equation (3) gives the condensed
the calculated values for the Pd–C1 and Pd–C3 bond
lengths (2.253 and 2.204 Å), with those measured in the X-
ray crystal structure obtained by Pregosin and Drago,
µ
ν
Fukui functions for electrophylic (α = –) and for nucleo-
philic (α = +) attacks, whilst an average is considered for
radical attacks.
[
10]
which were 2.238 and 2.221 Å, respectively. The good agree-
ment seen here supports the suitability of the PM6 method
to determine molecular geometries for this family of com-
plexes.
After completing this step, the next step was the calcula-
tion of the Fukui Function (FF) on the Pd-allyl complexes
of 2a and 2b (without the chloride ion) to establish the most
likely site of nucleophilic attack. The Fukui function is one
of the more common descriptors widely used to predict rel-
ative site reactivities[ and in fact has been used to provide
a more solid theoretical base for the local HSAB prin-
We were particularly interested in clarifying where the
preferential site of attack would occur in this alkylation re-
action and thus have calculated the FF values for this mole-
cule knowing that nucleophilic addition will preferentially
+
occur at the carbon atom with the highest fk value.
Both the Pd-allyl complex 2a (Figure 3) and 2b (Fig-
ure 4) were studied in the gas phase using this technique. In
both cases, it was calculated that the C-1 site had the largest
19]
+
fk value and thus the greatest susceptibility for nucleo-
philic attack. This was fine in the case of 2a where it has
[
20]
ciple.
The Fukui function has been defined as the elec-
been experimentally shown that the preferential site of at-
tron density derivative with respect to the electron number,
at constant external potential ν(r) generated by the nuclei
acting on the electrons,
[
10]
tack is at C-1, however in the case of 2b this can not be
the case and warranted a series of further studies to under-
stand this result.
f(r) = [Ѩρ(r)/ѨN]ν(r)
(1)
where, ρ(r) is the ground state electronic density of the sys-
tem at point r, and N is the total electron number.
Yang and Mortier[
function indices,
21]
proposed the condensed Fukui
fk+ ≈ q (N + 1) – q (N)
k
k
–
f
k
≈ q
k
(N) – q
k
(N – 1)
(2)
fk0 ≈ q (N + 1) – q (N – 1)
k
k
where, the Fukui function is condensed in the individual Figure 3. Calculated structure of (R,R)-Me-DUPHOS-Pd-
+
(
PhCHCHCHPh) (2a) (without chloride ion), with f
k
values for
atoms. Here q (N + 1), q (N – 1) and q (N) are the atomic
k
k
k
C1 and C3 and Pd–C1 and Pd–C3 distances [Å].
populations of atom k in the N + 1, N and N – 1 electron
+
systems. The indice fk reflects the capacity of atom k to
accommodate an extra electron and is the indice most
We considered the possibility of the formation of two
–
suited for studies of nucleophilic attack, fk describes the other active catalysts – trans-2b and trans-2bЈ which are
ability of the atom to donate an electron and is appropriate both diastereomers of cis-2b (Scheme 2) – via the well
0
3
1
3
for studies of electrophilic attack and fk is an indicator for known π-σ-π (η -η -η ) isomerisation mechanism, which
[
1]
radical reactivity. Parr and Yang proposed that the higher allows for syn-anti interconversions to occur (Scheme 2).
Fukui function values are related with increased reactivity In the case of trans-2b (Figure 5), once again it was shown
on that site. So, a larger condensed Fukui function on an that C1 was the more susceptible site for nucleophilic at-
atom indicates an increased reactivity at that particular tack, thus ruling out the possibility of this complex to be
atom.
the active catalyst.
Eur. J. Org. Chem. 2009, 6311–6317
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6313

V. R. Marinho, J. P. P. Ramalho, A. I. Rodrigues, A. J. Burke
FULL PAPER
Figure 4. Calculated structure of (R,R)-DUPHOS-iPr-Pd-
+
Figure 6. Calculated structure of (R,R)-DUPHOS-iPr-Pd-
(
PhCHCHCHPh) (cis-2b) (without chloride ion), with f
k
values
+
(
PhCHCHCHPh) (trans-2bЈ), with f
k
values for C1 and C3 and
for C1 and C3 and Pd–C1 and Pd–C3 distances [Å].
Pd–C1 and Pd–C3 distances [Å].
Table 2. Calculated energies for complexes cis-2b, trans-2b and trans-2bЈ
at the B3LYP/G-31G*//PM6 level.
Complex
Energy
cis-2b
trans-2b
trans-2bЈ
–1735.04774
–1735.04287
–1735.03395
In those reactions where the ee values were lower (Table 1,
entries 1, 2, 6 and 7) it is likely that there was nucleophilic
attack at C-1.
Figure 5. Calculated structure of (R,R)-DUPHOS-iPr-Pd-
+
(
PhCHCHCHPh) (trans-2b) (without chloride ion), with f
k
values
for C1 and C3 and Pd–C1 and Pd–C3 distances [Å].
Kinetic Studies
It was of interest to compare quantitatively the reactivit-
In the case of the trans-2bЈ (Figure 6) it was C-3 that had ies of complex 2a with 2b in the early stages of the reaction.
+
the greater fk value for the first time. We then calculated Thus the asymmetric allylic alkylation using rac-1,3-di-
the energies for each of the three complexes, cis-2b, trans- phenylpropenyl acetate as substrate and [Pd(allyl)Cl] as
2
2b and trans-2bЈ (Table 2) in the gas phase.
pre-catalyst was conducted under standard conditions (see
It was observed that cis-2b was calculated to be the most Table 1) using ligands 1a and 1b. During the first 4.5 hours
stable, followed closely by trans-2b, with trans-2bЈ the most of the reaction a sample was removed at 30 min intervals
unstable. This has led us to postulate that due to stereo- and analysed by HPLC. The results are shown in Figure 7.
chemical hindrance the maloante attacks preferentially C-3 The ee values at each stage of the reaction were also deter-
despite the fact that electronically it is C-1 that is favoured. mined.
Scheme 2. Possible interconversions of the 2b complex.
6314
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2009, 6311–6317

0
Pd -Catalysed Asymmetric Allylic Alkylation Reaction
this time period. Although this result indicated that kinetics
for formation of complexes 4a and 4b did not intefere with
the overall kinetics of the reaction, it still did not indicate
if steric hindrance or a slower rate of formation of 2b from
the pre-allyl palladium complex 4b was the reason for the
slower overall reaction kinetics with 1b. Thus a second ki-
netic study was conducted using ligands 1a and 1b. Both
ligands were refluxed with [Pd(allyl)Cl] to form the respec-
2
tive allyl chloride complexes 4a and 4b (Figure 8) in situ
with subsequent addition of rac-1,3-diphenylpropenyl ace-
tate to form complexes 2a and 2b (Figure 2). It was ob-
served that in fact (see Figure 10) the consumption of rac-
1,3-diphenylpropenyl acetate with 1a was faster than with
1b (after 1 h 37% of the substrate was consumed with 1a
whereas only 27% was consumed with 1b). This strongly
indicated that the reaction rate for formation of complex 2a
Figure 7. Kinetic studies using complexes 2a and 2b formed in situ. was faster. Overall these results were an indication that the
faster overall kinetics for the reaction involving 2a was
probably a consequence of both the rapid formation of the
This study showed that the reaction which involved in
1
,3-diphenylallyl palladium complex 2a coupled with a re-
situ formation of complex 2a went to completion within
the first hour whereas that involving complex 2b was very
sluggish throughout the reaction course. This observation
supports the conjecture that either, (1) there is stereochemi-
cal hindrance during the malonate nucleophilic addition
step which is perhaps the reason for the configurational
switch in the product configuration on going from 2a to 2b
or (2) the rates of ionization and formation of complexes
duced level of stereochemical hindrance on attack by ma-
lonate nucleophile. The fact that the consumption of rac-
1
3
,3-diphenylpropenyl acetate only reached ca. 37% and
5% for both 1b and 1a, after 20 h, strongly indicated the
presence of an equilibrium process and the requirement of
the malonate nucleophile to push the equilibrium towards
product formation and rac-1,3-diphenylpropenyl acetate
consumption, this is particularly the case with 1a. In the
case of 1b, it seems that the overall reaction conversion mir-
rors the substrate consumption value over the 20 h period.
2a and 2b are different, with 2a manifesting more rapid re-
action kinetics. The fact, that the ee values remained con-
stant throughout was an indication that one principle cata-
lyst was catalysing the reaction and it was stable. In order
to determine the most likely reason for these differential
reaction rates, we decided to conduct two key kinetic ex-
periments. Due to the difficulty in preparing the triflate
salts of both 2a and 2b in accordance with the procedure
of Pregosin and Drago,[ we opted for a study to compare
the relative reactivities of the pre-complexes of both 2a and
10]
2b but used the trifluoroborate salts 3a and 3b (Figure 8).
Figure 8. Isolated Pd-allyl complexes 3a and 3b formed using
AgBF and in situ formed complexes 4a and 4b.
4
Figure 9. Kinetic studies using pre-formed complexes 3a and 3b.
On performing kinetic studies with both 3a and 3b in the
same way as previously, but only over a 2 h period we no-
ticed that although the enantioselectivities were comparable
to those obtained using in situ formed 2a and 2b the con-
versions were much lower over the same period as for the
in situ kinetic study described above (Figure 9). This we at-
tribute to the presence of the tetrafluoroborate ion and the
fact that the complexes had to be isolated in air and light,
and thus some decomposition was inevitable. However, it
was indeed the reaction that involved complex 3a that was
quicker. The reaction involving 3b was very sluggish over Figure 10. Kinetic studies using ligands 1a and 1b.
Eur. J. Org. Chem. 2009, 6311–6317
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6315

V. R. Marinho, J. P. P. Ramalho, A. I. Rodrigues, A. J. Burke
BSA (0.3 mL, 3 equiv.) and KOAc (0.4 mg, 1 mol-%) were added
FULL PAPER
Conclusions
[
11]
(
In the case of the reaction with NaH the procedure given in ref.
The commercial ligand, (R,R)-DUPHOS-iPr 1b was was followed). The mixture was maintained at room temp. (or, in
tested in the asymmetric allylic alkylation using rac-1,3-di- some cases, refluxed) for 20 h, after which it was filtered using a
pad of celite and silica gel and washed with a mixture of hexane/
phenylpropenyl acetate as substrate, malonate as nucleo-
EtOAc (2:1). The solvent was evaporated under vacuum and the
resulting mixture analysed by HPLC.
phile and a variety of Pd pre-catalysts under standard con-
ditions. Excellent ee values could be achieved with 1b but
the conversions were moderate (except when NaH was used Preparation of the [AllylPd-1b]CF
as base and the reaction heated – entry 3, Table 1 and when (2.1 mg, 2.5 mol-%) and [Pd(C )Cl]
(5 mL) and stirred at room temp., under a
nitrogen atmosphere. After 30 min AgCF SO (0.5 mg, 1 equiv. to
3
3
SO Complex 2c: Ligand 1b
3
H
5
2
(0.7 mg, 1 mol-%) were dis-
2 2
solved in dry CH Cl
we increased the loading of palladium pre-catalyst by one
order of magnitude – entry 8, Table 1) and inferior to those
reported with (R,R)-Me-DUPHOS 1a under similar condi-
tions. This was attributed to both slower reaction kinetics
in the formation of the 1,3-diphenylallyl palladium complex
3
3
Pd) was added. The reaction mixture was stirred for 45 min and
then rac-1,3-diphenylpropenyl acetate (50 mg, 100 mol-%) was
added. After 20 h stirring at room temp., the mixture was filtered
over a pad of celite and silica. The filtrate was concentrated in
vacuo giving a green oil. MS (MaldiTOF, m/z): 527.22 (Pd-1b + 3,
2b which affords the alkylated malonate product after at-
tack by the malonate nucleophile and to increased stereo-
31
3
cleavage). P NMR (CDCl , 121.5 MHz): δ = 64.35 (d, J = 31 Hz,
1
chemical hindrance during the attack of the malonate nu- P1) and 62.55 (d, J = 31 Hz, P2) ppm. H NMR (CDCl , 300 MHz,
3
cleophile on 2b. The surprising change of configuration in substrate signals were detected, but are ignored here): δ = 0.45–
the malonate product on going from 2a to 2b was rational- 0.47 (d, J = 6.5 Hz, 3 H, CH
3
isopropyl), 0.54–0.56 (d, J = 6.5 Hz,
isopropyl), 0.72–0.84 (m, 12 H, 4 CH , 2 isopropyl),
.95–0.98 (d, J = 6.7 Hz, 3 H, CH isopropyl), 1.0–1.1 (d, J =
.6 Hz, 3 H, CH isopropyl), 1.2–1.3 (m, 8 H, cyclopentane), 3.5
3
0
6
H, CH
3
3
ised on the basis of a significant stereochemical hindrance
event at C-1 of 2b which resulted in preferential attack at
C-3. Further theoretical studies will be undertaken to sub-
stantiate this postulate.
3
3
(
7
d, J = 6.3 Hz, 1 H, –CHPd), 6.3–6.4 (m, 2 H, –CH=CH) and 7.2–
ar
.6 (m, 14 H, H ) ppm.
Preparation of the [AllylPd-1a]BF
4
Complex 3a: Ligand 1a (30 mg,
(17.9 mg, 0.5 equiv.) were dis-
–
2
9.8ϫ10 mmol) and [Pd(C
3
H
5
)Cl]
2
Experimental Section
General Remarks: Both (R,R)-Me-DUPHOS 1a and (R,R)- nitrogen atmosphere. After 30 min AgBF
DUPHOS-iPr 1b were obtained from Strem chemicals. All other
reagents were obtained from Aldrich, Fluka, Alfa Aesar or Acros.
Solvents were dried using common laboratory methods.[23]
solved in dry CH Cl
2
2
(5 mL) and stirred at room temp., under a
(19.1 mg, 1 equiv.) was
added. The reaction mixture was stirred for 45 min. AgCl was re-
moved and the solvent was reduced in vacuo. The isolated complex
4
2
31
3a was obtained as a white solid (32 mg, 60%). P NMR (CDCl
3
,
1
21.5 MHz): δ = 74.15–74.45 (d, J = 36 Hz, P1) and 75.00–75.29
High Performance Liquid Chromatography (HPLC) analysis was
performed on an Agilent 1100 series instrument. The following
conditions were used: pmax = 50 bar, flux = 1 mL/min, detector =
DAD (λ = 210.10 nm), eluent: n-hexane/2-propanol (98:2). To cal-
culate the reaction conversion (ratio substrate to product), given
that both the substrate and malonate product have different molar
extinction coefficients a correction factor of 0.65 was introduced in
order to correct the product peak area.
1
(
d, J = 35 Hz, P2) ppm. H NMR (CDCl
signals were detected, but are ignored here): δ = 0.87–0.93 (m, 3
H, CH ), 1.13–1.30 (m, 6 H, 2 CH ), 1.33–1.39 (m, 3 H, CH ),
.79–1.93 (m, 8 H, cyclopentane), 4.7 (m, 1 H, CHPd), 5.50–5.65
3
, 300 MHz, substrate
3
3
3
1
ar
(
m, 2 H, CH=CH) and 7.8–7.9 (m, 14 H, H ) ppm.
4
Preparation of the [AllylPd-1b]BF Complex 3b: As described above
for 3a but using ligand 1b (40 mg, 9.8ϫ10 mmol). The isolated
–
2
3
1
complex 3b was obtained as a white solid (43 mg, 68%). P NMR
CDCl , 121.5 MHz): δ = 61.84–62.09 (d, J = 31 Hz, P1) and
3.63–63.89 (d, J = 31 Hz, P2) ppm. H NMR (CDCl
substrate signals were detected, but are ignored here): δ = 0.46–
The column used was a Chiralcel OD-H (0.46 cmϫ25 cm) fitted
with a guard column composed of the same stationary phase. In
all cases, the reaction conversions were calculated by simply de-
termining the ratio of the peak areas for the substrate and the alkyl-
ated product.
(
6
3
1
3
, 300 MHz,
0
3
.48 (d, J = 6.5 Hz, 3 H, CH
H, -CH isopropyl), 0.73–0.85 (m, 12 H, 4 CH
isopropyl), 1.05–1.07 (d, J =
isopropyl), 1.69–1.87 (m, 8 H, cyclopentane), 4.7
3
isopropyl), 0.55–0.57 (d, J = 6.5 Hz,
3
3
, 2 isopropyl),
In the case of the kinetic studies the reactions were monitored by
collecting 1 mL samples at 30 min intervals for the first 1.5–4.5 h
and for the last hour. The samples were analyzed using HPLC. For
the kinetic studies involving complexes 3a and 3b the conversions
were determined as described above. In the case of the kinetic study
that involved determining the rate of consumption of the rac-1,3-
diphenylpropenyl acetate substrate this was calculated using a cali-
0.97–0.99 (d, J = 6.6 Hz, 3 H, -CH
6.5 Hz, 3 H, CH
(m, 1 H, –CHPd), 5.46–5.60 (m, 2 H, –CH=CH) and 7.8–7.9 (m,
3
3
ar
14 H, H ) ppm.
General Procedure for Kinetic Studies with Ligands 1a and 1b in
the Presence of Dimethyl Malonate and Base: [Pd(allyl)Cl]
mol-%) and the chiral ligand 1a (15.2 mg, 2.5 mol-%) or 1b
20.7 mg, 2.5 mol-%) were placed in a flask with dry CH Cl
5 mL) under a nitrogen atmosphere and refluxed for 2 h. The tem-
2
(7.2 mg,
1
(
(
–
3
bration curve for rac-1,3-diphenylpropenyl acetate between 4.2·10
to 40.0·10–3  (y = 1227.7x + 36.677; R = 0.9495).
2
2
2
General Procedure for the Catalytic Asymmetric Allylic Alkylation
Reactions Using Ligands 1a and 1b: The Pd pre-catalyst (0.5 or phenylpropenyl acetate (500 mg, 1.9 mmol) dissolved in dry
mol-%) and the chiral ligand (1.2 or 2.5 mol-%) were placed in a CH Cl (2 mL) was added. Dimethyl malonate (0.7 mL, 3 equiv.),
perature was then reduced to room temp. after which rac-1,3-di-
1
2
2
flask with dry solvent (2 mL) under a nitrogen atmosphere and
refluxed for 2 h. The temperature was then reduced to room temp.
after which the substrate (100 mg, 0.39 mmol) dissolved in a dry
BSA (1.5 mL, 3 equiv.) and KOAc (2 mg, 1 mol-%) were added.
The mixture was maintained at room temp. During the first 4.5 h
and the last hour of the reaction a sample was removed at intervals
solvent (2 mL) was added. Dimethyl malonate (0.15 mL, 3 equiv.), of 30 min and analysed by HPLC.
6
316 Eur. J. Org. Chem. 2009, 6311–6317
www.eurjoc.org
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

0
Pd -Catalysed Asymmetric Allylic Alkylation Reaction
General Procedure for Kinetic Studies with Complexes 3a and 3b: [4] C. S. Marques, A. J. Burke, Tetrahedron: Asymmetry 2007, 18,
Complexes 3a (5.4 mg, 10–3 mmol) or 3b (6.5 mg, 9.9ϫ10 mmol)
–3
1804–1808.
[
5] C. S. Marques, A. J. Burke, Synth. Commun. 2008, 38, 4207–
were placed in a flask with dry CH
atmosphere. rac-1,3-Diphenylpropenyl acetate (100 mg, 0.39 mmol)
dissolved in dry CH Cl (2 mL), dimethyl malonate (0.15 mL,
equiv.), BSA (0.3 mL, 3 equiv.) and KOAc (0.4 mg, 1 mol-%) were
2 2
Cl (2 mL) under a nitrogen
4214.
[
[
6] V. M. Marinho, A. J. Burke, Synth. Commun., in press.
2
2
7] W. Li, J. P. Waldkirch, X. Zhang, J. Org. Chem. 2002, 67, 7618–
3
7623.
added. The mixture was maintained at room temp. for 20 h. During
the first 3 h and the last hour of the reaction a sample was removed
at intervals of 30 min and analysed by HPLC.
[
8] T. Sturm, W. Weissensteiner, F. Spindler, Adv. Synth. Catal.
2003, 345, 160–164.
[9] M. J. Burk, J. E. Feaster, W. A. Nugent, R. L. Harlow, J. Am.
Chem. Soc. 1993, 115, 10125–10138.
General Procedure for Kinetic Studies with Ligands 1a and 1b in
[10] D. Drago, P. S. Pregosin, J. Chem. Soc., Dalton Trans. 2000,
3191–3196.
the Absence of Dimethyl Malonate and Base: [Pd(allyl)Cl]
2
(22 mg,
.06 mmol) and the chiral ligand 1a (36 mg, 2 equivs) or 1b (49 mg,
.12 mmol) were placed in a flask with dry CH Cl (5 mL) under
[
[
[
11] G. Malaisé, S. Ramdeehul, J. A. Osborne, L. Barloy, N. Kyrit-
0
0
sakas, R. Graff, Eur. J. Inorg. Chem. 2004, 3987–4001.
2
2
12] A. Togni, C. Breutel, A. Schnyder, F. Spindler, H. Landert, A.
Tijani, J. Am. Chem. Soc. 1994, 116, 4062–4066.
a nitrogen atmosphere and refluxed for 2 h. The temperature was
then reduced to room temp. after which rac-1,3-diphenylpropenyl
acetate (30 mg, 0.12 mmol) dissolved in dry CH Cl (2 mL) was
2 2
13] Y. Fukuda, K. Kondo, T. Aoyama, Tetrahedron Lett. 2007, 48,
3389–3391.
added. The mixture was maintained at room temp. for 20 h. During
the first 1.5 h of the reaction a 1 mL aliquot was removed at inter-
vals of 30 min and analysed by HPLC.
[14] J. J. P. Stewart, J. Mol. Mod. 2007, 13, 1173–1213.
[15] MOPAC2007, James J. P. Stewart, Stewart Computational
Chemistry, Version 7.319L web: HTTP://OpenMOPAC.net.
[
[
[
16] A. D. Beck, J. Chem. Phys. 1993, 98, 5648–5652.
17] C. Lee, W. Yang, R. G. Parr, Phys. Rev. 1988, 37, 785.
18] M. W. Schmidt, K. K. Baldridge, J. A. Boatz, S. T. Elbert,
M. S. Gordon, J. H. Jensen, S. Koseki, N. Matsunga, K. A.
Nguyen, S. J. Su, T. L. Windus, M. Dupuis, J. A. Montgomery,
J. Comput. Chem. 1993, 14, 1347–1363.
Acknowledgments
We are grateful for financial support from the Fundação para a
Ciência e a Tecnologia (FCT) (project; POCI/QUI/55779/2004)
through POCI 2010, supported by the European community fund,
FEDER. FCT is also acknowledged for the award of a PhD grant
to V. R. M. (SFRH/BD/25111/2005).
[
[
[
19] a) R. G. Parr, W. Yang, J. Am. Chem. Soc. 1984, 106, 4049–
4050; b) J. Melin, F. Aparicio, V. Subramanian, M. Galván,
P. K. Chattaraj, J. Phys. Chem. A 2004, 108, 2487–2491.
20] a) F. Méndez, J. L. Gázquez, J. Am. Chem. Soc. 1994, 116,
9
9
298–9301; b) J. L. Gázquez, F. Méndez, J. Phys. Chem. 1994,
8, 4591–4593.
21] W. Yang, W. J. Mortier, J. Am. Chem. Soc. 1986, 108, 5708–
[
[
[
1] For a key review, see: B. M. Trost, C. Lee, in: Catalytic Asym-
5711.
nd
metric Synthesis (Ed.: I. Ojima), 2 ed., Wiley-VCH, 2000.
[22] R. Contreras, P. Fuentealba, M. Galvan, P. Perez, Chem. Phys.
Lett. 1999, 304, 405–413.
2] B. M. Trost, P. E. Strege, J. Am. Chem. Soc. 1977, 99, 1649–
1651.
[23] W. L. F. Armarego, D. D. Perrin, Purification of laboratory
th
3] G-Q. Lin, Y-M. Li, A. S. C. Chan, Principles and applications
of Asymmetric Synthesis, Wiley-Interscience, 2001, pp. 331, and
references cited therein.
chemicals, 4 ed., Butterworth Heinmann, Oxford, 1996.
Received: June 13, 2009
Published Online: November 9, 2009
Eur. J. Org. Chem. 2009, 6311–6317
© 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
6317