Mechanistic Insights into Fe Catalyzed α-C?H Oxidations of Tertiary Amines

Article abstract of DOI:10.1002/cctc.202001382

We report detailed mechanistic investigations of an iron-based catalyst system, which allows the α-C?H oxidation of a wide variety of amines. In contrast to other catalysts that effect α-C?H oxidations of tertiary amines, the system under investigation exclusively employs peroxy esters as oxidants. More common oxidants (e. g. ^tBuOOH) previously reported to affect amine oxidations via free radical pathways do not provide amine α-C?H oxidation products in combination with the described catalyst system. The investigations described herein employ initial rate kinetics, kinetic profiling, DFT calculations as well as Eyring, kinetic isotope effect, Hammett, ligand coordination, and EPR studies to shed light on the Fe catalyst system. The obtained data suggest that the catalytic mechanism proceeds through C?H abstraction at a coordinated substrate molecule. This rate-determining step occurs either through an Fe(IV) oxo pathway or a 2-electron pathway at an Fe(II) intermediate with bound oxidant. DFT calculations indicate that the Fe(IV) oxo mechanism will be the preferred route of these two possibilities. We further show via kinetic profiling and EPR studies that catalyst activation follows a radical pathway, which is initiated by hydrolysis of PhCO₃^tBu to ^tBuOOH. Overall, the obtained mechanistic data support a non-classical, Fe catalyzed pathway that requires substrate binding, inducing selectivity for α-C?H functionalization.

Full text of DOI:10.1002/cctc.202001382

The European Society Journal for Catalysis
Supported by
Accepted Article
Title: Mechanistic Insights into Fe Catalyzed α-C-H Oxidations of
Tertiary Amines
Authors: Christopher J. Legacy, Taylor O. Hope, Yohann Gagné,
Frederick T. Greenaway, Mathieu Frenette, and Marion H.
Emmert
This manuscript has been accepted after peer review and appears as an
Accepted Article online prior to editing, proofing, and formal publication
of the final Version of Record (VoR). This work is currently citable by
using the Digital Object Identifier (DOI) given below. The VoR will be
published online in Early View as soon as possible and may be different
to this Accepted Article as a result of editing. Readers should obtain
the VoR from the journal website shown below when it is published
to ensure accuracy of information. The authors are responsible for the
content of this Accepted Article.
To be cited as: ChemCatChem 10.1002/cctc.202001382
Link to VoR: https://doi.org/10.1002/cctc.202001382
0
1/2020

ChemCatChem
10.1002/cctc.202001382
Mechanistic Insights into Fe Catalyzed α-C-H Oxidations of Tertiary
Amines
a
b
b
Dr. Christopher J. Legacy, Ms. Taylor O. Hope, Mr. Yohann Gagné, Prof. Dr. Frederick T.
Greenaway, Prof. Dr. Mathieu Frenette, * Prof. Dr. Marion H. Emmert *
c
b
a,d
aDepartment of Chemistry and Biochemistry, Worcester Polytechnic Institute, 100 Institute Road, Worcester, MA 01609,
USA
bDépartement de Chimie, Université du Québec à Montréal, Case Postale 8888, Succursale Centre-Ville, Montréal, Québec
H3C 3P8, Canada
cGustaf H. Carlson School of Chemistry, Clark University, 950 Main Street, Worcester, MA 01610, USA
dCurrent Address: Department of Process Research & Development, Discovery Process Chemistry, MRL, Merck & Co., Inc.,
7
70 Sumneytown Pike, West Point, PA 19486, USA
ABSTRACT: We report detailed mechanistic investigations of an iron-based catalyst system, which allows the α-C-H oxidation of
a wide variety of amines. In contrast to other catalysts that effect α-C-H oxidations of tertiary amines, the system under investigation
t
employs exclusively peroxy esters as oxidants. More common oxidants (e.g. BuOOH) previously reported to affect amine oxidations
via free radical pathways do not provide amine α-C-H oxidation products in combination with the herein described catalyst system.
The investigations described herein employ initial rate kinetics, kinetic profiling, DFT calculations as well as Eyring, kinetic isotope
effect, Hammett, ligand coordination, and EPR studies to shed light on the Fe catalyst system. The obtained data suggest that the
catalytic mechanism proceeds through C-H abstraction at a coordinated substrate molecule. This rate-determining step occurs either
through an Fe(IV) oxo pathway or a 2-electron pathway at an Fe(II) intermediate with bound oxidant. DFT calculations indicate the
Fe(IV) oxo mechanism will be the preferred route of these two possibilities. We further show via kinetic profiling and EPR studies
t
t
that catalyst activation follows a radical pathway, which is initiated by hydrolysis of PhCO Bu to BuOOH. Overall, the obtained
3
mechanistic data support a non-classical, Fe catalyzed pathway that requires substrate binding, inducing selectivity for α-C-H
functionalization.
t
Introduction
employing BuOOH as oxidant likely promote oxidations via
free radical pathways and hemiaminal intermediates, regardless
C-H oxidations of organic molecules frequently occur in
[19]
[
1]
[2]
of the used metal catalyst used. This study was performed
with aniline-type substrates and the catalysts which were
investigated included both precious metal systems (Rh
RuCl ) and base-metal catalysts (CuBr, FeCl , Co(OAc)
Importantly, the presence of O influenced the kinetic isotope
effects observed, suggesting that reaction pathways may differ
depending on the oxidant employed. Unfortunately, the
question of whether the mechanistic insights reported by Doyle
and coworkers are transferrable to α-C-H functionalizations of
aliphatic amines has not yet been resolved, as many catalysts
active in aniline α-C-H oxidations are unreactive towards
biosynthetic pathways as well as in drug metabolism. Many
of these reactions are catalyzed by enzymes bearing Fe
cofactors. Examples are the cytochrome P450 family exhibiting
2
(cap)
4
,
[2-4]
3
3
2
).
Fe-porphyrin cofactors
or non-heme dioxygenases,
[5]
2
hydroxylases, and halogenases. The majority of these systems
catalyze oxidations via Fe(IV) oxo intermediates, which attack
aliphatic C-H bonds of a substrate in a radical rebound fashion,
cleaving the C-H bond via
a
hydrogen abstraction
[2,5-9]
mechanism.
For enzyme-catalyzed amine α-C-H
oxidations an alternative mechanistic possibility to Fe oxo
pathways has been proposed, based on model compound
[20]
[10]
tertiary aliphatic amines.
studies. This mechanism proceeds through single electron
and proton transfers, forming iminium or imine products via
radical intermediates.
Motivated by the absence of such mechanistic information
regarding the α-C-H oxidations of aliphatic amines, we set out
to establish mechanistic hypotheses supported by experimental
data for an Fe-based catalyst system previously developed in
our lab (see Scheme 1). This reaction effects oxidations of
aromatic and aliphatic amines and is selective for α-C-H
functionalizations of amines, even in complex molecule settings
such as active pharmaceutical ingredients.
many common oxidants previously used in Fe oxidation
catalysis do not affect this reaction, with the exception of
The mechanistic framework of synthetic systems that
functionally mimic the metabolic pathways of amine α-C-H
oxidations in combination with O
t
2
, H
2
O
2
, or BuOOH as
oxidants has long been under discussion, with proposed
pathways proceeding via free or solvent-caged radicals,
hemiaminals, peroxyhemiaminals, and/or metal-coordinated
[20]
Surprisingly,
[11-18]
iminium ions.
Recently, a seminal publication by Doyle
and coworkers has shown that the majority of the catalysts
1
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
peroxyesters that afford between 21 and 71% yield. Yields are
strongly dependent on the steric bulk of the amine substrate. In
contrast to other protocols proceeding via free radical
pathways, the established reaction conditions are selective
for oxidation of C-H bonds of acyclic substrates, while cyclic
α-C-H bonds are not attacked.
(RR, Scheme 2, top), providing direct access to hemiaminal 2.
Alternatively, a proton-coupled electron transfer (PCET) or
hydride transfer (HT; Scheme 2, middle) or a sequence of
electron and proton transfers (ET/PT, bottom) may form the
[21]
iminium intermediate 3, which in turn is attacked by H
provide hemiaminal 2.
2
O to
Scheme 1. Optimized Conditions for Fe Catalyzed α-C-H
oxidation of Tertiary Aliphatic Amines.
Scheme 2. General Reaction Framework. RR = radical
rebound; PCET = proton-coupled electron transfer; HT =
hydride transfer; ET = electron transfer; PT = proton transfer.
Based on the studies detailed in this manuscript, we propose
that α-C-H cleavage proceeds through a radical rebound or
concerted mechanism instead of a free radical mechanism often
proposed with simple Fe catalysts such as in Fenton-type
chemistries. Interestingly, the experimental values for Eyring
parameters and kinetic isotope effect (KIE) are more akin to
values typically observed with ß-hydride elimination (formal
hydride shift) and concerted metalation deprotonation
[22]
In the presence of large amounts of H O in the reaction
2
solution, dealkylation of amine substrate to afford 4 is likely
driven by formation of the aldehyde hydrate; alternatively, a
second α-C-H oxidation of hemiaminal 2 would result in
formation of the amide product. These preliminary mechanistic
investigations did not provide any insight into the first, most
crucial part of amine α-C-H oxygenation: the process of α-C-H
bond breaking and the role of the catalyst in this process. As
such, our subsequent studies aimed to elucidate this question.
(involvement of a heteroatom of the oxidant). Such two-
electron processes are most often invoked for C-H cleavage
[23,24]
processes in precious metal catalysts (e.g. Pd, Ir
). This
mechanistic proposal provides insight into the need for
peroxyester oxidants, the role of each reaction component under
the empirically established optimal catalytic conditions, and the
sources of side products observed in the reaction mixture.
Overall, the mechanistic pathway provides a rationale for the
selectivity of amine α-C-H oxidation with the established Fe
catalyst system.
B. Initial Rate Kinetics. We first sought to establish an
experimental kinetic rate law, using the method of initial rates.
We reasoned that this approach would provide a suitable
framework for further mechanistic studies, eventually leading
to rational optimization of the amine α-C-H oxidation protocol
and new catalyst developments. Due to the paramagnetic
properties of the Fe catalyst used in amine α-C-H oxidation, the
use of NMR techniques to follow the kinetic profile of the
reactions was deemed unsuitable. Instead, in-situ FTIR was
Results and Discussion
A. General Mechanistic Framework for Fe Catalyzed
[20]
Amine to Amide Conversion. Our laboratory has previously
reported several key mechanistic features of the Fe catalyzed
α-C-H oxidation of amines (see Scheme 2). (1) The
employed and the disappearing C=O band of the oxidant
concentration of H
conversion, with loadings of less than 5 equiv. and more than
5 equiv. significantly reducing amide yields. (2) In reactions
with large amounts of H O added, the major side product is
dealkylated amine 4; lowering the oxidant (PhCO
2
O had been found to be crucial for high
t
(PhCO
3
Bu) was observed to provide well-characterized kinetic
profiles.
1
We first explored the role of the added ligand 1 for catalysis.
2
t
t
To this end, the oxidant PhCO
solvent and an in-situ FTIR signal was established. This signal
represented by the green line in Scheme 3) showed initially a
curve without any slope, signifying the absence of reactivity.
Next, substrate (NPr ), H O, and FeCl ·6H O were sequentially
added into the mixture. With each addition, the IR signal for
3
Bu was added to pyridine as
3
Bu) loading
leads to formation of up to 64% of 4. Furthermore, aldehyde
(
side products can be observed. (3) Reactions performed in the
1
8
presence of H
2
O as additive revealed that the oxygen atom
O (and not O
Bu). (4) Experiments performed in the presence of
KCN led to α-C-H cyanation instead of α-C-H oxygenation,
suggesting that iminium compounds are potential
intermediates or side products in the reaction pathway. (5)
Finally, reacting HNPr and EtCHO under analogous reaction
3
2
3
2
incorporated in the amide product stems from H
2
2
t
or PhCO
3
t
PhCO Bu decreased due to dilution, but remained flat,
3
indicative of a constant concentration of the oxidant.
3
Finally, 2-picolinic acid (1) was added to the mixture,
2
resulting in kinetic traces with increasing/decreasing slopes
t
conditions to amine α-C-H oxidation conditions also resulted in
formation of the corresponding amide. Collectively, these
findings supported the proposal of a mechanistic framework for
amine α-C-H oxidation as formulated in Scheme 2 in analogy
to aniline oxidation pathways affected by Fe-heme
(green: PhCO
3
Bu; blue: reaction product PhCO H; Scheme 3).
2
Overall, these data suggest that ligand 1 is necessary for
catalytic turnover and that α-C-H oxidation does not occur in
its absence. Importantly, the form of the kinetic traces
characterizes a catalytic reaction with an initiation period.
Therefore, all initial rate studies discussed below measured the
maximum reaction rate (maximum slope) of the kinetic curve.
[10]
compounds.
In the proposed mechanistic framework, the initial C-H bond
cleavage may proceed through a radical rebound mechanism
2
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
Initiation of the catalytic reaction by addition of 1 as detailed
conditions. Scheme 5 shows a plot of the resulting initial rates
above further suggests that the empirically optimized ratio of
plotted versus the pK of the corresponding acid of the X-type
a
[20]
Fe/1 (1:1) should also correspond to the highest observed
reaction rate, if only one molecule of 1 is bound in Fe in the
catalytically active species. This hypothesis was tested by
measuring the initial rate with varying amounts of ligand 1 (2.5
to 10 mol %; Scheme 4). As expected, the highest initial rate
was observed at a Fe/1 ratio of 1:1, while higher and lower
ratios resulted in decreased initial rates. This demonstrates that
the conditions providing the highest yields are also the
kinetically fastest conditions, suggesting that the active catalyst
does not change throughout the course of the reaction.
ligands. Notably, the initial rates do not change significantly,
implying that X-type ligands are not coordinated to the Fe
center in the catalytically active species.
We then turned our attention to establishing the kinetic orders
of the different components of the catalytic system; the
respective data are shown in Schemes 6 to 8. At low
(catalytically relevant) concentrations of Fe/1 (<5 mol %), a
nearly linear relationship was observed, suggesting that the
reaction is first order in [Fe/1] under these conditions. Towards
higher catalyst loadings, the measured initial rates go through a
maximum. This implies that off-cycle Fe species exist that are
favored under these conditions. Such off-cycle species may be
polymeric or dimeric [Fe/1]
n
compounds, similar to compounds
that have previously been characterized in analogous Fe/1
[18]
t
mixtures.
Kinetic orders in [PhCO
3
Bu] and [NPr ] were
3
determined in an analogous fashion. In both cases, the observed
rates showed a first-order dependence, as characterized by a
linear increase of rate with substrate or oxidant concentrations
(
see Schemes 7 and 8). At higher (not catalytically relevant)
t
PhCO
3
Bu and NPr loadings, saturation kinetics are observed
3
(see SI).
Scheme 3. In-situ FTIR Reaction Profiles of Oxidant
t
(PhCO
3
Bu)
and
Reaction
Product
(PhCO H)
2
Demonstrating Importance of Ligand 1 for Catalysis.
Conditions: 50 °C, pyridine (50 eq.), NPr (1 eq.), H O (9 eq.),
FeCl ·6H O (5.0 mol %), 1 (5.0 mol %), PhCO
tBu (3 eq.).
3
2
3
2
3
Scheme 5. Dependence of Initial Rate on X-Type Ligand.
Scheme 4. Dependence of Initial Rate on Ligand (1) Loading
at Fe Loading of 5.0 mol %.
Next, we explored the influence of X-type ligands. We
reasoned that the presence of these ligands can be expected to
influence catalytic activity, if they remain coordinated to the Fe
center. To examine this question, we measured the initial rates
of the reaction with a series of Fe(III) [Fe(OTf)
3
, FeCl
3
,
Scheme 6. Initial Rate Versus Catalyst Loading.
Fe(NO
3
)
3
, Fe
2
(SO
4
)
3
, Fe(acac)
3
] under otherwise unchanged
3
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
We then investigated if C-H bond cleavage may be occurring
in the rds (Scheme 9) by independently measuring the initial
Next, the kinetic orders in [H
determined; both components of the reaction system have
empirically been shown to be crucial for reactivity (Schemes 10
2
O] and [pyridine] were
rates of oxidation with NEt
3
and NEt
3
-D15 as substrates. The
[
20]
obtained ratio k /k was determined to be 1.7 ± 0.1, indicative
H
D
and 11).
increasing [H
either a negative 2 order dependence (R = 0.99; Scheme 10A)
Interestingly, saturation kinetics are observed at
of a primary kinetic isotope effect and consistent with rate-
determining C-H cleavage. Interestingly, similar values of KIEs
have been observed for rate-determining ß-hydride elimination
2
O]. At low [H O], our data analysis suggests
2
nd
2
st
2
or a negative 1 order dependence on [H O] (R = 0.90; Scheme
10B). The R values for 1 order and 2 order kinetics are both
in the acceptable range (0.99 and 0.90) when extrapolating a
2
[
25]
2
st
nd
by Ru catalysts.
effects have been observed in reactions between Pd complexes
and benzene/C , which have been proposed to bind benzene
Furthermore, similar ranges of isotope
st
6
D
6
line from the first four data points. This suggests that either a 1
[26]
nd
before C-H cleavage. Comparing these values is consistent
with a mechanism in which the amine substrate binds to the Fe
catalyst before C-H scission.
order or a 2 order dependence on [H
2
O] occurs at low [H O].
2
Distinguishing between the two regimes is complicated by the
th
saturation kinetics approaching 0 order dependence on [H O]
2
at higher loadings; thus both scenarios need to be considered in
formulating a mechanistic hypothesis (see Scheme 28 and
th
associated discussion). Furthermore, a 0 order dependence on
[pyridine] was observed, seemingly contradictory to the
experimental observation that the absence of pyridine results in
very low (EtOAc solvent) or no reactivity (all other tested
solvents).
Therefore, we hypothesized that the number of coordinated
pyridine ligands on Fe does not change between the resting state
and the transition state structure, which is in agreement with a
th
0
order dependence on [pyridine]. To test this hypothesis, we
performed the reaction in different substituted pyridines as
solvents (50 eq. each), reasoning that coordination of pyridine-
type ligands to Fe should result in a significant influence of the
ligand electronics on the catalytic rate. Such a dependence was
indeed observed (Scheme 12). Interestingly, both electron-
withdrawing and electron-donating substituents on pyridine
result in slower initial reaction rates than pyridine, indicating a
change in rds for pyridines with electron-withdrawing and
electron-donating substituents.
Scheme 7. Initial Rate Versus Oxidant Loading.
Scheme 8. Initial Rate Versus Amine Substrate Loading.
Scheme 9. Kinetic Isotope Effect Study: Initial Rate for
NEt and NEt -D15. Conditions: 50 °C, pyridine (50 eq.), NEt
or NEt -D15 (1 eq.), H O (9 eq.), FeCl ·6H O (5.0 mol %), 1 (5.0
tBu (3 eq.).
3
3
3
3
2
3
2
mol %), PhCO
4
3
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
Scheme 10. A. Inverse Negative Initial Rate Versus H O
2
Loading. B. Initial Rate Versus H
2
O Loading.
Finally, we aimed to document the kinetic orders for PhCO
2
H
-
and PhCO
2
. PhCO H is a product of the reaction, which may
2
react with the substrate in an acid-base reaction, and thus
influence reaction rates. Indeed, plotting the initial rates
obtained with different loadings of PhCO
2
H showed a negative
st
1
order dependence of the rate on [PhCO H] (Scheme 13).
2
-
However, all attempts to introduce PhCO
resulted either in precipitation of alkali benzoates or a rapid,
non-catalyzed background reaction (NBu PhCO ) with the
2
to the reaction
4
2
Scheme 13. Initial Rate Versus PhCO H Loading
2
oxidant. Both conditions prevent the acquisition of valid kinetic
data.
C. Further Insight into the Rate-Determining Step:
Eyring Study and Oxidant Model Substrate. To further
investigate the nature of the rds, the initial rate of the reaction
was measured as a function of temperature (Scheme 14) and
Eyring analysis was performed to determine the activation
‡
‡
-1
-1
parameters (ΔH = 14.1 kcal/mol; ΔS = -10.8 cal K mol ; for
details, see the SI). The obtained value for the activation
‡
entropy ΔS is in agreement with a highly ordered transition
state structure and is similar to values observed in concerted C-
[27,28]
H bond cleavage processes.
Furthermore, both activation
parameters are in a similar range as activation parameters
measured for rate-determining H atom transfer from substrates
‡
with weak C-H bonds to an Fe-oxo model compound (ΔH =
‡
-1
-1
1
2.7 kcal/mol; ΔS = -9 cal K mol ), albeit in combination
with a significantly larger kinetic isotope effect (KIE = 5.7 for
[29]
xanthene/9,10-d
2
-xanthene).
Generally, many metal-oxo
initiated C-H bond abstractions similar to cytochrome P450
Scheme 11. Initial Rate Versus Pyridine Loading.
(
radical rebound) mechanisms show similar negative values for
Conditions: Conditions: 50 °C, EtOAc (50 eq.), NPr
O (9 eq.), FeCl ·6H O (5.0 mol %), 1 (5.0 mol %), PhCO
3 eq.), pyridine (2.5 to 10 eq.).
3
(1 eq.),
[27,30]
t
the activation entropy.
cleavage of peroxyesters
In contrast, homolytic O-O bond
H
2
3
2
3
Bu
[
31]
or reactions proceeding via
(
dissociation of substrate from the metal catalyst prior to C-H
[32]
bond cleavage are typically associated with more positive
activation entropies, corresponding to a less ordered transition
state structure.
To probe the nature of oxidant participation in amine α-C-H
oxidation, a modified peroxyester 6 (Scheme 15) was employed
as oxidant. 6 and related compounds are known to allow
differentiation between pathways that undergo homolytic and
[
33-36]
heterolytic O-O bond cleavage.
Scheme 12. Hammett Studies with Different Pyridines as
Solvent.
5
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
Scheme 14. Eyring Study and Resulting Thermodynamic
Activation Parameters.
Scheme 15. Oxidation of NPr
Substrate. Conditions: FeCl ∙6H O (5.0 mol %), 1 (5.0 mol %),
pyridine (2 mL), H O (9.0 eq.), 6 (3.0 eq.), NPr (1.0 eq.); yields
3
with Oxidant Model
Scheme 16. (A) In-situ FTIR Analysis Result of Titration of
FeCl ∙6H O into Aqueous Pyridine Solution. (B)
Corresponding in-situ FTIR Spectra.
3
2
3
2
2
3
obtained after 75% conversion of 6. No amide product was
obtained, but products of dealkylation were observed in the
reaction mixture.
We next investigated binding of picolinic acid (1) to Fe.
When aliquots corresponding to 0.5 eq. of 1 were sequentially
As outlined in Scheme 15, homolytic cleavage of the O-O
bond would result in formation of acetone and a benzylic
radical; interestingly, acetone was found only as a minor
product (7%). The major product, alcohol 7, was obtained in
added to a solution of 1.0 eq. FeCl
3
∙6H O in pyridine (Scheme
2
17A), in situ FTIR analysis showed binding with a maximum
stoichiometry of 2:1 (1:Fe); addition of higher amounts of 1 led
to the observation of a signal for non-coordinated ligand 1
43% yield, suggesting that heterolytic O-O bond cleavage is the
(yellow trace in Scheme 17). Notably, two distinct spectra for
major reaction pathway under the established reaction
conditions. This is particularly interesting, as heterolytic
cleavage of the O-O bond suggests a two electron or metal-oxo
pathway, which stands in contrast to free radical pathways often
proposed and documented for simple Fe catalysts in the
Fe complexes of 1 were observed prior to detection of non-
coordinated 1. Both spectra are characterized by the shift of the
-1
-1
C=O signal from 1730 cm (free ligand 1) to 1690 cm . The
first spectrum (orange trace in Scheme 17) occurs exclusively
at low concentrations of 1 and constitutes the major species in
solution until 1.5 eq. of 1 and disappears in the presence of 2.0
eq. of 1. This implies that the first spectrum (orange trace)
[17, 22b, 37-38]
presence of other peroxy oxidants.
D. Coordination Studies. To provide more insight into the
resting state structure and allow future design of more reactive
and selective catalyst structures, we initiated coordination
studies geared towards establishing the ligand environment
around the Fe catalyst precursor.
corresponds to a Fe complex [Fe(1)L
of 1 coordinated to the metal center. The second spectrum
(purple trace in Scheme 17) characterizes a complex [Fe(1)
with two molecules of 1 as ligands. These data establish that the
majority of Fe is present as [Fe(1)L ] at the catalytically
relevant 1:1 ratio of FeCl ∙6H O and 1.
Finally, oxidant, substrate, and product binding to [Fe(1)L
were investigated. Interestingly, no significant binding to
n
] with only one molecule
2
n
L ]
n
3
2
First, we sought to confirm pyridine binding (as suggested by
the Hammett studies) to the Fe center. This was achieved by
n
]
titrating aliquots of a solution of FeCl
3
∙6H O into an aqueous
2
t
solution containing pyridine. The FTIR signal representing free,
non-coordinated pyridine proportionally diminished with the
addition of the first two aliquots, and then disappeared with the
addition of a third aliquot (Scheme 16). This result, supported
by data obtained via ESI-MS analysis (see SI), indicates a
pyridine : Fe binding stoichiometry of 3:1. Interestingly, the
FTIR bands representing coordinated pyridine (light green
spectrum labeled 1.0-1.5 eq. Fe; Scheme 16B) are very broad in
comparison to the non-coordinated pyridine signals, possibly
showcasing the fluxional nature of ligand coordination.
[Fe(1)L
propenamide (see SI for details). In contrast, NPr
[Fe(1)L ] was clearly observed, as shown in Scheme 18. Upon
addition of up to 1 eq. NPr to a solution of [Fe(1)L ] in
pyridine, a new species was detected, characterized by a C=O
n
] was observed for PhCO
3
Bu and N,N-dipropyl
3
binding to
n
3
n
-1
band at 1680 cm and several peaks in the fingerprint region of
-1
the FTIR spectrum (1320 to 950 cm , green trace in Scheme
18). This complex converts to a second species (red trace) upon
addition of more NPr
detected until addition of 3.0 eq. of NPr
3
; however, the first complex can still be
is completed. The
3
second complex (red trace) exhibits FTIR bands in the
fingerprint region that are similar to the first species but shows
-1
characteristic new bands at 1660 and 1350 cm (Scheme 18B).
Further addition of NPr beyond 3.0 eq. does not result in
detection of another species with distinct FTIR signals.
3
6
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
Scheme 17. (A) Titration of 1 into FeCl
3
∙6H
2
O in pyridine.
Scheme 18. (A) Titration of NPr
in pyridine. (B) Comparison of FTIR Spectra of Formed Fe
Complexes, non-coordinated NPr , and non-coordinated 1.
3
into solution of [Fe(1)L ]
n
(B) Comparison of observed FTIR spectra representing
different [Fe(1) ] species and free picolinic acid (1). Color
n
L
m
3
coding relates the disappearance and appearance of
different spectra throughout the titration experiment.
E. Investigation of Catalyst Initiation via Reaction
Progress Kinetic Analysis. One feature of the kinetic profiles
observed in all kinetic studies discussed above is a notable
initiation period. Literature precedent suggests generally two
mechanistic causes for an initiation period: (i) Promotion of
catalysis by reaction products or (ii) an initial, non-catalyzed
reaction, which assembles the catalytically active species.
Due to the stoichiometry required to form each species, we
propose that the first signal (green trace) corresponds to a
[39]
complex [Fe(1)(NPr
characterizes [Fe(1)(NPr
ligands. We therefore propose that up to 2 eq. of NPr
3
)L
n
], while the second (red) trace
] with two amine substrates as
can
3
)
L
2 n
3
coordinate to Fe under catalytic conditions, which are
characterized by a 20-fold excess of substrate to Fe catalyst.
Interestingly, the data shown in Scheme 18A imply different
binding constants for the first and second equivalent of amine:
To test if the reaction at hand is promoted by its products, a
kinetic trace obtained under standard conditions (black trace in
Scheme 19) was compared to kinetic traces observed upon
The maximum amount of [Fe(1)(NPr
eq. of NPr added, while addition of 2 more equivalents of
NPr is required to completely convert [Fe(1)(NPr )L ] to
Fe(1)(NPr ].
3
)L
n
] is observed with only
addition of
Pr NC(=O)Et, HNPr
products such as HNPr
in all tested reactions ). No significant effect on either the
maximum rate or the length of the initiation period was
observed, as shown by the near overlap of the corresponding
kinetic traces in Scheme 19. This suggests that reaction
products do not promote catalysis.
5
mol
%
of different reaction products
, BuOH, PhCO H; amine dealkylation
are typically observed in trace amounts
1
3
t
(
2
2
2
3
3
n
2
[
3
)
2
L
n
[20]
Having ruled out the possibility of self-promotion by reaction
products, we then sought to gain insight into potential off-cycle
reactivity that may lead to the assembly of the catalytically
active species. We hypothesized that different concentrations of
reagents that take part in the non-catalyzed initiation reaction
should result in differing initiation period lengths. Therefore,
we postulated that comparing reaction profiles obtained at
different concentrations of reagents would identify reagents
involved in catalyst activation.
To this end, kinetic reaction profiles obtained with different
reagent and catalyst concentrations were plotted as reaction rate
vs. time and conversion vs. time plots (Schemes 20 to 23). This
analysis shows that initiation periods are shortened at higher
[
oxidant] and [substrate] and at lower [H
2
O] and [PhCO H].
2
7
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
These observations suggest that these four reagents are involved
in the initiation (or catalyst activation) reaction.
In contrast, no clear trend was observed with different
loadings of Fe/1 (see SI), implying that the Fe catalyst
concentration has no significant influence on the rate of the
initiation reaction. This result agrees with the previously stated
postulate that the initiation reaction is a non-catalyzed reaction
aiding in assembly of the catalytically active species.
Scheme 21. Effect of Amine Substrate Loading on Length of
Initiation Period.
Scheme 19. Effect of Product Addition on Length of
Initiation Period.
Scheme 22. Effect of H
Period.
2
O Loading on Length of Initiation
Scheme 20. Effect of Oxidant Loading on Length of
Initiation Period.
Scheme 23. Effect of PhCO
Initiation Period.
2
H Loading on Length of
Based on these data, we hypothesized that hydrolysis of the
t
-
t
peroxyester PhCO
3
Bu to form PhCO and BuOOH is the non-
2
8
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
t
catalyzed reaction occurring during the initiation period
small amounts of BuO• were detected upon addition of
t
t
t
t
(
Scheme 24). BuOOH can go on to form BuO· radicals in the
PhCO
3
Bu. This suggests that BuOO• may be the product of the
catalyst activation pathway. Furthermore, the data imply that
small amounts of PhCO Bu can undergo homolytic O-O bond
[40-
presence of pyridine, as previously described in the literature,
4
2]
t
thereby forming at least small amounts of bipyridines. This
3
is consistent with the observation of small amounts of
bipyridines by GCMS analysis in all amine α-C-H oxidation
reactions (see SI).
scission in the presence of Fe(III), even though the majority of
turnover in C-H oxidation stems from heterolytic O-O bond
cleavage, as evidenced by the oxidant model study described
further above.
To further test the hypothesis of catalyst activation by
t
reaction between BuOOH formed via hydrolysis and the Fe(III)
t
catalyst precursor, we added 5 mol % of BuOOH to the reaction
mixture before adding amine substrate. Gratifyingly, the kinetic
trace obtained completely lacked an initiation period and the
reaction initiated with a maximum rate (see Scheme 25). This
t
suggests that formation of BuOOH via hydrolysis of the
t
oxidant PhCO
3
Bu is indeed responsible for the initiation
t
reaction. A proposed pathway how BuOOH may lead to
catalyst initiation is discussed vide infra.
Scheme 26. Summary of Results from EPR Studies.
Scheme 24. Proposed Reaction Leading to Catalyst
t
EPR studies to elucidate the spin state of the Fe center were
performed of solutions frozen in liquid nitrogen immediately
after preparation. EPR clearly shows the binding of ligand 1
Activation via Formation of BuOOH and Side Reactivity of
t
BuOOH.
upon addition of 1 to a FeCl
3
solution in pyridine/H
2
O, with a
new high-spin signal (S = 5/2). Addition of NPr
3
to this solution
resulted in a low-spin Fe species (S = 1/2) with a very broad
signal at g = 2. Both of these findings are in agreement with
ligand binding to Fe(III) as proposed by FTIR titration (see
t
above). Interestingly, when PhCO
3
Bu is added to the resulting
mixture, no EPR signal is observed, suggesting the formation
of an EPR silent species as the resting state of the reaction.
G. Discussion of Mechanistic Proposals Based on
Experimental Data. Based on the experimental mechanistic
data obtained, we propose (i) a mechanistic pathway for catalyst
activation and (ii) two catalytic cycles for catalyst turnover.
First, catalyst activation is clearly achieved by the interaction
t
of the Fe(III) catalyst precursor 8 and BuOOH. As evidenced
by EPR studies, the formation of tBuOO• likely occurs in this
reaction, suggesting a reduction from Fe(III) to Fe(II) in a
[22]
reaction step precedented in Fenton chemistry.
In
t
combination with the ligand coordination studies, this leads to
a proposed resting state structure 9 as shown in Scheme 27.
Scheme 25. Effect of BuOOH Addition on Length of
Initiation Period.
Based on this resting state structure, two different catalytic
cycles (Scheme 28A/B) can be proposed that fulfill the
empirical rate law (Scheme 28C), the requirement for
F. EPR Studies. To gain insight into the presence of radical
species and possible oxidation states of Fe in the reaction
mixture, EPR studies were performed (for complete details, see
SI). EPR studies at room temperature were designed to
elucidate the proposed presence of radicals in the reaction
mixture. In these studies, no radicals were observed in the
1
8
heterolytic O-O bond cleavage, O labeling being introduced
1
8
[20]
to the product from O-labeled water, and C-H bond scission
in the rds. Both catalytic cycles (A and B) start with resting state
9
. Due to the catalytic relevance of ligand 1, we propose that it
remains coordinated to Fe throughout the complete catalytic
cycle. Coordination of pyridine, two H O ligands, and an
2
absence of added FeCl
presence of the spin trap phenyl-tert-butylnitrone.
In a mixture of FeCl , 1, pyridine, and water (mimicking the
catalytic reaction mixture), addition of either BuOOH or
3
, either via direct observation or in the
undefined ligand L complete the resting state structures. Due to
the significantly lower reaction rates at high [Fe/1], we propose
that each resting state is in equilibrium with an off-cycle,
3
t
oligomeric or polymeric species [Fe(1)L
x
]
n
. The key difference
t
PhCO Bu resulted in formation of O-centered radicals, but in
3
between mechanisms A and B lies in (i) the Fe species affecting
different amounts and selectivity (Scheme 26A). With
t
t
BuOOH, the major radical detected was BuOO•, while only
9
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
the C-H scission and (ii) oxidation state changes throughout the
catalytic cycle.
Mechanism
A
is characterized by
a
series of
coordination/dissociation steps between resting state 9 and rds,
with the rds proceeding through a hybrid mechanism between
β-hydride elimination and concerted metalation/deprotonation,
which also cleaves the peroxide O-O bond. The resulting
carboxylate binds the Fe center while the iminium product
leaves the cycles. H O addition to the iminium to form the
2
hemiaminal product 2 and ligand exchanges at Fe by a total of
three water molecules completes the mechanism. During the
complete cycle A, Fe remains in the oxidation state +2.
Mechanism
B
also proceeds through
a
series of
coordination/dissociation steps between the resting state 9 and
the rds, in addition to a two electron-oxidation of Fe(+2) to
Fe(+4). This oxidation proceeds upon heterolytic O-O bond
t
cleavage of PhCO
3
Bu, in agreement with the model oxidant
studies discussed above. The rds in mechanism B is formulated
analogous to typical radical-rebound mechanisms at Fe-oxo
species, consisting of a homolytic C-H bond cleavage at Fe(+4)
intermediate 13, followed by a fast rebound step to directly
produce the bound hemiaminal intermediate.
Scheme 27. Catalyst Activation and Proposed Resting State
Structure 9.
Under the reaction conditions, the hemiaminal product 2
generated by both catalytic cycles will oxidize to yield the final
amide product. This final transformation is assumed to be rapid
and will not contribute to the initial reaction kinetics.
H. Evaluation of Mechanisms A and B using DFT
Calculations. Both mechanisms A and B appear to fit the
experimental observations. Thus, DFT calculations were
employed to gather support for the preferred catalytic pathway
for this reaction. For both reaction mechanisms, we set out to
calculate the thermodynamic barriers for the proposed rate
limiting step. The level of theory used for the results discussed
in the main text is B3LYP/6-311+g(2d,2p)//CPCM(pyridine)
that was shown to accurately model energy barriers for similar
[43]
Fe(IV)-oxo mechanisms. See SI for details and for results
using other levels of theory.
For mechanism A, we attempted to calculate the kinetic
barrier for the rate-determining step (rds) 11 → 12 as shown in
Scheme 28A. For both high spin (S = 2) and low spin (S = 0)
electronic configurations, the energy barrier to undergo a β-
hydride elimination was studied by forcing an approach of the
leaving hydride towards the peroxide oxygen of compound 11.
In both cases, the energy barrier exceeded 35 kcal/mol (see SI).
The β-hydride elimination could also occur via a reaction
with the other peroxide oxygen, i.e., forming a 7-membered ring
instead of the 8-membered transition state shown in Scheme
28A. The barrier for an approach of the β-hydride towards this
second peroxide oxygen gave a similarly large barrier (>35
kcal/mol). These data imply that mechanism A is unlikely to
occur as the barrier for the rds is prohibitively large.
DFT calculations investigating mechanism B were more
fruitful. In these calculations, the trialkylamine reagent was
simplified to dimethyl-ethylamine and the unspecified ligand L
was chosen to be NH . Intramolecular -C-H-abstraction from
3
the Fe(IV) oxo species 13 afforded activation energy
parameters that closely matched those obtained experimentally:

∆
G was determined to be 19.4 kcal/mol by DFT calculations
for the high spin quintet form (S = 2), whereas the experimental
results show a value of 17.3 kcal/mol (Scheme 29). At this level
of theory, the intermediate spin triplet form (S = 1) and the low
spin singlet form (S = 0) for compound 13 were higher in energy
than the quintet form by 3.1 and 32.5 kcal/mol, respectively.
Scheme 28. Proposed Catalytic Cycles Based on
Experimental Mechanistic Studies.
10
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
Intramolecular -C-H abstraction by the Fe(IV) oxo
generates intermediate 14 and the subsequent hydroxyl transfer
occurred with no kinetic barrier to form 15. The intramolecular
H-abstraction/hydroxyl transfer step from 13 → 15 is highly
favorable with ∆G = -50.3 kcal/mol. Several ligand exchange
steps are then required to form the resting state 9; this
favorable, the exact mechanism for this transformation remains
under investigation.
Several general conclusions can be drawn from the two
proposed mechanisms: (1) Both pyridine and ligand 1 do not
dissociate from the catalytically active species throughout the
reaction cycle. This suggests a path forward in the journey
towards more efficient catalysts (and away from pyridine in
solvent-quantities): The design of a ligand in which 1 is
covalently associated to a pyridine moiety. (2) Kinetic profiling
0
compound also favors a high spin quintet electronic structure (S
=
+
2). The oxidation of the Fe center from an oxidation state of
t
2 to +4 (9 to 13) using PhCO Bu as the oxidant is
3
0
thermodynamically favorable (∆G = -17.5 kcal/mol); however,
t
and EPR studies allowed the proposal of BuOO· as the reagent
the exact mechanism for this step is not yet known. The overall
providing access to the catalytically active species 9. This
suggests that one-electron reduction is required to access the
catalytically active species, which in turn might simply be
achieved by the use of Fe(+2) as catalyst precursor species. (3)
Mechanism B, which is consistent with both experimental data
and reaction energies as determined by DFT, does not suggest
a role for free radicals in the reaction mechanism outside of the
catalyst activation pathway. This fundamentally distinguishes
the investigated catalyst system from the systems investigated
0
∆
G for the catalytic cycle in mechanism B is -60.5 kcal/mol.
[19]
previously by Doyle and coworkers. (4) The presented work
provides an explanation for why the selective α-C-H oxidation
of a wide variety of tertiary aliphatic amines is possible with
this catalyst system: coordination to Fe activates the substrates
to undergo α-C-H cleavage. In contrast, previous systems
proceeding through free radical mechanisms are restricted to
more activated substrates (secondary amines, benzylic/aniline-
type substrates). Thus, the obtained mechanistic data support a
non-classical amine α-C-H oxidation pathway proceeding via
substrate binding, which concurrently induces selectivity and
enables reactivity.
ASSOCIATED CONTENT
Supporting Information. Detailed procedures for kinetic and
coordination studies, examples of kinetic traces, ESI-MS, GC-MS
analyses and DFT studies. This material is available free of charge
via the Internet at http://pubs.acs.org.
Scheme 29. DFT-calculated free energies of the proposed
Fe(IV) oxo mechanism shown in Scheme 28B using
B3LYP/6-311+g(2d,2p)//CPCM(pyridine).
High-spin
quintet (S = 2) is the favored electronic configuration
throughout the mechanism.
AUTHOR INFORMATION
Corresponding Authors
While this proposed mechanism involves an intramolecular
α-C-H abstraction between an Fe(IV) oxo species and the alkyl
amine, intermolecular reactivity was also considered (see SI).
*
mhemmert@wpi.edu; marion.emmert@merck.com;
emmertlab
frenette.mathieu@uqam.ca; @mat_frenette
@
*

Briefly, the transition state entropy ∆S for intermolecular
reactivity, as calculated by DFT, was -33.7 cal/(mol•K). This

deviated significantly from the experimentally-determined ∆S
of -10.8 cal/(mol•K). For comparison, the DFT-calculated ∆S

Author Contributions
for the intramolecular α-C-H abstraction shown in Scheme 28B
was -5.1 cal/(mol•K) and thus, in better agreement with the
experimentally measured value.
The manuscript was written through contributions of all authors.
All authors have given approval to the final version of the
manuscript.
ACKNOWLEDGMENT
Summary and Conclusions
We thankfully acknowledge R. Grimm (WPI) for assistance with
ESI-MS measurements, D. Hebrault and M. Burns (both Mettler-
Toledo) for assistance with acquiring in-situ FTIR analyses. We
acknowledge Worcester Polytechnic Institute for financial support
of this work. M.F., T.O.H and Y.G. acknowledge Compute Canada
and Gaussian for computational resources as well as NanoQAM,
NSERC and FRQNT for funding.
Overall, the results of the mechanistic studies can be
graphically summarized in detailed proposed mechanisms as
shown in Scheme 28. DFT-calculations conclusively reject the
β-hydride elimination mechanism A; the β-hydride elimination
gave prohibitively large energy barriers (>35 kcal/mol) for the
rds shown in Scheme 28A. Concurrently, DFT calculations lend
strong evidence for the Fe(IV) oxo mechanism B.
Experimentally-determined activation energies and those
calculated for the Fe(IV) oxo in the rds of mechanism B were
REFERENCES
[1] K. Chen, P. S. Baran, Nature 2009, 459, 824-828.
in good agreement. While the formation of the Fe(IV) oxo
t
[2] J. Genovino, D. Sames, L. G. Hamann, B. B. Touré, Angew.
Chem. Int. Ed. 2016, 55, 14218-14239
species by reaction with PhCO
3
Bu is thermodynamically
11
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
.
[
[24] Y. Boutadla, D. L. Davies, S. A. Macgregor, A. I. Poblador-
Bahamonde, Dalton Trans. 2009, 5820-5831.
3] M. Oscarson, Drug Metab. Dispos. 2001, 29, 91-95.
[
4] A. A. Walsh, G. D. Szklarz, E. E. Scott, J. Biol. Chem. 2013,
[25] L. K. Thalen, D. Zhao, J. B. Sortais, J. Paetzold, C. Hoben, J.
E. Backvall, Chem. Eur. J. 2009, 15, 3403-3410.
[26] H. A. Zhong, J. A. Labinger, J. E. Bercaw, J. Am. Chem. Soc.
2002, 124, 1378-1399.
2
88, 12932-12943.
[5] W. Nam, Acc. Chem. Res. 2007, 40, 522-531.
[6] I. Schlichting, J. Berendzen, K. Chu, A. M. Stock, S. A. Maves,
D. E. Benson, R. M. Sweet, D. Ringe, G. A. Petsko, S. G. Sligar,
Science 2000, 287, 1615-1622.
[27] K. Gardner, J. Mayer, Science 1995, 269, 1849-1851.
[28] G. D. Roiban, E. Serrano, T. Soler, G. Aullon, I. Grosu, C.
Cativiela, M. Martinez, E. P. Urriolabeitia, Inorg. Chem. 2011, 50,
8132-8143.
[29] K. Cho, P. Leeladee, A. J. McGown, S. DeBeer, D. P. Goldberg,
J. Am. Chem. Soc. 2012, 134, 7392-7399.
[
[
[
[
7] K. Chen, L. Que, J. Am. Chem. Soc. 2001, 123, 6327-6337.
8] J. T. Groves, Nature Chem. 2014, 6, 89-91.
9] X. Huang, J. T. Groves, J. Biol. Inorg. Chem. 2017, 22, 185-207.
10] B. Chiavarino, R. Cipollini, M. E. Crestoni, S. Fornarini, F.
Lanucara, A. Lapi, J. Am. Chem. Soc. 2008, 130, 3208-3217.
11] Z. Li, D. S. Bohle, C.-J. Li, Proc. Natl. Acad. Sci. 2006, 103,
928-8933.
12] A. J. Catino, J. M. Nichols, B. J. Nettles, M. P. Doyle, J. Am.
Chem. Soc. 2006, 128, 5648-5649.
13] E. Boess, C. Schmitz, M. Klussmann, J. Am. Chem. Soc. 2012,
34, 5317-5325.
14] J. Boivin, D. Gaudin, D. Labrecque, K. Jankowski,
Tetrahedron Lett. 1990, 31, 2281-2282.
15] D. H. R. Barton, J. Boivin, D. Gaudin, K. Jankowski,
Tetrahedron Lett. 1989, 30, 1381-1382.
[30] G. K. Cook, J. M. Mayer, J. Am. Chem. Soc. 1995, 117, 7139-
7156.
[31] S. S. Kim, I. S. Baek, A. Tuchkin, K. M. Go, J. Org. Chem.
2001, 66 , 4006-4011.
[32] W. D. Jones, F. J. Feher, Acc. Chem. Res. 2002, 22, 91-100.
[33] E. Hedaya, S. Winstein, J. Am. Chem. Soc. 1967, 89, 1661-
1672.
[34] T. L. Foster, J. P. Caradonna, J. Am. Chem. Soc. 2003, 125,
3678-3679.
[
8
[
[
1
[
[
[35] G. T. Rowe, E. V. Rybak-Akimova, J. P. Caradonna, Chem.
Eur. J. 2008, 14, 8303-8311.
[
[
[
16] D. H. R. Barton, D. Doller, Acc. Chem. Res. 1992, 25, 504-512.
17] M. J. Perkins, Chem. Soc. Rev. 1996, 25 (4), 229-236.
18] S. Kiani, A. Tapper, R. J. Staples, P. Stavropoulos, J. Am.
[36] O. V. Makhlynets, E. V. Rybak-Akimova, Chem. Eur. J. 2010,
16, 13995-14006.
[37] D. T. Sawyer, A. Sobkowiak, T. Matsushita, Acc. Chem. Res.
1996, 29, 409-416.
[38] MacFaul, P. A.; Wayner, D. D. M.; Ingold, K. U., A Radical
Account of “Oxygenated Fenton Chemistry” 1. Acc. Chem. Res.1998,
31, 159-162.
[39] D. G. Blackmond, J. Am. Chem. Soc.2015, 137, 10852-10866.
[40] N. V. Zolotova, E. T. Denisov, Bull. Acad. Sci. USSR, Div.
Chem. Sci. 1966, 15, 736-738.
[41] S. Steenken, P. O'Neill, J. Phys. Chem. 1978, 82 (3), 372-374.
[42] C. Maillard-Dupuy, C. Guillard, H. Courbon, P. Pichat,
Environ. Sci. Technol. 1994, 28, 2176-2183.
[43] A. Decker, J. U. Rohde, E. J. Klinker, S. D. Wong, L. Que, E.
I. Solomon, J. Am Chem. Soc. 2007, 129, 5983-5996.
Chem. Soc. 2000, 122, 7503-7517.
19] M. O. Ratnikov, M. P. Doyle, J. Am. Chem. Soc. 2013, 135,
549-1557.
20] C. J. Legacy, A. Wang, B. J. O'Day, M. H. Emmert, Angew.
Chem. Int. Ed. 2015, 54, 14907-14910.
21] O. Yilmaz, M. S. Oderinde, M. H. Emmert, J. Org. Chem. 2018,
3, 11089-11100.
22] a) D. T. Sawyer, A. Sobkowiak, T. Matsushita, Acc. Chem Res.
996, 29, 409-416; b) P. A. MacFaul, D. D. M. Wayner, K. U. Ingold,
[
1
[
[
8
[
1
Acc. Chem. Res. 1998, 31, 159-162; c) S. Goldstein, D. Meyerstein,
Acc. Chem. Res. 1999, 32, 547-550.
[23] D. L. Davies, S. M. Donald, S. A. Macgregor, J. Am. Chem.
Soc. 2005, 127, 13754-13755.
Table of Contents
Amine α-C-H oxidations provide access to amides and secondary amines in a biomimetic fashion. Detailed experimental and
computational studies reveal amine binding as key feature for α-C-H functionalization in the studied non-heme Fe catalyst
system. These findings explain why the catalyst successfully employs any tertiary amine as substrate, instead of being restricted
to functionalization of activated (benzylic, aniline-type) amines. While catalyst formation proceeds through a radical process,
free radical mechanisms are not involved during catalyst turnover.
12
This article is protected by copyright. All rights reserved.

ChemCatChem
10.1002/cctc.202001382
1
3
This article is protected by copyright. All rights reserved.