Angewandte
Communications
Chemie
by using preparative HPLC on a chiral stationary phase for
racemate separation on gram scale or by using a known
enantioselective [6+3] cycloaddition.[6c] With this optically
pure ligand in hand, we investigated the importance of the
protective group on the secondary amine. Several N-alkylated
ligands were readily synthesized by reductive amination with
various aldehydes. Among those, the N-methylated catalyst 1i
afforded the desired product in 90% yield and with 83% ee.
Further enhancement of steric hindrance by introduction of
an ethyl or an isobutyl substituent led to a decrease in both
reactivity and enantioselectivity (Table S3, entries 11–13).
The structure and absolute configuration of catalyst 1i were
unambiguously determined by crystal-structure analysis (Fig-
ure S3).[9]
Having an optimized catalyst (1i) in hand, we optimized
the reaction conditions (Table S4). The solvent had no
obvious effect on the enantioselectivity, but significantly
influenced the reactivity. Moreover, up to 90% ee could be
reached by lowering the temperature to À108C (Figure 2 and
Table S4, entry 3). Having established optimized reaction
conditions, the substrate scope was explored. As shown in
Figure 2, various styrenes, even those with heterocyclic
substituents or cyclic alkenes, are tolerated in this enantiose-
lective transformation, yielding the desired products with
excellent regio- and enantioselectivity (see 5a–5j and 5k–5m,
respectively). Aryl hydroxamates with substituents with
different electronic and steric properties proved to be suitable
for this reaction and yielded the desired products 5n–5v. The
absolute configuration of product 5m was assigned by
vibrational circular dichroism (VCD) spectroscopy (Fig-
ure S4).[10]
Figure 3. Enantioselective allylation of benzamides.[10] Method J: Ben-
zamides 6 (0.12 mmol, 1.20 equiv) and allenes 7 (0.10 mmol,
1.00 equiv) were used with 0.5 mL of DCM/TFE mixture as the solvent.
Method K: Benzamides 6 (0.10 mmol, 1.00 equiv) and allenes 7
(0.12 mmol, 1.20 equiv) were used. Method L: 0.1 mL of DCM/TFE.
reaction and decreased enantioselectivity (Table S5,
entry 11). Ultimately, catalyst 2l turned out to be best.
Subsequent optimization of the reaction conditions revealed
that the solvent has a dramatic effect on the reactivity, but
only a slight influence on the enantioselectivity (Table S6). A
4:1 combination of dichloromethane (DCM) and trifluoro-
ethanol (TFE) proved to be optimal, affording the desired
product in 85% yield and with 90% ee within 18 h at À208C
(Figure 3). Examination of the possible difference between
the two diastereomers of the RhI complex surprisingly
revealed that both diastereomers of catalyst 2l yielded the
product with the same stereoselectivity. This finding was
confirmed for catalyst 2n (Figure S2). Therefore, RhI com-
plex 2l was employed as a mixture of two diastereomers in all
further transformations. Evaluation of the substrate scope
(Figure 3) showed that benzamides bearing different substi-
tution patterns were well tolerated, affording products 8a–8h
with good to excellent ee values and yields. Furthermore,
various allenes can be employed in this reaction (8i–8k).
To confirm that our approach also enables the discovery
To demonstrate the generality of the approach and the
flexible applicability of our ligand library, we investigated the
À
asymmetric C H allylation of benzamides (Figure 3) that had
previously been rendered enantioselective by Cramer and co-
workers with chiral binaphthyl-substituted Cp ligands.[4c] This
seminal work showed that the cyclohexyl-substituted chiral
Cp ligands successfully applied in the synthesis of isoquino-
linones[4b] (Figure 2) failed to induce high enantioselectivity in
this transformation.[4c] A preliminary experiment showed that
only 24% ee could be achieved with catalyst 1i, which had
proven best in steering the reaction described above (Figure 2
and Table S5, entry 3). Gratifyingly, screening with our ligand
collection rapidly revealed that catalyst 2a, which is based on
a chiral Cp ligand obtained from exo-selective [6+3] cyclo-
addition and equipped with two aryl groups, is very efficient
in this process, yielding up to 86% CTwithout diminishing the
reactivity (Table S5, entry 6). Although RhI complex 2a was
obtained as a 60:40 mixture of diastereomers owing to the
face selectivity of Rh complexation (Table S2), we decided to
directly use this mixture of RhI complexes for further ligand
optimization to streamline the catalyst screening process.
Systematic variation of the R1 and R2 groups in the ligand
demonstrated that excellent enantioselectivity was observed
if ligands with para- or meta-substituted aryl groups were used
(Table S5). The R3 group did not have a significant effect on
the enantioselectivity (Table S5, entries 8, 17, and 18). In
sharp contrast to catalyst 1i, which performed best in the
isoquinolinone synthesis (Figure 2), R4 = Me led to a sluggish
À
of chiral catalysts for unprecedented reactions, a novel C H
activation reaction was realized that yields valuable axially
chiral biaryl compounds with excellent enantioselectivity
(Figure 4). Despite the prevalence of atropisomerism in
natural products,[11] pharmaceuticals,[12] and chiral catalysts/
ligands,[13] the corresponding catalytic enantioselective
approaches have not reached the state of the art for the
establishment of chirality established for other compound
classes, and novel methods are in high demand.[14] Especially
in the field of asymmetric transition-metal catalysis enabled
by chiral Cp ligands, only two classes of transformations have
been documented.[2d-f,5c,k] In contrast, biaryl compounds have
Angew. Chem. Int. Ed. 2017, 56, 1 – 7
ꢀ 2017 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!