Matching the chirality of monodentate N-heterocyclic carbene ligands: A case study on well-defined palladium complexes for the asymmetric α-arylation of amides

Article abstract of DOI:10.1021/ol8021808

(Chemical Equation Presented) N-Heterocyclic carbene ligands derived from C₂-symmetric diamines with naphthyl side chains are introduced as chiral monodentate ligands, and their palladium complexes (NHC)Pd(cin)C) are prepared. These compounds e

Full text of DOI:10.1021/ol8021808

ORGANIC
LETTERS
2008
Vol. 10, No. 24
5569-5572
Matching the Chirality of Monodentate
N-Heterocyclic Carbene Ligands: A Case
Study on Well-Defined Palladium
Complexes for the Asymmetric
r-Arylation of Amides
Xinjun Luan, Ronaldo Mariz, Carine Robert, Michele Gatti, Sascha Blumentritt,
Anthony Linden, and Reto Dorta*
Institute of Organic Chemistry, UniVersity of Zurich, Winterthurerstrasse 190,
CH-8057 Zurich, Switzerland
dorta@oci.uzh.ch
Received October 20, 2008
ABSTRACT
N-Heterocyclic carbene ligands derived from C2-symmetric diamines with naphthyl side chains are introduced as chiral monodentate ligands,
and their palladium complexes (NHC)Pd(cin)Cl are prepared. These compounds exist as a mixture of diastereomers, and the palladium complexes
can be successfully separated and their absolute stereochemistry assigned. When used in the asymmetric intramolecular r-arylation of amides,
oxindoles with quaternary carbon centers can be obtained in high yield and selectivity when correctly matching the chirality of the NHC
complexes.
N-Heterocyclic carbenes often show increased reactivity and
stability when compared to the commonly employed phos-
phine ligands.1 Because of their tight metal binding and
robustness, monodentate, chiral N-heterocyclic carbene
ligands appear to be a very promising ligand class for
asymmetric catalysis.2 Development in this area has been
relatively slow,3-5 and versatile chiral NHCs remain elusive.
Two reasons might account for this. First, the overwhelming
majority of chiral NHC salts is deprotonated and used in
(2) Cesar, V.; Bellemin-Laponnaz, S.; Gade, L. H. Chem. Soc. ReV. 2004,
33, 619.
(3) (a) Ma, Y.; Song, C.; Ma, C.; Sun, Z.; Chai, Q.; Andrus, M. B.
Angew. Chem., Int. Ed. 2003, 42, 5871. (b) Song, C.; Ma, C.; Ma, Y.; Feng,
W.; Ma, S.; Chai, Q.; Andrus, M. B. Tetrahedron Lett. 2005, 46, 3241.
(4) (a) Seiders, T. J.; Ward, D. W.; Grubbs, R. H. Org. Lett. 2001, 3,
3225. (b) Funk, T. W.; Berlin, J. M.; Grubbs, R. H. J. Am. Chem. Soc.
2006, 128, 1840. (c) Berlin, J. M.; Goldberg, S. D.; Grubbs, R. H. Angew.
Chem., Int. Ed. 2006, 45, 7591. For a computational study, see: (d) Costabile,
C.; Cavallo, L. J. Am. Chem. Soc. 2004, 126, 9592.
(1) (a) N-Heterocyclic Carbenes in Synthesis; Nolan, S. P., Ed.; Wiley-
VCH: Weinheim: Germany, 2006. (b) N-Heterocyclic Carbenes in Transi-
tion Metal Catalysis; Glorius, F., Ed.; Topics in Organometallic Chemistry,
Vol. 21; Springer: Berlin, 2007.
10.1021/ol8021808 CCC: $40.75
Published on Web 11/19/2008
2008 American Chemical Society

situ,6 a circumstance that precludes a detailed discussion on
the reaction/selection pathway of a given transformation and
hinders development of optimized ligand structures. Another
problem consists in building chirality without losing the
reactivity/versatility of the most successful achiral ligand
architectures, which incorporate aromatic N-substituents (for
instance, IMes/SIMes and IPr/SIPr). Elegant work by Grubbs
et al. has shown that NHCs derived from C2-symmetric
diamines and mono-ortho-substituted aryl halide side chains
provide successful ligand systems in some ruthenium-
catalyzed asymmetric metathesis reactions.4 Similar archi-
tectures with unsymmetrically 2,6-disubstituted phenyl side
chains have been reported more recently showing encourag-
ing results.5 In all of these cases, no mention is made
regarding the possibly different orientations of the side chains
or their ease of rotation with respect to the chiral N-
heterocyclic backbone and the resulting impact on selectivity
or reactivity in catalysis.7
Figure 1. New chiral N-heterocyclic carbene precursors 1a-d.
cyclohexyl; R2 ) isopropyl or hydrogen) were introduced,
and four different NHC salts were prepared (Figure 1).
Careful analysis of the N-heterocyclic proton signals by 1H
NMR showed three different diastereomers for all four
ligands (1a-d).
In the course of studying a series of NHCs that incorporate
2-substituted naphthyl side chains,8 we found that NHCs with
R1 ) methyl/isopropyl have two atropisomers with C2-
symmetric (anti orientation of side chains) and Cs-symmetric
(syn) conformations (eq 1). Logically, if the chiral regime
of C2-symmetric diamines is introduced to the heterocyclic
backbone of these NHCs, three diastereomers should be
generated. Before embarking on the present study, we
reasoned that ligand systems with bulky R1 groups would
be the best candidates because of their rotational stability
(eq 1) and excellent catalytic behavior.8 Moreover, building
upon the successful separation of anti/syn palladium com-
plexes in our previous study, we decided to concentrate our
preliminary efforts on the palladium-catalyzed asymmetric
intramolecular R-arylation of amides to give chiral quaternary
carbon centers.9
Table 1. Asymmetric Intramolecular R-Arylation of Amide 3a
with 1a-d/Palladium Source
entry L*
[Pd]
T (°C) time (h) % yielda (% eeb, c
)
1
2
3
4
5
6
1a Pd(dba)2
1a [Pd(cin)Cl]2
1a Pd(OAc)2
1b Pd(dba)2
1c Pd(dba)2
1d Pd(dba)2
23
23
23
50
50
23
12
12
24
24
24
12
98 (71)
98 (69)
96 (66)
98 (59)
91 (47)
92 (-6)
a Isolated yields. b Determined by chiral HPLC. c Absolute stereochem-
istry determined as the (R)-configuration; see ref 9d.
Ligand precursors 1a-d were then tested in the intramo-
lecular R-arylation of 3a following Hartwig’s in situ method
(Table 1).9a Results showed that Pd(dba)2, Pd(OAc)2, and
[Pd(cin)Cl]2 (cin ) cinnamyl) could be used as palladium
sources. Although diastereomeric mixtures of ligands 1a-d
were employed, oxindole 4a was obtained with selectivities
of up to 71% ee with ligand 1a (entry 1), whereas using 1d
resulted in almost racemic product (entry 6). This result
already indicates that the chiral information is probably
transferred to the substrate from the chiral diamine part of
the N-heterocycle.
To get the best insight possible into factors that govern
the selectivity and reactivity of these new ligands, variations
in both the chirality of the imidazolinium backbone (R3) and
the steric properties of the side chains (R1 ) isopropyl or
(5) For examples using C2-symmetric chiral NHCs with aromatic side
chains, see: (a) Martin, D.; Kehrli, S.; d’Augustin, M.; Clavier, H.; Mauduit,
M.; Alexakis, A. J. Am. Chem. Soc. 2006, 128, 8416. (b) Chaulagain, M. R.;
Sormunen, G. J.; Montgomery, J. J. Am. Chem. Soc. 2007, 129, 9568. (c)
Matsumoto, Y.; Yamada, K.; Tomioka, K. J. Org. Chem. 2008, 73, 4578.
(6) In some instances, the conditions used for deprotonation are such
that alternative binding modes, for instance η6-coordination of the aromatic
side chains, appear just as likely as coordination of the presumably
deprotonated carbene moiety; see ref 3.
Diastereomers of 1a-c were then used for the synthesis
of (NHC)Pd(cin)Cl complexes 2a-c.10,11 As expected from
our previous studies on similar ligands (eq 1), crude mixtures
of the complexes maintained the ratio between the diaster-
(7) Different rotamers of NHC salts and precatalysts seem to exist; see
the Supporting Information of refs 4a and 5b.
(8) Luan, X.; Mariz, R.; Gatti, M.; Costabile, C.; Poater, A.; Cavallo,
L.; Linden, A.; Dorta, R. J. Am. Chem. Soc. 2008, 130, 6848.
(9) (a) Lee, S.; Hartwig, J. F. J. Org. Chem. 2001, 66, 3402. (b) Glorius,
F.; Altenhoff, G.; Goddard, R.; Lehmann, C. Chem. Commun. 2002, 2704.
(c) Arao, T.; Kondo, K.; Aoyama, T. Chem. Pharm. Bull. 2006, 54, 1743.
(d) Ku¨ndig, E. P.; Seidel, T. M.; Jia, Y. X. Angew. Chem., Int. Ed. 2007,
46, 8484.
(10) Because 1d performed poorly in the in situ catalysis, it was not
employed further
.
(11) Marion, N.; Navarro, O.; Mei, J.; Stevens, E. D.; Scott, N. M.;
Nolan, S. P. J. Am. Chem. Soc. 2006, 128, 4101
.
5570
Org. Lett., Vol. 10, No. 24, 2008

Table 2. Asymmetric Intramolecular R-Arylation of Amides
with 2a-c
% yielda and (% eeb, c
)
entry
R
Ar
Ph
Ph
Ph
Ph
Ph
p-Tol
p-Tol
p-Tol
o-Tol
m-Tol
m-Tol
1-Nap
1-Nap
1-Nap
cat.
Ra,Ra
Ra,Sa
Sa,Sa
7
Me
Me
Me
Bn
Bn
Me
Me
Me
Me
Me
Me
Me
Me
Bn
2a
2b
2c
2a
2b
2a
2b
2c
2a
2a
2b
2a
2b
2a
98 (86)
94 (67)
98 (51)
98 (88)
89 (76)
99 (87)
98 (84)
98 (62)
99 (85)
99 (80)
98 (71)
93 (85)
92 (77)
93 (82)
98 (67)
95 (52)
98 (46)
98 (45)
96 (61)
98 (79)
98 (74)
98 (70)
97 (89)
96 (65)
98 (49)
95 (80)
95 (68)
96 (66)
98 (20)
95 (34)
8d
9d
10
11
12
13
14
15
16
17d
18
19d
20
97 (50)
97 (44)
98 (44)
98 (24)
Figure 2. Complexes 2a-c (top, yields in parentheses) and X-rays
95 (34)
93 (49)
of [(Ra,Ra)-(4S,5S)-DiPhSIPrNap]Pd(cin)Cl (bottom left, (Ra,Ra)-
2a) and [(Ra,Sa)-(4S,5S)-DiPhSI2PrNap]Pd(cin)Cl (bottom right,
(Ra,Sa)-2b).
a Isolated yields. b Determined by chiral HPLC. c Absolute stereochem-
istry determined as (R)-configuration, see ref 9d. d Reaction run at 50 °C.
eomers of the NHC salts. For all three compounds, we were
able to separate these diastereomeric compounds via column
chromatography and assign the absolute configurations
unequivocally by using a combination of both NMR spec-
troscopy and X-ray structural determinations (Figure 2). The
main anti ligand isomer orients its more hindered half of
the naphthyl moiety onto the protruding phenyl ring of the
backbone, an intriguing result in view of the previously
assumed prevalence of (Sa,Sa)-isomers in this ligand class.12
from the minor diastereomeric form of the ligands (Sa,Sa)
failed to give products with more than 50% ee. Both (Ra,Sa)-
and (Ra,Ra)-isomers proved superior in terms of selectivity.
Enantiomeric excesses when the (Ra,Sa)-complexes were
employed ranged from moderate (45% ee) to high (89% ee)
and showed a marked substrate-dependence. Of all the
precatalysts tested, (Ra,Ra)-2a outperformed its congeners and
gave high and very evenly distributed chiral induction
(80-88% ee) for all substrates. As such, (Ra,Ra)-2a compares
well with the best systems described recently by Ku¨ndig et
al.,9d giving generally better yields and slightly higher ee’s
in two cases (entries 10 and 18).
Overall, the results in Table 2 validate the assumption
made above that it is the chiral groups on the N-heterocycle
that determine the absolute configuration of the carbon center
in products 4a-g (all R-configured). Finally, it should also
be noted that adding up the results of entry 7 of Table 2
with (Ra,Ra)-2a (86% ee), (Ra,Sa)-2a (67% ee), and (Sa,Sa)-
2a (20% ee) and taking into account the relative abundance
of diastereomers in the NHC salt 1a essentially replicates
the in situ performance obtained in entry 2 of Table 1.
A viable model for the differences in selectivity between
the three diastereomeric palladium complexes of 2a is
depicted in Scheme 1. The steric pressure exerted by the
chiral NHC-phenyl groups onto the naphthyl side chains and
subsequently onto the enantiodiscriminating side of the
substrate would predominantly give intermediates 5 and, after
reductive elimination, products with the observed (R)-
configuration. Depending on the amount of steric pressure
exerted onto the naphthyl side chains, either higher or lower
proportions of the depicted intermediate 5 would therefore
be generated in the enantiodiscriminating carbon-metal
1
Numerous broad signals in the H NMR spectra of 2a-c
reflect the important steric crowding and the resulting
restricted rotation of both the cinnamyl and NHC side chains
in these complexes. Nonetheless, very diagnostic upfield
shifts are found for some of the R1 protons and arise from
strong σ-π interactions with the chiral backbone phenyl
moieties. Together with the X-ray analyses and the proton
signals of the N-heterocyclic backbone, the presence [twice
for (Ra,Ra)-2, once for (Ra,Sa)-2] or absence [for (Sa,Sa)-2]
of these upfield signals led to the correct assignment of the
different diastereomers.
The series of pure precatalysts was then applied to the
asymmetric oxindole synthesis starting from different 2-bro-
moanilides, and the catalytic results are given in Table 2.
Yields of isolated products are excellent, and reaction times
of less than 24 h were observed. Minor discrepancies in
reactivity exist between catalysts (2a more active than 2b/
2c),13 as well as between the three diastereomers of 2a (see
the Supporting Information). On the other hand, selectivity
differences are very pronounced and reflect the crucial
importance of the ligand architecture. Precatalysts derived
(12) Ra/Sa denote the axial chirality between the N-heterocycle and the
naphthyl side chains.
(13) Compound 2a performs better than the reference system (SIPr)P-
d(cin)Cl; see the Supporting Information.
Org. Lett., Vol. 10, No. 24, 2008
5571

In conclusion, we report the synthesis of new NHC ligands
with a chiral N-heterocycle and naphthyl side chains. Careful
analysis of the imidazolinium salts showed the existence of
three different isomers in such NHC structures. Pure pal-
ladium complexes incorporating these ligands were obtained
after successful separation of their diastereomeric mixtures.
The resulting compounds were tested in the asymmetric
intramolecular R-arylation of amides and lead to the iden-
tification of a precatalyst [(Ra,Ra)-2a] that formed oxindoles
in high yield and high enantiomeric purity. Analysis of the
catalytic data demonstrates the dramatic effects on selectivity
the orientation of the aromatic side chains can have in
catalysis. In view of the experimental results presented here,
the assumption that (Sa,Sa)-isomers are responsible for
enantioselectivity in this chiral NHC subclass should be
questioned. Obviously, direct access to diastereomerically
pure ligand precursors 1a–c and their derivatives would
provide an easier entry to their optimized use in catalysis
and should also facilitate identification of even more effective
and widely applicable ligand architectures. Studies with this
aim are ongoing.
Scheme 1
.
Model Explaining the Differences in Selectivity
(Derived from Catalyst 2a)
bond-forming step preceding intermediate 5. The higher steric
pressure in the (Ra,Ra)-isomer means that the relative
orientation of methyl vs aromatic group of the substrate as
shown in (Ra,Ra)-5(R) is highly favored over (Ra,Ra)-5(S),
while the (Sa,Sa)-isomer gives substantial amounts of (Sa,Sa)-
5(S) and low enantiomeric excess of the (R)-4 products. The
lack of a plane of symmetry in (Ra,Sa)-5(R) means that two
intermediates leading to the major isomer (R)-4 may be
operative, giving either high (left) or moderate (right) degrees
of selectivity. In this case again, the model that we propose
correctly reflects the observed experimental results that show
an apparently random distribution of high and moderate
selectivities when precatalysts (Ra,Sa)-2 are employed.
Acknowledgment. R.D. holds an Alfred Werner Assistant
Professorship and thanks the foundation for generous finan-
cial support. We thank SNF and UZH (Drittmittelkredit for
X.L.) for support.
Supporting Information Available: Experimental pro-
cedures and CIFs for (Ra,Ra)-2a and (Ra,Sa)-2b. This material
is available free of charge via the Internet at http://pubs.acs.org.
OL8021808
5572
Org. Lett., Vol. 10, No. 24, 2008