Divergence between organometallic and single-electron-transfer mechanisms in copper(II)-mediated aerobic c-h oxidation

Article abstract of DOI:10.1021/ja4026424

Copper(II)-mediated C-H oxidation is the subject of extensive interest in synthetic chemistry, but the mechanisms of many of these reactions are poorly understood. Here, we observe different products from Cu^II-mediated oxidation of N-(8-quinoli

Full text of DOI:10.1021/ja4026424

Article
pubs.acs.org/JACS
Divergence between Organometallic and Single-Electron-Transfer
Mechanisms in Copper(II)-Mediated Aerobic C−H Oxidation
Alison M. Suess,† Mehmed Z. Ertem,*,§ Christopher J. Cramer,§ and Shannon S. Stahl*,†
†Department of Chemistry, University of Wisconsin−Madison, 1101 University Avenue, Madison, Wisconsin 53706, United States
§Department of Chemistry, Chemical Theory Center, and Supercomputing Institute, University of Minnesota, 207 Pleasant Street SE,
Minneapolis, Minnesota 55455, United States
S
* Supporting Information
ABSTRACT: Copper(II)-mediated C−H oxidation is the subject of
extensive interest in synthetic chemistry, but the mechanisms of many of
these reactions are poorly understood. Here, we observe different products
from CuII-mediated oxidation of N-(8-quinolinyl)benzamide, depending on
the reaction conditions. Under basic conditions, the benzamide group
undergoes directed C−H methoxylation or chlorination. Under acidic
conditions, the quinoline group undergoes nondirected chlorination.
Experimental and computational mechanistic studies implicate an organo-
metallic C−H activation/functionalization mechanism under the former
conditions and a single-electron-transfer mechanism under the latter
conditions. This rare observation of divergent, condition-dependent mechanisms for oxidation of a single substrate provides a
valuable foundation for understanding CuII-mediated C−H oxidation reactions.
activation is not well established.8 Many of these reactions
appear similar to those catalyzed by Pd and other noble metals,
INTRODUCTION
Copper(II) catalyzes a wide variety of aerobic C−H oxidation
■
reactions.1 Many of these transformations involve electron-rich
and we considered whether some of these aerobic oxidation
substrates that undergo initial one-electron oxidation by CuII.
reactions could involve organometallic intermediates. Stoichio-
metric C−H activation by CuII had been identified for
macrocyclic substrates,9 including the triazamacrocyclic arene
1, reported by Ribas et al. (eq 2).10,11 Several of these studies
Examples include oxidative dimerization of naphthol,2 oxy-
chlorination of phenols and other electron-rich arenes (eq 1),3,4
and cross-dehydrogenative coupling reactions of aromatic and
benzylic amines.1b,5 Yu and co-workers proposed a similar
single-electron-transfer (SET) mechanism for chelate-directed,
aerobic oxidative functionalization of 2-phenylpyridines
(Scheme 1),6 and support for this mechanism was obtained
from the lack of a kinetic isotope effect (KIE).7
Numerous other Cu-catalyzed aerobic C−H oxidation
reactions, including chelated-directed and nondirected exam-
ples, have been reported for which the mechanism of C−H
revealed an unusual CuII disproportionation pathway that yields
aryl-CuIII and CuI products, as shown in eq 2. We recently
demonstrated the relevance of this reactivity to catalytic C−H
oxidation by showing that CuII salts, CuBr2 and Cu(ClO4)2
catalyze aerobic oxidation of the arene C−H bond in 1 to C−O
and C−N bonds in the presence of methanol and pyridone,
respectively (eq 3).12 An aryl-CuIII species, which undergoes
facile C−O or C−N reductive elimination to afford the
functionalized product,13 was directly observed and established
as a kinetically competent intermediate. This work provided the
first direct evidence for an organometallic pathway in Cu-
catalyzed aerobic C−H oxidation.
Scheme 1. Proposed SET Mechanism for Oxidative
Chlorination of 2-Phenylpyridine
The proof-of-principle studies depicted in eq 3 have an
important caveat: the macrocyclic nature of the substrate could
Received: March 14, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja4026424 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
a
Table 1. Directed C−H Methoxylation of the Benzamide
base
atm
%
artificially stabilize the aryl-CuIII intermediate and enforce an
otherwise-unfavorable organometallic pathway. In order to
explore the potential role of organometallic intermediates with
a less biased substrate, we investigated the reactivity of substrate
2a, which features an amidoquinoline directing group first used
by Daugulis and co-workers in Pd-catalyzed C−H functional-
ization reactions.14,15 Here, we show that substrate 2a exhibits
divergent reactivity in CuII-mediated C−H functionalization,
depending on the reaction conditions. Directed C−H oxidation
occurs under basic conditions, while nondirected oxidation of
the quinoline ring occurs under acidic conditions. Experimental
and computational mechanistic studies provide fundamental
insights into the organometallic and SET pathways that give
rise to these distinct outcomes. These results have important
implications for the ongoing development of CuII-catalyzed
aerobic oxidation reactions.
entry
[Cu]
(1 equiv)
(1 atm)
3a
% 4a
b
1
10 mol %
Cu(ClO4)2·6H2O
10 mol %
Cu(ClO4)2·6H2O
2 equiv Cu(OAc)2
2 equiv Cu(OAc)2
1 equiv CuCl2
none
O2
0.0
n/a
b
2
Cs2CO3
O2
18
n/a
c
3
Cs2CO3
Cs2CO3
Na2CO3
Na2CO3
O2
N2
O2
56
46
31
2.0
n/a
n/a
c
4
d
5
10.8
41.2
d
6
2 equiv CuCl2
O2
a
b
1
Yields determined by H NMR spectroscopy. Conditions: 50 mM
2a, 5 mM [Cu], 0 or 50 mM base, 1 mL MeOH, 60 °C, 17 h. 50 mM
2a, 100 mM [Cu], 50 mM base, 1 mL 5.2:1 MeOH/pyridine, 50 °C,
24 h. 50 mM 2a, 50 or 100 mM [Cu], 50 mM base, 1 mL MeOH, 50
c
d
°C, 17 h.
(Table 1), the quantity of 3a increased only 8.6% after 5 h,
whereas use of 2a alone affords a 25% yield of 3a over the same
time period. Analogous product inhibition has been reported
for Cu-catalyzed amination of 2-phenylpyridine.8a
The reactivity of 2a was then evaluated under conditions
similar to those used for the oxidative chlorination of electron-
rich arenes (cf. eq 1). In the presence of 1 atm O2 and 2 equiv
LiCl in AcOH at 100 °C, catalytic CuCl or CuCl2 promoted
chlorination of 2a. In contrast to the methoxylation results
described above, the product arose from chlorination of the
quinoline ring (5a), not the benzamide ring (Table 2). Higher
a
RESULTS AND DISCUSSION
Table 2. C−H Chlorination of the Quinoline
■
CuII-Mediated C−H Oxidation of N-(8-quinolinyl)-
benzamide under Different Conditions. Our investigation
of the reactivity of 2a was initiated by evaluating CuII-catalyzed
C−H oxidation conditions similar to those reported previously.
No reaction was observed under the conditions used for
oxidative methoxylation of the triazamacrocyclic arene 1 (cf. eq
3; Table 1, entry 1). Addition of a stoichiometric base
(Cs2CO3) to the reaction resulted in small quantities of product
3a, arising from methoxylation of the ortho position of the
benzamide (entry 2). The yield of 3a could be increased up to
56% by employing 2 equiv of Cu(OAc)2 as the CuII source. No
difunctionalization of the arene was observed under these
conditions, and the presence of O2 enabled a somewhat higher
yield relative to anaerobic reaction conditions (entries 3 and 4).
Use of CuCl2 resulted in a mixture of methoxylation and
chlorination products 3a and 4a (entries 5 and 6). Control
experiments suggest that the 3a formed under these conditions
arises from direct C−H methoxylation, not C−H chlorination
followed by methoxide substitution of chloride in 4a.
Subjection of 4a to the oxidative methoxylation conditions in
entry 3 resulted in near-complete recovery of 4a with no
formation of 3a.
entry
[Cu]
additive
% 5a % 4a
1
2
3
4
5
20 mol % CuCl2
20 mol % CuCl
none
2 equiv LiCl
2 equiv LiCl
none
34
81
0.0
0.0
88
0.0
2.1
0.0
0.0
0.0
none
2 equiv LiCl
20 mol % CuCl2, 20 mol % LiOAc 2 equiv LiCl
a
Yields determined by 1H NMR spectroscopy. Conditions: 50 mM 2a,
10 mM [Cu], 100 mM LiCl in 1 mL AcOH, 1 atm O2, 100 °C, 17 h.
yields were obtained with CuCl rather than CuCl2 (entries 1
and 2), and no conversion was observed under these conditions
in the absence of Cu (entries 3 and 4). The reactivity with
CuCl2 was nearly identical to that of CuCl when an equal
amount of LiOAc (20 mol %) was included in the reaction
mixture (entries 2 and 5). This effect is attributed to the role of
a Brønsted base to promote substrate binding to the Cu center:
The use of a CuI catalyst precursor generates an equivalent of
base in situ upon reaction with O2 to afford H2O2 or H2O.16
These product yields are not sufficiently high to be
synthetically useful, but they provide clear evidence for directed
C−H oxidation, mediated by CuII. Control experiments show
that the product inhibits the reaction. For example, when a 1:1
mixture of 2a:3a was combined under the conditions of entry 3
B
dx.doi.org/10.1021/ja4026424 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 2. Divergent Reactivity in CuII-Mediated C−H
Oxidation of 2a
Scheme 4. Deuteration and Isotope Effect Experiments for
the Benzamide C−H Methoxylation Reaction
Overall, these observations, summarized in Scheme 2,
highlight an unusual condition-dependent selectivity for CuII-
mediated C−H oxidation. With these empirical data in hand,
we turned our attention to understanding the mechanistic basis
for these observations.
Mechanistic Study of CuII-Mediated C−H Oxidation of
the Benzamide Ring. In order to gain further insight into the
methoxylation reaction, substrates with electronically varied
substituents on the benzamide ring were prepared (2a−d,
Scheme 3). Initial rates were measured for the Cu(OAc)2-
the independent rates of the H- and D-labeled substrates
(Scheme 4c). Reaction of the monodeuterated substrate 2e led
to an 84:16 ratio of product isotopologues, favoring reaction of
the C−H bond (KIE = 5.25
0.06). A similar KIE was
obtained from the comparison of the initial rates of 2a and the
aryl-d5 substrate 2f (KIE = 5.7 0.8; Scheme 4c).
Scheme 3. CuII-Mediated Methoxylation from Methanol to
Yield Substituted Anisoles
When the N-methyl substrate 6a was submitted to the
reaction conditions, no C−H functionalization was observed
after 24 h. Most of the starting material was recovered cleanly,
together with an approximately 20% yield of methyl benzoate
arising from methanolysis of the tertiary amide (Table S3).
promoted methoxylation of these substrates (Table S6), and
faster rates were observed with substrates bearing electron-
withdrawing groups. A Hammett plot using σmeta parameters
reveals a small, but distinct, positive slope reflecting this
electronic effect (ρ = 0.6 0.1; Figure 1).
When the reaction is performed in CH3OD, no deuterium is
incorporated into the recovered starting material or the product
(Scheme 4a), suggesting that C−H activation is irreversible.
Kinetic isotope effects were obtained by an intramolecular
competition experiment (Scheme 4b) and by comparison of
Mechanistic Study of Cu-Catalyzed C−H Chlorination
of the Quinoline Ring. When the quinoline chlorination
reaction was performed in CD3CO2D, no deuterium
incorporation was observed in the recovered starting material
or the product (Scheme 5a). This result indicates that cleavage
of the C−H bond is irreversible. Kinetic isotope effects were
obtained by an intermolecular competition experiment and by
comparing the independent rates of the H- and D-labeled
substrates. A 1:1 mixture of 2a and the dideuterated substrate
2g was submitted to the reaction conditions for 30 min,
resulting in ∼5% product yield (Scheme 5b). Analysis of the
H/D ratio at the 7-position of the quinoline revealed a nearly
equal mixture of isotopes, corresponding to a KIE of 1.04
0.05. Independent reactions of 2a and 2g exhibit nearly
identical time courses (Figure S9), corresponding to a KIE ∼ 1
(Scheme 5c).
Coordination of 2a to CuII as an amidate prior to
chlorination, even under these acidic conditions, is supported
by the lack of reaction of the N-methyl substrate 6a. This
substrate is fully recovered after 17 h under the standard
reaction conditions. This observation aligns with our
interpretation of the beneficial effect of a Brønsted base on
the chlorination reaction (Table 2, entry 5).
Figure 1. Hammett plot from initial rates of 2a−d methoxylation.
C
dx.doi.org/10.1021/ja4026424 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
These considerations provided the basis for the use of
density functional theory (DFT) calculations to gain further
insights into both of the proposed mechanisms.
Scheme 5. Deuteration and Isotope Effect Experiments for
the Quinolinyl C−H Chlorination Reaction
Computational Analysis of an Organometallic Path-
way for CuII-Mediated C−H Oxidation. DFT methods were
used to assess the energetic viability of an organometallic C−H
functionalization pathway originating from the CuII-2a
complex. The calculations were performed at the M06-L level
of density functional theory,20 incorporating solvation effects
via the appropriate SMD continuum solvation model21 (see
Supporting Information for details). A number of CuII
complexes bearing the 8-amidoquinoline ligand were optimized
to evaluate the relative energies of possible ground-state
structures (Table 3). The square-planar carbonate complex
Table 3. Calculated Stability of Possible Cu-2a Complexes
under the Benzamide C−H Methoxylation Conditions
[LCuII(κ2-CO3)]− (I) (L = deprotonated 2a) was the most
stable structure identified. Species in which the carbonate was
replaced with an acetate ligand, methoxide ligands, or five-
coordinate structures with a coordinated pyridine ligand were
less stable.
Preliminary Mechanistic Assessment. The results out-
lined above show that CuII-mediated C−H oxidation of 2a
affords different products and exhibits significantly different
mechanistic features under different reaction conditions. On
the basis of literature precedents (see Introduction), we
attribute these results to the operation of organometallic and
SET C−H functionalization pathways for the benzamide
methoxylation and quinoline chlorination, respectively.
Starting from the [LCuII(κ2-CO3)]−complex I, a C−H
activation pathway was identified that proceeds via transition
state I-TS (Scheme 6). In this mechanism, formation of the
aryl-CuII complex II involves deprotonation of the arene C−H
bond by the carbonate ligand with concomitant Cu−C bond
formation, analogous to the CMD pathways that have been
elaborated for PdII-mediated C−H activation.19 The activation
free energy (ΔG‡) for this step is 25.9 kcal/mol, which
compares very favorably to the experimental ΔG‡ of ∼26 kcal/
mol calculated from the reaction half-life (9−24 h) at 50 °C.
The computed KIE for this step (kH/kD = 4.9) is also very
similar to the experimental value (kH/kD = 5.7). The
[LCuII(OMe)2]− complex is 6.2 kcal/mol less stable than the
carbonate complex (Table 3), but this species mediates C−H
activation with a barrier of only 20.4 kcal/mol (Figure S10).
This comparatively low barrier is consistent with a beneficial
effect of the basic methoxide ligand in the CMD C−H
activation step. The net activation free energy of 26.6 kcal/mol,
which accounts for the ground-state destabilization of this
complex, is only slightly higher than the pathway involving the
carbonate complex I. An analogous pathway involving the
acetate complex LCuII(κ2-OAc) is significantly higher in energy
(net ΔG‡ = 45.5 kcal/mol), largely because the acetate complex
starts 16.6 kcal/mol higher in energy than the carbonate
complex (Table 3).
The positive Hammett slope observed for chelate-directed
methoxylation of the benzamide ring (Figure 1) is inconsistent
with an SET mechanism, which strongly favors electron-rich
substrates.17 Organometallic C−H activation reactions also
typically favor electron-rich substrates, albeit to a lesser extent
than SET reactions.18 Several studies of PdII-mediated C−H
activation, however, show that some of these reactions can
proceed more rapidly with electron-deficient substrates.19 For
example, Fagnou and Gorelsky demonstrated C−H arylation
reactions of pyridine N-oxides that proceed via rate-limiting C−
H activation (kH/kD = 3.3) and exhibit a positive Hammett
slope (ρ = 1.53).19f These results have been interpreted within
the framework of a “concerted metalation-deprotonation”
(CMD) mechanism in which C−H activation involves
simultaneous electrophilic activation of the arene and
deprotonation of the C−H bond. A positive Hammett slope
can be observed for reactions in which C−H deprotonation
contributes strongly to the transition state, and this pathway is
consistent with the large KIE that we observe in the CuII-
mediated C−H methoxylation reaction (KIE ∼ 5−6).
The nonchelate-directed chlorination of the quinoline ring of
2a is consistent with SET reactivity. No KIE is observed for this
reaction (KIE ∼ 1), and the regioselectivity is consistent with
the electronic directing effect of benzamide nitrogen.
Deprotonation of the amide group (i.e., upon coordination to
the Cu center) would enhance this electronic effect. Similar
electronic effects have been observed in SET-based chlorination
of other electron-rich arenes.3,4
The energetics of one-electron oxidation of the aryl-CuII
complex II to an aryl-CuIII complex were evaluated by
considering possible CuII/CuI redox couples. A number of
plausible CuII and CuI species that could be present in the
reaction solution were examined (Table 4). Assuming ligand
exchange in solution is facile, the most stable CuII and CuI
species, CuII(py)2(CO3) and CuI(py)(OMe), provide the
D
dx.doi.org/10.1021/ja4026424 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 6. Calculated Mechanism for the CuII-Mediated Benzamide C−H Methoxylation Reaction
sphere following oxidation of CuII to CuIII. This sequence may
vary with other substrates, for example, with those lacking the
conformational constraints of the chelating ligand. The
reductive elimination transition state VI-TS is similar in energy
to the C−H activation transitions state I-TS. The calculated
pathway is consistent with the experimental observation of rate-
limiting C−H activation given the expected accuracy of the
computational model. Uncertainties in the computed energies
derive in part from the relative pKa values of methanol and
carbonate, and they may be amplified by the multiple
deprotonations of MeOH by carbonate that are performed
along the reaction path.
Table 4. Redox Couples Calculated for Oxidation of Aryl-
CuII to Aryl-CuIII in the Benzamide C−H Methoxylation
Reaction in Methanol
[CuII]
[CuI]
relative
ΔG
relative
ΔG
(CuII, CO32−, 2 MeOH,
2 OAc−, 2 py)
(CuI, CO32−, 2 MeOH,
2 OAc−, 2 py)
(kcal/mol)
(kcal/mol)
CuII(py)2(CO3)
CuII(py)2(OMe)2
CuII(py)(CO3)
0.0
4.2
CuI(py)(OMe)
CuI(py)(OAc)
CuI(py)(HCO3)
CuI(py)2(OAc)
0.0
7.4
4.3
10.2
11.6
13.3
16.7
CuII(OMe)2(MeOH)2
CuII(MeOH)2(CO3)
8.5
−
8.7
CuI(py)(OAc)2
C−H activation by CuIII was also considered (Scheme 7,
lower pathway)22 and found to have an activation barrier 5.6
CuII(OMe)(HCO3)
(py)2
14.5
CuI(MeOH)(OAc)
CuI(MeOH)2(CO3)−
CuI(MeOH)2(OAc)
CuII(OMe)(HCO3)
(MeOH)2
15.1
21.1
25.5
Scheme 7. CuII vs CuIII C−H Activation Mechanisms
CuII(py)2(OAc)2
16.4
16.7
18.0
CuII(MeOH)2(OAc)2
CuII(py)(MeOH)
(OAc)2
relevant equilibrium redox couple for oxidation of aryl-CuII to
aryl-CuIII.
Oxidation of aryl-CuII complex II, which has a bicarbonate
ligand, is energetically disfavored (III, ΔG = 28.0 kcal/mol
relative to I; Scheme 6). Exchange of the bicarbonate ligand
with methoxide to form IV, however, is favorable, and
subsequent oxidation of IV to afford aryl-CuIII species V is
uphill by only 1.2 kcal/mol. These observations can be
rationalized by the ability of the methoxide ligand to stabilize
CuIII relative to a bicarbonate ligand.
Reductive elimination from the 4-coordinate CuIII(aryl)-
(OMe) complex V was found to be prohibitive (Figure S11,
ΔG‡ = 44.2 kcal/mol). The five-coordinate dimethoxide
complex VI is energetically accessible, however, and reductive
elimination from this species is more favorable (Scheme 6, ΔG‡
= 28.2 kcal/mol relative to I). This observation suggests that
the reactive methoxide nucleophile enters the coordination
kcal/mol lower in energy than the corresponding CuII-mediated
step (ΔG‡Cu(III) = 20.3 kcal/mol). Nevertheless, this pathway is
strongly disfavored by the highly unfavorable energy for
oxidation of the CuII center to CuIII (ΔG = 39.4 kcal/mol)
prior to C−H activation. A concerted proton-coupled electron-
transfer process cannot be excluded on the basis of our results,
E
dx.doi.org/10.1021/ja4026424 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
but the computational data presented here are consistent with a
stepwise CuII-mediated C−H activation step followed by one-
electron oxidation: I → IV → V (Schemes 6 and 7).
and were used to calculate the energy of the SET step. One-
electron oxidation of VIII to the CuII complex IX, bearing a
radical-cation amidoquinoline ligand, is thermodynamically
uphill by 8.6 kcal/mol (Scheme 8). This step is analogous to
the CuII disproportionation step that affords the aryl-CuIII
species in the organometallic mechanism (i.e., IV → V in
Scheme 6); however, in the present case, ligand oxidation is
favored over oxidation of the Cu center based on analysis of the
computed spin and charge densities.
Computational Analysis of a SET Pathway for CuII-
Mediated C−H Oxidation. Chlorination of the quinoline ring
takes place under conditions similar to those for reactions that
have been proposed to proceed by an SET mechanism. Our
initial studies probed the possibility of an intramolecular SET
step in complex VIII (eq 4), similar to that illustrated in
This 8.6 kcal/mol thermodynamic barrier for the SET step is
much lower than the excitation energy for the intramolecular
SET described above, and it is also much lower than one-
electron oxidation of the noncoordinated neutral substrate 2a
(ΔG = 31.0 kcal/mol; eq 5). The latter observation shows that
coordination of the substrate to CuII as an amidate ligand
greatly facilitates intermolecular SET.
Scheme 1. The product corresponds to an electronic excited
state of the ground-state species, and the energy of this state
was evaluated by performing TD-DFT calculations. The lowest-
energy internal SET product arises from transfer of a β-spin
electron from a mixture of two doubly occupied orbitals, one
associated with the chloride lone pairs (GS-A) and the other
with quinoline π electrons (GS-B), to the singly occupied
2
2
orbital localized on Cu (i.e., dx −y ) (Figure 2 and Table S7).
Various pathways were considered for chlorination of the
radical-cation complex IX. The lowest barrier was obtained
from a process involving chlorine-atom transfer from CuCl2,
with explicit modeling of an equivalent of chloride and acetic
acid (Scheme 8). This step proceeds via transition state IX-TS
and exhibits an activation energy of ΔG‡ = 29.8 kcal/mol.
Deprotonation of the resulting Wheland-type intermediate X
affords the quinoline chlorination product XI. The overall
barrier for this computed pathway exhibits good agreement
with the estimated experimental barrier of ∼30 kcal/mol
derived from the reaction half-life at 100 °C (8.5−17 h; energy
estimated based on a pseudo first-order reaction). The
computed KIE for this step, kH/kD = 0.91 is also similar to
the value determined experimentally (kH/kD = 1).
The similarity between the experimental and computed
barriers is almost certainly fortuitous, considering the nature of
our redox-couple calculation. For example, it is possible that
more stable CuII species may be present in solution than those
found in Table 5, thereby increasing the thermodynamic barrier
for the SET step. A considerably more sophisticated treatment
of aggregates and specific solvation effects would be required to
assess this point further, but we consider it unlikely that such
results would increase the energies of the species in Scheme 8
by more than 3 or 4 kcal/mol, which is within the limits of the
expected uncertainty for these calculations.
Figure 2. The lowest excited-state doublet derives from excitation of a
beta electron from the nominally doubly occupied HOMO-1 (GS-A)
and HOMO (GS-B), contributing with equal weight, into the singly
occupied SOMO (ES). The resulting excited state is 31.4 kcal/mol
above the ground state, corresponding to a photon energy of 909.4
nm.
The excitation energy for this transition is 31.4 kcal/mol.
Because chlorination of the aromatic ring would add a further
energy barrier, however, we considered whether an intermo-
lecular SET process might exist that provides access to a lower
energy pathway.
Various CuII/CuI redox couples were considered for the
intermolecular SET reaction (Table 5). CuIICl2(AcOH)2 and
CuICl2− were the lowest-energy CuII and CuI species identified
Reasons for the Mechanistic Divergence under
Different Reaction Conditions. The experimental and
computational results described above provide several insights
that can account for the change in mechanism and product
selectivity under different reaction conditions. Carbonate is an
important additive in the benzamide C−H methoxylation
reaction (Table 1), and the computational results draw
attention to at least two important roles for carbonate. First,
a basic ligand is essential for activation of the arene C−H bond
by an organometallic CMD mechanism. Acetate could also
perform such a role, but carbonate (or methoxide, Figure S10)
appears to be much more effective, probably owing to its higher
basicity. Second, the carbonate dianion is a good ligand that
stabilizes CuII and CuIII species relative to complexes with
Table 5. Redox Couples Calculated for the Oxidation of CuII
to Aryl-Radical-Cation-Ligated CuII Complex during
Quinoline C−H Chlorination in Acetic Acid
[CuII]
[CuI]
(CuII, 4CI−, 2
relative ΔG
(CuI, 4CI−, 2
AcOH)
relative ΔG
AcOH)
(kcal/mol)
(kcal/mol
CuII(CI)2(AcOH)2
CuIICI3(AcOH)−
0.0
1.6
CuICl2
0.0
2.0
−
CuICI2(AcOH)−
CuICI(AcOH)
CuICI(AcOH)2
CuICI
CuIICI4
4.0
10.4
11.9
20.3
2−
CuIICI2(AcOH)
CuIICI2
5.9
18.6
F
dx.doi.org/10.1021/ja4026424 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 8. Calculated Mechanism for the CuII-Mediated Quinoline C−H Chlorination Reaction
acetate or bicarbonate ligands (cf. Table 3 and Scheme 6). This
effect is particularly important to access the aryl-CuIII
intermediate that can undergo facile C−O bond formation.
The carbonate ligand properties that favor organometallic
C−H oxidation disfavor SET C−H oxidation. Specifically,
coordination of carbonate to CuII will lower the CuII/CuI
reduction potential and disfavor SET from the aromatic
substrate. The acidic reaction conditions associated with
chlorination of the quinoline ring lack strong donor ligands,
and CuII will be present in the form of more-oxidizing halide
species. Redox couples were computed for various Cu species
to provide further insight into this issue. As shown in Table 6,
changing from basic to acidic reaction conditions. The presence
of a Brønsted basic ligand on the CuII center facilitates C−H
activation by an organometallic mechanism, while acidic
conditions enhance the CuII reduction potential, thereby
favoring SET. The results of this study show that a macrocyclic
chelate (as in 1; eq 2) is not required to achieve organometallic
C−H activation by CuII. The experimental and computational
results highlight the proton-transfer component associated with
CuII-mediated C−H activation. On the basis of this insight, it
seems reasonable to speculate that CuII-catalyzed oxidations of
substrates containing acidic C−H bonds follow pathways
analogous to the organometallic pathway shown in Scheme 6.
Many precedents for C−H oxidations of this type, including
reactions with alkynes and various aromatic heterocycles, are
accomplished with substrates that lack directing groups
altogether. Collectively, these insights provide a valuable
foundation for continued efforts to expand the scope of Cu-
catalyzed aerobic C−H oxidation reactions.
Table 6. Computed Redox Potentials of Species Available in
Acidic and Basic Reaction Conditions
ΔG
E (V) vs
NHE
redox couples
(kcal/mol)
Acidic Conditions:
CuIICl2(AcOH)2 + e− → CuICl2 + 2AcOH
−102.8
−111.4
0.18
0.55
−
ASSOCIATED CONTENT
VIII + e− → IX
■
Basic Conditions:
S
* Supporting Information
CuII(py)2CO3−+ CH3OH + e− → CuI(py)(OCH3) +
−80.2
−0.80
Synthetic procedures, characterization, additional data, and
computational data. This material is available free of charge via
the Internet at http://pubs.acs.org
py + HCO3
III + e− → II
V + e− → IV
−90.7
−81.4
−0.34
−0.75
AUTHOR INFORMATION
the CuII/CuI reduction potentials computed for the major
species present under acidic and basic reaction conditions (the
top redox couple in each category) differ by nearly 1 V. While
these redox potentials have considerable uncertainty, the
qualitative trend is clear: acidic conditions are much more
amenable to SET-based oxidation of electron-rich substrates.
■
Corresponding Author
mzertem@gmail.com; stahl@chem.wisc.edu
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
CONCLUSIONS
■
■
In summary, we have identified an unusual switch in
mechanism and product identity in CuII-mediated C−H
oxidation of the amidoquinoline substrate 2a. The experimental
and computational results are consistent with a switch between
organometallic and SET-based C−H oxidation pathways upon
A.M.S. and S.S.S. thank Anh T. Nguyen for experimental
assistance. Financial support was provided by the DOE (A.M.S.
and S.S.S.: DE-FG02-05ER15690), and the University of
Minnesota and the National Science Foundation (CHE09-
52054) (M.Z.E. and C.J.C.).
G
dx.doi.org/10.1021/ja4026424 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(12) King, A. E.; Huffman, L. M.; Casitas, A.; Costas, M.; Ribas, X.;
Stahl, S. S. J. Am. Chem. Soc. 2010, 132, 12068−12073.
(13) For fundamental studies of the C−O/C−N reductive
elimination reactions, see: (a) Huffman, L. M.; Stahl, S. S. J. Am.
Chem. Soc. 2008, 130, 9196−9197. (b) Huffman, L. M.; Casitas, A.;
Font, M.; Canta, M.; Costas, M.; Ribas, X.; Stahl, S. S. Chem.Eur. J.
2011, 17, 10643−10650.
(14) For leading references, see: (a) Zaitsev, V. G.; Shabashov, D.;
Daugulis, O. J. Am. Chem. Soc. 2005, 127, 13154−13155. (b) Daugulis,
O.; Do, H.-Q.; Shabashov, D. Acc. Chem. Res. 2009, 42, 1074−1086.
(15) The first application of this ligand to Cu-mediated C−H
functionalization (non-aerobic) was reported recently: Tran, L. D.;
Popov, I.; Daugulis, O. J. Am. Chem. Soc. 2012, 134, 18237−18240.
(16) For a similar beneficial catalytic effect arising from in situ
generation of a base upon aerobic oxidation of CuI in alcohol oxidation
reaction, see: Hoover, J. M.; Ryland, B. L.; Stahl, S. S. J. Am. Chem. Soc.
2013, 135, 2357−2367.
(17) For example, in Cu-catalyzed oxidative chlorination (ref 4),
benzene, toluene, and anisole do not react, while dimethoxybenzene,
which is significantly more electron-rich, does react. As another
example, the reaction of PhOAc is 1000-fold faster than PhO2CCF3 in
CoIII-mediated oxidative trifluoroacetoxylation (ref 7), which is
proposed to proceed via SET.
(18) For a comparison of the electronic effects of ″electrophilic″ C−
H activation by PdII and other more-classical electrophiles, see: Stock,
L. M.; Tse, K.-t.; Vorvick, L. J.; Walstrum, S. A. J. Org. Chem. 1981, 46,
1757−1759.
REFERENCES
■
(1) (a) Wendlandt, A. E.; Suess, A. M.; Stahl, S. S. Angew. Chem., Int.
Ed. 2011, 50, 11062−11087. (b) Shi, Z.; Zhang, C.; Tang, C.; Jiao, N.
Chem. Soc. Rev. 2012, 41, 3381−3430. (c) Punniyamurthy, T.; Rout, L.
Coord. Chem. Rev. 2008, 252, 134−154. (d) Gamez, P.; Aubel, P. G.;
Driessen, W. L.; Reedijk, J. Chem. Soc. Rev. 2001, 30, 376−385.
(e) Liu, C.; Zhang, H.; Shi, W.; Lei, A. Chem. Rev. 2011, 111, 1780−
1824.
(2) (a) Brunel, J. M. Chem. Rev. 2005, 105, 857−898. (b) Sainsbury,
M. Tetrahedron 1980, 36, 3327−3359. (c) Ashenhurst, J. A. Chem. Soc.
̌
́
̌ ̌ ̌
Rev. 2010, 39, 540−548. (d) Kocovsky, P.; Vyskocil, S.; Smrcina, M.
Chem. Rev. 2003, 103, 3212−3245. (e) Rosini, C.; Franzini, L.;
Raffaelli, A.; Salvadori, P. Synthesis 1992, 1992, 503−517. (f) Wang, H.
Chirality 2010, 22, 827−837.
(3) (a) Menini, L.; Gusevskaya, E. V. Chem. Commun. 2006, 209−
211. (b) Menini, L.; Gusevskaya, E. V. Appl. Catal., A 2006, 309, 122−
128. (c) Menini, L.; Parreira, L. A.; Gusevskaya, E. V. Tetrahedron Lett.
2007, 48, 6401−6404. (d) Menini, L.; da Cruz Santos, J. C.;
Gusevskaya, E. V. Adv. Synth. Catal. 2008, 350, 2052−2058.
(4) Yang, L.; Lu, Z.; Stahl, S. S. Chem. Commun. 2009, 45, 6460−
6462.
(5) For leading references, see: (a) Boess, E.; Schmitz, C.;
Klussmann, M. J. Am. Chem. Soc. 2012, 134, 5317−5325. (b) Li, C.-
J. Acc. Chem. Res. 2009, 42, 335−344. (c) Li, C.-J.; Li, Z. Pure Appl.
Chem. 2006, 78, 935−945. (d) Li, Z.; Bohle, D. S.; Li, C.-J. P. Natl.
Acad. Sci. 2006, 103, 8928−8933.
(6) Chen, X.; Hao, X.-S.; Goodhue, C. E.; Yu, J.-Q. J. Am. Chem. Soc.
2006, 128, 6790−6791.
(19) See, for example: (a) Ryabov, A. D.; Sakodinskaya, I. K.;
Yatsimirsky, A. K. J. Chem. Soc., Dalton Trans. 1985, 14, 2629−2638.
(b) Karig, G.; Moon, M.-T.; Thasana, N.; Gallagher, T. Org. Lett.
2002, 4, 3115−3118. (c) Davies, D. L.; Donald, S. M. A.; Macgregor,
S. A. J. Am. Chem. Soc. 2005, 127, 13754−13755. (d) García-
Cuadrado, D.; Braga, A. A. C.; Maseras, F.; Echavarren, A. M. J. Am.
Chem. Soc. 2006, 128, 1066−1067. (e) Gorelsky, S. I.; Lapointe, D.;
Fagnou, K. J. Am. Chem. Soc. 2008, 130, 10848−10849. (f) Balcells, D.;
Clot, E.; Eisenstein, O. Chem. Rev. 2010, 110, 749−823. (g) Sun, H.-
Y.; Gorelsky, S. I.; Stuart, D. R.; Campeau, L.-C.; Fagnou, K. J. Org.
Chem. 2010, 75, 8180−8189. (h) Ackermann, L. Chem. Rev. 2011, 111,
1315−1345. (i) Gorelsky, S. I.; Lapointe, D.; Fagnou, K. J. Org. Chem.
2012, 77, 658−668.
(7) The mechanism was further supported by analogy to CoIII-
promoted arene C−H oxidation: Kochi, J. K.; Tang, R. T.; Bernath, T.
J. Am. Chem. Soc. 1973, 95, 7114−7123.
(8) For an extensive review of such reactions, see ref 1a. For selected
examples, see: (a) Uemura, T.; Imoto, S.; Chatani, N. Chem. Lett.
2006, 35, 842−843. (b) Brasche, G.; Buchwald, S. L. Angew. Chem., Int.
Ed. 2008, 47, 1932−1934. (c) Ueda, S.; Nagasawa, H. Angew. Chem.,
Int. Ed. 2008, 47, 6411−6413. (d) Hamada, T.; Ye, X.; Stahl, S. S. J.
Am. Chem. Soc. 2008, 130, 833−835. (e) Do, H.-Q.; Daugulis, O. J.
Am. Chem. Soc. 2009, 131, 17052−17053. (f) Monguchi, D.; Fujiwara,
T.; Furukawa, H.; Mori, A. Org. Lett. 2009, 11, 1607−1610. (g) Wang,
Q.; Schreiber, S. L. Org. Lett. 2009, 11, 5178−5180. (h) Mizuhara, T.;
Inuki, S.; Oishi, S.; Fujii, N.; Ohno, H. Chem. Commun. 2009, 3413−
3415. (i) Zhang, C.; Jiao, N. J. Am. Chem. Soc. 2010, 132, 28−29.
(j) Wang, W.; Luo, F.; Zhang, S.; Cheng, J. J. Org. Chem. 2010, 75,
2415−2418. (k) Wang, H.; Wang, Y.; Peng, C.; Zhang, J.; Zhu, Q. J.
Am. Chem. Soc. 2010, 132, 13217−13219. (l) Matsuyama, N.; Kitahara,
M.; Hirano, K.; Satoh, T.; Miura, M. Org. Lett. 2010, 12, 2358−2361.
(m) John, A.; Nicholas, K. M. J. Org. Chem. 2011, 76, 4158−4162.
(n) Guru, M. M.; Punniyamurthy, T. J. Org. Chem. 2012, 77, 5063−
5073. (o) Ding, S.; Yan, Y.; Jiao, N. Chem. Commun. 2013, 49, 4250−
4252.
(20) Zhao, Y.; Truhlar, D. G. J. Chem. Phys. 2006, 125, 194101−
194118.
(21) Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B
2009, 113, 6378−6396.
(22) C−H activation by CuIII has been proposed in the presence of
hypervalent iodine reagents: (a) Phipps, R. J.; Grimster, N. P.; Gaunt,
M. J. J. Am. Chem. Soc. 2008, 130, 8172−8174. (b) Phipps, R. J.;
Gaunt, M. J. Science 2009, 323, 1593−1597. (c) Cho, S. H.; Yoon, J.;
Chang, S. J. Am. Chem. Soc. 2011, 133, 5996−6005. (d) Chen, B.;
Hou, X.-L.; Li, Y.-X.; Wu, Y.-D. J. Am. Chem. Soc. 2011, 133, 7668−
7671.
(9) For leading references, see: (a) Furuta, H.; Maeda, H.; Osuka, A.
J. Am. Chem. Soc. 2000, 122, 803−807. (b) Chmielewski, P. J.; Latos-
́
Grazynski, L.; Schmidt, I. Inorg. Chem. 2000, 39, 5475−5482.
̇
(c) Furuta, H.; Ishizuka, T.; Osuka, A.; Uwatoko, Y.; Ishikawa, Y.
Angew. Chem., Int. Ed. 2001, 40, 2323−2325. (d) Maeda, H.; Ishikawa,
Y.; Matsuda, T.; Osuka, A.; Furuta, H. J. Am. Chem. Soc. 2003, 125,
11822−11823. (e) Pawlicki, M.; Kan
Chem. 2007, 46, 6575−6584.
ska, I.; Latos-Grazynski, L. Inorg.
̇
(10) (a) Ribas, X.; Jackson, D. A.; Donnadieu, B.; Mahía, J.; Parella,
T.; Xifra, R.; Hedman, B.; Hodgson, K. O.; Llobet, A.; Stack, T. D. P.
Angew. Chem., Int. Ed. 2002, 41, 2991−2994. (b) Xifra, R.; Ribas, X.;
̀
Llobet, A.; Poater, A.; Duran, M.; Sola, M.; Stack, T. D. P.; Benet-
Buchholz, J.; Donnadieu, B.; Mahía, J.; Parella, T. Chem.Eur. J. 2005,
11, 5146−5156. (c) Ribas, X.; Calle, C.; Poater, A.; Casitas, A.;
Gom
Mitrikas, G.; Sola,
2010, 132, 12299−12306.
(11) For a related macrocyclic arene, see: Yao, B.; Wang, D.-X.;
Huang, Z.-T.; Wang, M.-X. Chem. Commun. 2009, 45, 2899−2901.
́
ez, L.; Xifra, R.; Parella, T.; Benet-Buchholz, J.; Schweiger, A.;
̀
M.; Llobet, A.; Stack, T. D. P. J. Am. Chem. Soc.
H
dx.doi.org/10.1021/ja4026424 | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX