Mechanism of the Pd-catalyzed decarboxylative allylation of α-imino esters: Decarboxylation via free carboxylate ion

Article abstract of DOI:10.1002/chem.201201425

The Pd-catalyzed decarboxylative allylation of α-(diphenylmethylene) imino esters (1) or allyl diphenylglycinate imines (2) is an efficient method to construct new C(sp³)-C(sp³) bonds. The detailed mechanism of this reaction was stud

Full text of DOI:10.1002/chem.201201425

FULL PAPER
DOI: 10.1002/chem.201201425
Mechanism of the Pd-catalyzed
N
ACHTUNGTRENaNGNU -imino Esters:
Zhe Li,[a] Yuan-Ye Jiang,[b] Andrew A. Yeagley,[c] James P. Bour,[c] Lei Liu,[b, d]
Jason J. Chruma,*[c, e] and Yao Fu*[b]
Abstract: The Pd-catalyzed decarboxy-
lative allylation of a-(diphenylmethyle-
ne)imino esters (1) or allyl diphenylgly-
tion barrier of +9.1 kcalmolꢀ1. The fol-
lowing rate-determining decarboxyla-
tion proceeds via a solvent-exposed a-
imino carboxylate anion rather than an
O-ligated allylpalladium carboxylate
with an activation barrier of +22.7 kcal
molꢀ1. The 2-azaallyl anion generated
by this decarboxylation attacks the face
of the allyl ligand opposite to the
Pd center in an outer-sphere process to
produce major product 3, with a lower
activation barrier than that of the
minor product 4. A positive linear
Hammett correlation [1=1.10 for the
PPh3 ligand] with the observed regiose-
lectivity (3 versus 4) supports an outer-
sphere pathway for the allylation step.
When Pd combined with the bis(diphe-
nylphosphino)butane (dppb) ligand is
employed as a catalyst, the decarboxy-
lation still proceeds via the free carbox-
ylate anion without direct assistance of
the cationic Pd center. Consistent with
experimental observations, electron-
withdrawing substituents on 2 were cal-
culated to have lower activation barri-
ers for decarboxylation and, thus, ac-
celerate the overall reaction rates.
cinate imines (2) is an efficient method
3
ꢀ
to construct new C
(sp ) C
(sp3) bonds.
The detailed mechanism of this reac-
tion was studied by theoretical calcula-
tions [ONIOM(B3LYP/LANL2DZ+
p:PM6)] combined with experimental
observations. The overall catalytic
cycle was found to consist of three
steps: oxidative addition, decarboxyla-
tion, and reductive allylation. The oxi-
Keywords: carboxylate ions · decar-
boxylative allylations · density func-
tional calculations · palladium · re-
action mechanisms
dative addition of
1 to [(dba)Pd-
ACHTUNGTRENNUNG(PPh3)2] (dba=dibenzylideneacetone)
produces an allylpalladium cation and
a carboxylate anion with a low activa-
3
Introduction
(sp ) C
(sp3) bonds.[1] Similar to other decarboxylative cross-
ꢀ
coupling reactions,[2] DcA reactions utilize ubiquitous car-
boxylate derivatives, and produce nonflammable and non-
toxic CO2 as the only by-product. This feature compares fa-
vorably with the cross-coupling reactions of aryl/alkyl hal-
ides with organometallic nucleophiles, a process which inevi-
tably produces potentially hazardous metal halides as stoi-
chiometric by-products.[3] Furthermore, as a subset of the
Tsuji-Trost allylation,[3e,4] DcA reactions do not require stoi-
chiometric preformed organometallic reagents or exogenous
bases. Accordingly, DcA reactions typically proceed under
neutral conditions; both the nucleophile and electrophile
are generated in minute concentrations commensurate with
the amount of catalyst present. Recent advances in DcA
transformations include the enantioselective DcA of allyl b-
ketoesters[5] and allyl vinyl carbonates,[6] the development of
complementary Ru-[7] and Ir-catalyzed transformations,[8]
and the extension of the substrate scope to include non-eno-
late nucleophiles.[9] In this last regard, Tunge[10] and
Chruma[11] independently reported a complimentary base-
free[12] Pd-catalyzed DcA process to synthesize homoallylic
imines from allyl a-imino esters (Scheme 1). Homoallylic
imines are an important building block for a large number
of natural products including aza-heterocycles[13] and alka-
loids.[14] The Tunge approach employs benzophenone imines
of various allyl a-amino esters (1) while the Chruma proto-
col utilizes a variety of allyl diphenylglycinate imines (2).
Palladium-catalyzed decarboxylative allylation reactions
(DcA) are a versatile tool for the construction of new C-
[a] Dr. Z. Li
Department of Applied Chemistry, China Agricultural University
Beijing 100193 (P. R. China)
[b] Y.-Y. Jiang, Prof. L. Liu, Prof. Y. Fu
Department of Chemistry
University of Science and Technology of China
Hefei 230026 (P. R. China)
Fax : (+86)551-3606689
E-mail: fuyao@ustc.edu.cn
[c] Dr. A. A. Yeagley, J. P. Bour, Prof. J. J. Chruma
Department of Chemistry, University of Virginia
P.O. Box 400319, Charlottesville, VA 22904-4319 (USA)
Fax : (+1)434-924-3710
E-mail: jjc5p@virginia.edu
[d] Prof. L. Liu
Department of Chemistry, Tsinghua University
Beijing 100084 (P. R. China)
[e] Prof. J. J. Chruma
Current address: College of Chemistry
Sichuan University, No. 29, Wangjiang Road
Chengdu, Sichuan, 610064 (P. R. China)
Fax : (+86)26-85415886
E-mail: chruma@scu.edu.cn
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201201425.
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
1
&
ÞÞ
These are not the final page numbers!

Scheme 1. The Tunge–Chruma DcA of a-imino allyl esters.
Both transformations apparently proceed via the same puta-
tive 2-azaallyl anion and p-allyl PdII-intermediates to afford
a mixture of isomeric homoallylic imines 3 and 4 in moder-
ate to excellent combined yield, with the former benzophe-
none imine predominating. These reactions provide access
to substituted homoallylic imines under mild conditions with
CO2 as the only main by-product.
Unlike the considerable mechanistic studies for the DcA
of enolates,[6d,15] there remains to be a concerted effort
toward developing a mechanistic understanding for this new
decarboxylative allylation manifold, not to mention a thor-
ough explanation for the observed regioselectivity between
the major and minor products 3 and 4, respectively. Tunge
proposed a plausible mechanism that consists of three steps
(Scheme 2, Path 1): oxidative addition of substrate 1 to form
the allylpalladium cation and the a-imino carboxylate anion,
decarboxylation of the carboxy-
Scheme 2. Two possible mechanistic pathways.
late moiety with Pd coordinated
to the imine nitrogen, and re-
ductive elimination of allyl and
imino groups to form the final
product. It is also possible, how-
ever, that the decarboxylation
and reductive elimination adopt
an ionic pathway without re-
quiring PdII-imine coordination
(Scheme 2, Path 2). Mixed
mechanisms including charac-
teristics of both Path 1 and
Path 2 are also possible because
the ion pairs can be in fast
Scheme 3. Crossover during DcA.
equilibrium with the neutral in-
termediates throughout the cat-
alytic cycle. We observed com-
plete cross-over when imino esters 2e and 5 were jointly
submitted to standard Pd-catalyzed decarboxylative allyla-
tion conditions (Scheme 3), but this observation alone
cannot unambiguously distinguish between the two path-
ways presented in Scheme 2 because the ion pair of the al-
lylpalladium cation and the a-imino carboxylate anion can
also lead to complete cross-over before decarboxylation.
Thus, a full analysis of the structures and energies of the in-
termediates and transition states in these proposed mecha-
nisms is needed to determine the most favored mechanism.
Furthermore, many mechanistic questions for this DcA
process remain unanswered. For example, which step in the
process is rate-determining, why do electron withdrawing
substituents accelerate the reaction rate, and what factors
control the regioselectivity of allylation? The answers to
these and related questions should provide a clear mechanis-
tic vision for this reaction and help to improve the efficien-
cy, selectivity, and to extend the substrate scope.
Herein, we present a detailed mechanistic study of the
Tunge–Chruma DcA of allyl a-iminoesters carried out using
quantum chemistry calculations. The neutral and ionic
mechanistic manifolds are discussed thoroughly on the basis
&
2
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Pd-catalyzed Decarboxylative Allylation of a-imino Esters
FULL PAPER
ꢀ
of the calculated structures and the energetics of possible in-
termediates and transition states. Furthermore, the substitu-
ent effect, ligand effect, and the origin of regioselectivity are
discussed in detail. Our results are expected to have impor-
tant implications for the development of new DcA reactions
utilizing a-imino acid esters and other non-enolate nucleo-
philes.
tion/allylation, in which the new C C bond is formed be-
tween the a-imino anion and the allyl moiety to generate
the product and regenerate the Pd0 catalyst. We will discuss
these three steps separately in detail.
Oxidative addition: It is necessary to check the energy of
[(dba)PdACTHNUGRTNEUNG(PPh3)2] (6) because 6 is reported to be the major
complex of the catalyst[25] used in experimental studies.[10,11]
The free energy of 6 is 17.4 kcalmolꢀ1 higher than the h2-
complex 7 (Figure 2). Thus, 7 is set as the reference for the
free-energy profile of the catalytic cycle. Previous studies on
the oxidative addition of allyl esters to Pd0 showed that, de-
pending on ligands and substrates, the stable product could
be either the ion pair consisting of an allylpalladium cation
and a carboxylate anion[26] or the neutral carboxylate h1-al-
lylpalladium complex.[27] Thus, two possible intermediate
products were considered (Scheme 4). One product is an ion
Results and Discussion
Model reaction: Based on the substrate scope and catalysts
used in the reported experiments,[10,11] our computational
model of the catalytic cycle begins with the complex of
imine 1a (R=Ph) and [PdACHTNUGTRNEGNU(PPh3)2]. It is known that the full
sized ligand should be used to model the reliable free
energy profile of Pd-catalyzed reactions.[16] As a compromise
between credibility and computational efficiency, a two-lay-
ered ONIOM method was used to optimize the geometries
of the intermediates and the transition states.[17] In the
ONIOM calculation, the phenyl rings of the triphenylphos-
phine ligand, and the substrate were put into the lower layer
that is handled by the PM6
method,[18,19] and the remaining
atoms belong to the high layer
that is handled by B3LYP[20]
/
LANL2DZ+p[21]
method
(Figure 1). The assignment of
atoms for ONIOM layer re-
mains constant in the catalytic
cycle. To obtain reliable results,
single point energies were cal-
culated by full B3LYP method
with SDD basis set for Pd and
Figure 1. ONIOM layers as-
signment. The atoms in the
dashed area belong to the high
layer, and the atoms outside
the rectangle are in the low
layer.
6-311+GACTHUNGTRENUN(NG 2d,p) basis set for
other atoms. In the catalytic
cycle, solvent-exposed 2-azaallyl anions are possible inter-
mediates. The geometries and the corresponding energies of
ions in solution usually have significant differences with
those in the gas phase.[22] As a result, we chose to optimize
the geometries in THF solution with the IEF-PCM
model.[23] All the calculations were performed using the
Gaussian 09 software.[24]
Figure 2. A) Free-energy profile (in kcalmolꢀ1) of the oxidative addition
step. B) Optimized structures of selected stationary points along with key
bond lengths (in ꢁ). All H atoms are omitted for clarity.
In accord with previous
mechanistic studies of the Pd-
catalyzed allylic alkylation of
allyl enol carbonates and a-ke-
toesters,[6d,16] the catalytic cycle
was proposed to consist of
three steps: 1) oxidative addi-
tion, in which the allyl moiety
of the substrate transfers from
the substrate O to the Pd of the
catalyst, 2) decarboxylation, in
which the carboxylate moiety
dissociates to CO2 and 2-azaall-
yl anion, 3) reductive elimina- Scheme 4. Oxidative addition of 1a.
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
3
&
ÞÞ
These are not the final page numbers!

J. J. Chruma, Y. Fu et al.
pair consisting of the cationic allyl-palladium 8 and anionic
carboxylate 9. The other possible intermediate is the neutral
four-coordinated complex 10 with a carboxylate anion coor-
dinating to the Pd. Accordingly, two unique transition states
were considered for oxidative addition, that is, TS7–8 and
TS7–10. A viable transition state TS7–10 could not be generat-
ed, however, by placing the C and O atoms of 1a and Pd in
any bonding distance; this treatment always led to the disso-
ciation of 1a presumably due to the strong steric repulsion
between the two PPh3 ligands. On the other hand, we were
able to find the dissociative transition state, TS7–8. The free
energy barrier of TS7–8 was calculated to be +9.1 kcalmolꢀ1
relative to 7. The free energy of free ion pair (8+9) is
+3.8 kcalmolꢀ1. These calculations suggest that the oxida-
tive addition step is uphill in energy and reversible, which is
qualitatively consistent with previous studies.[28]
without an a-H in their respective DcA reactions.[10,11] De-
carboxylation could proceed from any of the three inter-
mediates 9, 10, or 12, and viable transition states were locat-
ed for each (Scheme 5). From the bisphosphine 10, the tran-
sition state TS10–13 has a free energy barrier as high as
+65.3 kcalmolꢀ1 relative to 7. Note that the decarboxylative
allylation of 1a (R=Ph) takes place at room temperature.[10]
Accordingly, a decarboxylation pathway involving TS10–13
can be excluded.
The transition state for decarboxylation of 12 is TS12–16
with a free energy barrier of +36.5 kcalmolꢀ1, which also is
prohibitively high for a reaction that proceeds at room tem-
perature (Figure 3). Another possible pathway for the decar-
The neutral complex 10 is in coordination equilibrium
with the ion pair, 8+9. The free energy of 10 is +30.5 kcal
molꢀ1, indicating that 10 is not favored in the equilibrium. It
is worth noting, however, that Stoltz and co-workers isolated
a neutral h1-allyl Pd-carboxylate species similar to 10, but
with a bidentate P,N ligand, and they suggest that this spe-
cies represents the resting state for the decarboxylative ally-
lation of an allyl a,a-dialkyl-b-ketoester.[27] Similarly, our
calculations indicate that the neutral h1-allyl Pd-carboxylate
with a bidentate bisphosphine ligand, bis(diphenylphosphi-
no)butane (dppb, compound 27) is energetically more acces-
sible than 10 (see below). The allyl group in 10 is h1-coordi-
nated to Pd due to steric hindrance with the two PPh3 li-
gands. This indicates that dissociation of one PPh3 ligand
from 10 to form the monophosphine allylpalladium 12
should be favored. Although the high calculated relative
free energy of 10 precludes direct transformation to 12, the
combination of 11 and 9 is a possible pathway to produce
12. The formation of 11 from 8 via PPh3 dissociation only
causes an increase in free energy of 1.2 kcalmolꢀ1. The dis-
sociation of one phosphine ligand from an allylpalladium in-
termediate is often encountered
Figure 3. A) Free-energy profile (in kcalmolꢀ1
states of decarboxylation of 12 (L=PPh3). B) Optimized structures of se-
lected stationary points along with key bond lengths (in ꢁ). All H atoms
are omitted for clarity.
) of possible transition
in the mechanisms for allylative
dearomatization reactions.[17a,29]
Indeed, the calculated free
energy of monophosphine and
carboxylate ligated species 12 is
ꢀ15.8 kcalmolꢀ1, relative to 7,
indicating that 12 is the most
stable intermediate after oxida-
tive addition. Crystal structures
of palladium complexes with
one phosphine, an allyl group
and a carboxylate ligand similar
to 12 have been reported.[30]
Decarboxylation: The decar-
boxylation step should occur
prior to allylation given that
both Tunge and Chruma suc-
cessfully employ substrates Scheme 5. Possible pathways for decarboxylation step (L=PPh3).
&
4
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Pd-catalyzed Decarboxylative Allylation of a-imino Esters
FULL PAPER
boxylation of 12 is similar to the proton-induced decarboxy-
lation of a-imino acids (Scheme 5).[31] O-Ligated 12 can iso-
merize to the zwitterionic complex 14 with the free energy
of +4.9 kcalmolꢀ1. Two isomeric transition states for the de-
carboxylation from 14, TS14–15a and TS14–15b, were investigat-
ed. The structure of TS14–15a is a five-membered ring involv-
ing Pd and the dissociating CO2 (Figure 3). The barrier of
TS14–15a relative to 12 is +41.4 kcalmolꢀ1, which is not possi-
ble for a reaction at room temperature. The free energy bar-
rier of TS14–15b is +33.2 kcalmolꢀ1 relative to 12, which is
lower than that of TS14–15a by 8.2 kcalmolꢀ1. The geometry
of TS14–15b is such that the dissociating CO2 and the Pd are
on opposite faces of the incipient 2-azaallyl anion species.
The transition state for decarboxylation from the solvent
separated anion 9, that is, TS9–18, was also examined. The ac-
tivation barrier for TS9–18 was determined to be +28.3 kcal
molꢀ1 relative to 12 (Figure 4). This barrier is 4.9 kcalmolꢀ1
with the p-allyl PdII cation 11 to produce intermediates 15,
16, and/or 17 (Scheme 5). Thus these four intermediates, to-
gether with bisphosphine complex 13, are expected to be in
equilibrium via fast ligand exchange. Among these inter-
mediates, 17 has the lowest calculated free energy (ꢀ9.8 kcal
molꢀ1) in solution. Given that decarboxylation will prove to
be the rate-limiting step (see below), the observed crossover
(Scheme 3) most likely occurs at the ion pair 8+9, prior to
decarboxylation.
Product formation: Similar to the Pd-catalyzed Tsuji-Trost
allyllation reaction,[4] there are two pathways to form the
final products. One is an inner-sphere pathway, in which the
2-azaallylic and allylic moieties undergo reductive elimina-
tion from the Pd center. The other is an outer-sphere path-
way, in which the 2-azaallyl anion dissociates from the
Pd center to form either a solvent separated or a contact ion
pair that then attacks the allyl ligand on the face opposite to
the Pd metal center (Scheme 6).
Figure 4. A) Free-energy profile (in kcalmolꢀ1) of decarboxylation step
(L=PPh3). B) Optimized structures of selected stationary points along
with key bond lengths (in ꢁ) and bond angle (in degree). All H atoms
are omitted for clarity.
Scheme 6. Possible pathways for product formation.
After a comprehensive search for possible pathways for
reductive elimination, three inner-sphere pathways were un-
covered (Figure 5). The core structure of TS17–19a is a seven-
membered ring, which is similar to the corresponding transi-
tion state predicted by Goddard III and co-workers for the
lower than that of TS14–15b. As a result, according to the cal-
culated free energy, TS9–18 is the most favored transition
state for decarboxylation. To be careful with our conclusion,
we also tried other solvation models to calculate the relative
free energy of TS14–15b and TS9–18, based on full geometries
optimized by the B3LYP methods (see Table S1 in the Sup-
porting Information). All of the examined methods show
that the anionic pathway for decarboxylation is favored. Ac-
tually, although it is reported that the decarboxylations of
allyl b-ketocarboxylates proceed via an O-ligated allyl palla-
dium complex,[32] the decarboxylation of diallyl 2-alkyl-2-ar-
ylmalonates is suggested to occur via a solvent separated ion
pair corresponding to a carboxylate anion and an allylpalla-
dium cation.[33]
Tsuji allylation of enolates employing
a P,N bidentate
PHOX ligand.[15a] The barrier of TS17–19a is only +11.2 kcal
molꢀ1 relative to 17. TS17–19b is a three-centered transition
state with a barrier of +26.5 kcalmolꢀ1. This transition state
was also found to have a high energy barrier in the study of
Goddard III.[15a] A six-centered transition state, TS17–19c was
also located, with a barrier of +27.4 kcalmolꢀ1. Among
these three transition states, the seven-centered TS17–19a with
the lowest free energy is the most likely for an inner-sphere
allylation pathway.
The intermediate 17 may dissociate to produce the free 2-
azaallyl anion 18 and the monophosphine allylpalladium
cation 11 as either a solvent separated or a contact ion pair.
While it is nontrivial to calculate an accurate energy value
The free energy of 8+18 is 5.2 kcalmolꢀ1 lower than 8+9,
although 18 is more unstable than 9. This is prompted by
the loss of solvated CO2. Intermediate 18 can recombine
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
5
&
ÞÞ
These are not the final page numbers!

J. J. Chruma, Y. Fu et al.
ion pair 11+18. This does not necessarily mean that the
energy of transition state TS18–19 is even lower than the reac-
tants. In fact, an energy scan along the increase of the form-
ꢀ
ing C C bond in TS18–19 leads to an energy minima (IP11–18
)
that is 2.5 kcalmolꢀ1 lower than TS18–19. Thus, IP11–18 can be
considered as a certain structure of the contact ion pair be-
tween 11 and 18. This result indicates that the solvent sepa-
rated ion pair of 11 and 18 is not a preferred mode for
anion alkylation in this reaction system. As a result, it is pre-
sumed that this outer-sphere allylation occurs via the con-
tact ion pair and not the solvent separated ions.
The 2-azaallyl anion could also attack the bisphosphine p-
allylpalladium cation 8; the free energy of this transition
state (TS18–3) is +3.1 kcalmolꢀ1, which is 8.4 kcalmolꢀ1
higher than its monophosphine counterpart TS18–19. The bar-
rier of TS18–19 is also 6.7 kcalmolꢀ1 lower than that of TS17–19a
in the inner-sphere pathway. Thus, according to the calculat-
ed activation barriers, the outer-sphere pathway proceeding
via the contact ion pair of monophosphine complex 11 and
2-azaallyl anion 18 toward allylation is favored. Attempts to
provide empirical support for the predominance of an outer-
sphere allylation manifold using a homochiral 1,3-disubsti-
tuted allyl ester have been repeatedly rebuffed by the negli-
gible reactivity of the requisite substrates under the reaction
conditions and the current lack of an appropriate chiral
ligand to catalyze the DcA with high enantioselectivity at
the newly generated stereogenic center.[35]
Figure 5. A) Free-energy profile (in kcalmolꢀ1) of three possible transi-
tion states for inner sphere reductive elimination (L=PPh3). B) Opti-
mized structures of selected stationary points along with key bond
lengths (in ꢁ). All H atoms are omitted for clarity.
for a contact ion pair,[34] the calculated value for the solvent
separated ion pair 11+18, ꢀ0.2 kcalmolꢀ1, represents an
upper energetic limit. Attack by the 2-azaallyl anion nucleo-
phile onto the allyl ligand opposite to the Pd center would
produce the product 3 (Figure 6). By decreasing the distance
of 11 and 18, a transition state for allylation was located as
TS18–19 with an activation barrier of +4.5 kcalmolꢀ1 relative
to the 17, but 5.1 kcalmolꢀ1 lower than the solvent separated
Regioselectivity - Hammett correlation studies and calcula-
tions: Both Tunge[10] and Chruma[11] reported that an addi-
tional allylation isomer 4 is obtained as a minor product.
From these initial reports, the ratio of major to minor allyla-
tion isomer (3/4, respectively) appears to depend on the
electronic nature of the R group in either imino esters 1 or
ꢀ
2. Note that the C C bond-forming reductive allylation step
is the same for both 1 in Tungeꢂs protocol and 2 in Chrumaꢂs
protocol (Scheme 1). We decided to investigate this phe-
nomenon systematically following the Chruma protocol,[11]
given that this strategy is more amenable to variation of the
R substituent. Accordingly, a series of allyl diphenylglyci-
nate imines 2 were subjected to [PdACTHNUTRGNEUNG(dba)2] (10 mol%) with
PPh3 (20 mol%) in CH2Cl2 at 358C overnight. The reaction
mixtures were concentrated and immediately analyzed by
1H NMR spectroscopy to determine the resultant isomeric
product ratio, 3/4 (Table 1). To accommodate the precision
1
limitations of H NMR spectroscopy,[36] the substrate scope
was limited to imines 2 that provided <20:1 isomeric ratios.
The resulting isomeric ratios were then compared to the
Hammett substituent constants (s)[37] for the various sub-
stituents on the aromatic imine moiety using a standard
Hammett plot, that is, logACTHNURGTNE(NUG 3/4) versus s (Figure 7). Under
these reaction conditions and constraints, a positive linear
Hammett correlation was observed with a reaction constant
(1) of 1.10. These results are in accord with a reaction mech-
anism in which regioselectivity is influenced by the stabiliza-
tion of a negative charge on the imine carbon, that is, the
carbon adjacent to the R group in imines 2. It is important
Figure 6. A) Free-energy profile (in kcalmolꢀ1) of product formation step
(L=PPh3). B) Optimized structures of selected stationary points along
with key bond lengths (in ꢁ). All H atoms are omitted for clarity.
&
6
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Pd-catalyzed Decarboxylative Allylation of a-imino Esters
FULL PAPER
Table 1. Comparison of Hammett constant (s) and observed isomeric
ratio (3/4) for decarboxylative allylation of imines 2.
(Scheme 7). Accordingly, production of 3 would be ex-
19
pected to occur much faster than 4, as was observed empiri-
cally. This regiochemistry can also be reasoned by the charg-
es on 2-azaallyl anions. The NBO charge[39] of C1 in 18 is
more negative than C3 by ꢀ0.151e, indicating that C1 is
more nucleophilic than C3 (Scheme 8). Thus the formation
of product 3 is favored. The charge difference between C1
and C3 in 18b (from substrate 2b) is ꢀ0.142e, which is
smaller than that of 18. This is why the product ratio of 3/4
from substrate 2b is lower than 2a. On the other hand, the
charge difference in 18 f is larger than that in 18, which is
consistent with a higher product ratio of 3/4 for 2 f versus
2a. In addition, the product ratio of 4,4’-dichlorobenzophe-
none derivative of phenyl glycine is 2:1,[10] which is consis-
tent with a smaller charge difference of 18g than 18. The
above data can be summarized as the following: electron-
withdrawing groups increase the negative charge on attach-
ed carbon to increase its nucleophilicity and the proportion
of the corresponding product isomer, whereas electron-do-
nating groups decrease the nucleophilicity of the attached
carbon and the proportion of corresponding product.
Substrate
Hammett constant (s)[a]
Isomeric ratio (3/4)[b,c]
2a (R=H)
0.00
ꢀ0.27
ꢀ0.20
0.12
0.23
0.24
4.1:1
1.9:1
2.1:1
5.3:1
6.2:1
6.7:1
7.1:1
2b (R=4-MeO)
2c (R=4-tBu)
2d (R=3-MeO)
2 f (R=4-Br)
2g [R=3,5-(MeO)2]
2h (R=3-PhO)
0.25
[a] Hammett constants (s) as reviewed previously.[37a] [b] Reaction condi-
tions A: Pd(dba)2 (10 mol%), PPh3 (20 mol%), CH2Cl2 (0.1m), 358C,
ACHTUNGTRENNUNG
10 h. [c] Determined by comparison of the integration values for the al-
lylic methylenes (CH2) of 3 and 4 in the corresponding 1H NMR spectra
of the concentrated reaction mixtures (CDCl3). All reported values rep-
resent an average of at least three runs.
to note that a significant posi-
tive deviation from linearity
was observed with electron-
withdrawing groups para to the
imine carbon (data not shown
due to lack of precision and ac-
curacy in detection of isomeric
ratios ꢁ20:1).[36] For example,
benzaldimine 2i (R=4-CO2Me,
s=0.45), is predicted to have
an isomeric ratio, 3i/4i, of
11.6:1. Instead, the observed
isomeric ratio was beyond the
limits of accurate detection by
1H NMR, ꢁ20:1. It is plausible
that as the stability of the 2-
azaallyl anion intermediate 18
increases, that is, the pKa of the
conjugate acid decreases, the al-
lylation mechanism switches to
an alternate pathway. Such
transitions between two differ-
ent mechanistic pathways de-
pending of the strength (hard
vs. soft) of the nucleophile were
described for other Pd-cata-
lyzed Tsuji-Trost reactions.[3e]
Benzaldimine 4 can be pro-
duced by attack at the diphe-
nylmethylene terminus of the 2-
azaallyl anion instead of the
less substituted position as in 3.
This transition state, TS18–20, has
Figure 7. Hammett correlation plot comparing isomeric ratio [log
tion constant, 1, corresponds to the slope of the best-fit line.
(3/4)] to Hammett constant (s).[37a] The reac-
ACHTUNGTRENNUNG
a
relative free energy of
ꢀ2.8 kcalmolꢀ1,
which
is
2.5 kcalmolꢀ1 higher than TS18– Scheme 7. Competitive formation of two products (L=PPh3).
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
7
&
ÞÞ
These are not the final page numbers!

J. J. Chruma, Y. Fu et al.
+22.7 kcalmolꢀ1 by full B3LYP
calculation, see below), is well
within the range expected for a
reaction that proceeds at ambi-
ent temperature.
Limited decarboxylation of
sodium salt 21-Na: Our calcula-
tions point out that decarboxy-
lation occurs via a free, solvent
Scheme 8. Charges on substituted 2-azaallyl anions.
separated anionic transition
Overall mechanism: Inspired by the above results, a free-
energy profile for the whole catalytic cycle can be construct-
ed, as depicted in Figure 8. The overall catalytic cycle con-
sists of three main steps: oxidative addition, decarboxyla-
tion, and reductive allylation. In the oxidative addition step,
the reactant complex 7 readily dissociates to produce the
carboxylate anion 9 and the cationic h3-allyl palladium 8.
The decarboxylation of solvent separated anion 9 via TS9–18
state without direct assistance
by the cationic PdII center. Burger and Tunge reported, how-
ever, that 21-Na alone did not undergo decarboxylation
unless an equivalent amount of a cationic p-allyl PdII species
was added (Scheme 9).[10] This observation does not necessa-
rily contradict our conclusions.
has a lower free energy barrier than the PdII-assisted TS12–16
,
TS14–15a or TS14–15b. After decarboxylation of 9 and rapid con-
tact ion pair formation, the resulting 2-azaallyl anion 18 at-
tacks the p-allyl PdII moiety from the other side of the
Pd center to produce the major isomeric product 3. The
minor isomer 4 is produced via TS18–20, which is calculated
to be 2.5 kcalmolꢀ1 higher than TS18–19 for R=Ph. It is
worth noting that while this difference in transition state en-
ergies correctly predicts which isomer will predominate, it is
greater than implied by the observed regioselectivity
(ca. 5:1). This is not surprising given the current limits to
precision inherent to our calculation methods. The highest
free energy in the catalytic cycle is observed for TS9–8. This
suggests that decarboxylation is the rate-determining step.
The calculated substituent effect for this step is consistent
with experimental observations (see below).[10,11] Additional-
ly, the calculated free energy for this rate-determining step,
Scheme 9. Role of cation in decarboxylation.
The activation barrier of decarboxylation of 21-anion is
only +18.4 kcalmolꢀ1, which means that the decarboxylation
should be fast at 1108C.[40] However, there are two factors
impeding the decarboxylation of 21-Na. First, 21-Na is not
soluble in toluene,[10] which causes the concentration of 21-
anion to be prohibitively low. In other words, the formation
of a solvent separated ion pair for 21-Na in toluene, which is
necessary for decarboxylation, is prohibitively disfavored.
Addition of superstoichiometric amounts of the allyl
ꢀ
+28.3 kcalmolꢀ1 relative to
6
(or more accurately,
PdII cation also introduces the BF4 anion, both of which
can bring 21-Na into solution.
Secondly, the reaction free
energy for the decarboxylation
of 21-anion is +7.6 kcalmolꢀ1.
According to the microscopic
reversibility principle, the re-
combination of CO2 with 22-
anion is much faster than decar-
boxylation. Thus 21-anion and
22-anion are in fast equilibrium
in solution (Scheme 10). The
stoichiometric reaction between
a
potassium 2-azaallyl anion
and CO2 to produce a potassi-
um a-imino carboxylate has
been reported previously.[41]
Additionally, the recombination
of CO2 with a carbanion in the
decarboxylation of mandelylth-
iamin was observed experimen-
Figure 8. Overall mechanism of the decarboxylation of 1a.
&
8
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Pd-catalyzed Decarboxylative Allylation of a-imino Esters
FULL PAPER
tive addition are anionic 9 and cationic 26.
Similar to the case with the monodentate ligand (PPh3),
two pathways for decarboxylation were considered. In the
first pathway, association of 9 and 26 produces the five-coor-
dinate complex 27. The free energy of 27 is 12.5 kcalmolꢀ1
higher than the solvent-separated ion pair of 9+26. This in-
dicates that the combination of 9 and 26 to form 27 is ener-
getically feasible for a reaction that proceeds at ambient
temperature. It should be noted that the corresponding
energy difference between 10 and 9+11 for the monoden-
tate (PPh3) ligand system is 25.5 kcalmolꢀ1. This implies that
bidentate ligands such as dppb stabilize h1-allyl Pd-carboxy-
late complexes better than monodentate ligands such as
PPh3. This is in line with reports by Stoltz et al. in which the
neutral 4-coordinate h1-allyl Pd-carboxylate complex with a
bidentate P,N ligand appears to be the resting state for the
decarboxylative allylation of an allyl a,a-dialkyl-b-ketoest-
er.[27] However, the free energy of 27 is 3.8 kcalmolꢀ1 higher
than the transition state TS9–18 for decarboxylation in the
ion pair pathway. Thus, the decarboxylation catalyzed by a
Pd/dppb complex still prefers an anionic pathway via TS9–18
to produce the corresponding ion pair of 18+26. The follow-
ing allylation between 18 and 26 proceeds via TS18–28 with a
barrier of +7.6 kcalmolꢀ1.
As with the monodentate ligand system, the rate deter-
mining step appears to be the decarboxylation of dissociated
anion 9 in the reaction catalyzed by Pd0/dppb. The activa-
tion barrier is calculated to be about 20 kcalmolꢀ1 lower
than the corresponding barrier of the decarboxylation cata-
lyzed by Pd0/PPh3. This implies that bisphosphine ligands
should be much more effective than the monophosphine
ligand PPh3 in the Pd-catalyzed decarboxylative allylation of
1a and, presumably 2a. An explanation for this significant
difference is that the solvent separated ion pair of 9 and 26
is the resting state in the reaction catalyzed by Pd0 with
Scheme 10. The decarboxylation of 21-anion and the effect of L2Pd-
ACHTUNGTRENNUNG
(allyl)+ (L=PPh3).
tally by Kluger et al.[42] The reaction free energy for the as-
sociation of 22-anion with allyl palladium, however, is
ꢀ23.0 kcalmolꢀ1. Thus when the cationic allylpalladium spe-
cies is added, the equilibrium will be driven towards the for-
mation of 23.[43] This is how the Pd “catalyzes” the decar-
boxylation of 21-Na.[44]
Previously, we noted that addition of equimolar amounts,
in relation to the Pd catalyst, of oxophilic Lewis acid, for ex-
ample, LiCl and ScACHTUNGTRENNUNG(OTf)3, completely inhibit the Pd-cata-
lyzed DcA of allyl esters 2 at room temperature.[11b] This is
in accord with the above discussion since these Lewis acids
can compete with the less oxophilic allypalladium cation for
the carboxylate anion, thus forming a tighter (less solvent
exposed) ion pair. We have since discovered that the inhibi-
tory effects of oxophilic Lewis acids can be overcome by
heating the reaction mixture (1508C in DMF for 2e).[35]
Decarboxylative allylation employing a bidentate ligand
(dppb): Chelating bisphosphine ligands such as dppe, dppb,
and dppf[10,11] are also able to catalyze the decarboxylation
of allyl a-imino esters with Pd0. To study how the bidentate
ligands affect the rate-determining decarboxylation step, we
reinvestigated our reaction mechanism calculations using
dppb as a model bidentate ligand. The proposed mechanism
for decarboxylation with dppb as a ligand consists of the
same three steps explored for the monodentate PPh3: oxida-
tive addition, decarboxylation, and reductive allylation.
In reported protocols the cat-
alyst precursors are mixtures of
Pd2dba3 and dppb.[10,11] The free
energy of (dba)PdACHTUNGTRENNUNG(dppb) (24) is
+4.8 kcalmolꢀ1 greater than the
reactant complex 25 (Figure 9).
The following transition state
for oxidative addition, TS25–9
,
has a free energy barrier of
+6.0 kcalmolꢀ1. This value is
3.1 kcalmolꢀ1 lower than TS7–8
using the monodentate ligand
PPh3. The reason is presumably
that the dppb ligand better sta-
bilizes the accumulating posi-
tive charge on the Pd center;
the Mulliken charges on Pd in
TS7–8 and TS25–9 are +1.265e
and +0.843e, respectively. The
resultant products after oxida-
Figure 9. Mechanism of decarboxylation catalyzed by Pd0 with dppb ligand.
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
9
&
ÞÞ
These are not the final page numbers!

J. J. Chruma, Y. Fu et al.
Table 2. Free energies for intermediates and transition states of substitut-
dppb, whereas the neutral complex 12 is dominant in the
ed imino acid esters (kcalmolꢀ1).
catalytic flux of Pd0 with PPh3 ligand.
¼
Substrate
IN1
IN2
TSdeca
DG
Texptl
Impact of allyl a-imino ester substituents: Because the rate-
determining step is the decarboxylation of the a-imino car-
boxylates, the free-energy barriers of different substrates
should reflect their relative propensities toward the DcA re-
action. To study the substituent effect on decarboxylation,
we calculated the barriers to decarboxylation for a series of
substrates (Figure 10).[44]
ꢀ0.3
+11.1
+22.3
+22.7
258C
1a
29
+1.4
+2.1
+12.1
+13.6
+23.3
+32.0
+23.3
+32.0
408C
1108C
1b
2a
30
ꢀ1.6
ꢀ1.6
+7.7
+9.4
+17.4
+19.0
+19.0
+20.6
238C
408C
ꢀ0.6
ꢀ2.6
+8.8
+5.6
+19.5
+12.6
+20.1
+15.2
238C
238C
Figure 10. Free-energy profile for decarboxylation step of substituted
allyl a-imino ester. (L=PPh3).
2b
2j
Although specific reaction rates were not reported experi-
mentally, the relative activities could be estimated from the
requisite reaction temperatures. The 2-Me substituent on
the allyl moiety of substrate 29 increases the activation bar-
rier to +23.3 kcalmolꢀ1 (Table 2). This is consistent with the
experimental observation that the decarboxylative allylation
of 29 requires an elevated temperature of 408C.[10] The
benzyl substituent of substrate 1b increases the activation
barrier to +32.0 kcalmolꢀ1, that is, 9.3 kcalmolꢀ1 higher than
1a. This explains why phenylalanine 1b requires a much
higher reaction temperature than phenylglycine 1a, that is,
1108C versus room temperature, respectively.[10] The activa-
tion barrier for benzaldimine 2a (R=Ph) is +19.0 kcal
molꢀ1. Similar to (2-Me)allyl ester 29, the 2-methyl substitu-
ent on allyl ester 30 increases the activation barrier for de-
carboxylation to +20.6 kcalmolꢀ1, 1.6 kcalmolꢀ1 higher than
2a. Accordingly, ester 30 requires an elevated temperature,
408C, to achieve decarboxylative allylation.[11a] The p-me-
thoxybenzaldimine 2b exhibits an increase in the activation
barrier by 1.1 kcalmolꢀ1, compared to 2a, whereas the elec-
tron-deficient p-cyanobenzaldimine 2j decreases the barrier
for decarboxylation by 3.8 kcalmolꢀ1 versus 2a. This is con-
sistent with experimental reports that electron-withdrawing
substituents on the imines accelerate the reaction whereas
the electron-rich substituents decrease the reaction rate.
Indeed, we have observed complete conversion of imine 2j
within 24 h at ꢀ108C, whereas little-to-no conversion of p-
methoxybenzaldimine 2b was observed at this relatively low
temperature.[35]
Conclusion
The palladium-catalyzed decarboxylative allylation of allyl
a-imino esters represents a mild and potentially important
tool for the construction of new C
3
ꢀ
(sp ) C
A
work, we presented a detailed theoretical investigation into
the reaction mechanism starting from either a-(diphenylme-
thylene)imino esters (1) or allyl diphenylglycinate imines
(2). Several conclusions can be drawn from the above dis-
cussions as follows:
The complete catalytic cycle consists of three major steps:
oxidative addition, decarboxylation, and reductive allylation.
Decarboxylation represents the rate-determining step and
proceeds via a solvent exposed free anionic transition state
without direct assistance from the Pd-centered cationic com-
ponent. The overall barrier to decarboxylation, and thus the
activation energy for the overall reaction was calculated to
be +22.7 kcalmolꢀ1 relative to 12 for the phenylglycine de-
rivative 1a (R=Ph). This is in accord with reports that the
reaction proceeds at room temperature.
The nucleophilic a-imino anion (2-azaallyl anion) inter-
mediate attacks the cationic p-allylpalladium species via an
outer-sphere pathway from the face opposite to the h3-
bound Pd center resulting in regeneration of a Pd0 species
&
10
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Pd-catalyzed Decarboxylative Allylation of a-imino Esters
FULL PAPER
ꢀ
Chem. 2007, 5, 3571; d) J. A. Tunge, E. C. Burger, Eur. J. Org.
Chem. 2005, 1715.
and C C bond formation. The regioselectivity, in regard to
which carbon of the 2-azaallyl anion is allylated, is deter-
mined by this reductive alkylation step. The free energy bar-
rier for the transition state leading to the major product (1a,
R=Ph) is lower than the corresponding activation barrier of
the minor product 4. The electronic effects of the specific
R groups demonstrate a positive linear Hammett correlation
with the observed regioselectivities, at least starting from
imines 2.
[2] Recent advances for decarboxylative coupling reactions: a) L. J.
Goossen, N. Rodriguez, K. Goossen, Angew. Chem. 2008, 120, 3144;
Angew. Chem. Int. Ed. 2008, 47, 3100; b) L. J. Goossen, B. Zimmer-
mann, T. Knauber, Angew. Chem. 2008, 120, 7211; Angew. Chem.
Int. Ed. 2008, 47, 7103; c) L. J. Goossen, N. Rodriguez, L. Christo-
phe, J. Am. Chem. Soc. 2008, 130, 15248; d) R. Shang, Y. Fu, J. B.
Li, S. L. Zhang, Q. X. Guo, L. Liu, J. Am. Chem. Soc. 2009, 131,
5738; e) R. Shang, Y. Fu, Y. Wang, Q. Xu, H. Z. Yu, L. Liu, Angew.
Chem. 2009, 121, 9514; Angew. Chem. Int. Ed. 2009, 48, 9350; f) H.-
P. Bi, L. Zhao, Y.-M. Liang, C.-J. Li, Angew. Chem. 2009, 121, 806;
Angew. Chem. Int. Ed. 2009, 48, 792; g) C. Wang, I. Piel, F. Glorius,
J. Am. Chem. Soc. 2009, 131, 4194; h) L. Yin, M. Kanai, M. Shibasa-
ki, J. Am. Chem. Soc. 2009, 131, 9610; i) L. J. Goossen, N. Rodri-
guez, P. P. Lange, C. Linder, Angew. Chem. 2010, 122, 1129; Angew.
Chem. Int. Ed. 2010, 49, 1111; j) J. Lindh, P. J. R. Sjoeberg, M.
Larhed, Angew. Chem. 2010, 122, 7899; Angew. Chem. Int. Ed. 2010,
49, 7733; k) L. Zhang, J. Cheng, T. Ohishi, Z. Hou, Angew. Chem.
2010, 122, 8852; Angew. Chem. Int. Ed. 2010, 49, 8670; l) C. Wang,
S. Rakshit, F. Glorius, J. Am. Chem. Soc. 2010, 132, 14006; m) P.
Fang, M. Li, H. Ge, J. Am. Chem. Soc. 2010, 132, 11898; n) F.
Zhang, M. F. Greaney, Angew. Chem. 2010, 122, 2828; Angew.
Chem. Int. Ed. 2010, 49, 2768; o) S. L. Zhang, Y. Fu, R. Shang, Q. X.
Guo, L. Liu, J. Am. Chem. Soc. 2010, 132, 638; p) R. Shang, Z.-W.
Yang, Y. Wang, S.-L. Zhang, L. Liu, J. Am. Chem. Soc. 2010, 132,
14391; q) R. Shang, D.-S. Ji, L. Chu, Y. Fu, L. Liu, Angew. Chem.
2011, 123, 4562; Angew. Chem. Int. Ed. 2011, 50, 4470; r) R. Shang,
L. Liu, Sci. China Chem. 2011, 54, 1670.
When bidentate ligands, such as dppb, are employed, de-
carboxylation still proceeds via a solvent exposed carboxy-
late without direct influence by the palladium. Accordingly,
the bidentate ligand dppb exhibits superior reaction rates to
monodentate ligand PPh3 since the resting state of the dppb/
catalyst complex is the solvent-separated ion pair consisting
of a-imino carboxylate anion and allyl PdII cation, rather
than the neutral complex favored by the monodentate
ligand.
Electron-withdrawing substituents on the R groups in the
2-azaallyl anion are calculated to result in lower activation
barriers for decarboxylation, indicating an acceleration in
the reaction rate. This is consistent with circumstantial ex-
perimental observations.[10,11]
[3] a) C. Torborg, M. Beller, Adv. Synth. Catal. 2009, 351, 3027; b) G. P.
McGlacken, I. J. S. Fairlamb, Eur. J. Org. Chem. 2009, 4011; c) K. C.
Nicolaou, P. G. Bulger, D. Sarlah, Angew. Chem. 2005, 117, 4516;
Angew. Chem. Int. Ed. 2005, 44, 4442; d) I. P. Beletskaya, A. V. Che-
prakov, Chem. Rev. 2000, 100, 3009; e) B. M. Trost, D. L. Van Vrank-
en, Chem. Rev. 1996, 96, 395.
[4] a) J. Tsuji, in Handbook of Organopalladium Chemistry in Organic
Synthesis, Vol. 2, (Eds.: E.-i. Negishi, A. Meijere), Wiley & Sons,
New York, 2002, p. 1669; b) B. M. Trost, M. L. Crawley, Chem. Rev.
2003, 103, 2921; c) M. Braun, T. Meier, Angew. Chem. 2006, 118,
7106; Angew. Chem. Int. Ed. 2006, 45, 6952.
[5] a) E. C. Burger, J. A. Tunge, Org. Lett. 2004, 6, 4113; b) J. T. Mohr,
D. C. Behenna, A. M. Harned, B. M. Stoltz, Angew. Chem. 2005,
117, 7084; Angew. Chem. Int. Ed. 2005, 44, 6924; c) B. Yan, C. D.
Spilling, J. Org. Chem. 2008, 73, 5385.
[6] a) D. C. Behenna, B. M. Stoltz, J. Am. Chem. Soc. 2004, 126, 15044;
b) B. M. Trost, J. Xu, J. Am. Chem. Soc. 2005, 127, 2846; c) B. M.
Trost, J. Xu, T. Schmidt, J. Am. Chem. Soc. 2008, 130, 11852;
d) B. M. Trost, J. Xu, T. Schmidt, J. Am. Chem. Soc. 2009, 131,
18343; e) J. Streuff, D. E. White, S. C. Virgil, B. M. Stoltz, Nat.
Chem. 2010, 2, 192.
Experimental Section
Experimental details: All non-aqueous reactions were carried out in
oven-dried glassware under an inert atmosphere (argon or nitrogen). All
argon was purified by sequential passage through an OXICLEARTM de-
oxygenation column then Drierite and NaOH. All nitrogen was purified
by passage through Drierite. All anhydrous solvents were purchased
from EMD and used as received and transferred via syringe under an
inert atmosphere. Except for PPh3 (TCI America), all ligands and metals
were purchased from Strem and used as received. All other purchased
chemicals were used as received. Imino esters 2a–j were obtained by
imine condensation between allyl diphenylglycinate and the correspond-
ing aldehyde following previously reported protocols.[11] Similarly, imino
ester 5 was obtained following known methods.[11b] Reaction progress was
monitored by analytical thin layer chromatography using 250 mm thick
Partisil K6F 60 ꢁ pre-coated silica gel plates (Whatman) or 1H NMR
spectroscopic analysis. Developed TLC plates were analyzed by UV light
(254 nm) and/or potassium permanganate stain. All flash chromatogra-
phy used Purasil 60 ꢁ 230–400 mesh silica gel (Whatman). 1H NMR and
13C NMR spectra were recorded on a Varian MercuryPlus 300 or Varian
UnityNova 500 at 300 K, as indicated. Chemical shifts are reported in
ppm (d) values relative to the solvent residual peak (CHCl3) [d 7.26 for
1H NMR and d 77.0 for 13C NMR] as a standard.
[7] a) S. Constant, S. Tortoioli, J. Mueller, D. Linder, F. Buron, J.
Lacour, Angew. Chem. 2007, 119, 9137; Angew. Chem. Int. Ed. 2007,
46, 8979; b) S. Constant, S. Tortoioli, J. Mueller, J. Lacour, Angew.
Chem. 2007, 119, 2128; Angew. Chem. Int. Ed. 2007, 46, 2082.
[8] a) S. T. Madrahimov, D. Markovic, J. F. Hartwig, J. Am. Chem. Soc.
2009, 131, 7228; b) O. V. Singh, H. Han, J. Am. Chem. Soc. 2007,
129, 774.
[9] a) J. D. Weaver, B. J. Ka, D. K. Morris, W. Thompson, J. A. Tunge, J.
Am. Chem. Soc. 2010, 132, 12179; b) R. R. P. Torregrosa, Y. Ariyar-
athna, K. Chattopadhyay, J. A. Tunge, J. Am. Chem. Soc. 2010, 132,
9280; c) B. M. Trost, K. Lehr, D. J. Michaelis, J. Xu, A. K. Buckl, J.
Am. Chem. Soc. 2010, 132, 8915; d) R. Shintani, S. Park, F. Shirozu,
M. Murakami, T. Hayashi, J. Am. Chem. Soc. 2008, 130, 16174; e) C.
Wang, J. A. Tunge, J. Am. Chem. Soc. 2008, 130, 8118; f) S. R. Waet-
zig, J. A. Tunge, J. Am. Chem. Soc. 2007, 129, 14860, g) S. R. Waet-
zig, J. A. Tunge, J. Am. Chem. Soc. 2007, 129, 4138.
Acknowledgements
We thank the NSFC (Nos. 20972148, 21002055), China Postdoctoral Sci-
ence Foundation (No. 20090460298, 201003102), Chinese Universities Sci-
entific Fund (Project No. 2011 JS027) and the Thomas F. and Kate Miller
Jeffress Memorial Trust (J-808) for financial support.
[10] E. C. Burger, J. A. Tunge, J. Am. Chem. Soc. 2006, 128, 10002.
[11] a) A. A. Yeagley, J. J. Chruma, Org. Lett. 2007, 9, 2879; b) W. H.
Fields, A. K. Khan, M. Sabat, J. J. Chruma, Org. Lett. 2008, 10, 5131;
[1] a) J. D. Weaver, A. Recio III, A. J. Grenning, J. A. Tunge, Chem.
Rev. 2011, 111, 1846; b) J. T. Mohr, B. M. Stoltz, Chem. Asian J.
2007, 2, 1476; c) J. T. Mohr, D. C. Ebner, B. M. Stoltz, Org. Biomol.
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
11
&
ÞÞ
These are not the final page numbers!

J. J. Chruma, Y. Fu et al.
c) A. A. Yeagley, M. A. Lowder, J. J. Chruma, Org. Lett. 2009, 11,
4022.
[26] a) C. Amatore, A. A. Bahsoun, A. Jutand, L. Mensah, G. Meyer, L.
Ricard, Organometallics 2005, 24, 1569; b) L. A. Evans, N. Fey, J. N.
Harvey, D. Hose, G. C. Lloyd-Jones, P. Murray, A. G. Orpen, R. Os-
borne, G. J. J. Owen-Smith, M. Purdie, J. Am. Chem. Soc. 2008, 130,
14471.
[27] N. H. Sherden, D. C. Behenna, S. C. Virgil, B. M. Stoltz, Angew.
Chem. 2009, 121, 6972; Angew. Chem. Int. Ed. 2009, 48, 6840.
[28] a) C. Amatore, S. Gamez, A. Jutand, Chem. Eur. J. 2001, 7, 1273;
b) C. Amatore, A. Jutand, L. Mensah, G. Meyer, J.-C. Fiaud, J.-Y.
Legros, Eur. J. Org. Chem. 2006, 1185; c) P. Fristrup, T. Jensen, J.
Hoppe, P.-O. Norrby, Chem. Eur. J. 2006, 12, 5352; d) N. Svensen, P.
Fristrup, D. Tanner, P.-O. Norrby, Adv. Synth. Catal. 2007, 349, 2631.
[29] M. Bao, H. Nakamura, Y. Yamamoto, J. Am. Chem. Soc. 2001, 123,
759.
[30] G. J. P. Britovsek, W. Keim, S. Mecking, D. Sainz, T. Wagner, J.
Chem. Soc. Chem. Commun. 1993, 1632.
[31] a) R. Grigg, S. Thianpatanagul, J. Chem. Soc. Chem. Commun. 1984,
180; b) Another possible mechanism for high temperature decarbox-
ylation reaction: R. Grigg, J. Idle, P. McMeekin, D. Vipond, J.
Chem. Soc. Chem. Commun. 1987, 49.
[12] For base promoted a-allylation of imines, see a) J. P. Genet, D. Fer-
roud, S. Juge, J. R. Montes, Tetrahedron Lett. 1986, 27, 4573; b) J. P.
Genet, S. Juge, S. Achi, S. Mallart, J. R. Montes, G. Levif, Tetrahe-
dron 1988, 44, 5263; c) M. Nakoji, T. Kanayama, T. Okino, Y. Take-
moto, J. Org. Chem. 2002, 67, 7418.
[13] a) C. O. Puentes, V. Kouznetsov, J. Heterocycl. Chem. 2002, 39, 595;
b) W.-H. Chiou, N. Mizutani, I. Ojima, J. Org. Chem. 2007, 72, 1871;
c) P. V. Ramachandran, T. E. Burghardt, L. Bland-Berry, J. Org.
Chem. 2005, 70, 7911; d) S. Gille, A. Ferry, T. Billard, B. R. Langlois,
J. Org. Chem. 2003, 68, 8932.
[14] a) M. Atobe, N. Yamazaki, C. Kibayashi, Tetrahedron Lett. 2005, 46,
2669; b) S. Randl, S. Blechert, J. Org. Chem. 2003, 68, 8879; c) G.
Kim, S.-d. Jung, E.-j. Lee, N. Kim, J. Org. Chem. 2003, 68, 5395;
d) D. L. Wright, J. P. Schulte II; M. A. Page, Org. Lett. 2000, 2, 1847;
M. A. Page, Org. Lett. 2000, 2, 1847; e) B. Danieli, G. Lesma, D. Pas-
sarella, A. Sacchetti, A. Silvani, A. Virdis, Org. Lett. 2004, 6, 493.
[15] a) J. A. Keith, D. C. Behenna, J. T. Mohr, S. Ma, S. C. Marinescu, J.
Oxgaard, B. M. Stoltz, W. A. Goddard III, J. Am. Chem. Soc. 2007,
129, 11876; b) R. Shintani, T. Tsuji, S. Park, T. Hayashi, J. Am.
Chem. Soc. 2010, 132, 7508; c) K. Chattopadhyay, R. Jana, V. W.
Day, J. T. Douglas, J. A. Tunge, Org. Lett. 2010, 12, 3042.
[16] a) T. E. Barder, S. L. Buchwald, J. Am. Chem. Soc. 2007, 129, 12003;
b) A. Ariafard, B. F. Yates, J. Am. Chem. Soc. 2009, 131, 13981.
[17] ONIOM method has been validated in previous studies on cross
coupling reactions: a) A. Ariafard, Z.-Y. Lin, J. Am. Chem. Soc.
2006, 128, 13010; b) H. Hirao, K. Morokuma, J. Am. Chem. Soc.
2009, 131, 17206; c) X. Li, L. W. Chung, P. Paneth, K. Morokuma, J.
Am. Chem. Soc. 2009, 131, 5115; d) M. A. Carvajal, S. Kozuch, S.
Shaik, Organometallics 2009, 28, 3656; e) N. J. Brookes, D. C.
Graham, G. Christan, R. Stranger, B. F. Yates, J. Comput. Chem.
2009, 30, 2146; f) Z. Li, Y. Fu, S.-L. Zhang, Q.-X. Guo, L. Liu,
Chem. Asian J. 2010, 5, 1475; g) Z. Li, S.-L. Zhang, Y. Fu, Q.-X.
Guo, L. Liu, J. Am. Chem. Soc. 2009, 131, 8815.
[32] T. Tsuda, Y. Chujo, S. Nishi, K. Tawara, T. Saegusa, J. Am. Chem.
Soc. 1980, 102, 6384.
[33] D. Imao, A. Itoi, A. Yamazaki, M. Shirakura, R. Ohtoshi, K. Ogata,
Y. Ohmori, T. Ohta, Y. Ito, J. Org. Chem. 2007, 72, 1652.
[34] a) Y. Okuno, J. Am. Chem. Soc. 2000, 122, 2925; b) K. S. Peters,
Chem. Rev. 2007, 107, 859; c) N. Deora, P. R. Carlier, J. Org. Chem.
2010, 75, 1061.
[35] See the Supporting Information.
[36] M. Wernerova, T. Hudlicky, Synlett 2010, 2701.
[37] a) C. Hansch, A. Leo, R. W. Taft, Chem. Rev. 1991, 91, 165; b) R. W.
Taft, in Steric Effects in Organic Chemistry, (Ed.: M. S. Newman),
Wiley & Sons, New York, 1956, p. 595 (Table V); c) R. W. Taft, I. C.
Lewis, J. Am. Chem. Soc. 1958, 80, 2436.
[38] The pKa(DMSO) of 18-H, which represents the conjugate acid of the 2-
azaallyl anion derived from imino ester 2a, is reported to be 24.3:
M. J. O’Donnell, W. D. Bennett, W. A. Bruder, W. N. Jacobsen, K.
Knuth, B. LeClef, R. L. Polt, F. G. Bordwell, S. R. Mrozack, T. A.
Cripe, J. Am. Chem. Soc. 1988, 110, 8250.
[39] J. P. Foster, F. Weinhold, J. Am. Chem. Soc. 1980, 102, 7211.
[40] The geometries were optimized by full B3LYP/LANL2DZ method
in this section for higher calculation precision.
[41] C. R. Hauser, I. C. Flur, S. W. Kantor, J. Am. Chem. Soc. 1949, 71,
294.
[18] J. J. P. Stewart, J. Mol. Model. 2007, 13, 1173.
[19] Recent research using PM6 method: a) C. Bonnot, E. Aubert, N.
Banerji, J. Lacour, E. Espinosa, J.-C. Chambron, Chem. Eur. J. 2010,
16, 5706; b) G. Wagner, A. Marchant, J. Sayer, Dalton Trans. 2010,
39, 7747; c) T. Cusati, G. Granucci, M. Persico, J. Am. Chem. Soc.
2011, 133, 5109; d) V. Garg, G. Kodis, M. Chachisvilis, M. Ham-
bourger, A. L. Moore, T. A. Moore, D. Gust, J. Am. Chem. Soc.
2011, 133, 2944.
[20] a) A. D. Becke, J. Chem. Phys. 1993, 98, 5648; b) C. Lee, W. Yang,
R. G. Parr, Phys. Rev. B 1988, 37, 785; c) B. Miehlich, A. Savin, H.
Stoll, H. Preuss, Chem. Phys. Lett. 1989, 157, 200.
[21] a) W. R. Wadt, P. J. Hay, J. Chem. Phys. 1985, 82, 284; b) A. Hçll-
warth, M. Bohme, S. Dapprich, A. W. Ehlers, A. Gobbi, V. Jonas,
K. F. Kohler, R. Stegmann, A. Veldkamp, G. Frenking, Chem. Phys.
Lett. 1993, 208, 237.
[22] a) H. Hagelin, B. Akermark, P.-O. Norrby, Chem. Eur. J. 1999, 5,
902; b) D. Sicinska, D. G. Truhlar, P. Paneth, J. Am. Chem. Soc.
2001, 123, 7683; c) H. M. Senn, T. Ziegler, Organometallics 2004, 23,
2980; d) P. Fristrup, M. Ahlquist, D. Tanner, P.-O. Norrby, J. Phys.
Chem. A 2008, 112, 12862; e) C. P. Butts, E. Filali, G. C. Lloyd-
Jones, P.-O. Norrby, D. A. Sale, Y. Schramm, J. Am. Chem. Soc.
2009, 131, 9945.
[23] a) J. Tomasi, B. Mennucci, R. Cammi, Chem. Rev. 2005, 105, 2999;
b) Y. Fu, L. Liu, R.-Q. Li, R. Liu, Q.-X. Guo, J. Am. Chem. Soc.
2004, 126, 814.
[42] a) S. O. C. Mundle, S. Rathgeber, G. Lacrampe-Couloume, B. S.
Lollar, R. Kluger, J. Am. Chem. Soc. 2009, 131, 11638; b) O. M.
Gonzalez-James, D. A. Singleton, J. Am. Chem. Soc. 2010, 132, 6896.
[43] The Pd cation may also facilitate the decarboxylation by forming a
loose ion pair with TS21–22 to stabilize the nascent negative charge.
[44] In Tungeꢂs report (Ref. [10]), the decarboxylation allylation of the
allyl ester of 21-Na produces an aziridine as the major product. The
mechanism remains unclear and in our computation all proposed
pathways leading to the aziridine show higher free energy barriers
than the pathway leading to homoallylic imine (see the Supporting
Information). This is presumably because the ligand used in our cal-
culation is PPh3 ligand whereas the ligands used in ref. 10 are dppb
or dppf ligand. Further study on the aziridine formation pathway is
needed, but this is beyond the scope of the present work.
[45] In this section, the geometries were optimized by full B3LYP/
LANL2DZ method in solution phase to give more accurate activa-
tion barriers.
[24] M. J. Frisch, Gaussian 09, Revision B.01; Gaussian, Inc., Wallingford
CT, 2009. Full citation and computational details are in the Support-
ing Information.
Received: April 25, 2012
Revised: July 9, 2012
[25] C. Amatore, A. Jutand, Coord. Chem. Rev. 1998, 178, 511.
Published online: && &&, 0000
&
12
&
www.chemeurj.org
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 0000, 00, 0 – 0
ÝÝ
These are not the final page numbers!

Pd-catalyzed Decarboxylative Allylation of a-imino Esters
FULL PAPER
Decarboxylation Reactions
Z. Li, Y.-Y. Jiang, A. A. Yeagley,
J. P. Bour, L. Liu, J. J. Chruma,*
Y. Fu* .......................... &&&& — &&&&
Mechanism of the Pd-catalyzed
ACHUTNGRENDNUG ecarboxylative Allylation of ACHTUNGTRENNUNG
Esters: Decarboxylation via
Carboxylate Ion
ACHTUNGTRENNUNG
Decarboxylation via carboxylate ions:
The Pd-catalyzed decarboxylative ally-
lation of a-imino esters has been stud-
ied using theoretical calculations and
experimental investigations. The cata-
lytic cycle consists of three steps: oxi-
dative addition, decarboxylation, and
reductive allylation. The decarboxyla-
tion step is rate-determining, occurring
via a solvent-exposed a-imino carboxy-
late anion (see scheme).
Chem. Eur. J. 2012, 00, 0 – 0
ꢀ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
&
13
&
ÞÞ
These are not the final page numbers!