Mechanism of the palladium-catalyzed arene C-H acetoxylation: A comparison of catalysts and ligand effects

Article abstract of DOI:10.1021/jacs.5b00238

This article describes detailed mechanistic studies focused on elucidating the impact of pyridine ligands on the Pd-catalyzed C-H acetoxylation of benzene. Three different catalysts, Pd(OAc)₂, Pd(OAc)₂/pyridine (1:1), and Pd(OAc)₂/pyridine (1:2), are compared using a combination of mechanistic tools, including rate and order studies, Hammett analysis, detailed characterization of catalyst resting states, and isotope effects. The data from these experiments implicate C-H activation as the rate-limiting step in all cases. The major difference between the three catalysts is proposed to be the resting state of Pd. Under the reaction conditions, Pd(OAc)₂ rests as an acetate bridged dimer, while the Pd(OAc)₂/pyridine (1:2) catalyst rests as the monomer (pyridine)₂Pd(OAc)₂. In contrast, a variety of experiments suggest that the highly active catalyst generated from the 1:1 combination of Pd(OAc)₂ and pyridine rests as the dimeric structure [(pyridine)Pd(OAc)₂]₂.

Full text of DOI:10.1021/jacs.5b00238

Article
pubs.acs.org/JACS
Mechanism of the Palladium-Catalyzed Arene C−H Acetoxylation: A
Comparison of Catalysts and Ligand Effects
Amanda K. Cook and Melanie S. Sanford*
Department of Chemistry, University of Michigan, 930 North University Avenue, Ann Arbor, Michigan 48109, United States
*
S Supporting Information
ABSTRACT: This article describes detailed mechanistic studies focused
on elucidating the impact of pyridine ligands on the Pd-catalyzed C−H
acetoxylation of benzene. Three different catalysts, Pd(OAc) , Pd(OAc) /
2
2
pyridine (1:1), and Pd(OAc) /pyridine (1:2), are compared using a
2
combination of mechanistic tools, including rate and order studies,
Hammett analysis, detailed characterization of catalyst resting states, and
isotope effects. The data from these experiments implicate C−H activation
as the rate-limiting step in all cases. The major difference between the
three catalysts is proposed to be the resting state of Pd. Under the reaction
conditions, Pd(OAc) rests as an acetate bridged dimer, while the Pd(OAc) /pyridine (1:2) catalyst rests as the monomer
2
2
(
pyridine) Pd(OAc) . In contrast, a variety of experiments suggest that the highly active catalyst generated from the 1:1
2 2
combination of Pd(OAc) and pyridine rests as the dimeric structure [(pyridine)Pd(OAc) ] .
2
2 2
INTRODUCTION
Over the past 15 years, there have been major advances in the
Over the past several years, our group has focused on
■
developing Pd catalysts for the non-directed C−H acetox-
8
1
ylation of arenes. The Pd-catalyzed C−H acetoxylation of
field of Pd-catalyzed C−H functionalization. Palladium
benzene was originally reported in the 1970s using simple Pd
catalysis enables the introduction of a diverse array of
functional groups, often in the context of complex organic
9
−11
salts and oxidants such as K
Cr
2
O
2
and K
S
2
O
2
.
8
While
7
2
1,2
molecules. However, the vast majority of synthetic and
these early studies provided proof-of-principle for the feasibility
of this transformation, the catalyst turnover numbers (TON)
were too low for practical utility (typically ranging between <1
and 7). In 1996, Crabtree reported that the combination of
Pd(OAc)2 as catalyst and PhI(OAc)2 as oxidant afforded
dramatically improved results, with up to 127 turnovers in the
3
mechanistic efforts in this area have focused on substrates
bearing directing groups (DG). These directing groups render
the reactions kinetically fast as well as highly site-selective for
C−H functionalization proximal to the directing groups
3
p,4
(
Scheme 1a).
In contrast, analogous non-directed Pd-
12
C−H acetoxylation of naphthalene. More recently, our group
has demonstrated that the rate and TON of Crabtree’s reaction
can be increased dramatically through the use of pyridine (pyr)
Scheme 1. Comparison of Directed and Non-Directed C−H
Functionalization
8c
as a supporting ligand. The ratio of Pd(OAc) to pyridine was
2
critical in this system, with a 1:1 ratio proving optimal. This
second-generation catalyst system provided an approximately
1
0-fold increase in reaction rate versus Pd(OAc) , and TONs of
2
>
4500 were achieved (Scheme 2). In addition, the site
Scheme 2. Pyridine Ligand Effects in the Pd-Catalyzed C−H
Acetoxylation of Benzene
catalyzed C−H functionalization processes have been much
5
less developed in terms of both synthetic applications and
3f,6
mechanistic analysis
(Scheme 1b). The identification of
efficient and selective catalysts for non-directed C−H
7
functionalization is of particular importance, because many
Received: January 8, 2015
synthetic targets do not contain directing groups.
©
XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
selectivity of this transformation could be tuned through
from 140 to 840 mM. Under these conditions, the reaction is
8a
modification of the pyridine ligand. Despite these advances,
the mechanistic role of the pyridine ligand remains poorly
understood. Notably, similar pyridine effects have been
approximately first order in benzene (1.5 ± 0.3) and zero order
(0.1 ± 0.1) in PhI(OAc) (Figures S18 and S21).
2
Finally, the kinetic isotope effect (KIE) was determined by
13
observed in related Pd-catalyzed C−H functionalization and
comparing the initial reaction rate with C H to that with C D .
6
6
6
6
14
oxidation reactions, underscoring the significance of mecha-
nistic understanding of ligand effects in these transformations.
This report describes a detailed mechanistic investigation
focused on elucidating the impact of pyridine ligands on Pd-
catalyzed benzene C−H acetoxylation. We compare three
catalyst systems, Pd(OAc) , Pd(OAc) /pyridine (1:1), and
A KIE (k /k ) of 4.5 ± 0.4 was obtained from these
H D
experiments (Figure S19). This value is similar to that observed
1
2a
under Crabtree’s conditions (k /k = 4.1) and is consistent
with a 1° isotope effect.
H
D
16
On the basis of all of the data presented above, we propose
the catalytic cycle in Scheme 3 for the Pd(OAc) -catalyzed C−
2
2
2
Pd(OAc) / pyridine (1:2), using a combination of mechanistic
2
tools, including rate and order studies, Hammett analysis,
detailed characterization of catalyst resting states, and isotope
effects. These investigations provide insights into the
similarities and differences between the three catalyst systems
that explain their dramatically different reactivities.
Scheme 3. Proposed Mechanism of the Pd(OAc) -Catalyzed
C−H Acetoxylation of Benzene
2
RESULTS AND DISCUSSION
Mechanistic Investigation of the Pd(OAc) -Catalyzed
■
2
C−H Acetoxylation of Benzene. We first assessed the order
in Pd for the Pd(OAc) catalyst system. In Crabtree’s original
2
12a
study, a half-order kinetic dependence on Pd was observed.
This result was rationalized based on a resting state dimer,
Pd(OAc) ] , that breaks up into monomeric Pd(OAc) to
[
2
2
2
effect catalysis. However, the conditions developed in our lab
are somewhat different than those reported by Crabtree (e.g.,
different concentration of benzene, different solvent system).
Thus, we needed to establish whether the same kinetic
dependence on Pd is observed in our system.
The method of initial rates was used to determine the rate of
product formation as a function of Pd concentration over a
concentration range of 0.28−56 mM. The reactions were
monitored using GC-FID and were run to approximately 10%
conversion. Hexafluorobenzene was employed as a co-solvent
in order to mitigate changes in the polarity of the reaction
medium upon varying the concentration of benzene (during
benzene order studies, vide infra). Under our reaction
H acetoxylation of benzene. The mechanism begins with a
II
dimeric Pd precatalyst, [Pd(OAc) ] (1a), as the resting state.
2 2
Complex 1a undergoes reversible dissociation into the
monomer 1b, which lies on the catalytic cycle. C−H activation
of benzene at 1b is the rate-determining step (RDS) and forms
II
the Pd -aryl intermediate 2. This complex is then oxidized by
PhI(OAc) to form a high-valent Pd intermediate 3, which
2
17
undergoes reductive elimination to release PhOAc and
regenerate 1b.
The rate expression for this mechanism can be derived as
shown below in eqs 1−3. This rate expression is fully consistent
with the experimental data, as it predicts a half-order
dependence on Pd, a first-order dependence on benzene, and
conditions, the Pd(OAc) -catalyzed C−H acetoxylation of
2
a zero-order dependence on PhI(OAc) . Additionally, since C−
benzene showed a half-order dependence on Pd (0.49 ± 0.06),
analogous to that observed by Crabtree (see Figure S23 in the
Supporting Information). This is consistent with the Pd resting
as a dimer in our system.
2
H activation is the RDS, a 1° KIE is expected.
d(PhOAc)
rate =
= k [1b][benzene]
2
dt
(1)
To further probe the impact of Pd aggregation state, we
examined the kinetic order in Pd using Pd(OTFA) (OTFA =
trifluoroacetate) as the catalyst. In contrast to Pd(OAc)2,
2
[1b]
1a]
K1 =
1/2
[
(2)
(3)
Pd(OTFA) is known to exist as a monomer in organic
2
1
5
solvents. Thus, we predicted that a first-order dependence on
Pd] should be observed with this catalyst. Indeed, under
rate = K k [1a]1/2[benzene]
1
2
[
otherwise analogous conditions, the Pd(OTFA) -catalyzed C−
Mechanistic Investigation of the Pd(OAc) /Pyr (1:2)-
Catalyzed C−H Acetoxylation of Benzene. We next
studied the catalyst generated upon combining a 1:2 ratio of
2
2
H acetoxylation of benzene was first order in [Pd] (1.0 ± 0.1;
see Figure S25). Notably, at 5.6 mM [Pd], the rate of C−H
acetoxylation using Pd(OTFA) is almost 4-times faster than
Pd(OAc) to pyridine (pyr). To assess the resting state of Pd
2
2
1
with Pd(OAc) (Δ[PhOAc]/Δt = 6.5 mM/h for Pd(OTFA)
during catalysis, the reaction was monitored by H NMR
spectroscopy. The only Pd species observed at the beginning
2
2
18
and Δ[PhOAc]/Δt = 1.7 mM/h for Pd(OAc) ).
2
We next assessed the kinetic order of the Pd(OAc) -
(Figure 1a), middle (Figure 1b), and end (Figure 1c) of the C−
2
19
catalyzed reaction in benzene and PhI(OAc) . The order in
H acetoxylation reaction is (pyr) Pd(OAc) (4). The identity
2
2 2
benzene was determined by varying the concentration of
benzene from 0.56 to 2.8 M. Hexafluorobenzene was used as a
co-solvent, and the sum of the volumes of C H and C F was
held constant during these experiments in order to minimize
solvent effects on the observed rates. The order in PhI(OAc)2
was determined by varying the concentration of PhI(OAc)2
of 4 was verified via independent synthesis of this complex
20
(Figure 1d). These data implicate 4 as the catalyst resting
state in this system.
6
6
6 6
The lability of the pyridine ligands in complex 4 was assessed
using rotating frame nuclear Overhauser effect NMR spectros-
copy (ROESY). As shown in Figure 2, exchange between free
B
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 3. Plot of initial rate versus [Pd] with the (pyr) Pd(OAc)
2
2
catalyst in Regime 1.
1
Figure 1. C−H acetoxylation of benzene monitored by H NMR
spectroscopy (aromatic region shown) using Pd(OAc) /pyr (1:2) as
2
catalyst. The red spectrum (a) was acquired after heating for 1.5 h, the
green spectrum (b) after heating for 68 h, and the blue spectrum (c)
#
after heating for 97 h. *PhI (consumed oxidant) peaks. PhI(OAc)2
peaks.
Figure 4. Plot of initial rate versus [benzene] with the (pyr) Pd-
2
(
OAc) catalyst in Regime 1.
2
Figure 2. ROESY NMR experiment showing exchange between free
(
8.48 ppm) and bound (8.59 ppm) pyridine. Spectrum acquired in
C D /CD CO D (1:1) at 80 °C.
6
6
3
2
and bound pyridine is observed at 80 °C (20 °C below the
temperature used for catalytic C−H acetoxylation). This
demonstrates the feasibility of pyridine ligand dissociation/
exchange during catalysis.
Using the method of initial rates, we next determined the
order of the reaction in each reagent. Importantly, all of these
experiments were conducted with a Pd:pyr ratio of 1:2 (i.e., no
extra pyidine was added to the reactions). Interestingly, this
reaction showed a half-order dependence on [Pd] (0.53 ± 0.07;
Figure 5. Plot of initial rate versus [PhI(OAc) ] with the
2
(
pyr) Pd(OAc) catalyst in Regime 1.
2
1
2 2
Figure 3), despite the fact that the catalyst resting state
2
2
appears to be a monomer. In addition, a first-order
dependence on [benzene] (1.01 ± 0.07; Figure 4) and a
zero-order dependence on [PhI(OAc) ] (−0.08 ± 0.04; Figure
rate with a series of 3- and 4-substituted pyridine derivatives. As
shown in Figure 6, a Hammett ρ value of +0.64 was obtained,
indicative of increasing reaction rate with more electron-
deficient pyridine ligands.
2
5
±
) were observed. The value of k /k for C H /C D was 3.6
0.3 with this catalyst system (Figure S37). Finally, a
H D 6 6 6 6
Hammett plot was constructed by examining the initial reaction
C
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
K3[4]
pyr]
[
5] = [
(6)
k K [4][benzene]
4
3
rate =
[pyr]
(7)
On the basis of the expression in eq 7, we would initially
anticipate a first-order dependence on [Pd]. However, since all
of the kinetic orders were determined under conditions where
no exogeneous pyridine is added, we can approximate that
[pyr] = [5] and hence reduce eq 5 to eq 8. This will be referred
to as Regime 1 for the rest of the discusson. Using this
approximation, the rate expression reduces to eq 9, which
predicts a half-order dependence on [Pd], first-order depend-
ence on benzene, and zero-order dependence on PhI(OAc)2,
which are all in line with experimental observations.
5]2
K = [
;
[5] = K 21 /2[4]1/2
3
Figure 6. Hammett plot showing the effect of pyridine electronics for
the (pyr) Pd(OAc) catalyst in Regime 1.
[4]
(8)
(9)
2
2
rate = k4K 31 /2[4]1/2[benzene]
Scheme 4. Proposed Mechanism for the (pyr) Pd(OAc) -
2
2
Catalyzed C−H Acetoxylation of Benzene
The approximation that [pyr] = [5] (Regime 1) will break
down under conditions where exogeneous pyridine is added to
the catalytic reaction. When [pyr] ≫ [5] (referred to as
Regime 2 for the remaining discussion), the rate expression in
eq 7 should be in operation. Here, a first-order dependence on
[Pd], first-order dependence on benzene, and inverse first-
order dependence on pyridine are expected.
We next sought to experimentally determine the kinetic
orders in [Pd], benzene, and pyr in the presence of added
pyridine (Regime 2). However, with catalyst 4, the addition of
pyridine resulted in reaction rates that were too slow to
measure accurately and reliably, even at elevated temperatures.
To address this issue, we moved to the analogous 3-
nitropyridine-containing catalyst 4-NO . In Regime 1, this
2
catalyst reacts approximately 3-times faster than its pyr
analogue (Figure 6), and we hypothesized that this would
translate to enhanced reactivity and more reproducible rates in
Regime 2.
Based on the data discussed above, we propose the
mechanism outlined in Scheme 4 for the Pd(OAc) /pyr
2
(
1:2) catalyst system. The NMR data in Figure 1 implicate
The 4-NO -catalyzed conversion of benzene to phenyl
2
monomeric complex 4 as the catalyst resting state. This
complex then undergoes reversible pyridine dissociation to
generate monopyridine complex 5. The ROESY experiment in
Figure 2 supports the feasibility of this step. The positive
Hammett ρ value is also consistent, as more electron-deficient
pyridine ligands should provide a more favorable equilibrium
toward the on-cycle intermediate 5. C−H activation at 5 to
generate 6 is then the RDS. The observation of a first-order
dependence on benzene and a 1° KIE is further consistent with
acetate was first examined under Regime 1 to compare the
catalyst resting state and orders in [Pd], benzene, and
PhI(OAc) to those obtained with catalyst 4. As shown in
2
Figure 7, the bis-pyridine complex (3-NO pyr) Pd(OAc) (4-
2
2
2
NO ) was the only Pd-containing species detected during
2
catalysis, consistent with this complex as the catalyst resting
state. In the absence of added ligand (Regime 1), we observed
an approximately half-order dependence on [Pd] (0.67 ± 0.03;
Figure 8), approximately first-order dependence on benzene
−
this proposal. Finally, 2e oxidation of 6 would form 7 (or an
(
1.4 ± 0.2), and zero-order dependence on PhI(OAc) (−0.06
2
isomer thereof), and C−O bond-forming reductive elimination
would release PhOAc and regenerate 5.
The rate expression for this sequence is shown in eq 4:
±
0.05) (Figures S49 and S51). These results are closely
analogous to those obtained with catalyst 4 and thus implicate
similar mechanisms for 4 and 4-NO in Regime 1.
2
We next evaluated catalyst 4-NO2 in the presence of
exogeneous 3-nitropyridine (28 mM). These conditions
correspond to Regime 2; thus the rate expression in eq 7 is
expected to be operative. Indeed, an approximately first-order
dependence on [Pd] (1.2 ± 0.1; Figure 9), approximately
inverse first-order dependence on 3-nitropyridine (−1.4 ± 0.1;
Figure 10), first-order dependence on benzene (1.0 ± 0.1;
d(PhOAc)
rate =
= k [5][benzene]
4
dt
(4)
Using the equilibrium constant for the dissociation of
pyridine from 4 provides eq 5, which can be rearranged to eq
. Subsitution for [5] then provides eq 7.
6
K = [
5][pyr]
Figure S57), and zero-order dependence on PhI(OAc)
0.08; Figure S59) were observed. These results are consistent
(0.05
3
2
[4]
(5)
±
D
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 9. Plot of initial rate versus [Pd] for the (3-NO pyr) Pd(OAc)
2
2
2
catalyst in Regime 2.
1
Figure 7. C−H acetoxylation of benzene monitored by H NMR
spectroscopy (aromatic region shown) using 4-NO as catalyst. The
2
red spectrum (a) was acquired prior to heating, the green spectrum
(
1
b) after heating for 3.5 h, and the blue spectrum (c) after heating for
#
8.5 h. *PhI (consumed oxidant) peaks. PhI(OAc) peaks.
2
Figure 10. Plot of initial rate versus [excess 3-NO Pyr] for the (3-
2
NO pyr) Pd(OAc) catalyst in Regime 2.
2
2
2
combined equimolar quantities of pyridine and Pd(OAc) in
2
1
C D /CD CO D. H NMR spectroscopic analysis at room
6
6
3
2
temperature showed a mixture of two Pd-pyridine adducts:
complex 4 [(pyr) Pd(OAc) ] along with a second species of
2
2
unknown structure (8) (Figure 11). As shown in Figure 12,
both complexes 4 and 8 are observed throughout the catalytic
reaction. Since the reaction rate under these conditions is
significantly higher than that with 4 alone, we propose that 8 is
likely the predominant active catalyst (or resting state of the
25
active catalyst) in this system.
Figure 8. Plot of initial rate versus [Pd] for the (3-NO pyr) Pd(OAc)
2
catalyst in Regime 1.
2
2
As shown in Figure 11, complex 8 could be a monomer (8a),
dimer (8b), trimer (8c), or a larger oligomeric structure with a
Pd:pyr stoichiometry of 1:1. A diffusion-ordered spectroscopy
(DOSY) NMR experiment was performed to obtain the
approximate molecular weight of 8 and thus to preliminarily
with the mechanism shown in Scheme 4 as well as the rate
expression in eq 7. Notably, these results are analogous to work
2
3
24
26
by Stahl and Hartwig, in which half order in catalyst was
obtained in a regime in which ligand dissociation occurs prior
to the RDS. In both systems upon addition of exogenous
ligand, the catalysts no longer showed half-order dependencies.
differentiate between these possibilities. This experiment
involves generating 8 in situ in the presence of a series of
different molecules that serve as molecular weight standards.
These standards (which include small organic molecules and
large Pd complexes) were selected based on three criteria: (i)
they represent a wide range of molecular weights (MW), (ii)
they are all expected to be unreactive with 8 and with each
Mechanistic Investigation of the Pd(OAc) /Pyridine
1:1) Catalyst System. Finally, we conducted a detailed
2
(
investigation of the most active catalyst system, which is
1
generated by combining a 1:1 ratio of Pd(OAc) to pyridine.
other, and (iii) they have distinct H NMR signals. The DOSY
2
To assess the catalyst resting state under these conditions, we
NMR spectrum of this mixture was then obtained at 25 °C in
E
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 13. DOSY data used to approximate the molecular weight of
[
(pyr)Pd(OAc) ] (blue data point) in solution. Standards used are
2 n
1
benzene, cyclooctane, 18-crown-6, 4,4′-(t-Bu)
2
biphenyl, 1,3,5-
Figure 11. New species (8) observed by H NMR spectroscopy along
(
(
CF ) Ph, (pyr) Pd(OAc) , Si(Me) (C F ) , and (IMes) Pd(OAc)
3
3
2
2
2
6
5
2
2
2
with 4 when combining a 1:1 ratio of Pd(OAc) :pyr.
2
IMes = 1,3-bis(2,4,6-trimethylphenyl)-1,3-dihydro-2H-imidazol-2-
ylidene).
three compounds are expected in a 1:1:2 ratio: [(X)Pd-
(
OAc) ] , [(Y)Pd(OAc) ] , and (X)(Y)Pd (OAc) . Finally, if 8
2 2 2 2 2 4
is a trimer (8c), four compounds are expected in a 1:1:3:3 ratio:
[
(X)Pd(OAc) ] , [(Y)Pd(OAc) ] , (X) (Y)Pd (OAc) , and
2 3 2 3 2 3 6
(
X)(Y) Pd (OAc) (for a pictorial explanation, see Supporting
2
3
6
27
Information, Section 9). We chose 4-methoxypyridine and 4-
tert-butylpyridine for X and Y since they are expected to have
1
similar binding affinities to Pd(OAc) yet their H NMR
2
1
chemical shifts are well resolved. Upon mixing, the H NMR
spectrum shown in Figure 14c was obtained (only the signals
associated with the 2-position of the pyr derivatives are shown).
Both [(4-OMe-pyr)Pd(OAc) ] (see Figure 14a (green) for
2
n
1
independently acquired H NMR spectrum) and [(4-t-Bu-
pyr)Pd(OAc) ] (see Figure 14b (blue) for independently
2
n
1
acquired H NMR spectrum) are observed along with signals
associated with at least one additional compound. Since >2
distinct species are detected, monomer 8a can be eliminated as
a possible structure. Two new signals, labeled as 11, are
observed in a 1:1 ratio. This is inconsistent with the formation
of 8c (mixed trimers). On the basis of this data, we propose
that the resting state of the catalyst is the dimer [(pyr)Pd-
1
Figure 12. C−H acetoxylation of benzene monitored by H NMR
spectroscopy (aromatic region shown) using a Pd(OAc) :pyr ratio of
:1. The red spectrum (a) was acquired at t = 0, the green spectrum
b) at t = 1.5 h, and the blue (c) at the completion of the reaction (t =
2
1
(
3
28
#
.5 h). *PhI (consumed oxidant) peaks. PhI(OAc) peaks.
2
(
OAc) ] (8b).
2 2
C D /CD CO D (1:1). The raw spectrum is provided in
Rate studies were next conducted to assess the order in each
6
6
3
2
Figure S68, and the resulting plot of log(MW) versus log(D)
D = diffusion coefficient) is shown in Figure 13. On the basis
reaction component with Pd(OAc) /pyridine (1:1) as catalyst.
2
(
Under these conditions, the reaction shows a half-order
dependence on Pd (0.55 ± 0.02; Figure 15), a first-order
dependence on benzene (0.94 ± 0.06; Figure 16), and a zero-
of this data, the molecular weight of 8 is calculated to be 755 g/
mol (blue data point in Figure 13). This value is in between
that of the dimer 8b (607 g/mol) and the trimer 8c (911 g/
mol). Importantly, the DOSY experiment requires assumptions
about the size and shape of the standards compared to 8;
therefore, the calculated MW is only an approximation.
Nonetheless, we believe that these data suggest that complex
order dependence on PhI(OAc) (0.00 ± 0.06; Figure 17). In
2
addition, a relatively large primary KIE (k /k = 3.9 ± 0.2)
H
D
with C H /C D is observed (Figure S28). These data are
6
6
6
6
consistent with a mechanism in which a dimeric precatalyst
(8b) enters the catalytic cycle by dissociation to a monomer
(5) (Scheme 5). Rate-limiting C−H activation of benzene at 5
would then form the σ-aryl complex 6, and fast oxidation and
reductive elimination steps would release the product and
regenerate the catalyst. The rate expression for the mechanism
proposed in Scheme 5 can be derived as shown in eqs 10−13. It
is in full accord with the experimental data.
8
is unlikely to be a monomer (8a).
To gain further insights into the structure of 8, we perfomed
an NMR experiment using 1 equiv of Pd(OAc) and 0.5 equiv
2
each of two distinct pyridine derivatives (X and Y). If the 8 is a
monomer (8a), two different complexes should be formed in a
1
:1 ratio: (X)Pd(OAc) and (Y)Pd(OAc) . If 8 is a dimer (8b),
2 2
F
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 16. Plot of initial rate versus [benzene] with Pd(OAc) /pyr
2
(1:1) as catalyst.
Figure 14. Mixed pyridines experiment. (a) Independently measured
1
H NMR spectrum when Pd(OAc) and OMe-pyr (1:1) are mixed.
2
1
(
b) Independently measured H NMR spectrum when Pd(OAc) and
2
1
t-Bu-pyr (1:1) are mixed. (c) H NMR spectrum when Pd(OAc) , t-
2
Bu-pyr, and OMe-pyr (1:0.5:0.5) are mixed. Two new peaks B and C
are observed. (Spectra shown are the region corresponding to the
signals of the 2-position of the monopyridine complexes. For full
spectra, see Supporting Information, section 9).
Figure 17. Plot of initial rate versus [PhI(OAc) ] with Pd(OAc) /pyr
2
2
(1:1) as catalyst.
Scheme 5. Proposed Mechanism with Pd(OAc) /pyr (1:1)
2
as Catalyst
Figure 15. Plot of initial rate versus [Pd] with Pd(OAc) /pyr (1:1) as
2
catalyst.
d(PhOAc)
rate =
= k [benzene][5]
6
dt
(10)
We next probed the impact of substituted pyridine ligands on
the rate of the Pd(OAc) /pyridine (1:1) catalyzed C−H
2
[
8b]
5]
acetoxylation of benzene. As shown in the Hammett plot in
Figure 18, substitution of the pyridine ligand had a negligible
impact on the overall reaction rate (ρ ∼ 0). This result can be
rationalized based on counterbalancing electronic effects of
pyridine substituents on the dissociation of dimer 8b-X to
monomer 5-X (Scheme 6, step 1) and the coordination of
K5 = [
1
/2
(11)
(12)
(13)
1
/2
[
5] = K [8b]
5
1
/2
rate = k K [benzene][8b]
2
6
5
benzene to form the η -benzene intermediate 12-X (Scheme 6,
G
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
breaks up into a monomer. This study provides valuable
insights into the chemistry of Pd(OAc) /pyr-based catalyst
2
systems for C−H bond acetoxylation. Additionally, they have
broader implications in catalysis, since Pd/pyr catalysts are
employed for a variety of transformations including alcohol
1
4a,d
14b
13c
oxidation,
alkene amination, indole arylation, and the
13a,b
Fujiwara−Moritani reaction.
ASSOCIATED CONTENT
Supporting Information
■
*
S
Experimental and spectral details for all new compounds as well
as detailed experimental procedures and data for all kinetic and
NMR experiments. This material is available free of charge via
the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
Corresponding Author
mssanfor@umich.edu
■
*
Figure 18. Hammett plot showing the effect of pyridine electronics
Notes
with Pd(OAc) /pyr (1:1) as catalyst.
2
The authors declare no competing financial interest.
Scheme 6. Explanation for Lack of Pyridine Electronic Effect
with Catalyst 8b
ACKNOWLEDGMENTS
■
This work was supported by NSF under the CCI Center for
Enabling New Technologies through Catalysis (CENTC)
Phase II Renewal, CHE-1205189. A.K.C. thanks the Rackham
Graduate School for a predoctoral fellowship. We thank
Professors Tom Cundari, Karen Goldberg, Bill Jones, Elon
Ison, and John Hartwig for valuable discussions. Dr. Marion H.
Emmert is acknowledged for preliminary studies, particularly an
initial DOSY experiment with [(pyr)Pd(OAc) ] as well as an
2
2
initial investigation of pyridine electronic effects.
REFERENCES
■
(
1) (a) Lyons, T. W.; Sanford, M. S. Chem. Rev. 2010, 110, 1147.
b) Muniz, K. Angew. Chem., Int. Ed. 2009, 48, 9412.
(
(
2) (a) Noisier, A. F. M.; Brimble, M. A. Chem. Rev. 2014, 114, 8775.
(b) Yamaguchi, J.; Yamaguchi, A. D.; Itami, K. Angew. Chem., Int. Ed.
014, 51, 8960. (c) Wencel-Delord, J.; Glorius, F. Nat. Chem. 2013, 5,
369. (d) Chen, D. Y.-K.; Youn, S. W. Chem. Eur, J. 2012, 18, 9452.
step 2). Step 1 is expected to be accelerated by electron-rich
pyridine ligands (e.g., 4-methoxypyridine), while step 2 is
2
9
2
expected to be fastest with electron-deficient pyridines, such as
3
0
3
-nitropyridine. Rate-determining C−H activation would
(
(
e) Gutekunst, W. R.; Baran, P. S. Chem. Soc. Rev. 2011, 40, 1976.
f) McMurray, L.; O’Hara, F.; Gaunt, M. J. Chem. Soc. Rev. 2011, 40,
885. (g) Godula, K.; Sames, D. Science 2006, 312, 67.
then occur at intermediate 12-X. Literature precedent strongly
suggests that this step proceeds via a concerted, acetate-assisted
1
4
metalation-deprotonation transition state. This transition state
(3) For examples of mechanistic studies of ligand-directed Pd-
2
is cyclic, and minimal charge is accumulated. As such, we
anticipate that substitution on the pyridine ligand would have
minimal impact on the energy of this transition state. Overall,
the observed Hammett ρ value of ∼0 is consistent with the
electronic effects on steps 1 and 2 essentially canceling one
another.
catalyzed C(sp )−H functionalization, see: (a) Yang, Y.-F.; Cheng, G.-
J.; Liu, P.; Leow, D.; Sun, T.-Y.; Chen, P.; Zhang, X.; Yu, J.-Q.; Wu, Y.-
D.; Houk, K. N. J. Am. Chem. Soc. 2014, 136, 344. (b) Baxter, R. D.;
Sale, D.; Engle, K. M.; Yu, J.-Q.; Blackmond, D. G. J. Am. Chem. Soc.
́ ́ ́
̃
012, 134, 4600. (c) Iglesias, A.; Alvarez, R.; de Lera, A. R.; Muniz, K.
2
Angew. Chem., Int. Ed. 2012, 51, 2225. (d) Powers, D. C.; Ritter, T.
Acc. Chem. Res. 2012, 45, 840. (e) Ackermann, L. Chem. Rev. 2011,
1
11, 1315. (f) Lyons, T. W.; Hull, K. L.; Sanford, M. S. J. Am. Chem.
CONCLUSIONS
This article describes a detailed exploration of differences
between three catalysts Pd(OAc) , [(pyr)Pd(OAc) ] , and
■
Soc. 2011, 133, 4455. (g) Engle, K. M.; Wang, D.-H.; Yu, J. Q. J. Am.
Chem. Soc. 2010, 132, 14137. (h) Powers, D. C.; Xiao, D. Y.; Geibel,
M. A.; Ritter, T. J. Am. Chem. Soc. 2010, 132, 14530. (i) Deprez, N. R.;
Sanford, M. S. J. Am. Chem. Soc. 2009, 131, 11234. (j) Powers, D. C.;
Ritter, T. Nat. Chem. 2009, 1, 302. (k) Stowers, K. J.; Sanford, M. S.
Org. Lett. 2009, 11, 4584. (l) Desai, L. V.; Stowers, K. J.; Sanford, M. S.
J. Am. Chem. Soc. 2008, 130, 13285. (m) Li, J.-J.; Giri, R.; Yu, J.-Q.
Tetrahedron 2008, 64, 6979. (n) Hull, K. L.; Lanni, E. L.; Sanford, M.
S. J. Am. Chem. Soc. 2006, 128, 14047. (o) Yu, J.-Q.; Giri, R.; Chen, X.
Org. Biomol. Chem. 2006, 4, 4041. (p) Kalyani, D.; Deprez, N. R.;
Desai, L. V.; Sanford, M. S. J. Am. Chem. Soc. 2005, 127, 7330.
2
2 2
(
pyr) Pd(OAc) for the C−H acetoxylation of benzene. With
2
2
all three catalyst systems, the reaction appears to involve rate-
limiting C−H bond cleavage. NMR and kinetic studies suggest
that the most active catalyst (8) rests as a dimer in solution. A
comparison of the mechanisms with bis-pyridine complex 4 and
monopyridine complex 8 implicates C−H activation occurring
at the same intermediate 5. The difference in reactivity between
the bis- and monoligated catalysts, therefore, lies in how they
enter into the catalytic cycle. Bis-pyridine complex 4 dissociates
a pyridine ligand, while the dimeric monopyridine complex 8a
(
4) For reviews of computational studies on the mechanism of CMD-
type C−H activation, see: (a) Balcells, D.; Clot, E.; Eisenstein, O.
Chem. Rev. 2010, 110, 749. (b) Lapointe, D.; Fagnou, K. Chem. Lett.
H
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
2
010, 39, 1118. (c) Boutadla, Y.; Davies, D. L.; Macgregor, S. A.;
Poblador-Bahamonde, A. I. Dalton Trans. 2009, 5820.
5) For a review on methodological studies of C−H activation/
functionalization with substrates that do not bear directing groups, see:
a) Kuhl, N.; Hopkinson, M. N.; Wencel-Delord, J.; Gloruis, F. Angew.
(10) For Pd-catalyzed C−H acetoxylation of arenes using chromates,
nitrogen oxides, and dioxygen as oxidants, see: ref 9a−c. (a) Eberson,
L.; Gomez-Gonzales, L. J. Chem. Soc. D 1971, 263. (b) Stock, L. M.;
Tse, K.; Vorvick, L. J.; Walstrum, S. A. J. Org. Chem. 1981, 46, 1757.
(c) Henry, P. M. J. Org. Chem. 1971, 36, 1886. (d) Tisue, T.; Downs,
W. J. J. Chem. Soc. D 1969, 410a. (e) Davidson, J. M.; Triggs, C. J.
Chem. Soc. A 1968, 1331.
(
(
Chem., Int. Ed. 2012, 51, 10236. For a subsequent report of C−H
amination with substrates that do not bear directing groups, see:
(
b) Shrestha, R.; Mukherjee, P.; Tan, Y.; Litman, Z. C.; Hartwig, J. F. J.
(11) For related work on Pd-catalyzed C−H hydroxylation of arenes,
see: (a) Guo, H.; Chen, Z.; Mei, F.; Zhu, D.; Xiong, H.; Yin, G.
Chem.Asian J. 2013, 8, 888. (b) Liang, P.; Xiong, H.; Guo, H.; Yin,
G. Catal. Commun. 2010, 11, 560. (c) Lee, J. H.; Yoo, K. S.; Park, C.
P.; Olsen, J. M.; Sakaguchi, S.; Prakash, G. K. S.; Mathew, T.; Jung, K.
W. Adv. Synth. Catal. 2009, 351, 563. (d) Liu, Y.; Murata, K.; Inaba, M.
J. Mol. Catal. A 2006, 256, 247. (e) Shibahara, F.; Kinoshita, S.;
Nozaki, K. Org. Lett. 2004, 6, 2437. (f) Passoni, L. C.; Cruz, A. T.;
Buffon, R.; Schuchardt, U. J. Mol. Catal. A 1997, 120, 117. (g) Jintoku,
T.; Takaki, K.; Fujiwara, Y.; Fuchita, Y.; Hiraki, K. Bull. Chem. Soc. Jpn.
Am. Chem. Soc. 2013, 135, 8480. For subsequent reports on C−H
arylation with substrates that do not bear directing groups, see:
(
c) Durak, L. J.; Lewis, J. C. Organometallics 2014, 33, 620. (d) Storr,
T. E.; Namata, F.; Greaney, M. F. Chem. Commun. 2014, 50, 13275.
e) Wang, S.; Liu, W.; Cen, J.; Liao, J.; Huang, J.; Zhan, H. Tetrahedron
Lett. 2014, 55, 1589. (f) Yan, T.; Zhao, L.; He, M.; Soule, J.-F.;
(
́
Bruneau, C.; Doucet, H. Adv. Synth. Catal. 2014, 356, 1586. (g) Zhao,
L.; Yan, T.; Bruneau, C.; Doucet, H. Catal. Sci. Technol. 2014, 4, 352.
(
h) Cambeiro, X. C.; Boorman, T. C.; Lu, P.; Larrosa, I. Angew. Chem.,
1
1
990, 63, 438. (h) Jintoku, T.; Taniguchi, H.; Fujiwara, Y. Chem. Lett.
987, 1865.
Int. Ed. 2013, 52, 1781. (i) Wagner, A. M.; Sanford, M. S. J. Am. Chem.
Soc. 2013, 135, 15110. (j) Ren, X.; Wen, P.; Shi, X.; Wang, Y.; Li, J.;
Yang, S.; Yan, H.; Huang, G. Org. Lett. 2013, 15, 5194. (k) Storr, T. E.;
Greaney, M. F. Org. Lett. 2013, 15, 1410. (l) Wencel-Delord, J.;
Nimphius, C.; Wang, H.; Glorius, F. Angew. Chem., Int. Ed. 2012, 51,
(
12) (a) Yoneyama, T.; Crabtree, R. H. J. Mol. Catal. A 1996, 108, 35.
For subsequent work on Pd-catalyzed C−H acetoxylation using
PhI(OAc) , see: (b) Tato, F.; García-Domínguez, A.; Cardenas, D. J.
́
2
Organometallics 2013, 32, 7487. (c) Wang, N.; McCormick, T. M.; Ko,
S.-B.; Wang, S. Eur. J. Inorg. Chem. 2012, 4463.
1
3001. (m) Zhou, L.; Lu, W. Organometallics 2012, 31, 2124. For
subsequent reports on C−H olefination with substrates that do not
bear directing groups, see: (n) Gigant, N.; Backvall, J.-E. Org. Lett.
kvall, J.-E. Org. Lett. 2014, 16,
(13) (a) Kubota, A.; Emmert, M. H.; Sanford, M. S. Org. Lett. 2012,
̈
1
1
4, 1760. (b) Zhang, Y.-H.; Shi, B.-F.; Yu, J.-Q. J. Am. Chem. Soc. 2009,
31, 5072. For an additional report in which a monopyridine/Pd
2
1
6
3
014, 16, 4432. (o) Gigant, N.; Bac
̈
664. (p) Jin, W.; Wong, W.-T.; Law, G.-L. Chem. Catal. Chem. 2014,
, 1599. (q) Pham, M. V.; Cramer, N. Angew. Chem., Int. Ed. 2014, 53,
484. (r) Vora, H. U.; Silvestri, A. P.; Engelin, C. J.; Yu, J.-Q. Angew.
complex is implied as an intermediate, see: (c) Stuart, D. R.; Fagnou,
K. Science 2007, 316, 1172.
(14) (a) Steinhoff, B. A.; Stahl, S. S. Org. Lett. 2002, 4, 4179. For
Chem., Int. Ed. 2014, 53, 2683. (s) Ying, C.-H.; Yan, S.-B.; Duan, W.-L.
Org. Lett. 2014, 16, 500. (t) Zhou, L.; Lu, W. Chem.Eur. J. 2014, 20,
additional reports in which monopyridine/Pd complexes are implied
as intermediates, see: (b) Ye, X.; Liu, G.; Popp, B. V.; Stahl, S. S. J. Org.
Chem. 2011, 76, 1031. (c) Popp, B. V.; Stahl, S. S. Chem.Eur. J.
6
34. (u) Min, M.; Choe, H.; Hong, S. Asian J. Org. Chem. 2012, 1, 47.
(
v) Wencel-Delord, J.; Nimphius, C.; Patureau, F. W.; Glorius, F.
2
009, 15, 2915. (d) Schultz, M. J.; Adler, R. S.; Zierkiewicz, W.;
Chem.Asian J. 2012, 7, 1208. For a subsequent report on C−H
Privalov, T.; Sigman, M. S. J. Am. Chem. Soc. 2005, 127, 8499.
15) Stephenson, T. A.; Morehouse, S. M.; Powell, A. R.; Heffer, J.
P.; Wilkinson, G. J. Chem. Soc. 1965, 3632.
16) Simmons, E. M.; Hartwig, J. F. Angew. Chem., Int. Ed. 2001, 51,
066.
carboxylation with substrates that do not bear directing groups, see:
(
(
2
w) Suga, T.; Mizuno, H.; Takaya, J.; Iwasawa, N. Chem. Commun.
014, 50, 14360.
6) For mechanistic studies on C−H activation/functionalization
(
3
(
with substrates that do not bear directing groups, see: (a) Cheng, C.;
Hartwig, J. F. J. Am. Chem. Soc. 2014, 136, 12064. (b) Wang, D.; Izawa,
Y.; Stahl, S. S. J. Am. Chem. Soc. 2014, 136, 9914. (c) Sanhueza, I. A.;
Wagner, A. M.; Sanford, M. S.; Schoenebeck, F. Chem. Sci. 2013, 4,
(17) Gary, J. B.; Sanford, M. S. Organometallics 2011, 30, 6143.
(18) Previous studies from our group suggested the possibility of
interactions between the pyridine and hypervalent iodine oxidant [ref
i]. To test this, varying mol fractions of 4-F-PhI(OAc)2 and 3-
3
2
767. (d) Wagner, A. M.; Hickman, A. J.; Sanford, M. S. J. Am. Chem.
1
fluoropyridine were combined in C D /CD CO D. However, H and
6
6
3
2
Soc. 2013, 135, 15710. (e) Lyons, T. W.; Hull, K. L.; Sanford, M. S. J.
Am. Chem. Soc. 2011, 133, 4455. (f) Tan, Y.; Hartwig, J. F. J. Am.
Chem. Soc. 2011, 133, 3308. (g) Zhang, S.; Shi, L.; Ding, Y. J. Am.
Chem. Soc. 2011, 133, 20218. (h) Hull, K. L.; Sanford, M. S. J. Am.
Chem. Soc. 2009, 131, 9651. (i) Biswas, B.; Sugimoto, M.; Sakaki, S.
Organometallics 2000, 19, 3895.
7) For reviews summarizing various approaches to non-directed C−
H functionalization, see: (a) ref 5a. (a) Neufeldt, S. R.; Sanford, M. S.
Acc. Chem. Res. 2012, 45, 936. (b) Bruckl, T.; Baxter, R. D.; Ishihara,
19
F NMR spectroscopic analysis showed no changes in the chemical
shifts of either species at room temperature or at 80 °C, indicating that
there is no reaction between pyridine and the oxidant. See Supporting
Information, section 7.
(19) Notably, complex 8 is not observed during catalysis when
Pd(OAc) :Pyr (1:2) is used.
2
(
(
20) Trend, R. M.; Ramtohul, Y. K.; Stoltz, B. M. J. Am. Chem. Soc.
005, 127, 17778.
21) Half order is observed when 4 is generated in situ (as in Figure
) and when using the isolated complex 4 (see Supporting
Information, section 6.5.4.).
22) There is a possibility that 8 is the active catalyst under these
2
(
3
̈
Y.; Baran, P. S. Acc. Chem. Res. 2012, 45, 826. (b) Newhouse, T.;
Baran, P. S. Angew. Chem., Int. Ed. 2011, 50, 3362. (c) Shul’pin, G. Org.
Biomol. Chem. 2010, 8, 4217. (c) Ref 2g.. (d) Dick, A. R.; Sanford, M.
S. Tetrahedron 2006, 62, 2439. For a perspective on the role of ligands
in C−H activation, see: (e) Engle, K. M.; Yu, J.-Q. J. Org. Chem. 2013,
(
conditions; however, because no signals corresponding to 8 are
observed in the spectra of Figure 1a−c, this is an unlikely scenario.
The possibility of 8 forming as an intermediate along the path from 4
to the RDS can not be not excluded.
(23) (a) Steinhoff, B. A.; Guzei, I. A.; Stahl, S. S. J. Am. Chem. Soc.
2004, 126, 11268. (b) King, A. E.; Ryland, B. L.; Brunold, T. C.; Stahl,
S. S. Organometallics 2012, 31, 7948.
7
8, 8927.
8) (a) Cook, A. K.; Emmert, M. H.; Sanford, M. S. Org. Lett. 2013,
5, 5428. (b) Gary, J. B.; Cook, A. K.; Sanford, M. S. ACS Catal. 2013,
, 700. (c) Emmert, M. H.; Cook, A. K.; Xie, Y. J.; Sanford, M. S.
(
1
3
Angew. Chem., Int. Ed. 2011, 50, 9409. (d) Emmert, M. H.; Gary, J. B.;
Villalobos, J. M.; Sanford, M. S. Angew. Chem., Int. Ed. 2010, 49, 5884.
(24) Boller, T. M.; Murphy, J. M.; Hapke, M.; Ishiyama, T.; Miyaura,
N.; Hartwig, J. F. J. Am. Chem. Soc. 2005, 127, 14263.
(
9) For Pd-catalyzed C−H acetoxylation of arenes using K S O as
2
2
8
oxidant, see: (a) Eberson, L.; Jonsson, L. Liebigs Ann. Chem. 1977, 233.
(25) By mass balance, Pd(OAc) must also be formed under these
2
(b) Eberson, L.; Jonsson, L. Acta Chem. Scand. B 1976, 30, 361.
(c) Eberson, L.; Jonsson, L. Acta Chem. Scand. B 1974, 28, 771.
(d) Eberson, L.; Jonsson, L. J. Chem. Soc., Chem. Commun. 1974, 885.
conditions. However, this is likely also not the active catalyst because
the rates of reactions with 8 are much higher than that with Pd(OAc)2
alone.
I
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(
26) (a) Macchioni, A.; Ciancaleoni, G.; Zuccaccia, C.; Zuccaccia, D.
Supramolecular Chemistry: From Molecules to Nanomaterials; Wiley:
Hoboken, NJ, 2012; p 1. (b) Li, D.; Keresztes, I.; Hopson, R.; Willard,
P. G. Acc. Chem. Res. 2009, 42, 270.
(
27) For a review on the method of continuous variations, a related
technique used to quantify aggregation, see: Renny, J. S.; Tomasevich,
L. L.; Tallmadge, E. H.; Collum, D. B. Angew. Chem., Int. Ed. 2013, 52,
11998.
(
28) The integral ratios also support formation of a dimer: two new
signals labelled as 11 are observed in a 1:1 ratio; since the integral ratio
of [(4-t-Bu-pyr)Pd(OAc) ] and [(4-OMe-pyr)Pd(OAc) ] is twice
2
2
2 2
that of the mixed compound (4-t-Bu-pyr)(4-OMe-pyr)Pd (OAc) , the
2
4
new complexes are formed in a 1:1:2 ratio (9:10:11), and this is
consistent with the formation of 8b (dimer).
(
29) For a study on the trans influence as a function of pyridine
substituents, see: Woods, C.; Daffron, C. Inorg. Chim. Acta 1985, 101,
3.
30) Appleton, T. G.; Clark, H. C.; Manzer, L. E. Coord. Chem. Rev.
973, 10, 335.
1
(
1
J
DOI: 10.1021/jacs.5b00238
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX