Exception to the ortho effect in palladium/norbornene catalysis

Article abstract of DOI:10.1002/anie.201104356

Out of the norm: The first deviation from the ortho effect in palladium/norbornene catalysis, as evidenced by the resulting products, is reported (see scheme). DFT calculations indicate that this deviation is likely to originate from a distortion, caused

Full text of DOI:10.1002/anie.201104356

DOI: 10.1002/anie.201104356
ꢀ
C C coupling
Exception to the ortho Effect in Palladium/Norbornene Catalysis**
Marie-Hꢀlꢁne Larraufie, Giovanni Maestri, Aurore Beaume, ꢂtienne Derat, Cyril Ollivier,
Louis Fensterbank, Christine Courillon, Emmanuel Lacꢃte,* Marta Catellani, and
Max Malacria*
Dedicated to Professor Alfredo Ricci
The association of palladium and norbornene opens access to
a myriad of polycyclic and heterocyclic frameworks through
sequential coupling reactions of aryl halides or triflates.[1] In
order for the norbornene cocatalysis to lead to the desired
products, one needs ortho substituents on the aryl halides (the
so-called ortho effect).[2]
Density functional theory (DFT) calculations run in our
group as well as in the Catellani group indicate that the role of
these ortho substituents is to prevent reductive elimination
from the bimetallic PdII–PdII intermediates of type D
(Scheme 1). Starting from the initially formed palladacycle
A, wherein an ortho group is present, the key PdIV inter-
mediate B is formed and undergoes regioselective biaryl
formation to give C rather than E. Our study showed that
steric factors rather than electronic ones are predominant for
the selectivity, and is in agreement with the variety of
synthetic applications reported so far.[3,4]
Scheme 1. The ortho effect in palladium/norbornene catalysis.
Together with the Catellani group,[5] we now report the
formation of adducts derived from the previously unobserved
sp2–sp3 reductive elimination,[2,3c] that is, exceptions to the
ortho rule. In our hands, this sp2–sp3 reductive elimination
serendipitously happened when we worked with 2-bromo-
phenylacetamide (Scheme 2), thus suggesting that coordinat-
ing substituents might induce different reactivities at the
palladium center, and that chelation might therefore be a tool
to achieve regiocontrol in palladium/norbornene catalysis.
Scheme 2. Chelation approach to regiocontrol; n=1, Y=(C(=O)NH2)
for sp2–sp3 coupling (this work). n=0, Y=NH2 for sp2–sp3 coupling
(Ref. [5]).
[*] M.-H. Larraufie, Dr. G. Maestri, A. Beaume, Dr. ꢀ. Derat,
Dr. C. Ollivier, Prof. L. Fensterbank, Prof. C. Courillon, Dr. E. Lacꢁte,
Prof. M. Malacria
UPMC Univ Paris 06, Institut Parisien de Chimie Molꢂculaire
(UMR CNRS 7201), 4 place Jussieu, C. 229, 75005 Paris (France)
E-mail: emmanuel.lacote@icsn.cnrs-gif.fr
max.malacria@upmc.fr
Herein we present our synthetic results and our proposed
mechanistic rationalization that arises from computational
analysis of several reactive pathways.
Homepage: http://www.ipcm.fr/
In a typical experiment, an equimolar amount of 2-
iodotoluene, norbornene, and 2-bromophenylacetamide were
reacted in the presence of Pd(OAc)2 (10 mol%), tri(2-
furyl)phosphine (20 mol%), and Cs2CO3 in dry DMF. Grat-
ifyingly, this selectively yielded 75% of the norbornene-
containing dihydrophenanthrene 1a (Table 1), which is not
the product expected given the ortho rule.[3c,6] We sought to
extend the scope of this peculiar reactivity with differently
substituted aryl iodides by varying both their steric and
electronic properties. Good results were obtained with alkyl
or alkoxy substitutents for the formation of compounds 1a–g
(entries 1–7). Polyaryl iodides (entries 8 and 9) or aryl iodides
bearing chloride or trifluoromethyl substituents (entries 10–
Dr. E. Lacꢁte
ICSN CNRS, Avenue de la Terrasse
91198 Gif-sur-Yvette Cedex (France)
Prof. M. Catellani
Dipartimento di Chimica Organica e Industriale, Universitꢃ degli
Studi di Parma, Parco Area delle Scienze, 43124 Parma (Italy)
[**] We thank the UPMC, CNRS, IUF (M.M., L.F.), and ANR (grant
Credox) for funding of this work. M.-H.L. thanks the rꢂgion Ile-de-
France and A.B. thanks Pierre Fabre for stipends. Analytical
equipment was generously offered through FR 2769. We thank
Geoffrey Gontard for the X-ray diffraction analysis.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201104356.
Angew. Chem. Int. Ed. 2011, 50, 12253 –12256
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
12253

Communications
Table 1: Multicomponent synthesis of dihydrophenanthrene.
general although yields remained limited (ranging from 28%
to 65%).
On the basis of both our theoretical results and previous
studies,
a
hypothetical mechanism was considered
(Scheme 3). The ortho-substituted aryl iodide and norbor-
Entry[a]
Aryl iodide
Z
Yield [%][b]
1
2
3
4
5
6
7
8
R1 =Me; R2 =R3 =H
R1 =Et; R2 =R3 =H
H
H
H
H
H
H
H
H
H
H
H
CF3
CF3
1a: 75
1b: 57
1c: 60
1d: 60
1e: 82
1 f: 65
1g: 60
1h: 61
1i: 81
1j: 47
1k: 71
1l: 45
1m: 52
R1 =CH2-OMe; R2 =R3 =H
R1 =Me; R2 =Me; R3 =H
R1 =Me; R2 =H; R3 =Me
R1 =Me; R2 =H; R3 =OMe
R1 =Me; R2 =R3 =OMe
1-iodonaphtalene
9
9-iodophenanthrene
10
11
12
13
R1 =Cl; R2 =R3 =H
R1 =CF3; R2 =R3 =H
R1 =Me; R2 =R3 =H
R1 =Me; R2 =H; R3 =Me
[a] Reaction conditions: ArBr (0.36 mmol, 0.045m in DMF), ArI
(0.36 mmol), norbornene (0.36 mmol), Pd(OAc)2 (0.036 mmol), TFP
(0.072 mmol), Cs2CO3 (0.72 mmol), 1308C. [b] Yield of isolated product.
DMF=dimethylformamide.
11) were also reacted successfully and led to the adducts 1h–
k. 2-Bromo-5-trifluoromethyl phenylacetamide followed a
similar trend and delivered 1l–m in reasonable yields
(entries 12 and 13). Overall, the exception to the ortho rule
proved to be at work with a variety of substituents.
Scheme 3. Proposed mechanistic pathways.
nene should lead to metallacycle A as usual.[8] A would then
oxidatively add to 2-bromophenylacetamide to generate the
PdIV complex B.[3a] At this point, coordination of the carbonyl
moiety would form a chelate that might favor sp –sp C C
bond-forming reductive elimination to 1,2 diaryl norbornene
Upon determining the scope of the chelation-controlled
sequence, we observed that when the substrates were not
carefully dried, an appreciable amount of the spiro derivative
3a was isolated as a by-product together with the now familiar
1a (Table 2).[7] When excess water (11 equiv relative to
palladium) was added, 3a was obtained as the sole product
of the reaction (entry 3). An additional increase of the water
content led to degradation. It is noticeable that at first sight,
the formation of 3a arises as a result of the ortho effect.
We investigated the behavior of various aryl iodides in the
presence of water (eight examples, see Table S1 in the
Supporting Information). Pleasingly, the palladium-mediated
dearomatization leading to spirocyclic adducts proved quite
2
3
ꢀ
ꢀ
derivative E, which would undergo a second C H activation
step that would release 1a–m and the Pd0 complex for
turnover (path a, Scheme 3).
Somehow, water returns the reactive path back toward the
traditional manifold, either through an apical ligand
exchange, hydrogen bonding, or stabilization of a pentavalent
PdIV species. Further on in the reaction pathway, water also
appears to hamper the usual norbornene extrusion[1a,b]
through b-hydride elimination, thus favoring dearomatization
instead.[6]
The PdIV intermediates involved are extremely prone to
reductive eliminations, and it has not been possible to isolate
any putative intermediate from the reaction mixtures. We
thus turned to DFT modeling to gain insight into the
reactivity.
Table 2: Effect of water on the selectivity of the reaction.
To check the feasibility of the PdIV complex B, we
modeled oxidative additions of the aryl bromide onto the
starting PdII metallacycle (with and without amide chelation)
and we compared them with a bimetallic pathway alternative
to PdIV. The energetic convenience of PdIV formation was
confirmed, in accordance with our previous calculations.[3a] In
our system the chelating substituent in the ortho position of
the aryl bromide certainly increases the energy gap between
these pathways, by both favoring the oxidative addition
leading to B (as demonstrated recently by Vicente et al.)[3d]
Entry[a]
n
1a/3a[b]
1
2
3
0
2.5
11
1:0 (1a only)
1:1
0:1 (3a only; 65% yield)[c]
[a] Reaction conditions: as in Table 1 but at half the scale. [b] Determined
by 1H NMR analysis. [c] Yield of isolated product.
12254 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 12253 –12256

and increasing the steric hindrance in the bimetallic pathway
a very short distance between the hydrogen atom ortho to the
(PdIV formation results favored by more than 20 kcalmolꢀ1
see Figures S1 and S2 in the Supporting Information).
;
C Pd bond on the benzylamide and the benzylic hydrogen
ꢀ
atom of the norbonene b to palladium (2.01 ꢀ; Figure S8).
This second effect additionally contributes to increasing the
activation energy for the sp2–sp2 coupling. The combined
effects lead the usually disfavored sp2–sp3 coupling to prevail.
Water can efficiently replace the phosphine as the apical
ligand and the corresponding cis Br/norbornene PdIV complex
is stabilized because of the steric strain released in the ligand
exchange (DG = ꢀ7.6 kcalmolꢀ1, L = H2O).[9] In this case
(Figure 1, levels marked with dashed lines), the coordination
of palladium is similar to that in the path a transition state
relative to the phosphine case, but the geometry changes
significantly in the path b transition state (sp2–sp2 coupling).
The steric strain is partially released since the shortest H–H
distance is now 2.08 ꢀ. Moreover, if chelation is still present,
a partial displacement of water is nonetheless observed (the
The PdIV intermediate calc-B obtained from oxidative
addition of bromophenyl acetamide to the cyclopalladated
complex A converged systematically on a minimum energy in
which a chelate to the palladium center forms through the
oxygen atom of the amide (L = tris(2-furyl)phosphine;
Figure 1). In contrast to previous calculations on the octa-
hedral PdIV complexes with an apical aliphatic or benzylic
ꢀ
Pd O distance goes up to 2.63 ꢀ from 2.34 ꢀ in the path a
transition state (Figure S8). Thus the metal resembles a more
reactive pentacoordinated species. These two features com-
bined restore the usual and expected energetic preference of
the sp2–sp2 coupling over the sp2–sp3 coupling (DDG of
12.1 kcalmolꢀ1), which is in perfect agreement with the
experimental results. The inversion in the favored reductive
elimination was found only when water replaced the phos-
phine. Alternative mechanisms based on pentacoordinated
high-valent palladium, hydrogen bonding of water to the
amide, and scrambling of the axial ligands were considered
but cannot explain the reaction outcome (see the Supporting
Information for details).[10]
Figure 1. Calculations of the barrier energies for the sp3–sp2 (path a)
and sp2–sp2 (path b) couplings starting from complex B having either
a phosphine ligand (solid energy levels) or water (dashed levels).
Values in kcalmolꢀ1. See the Supporting Information for computational
details.
This reaction mechanism also agrees well with the results
of Table 2. Since B is a closed-shell, 18-electron complex,
replacement of the phosphine with water likely occurs
through a dissociative mechanism,[11] and the two complexes
are in thermodynamic equilibrium. We postulate that a 1:1
mixture is formed when the two ligands are present in a nearly
equimolar amount (Table 2, entry 2). Formation of 1 is
excluded only when the excess of water is large enough to
group,[3a] the chelating acetamido moiety also forces the
halide to be cis to the norbornane unit (Figure 1, and see
Figure S3 in the Supporting Information). Without water, the
2
3
ꢀ
activation barrier to the sp –sp C C bond formation is lower
than the barrier leading to the usual sp2–sp2 C C bond
ꢀ
formation (3.0 kcalmolꢀ1, Figure 1, levels marked with solid
lines).
ꢀ
overcome the contribution of the free enthalpy of the Pd P
bond.
The chelation of the amide is released in the transition
state leading to the sp2–sp3 coupling in the calculated
phosphine-containing system (Figure 1, path a, and see Fig-
ure S8 in the Supporting Information). Therefore, the metal
center is pentacoordinated. On the contrary, the chelation
remains in place for the transition state of the compound
showing an ortho effect (path b); that is the palladium atom is
hexacoordinated in that case (Figure S8). We know that
barriers for reductive eliminations from octahedral PdIV
complexes are in general higher than those of pentacoordi-
nated PdIV complexes,[3a] which gives a first rationale for the
observed exception. To further strengthen the validity of our
model, we calculated the barriers from a chelated PdIV species
analogous to B but arising from oxidative addition of
bromobenzylamine. This cascade was proven to follow the
general ortho rule.[4] Results correlate with experiments since
sp2–sp2 coupling is favored in this case by 5.2 kcalmolꢀ1
(Figures S4 and S5 in the Supporting Information).
According to the DFT calculations, the amide moiety is
therefore crucial for leading to the exception to the ortho rule.
Chelation forces the PdIV complex into a cis Br/norbornene
geometry, and it also likely hampers the transition state
leading to sp2–sp2 coupling. Water can overcome this trend by
reducing the steric strain and increasing the flexibility of the
complex.
ꢀ
Downstream from the C C coupling, 5-exo migratory
insertion or tandem deprotonation/nucleophilic attack at the
palladium center could explain the dearomatization of
complex C (Figure 1) and the formation of products 3 via F.
Our DFT results suggest that the former option is favored by
5.6 kcalmolꢀ1 (see Figure S9 in the Supporting Information).
Formation of 3 from E in the presence of water could be ruled
out on the basis of the large energetic gap between the two
nonreversible reductive eliminations from the complex B.[7b,12]
The relative ease of this diastereoselective dearomatization
(calculated barrier of + 13.3 kcalmolꢀ1 for C!F) is addition-
ally confirmed since it experimentally happens faster than the
In addition, steric strain appears to develop more in the
ortho-effect transition state leading to calc-C, as evidenced by
Angew. Chem. Int. Ed. 2011, 50, 12253 –12256
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 12255

Communications
usual norbornene extrusion (+ 20.5 kcalmolꢀ1 barrier; see
A. M. Echavarren, J. Am. Chem. Soc. 2006, 128, 5033 – 5040;
c) M. Livendahl, A. M. Echavarren, Isr. J. Chem. 2010, 50, 630 –
651; d) J. Vicente, A. Arcas, F. Juliꢁ-Hernꢁndez, D. Bautista,
Angew. Chem. 2011, 123, 7028 – 7031; Angew. Chem. Int. Ed.
2011, 50, 6896 – 6899.
Figure S11 in the Supporting Information).[1,2] The latter step
is usually favored by steric hindrance.[1a,b] We thus suggest that
water is crucial not just for switching the favored reductive
elimination from B but also for triggering the subsequent
dearomatization, owing to its minimal size.
[4] G. Maestri, M. H. Larraufie, E. Derat, C. Ollivier, L. Fenster-
bank, E. Lacꢂte, M. Malacria, Org. Lett. 2010, 12, 5692 – 5695.
[5] N. Della Ca, G. Maestri, M. Malacria, E. Derat, M. Catellani,
Angew. Chem. 2011, 123, 12456 – 12469; Angew. Chem. Int. Ed.
2011, 50, 12557 – 12261.
We found a barrier of + 19.6 kcalmolꢀ1 for the formation
of the other diastereomer of 3, and it is due to a clear steric
interaction between the amide group and the norbornene unit
(see Figure S10 in the Supporting Information). The fact that
DFT predicted the observed diastereostereoselectivity for the
dearomatization supports the proposed mechanism.
To conclude, we have reported the first deviation from the
ortho effect in palladium/norbornene catalysis. DFT calcu-
lations indicate that this likely originates from the role played
by a suitably placed amide group in the reductive elimination
pathway involving an initially formed chelated PdIV inter-
mediate. Addition of water restores the reaction to its normal
(ortho rule in effect) selectivity, however it led to dearoma-
tization. Additional work will seek to expand the synthetic
applications of the chelate approach in PdIV chemistry.
[6] The outcome of the reaction is completely different from the one
obtained by using o-bromobenzamides: R. Ferraccioli, D. Ca-
renzi, O. Rombola, M. Catellani, Org. Lett. 2004, 6, 4759 – 4762.
[7] For a recent review, see: a) S. P. Roche, J. A. Porco Jr, Angew.
Chem. 2011, 123, 4154 – 4179; Angew. Chem. Int. Ed. 2011, 50,
4068 – 4093; a similar reactivity was observed with phenols,
which are much more prone to dearomatization: b) M. Catellani,
F. Cugini, G. Bocelli, J. Organomet. Chem. 1999, 584, 63 – 67; for
other reported pallado-catalyzed dearomatizations, see: c) R. B.
Bedford, C. P. Butts, M. F. Haddow, R. Osborne, R. F. Sankey,
Chem. Commun. 2009, 4832 – 4834; d) R. B. Bedford, N. Fey,
M. F. Haddow, R. F. Sankey, Chem. Commun. 2011, 47, 3649 –
3651; e) J. Garcꢃa-Fortanet, F. Kessler, S. L. Buchwald, J. Am.
Chem. Soc. 2009, 131, 6676 – 6677; f) S. Lu, Z. Xu, M. Bao, Y.
Yamamoto, Angew. Chem. 2008, 120, 4438 – 4441; Angew. Chem.
Int. Ed. 2008, 47, 4366 – 4369.
Received: June 23, 2011
[8] a) A. H. Roy, J. F. Hartwig, J. Am. Chem. Soc. 2003, 125, 13944 –
13945; b) C. Amatore, A. Jutand, Acc. Chem. Res. 2000, 33, 314 –
321; c) C.-S. Li, C.-H. Cheng, F.-L. Liao, S.-L. Wang, J. Chem.
Soc. Chem. Commun. 1991, 710 – 712; d) M. Portnoy, Y. Ben-
David, I. Rousso, D. Milstein, Organometallics 1994, 13, 3465 –
3479; e) M. Catellani, C. Mealli, E. Motti, P. Paoli, E. Perez-
Carreno, P. S. Pregosin, J. Am. Chem. Soc. 2002, 124, 4336 – 4346;
f) C.-H. Liu, C.-S. Li, C.-H. Cheng, Organometallics 1994, 13,
18 – 20; g) G. Dyker, Angew. Chem. 1999, 111, 1808 – 1822;
Angew. Chem. Int. Ed. 1999, 38, 1698 – 1712; h) F. Kakiuchi, N.
Chatani, Adv. Synth. Catal. 2003, 345, 1077 – 1101; i) D. I. Chai, P.
Thansandote, M. Lautens, Chem. Eur. J. 2011, 17, 8175-8188.
[9] For other examples of intermediate stabilization by water, see: S.
Baumgarten, D. Lesage, V. Gandon, J.-P. Goddard, M. Malacria,
J.-C. Tabet, Y. Gimbert, L. Fensterbank, ChemCatChem 2009, 1,
138 – 143, and references therein.
Revised: August 24, 2011
Published online: October 21, 2011
ꢀ
ꢀ
Keywords: C C coupling · C H activation · dearomatization ·
density functional calculations · palladium
.
[1] Reviews: a) M. Catellani, E. Motti, N. Della Ca, Acc. Chem. Res.
2008, 41, 1512 – 1522; b) A. Martins, B. Mariampillai, M.
Lautens, Top. Curr. Chem. 2010, 292, 1 – 33; c) P. Sehnal,
R. J. K. Taylor, I. J. S. Fairlamb, Chem. Rev. 2010, 110, 824 –
889; d) A. R. Dick, M. S. Sanford, Tetrahedron 2006, 62, 2439 –
2463; e) K. MuÇiz, Angew. Chem. 2009, 121, 9576 – 9588; Angew.
Chem. Int. Ed. 2009, 48, 9412 – 9423.
[2] Recent examples: a) Y.-B. Zhao, B. Mariampillai, D. A. Candito,
B. Laleu, M. Z. Li, M. Lautens, Angew. Chem. 2009, 121, 1881 –
1884; Angew. Chem. Int. Ed. 2009, 48, 1849 – 1852; b) D. A.
Candito, M. Lautens, Angew. Chem. 2009, 121, 6841 – 6844;
Angew. Chem. Int. Ed. 2009, 48, 6713 – 6716; c) M. Blanchot,
D. A. Candito, F. Larnaud, M. Lautens, Org. Lett. 2011, 13,
1486 – 1489; d) G. Maestri, N. Della Ca, M. Catellani, Chem.
Commun. 2009, 4892 – 4894; e) N. Della Ca, G. Maestri, M.
Catellani, Chem. Eur. J. 2009, 15, 7850 – 7853.
[10] In analogy with the reactions without water, only the PdIV
pathway was considered in the presence of water. Other
potential pathways cannot be formally excluded at this stage.
[11] Reviews: a) C. A. Tolman, Chem. Soc. Rev. 1972, 1, 337 – 353;
b) J. A. S. Howell, P. M. Burkinshaw, Chem. Rev. 1983, 83, 557 –
599.
[12] Detailed computational results supporting our mechanistic
rationalization in the dearomatization reaction are available in
the Supporting Information.
[3] a) G. Maestri, E. Motti, N. Della Ca, M. Malacria, E. Derat, M.
Catellani, J. Am. Chem. Soc. 2011, 133, 8574 – 8585. For other
computational studies, see: b) D. J. Cꢁrdenas, B. Martin-Matute,
12256 www.angewandte.org
ꢀ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2011, 50, 12253 –12256