Deconvoluting the memory effect in Pd-catalyzed allylic alkylation: Effect of leaving group and added chloride

Article abstract of DOI:10.1002/chem.200600152

An analysis of product distributions in the Tsuji-Trost reaction indicates that several instances of reported "memory effects" can be attributed to slow interconversion of the initially formed syn- and anti- [Pd(η³-allyl)] complexes. Addition of chloride triggers a true memory effect, in which the allylic terminus originally bearing the leaving group has a higher reactivity. The latter effect, termed regioretention, can be rationalized by ionization from a palladium complex bearing a chloride ion, forming an unsymmetrically substituted [Pd(η³-allyl)] complex. DFT calculations verify that the position trans to the phosphine ligand is more reactive both in the initial ionization and in the subsequent nucleophilic attack.

Full text of DOI:10.1002/chem.200600152

DOI: 10.1002/chem.200600152
Deconvoluting the Memory Effect in Pd-Catalyzed Allylic Alkylation: Effect
of Leaving Group and Added Chloride
[
a]
Peter Fristrup, Thomas Jensen, Jakob Hoppe, and Per-Ola Norrby*
Abstract: An analysis of product distri-
butions in the Tsuji–Trost reaction indi-
cates that several instances of reported
allylic terminus originally bearing the
leaving group has a higher reactivity.
The latter effect, termed regioretention,
can be rationalized by ionization from
a palladium complex bearing a chloride
ion, forming an unsymmetrically substi-
3
“memory effects” can be attributed to
tuted [Pd
A
H
R
U
G
slow interconversion of the initially
culations verify that the position trans
to the phosphine ligand is more reac-
tive both in the initial ionization and in
the subsequent nucleophilic attack.
Keywords: allylic compounds · ha-
3
formed syn- and anti-[Pd(h -allyl)]
A
H
R
U
G
AHCTREUNG
complexes. Addition of chloride trig-
gers a true memory effect, in which the
Introduction
The palladium-catalyzed allylic alkylation (Tsuji–Trost reac-
tion) is a popular tool in modern organic synthesis. The re-
[1]
action has been the subject of numerous reviews, and in
particular the asymmetric version has received much recent
attention. The generally accepted mechanism involves
[2]
Scheme 1. Basic mechanism for the Tsuji–Trost reaction.
complexation with Pd(0), displacement of an allylic leaving
3
group (X) leading to the formation of an [Pd
A
H
R
U
G
[
6–8]
cies, and finally, attack by nucleophiles at either terminus of
the allyl moiety, releasing Pd(0) and closing the catalytic
cycle.
for unsymmetrically substituted allyls,
and the term
“memory effect” has sometimes been applied to this phe-
nomenon. The mechanism of the memory effect has been
studied, and is in many cases well understood. Particu-
larly clear cases can be noted in the presence of chiral li-
in which initial ionization can give different
points of entry into the [Pd(h -allyl)] manifold for enantio-
topic substrates. Another important class of memory effects
is observed upon formation of unsymmetrically substituted
h -allyls, for which the product distribution can be depend-
ent on both the regio- and stereochemistry of the starting
material. The latter class can be observed also with nonchi-
ral ligands, and is the subject of the current study. In reac-
(h -allyl)] intermediates,
three isomeric starting materials (cis, trans, and internal) can
[9–12]
From this mechanism it can be predicted that two isomer-
ic allylic substrates that can yield the same intermediate,
such as the isomeric allylic substrates shown in Scheme 1, or
enantiomeric substrates leading to the same symmetric inter-
mediate, should yield the same product distribution. Howev-
er, more than two decades ago, Fiaud and Malleron report-
ed that enantioenriched cyclohexenyl acetate yields product
with partial retention of the optical activity, in conflict with
[9,11,12]
gands,
3
A
H
R
U
G
3
[3]
the accepted mechanism. The experimental results were
[4]
[8]
immediately questioned by Trost and later also by Bos-
[5]
3
nich. However, a substrate-dependent product distribution
that cannot be rationalized by using a single common inter-
mediate has been noted on several occasions, in particular
tions involving monosubstituted [Pd
A
H
R
U
G
3
lead to two isomeric [Pd
A
H
R
U
G
that can equilibrate and, in turn, can produce three isomeric
[8]
products (Scheme 2).
It has been suggested that the fast dynamics of the [Pd-
[
a] P. Fristrup, T. Jensen, J. Hoppe, Prof. Dr. P.-O. Norrby
Department of Chemistry, Technical University of Denmark
Building 201, Kemitorvet, 2800 Kgs. Lyngby (Denmark)
Fax : (+45)459-33-968
[5]
[13]
3
A
H
R
U
G
ates. However, a series of papers from the Škermark group
clearly showed that the isomerization can, in some instances,
E-mail: pon@kemi.dtu.dk
5352
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 5352 – 5360

FULL PAPER
termediate, favoring reactivity of the terminus that was con-
nected to the leaving group before initial ionization. Many
explanations for the reactivity difference between the termi-
ni have been advanced, including a tight ion pair with the
[
12]
1
leaving group,
reaction through [Pd
A
H
R
U
G
[17]
ates, and unequal trans effects arising from unsymmetrical
ligation. In the case in which chiral ligands are employed,
[18]
the available evidence indicates that enantiotopic substrates
Scheme 2. Isomers in the Tsuji–Trost reaction of butenyl substrates.
3
ionize to diastereomeric [Pd
A
H
R
U
G
which the interactions of the chiral ligand with the allyl
moiety lead to differences in reactivity and, thus, to unequal
product distributions. In nonchiral systems, attention has
be rendered slow enough to allow isolation of pure syn and
3
[9]
anti configurations of the [Pd
A
H
R
U
G
thermore, that the reactivities of these configurations differ
focused on branched-versus-linear products, and several
studies have shown that branched and linear substrates
indeed give different product distributions. However, most
of these studies have employed only two substrates, com-
monly trans linear and branched butenyl substrates. We note
that the mechanism depicted in Scheme 2 is well able to ra-
tionalize these results, because the branched substrate is ex-
[8]
considerably. In short, nucleophilic attack is disfavored at
[8]
syn-substituted positions.
In the reaction depicted in
Scheme 2, the anti complex produces cis linear and
branched products in almost equal amounts as the reactivi-
ties of the two termini are similar, whereas the syn complex
[8]
gives predominantly the trans linear product. Thus, it is
3
clear that slow equilibration of the intermediate will give
pected to ionize at least partially to an anti-[Pd(h -alkenyl)]
[14,15]
rise to memory effects.
A trans substrate must ionize in-
intermediate, with a product profile markedly different from
3
itially to a syn complex (usually, but not always, the most
that of the syn-[Pd(h -alkenyl)] complex initially formed
[16]
[8]
stable isomer ) that will then give preferential formation
of trans product, due to the inherent reactivity difference
between the syn-substituted and -unsubstituted termini. A
cis substrate, on the other hand, ionizes to the anti complex,
and usually yields a mixture of products, depending on the
particular substituent and the rate of isomerization. Finally,
the branched substrate would be expected to give a result
intermediate between that of the cis and trans substrates, en-
tirely dependent on the initial ionization preference (anti/
syn). In fact, if no memory effects beyond those implied in
Scheme 2 are active, the product distribution obtained from
the branched substrate must be a linear combination of the
distributions obtained by using cis and trans substrates. We
propose that by investigating the product distribution from
each of the three isomeric substrates in Scheme 2, it is possi-
ble to elucidate the degree of isomerization of the inter-
mediate and the ionization preference of the branched sub-
strate, as well as to detect memory effects not included in
Scheme 2 by deviations from the linear relationship between
the three product distributions. Herein, we use the term ster-
eoretention for unequal product distributions caused by slow
isomerization between syn and anti complexes. Experimen-
tally, this is detected by comparing the distribution of linear
products from cis and trans linear substrates, respectively
from the trans linear substrate. Herein, we show that inclu-
sion of the cis linear substrate in the study allows a detec-
tion of memory effects beyond those apparent from the ste-
A
H
R
U
G
meric intermediates in Scheme 2. We use the term regiore-
tention exclusively for cases in which it can be proven that
the intermediate Pd–allyl complex retains a “memory” of
the position of the leaving group beyond the inherent reac-
tivity difference between syn and anti complexes. Experi-
mentally, we detect this effect by comparing the product dis-
tributions from all three isomeric substrates in Scheme 2. If
the product distribution obtained from the branched sub-
strate is a linear combination of those obtained from cis and
[19]
trans linear substrates, we have no regioretention.
Herein, we have studied the behavior of butenyl sub-
3
strates that ionize to simple methyl-substituted [Pd
A
H
R
U
G
intermediates (Figure 1). It is of critical importance that all
3
three substitution patterns leading to the same [Pd(h -allyl)]
A
H
R
U
G
[8]
manifold are included.
Alcohols 1a–c were synthesized and converted into sub-
strates 2–4 by using standard methods. The substrates were
reacted with two equivalents of the sodium salt of diethyl
methyl malonate, in the presence of 2.5% [Pd (dba) ]
A
C
H
T
R
E
U
N
G
2 3
(dba=(E,E)-dibenzylideneacetone) and 10% PPh , to gen-
3
(
any branched product can be ignored in this analysis). If
erate 5% of the catalytically active [Pd(PPh ) ] complex.
A
H
R
U
G
3 2
the cis and trans substrates yield identical linear-product dis-
tributions, there is no stereoretention. Conversely, if there is
no crossover, that is, if only cis linear product is obtained
from cis substrate and trans linear product is obtained from
The product distribution was determined by GC and NMR
trans substrate, we have full stereoretention. We expect ad-
3
ditives that increase the rate of isomerization of the [Pd
allyl)] intermediate, such as halides,
degree of stereoretention.
A
H
R
U
G
[13]
to decrease the
The “memory effect” has been frequently rationalized in
3
terms of an unsymmetrical reactivity in the [Pd
A
H
R
U
G
Figure 1. Allylic substrates and corresponding products studied.
Chem. Eur. J. 2006, 12, 5352 – 5360
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5353

P.-O. Norrby et al.
spectroscopy. Halide ions were rigorously excluded in the in-
itial part of the study to avoid complications caused by the
tention, here defined as the amount that reacts in the initial-
[19]
ly formed isomeric form of the intermediate. By using the
above ratios in analyzing the product distribution from
branched acetate 2b, we can conclude that the initial ioniza-
tion yields approximately 70% syn complex. However, we
also see a slight excess of internal product relative to what
would be expected from the observed amounts of cis and
trans product (observed: 26%, predicted: 19%). This could
be a slight regioretention, that is, a preference for the nucle-
ophile to attack the carbon that originally carried the leav-
[20]
“halide effect”.
In the second part of the study, we
wanted to investigate the influence of chloride ions, without
adding a new type of countercation to the reaction. This was
achieved by addition of a small amount of allyl chloride,
which reacts more rapidly than the allylic acetates, liberating
a controlled amount of chloride ions.
[19]
Results and Discussion
ing group. The effect here is weak, and may not be signifi-
[22]
cantly larger than the error in the measurements.
Notably, branched products arise primarily from anti-[Pd-
Allylic alkylation of acetates 2a–c: The acetate leaving
group is expected to be only weakly coordinating to palladi-
um. As for any coordinating anion, it will accelerate equili-
3
AHCTREUNG
3
3
[16]
bration of the h -allyl intermediate, though not as strongly
of monosubstituted [Pd
A
H
R
U
G
[21]
as, for example, a chloride anion. The results from allylic
alkylation of the acetates 2a–c are listed in Table 1 (en-
tries 1–3).
an improved chance of giving branched, chiral product in
[2]
the asymmetric version of the title reaction. In addition, it
is clear that successful rationalization of enantioselectivity
[23]
requires consideration of anti complexes.
The strong stereoretention effect observed here, resulting
from a slow interconversion of intermediates, can be used to
Table 1. Results from allylic alkylation.
[6]
Entry
Substrate
5a
5b
[%]
5c
[%]
Chloride
[mol %]
rationalize the observed memory effects of Williams, Hay-
[7]
[10]
[%]
ACHTREUNG
ashi, and Faller, by postulating that a significant propor-
tion of the internal allylic substrates utilized in those studies
ionized initially to the anti complex. As no cis substrates
were employed, it is not possible to state if also a regiore-
tention mechanism was operative.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
2a
2b
2c
3a
3b
3c
4a
4b
4c
4c
2a
2b
2c
2a
2b
2c
77
59
15
81
55
52
67
61
47
36
76
36
45
81
43
53
16
26
38
17
30
30
22
26
34
38
16
51
21
17
46
22
7
–
–
–
–
–
–
–
–
–
14
47
3
15
18
11
12
20
26
8
13
34
2
11
25
Notably, a wide range of relative rates of isomerization
and nucleophilic attack can be found in the literature. For
example, almost complete stereoretention can be achieved
[24]
by increasing the reactivity of the nucleophile,
whereas
0
1
2
3
4
5
6
–
0.5
0.5
0.5
5
5
5
complete isomerization within an isolated manifold coupled
with substrate-directed regioselectivity forms the basis of a
[25]
recent enantioconvergent procedure.
New leaving groups: Two alternative types of leaving groups
were included in the current studies, isoureas and imidates.
[26]
Dicyclohexylisoureas were employed previously, however,
in our hands, it was difficult to isolate these in pure form.
The corresponding diisopropylisoureas were found to be
more tractable (3a–c). In addition, we wanted to test tri-
chloroacetimidates as leaving groups (4a–c), as these are
easily synthesized from the corresponding alcohols, and
could also function as bases upon liberation. For the latter
substrates, a one-pot tandem procedure was developed.
Thus, the sodium salt of an allylic alcohol (1a–c) was treated
We analyze the results in terms of the mechanism depict-
ed in Scheme 2. If nucleophilic attack is slow relative to the
isomerization of the intermediate, the same product distri-
bution should be expected from all three substrates. This is
clearly not the case; trans-acetate 2a yields mostly trans
product 5a from terminal attack on the syn intermediate,
whereas cis-acetate 2c yields almost equal amounts of
branched and cis products, 5b and 5c, respectively, in good
agreement with a previous study from the Škermark
with CCl CN, followed by the catalyst and the malonate nu-
3
cleophile in neutral form, to yield directly products 5a–c.
The results obtained with the new leaving groups are
shown in Table 1 (entries 4–10). There are some distinct dif-
ferences from the allylic acetates. First of all, it is clear that
the isomerization is now faster (i.e., there is less stereoreten-
tion), as shown by the high proportion of trans product 5a
from the cis substrates 3c and 4c (compare entries 6 and 9
to entry 3). Furthermore, the position of the syn/anti equili-
brium has clearly shifted; despite the higher rate of isomeri-
[8]
group. A more detailed analysis of the product distribution
from 2a and 2c shows that under the current reaction condi-
tions, only 12% of the syn complex (from 2a) and 17% of
the anti complex (from 2c) isomerize before reaction with
nucleophile, and that the ratio of terminal:internal attack is
8
8:12 for syn, and 57:43 for anti. (See Experimental Sec-
tions for the iterative formulae used to derive these ratios.)
Thus, for both substrates, there is more than 80% stereore-
5354
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 5352 – 5360

Pd-Catalyzed Allylic Alkylation
FULL PAPER
zation of the intermediate, trans-isourea 3a yields only ap-
proximately 3% of cis product, compared to 7% for the
acetate 2a, and 11% for the imidate 4a (entries 4, 1, and 7).
This indicates that the leaving groups function to some
extent as Pd ligands, influencing the equilibrium by coordi-
nation. The experiment starting from 4c was repeated (en-
tries 9 and 10), and it is clear that the relative rates of iso-
merization and nucleophilic attack are very sensitive to re-
action conditions. However, none of our conclusions are
sensitive to variations of this moderate magnitude.
For both new leaving groups, the degree of regioretention
is negligible. The product distribution from 3b closely
matches 3c (entries 5 and 6), whereas 4b instead gives a
result more similar to 4a (entries 8 and 7) showing that the
position of the leaving group is immaterial, and also that the
two leaving groups have opposite ionization preferences.
Isourea 3b, like 3c, forms the anti-allyl complex upon ioni-
zation. To our knowledge, this is the first report of a leaving
group that ionizes to the thermodynamically less-favored
anti complex. Imidate 4b, on the other hand, gives mainly
the syn complex and, thus, matches 4a. As a corollary, isour-
eas should be better leaving groups than acetates if a
branched-to-branched reaction is desired, whereas imidates
should be preferred for isomerization to linear product.
Scheme 3. Plausible intermediates rationalizing regioretention in the
presence of chloride.
3
The first is a neutral [Pd
A
T
E
N
(h -allyl)] complex with two un-
ꢀ
equal ligands, PPh and Cl , in which both ionization and
3
nucleophilic attack would be expected to occur primarily
[18]
trans to phosphorus. The second is a likewise neutral [Pd-
1
A
H
R
U
G
A
H
R
U
G
3
2
[
3,17]
experimental results indicate that [Pd(h -allyl)] complexes
react with electrophiles rather than nucleophiles. DFT mod-
eling also indicates that the former explanation is well able
to rationalize the observed results (see below).
The chloride effect shows an early saturation behavior.
ꢀ
The effect is dramatic already at 0.5% Cl (entries 11–13),
ꢀ
and then does not change much upon increasing to 5% Cl
(equimolar with Pd, entries 14–16). This is a strong indica-
tion that under the present reaction conditions, the resting
state of the catalyst is Pd(0), and that less than 10% of the
II
ꢀ
Pd is present in Pd form at low levels of Cl . In addition,
ꢀ
the ionization event must be strongly accelerated by Cl , as
shown by Amatore, Jutand, and co-workers, as an excess
of Pd(0) over Cl has little effect.
[
20,27]
[17]
Effect of chloride ions: To study the “halide effect”,
the
ꢀ
allylic alkylation of the acetates 2a–c was also performed by
the addition of small amounts of allyl chloride (either 10%
or 100% relative to Pd). In the presence of Pd(0) the very
reactive allyl chloride will immediately liberate a controlled
amount of chloride ions into the reaction mixture. We ex-
pected the main effect of the chloride ions to be an accelera-
Computational study: To verify some of the conclusions
reached in the mechanistic study we undertook a computa-
tional study of the selectivity-determining features in the
transition state of the title reaction, assuming that initial
ionization occurs by reaction of allylic acetate with an
[13]
tion of the syn/anti isomerization. With this in mind, the
product distributions from the three isomeric acetates 2a–c
should become more similar upon chloride addition. The re-
sults are shown in Table 1 (entries 11–16).
anionic [Pd(PR )X] complex (X=chloride or carboxylate),
A
H
R
U
G
3
[29]
in analogy with recent studies on oxidative addition. We
have shown previously that the transition state in the allyla-
[30]
Upon observation of the product distributions, the initial
expectation (i.e., faster syn/anti isomerization) is clearly ful-
filled, as shown by the higher amount of trans product 5a
from cis substrate 2c. However, the most drastic effect is a
strong regioretention. The results from 2a and 2c show the
two extremes of initial ionization, to the pure syn and anti
complex, respectively. In the absence of regioretention, the
product distribution from the branched substrate must be a
linear interpolation between these two extremes, however,
the results with 2b show almost twice as much internal
product as from either of the other substrates. This is proof
tion reaction is strongly affected by solvent. As in previ-
ous studies, we chose to represent the solvent with a contin-
[30,31]
uum model.
For the phosphine ligand, we used Me P as
3
a model. In analogy with previous studies of similar reac-
[18c,32]
tions,
the nucleophile is modeled by simple ammonia.
This model system is not expected to represent steric inter-
actions well, but should be able to uncover inherent regiose-
lectivity caused by the trans effect. We have also modeled
the initial ionization step, with the same phosphine model,
but including the full acetate leaving group. All DFT calcu-
[
33]
lations were performed with the B3LYP functional
in
[34]
[35]
that the intermediate in this case is not the symmetrically li-
combination with the LACVP* basis set in Jaguar. The
3
[36]
gated [Pd
A
C
H
T
R
E
U
N
G
(h -allyl)] complex shown in Scheme 2. Further-
solvent was represented by using the PB-SCRF model in
more, the observation that the regioretention is triggered by
addition of external chloride makes it very unlikely that in
this case the memory effect is due to tight ion pairing with
Jaguar, with parameters suitable for CH Cl (dielectricity
2
2
constant: epsout=9.08; probe radius: radprb=2.33237).
We first considered the influence of the trans effect on the
initial ionization step. For the model system depicted in
Scheme 4, ionization trans to the phosphine is favored by
[12]
the leaving group, acetate.
Somehow, the chloride ion
must interact with Pd to yield a less-symmetric intermediate.
Two plausible possibilities have been advanced in the litera-
ꢀ
1
approximately 6 kJmol , corresponding to a 10:1 ratio of
products under kinetic control at room temperature. Thus,
ionization in the presence of chloride would be expected to
0
ture, both depending on initial reaction with anionic [Pd -
ꢀ
[17,27]
A
C
H
T
R
E
U
N
G
(PPh ) Cl] ,
(Scheme 3).
3
x
Chem. Eur. J. 2006, 12, 5352 – 5360
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5355

P.-O. Norrby et al.
ments (Table 1). If this isomerization could be suppressed,
we expect that a very high degree of regioretention could be
[18d]
obtained.
Conclusion
The previously reported memory effects in the title reaction
can be divided into two distinct classes that we term regiore-
Scheme 4. Calculated trans effect in the initial ionization of allyl acetate.
tention and stereoretention. The latter is caused by a slow
3
isomerization of the intermediate [Pd
A
H
R
U
G
combination with the inherent stereochemical preference of
the initial ionization step. We also identified reaction with
3
give a large excess of [Pd
A
H
R
U
G
phine ligand trans to the allyl terminus that was originally
bonded to the leaving group.
From this result, it is clear that in the presence of chlo-
anionic [Pd(PR )Cl], influenced by a strong trans effect, as
A
H
R
U
3
one potential source for regioretention in the title reaction.
By changing the nature of the leaving group, the fate of
the initial ionization step can be controlled, directing the re-
action of branched allylic substrates either to an anti com-
plex with high branched preference or to a syn complex
with high linear preference of the subsequent nucleophilic
attack. This discovery holds great promise for further devel-
opment of the synthetic applications of the palladium-cata-
lyzed allylic alkylation. Also, the addition of anionic ligands,
A
C
H
T
R
E
U
N
G
ride, branched and linear substrates will not ionize to the
3
same [Pd(h -allyl)] complexes (Scheme 5). For all of the in-
A
C
H
T
R
E
U
N
G
ꢀ
such as Cl , enhances the branched-to-branched reactivity, a
fact that could be useful in future enantioselective imple-
mentations.
Experimental Section
General: All reactions were performed in glassware flame-dried under
vacuum and flushed with argon, except for the synthesis of the allylic al-
cohols 1a–c and the trichloracetimidates 4a–c. Solvents were distilled
prior to use. THF was distilled from Na/benzophenone and dichloro-
A
H
R
U
G
2 2
methane from CaH , both under N atmosphere. Diethyl ether, ethyl ace-
tate, and hexane used for flash chromatography were of HPLC grade.
Commercially available reagents were used as delivered unless men-
tioned. TLC was performed by using alumina plates coated with silica
Scheme 5. In all examples except the lower one, nucleophilic attack trans
ꢀ1
to phosphine is favored by more than 10 kJmol
.
(
Merck: silica gel 60 F254). The plates were visualized by using a 5%
phosphormolybdic acid solution in ethanol followed by warming with a
heat gun, which resulted in blue spots.
termediates, we calculated the barrier for reaction with the
model nucleophile at each allyl terminus, and in the three
first cases, we found a preference for the terminus trans to
phosphine of 11–13 kJmol , which would predict a regiore-
tention of about 99% for the internal and trans linear sub-
strates. Only in the anti complex, initially formed from the
cis substrate, will the inherent preference for internal attack
compensate the trans effect, and for the mismatched anti
complex we calculate that attack at the two allyl termini are
virtually isoenergetic.
Flash column chromatography was performed by using silica gel
(Amicon 85040) as described by Still et al. Solvents were removed by
using a rotary evaporator (about 17 mmHg at 20–308C).
[
37]
1
ꢀ
1
Products were identified by H NMR spectroscopy. Spectra were record-
ed by using a Varian Mercury 300 operating at 300 MHz. Chemical shifts
are given in ppm relative to CHCl
3
(7.27 ppm). Quartets are designated
by using “q” and apparent quintets are designated by using the abbrevia-
1
3
tion “k”. New compounds were characterized further by C NMR spec-
1
3
troscopy (Varian Mercury 300, 75 MHz). Chemical shifts in C NMR
data are given relative to CDCl (77.0 ppm). Microanalyses were con-
3
ducted by Mikroanalytisches Laboratorium am Institut für Physikalische
Chemie der Universität Wien.
From the results depicted in Scheme 5, we would expect a
very high degree of regioretention for the branched and
trans linear substrates. However, as stated earlier, the pres-
Gas chromatography (GC) of allylic alcohols 1a–c was performed by
using a Hewlett–Packard 5890 Series II gas chromatograph connected to
a Hewlett–Packard 3392A integrator on a 25 m”0.25 mm Chrompack
Chirasil-Dex column. H
2
was used as carrier gas in an isothermic run
ence of chloride and excess ligand will also increase the rate
(
508C, 20 min). Gas chromatography to determine the product distribu-
3
of isomerization of the [Pd
A
H
R
U
G
tion from the allylic substitutions were performed by using a Perkin–
Elmer Autosystem 1020 equipped with a 25 m”0.25 mm Chrompack CP-
SIL 8CB column. He was used as carrier gas in an isothermic run
the regioretention from the very high level implied by
Scheme 5 to the moderate 30–40% observed in the experi-
5356
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 5352 – 5360

Pd-Catalyzed Allylic Alkylation
FULL PAPER
1
(
1808C, 10 min). IR spectra were recorded by using a Perkin–Elmer 1600
(EtOAc); H NMR (300 MHz, CDCl
3
): d=5.75 (dqt, J=15.3, 6.3, 1.1 Hz,
Series FTIR with AgCl plates.
1H; CH), 5.63 (dtq, J=15.3, 5.7, 1.2 Hz, 1H; CH), 4.47 (app. dk, J=5.7,
[
38]
1.1 Hz, 2H; CH ), 3.78 (sp, J=6.6 Hz, 1H; CH), 3.41 (brd, J=6.7 Hz,
(
1
(
E)-2-Buten-1-ol (1a):
LiAlH
4
(4.5 g, 117.7 mmol) was suspended in
2
1
3
CH
H; NH), 3.16 (sp, J=6.1 Hz, 1H; CH), 1.72 (app. dq, J=6.3, 1.2 Hz,
H; CH ), 1.12 (d, J=6.6 Hz, 6H; CH ), 1.09 ppm (d, J=6.3 Hz, 6H;
).
,2-dimethoxyethane and cooled to 08C in an ice-bath. 2-Butyn-1-ol
6.9 g, 98.5 mmol) in 1,2-dimethoxyethane (30 mL) was added over a
period of 15 min. After addition, the solution was left at RT for 80 h. The
reaction mixture was quenched with water (6 mL), and washed with 14%
NaOH (6 mL) and water (16 mL). The gray precipitate was filtered off
and washed with diethyl ether. Product 1a was isolated as a colorless oil
3
3
3
O-(3-Buten-2-yl)-N,N’-diisopropyl isourea (3b): The crude product was
purified by Kugelrohr distillation (0.8 mmHg, 508C). Yield 81%
1
(801 mg). R
f
=0.30/0.49 (EtOAc); H NMR (300 MHz, CDCl
3
): d=5.90
by distillation through a Vigreaux column (yield 48%, 3.4 g). R
f
=0.25
(ddd, J=17.3, 10.6, 5.2 Hz, 1H; CH), 5.42 (app. qdt, J=6.3, 5.2, 1.3 Hz,
1H; CH), 5.22 (app. dt, J=17.3, 1.5 Hz, 1H; CH), 5.05 (app. dt, J=10.6,
1.4 Hz, 1H; CH), 3.78 (sp, J=6.5 Hz, 1H; CH), 3.37 (brd, J=6.6 Hz,
(
hexane/EtOAc 5:1); b.p. 1208C (ref. [39] 118–1228C); GC (508C, iso-
thermic): t
R
=8.77 min, analysis did not show any traces of (Z)-isomer;
): d=5.76–5.64 (m, 2H; CH), 4.10–4.07 (m,
), 1.72 (app. dq, J=3.6, 0.6 Hz, 3H; CH ), 1.42 ppm (brs, 1H;
1
H NMR (300 MHz, CDCl
H; CH
OH).
3
1H; NH), 3.16 (sp, J=6.2 Hz, 1H; CH), 1.12 (d, J=6.4 Hz, 6H; CH
1.29 (d, J=6.4 Hz, 3H; CH ), 1.07 (d, J=6.2 Hz, 3H; CH ), 1.06 ppm (d,
); C NMR (75 MHz, CDCl ): d=151.6, 127.7, 126.1,
60.7, 46.2, 43.3, 24.3, 24.0, 23.9, 13.3 ppm; elemental analysis calcd (%)
for C11 O: C 66.62, H 11.18, N 14.13; found: C 66.85, H 10.88, N
4.17.
3
),
2
2
3
3
3
1
3
J=6.3 Hz, 3H; CH
3
3
[
40]
(
(
Z)-2-Buten-1-ol (1c):
2-Butyn-1-ol (8.0 g, 114.1 mmol), quinoline
and doped with
22 2
H N
8 mL), and Lindlar catalyst (5% Pd mixed with CaCO
3
1
Pb, 879 mg, 0.41 mmol) was mixed in dichloromethane (72 mL). The hy-
drogenation was performed with 1 atm H at RT under stirring for two
weeks. Consumption of H was 3100 mL (129 mmol). The catalyst was fil-
tered off, and 1c was isolated as a colorless oil by distillation through a
Vigreaux column (yield 61%, 5.0 g). R =0.23 (hexane/EtOAc 5:1); b.p.
19–1218C (ref. [41] 119.8–120.58C, 747 torr); GC (508C, isothermic):
O-((Z)-2-Buten-1-yl)-N,N’-diisopropyl isourea (3c): The use of an oil
2
pump (0.8 mmHg, RT) for 2 h was necessary to remove traces of solvent.
2
1
Yield 88% (873 mg).
CDCl
sp, 6.5 Hz, 1H; CH), 3.40 (brd, J=6.6 Hz, 1H; NH), 3.18 (sp, J=6.2 Hz,
1H; CH), 1.71 (dm, J=5.4 Hz, 3H; CH ), 1.11 ppm (app. t, J=5.9 Hz,
R
f
=0.30/0.49 (EtOAc); H NMR (300 MHz,
3
2
): d=5.70–5.55 (m, 2H; CH), 4.62 (brd, J=5.4 Hz, 2H; CH ), 3.76
f
(
1
1
t
R
=12.11 min, analysis did not show any (E)-product; H NMR
300 MHz, CDCl ): d=5.65–5.53 (m, 2H; CH), 4.18 (dm, J=4.5 Hz, 2H;
), 1.64 ppm (dm, J=5.1 Hz, 3H; CH ).
General procedure for synthesis of acetates 2a–c: The allylic alcohol
4.4 mmol) was dissolved in dichloromethane (10 mL) and cooled to 08C
under stirring. Acetyl chloride (0.34 mL, 4.8 mmol) and pyridine
0.38 mL, 4.7 mmol) were added and the reaction was monitored by
TLC. After completion of the reaction (after 15 min) water was added
3
1
3
1
4
2H; CH
3.2, 24.3, 24.2, 24.0, 19.7 ppm; elemental analysis calcd (%) for
O: C 66.62, H 11.18, N 14.13; found: C 66.34, H 10.91, N 13.94.
3 3
); C NMR (75 MHz, CDCl ) d=150.3, 139.4, 114.0, 69.8, 46.1,
(
3
CH
2
3
11 22 2
C H N
[
42]
[
45]
General procedure for synthesis of allylic trichloracetimidates 4a–c:
(
The allylic alcohol (1.96 mmol) was dissolved in dichloromethane and
cooled to ꢀ158C. Aqueous KOH (50%, 2 mL) and tetrabutylammonium
hydrogensulfate (8.8 mmol, 3 mg) were added as the temperature was
maintained at ꢀ158C, then trichloracetonitrile (2.34 mmol, 236 mL) was
added dropwise. The resulting solution was stirred for 30 min at ꢀ158C,
followed by 30 min stirring at RT. The solution was diluted with 5 mL di-
chloromethane and water (10–15 mL). The organic phase was isolated,
and the water phase was extracted twice with dichloromethane (7.5 mL).
(
(
(
2.5 mL, 08C), the organic phase was separated and washed with brine
10 mL, 08C) and water (10 mL, 08C). The organic phase was dried over
Na
colorless oil.
E)-2-Butene-1-yl acetate (2a): Yield 96% (452 mg). R
2 4
SO , filtered, and evaporated to yield the corresponding acetate as a
(
f
=0.54 (hexane/
): d=5.79 (dqt, J=15.3, 6.6,
.2 Hz, 1H; CH), 5.58 (dtq, J=15.3, 6.6, 1.5 Hz, 1H; CH), 4.49 (app. dk,
2 4
The combined organic phases were dried over Na SO , concentrated to
1
EtOAc 5:1); H NMR (300 MHz, CDCl
3
about 10 mL, and filtered through silica gel (2 cm). The silica gel was
eluted with 10–15 mL of dichloromethane, and the product was isolated
as a colorless oil by evaporation of solvent.
1
2 3
J=6.3, 1.0 Hz, 2H; CH ), 2.05 (s, 3H; CH ), 1.71–1.75 ppm (m, 3H;
CH
3
).
(
f
(E)-2-Butene-1-yl)trichloracetimidate (4a): Yield 75% (319 mg). R =
1
3
-Butene-2-yl acetate (2b): Yield 91% (432 mg).
R
f
=0.51 (hexane/
): d=5.85 (ddd, J=17.3, 10.5,
.3 Hz, 1H; CH), 5.35 (brk, J=6.3, 1.3 Hz, 1H; CH), 5.25 (app. dt, J=
7.4, 1.3 Hz, 1H; CH), 5.15 (app. dt, J=10.5, 1.2 Hz, 1H; CH), 2.07 (s,
H; CH ), 1.32 ppm (d, J=6.6 Hz, 3H; CH ).
0
1
6
.59 (hexane/EtOAc 4:1); H NMR (300 MHz, CDCl
H; NH), 5.90 (dqt, J=15.3, 6.3, 1.2 Hz, 1H; CH), 5.72 (dtq, J=15.3,
.2, 1.5 Hz, 1H; CH), 4.74 (app. dk, J=6.3, 1.1 Hz, 2H; CH ), 1.77 ppm
).
3-Butene-2-yl)trichloracetimidate (4b): Yield 14% (61.4 mg). R
3
): d=8.28 (brs,
1
EtOAc 5:1); H NMR (300 MHz, CDCl
3
6
1
3
2
(
dm, J=6.4 Hz, 3H; CH
3
3
3
(
f
=0.56
(
Z)-2-Butene-1-yl acetate (2c): Yield 98% (460 mg). R
f
=0.49 (hexane/
): d=5.74 (dqt, J=10.8, 6.9,
.7 Hz, 1H; CH), 5.56 (dtq, J=10.8, 6.9, 1.7 Hz, 1H; CH), 4.64 (app. dk,
1
(
hexane/EtOAc 4:1); H NMR (300 MHz, CDCl ): d=8.31 (brs, 1H;
3
1
EtOAc 5:1); H NMR (300 MHz, CDCl
1
3
NH), 5.95 (ddd, J=17.3, 10.6, 5.4 Hz, 1H; CH), 5.47 (brk, J=6.2 Hz,
1
1
H; CH), 5.37 (app. dt, J=17.3, 1.3 Hz, 1H; CH), 5.21 (app. dt, J=10.6,
.2 Hz, 1H; CH), 1.45 ppm (d, J=6.5 Hz, 3H; CH ).
2 3
J=6.6, 0.90 Hz, 2H; CH ), 2.07 (s, 3H; CH ), 1.71 ppm (ddt, J=6.9, 1.7,
3
0
3
.8 Hz, 3H; CH ).
(
f
(Z)-2-Butene-1-yl)trichloracetimidate (4c): Yield 79% (335 mg). R =
1
General procedure for synthesis of allylic isoureas 3a–c: The literature
procedure for the formation of dicyclohexylisoureas could also be used
0
1
1
.55 (hexane/EtOAc 4:1); H NMR (300 MHz, CDCl
H; NH), 5.82 (dq, J=10.8, 6.9 Hz, 1H; CH), 5.69 (dtq, J=10.8, 6.3,
.4 Hz, 1H; CH), 4.87 (brd, J=6.3 Hz, 2H; CH ), 1.76 ppm (brd, J=
3
): d=8.30 (brs,
[
43]
to form the diisopropylisoureas.
added to a mixture of CuCl (0.08 mmol, 7.5 mg) and diisopropyl carbodi-
imide (5.0 mmol, 781 mL) under stirring at RT. After about 12 h, CuCl
0.08 mmol, 7.5 mg) was added again. Before use, CuCl was purified as
The allylic alcohol (5.0 mmol) was
2
6.3 Hz, 3H; CH ).
3
ACHTREUNG
General procedure for allylic alkylation of acetates 2a–c: NaH (60% sus-
pension in oil, 1.0 mmol) was dissolved in THF (1 mL) and degassed with
argon. Diethyl methyl malonate (1.2 mmol, 200 mL) was added and the
solution was stirred for 30 min. Triphenylphosphine (13.2 mg, 0.05 mmol)
(
[
44]
described by Österlçf. The reaction mixture was stirred until the ab-
sorption due to diisopropyl carbodiimide (2110–2120 cm ) was no longer
apparent from the IR spectrum (usually 24 h). The reaction mixture was
ꢀ
1
was dissolved in THF (0.5 mL) and degassed with argon. [Pd
2
-
diluted with hexane (15 mL), washed with water (10 mL), 25% NH
3”5mL), and water (3”10 mL). The organic phase was dried over
Na SO , filtered, and evaporated by using a rotary evaporator. This pro-
3
(aq)
A
H
R
U
G
3
3
(
solution turned brownish. The solution was stirred for 30 min and the
allyl acetate (0.5 mmol) was added, followed by the sodium enolate of di-
ethyl methyl malonate (1 mmol) in dry, degassed THF (1 mL). After stir-
ring for 24 h at RT the solution was diluted with diethyl ether (10 mL)
and washed with 1m HCl (5 mL) and brine (5 mL). The organic phase
was dried over Na SO , filtered, and the solvent was removed. This gave
2
4
cedure yielded the desired product as a colorless oil. The isoureas pro-
duced two spots on a TLC plate, in spite of high purity. For this reason,
both R
O-((E)-2-Buten-1-yl)-N,N’-diisopropyl isourea (3a): This synthesis was
performed on 2-mmol scale. Yield 68% (270 mg). =0.25/0.44
f
values were reported for each isourea.
2
4
a
R
f
a pale-yellow oil, which was purified by flash column chromatography
Chem. Eur. J. 2006, 12, 5352 – 5360
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5357

P.-O. Norrby et al.
(
hexane/Et
2
O 9:1). The product was a mixture of three isomers and the
washed with 1m HCl (5 mL) and brine (5 mL). The organic phase was
components were identified and characterized by GC (1808C, isothermic)
and H NMR spectroscopy. The yield and composition from each allylic
alkylation is given below (Table 2).
dried over Na
pale-yellow oil, which was purified by flash column chromatography
(hexane/Et O 9:1). The product was a mixture of three isomers and the
2 4
SO , filtered, and the solvent was removed. This gave a
1
2
components were identified and characterized by GC (1808C, isothermic)
1
and H NMR spectroscopy. The yield and composition from each allylic
alkylation is given below (Table 4).
Table 2. Yield and product distributions from allylic alkylation of 2a–c.
Substrate
Isolated yield [%]
5a [%]
5b [%]
5c [%]
2
2
2
a
b
c
77
65
68
77
59
15
16
26
38
7
14
47
Table 4. Yield and product distributions from allylic alkylation of 1a–c
via 4a–c.
Substrate
Isolated yield [%]
5a [%]
5b [%]
5c [%]
1
1
1c
a
b
46
38
35
67
61
47
22
26
33
11
12
20
Diethyl-((E)-2-butene-1-yl)methyl malonate (5a): GC:
t
R
=5.50 min;
1
H NMR (300 MHz, CDCl
3
): d=5.53 (dqt, J=15.1, 6.3, 1.1 Hz, 1H; CH),
.31 (dtq, J=15.1, 7.3, 1.5 Hz, 1H; CH), 4.19 (q, J=7.1 Hz, 1H, CH),
.18 (q, J=7.1 Hz, 2H; CH ), 2.54 (app. dk, J=7.3, 1.1 Hz, 2H; CH),
.65 (app. dq, J=6.3, 1.1 Hz, 3H; CH ), 1.37 (s, 3H; CH ), 1.25 ppm (t,
).
Diethyl-(3-butene-2-yl)methyl malonate (5b): GC:
5
4
1
2
3
3
General procedure for allylic alkylation with chloride ions: Two experi-
ments was performed with each of the three acetates 2a–c, the only dif-
ference being the amount of added allyl chloride (0.5 and 5 mol%).
J=7.1 Hz, 6H; CH
3
t
R
=5.24 min;
): d=5.79 (ddd, J=17.1, 10.3, 8.1 Hz, 1H;
CH), 5.08 (ddd, J=17.1, 1.9, 1.1 Hz, 1H; CH), 5.04 (ddd, J=10.3, 1.9,
1
H NMR (300 MHz, CDCl
3
NaH (60% suspension in oil, 1.0 mmol) was dissolved in THF (1 mL)
and degassed with argon. Diethyl methyl malonate (1.2 mmol, 200 mL)
was added and the solution was stirred for 30 min. Triphenylphosphine
(13.2 mg, 0.05 mmol) was dissolved in THF (0.5 mL) and degassed with
0
1
1
.8 Hz, 1H; CH), 4.17 (q, J=7.1, 4H; CH
.0 Hz, 1H; CH), 1.35 (s, 3H; CH ), 1.25 (t, J=7.1 Hz, 6H, CH
.07 ppm (d, J=6.9 Hz, 3H; CH ).
2
), 3.01 (app. k, J=7.0, 1.9,
3
3
),
3
A
H
R
U
G
3 3
argon. [Pd2(dba) ]·CHCl (13 mg, 0.0125 mmol) was then added, during
Diethyl-((Z)-2-butene-1-yl)methyl malonate (5c): GC:
t
R
=5.76 min;
): d=5.62 (dqt, J=10.9, 6.8, 1.5 Hz, 1H; CH),
.30 (1H, dtq, J=11.0, 7.5, 1.8 Hz), 4.18 (q, J=7.1 Hz, 4H; CH ), 2.64
ddq, J=7.6, 1.4, 0.8 Hz, 2H; CH ), 1.63 (dq, J=6.0, 0.9 Hz, 3H; CH ),
), 1.25 ppm (t, J=7.1 Hz, 6H; CH ).
which the solution turned brownish. The solution was stirred for 30 min
and allyl chloride (0.0025 or 0.025 mmol) was added followed by 10 min
of stirring. The allylic acetate (0.5 mmol) was added, followed by the
sodium enolate of diethyl methyl malonate (1 mmol) in dry, degassed
THF (1 mL). After stirring for 24 h at RT the solution was diluted with
diethyl ether (10 mL) and washed with 1m HCl (5 mL) and brine (5 mL).
1
H NMR (300 MHz, CDCl
3
5
2
(
2
3
1
.39 (s, 3H; CH
3
3
General procedure for allylic alkylation of isoureas 3a–c: NaH (60% sus-
pension in oil, 24 mg, 0.6 mmol) was dissolved in THF (0.3 mL). The sol-
ution was degassed with argon. Diethyl methyl malonate was added
The organic phase was dried over Na
removed. This gave a pale-yellow oil, which was purified by flash column
chromatography (hexane/Et O 9:1) in the case of 0.5 mol% allyl chlo-
ride. The product was a mixture of three isomers and the components
2 4
SO , filtered, and the solvent was
(
120 mL, 0.70 mmol) and the solution was stirred for 30 min. [Pd
(dba) ]·CHCl (7.8 mg, 0.0075 mmol) and triphenylphosphine (7.9 mg,
.03 mmol) were dissolved in THF (0.6 mL), which gave a brownish solu-
2
-
2
AHCTREUNG
A
C
H
T
R
E
U
N
G
3
3
were identified and characterized by GC (1808C, isothermic) and
0
1
H NMR spectroscopy. After the reaction was run with 5 mol% allyl
tion. The solution was purged with argon and stirred for 30 min. The al-
lylic isourea (0.3 mmol) was added followed by the sodium enolate of di-
ethyl methyl malonate (0.6 mmol) in dry, degassed THF (0.3 mL). The
solution was stirred for 24 h (RT), then diluted with diethyl ether
chloride the product distribution was determined by GC on the crude
product. The yield and composition from both experiments are given
below (Table 5).
(
10 mL) and washed with 1m HCl (5 mL) and brine (5 mL). The organic
phase was dried over Na SO , filtered, and the solvent was removed. This
gave a pale-yellow oil, which was purified by flash column chromatogra-
phy (hexane/Et O 9:1). The product was a mixture of three isomers and
2
4
Table 5. Yield and product distributions from allylic alkylation of 2a–c in
the presence of chloride ions.
2
Substrate
Isolated yield [%]
5a [%]
5b [%]
5c [%]
the components were identified and characterized by GC (1808C, isother-
[
a]
1
2a
49
59
62
–
–
–
76
36
45
81
43
53
16
51
21
17
46
22
8
mic) and H NMR spectroscopy. The yield and composition from each al-
[
a]
2
2
2
2
2
b
13
34
2
11
25
lylic alkylation is given below (Table 3).
[
a]
c
a
b
[
b]
[
b]
Table 3. Yield and product distributions from allylic alkylation of 3a–c.
[
b]
c
Substrate
Isolated yield [%]
5a [%]
5b [%]
5c [%]
[
a] 0.5 mol% allyl chloride. [b] 5 mol% allyl chloride.
3
3
3
a
b
c
57
61
54
81
55
52
17
30
30
3
15
18
Iterative determination of terminal:internal reactivity ratios: The ratio of
internal-to-terminal attack on each intermediate is denoted f. In the com-
plete absence of regioretention (the initial assumption), we have two
such ratios, one for the syn intermediate, f
s
, and one for anti intermediate,
General procedure for alkylation of in situ generated trichloroacetimi-
dates 4a–c: The allylic alcohol (36.1 mg, 0.5 mmol), trichloroacetonitrile
f
a
. Furthermore, we use brackets with indices “t” to label reactions start-
ing from trans substrates 2a–4a, “c” for cis substrates 2c–4c, and “i” for
internal substrates 2b–4b. We obtain initial estimates of all fractions f
by assuming complete stereoretention, that is, by assuming that trans sub-
strates react completely via syn complexes, and cis substrates via anti
complexes:
(
72.2 mg, 0.5 mmol), and NaH (60% suspension in oil, 40 mg, 1.0 mmol)
were dissolved in dry THF (1.0 mL). Triphenylphosphine (13.3 mg,
.05 mmol) and [Pd (dba) ]·CHCl (13.1 mg, 0.0125 mmol) were dissolved
0
0
2
A
C
H
T
R
E
U
N
G
3
3
in dry THF (0.5 mL), and the solution was degassed with argon, during
which it turned brownish. After stirring for 30 min the solution contain-
ing the allylic imidate was added. After another 30 min of stirring diethyl
methyl malonate (202.6 mg, 1.16 mmol) was added. The solution was stir-
red for 24 h (RT) and then diluted with diethyl ether (10 mL) and
0
fs ¼ ½5b=5aŠ
t
0
a
f
¼ ½5b=5cŠ
c
5358
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 5352 – 5360

Pd-Catalyzed Allylic Alkylation
FULL PAPER
We know that the final proportion of branched product 5b arises both
from syn and anti complexes, whereas trans product 5a must arise exclu-
sively from the syn complex, and cis product 5c can come from the anti
complex only. Thus, for all reactions:
Vyskocil, P. Kocovsky, Chem. Eur. J. 2000, 6, 4348–4357; c) G. C.
Lloyd-Jones, Synlett 2001, 2, 161–183; d) I. J. S. Fairlamb, G. C.
Lloyd-Jones, S. Vyskocil, P. Kocovsky, Chem. Eur. J. 2002, 8, 4443–
4453; e) L. Gouriou, G. C. Lloyd-Jones, S. Vyskovil, P. Kocovsky, J.
Organomet. Chem. 2003, 687, 525–537.
5
b ¼ f
s
5a þ f
a
5c
[10] J. W. Faller, N. Sarantopoulos, Organometallics 2004, 23, 2179–2185.
[
[
[
11] G. Poli, C. Scolastico, Chemtracts 1999, 12, 837–845.
12] B. M. Trost, R. C. Bunt, J. Am. Chem. Soc. 1996, 118, 235–236.
13] K. Vrieze in Dynamic Nuclear Magnetic Resonance Spectroscopy
For each set of three reactions, we then get a simple set of refinement
equations by using the ratios from the preceding iteration:
(
Eds.: L. M. Jackman, F. A. Cotton), Academic Press, New York,
1975.
[14] a) M. Ogasawara, K.-i. Takizawa, T. Hayashi, Organometallics 2002,
1, 4853–4861; b) J. W. Faller, J. C. Wilt, Organometallics 2005, 24,
nþ1
n
f
f
¼ ½ð5b ꢀ fa 5cÞ=5aŠ
t
s
nþ1
n
¼ ½ð5b ꢀ fs 5aÞ=5cŠ
c
a
2
The above expressions converge to the final ratios in a few iterations.
With the final ratios in hand, the expected amount of branched product
can now be calculated also for the reaction employing branced substrates
5076–5083.
[15] G. Poli, C. Scolastico, Chemtracts 1999, 12, 822–836.
[16] a) B. Škermark, S. Hansson, A. Vitagliano, J. Am. Chem. Soc. 1990,
112, 4587–4588; b) M. Sjçgren, S. Hansson, P.-O. Norrby, B. Šker-
mark, M. E. Cucciolito, A. Vitagliano, Organometallics 1992, 11,
2
b–4b. The excess observed branched-to-branched reactivity x
b
is then
given by:
3
954–3964; c) P.-O. Norrby, B. Škermark, F. Hæffner, S. Hansson,
x
b
¼ ½ðf
s
5a þ f
a
5c ꢀ 5bÞ=ð5a þ 5b þ 5cފ
i
M. Blomberg, J. Am. Chem. Soc. 1993, 115, 4859–4867.
17] a) C. Amatore, A. Jutand, M. A. MꢁBarki, G. Meyer, L. Mottier,
Eur. J. Inorg. Chem. 2001, 873–880; b) A. Jutand, Eur. J. Inorg.
Chem. 2003, 2017–2040.
18] a) J. Sprinz, M. Kiefer, G. Helmchen, M. Reggelin, G. Huttner, O.
Walter, L. Zsolnai, Tetrahedron Lett. 1994, 35, 1523–1526; b) J. M.
Brown, D. I. Hulmes, P. J. Guiry, Tetrahedron 1994, 50, 4493–4506;
c) P. E. Blçchl, A. Togni, Organometallics 1996, 15, 4125–4132; d) B.
Goldfuss, U. Kazmeier, Tetrahedron 2000, 56, 6493–6496.
[
[
With this definition, x
tion in the reaction. In all experiments, x is found to be positive, al-
though in the experiments without chloride, it is small, possibly within
experimental uncertainty (2–7%). In the experiments with added chlo-
b
(in %) is a measure of the degree of regioreten-
b
A
C
H
T
R
E
U
N
G
ride, it is significant, 33–40%.
[
1] a) B. M. Trost, M. L. Crawley, Chem. Rev. 2003, 103, 2921–2943;
b) J. Tsuji in Handbookof Organopalladium Chemistry for Organic
Synthesis, Vol. 2 (Eds.: E.-I. Negishi, A. de Meijere), John Wiley,
New York, 2002, pp. 1669–1688; c) L. Acemoglu, J. M. J. Williams
in Handbookof Organopalladium Chemistry for Organic Synthesis,
Vol. 2 (Eds.: E.-I. Negishi, A. de Meijere), John Wiley, New York,
[
19] Lloyd-Jones and co-workers use the terms stereochemical conver-
g
gence (sc) and global enantiomeric excess (ee ) to quantify memory
[9]
effects, . Faller and Sarantopoulos use the terms regiochemical-
[10]
and stereochemical-memory effect , but with definitions differing
slightly from our use of regioretention and stereoretention.
20] K. Fagnou, M. Lautens, Angew. Chem. 2002, 114, 26–49; Angew.
Chem. Int. Ed. 2002, 41, 26–47; see also refs. [18, 19d].
21] S. Hansson, P.-O. Norrby, M. P. T. Sjçgren, B. Škermark, M. E. Cuc-
ciolito, F. Giordani, A. Vitagliano, Organometallics 1993, 12, 4940–
[
[
2
002, pp. 1689–1706; d) A. Pfaltz, M. Lautens in Comprehensive
Asymmetric Catalysis I–III (Eds.: E. N. Jacobsen, A. Pfaltz, H. Ya-
mamoto), Springer Verlag, Berlin, 1999, pp. 833–884; e) G. Helm-
chen, J. Organomet. Chem. 1999, 576, 203–214; f) B. M. Trost, D. L.
Van Vranken, Chem. Rev. 1996, 96, 395–422; g) G. Consiglio, R. M.
Waymouth, Chem. Rev. 1989, 89, 257–276.
4
948.
[
[
[
22] We note that, because we are measuring the competition between
unimolecular and bimolecular reactions, the individual measure-
ments are sensitive to the exact concentrations and the temperatures
employed.
23] a) E. PeÇa-Cabrera, P.-O. Norrby, M. Sjçgren, A. Vitagliano, V. de-
Felice, J. Oslob, S. Ishii, B. Škermark, P. Helquist, J. Am. Chem.
Soc. 1996, 118, 4299–4313; b) J. D. Oslob, B. Škermark, P. Helquist,
P.-O. Norrby, Organometallics 1997, 16, 3015–3021.
[
2] a) B. M. Trost, J. Dudash, Jr., E. J. Hembre, Chem. Eur. J. 2001, 7,
1
619–1629; b) M. Kranenburg, P. C. J. Kamer, P. W. N. M. van Leu-
wen, Eur. J. Inorg. Chem. 1998, 25–27; c) R. J. van Haaren, H. Oev-
ering, B. B. Coussens, G. P. F. van Strijdonck, J. N. H. Reek, P. C. J.
Kamer, P. W. N. M. van Leuwen, Eur. J. Inorg. Chem. 1999, 1237–
1
241; d) R. J. van Haaren, G. P. F. van Strijdonck, H. Oevering,
J. N. H. Reek, P. C. J. Kamer, P. W. N. M. van Leuwen, Eur. J. Inorg.
Chem. 2001, 837–843; e) P. Dierkes, S. Ramdeehul, L. Barloy, A. D.
Cian, J. Fischer, P. C. J. Kamer, P. W. N. M. van Leeuwen, J. A.
Osborn, Angew. Chem. 1998, 110, 3299–3301; Angew . Chem. Int.
Ed. Engl. 1998, 37, 3116–3118; f) S. Ramdeehul, P. Dierkes, R.
Aguado, P. C. J. Kamer, P. W. N. M. van Leeuwen, J. A. Osborn,
Angew. Chem. 1998, 110, 3302–3304; Angew. Chem. Int. Ed. 1998,
24] U. Kazmeier, F. L. Zumpe, Angew. Chem. 2000, 112, 805–807;
Angew. Chem. Int. Ed. 2000, 39, 802–804.
[25] a) T. M. Pedersen, E. L. Hansen, J. Kane, T. Rein, P. Helquist, P.-O.
Norrby, D. Tanner, J. Am. Chem. Soc. 2001, 123, 9738–9742; b) D.
Strand, T. Rein, Org. Lett. 2005, 7, 199–202; c) D. Strand, T. Rein,
Org. Lett. 2005, 7, 2779–2781; d) D. Strand, P.-O. Norrby, T. Rein, J.
Org. Chem. 2006, 71, 1879–1891.
3
7, 3118–3121.
[
[
[
3] J.-C. Fiaud, J. L. Malleron, Tetrahedron Lett. 1981, 22, 1399–1402.
4] B. M. Trost, N. R. Schmuff, Tetrahedron Lett. 1981, 22, 2999–3000.
5] a) P. R. Auburn, P. B. Mackenzie, B. Bosnich, J. Am. Chem. Soc.
[26] R. Schobert, S. Siegfried, Synlett 2000, 5, 686–688.
[27] G. C. Lloyd-Jones, S. C. Stephen, Chem. Commun. 1998, 2321–2322.
[28] N. Solin, J. Kjellgren, K. J. Szabꢂ, Angew. Chem. 2003, 115, 3784–
3785; Angew. Chem. Int. Ed. 2003, 42, 3656–3658.
1
985, 107, 2033–2046; b) P. B. Mackenzie, J. Whelan, B. Bosnich, J.
Am. Chem. Soc. 1985, 107, 2046–2054.
[29] a) M. Ahlquist, G. Fabrizi, S. Cacchi, P.-O. Norrby, Chem. Commun.
[
[
6] A. J. Blacker, M. L. Clarke, M. S. Loft, J. M. J. Williams, Org. Lett.
2
005, 4196–4198; b) L. J. Goossen, D. Koley, H. L. Hermann, W.
1
999, 1, 1969–1971.
7] a) T. Hayashi, A. Yamamoto, T. Hagihara, J. Org. Chem. 1986, 51,
23–727; b) T. Hayashi, M. Kawatsura, Y. Uozomi, J. Am. Chem.
Thiel, Organometallics 2005, 24, 2398–2410; c) L. J. Goossen, D.
Koley, H. L. Hermann, W. Thiel, Organometallics 2006, 25, 54–67;
d) M. Ahlquist, P. Fristrup, D. Tanner, P.-O. Norrby, Organometallics
7
Soc. 1998, 120, 1681–1687.
8] a) M. P. T. Sjçgren, S. Hansson, B. Škermark, A. Vitagliano, Or-
2
006, in press; for related studies, see: e) H. M. Senn, T. Ziegler, Or-
[
ganometallics 2004, 23, 2980–2988; f) S. Kozuch, C. Amatore, A.
Jutand, S. Shaik, Organometallics 2005, 24, 2319–2330.
A
C
H
T
R
E
U
N
G
ganometallics 1994, 13, 1963–1971; b) M. P. T. Sjçgren, PhD Thesis,
The Royal Institute of Technology, Stockholm, Sweden, 1993.
[
30] H. Hagelin, B. Škermark, P.-O. Norrby, Chem. Eur. J. 1999, 5, 902–
[
9] a) G. C. Lloyd-Jones, S. C. Stephen, Chem. Eur. J. 1998, 4, 2539–
9
09.
2
549; b) G. C. Lloyd-Jones, S. C. Stephen, M. Murray, C. P. Butts, S.
Chem. Eur. J. 2006, 12, 5352 – 5360
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
5359

P.-O. Norrby et al.
[
31] a) P.-O. Norrby, M. M. Mader, M. Vitale, G. Prestat, G. Poli, Or-
ganometallics 2003, 22, 1849–1855; b) D. Madec, G. Prestat, E. Mar-
models in general, see: C. Cramer, Essentials of Computational
Chemistry: Theories and Models, Wiley, New York, 2002.
[37] W. C. Still, M. Kahn, A. Mitra, J. Org. Chem. 1978, 43, 2923–2925.
[38] S. E. Denmark, M. A. Harmata, K. S. White, J. Org. Chem. 1987, 52,
4031–4042.
ACHTREUNG
tini, P. Fristrup, G. Poli, P.-O. Norrby, Org. Lett. 2005, 7, 995–998.
32] F. Delbecq, C. Lapouge, Organometallics 2000, 19, 2716–2723.
33] a) C. Lee, W. Yang, R. G. Parr, Phys. Rev. B 1988, 37, 785–789;
b) A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652; c) P. J. Ste-
phens, F. J. Devlin, C. F. Chabalowski, M. J. Frisch, J. Phys. Chem.
[
[
[39] M. Ratier, M. Perꢃyre, A. G. Davies, R. Sutcliffe, J. Chem. Soc.
Perkin Trans. 2 1984, 1907–1915.
[
[
[
40] H. Lindlar, R. Dubuis, Org. Synth. 1966, 46, 89–92.
41] L. F. Hatch, J. Org. Chem. 1963, 28, 2400–2403.
42] J. H. Bateson, A. M. Quinn, T. C. Smale, R. Southgate, J Chem. Soc.
Perkin Trans. 1 1985, 2219–2234.
1
994, 98, 11623–11627.
[
34] LACVP* uses the 6–31G* basis set for all light elements, and the
Hay-Wadt ECP and basis set for Pd: P. J. Hay, W. R. Wadt, J. Chem.
Phys. 1985, 82, 299–310.
35] Jaguar 4.2, Schrçdinger, Portland, Oregon, 2000; for the most recent
version, see: http://www.schrodinger.com.
[
[
[
43] R. Schobert, S. Siegfried, Synlett 2000, 5, 686–688.
44] J. Österlçf, Acta Chem. Scand. 1950, 4, 374–385.
45] V. J. Patil, Tetrahedron Lett. 1996, 37, 1481–1484.
[
[
36] a) B. Marten, K. Kim, C. Cortis, R. A. Friesner, R. B. Murphy, M. N.
Ringnalda, D. Sitkoff, B. Honig, J. Phys. Chem. 1996, 100, 11775–
Received: February 3, 2006
Published online: April 25, 2006
1
1788; b) For an illuminating discussion about implicit solvation
5360
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2006, 12, 5352 – 5360