Catalytic Enantioselective Transformations Involving C−H Bond
Cleavage by Transition-Metal Complexes
Christopher G. Newton,
†
Shou-Guo Wang,
†
Caio C. Oliveira, and Nicolai Cramer*
Laboratory of Asymmetric Catalysis and Synthesis, Institute of Chemical Sciences and Engineering, E
cole Polytechnique Fe
de
rale de
Lausanne (EPFL), CH-1015 Lausanne, Switzerland
ABSTRACT: The development of new methods for the direct functionalization of
unactivated C−H bonds is ushering in a paradigm shift in the field of retrosynthetic
analysis. In particular, the catalytic enantioselective functionalization of C−H bonds
represents a highly atom- and step-economic approach toward the generation of
structural complexity. However, as a result of their ubiquity and low reactivity,
controlling both the chemo- and stereoselectivity of such processes constitutes a
significant challenge. Herein we comprehensively review all asymmetric transition-metal-
catalyzed methodologies that are believed to proceed via an inner-sphere-type
mechanism, with an emphasis on the nature of stereochemistry generation. Our
analysis serves to document the considerable and rapid progress within in the field, while
also highlighting limitations of current methods.
CONTENTS
1. Introduction 8908
2. Stereochemistry-Generating C−H Activation 8909
2.1. C(sp
2
)−H Functionalization 8909
2.1.1. Palladium Catalysis 8909
2.1.2. Rhodium Catalysis 8916
2.1.3. Iridium Catalysis 8918
2.2. C(sp
3
)−H Functionalization 8919
2.2.1. Palladium Catalysis 8919
2.2.2. Rhodium Catalysis 8925
2.2.3. Iridium Catalysis 8925
3. Stereochemistry-Generating Migratory Insertion 8926
3.1. C(sp
2
)−H Functionalization 8926
3.1.1. Alkenyl or Aryl C−H Functionalization 8926
3.1.2. Aldehyde C−H Functionalization 8940
3.1.3. Aldiminium C−H Functionalization 8951
3.1.4. Formamide C−H Functionalization 8952
3.2. C(sp
3
)−H Functionalization 8952
3.2.1. Hydroaminoalkylation 8952
4. Other Stereochemistry-Generating Steps 8953
5. Ambiguous or Unknown Stereochemistry-Gener-
ating Steps 8954
5.1. Synthesis of Atropisomers 8955
5.2. Allylic C(sp
3
)−H Functionalization 8956
5.2.1. Palladium Catalysis 8958
5.2.2. Rhodium Catalysis 8960
6. Other Reactions as Stereochemistry-Generating 8960
7. Kinetic Resolutions 8961
7.1. Standard Kinetic Resolutions 8961
7.2. Dynamic Kinetic Resolutions 8964
7.3. Parallel Kinetic Resolutions 8964
8. Conclusion and Future Outlook 8967
Author Information 8967
Corresponding Author 8967
ORCID 8967
Author Contributions 8967
Author Contributions 8967
Notes 8967
Biographies 8967
Acknowledgments 8968
Abbreviations Used 8968
References 8968
1. INTRODUCTION
Synthetic strategies toward the enantioselective construction of
carbon−carbon (C−C) and carbon−heteroatom (C−X) bonds
have traditionally relied upon the presence of either a heteroatom
or unsaturation to facilitate the transformation.
1−3
Transition-
metal-catalyzed cross-coupling approaches are among the most
powerful methodologies in the field, and over the past several
decades, the development of new reactions, ligand families, and
metal complexes have enabled myriad elegant and creative
applications in both academic and industrial settings.
4−9
However, despite their enormous utility, this class of reaction
is not without limitation. At least one of the coupling partners
requires prefunctionalization; thus, multistep preparative
sequences are often n ecessary. In addition, many of the
organometallic reactants (and their byproducts) are highly
toxic, air and moisture sensitive. Conceptually, the direct
functionalization of carbon−hydrogen (C−H) bonds represents
an elegant, atom- and step-economic solution,
10−13
with the
Special Issue: CH Activation
Received: October 8, 2016
Published: February 17, 2017
Review
pubs.acs.org/CR
© 2017 American Chemical Society 8908 DOI: 10.1021/acs.chemrev.6b00692
Chem. Rev. 2017, 117, 8908−8976
Downloaded via ECOLE POLYTECHNIC FED LAUSANNE on May 7, 2020 at 07:14:49 (UTC).
See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
potential to fundamentally alter the logic of retrosynthetic
analysis. In particular, the burgeoning field of enantioselective
transition-metal-catalyzed C−H functionalization holds great
promise as an enabling methodology for the rapid generation of
structural complexity from simple precursors.
14−17
A variety of systems for enantioselective transition-metal-
catalyzed C−H functionalization reactions have been developed,
and each can be classified by the general mechanism in operation
(Scheme 1).
18
Reactions that proceed via a C−H bond cleavage
event that leads to the formation of a discrete organometallic
intermediate can be considered “inner-sphere” mechanisms. In
contrast, “ outer-sphere” processes do not involve direct
interaction between the C−H bond and the metal center, but
rather, association with a coordinated ligand facilitates C−H
bond cleavage. These processes can proceed via either a
concerted insertion process, such as in the case of metal−
carbenoid and −nitrenoid insertions,
19
or via an H atom
abstraction/rebound sequence, as is observed for iron porphyrin
C−H hydroxylation.
20
As a result of the fundamental differences
between these mechanisms, the innate selectivity between each
methodology differs. Generally speaking, inner-sphere processes
are particularly sensitive to the steric environment of the C−H
bond, whereas outer-sphere are usually selective for the weakest
C−H bonds.
18
The scope of this review is limited to those examples that are
believed to proceed via an inner-sphere mechanism and involve
activation of a C−H bond with a pK
a
greater than 25, thus best
enabling a comparative analysis between related methodologies.
Accordingly, reactions that represent a formal C−H function-
alization (e.g., the Mizoroki−Heck,
21
Wacker−Tsuji oxidation,
22
Friedel−Crafts,
23
etc.) will not be discussed, although in cases
where the mechanism is ambiguous or debated we have erred on
the side of caution, and these examples are included. With the
exception of the Kharasch−Sosnovsky reaction, which has been
well-covered elsewhere,
16,24,25
we have made all attempts to
comprehensively review this area of research up until November
2016, and we have primarily organized the literature according to
the nature of the stereochemistry-generating step in each
methodology.
2. STEREOCHEMISTRY-GENERATING C−H
ACTIVATION
2.1. C(sp
2
)−H Functionalization
C(sp
2
)−H bonds are typically more sterically accessible and
acidic than their sp
3
-hybridized counterparts. Although these
inherent biases can be overcome, for example, via the
employment of directing groups, the asymmetric C−H
functionalization of C(sp
2
)−H bonds remains a considerably
more common process. To the best of our knowledge, all
methodologies that proceed via a stereochemistry-generating
C(sp
2
)−H activation involve desymmetrization of a prochiral
starting material via the selective C−H activation of an
enantiotopic aryl C−H bond, followed by either an inter- or
intramolecular functionalization to generate a molecule with
either central or planar chirality (Scheme 2).
2.1.1. Palladium Catalysis. Palladium is the most
commonly employed transition metal for reactions that proceed
through a stereochemistry-generating C−H activation, and
within this area there are three general mechanistic scenarios
(Scheme 3). The Pd(0)/Pd(II) catalytic cycle initiates with
oxidative addition of a Pd(0) complex into a carbon−
(pseudo)halogen bond, followed by a C−H activation event
that typically proceeds via a reversible carboxylate-assisted
concerted metalation−deprotonation (CMD) mechanism.
26,27
Scheme 1. Mechanistic Classification of Transition-Metal-
Catalyzed Enantioselective C−H Functionalization Reactions
Scheme 2. Nature of Chirality Generated by Stereochemistry-
Generating C(sp
2
)−H Activations
Scheme 3. General Mechanisms of Palladium-Catalyzed
Reactions Involving a Stereochemistry-Generating C−H
Activation
Chemical Reviews Review
DOI: 10.1021/acs.chemrev.6b00692
Chem. Rev. 2017, 117, 8908−8976
8909
Irreversible deprotonation of the released carboxylic acid by an
inorganic base renders the CMD enantiodete rmining and
regenerates the carboxylate cocatalyst.
28
Finally, reductive
elimination serves to release the coupled product and regenerate
the active Pd(0) catalyst. To date this process has only been
realized in an intramolecular sense, likely highlighting the
difficulty of controlling the enantioselectivity of an intermo-
lecular C−H activation event with today’s methodologies.
Conversely, enantioselective reactions proceeding via Pd(II)/
Pd(0) or Pd(II)/Pd(IV) catalysis have been conducted in both
an intra- and intermolecular sense. Both catalytic cycles begin
with a directed C−H activation, facilitated by a Pd(II) complex,
before bifurcation of the two processes. The former proceeds via
a transmetalation/reductive elimination sequence, providing the
coupled product and a reduced form of the palladium catalyst. In
this case, reoxidation via an external oxidant closes the catalytic
cycle. The Pd(II)/Pd(IV) cycle continues via an oxidative
addition process to generate an electron-poor Pd(IV) species
that reductively eliminates to complete the catalytic cycle.
29
Although the three mechanistic pathways described above are
the most common, a recent combined experimental and
computational study by Yu, Musaev, and their co-workers
indicates that for C−H iodination reactions employing molecular
iodine as the sole oxidant, a switch in mechanism can occur
depending on the nature of the Pd-directing group interaction
and the hybridization of the C−H bond.
30
More specifically,
C(sp
2
)−H bonds typically react via a Pd(II) redox-neutral
mechanism (Scheme 4). In this case, following C−H activation, a
concerted electrophilic Pd−C bo nd cleav age provi des the
iodinated product and regenerates the active catalyst. Notably,
C(sp
3
)−H iodination reactions tend to proceed through the
Pd(II)/Pd(IV) catalytic cycle described above; however this
preference can be reversed via the employment of weak directing
groups (currently only achieved in an achiral fashion).
2.1.1.1. Pd(0)/Pd(II). The first example of an enantioselective
Pd(0)/Pd(II) C(sp
2
)−H functionalization reaction was dis-
closed by the Cramer research group in 2009 and involved
intramolecular arylation of vinyl triflates 1, providing access to
chiral indanes 2 bearing an all-carbon quaternary stereocenter
(Scheme 5).
31
At the time only one example of a palladium-
catalyzed enantioselective reaction involving a stereochemistry-
generating C−H activation had been reported, by Yu and co-
workers, who in a pioneering study had successfully applied
mono-N-protected amino acids (MPAA) as ligands in a Pd(II)/
Pd(0)-catalyzed process (presented later in Scheme 15).
32
In
efforts directed toward the development of complementary
reaction pathways and the expansion of suitable ligand families
for enantioselect ive C−H functionalization, Cr amer et al.
discovered that monodentate phosphine ligands displayed high
reactivity in the envisioned transformation and that TADDOL-
derived phosphoramidites, in particular, provided excellent
enantiocontrol. Under optimized conditions, a room-temperate
C−H functionalization, facilitated by 4-tBu-C
6
H
4
substituted
ligand L1, yielded the desired indanes in high enantiopurity. A
variety of functionalities were tolerated in the reaction, including
aryl chlorides (2a), substituted thiophenes (2b), and Boc-
protected amines (2c). Although reaction of acyclic alkenyl
triflates was also tolerated (2d ), a significant reduction in
enantioselectivity was observed. The working mechanistic model
for the reaction was based on proposals by Fagnou,
23−25
and a
collaborative study from Maseras, Echavarren, and their co-
workers
33,34
in which the oxidative addition intermediate 3
undergoes ligand exchange with sodium bicarbonate to provide
4, followed by an enantiodetermining carboxylate-assisted CMD
to palladacycle 5.
In 2013, Saget and Cramer extended this methodology to the
arylation of amides 6, providing highly functionalized
dibenzazepinones 7 (Scheme 6).
35
In this case, the C−H
activation generates a rare eight-membered palladacycle, which at
the time was an unprecedented process in asymmetric catalysis.
Protection of the amide was determined to be necessary for
reactivity, and except for in the case of substituted pyridines (e.g.,
when X = N), all examples proceeded with high levels of
enantiocontrol. Notably, even when 4-methoxybenzyl-substi-
tuted amide 8 was employed, dibenzazepinone 9 was generated
in 97% yield, and no competing C−H functionalization to
dihydrophenanthridine 10 was detected (in this case via a seven-
membered palladacycle).
Several methodologies proceeding via a stereochemist ry-
generating C−H activation event to generate planar chiral
compounds have also been disclosed. All involve the
functionalization of metallocenes, enabling efficient access to
valuable chiral scaffolds with potential application as new ligands
or catalysts.
36
Early Pd(II)/Pd(0)-catalyzed studies on sub-
stit uted ferrocenes required external oxidants for catalytic
turnover (presented later in Scheme 17),
37−39
creating the
potential for undesired ferrocenium generation. In early 2014,
the You
40
and Gu
41
groups independently and concurrently
circumvented this potential problem via the development of an
intramolecular Pd(0)/Pd(II)-catalyzed arylation of aryl halides
11 (Scheme 7). This methodology was later extended by the You
Scheme 4. Redox-Neutral Pd(II)-Catalyzed C(sp
2
)−H
Iodination
Scheme 5. First Pd(0)/Pd(II) Enantioselective C(sp
2
)−H
Functionalization
Chemical Reviews Review
DOI: 10.1021/acs.chemrev.6b00692
Chem. Rev. 2017, 117, 8908−8976
8910
group to tolerate N-heterocyclic derivatives, allowing for the
synthesis of a small library of chiral metallocenes in excellent
yield and enantioselectivity (12a−12d ).
42
Notably, all three of
these methodologies employ Pd(OAc)
2
as a precatalyst, cesium
carbonate as an inorganic base, and the readily available (R)-
BINAP as ligand, demonstrating the generality of these
conditions.
Recently, Guiry and co-workers applied this methodology to
the synthesis of a new family of chiral ferrocenyl diols from
dibromide 13 (Scheme 8).
43
Both the prior reports from the
groups of Gu and You
40
and Kang and Gu,
41
had disclosed the
synthesis of the C
2
symmetric dione 14 in excellent
enantioselectivity via a double cyclization event. Further
optimization by Guiry et al. enabled a multigram, chromatog-
raphy-free synthesis of this key intermediate, while simulta-
neously decreasing catalyst and ligand loadings. These products
were derivatized via a stereoselective double nucleophilic
addition to provide diols 15, and studies exploring the
application of these new scaffolds as organocatalysts in an
asymmetric Diels−Alder reaction were reported.
In 2016, the Gu and You groups demonstrated that the
intramolecular C−H alkenylation of metallocenes 16 is also
possible under Pd(0)/Pd(II) catalysis, again employing BINAP
as the stereocontrolling element (Scheme 9).
44
A range of
functional groups were well-tolerated, and notably, the reaction
could be combined with an intramolecular C−H arylation,
enabling a diastereo- and enantioselective synthesis of planar
chiral ferrocenes 17.
Chiral phosphoric acids are also suited for the enantioselective
construction of planar chiral ferrocenes (Scheme 10).
45
In 2016,
Ye, Duan, and their co-workers reported an intramolecular
asymmetric C−H arylation of ketones 18 with BINOL-derived
phosphoric acid L3, delivering chiral ferrocenes 19 in up to 83%
ee. The authors hypothesize that in situ deprotonation of L3 with
cesium carbonate forms the corresponding chiral phosphate,
Scheme 6. Synthesis of Dibenzazepinones via an Intramolecular C−H Arylation
Scheme 7. Pd(0)/Pd(II)-Catalyzed Routes to Fluoreno- and
Pyridylmetallocenes
Scheme 8. Synthesis of Novel Ferrocenyl Diol Catalysts via a Double C−H Functionalization Strategy
Scheme 9. Pd(0)/Pd(II)-Catalyzed C −H Alkenylation of
Ferrocenes
Chemical Reviews Review
DOI: 10.1021/acs.chemrev.6b00692
Chem. Rev. 2017, 117, 8908−8976
8911
which serves as a coordinated base to the Pd center and facilitates
an enantioselective C−H activation through a CMD mechanism.
The synthesis of chiral quinilino-metallocenes has also been
achieved via a closely related enantioselective Pd(0)/Pd(II) C−
H functionalization, as reported by both the research groups of
Gu
46
and of Liu and Zhao
47
(Scheme 11). In these studies,
carboxamides 21 were converted to planar chiral derivatives 22
using Pd(OAc)
2
and cesium carbonate, with either Carreira’s O-
PINAP derivative L4 (Gu) or TADDOL-derived phosphor-
amidite L5 in the presence of pivalic acid (Liu and Zhao). The
latter system provided superior enantioselectivities (including
examples with identical substrates). Interestingly, the addition of
pivalic acid to Gu’s reaction with O-PINAP L4 had no effect on
enantioselectivity, suggesting that phosphoramidite L5 is the
primary contributor to the greater level of enantiocontrol in Liu
and Zhao’s work.
C−H functionalization strategies have also been employed to
access heteroatom-chiral molecules. The earliest enantioselective
approach toward Si-stereocenters, reported by Shintani, Hayashi,
and their co-workers in 2012,
48
is an extension of Shimizu’s
49
Pd(0)-catalyzed synthesis of achiral Si-bridged biaryls (Scheme
12). Desymmetrization of triarylsilanes 23 with Pd(OAc)
2
,an
amine base, and the electron-rich Josiphos-type ligand L6
proceeded with high chemo- and enantioselectivity, to yield Si-
stereogenic dibenzosiloles 24. Formation of the undesired
isomer 25 was largely retarded with all chiral ligands screened,
and enantioselectivities ranged from 75−97%. Alkyl (24a),
trifluoromethyl (24b), and indole (24c) triflates were all well-
tolerated, and differentiation between appropriately substituted
aryl groups proved possible (24d). Competition experiments
between the triflate precursors 24a and 24b demonstrated that
electron-poor substrates reacted significantly faster; these studies
in conjunction with kinetic isotope experiments led the authors
to propose that oxidative addition is likely the rate-determining
step.
Intramolecular Pd(0)/Pd(II) C−H functionalization reac-
tions have also been empl oyed to access P-stereoge nic
compounds. P-Chiral ligands have played a significant role in
the development of asymmetric catalysis,
50,51
and their efficient
construction is of great importance.
52,53
The cyclization of
prochiral phosphinic amides 26 to azaphosphinine oxides 27 was
reported concurrently by Duan and co-workers
54
and by Liu, Ma,
and their co-workers
55
via almost identical protocols (Scheme
13). Both research groups screened a variety of chiral phosphine
ligands, observing that TADDOL-derived tetraphenyl phosphor-
amidites L2 and L5 provided the highest levels of enantiose-
lectivity. Good yields were observed for various aryl substituents,
Scheme 10. Pd(0)/Pd(II)-Catalyzed Enantioselective Synthesis of Ferrocenes Using a Chiral Phosphoric Acid
Scheme 11. Synthesis of Quinilino−Metallocenes by Gu, and
Liu and Zhao
Scheme 12. An Intramolecular C−H Functionalization Route to Si-Stereogenic Dibenzosiloles
Chemical Reviews Review
DOI: 10.1021/acs.chemrev.6b00692
Chem. Rev. 2017, 117, 8908−8976
8912