Nonclassical Single-State Reactivity of an Oxo-Iron(IV) Complex Confined to Triplet Pathways

Article abstract of DOI:10.1021/jacs.7b03255

C-H bond activation mediated by oxo-iron (IV) species represents the key step of many heme and nonheme O₂-activating enzymes. Of crucial interest is the effect of spin state of the Fe^IV(O) unit. Here we report the C-H activation kinetics and corresponding theoretical investigations of an exclusive tetracarbene ligated oxo-iron(IV) complex, [L^NHCFe^IV(O)(MeCN)]²⁺ (1). Kinetic traces using substrates with bond dissociation energies (BDEs) up to 80 kcal mol^-1 show pseudo-first-order behavior and large but temperature-dependent kinetic isotope effects (KIE 32 at ?40 °C). When compared with a topologically related oxo-iron(IV) complex bearing an equatorial N-donor ligand, [L^TMCFe^IV(O) (MeCN)]²⁺ (A), the tetracarbene complex 1 is significantly more reactive with second order rate constants k′₂ that are 2-3 orders of magnitude higher. UV-vis experiments in tandem with cryospray mass spectrometry evidence that the reaction occurs via formation of a hydroxo-iron(III) complex (4) after the initial H atom transfer (HAT). An extensive computational study using a wave function based multireference approach, viz. complete active space self-consistent field (CASSCF) followed by N-electron valence perturbation theory up to second order (NEVPT2), provided insight into the HAT trajectories of 1 and A. Calculated free energy barriers for 1 reasonably agree with experimental values. Because the strongly donating equatorial tetracarbene pushes the Fe-d_x²-y² orbital above d_z², 1 features a dramatically large quintet-triplet gap of ～18 kcal/mol compared to ～2-3 kcal/mol computed for A. Consequently, the HAT process performed by 1 occurs on the triplet surface only, in contrast to complex A reported to feature two-state-reactivity with contributions from both triplet and quintet states. Despite this, the reactive Fe^IV(O) units in 1 and A undergo the same electronic-structure changes during HAT. Thus, the unique complex 1 represents a pure "triplet-only" ferryl model.

Full text of DOI:10.1021/jacs.7b03255

Article
Non-Classical Single-State Reactivity of an Oxo-
Iron(IV) Complex Confined to Triplet Pathways
Claudia Kupper, Bhaskar Mondal, Joan Serrano-Plana, Iris Klawitter,
Frank Neese, Miquel Costas, Shengfa Ye, and Franc Meyer
J. Am. Chem. Soc., Just Accepted Manuscript • Publication Date (Web): 30 May 2017
Downloaded from http://pubs.acs.org on May 30, 2017
Just Accepted
“Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted
online prior to technical editing, formatting for publication and author proofing. The American Chemical
Society provides “Just Accepted” as a free service to the research community to expedite the
dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts
appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been
fully peer reviewed, but should not be considered the official version of record. They are accessible to all
readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered
to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published
in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just
Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor
changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers
and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors
or consequences arising from the use of information contained in these “Just Accepted” manuscripts.
Journal of the American Chemical Society is published by the American Chemical
Society. 1155 Sixteenth Street N.W., Washington, DC 20036
Published by American Chemical Society. Copyright © American Chemical Society.
However, no copyright claim is made to original U.S. Government works, or works
produced by employees of any Commonwealth realm Crown government in the course
of their duties.

Page 1 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
NonꢀClassical SingleꢀState Reactivity of an OxoꢀIron(IV) Complex
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Confined to Triplet Pathways
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
†ǁ
‡ǁ
§
†
‡
Claudia Kupper, Bhaskar Mondal, Joan SerranoꢀPlana, Iris Klawitter, Frank Neese, Miquel Cosꢀ
§
‡*
†*
tas Shengfa Ye, and Franc Meyer
,
†
Universität Göttingen, Institut für Anorganische Chemie, Tammannstrasse 4, 37077 Göttingen, Germany.
MaxꢀPlanck Institut für Chemische Energiekonversion, Stiftstrasse 34ꢀ36, 45470 Mülheim an der Ruhr, Germany.
‡
§
Institut de Química Computacional i Catàlisi (IQCC), Departament de Quimica, Universitat de Girona, Campus Montilivi,
E17071 Girona, Catalonia, Spain.
ǁ
These authors contributed equally to this work.
ABSTRACT: C–H bond activation mediated by oxoꢀiron(IV) species represents the key step of many heme and nonꢀheme O ꢀ
activating enzymes. Of crucial interest is the effect of spin state of the Fe (O) unit. Here we report the C–H activation kinetics and
corresponding theoretical investigations of an exclusive tetracarbene ligated oxoꢀiron(IV) complex, [ LFe (O)(MeCN)] (1).
2
IV
NHC
IV
2+
ꢀ
1
Kinetic traces using substrates with bond dissociation energies (BDEs) up to 80 kcal mol show pseudoꢀfirst order behavior and
large but temperatureꢀdependent kinetic isotope effects (KIE 32 at ꢀ40°C). When compared with a topologically related oxoiron(IV)
TMC
IV
2+
complex bearing an equatorial Nꢀdonor ligand, [ LFe (O)(MeCN)] (A), the tetracarbene complex 1 is significantly more reacꢀ
tive with second order rate constants k’ that are 2ꢀ3 orders of magnitude higher. UVꢀvis experiments in tandem with cryospray
2
mass spectrometry evidence that the reaction occurs via formation of a hydroxoꢀiron(III) complex (4) after the initial Hꢀatom transꢀ
fer (HAT). An extensive computational study using a wave function based multireference approach, viz. complete active space selfꢀ
consistent field (CASSCF) followed by Nꢀelectron valence perturbation theory up to second order (NEVPT2), provided insight into
the HAT trajectories of 1 and A. Calculated free energy barriers for 1 reasonably agree with experimental values. Because the
strongly donating equatorial tetracarbene pushes the Feꢀdx²ꢀy² orbital above d , 1 features a dramatically large quintetꢀtriplet gap of
z²
~
18 kcal/mol compared to ~2–3 kcal/mol computed for A. Consequently, the HAT process performed by 1 occurs on the triplet
surface only, in contrast to complex A reported to feature twoꢀstateꢀreactivity with contributions from both triplet and quintet states.
IV
Despite this, the reactive Fe (O) units in 1 and A undergo the same electronicꢀstructure changes during HAT. Thus, the unique
complex 1 represents a pure ‘tripletꢀonly’ ferryl model.
experimental work on dinuclear iron complexes provided
support for this theoretical prediction, till date all S = 2 monꢀ
onuclear models available suffer from low stability and/or
their reactivity is affected by steric hindrance. The most thorꢀ
oughly studied high spin Fe (O) model is so sterically hinꢀ
dered by its TMG tren ligand system (TMG tren = 1,1,1ꢀ
tris{2ꢀ[N ꢀ(1,1,3,3ꢀtetramethylguanidino)]ethyl}amine) that it
shows reactivity similar to S = 1 models. Furthermore, the
so far most r2e+active high spin complex
[(TQA)Fe (O)(NCMe)] (TQA tris(quinolinꢀ2ꢀ
ylmethyl)amine) is only slightly more reactive than its most
INTRODUCTION
5
Highꢀvalent oxoꢀiron(IV) species occur in nature as key inꢀ
termediates in catalytic cycles of numerous heme and nonꢀ
heme iron enzymes, such as Cytochrome P450, TauD, CytC3
IV
1
and PheH. Since these species mediate challenging substrate
3
3
2
transformations including the oxygenation of unactivated C–H
bonds, they are of tremendous interest. Inspired by the biologꢀ
ical archetypes, synthetic oxoꢀiron(IV) complexes are increasꢀ
ingly employed in CꢀH bond activation chemistry. Over the
years, the reactivity of numerous model complexes featuring
multidentate chelating Nꢀdonor coordination has been investiꢀ
6,3d
IV
=
IV
2+
effective low spin counterpart [(Me NTB)Fe (O)] (Me NTB
3
3
2
gated in great detail. Those studies have shown that the reacꢀ
=
tris((1ꢀmethylꢀ1Hꢀbenzo[d]imidazoleꢀ2ꢀyl)methyl)amine),
tivity of oxoꢀiron(IV) complexes with respect to oxygen atom
transfer (OAT), hydrogen atom transfer (HAT), and electron
transfer (ET) processes is markedly affected by the supporting
and both species show the same reaction mechanisms and
reactivity patterns. Moreover, gas phase reactivity studies on
[(PyTACN)Fe=O(X)] compounds (PyTACN
7
=
1ꢀ(2′ꢀ
3
ligands and the spinꢀstate of iron.
pyridylmethyl)ꢀ4,7ꢀdimethylꢀ1,4,7ꢀtriazacyclononane;
X
=
–
–
–
–
OTf , ClO , CF COO , NO ) have shown that the nature of
However, the exact influence of the spin state in nonꢀheme
iron systems is still under current debate. Density functional
theory (DFT) calculations have predicted a higher reactivity of
highꢀspin ferryl species in the quintet state (S = 2) compared to
the triplet state (S = 1) due to the availability of an additional
channel for the approach of the substrate C–H bond to the
4
3
3
the coordinating anions X has a more dramatic effect on the
reactivity than the spin state.
8
Unambiguous comparison of the reactivity of models in triꢀ
plet and quintet ground states is usually hampered by the clasꢀ
sical ‘twoꢀstateꢀreactivity’ (TSR), a result of the small energy
9
IV
4
Fe (O) unit and a lower activation barrier. Although recent
gap between the triplet (S = 1) ground state and the quintet (S
ACS Paragon Plus Environment

Journal of the American Chemical Society
2) excited state in several oxoꢀiron(IV) model systems.
Page 2 of 13
=
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Initiated by interaction with the substrate, the spin state of the
system changes when approaching the transition state (TS) and
the reaction then proceeds via the quintet surface with a markꢀ
edly lower activation barrier. However, experimental methods
to test such computational predictions are very limited. Reꢀ
cently it has been proposed that the H/D kinetic isotope effect
We herein present a combined experimental and theoretical
(
KIE) may serve as a sensitive mechanistic probe of the spin
IV
investigation of the reactivity of the unique organometallic
state pathway during C–H activation by nonꢀheme Fe (O)
IV
Fe (O) complex 1 towards a set of substrates featuring varyꢀ
complexes, with very large nonꢀclassical KIEs indicating S = 1
1
0
ing C–H bond dissociation energies (BDE). In order to examꢀ
ine how the tripletꢀstate reactivity of the complex 1 compares
with the wellꢀestablished TSR, we chose the electronically
distinct complex A, a representative example for classical oxoꢀ
iron(IV) compounds in tetragonal ligand field that feature a
as the reactive spin state. Evaluation of KIEs for a series of S
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
IV
=
1 Fe (O) complexes with different supporting ligands and
in combinations with different substrates suggested various
factors to contribute to the extent of TSR as well as to extent
of tunneling, the latter rendering the S = 1 HAT process more
9
11
IV
rather small tripletꢀquintet energy gap allowing for TSR. In
favorable. In that situation, a benchmark nonꢀheme Fe (O)
complex whose reactivity is confined to singleꢀstate pathways
and that can serve as a ‘tripletꢀonly’ model seems highly deꢀ
sirable. Unfortunately though, all common intermediate spin
(S = 1) systems typically show a small energetic tripletꢀquintet
separation and are thus potentially amenable to TSR.
principle, the different electronic structures determined for
complexes 1 and A may lead to disparate reaction mechanisms
for a given transformation. Therefore, a fair comparison of
their reactivity is possible only if the HAT reactions with both
complexes follow the same mechanism. To this end, we perꢀ
formed a detailed computational study by using density funcꢀ
tional theory (DFT) and wavefunction based multireference
approaches to evaluate the HAT reactivity of complexes 1 and
A, and to dissect the electronicꢀstructure changes along their
HAT trajectories. As will be shown, comparison of these
closely related systems, viz. complexes 1 and A provides
direct insight into the importance of spinꢀstate on reactivity of
oxoꢀiron(IV) compounds, with 1 representing a benchmark
model for pure, intrinsic S = 1 reactivity of ferryl species.
In 2013, one of our groups prepared and characterized a
unique example of an oxoꢀiron(IV) compound ligated by a
NHC
IV
2+
macrocyclic tetracarbene ligand, [L Fe (O)(MeCN)] (1)
NHC
(L
=
3,9,14,20ꢀtetraazaꢀ1,6,12,17ꢀtetraazoniapentaꢀ
cyclohexacosaneꢀ1(23),4,6(26),10,12(25),15,17(24),21ꢀ
octaene), which is the only authenticated organometallic oxoꢀ
1
2
iron(IV) published so far. Complex 1 is topologically related
to the prominent cyclamꢀbased oxoꢀiron(IV) complex
TMC
IV
2+
TMC
[
L
Fe (O)(MeCN)] (A) (L
,4,8,11ꢀtetraazaꢀcyclotetradecane),
= 1,4,8,11ꢀtetramethylꢀ
but is electronically
1
3
1
distinct. A recent inꢀdepth electronicꢀstructure investigation by
using helium tagging infrared photodissociation (IRPD), abꢀ
sorption and magnetic circular dichroism (MCD) spectroscopy
in combination with DFT and highly correlated wave function
based multireference calculations revealed that the equatorial
tetracarbene ligand in 1 does barely affect the bonding in the
RESULTS AND DISCUSSION
C–H bond activation by complex 1. As shown in Scheme
2, tetracarbene coordinated oxoꢀiron(IV) complex 1 was genꢀ
erated in situ using the soluble iodosobenzene derivative 2ꢀ
(tBuSO )C H ꢀIO following procedures described previousꢀ
2 6 4
IV
12
Fe (O) unit. However, the very strong equatorial carbene σꢀ
ly. Complex 1 was found to react with various substrates
with relatively weak C–H bonds such as 1,4ꢀcyclohexadiene
donors push up the dx ꢀy
in energy than the dz . The Fe−O stretching vibrations of comꢀ
plex 1 are very similar to that of A, not only in the ground
2
2
orbital to an extent that it lies higher
(
CHD), 9,10ꢀdihydroanthacene (DHA), 9Hꢀxanthene and 9Hꢀ
fluorene. Reactions were performed in acetonitrile (1 mM) at –
40 °C in order to avoid any competing selfꢀdecay of 1 to the
2
ꢀ
1
−1 14,15
ꢀꢀ
state (1: 832 cm ; A: 834 cm )
but also in the excited state
oxo diiron(III) species 3; under these conditions the reaction
progress for the C–H activation can be conveniently moniꢀ
tored.
that arises from lifting an electron from the nonbonding d
xy
ꢀ
1
orbital into the Fe−O π* orbitals (d /d ) (1: 616 cm ; A: 610
xz yz
ꢀ
1 14,16
cm ).
More importantly, starting from the same electron
donating orbitals (EDOs), d /d , the excitations to dz
2
feature
xz yz
ꢀ
1
Scheme 2. Formation of 1 (a) and its reaction with C–H
substrates such as CHD and DHA (b); absorption maxima
of species 1 – 4 are indicated.
nearly identical energies for both complexes (~17000 cm ),
but the excitations to d are found to have a much higher
energy for complex 1 (24300 cm ) than that for A (12900 cm
2
2
x ꢀy
ꢀ
1
ꢀ
1
). All observations point to distinct ligandꢀfieldꢀsplitting
patterns for complexes 1 and A as sketched in Scheme 1. As a
consequence, the substantial geometric changes associated
*
with occupation of the Fe−O σ orbital (d
2
) on the quintet
z
excited state rationalize a very large tripletꢀquintet energy gap
(18.6 kcal/mol) calculated for 1. This can be expected to transꢀ
late into pure tripletꢀstate reactivity, because the quintet state
is essentially inaccessible.
Scheme 1. Structures and GroundꢀState Electronic Conꢀ
12,14
13
figuration of 1
(Left) and A (Right).
ACS Paragon Plus Environment

Page 3 of 13
Journal of the American Chemical Society
1
2
Changes observed in UVꢀvis absorption spectra during the
reaction of 1 with CHD (50 equiv) are shown in Figure 1 as a
representative example. The initial spectrum shows the characꢀ
teristic bands of 1 at 400 and 610 nm (light green). Upon
addition of substrate, the UVꢀvis features start changing and
two reaction steps can be differentiated while the reaction
proceeds. During the first 100 seconds, new features at
λmax = 367, 448 and 558 nm for an intermediate 4 appear and
reach their maximum absorbance. In a second stage, the bands
at λmax = 448 and 558 nm start decreasing whereas the band at
λmax = 367 nm keeps rising and a new band at 675 nm emerges
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Figure 2. Part of the cryoꢀESI mass spectrum of 1 (0.25 mM)
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
during the reaction with CHD (25 mM); peak assignments:
NHC II 2+
NHC IV
2+
+
[L Fe ]
at m/z = 202.1, [L Fe (O)]
at 210.0,
NHC III
2+
NHC
IV
(
300 s shown in Figure 1, right). This second process continꢀ
ues slowly and the absorbance at 448 nm almost vanishes
leading to a final spectrum that is characteristic for the ꢀoxo
[L Fe (OH)] at 210.5, [(L )Fe O(OTf)] at 569.1 and
NHC
III
+
[(L )Fe (OH)(OTf)] at 570.1. Peaks for the hydroxoꢀiron(III)
ꢀ
species 4 reach their maximum after around 100 s.
diiron(III) complex 3. Treatment of the data as two sequential
processes enables the detection of isosbestic points at 379 and
13 nm during the second step (Figure 1, right).
In the present case, deriving kinetic data from the UVꢀvis
spectral changes proved quite challenging. The most common
strategy reported in literature is to monitor the decay of the
active ferryl species upon addition of different amounts of
substrates. However, this is not a suitable approach for 1,
because the final decay species (i.e., ꢁꢀoxo complex 3) has
optical features in the same region as 1. However, the rather
unusual observation of two well separated sequential processꢀ
es in the UVꢀvis spectra for these CꢀH bond activation reacꢀ
tions allowed to focus on this first step only. It was thus decidꢀ
ed to follow the kinetics of the absorbance at 448 nm which
6
III
corresponds to the formation of the Fe (OH) intermediate (4).
This band shows a marked change in intensity reflecting the
formation and subsequent decay of 4 (see inset of Figure 1),
leading to consistent kinetic results for all sets of experiments,
Figure 1. UVꢀvis spectral changes observed upon reaction of 1 (1
mM) with 50 equivalents CHD at –40 °C in acetonitrile during the
initial step (about 100 s; left) and the second process (right). Inset:
Kinetic trace at 448 nm for this transformation and exponential fit
IV
both with different concentrations of Fe (O) complex 1 and of
the substrates. As mentioned above, under the applied reaction
conditions, selfꢀdecay of 1 to 3 is negligible on the experiꢀ
mental time scale. Thus, it is assumed that the decay of 1 is
caused exclusively by reaction with the added substrate via Hꢀ
(
red dashed line). * indicates isosbestic points.
Intermediate 4 is assigned to a hydroxoꢀiron(III) species
Scheme 2) which is generally accepted in literature to form
III
atom transfer (HAT) to form Fe (OH) intermediate 4 (inꢀ
(
crease of the absorbance at 448 nm). Therefore, the kobs deꢀ
rived from fitting the absorption trace to a single exponential
can be considered as the absolute value kobs for the decrease of
the active ferryl complex 1. Experiments at variable concenꢀ
trations of 1 with a constant concentration of CHD (60 eq)
showed kobs to be independent of the concentrations of 1, conꢀ
firming that this process follows pseudo first order kinetics,
and that we are dealing with a bimolecular reaction (Figure
S1).
during the first mechanistic step upon abstraction of a subꢀ
2,3b,17,18
strate Hꢀatom by the oxoꢀiron(IV) complex.
Supporting
evidence for this assignment comes from cryospray mass
spectrometry experiments (Figure 2). To that end, 1 was genꢀ
erated at −40 °C in acetonitrile and its initial mass spectrum
was recorded. Then, CHD (100 equiv) was added and the
evolution of the mass spectrum was followed over time. The
initial spectrum of 1 shows two main peaks at m/z = 210.1
NHC
2+
NHC
+
([L Fe(O)] ) and 569.1 ([L Fe(O)(OTf)] ). Upon subꢀ
strate addition, a significant increase of the peaks at
As depicted in Figure 3 (left), the plot of kobs values versus
different concentrations of substrate gave linear correlations.
m/z = 210.6 2+and 570.1 corresponding to
species
NHC
NHC
+
[
L
Fe(OH)] and [L Fe(OH)(OTf)] are observed and
Consequently, the observed log(k ’) of conversion for each
2
reach their maximum intensity after around 100 s, in accordꢀ
ance with the UVꢀvis spectral changes and indicative of the
formation of intermediate 4. For analyzing the changes in the
mass spectrum, it is important to know that the µꢀ
oxodiiron(III) species 3 is not stable under ESI conditions and
dissociates into 1 and 2. Therefore, the ESI mass spectrum of
the final product 3 (brown) shows peaks for its fragments
substrate (k ’ is the second order rate constant k divided by
2
2
the number of equivalent C−H bonds on the substrate; for
values see Table S1) were plotted versus the bond dissociation
1
9
energy (BDE) of the weakest C–Hꢀbonds (Figure 3, right).
This correlation gives a relatively good linear fit with a slope
of –0.39, which resembles the value obtained for previously
3b,3c, 20 , 21
reported oxoꢀiron(IV) complexes.
This finding supꢀ
NHC
2+
NHC
2+
[
L
Fe] and [L Fe(O)] , the former giving rise to the
ports the notion that the process of C–H bond activation by 1
emerging peak at m/z = 202.1 (Figure 2, left).
occurs via an HAT mechanism and is sensitive to the BDE
22
according to a BellꢀEvansꢀPolanyi (BEP) like correlation. No
reactivity of 1 was observed towards substrates with stronger
ꢀ
1
CꢀH bonds such as 2,3ꢀdimethybutane (BDE = 84 kcal mol )
ꢀ
1
or cyclohexane (BDE = 99 kcal mol ).
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 4 of 13
The reaction of 1 with 20 eq. of CHD, DHA and DHAꢀd4
was performed in the temperature range between −40 °C and 0
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
°C, and the experimentally determined k values were used to
2
derive activation parameters from an Eyring analysis. Plotting
ln(k /T) versus 1/T (Figures S3 and S4) gives good linear fits
2
‡
‡
with ꢁH = 8.5 kcal/mol and ꢁS = −22.9 cal/Kmol for CHD
‡
‡
and ꢁH = 8.5 kcal/mol and ꢁS = −22.4 cal/Kmol for DHA,
respectively (Table 1). Thus, the obtained activation parameꢀ
ters for the reaction of 1 with both substrates are very similar,
in agreement with the similar BDE. In accordance with the
higher BDE for a CꢀD bond, a significantly higher activation
enthalpy (13.0 kcal/mol, Figure S5) was determined for the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 3. Left: Kinetic data for 1 with various substrates at 233 K
in MeCN. Right: Plot of log(k ‘) against the weakest C–H BDE of
2
deuterated substrate. Surprisingly, C–H/C–D exchange is also
different substrates for 1.
accompanied by a decrease in entropy of activation. This
effect might be due to enthalpy/entropy compensation as obꢀ
served for intermolecular interactions involving hydrogen
To gain more insight into the mechanism of this process,
deuterated dihydroanthracene (DHAꢀd ) was used to study the
4
30
bonds.
kinetic isotope effect (KIE). As shown in Figure S2, the reacꢀ
tion with the deutero substrate is much slower than for the
23
Table 1. Thermodynamic parameters of the reaction of 1
protio substrate, leading to k /k = 32 ± 8 at −40 °C. This
H
D
with CHD, DHA and DHAꢀd derived from an Eyring Plot.
4
markedly high value is close to the range reported for enzymes
such as TauD and nonꢀheme iron compounds reacting with
1
1
1
with DHAꢀ
d4
ethylbenzole, whereas pMMO and most of the cyc20la,2m4,2ꢀ5
with CHD
with DHA
coordinated iron complexes show lower KIE’s (10ꢀ20).
‡
ꢁH / kcal/mol
8.5 ± 0.2
8.5 ± 0.2
13.0 ± 0.5
However, all these unusual high values are suggestive of nonꢀ
‡
20,7b
ꢁ
S / cal/K mol
ꢁG 20°C/ kcal/mol
–22.9 ± 0.8 –22.4 ± 0.7 –11.6 ± 2.0
15.2 ± 0.9
classical hydrogen tunneling effects
and are consistent with
‡
15.1 ± 0.8
16.4 ± 1.1
HAT as RDS. Importantly, the large KIE for 1 is in agreement
with reactivity occurring along an S = 1 pathway, though it
decreases with increasing temperature as previously observed
for other systems, from KIE = 32 at −40 °C to 18 at −20 °C
and 11 at 0 °C (see Table S2). Extrapolating the experimental
data, only a small a KIE (~5) is expected at 20°C, in agreeꢀ
ment with the similar free energy of activation at this temperaꢀ
ture (vide infra, Table 1). The temperature dependence reflects
the decreasing influence of the differences in activation enꢀ
thalpy for C–H and C–D with increasing temperature while
the changes in entropy become more relevant. Tunneling
contributions, which are barely affected by temperature, may
complicate these trends in KIE as has been pointed out in
3
a
Table 2. Comparison of kinetic parameters of 1 and A.
All reactions were performed at –40°C in MeCN.
3
a
A
1
2.2
0.76
0.48
6.4ꢂ10
ꢀ2
k2(xanthene)
k2(DHA)
k2(CHD)
6.6ꢂ10
2.5ꢂ10
6.4ꢂ10
–3
–4
ꢀ5
ꢀ3
k2(fluorene)
3.2ꢂ10
Reactivity comparison between 1 and A. The experimenꢀ
tally obtained bimolecular reaction rates allowed a direct comꢀ
parison with the literature reported values for complex A
10,26
literature.
3
a
under identical conditions (MeCN, –40°C). Rate constants
presented in Table 2 clearly show that 1 features significantly
higher HAT reactivity than A by about two to three orders of
magnitude depending on the substrate. Comparison with other
Additionally, we studied the reaction products resulting
from the oxidation of DHA by 1. As described in literature,
under stoichiometric conditions a nonꢀrebound HAT process
leads to a maximum formation of 0.5 eq. of anthracene per 1
IV
IV
literature known Fe (O) compounds classifies A to have
eq. of Fe (O) complex, as two Hꢀatoms need to be abstractꢀ
18
27
rather low and 1 to have intermediate reactivity. The so far
ed. In our case, the yield was determined to be already 0.32
most reactive oxoꢀiron(IV) complexes do not only show reacꢀ
tion rates up to four orders of magnitude higher than 1 but also
eq. of anthracene (64% of the theoretical maximum yield)
when the maximum in absorbance at 448 nm was reached after
around 100 s, and 0.37 eq. (74%) after 1 h. Consequently, the
major amount of anthracene (87% of the final yield) is already
generated during the initial process which is in line with the
the ability to activate substrates with stronger C–H bonds such
7b,31
as cyclohexane (BDE = 99 kcal/mol).
Numerous reports
have demonstrated the crucial influence of the supporting
ligands on the oxidation and oxygenation reactivity of oxoꢀ
III
proposed formation of a Fe (OH) intermediate in that first
2,3
iron(IV) complexes. For macrocyclic tetradentate scaffolds,
reaction step. Notably, anthracene was the only observed
product indicating that after the initial HAT no radical reꢀ
variations such as different axial coligands or different macroꢀ
cylic ring sizes can significantly increase or decrease the reacꢀ
tion rate. For example, the oxoꢀiron(IV) unit coordinated by
28
bound that might lead to any oxygenated product occurs. In
this context it is interesting to note that oxygen atom transfer
reactivity of 1 is essentially lacking. In fact, no significant
13ꢀTMC
(13ꢀTMC
=
1,4,7,10ꢀtetramethylꢀ1,4,7,10ꢀ
IV
tetraazacycloꢀtridecane) – a macrocyclic ligand system smaller
reactivity of Fe (O) complex 1 toward common substrates
3
by one CH group compared to 14ꢀTMC – is about 3ꢂ10 times
2
such as MeSPh or PPh was observed. This effect might be
3a
3
more reactive than A. Having this in mind, it is not surprising
due to the high basicity of the tetracarbene oxoiron(IV) system
that allows for moderate HAT reactivity while the 2ꢀelectron
that 1 and A show different reactivity in C–H bond activation
due to their differences in the supporting ligand system.
OAT process is not influenced by pK but is correlated with
a
2
9
Important parameters that determine the driving force of
HAT processes by oxoꢀiron(IV) species are the
the redox potential only (vide infra).
ACS Paragon Plus Environment

Page 5 of 13
Journal of the American Chemical Society
complexes 1 and A, respectively. Thus, one can anticipate that
iron(III)/iron(IV) redox potential as well as the basicity of the
3
2
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Fe(O) unit. The interplay of both factors determines the
ground state energy and thus the thermodynamics of the PCET
process, for which HAT can be viewed as a special case where
the proton and electron involved originate from the same
source, viz. the CꢀH bond being cleaved. Unfortunately, so far
more accurate values can be obtained by the calculations inꢀ
cluding both iron 4d and 3p shells into the active space, which
likely covers most of the important electronꢀcorrelations.
However, the resulting active space CAS(18,17) becomes too
large to be handled by usual configuration interaction proceꢀ
dures. Our earlier B3LYP calculations predicted that the
triplet ground state of complex 1 is energetically well separatꢀ
ed from the lowestꢀenergy quintet state by 18.6 kcal/mol,
whereas a vanishingly small quintetꢀtriplet gap of ~ 2 kcal/mol
is found for complex A. Taken together, our theoretical results
unanimously suggest that the quintet state of complex 1 is
thermally inaccessible; hence, the HAT process can only occur
on the triplet surface. While in the case of complex A the
marginal quintetꢀtriplet energy gap permits an easy access to
the quintet surface; thus, for A the HAT reaction may take
place on the triplet and quintet surfaces.
14
it was not possible to determine the pK value or an exact
a
0
value for E for the tetracarbeneꢀcoordinated system discussed
herein. As mentioned above, because of the negligible OAT
reactivity we assume that the pK of 1 might be rather high
a
III
IV
while the Fe /Fe redox potential is relatively low. Hence the
discussion will be focused on the kinetics and activation enerꢀ
gies of these HAT reactions. The key distinction when comꢀ
paring the tetracarbene coordinated system 1 with A is that the
electronic structure of 1 differs fundamentally from those of
all previously reported tetragonal oxoꢀiron(IV) complexes
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
2
with equatorial Nꢀdonor ligands. Because of the extremely
strong σꢀdonating carbenes in 1, its dx²ꢀy² orbital is lifted in
energy resulting in a reversed energetic order of the high lying
For HAT processes performed by classical ferryl models
such as complex A, the feasible electron accepting orbitals
(EAOs) are found to be dxz/yz and d , whereas d and dx²ꢀy²
dx²ꢀy² and d orbitals and in a much larger HOMOꢀLUMO gap,
z²
z²
xy
which was anticipated to have a decisive effect of the potential
involvement of the S = 1 and S = 2 reactive spin states.
function as inspectors because the latter two orbitals cannot
40
interact with the target CꢀH σꢀbond. As such, an oxoꢀ
iron(IV) species with a given spin multiplicity can employ two
channels to activate C–H bonds, of which one is referred to as
the πꢀtrajectory with the dxz/yz orbitals serving as the EAO and
Calculations. To evaluate the reactivity differences of comꢀ
plexes 1 and A and to gain deep insights into their reaction
mechanisms, we have carried out detailed computational studꢀ
ies on their HAT processes. For transition metal complexes,
DFT calculations often fail to provide a reliable and systematꢀ
ic accuracy in predicting relative energies that can vary up to
the other involving the d orbital is termed σꢀroute. Therefore,
2
z
there are in total four viable pathways for the reaction with
4
0a,41
complex A, as depicted in the right panel of Scheme 3.
On
1
0 kcal/mol with different choice of exchangeꢀcorrelation
the other hand, one can envision that only two tripletꢀchannels
can be adopted by complex 1 owing to the unavailability of
the quintet state (Scheme 3, left panel).
33
functionals. A recent benchmark study showed that among a
variety of density functionals B3LYP delivers the most accuꢀ
+ 34
rate energy profile for H activation by (FeO) . In addition to
2
B3LYP, complete active space selfꢀconsistent field
35
Scheme 3. Possible substrate attack trajectories and elecꢀ
tron transfer pathways for the HAT processes mediated by
complex 1 and A. Highlighted in red are the most favoraꢀ
ble pathways. The inset depicts the optimal attack geomeꢀ
try of the substrates for the σ and π pathways.
(
CASSCF) followed by Nꢀelectron valence perturbation
36
theory up to second order (NEVPT2) , a wavefunction based
multireference approach, has been used to calculate the accuꢀ
rate energies of the critical points on the potentialꢀenergy
surfaces (PESs).
As elaborated above, the relative energies of the triplet and
quintet states for complexes 1 and A are crucial for their reacꢀ
tivity. To compute reliable quintetꢀtriplet energy separations
(ꢁEQ–T), we have considered three different active spaces. Our
earlier work showed that a minimum balanced active space
should consist of 12 electrons in 9 orbitals (CAS(12,9)), viz.
the five iron 3d based orbitals, the three oxygen pꢀorbitals, and
1
4,37
the bonding counterpart of the Feꢀdx²ꢀy² orbital.
On top of
CAS(12,9), we added the iron 4d and 3p shells into the active
space, leading to CAS(12,14) and CAS(18,12), respectively.
The former active space accounts for the doubleꢀshell effect,
which considerably affects the spinꢀstate energetics of transiꢀ
tion metal complexes, as extensively documented in the literaꢀ
38
ture. Recently, Pierloot and coworkers proposed that the
coreꢀcorrelation might also have appreciable influence on the
39
spinꢀstate energetics, for which the latter active space was
designed.
The CASSCF(12,9)/NEVPT2 calculations estimate rather
large quintetꢀtriplet energy differences for both complexes
(complex 1, ꢁEQ–T = 30.4 kcal/mol, complex A, ꢁEQ–T = 14.9
We succeeded in locating the two relevant triplet TSs for
complex 1 by using B3LYP (the nature of the TSs (σ vs. π)
will be verified below.), in addition to the four pathways for
complex A. To compute reliable barriers at the
CASSCF/NEVPT2 level of theory, we made very careful
kcal/mol). Inclusion of the second dꢀshell changes the ꢁEQ–T
values by ~ 5 kcal/mol (complex 1, ꢁEQ–T = 34.9 kcal/mol;
complex A, ꢁEQ–T = 8.6 kcal/mol), and the coreꢀcorrelation
effect lowers the energy gaps to 28.7 and 5.0 kcal/mol for
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 6 of 13
selection of active spaces in order to properly describe breakꢀ
ing CꢀH and forming OꢀH bonds. During the reaction the
change in the metal oxidation state induces substantial adjustꢀ
complex (having a smaller tetraꢀNHC macrocycle with four
46
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
methylene bridges) suggests that the trans coordination of
47
MeCN slows down the reaction. We carefully examined this
hypothesis for complex 1. In fact, the Fe–NCMe bonding in 1
is rather weak, because the dissociation of MeCN from the Fe
4
2
ments of the radial electron correlation. It has been shown
that to account for this effect the active space needs to include
4
3
metal double dꢀshells; hence, we have to add key Fe 4d
orbitals instead of Fe 3p orbitals in the active space. Eventualꢀ
ly, an active space of CAS(14,14) for the TSs is required. A
detailed description of the active space construction is docuꢀ
mented in the experimental section. The CASSCF/NEVPT2
energy profiles are presented in Figure 4, which, to the best of
our knowledge, is the first HAT PES constructed by multiꢀ
center is computed to be almost thermoneutral (Figure S8).
3
The optimized geometry of TS shows that the FeꢀNCMe
σ
bond is completely broken with the Fe–NCMe distance of
3
3.929 Å (vs. 2.438Å in RC). Therefore, fiveꢀcoordinate comꢀ
plex 1′ may take part in the actual HAT process. Similar to
complex 1, complex 1′ with a large quintetꢀtriplet separation
of 15.6 kcal/mol (CASSCF/NEVPT2 value) also features
singleꢀstate reactivity (Figure 4b). We found that the
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
reference calculations. For the TSs and intermediates (Is), we
m
have used shorthand notations X , where X represents either
CASSCF(14,14) calculations on the B3LYPꢀoptimized geꢀ
n
3
TS or I, and m and n are the spin multiplicity (triplet or quinꢀ
tet) and the type of TS (σ or π), respectively. The free energy
difference between the reaction complex (RC) and the sepaꢀ
rated reactants (1 or A + CHD) is largely dictated by an unfaꢀ
vorable entropy term for a combination reaction, which is
partly compensated by weak nonꢀcovalent interactions. Such
effects mainly originate from dynamic electron correlation that
cannot be appropriately treated by the CASSCF approach,
which is designed to cover most of the static electron correlaꢀ
tion. In this regards, DFT calculations in combination with
explicit nonꢀcovalent corrections are expected to deliver reliaꢀ
ometry of TS always converge to a dominant electron conꢀ
π
3
3
figuration analogous to TS , thus indicating that TS is situꢀ
ated at a much higher energy than TS . In comparison with
σ
π
3
σ
complex 1, our theoretical results predict enhanced reactivity
of complex 1′. This could be readily attributed to the diminꢀ
ished energetic penalty required to promote an electron from
the σCꢀH bond, the EDO of HAT, to the stabilized Feꢀd orbital
z²
4
8
(EAO) in complex 1′. The overall HAT barrier for complex
1′ is 15.1 and 15.7 kcal/mol delivered by the
CASSCF/NEVPT2 and B3LYP computations, respectively
(Figures 4b). The corresponding values for complex 1 are 18.1
and 18.5 kcal/mol.
44
ble results. Hence, we did not calculate this energy gap by
using CASSCF.
Given the uncertainty of our calculations, the computed barꢀ
Our CASSCF(14,14)/NEVPT2 results for complex 1 reveal
riers are in reasonable agreement with the activation parameꢀ
3
that the σꢀpathway ( TS ) is more favorable than the πꢀroute
‡
σ
ters determined experimentally (ꢁG
= 15.2 ± 0.9
3
20°C
(
TS ) by ~ 4 kcal/mol (Figure 4a, upper panel), whereas a
π
kcal/mol), especially for complex 1′. In addition to CHD, we
also calculated the HAT reactivity of complex 1 toward DHA
(Figure S9). Our DFT results show that both reactions involve
similar barriers of ~ 19 kcal/mol owing to their nearly identiꢀ
barrier difference of ~2 kcal/mol, falling within the wellꢀ
45
established DFT error range (~3 kcal/mol), is estimated by
the B3LYP calculations (Figure 4a, lower panel). Despite this
subtle difference, the barriers calculated by the two theoretical
approaches are considerably lower than the corresponding
quintetꢀtriplet gaps, thereby corroborating the notion that the
reaction with complex 1 indeed exhibits pure tripletꢀstate
reactivity. A recent DFT study on a related tetracarbene ferryl
cal C–H bond dissociation energies (CHD: 78 kcal/mol; DHA:
19
7
7 kcal/mol ).
ACS Paragon Plus Environment

Page 7 of 13
Journal of the American Chemical Society
Figure 4. Computed free energy (∆G) profiles of the HAT processes from CHD for (a) complex 1 (b) complex 1′ (c) complex A.
The CASSCF/NEVPT2 profile is presented in the upper panel and the B3LYP profile is presented in the lower panel. Black line:
triplet pathways; Red lines: quintet pathways.
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
In case of complex A, our CASSCF/NEVPT2 results
reliable
spinꢀstate
energetics
(vide
supra),
our
demonstrate that the quintet σꢀpathway has the lowest barrier
CASSCF/NEVPT2 calculations on complex A may somewhat
5
3
(8.6 kcal/mol) and hence is the most efficient reaction channel
overestimate the relative energy of RC relative to RC. Howꢀ
5 3
to accomplish HAT (Figure 4c), while the triplet channels
ever, by neglecting the energy gap between RC and RC and
the spinꢀcrossover barrier, we can estimate the lowest limit of
the HAT barrier for complex A. The value of 16.1 kcal/mol
thus obtained is similar to the lowest barriers found for comꢀ
plexes 1 and 1′. Our CASSCF/NEVPT2 calculations thus
indicate that complex 1 is likely to be slightly more reactive
than complex A, consistent with the observed two to three
orders of magnitude difference in their reaction rates, which
translates into a reactionꢀbarrier separation of only 2–4
kcal/mol. In contrast to this, DFT calculations predict competiꢀ
tive reactivity for the two systems within the uncertainties of
DFT methods (~3 kcal/mol). This lends credence to our emꢀ
ployed CASSCF/NEVPT2 method and showcases the imꢀ
portance of multireference calculations for oxoꢀiron(IV) sysꢀ
tems. As elaborated below, although complex 1 features a
distinct electronic structure, its HAT reaction follows the same
mechanism as the triplet state HAT of complex A. As such,
our theoretical results show that complex 1, while exhibiting
singleꢀstateꢀreactivity, can activate C–H bonds at least with
paralleled efficiency to complex A, which features twoꢀstateꢀ
reactivity.
5
need to traverse barriers at least 4 kcal/mol higher. For TS ,
π
our CASSCF(14,14) computations couldn’t converge to the
5
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
correct electronic state, but to one corresponding to TS , an
σ
3
analogous situation found for TSπ of complex 1′. The
CASPT2 (multireference perturbation theory up to second
order based on CASSCF wavefunction) calculations on a
model oxoꢀiron(IV) system (the oxoꢀiron(IV) site in magnesiꢀ
4
9
umꢀdiluted Fe (dobdc) metalꢀorganic framework; dobdc = 2,5ꢀ
2
5
dioxoꢀ1,4ꢀbenzenedicarboxylate) suggest that TS is stabiꢀ
lized by 5.8 kcal/mol relative to TS . The barrier for the
σ
5
50
π
quintet σꢀpathway estimated by B3LYP is 4.9 kcal/mol, and
the closest triplet σꢀchannel is ~7 kcal/mol higher in energy,
both values comparable to those reported by Siegbahn and
51
coworkers. In comparison with the CASSCF/NEVPT2 reꢀ
5
sults, B3LYP considerably underestimate the TSσ barrier,
51
which has been observed in an earlier computational study
on the reactivity of complex A and attributed to the selfꢀ
52
interaction error of DFT calculations. Furthermore, to more
accurately evaluate the reactivity of complex A, we located the
minimumꢀenergy crossing point (MECP) between its triplet
and quintet surfaces, which can be regarded as the upper
bound of the corresponding spinꢀcrossover barrier. It turns out
Electronicꢀstructure analysis. To gain more mechanistic
insights about complex 1, we carried out a detailed analysis by
mapping the electronicꢀstructure changes along the reaction
coordinate. Oxoꢀiron(IV) complexes feature rather covalent
ironꢀoxo interactions, consisting of two πꢀbonds involving the
3
that the MECP is situated only 1.5 kcal/mol above RC, fallꢀ
ing within the barrier range of 1–4 kcal/mol calculated for
5
3
similar spinꢀcrossover processes. The optimized geometries
5
3
of TS and TS for complex A show that the axial Fe–N
σ
σ
Feꢀdxz/yz and Oꢀp atomic orbitals and one σꢀbond formed by
x/y
40a,16, 54
bonds are only slightly lengthened by 0.09 and 0.26 Å, respecꢀ
tively. Thus, dissociation of the trans axial ligand is unlikely
to happen for complex A, different from complex 1. Considerꢀ
ing the inadequacy of the employed active space in predicting
Feꢀdz² and Oꢀpz.
To properly describe such covalent
interactions, a balanced active space has to include not only
metalꢀcentered antibonding orbitals but also the corresponding
43
bonding orbitals.
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 8 of 13
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
Figure 5. Comparison of the electronic structures of the reactant complex (RC) and transition states (TS) involved in the HAT
processes by complex 1 and CHD. (a), (b), and (c) represents CASSCF natural orbitals along with the occupation numbers in paꢀ
3
3
3
renthesis below the orbital labels and atomic contributions for TS , RC and TS , respectively; (d), (e), and (f) represents
σ
π
3
3
3
CASSCF spinꢀdensity for TS , RC and TS , respectively. (green: positive density; magenta: negative density)
σ
π
*
As shown in Figure 5b, the occupation numbers (ONs) of
most valence orbitals are markedly greater than 0.02 and lower
than 1.98, in CASSCF calculations a virtual molecular orbital
O
N
o
f
t
h
e
a
n
t
i
b
o
n
d
i
n
g
σ
o
r
b
i
t
a
l
w
i
t
h
t
h
e
p
r
e
d
o
m
i
n
a
n
t
F
e
ꢀ
d
z
z
²
character, a formal VMO, appreciably differs from zero (0.13).
In the HAT TSs, the target CꢀH is partially broken. Followꢀ
ing the same line of the reasoning about how to construct a
balanced active space, we need to include the σC–H and σ
orbitals into the active space. The spinꢀdensity plot of TSσ
(VMO) typically having an ON of no more than 0.02 and a
*
doubly occupied molecular orbital (DOMO) having an ON of
no less than 1.98. This observation clearly indicates significant
multireference character in complex 1, and the results delivꢀ
ered by singleꢀdeterminant DFT approaches hence need be
viewed with caution. The sum of the ONs of the bonding and
antibonding orbitals is close to that expected for an electron
C–H
3
shows that some negative spin density develops in the subꢀ
strate, indicative of a σ TS (Figure 5d). In comparison with
complex 1, no substantial changes in the ONs of all valence
orbitals were found except σ and σ (Figure 5a). Surprisingly,
*
z
z
2
2
4
2
*
2
*
0
* 0
z
*
configuration of (σ ) (σ ) (π ) (nb d ) (π ) (σ ) (σ ) (eq
eq
z
x/y
xy
x/y
eq
the ON of σ z rises at the expense of that of σ , and conseꢀ
z
= equatorial, nb = nonꢀbonding). The computed spinꢀdensity
plot of complex 1 demonstrates that the oxo group has large
donutꢀlike positive spin density in the xyꢀplane and small yet
discernable negative spin density in the zꢀdirection (Figure 5e,
topꢀview). Due to the covalent ironꢀoxo bonding, the singly
occupied molecular orbitals (SOMOs) mainly with the Feꢀ
dxz/yz parentage contain substantial contributions from the Oꢀ
px/y atomic orbitals (~30%), thereby resulting in sizeable posiꢀ
tive spin density on the oxo ligand. Because the Fe center has
a much larger spin population than the oxo ligand, the exꢀ
change stabilization induces some positive spin density in the
Feꢀdz² orbital and accordingly some negative spin density in
quently the sum of the ONs for both orbitals remain two.
Thus, the deviations of the ONs from their ideal numbers for a
3
DOMO and a VMO become more pronounced in TS . This
σ
change signals the onset of homolytic cleavage of the iron–
oxo σꢀbond, similar to H homolysis. At the equilibrium geꢀ
2
ometry of H , one can anticipate that the ON of the σ
orꢀ
2
H–H
*
bital should be two, and that of σ
should be zero. As the
H–H
H–H bond lengthens, the ON of σ
decreases and simultaneꢀ
H–H
*
ously that of σ
increases. Eventually, both values are close
H–H
to one when the two hydrogen atoms are farther apart. This
interpretation rationalizes the much larger negative spin densiꢀ
3
ty in the Oꢀp orbital found for TS compared to complex 1.
z
σ
the Oꢀp orbital, a spinꢀpolarization process taking place. This
z
Therefore, our multireference calculations suggest that during
•–
analysis explains why the ON of the Oꢀp based σ orbital, a
z
z
the reaction a transient oxyl radical (O ) forms, the same
conversion found in the quintet σꢀpathway of HAT by classiꢀ
formal DOMO, considerably deviates from two (1.87), and the
ACS Paragon Plus Environment

Page 9 of 13
cal ferryl complexes.
Journal of the American Chemical Society
Originally in complex 1 the π orꢀ first ab initio study on the reactivity of oxoꢀiron(IV) species. A
4
0c,55
x
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
bital is Oꢀp based; during the reaction this orbital acquires
detailed analysis reveals that in the course of HAT, the
Fe (O) core in complex 1 undergoes the same electronicꢀ
x
IV
substantial Feꢀdxz character and has dominant Feꢀdxz parentage
3
in TS , this observation pointing to a transformation from
structure changes as those found for classical ferryl models
such as A, i.e. for both systems the reactions follow an identiꢀ
cal mechanism. As such, complex 1 is a trendꢀbreaking examꢀ
ple of an S = 1 oxoꢀiron(IV) species that shows efficient reacꢀ
tivity confined to singleꢀstate pathways, and that now repreꢀ
sents a pure ‘tripletꢀonly’ model.
π
III
•–
oxoꢀiron(IV) to oxylꢀiron(III) (Fe –O ). By analyzing the
lead electron configurations of the CASSCF wave functions of
3
3
TS and TS , one can reach the same conclusion (for details,
σ
π
5
0
see the supporting information, Tables S3,S4). The resulting
oxyl radical interacts with the EDO, leading to formation of a
pair of threeꢀcenter orbitals, viz. σCHO and σ CHO, and the forꢀ
*
mer orbital evolves to σO–H in hydroxylꢀiron(III) and the latter
EXPERIMENTAL SECTION
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
•
to a Cꢀradical, reminiscent of an organic radical, such as OH ,
General Considerations. All reactions were carried out in
UV/vis cuvettes prepared inside a glove box (MBRAUN
LabMaster) under nitrogen atmosphere with less than 0.1 ppm
O and H O and kept under a nitrogen atmosphere. Acetoniꢀ
54
abstracting an Hꢀatom from alkanes. Thus, the C–H bond
activation by ferryl intermediates can be essentially viewed as
a radical reaction, but at the early phase of the reaction the
2
2
56
oxyl radical is masked by the covalent ironꢀoxo interaction.
trile was dried and degassed by standard procedures before
Apparently, the reaction involves rather sophisticated electronꢀ
icꢀstructure changes, for which the electron shift diagram
depicted in Scheme 3 only gives a phenomenological picture
of it. In comparison with the HAT processes mediated by
use. Soluble iodosobenzene, DHAꢀd and compound 1 were
4
58,20,12
prepared according to literature.
Other chemicals were
purchased from commercial sources and used as received.
Cryo Mass spectrometry was recorded using a highꢀresolution
mass spectra (HRꢀMS) Bruker MicrOTOFꢀQ IITM instrument
using cryospray ionization sources using a cryospray attachꢀ
ment and setting the temperature of the nebulizing and drying
gas at ꢀ40ºC
5
1,57
IV
classical ferryl species,
the reactive Fe (O) unit in complex
1
undergoes the same electronicꢀstructure adjustments. Thereꢀ
fore, the reactivity of complex 1, while featuring a distinct
electronic configuration, represents the intrinsic triplet HAT
efficacy of ferryl intermediates.
UVꢀvis spectroscopy was performed using an Agilent 8453
UVꢀvis spectrophotometer. Absorption of 1 mM solutions of
the complexes in acetonitrile was measured from 200 nm to
800 nm with a resolution of 1 nm in a 1 cm UV quartz cells.
The low temperature control was performed with a cryostat
from Unisoku Scientific Instruments (Japan). The GC measꢀ
urement was performed by a Thermo Finnigan TRACE GC
with EI and Varian GC Capillary Column.
CONCLUSIONS
The organometallic
L
S
=
1
oxoꢀiron(IV) complex
Fe (O)(MeCN)] (1), while topologically related to
other tetragonal oxoꢀiron(IV) complexes such as the standard
NHC
IV
2+
[
TMC
IV
2+
system [L Fe (O)(MeCN)] (A), features a unique elecꢀ
tronic structure, as evidenced recently by an extensive MCD
14
study. The strongly σꢀdonating equatorial tetracarbene ligand
IV
LNHC
2+
IV
Formation of [Fe O(
)(MeCN)] and reaction with
in 1 does not affect the bonding within the Fe (O) unit but
II LNHC
2+
substrates. A 1 mM solution of [Fe (
)(MeCN) ] (OTf)
2
2
strongly destabilizes the dx²ꢀy² orbital, which leads to a very
large energetic tripletꢀquintet splitting and presumably prohibꢀ
its any spin crossover to the highꢀlying S = 2 state. To assess
the electronic structure impact on the reactivity, we have here
presented experimental results for the C–H activation reacꢀ
tions of 1 with a set of substrates featuring CꢀH BDEs ranging
from 75 to 80 kcal/mol. BEPꢀlike correlation of second order
(1.569 mg, 2 ꢃmol) in acetonitrile (2 mL) was cooled o –40
°
C. A precooled and freshly prepared solution of 2ꢀ(tBuSO )ꢀ
2
C H ꢀIO (1.905 mg, 5.60 ꢃmol, 2.8 eq.) in CH Cl (190.5 ꢃL)
6
4
2
2
was added and the cuvette was transferred out of the glove box
and the UV/vis absorption measurement at –40 °C was started.
The appropriate amount of substrate (10 – 160 eq) was disꢀ
solved in Acetonitrile (CHD) or CH Cl (DHA, Xanthene,
2
2
rate constants k’ with BDEs supports the notion that the proꢀ
2
Fluorene) to give 100 ꢃL (because of solubility problems
sometimes up to 200 ꢃL) which were then added to the cuvette
using a syringe. The subsequent increase of the characteristic
absorption band of the iron(III)ꢀhydroxo (448 nm) was directly
monitored and fitted to a single exponential function.
cess of C–H bond activation by 1 occurs via a HAT mechaꢀ
nism, and the putative hydroxoꢀiron(III) HAT intermediate has
been detected by cryoꢀESI mass spectrometry. No reaction
occurs with substrates having stronger CꢀH bonds, and also no
significant OAT reactivity of 1 is observed. Kinetic parameꢀ
ters for the HAT process and comparison with the large body
of known oxoꢀiron(IV) complexes shows that the reactivity of
Product analysis for the oxidation of DHA was performed
by gas chromatography. Therefore, a 1 mM solution of crysꢀ
talline material of 2 (2.21 mg, 2.91 ꢃmol) in acetonitrile
(2.91 mL,) was reacted with 10 eq of DHA (5.25 mg in 100
1
is moderate, but significantly higher than the reactivity of
related standard compound A that was proposed to follow a
TSR pathway under the same experimental conditions; second
ꢃL CH Cl ) as described previously. The yield was determined
2
2
after 3 min, which corresponds to maximum absorption at
48 nm, and after 1 h. Prior to injection, the solution was
order rate constants k’ are two to three orders of magnitude
2
4
higher in case of 1. The very large KIE observed for 1 is in
agreement with S = 1 being the reactive spin state.
CASSCF/NEVPT2 and DFT calculations confirmed that for 1
the reactions with CꢀH substrates such as CHD and DHA take
place on the triplet surface only, because the computed HAT
barrier on the triplet surface is substantially lower than the
tripletꢀquintet energy separation. The theoretical results delivꢀ
ered by the CASSCF/NEVPT2 calculations nicely explain the
observed reactivity difference between complexes 1 and A;
however, DFT fails. To the best of our knowledge, this is the
filtered through silica and washed with acetonitrile, ethyl
acetate and dichloromethane. Anthracene was the only obꢀ
served product and its amount was compared to leftover
amount of DHA since both substances show very similar
intensity in GC. After 3 min 0.93 ꢃmol (64% compared to the
maximum theoretical yield) anthracene was detected, after 1 h
1
.08 ꢃmol (74%).
Computational Details. All calculations were performed
59
by using the ORCA program package. Geometry optimizaꢀ
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 10 of 13
tions were undertaken by employing the hybridꢀGGA (GGA =
generalized gradient approximation) density functional,
Supplemental Figures on experimental results described in the
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
text, details about CASSCF and DFT calculations, additional
information on the reaction energetics for CHD and DHA subꢀ
strates and optimized Cartesian coordinates of all the species. This
material is available free of charge via the Internet at
http://pubs.acs.org.
60
61
B3LYP, in conjunction with def2ꢀTZVP(ꢀf) basis set withꢀ
out f polarization function for the first coordination sphere (Fe,
O, N and C) and C and H participating in the reaction and
62
def2ꢀSVP basis set for the rest of atoms. The scalar relativꢀ
6
3
istic zerothꢀorder regular approximation (ZORA) was apꢀ
plied in geometry optimizations as well as single point energy
calculations. To speed up the overall calculations, the
AUTHOR INFORMATION
Corresponding Authors
6
4
RIJCOSX approximation was applied for the expensive
integral calculations. To be consistent with the experiment, we
carried out all calculations by applying the conductorꢀlike
*Eꢀmail: franc.meyer@chemie.uniꢀgoettingen.de
*Eꢀmail: shengfa.ye@cec.mpg.de
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
polarizable continuum model (CPCM) with acetonitrile (ε =
Notes
65
3
6.6) being the solvent. Nonꢀcovalent interactions, important
The authors declare no competing financial interest.
for the reaction complexes and transition states, were accountꢀ
ed for by using atomꢀpairwise dispersion corrections with
ACKNOWLEDGMENT
Support by the Fonds der Chemischen Industrie (Kekuléꢀ Scholarꢀ
ship for C.K.), the Deutsche Forschungsgemeinschaft (IRTG 1422
4
4
BeckeꢀJohnson (D3BJ) damping. Subsequent numerical
frequency calculations were undertaken on the optimized
geometries, which revealed stationary points featuring no
imaginary frequency and firstꢀorder saddle points having one
imaginary frequency. The zeroꢀpoint vibrational energies,
thermal corrections and entropy terms were obtained from the
frequency calculations. Final single point energy calculations
were undertaken by using the B3LYP functional in conjuncꢀ
“Metal Sites in Biomolecules: Structures, Regulation and Mechaꢀ
nisms”) and COST Action CM1003 (ECOSTBio, STSM at UdG,
Spain for C.K.) is gratefully acknowledged. B.M., F.N. and S.Y.
gratefully acknowledge financial support from the MaxꢀPlanck
society.
6
6
tion with a more flexible basisꢀset def2ꢀTZVPP.
Multiconfigurational calculations based on complete active
3
5
REFERENCE
space selfꢀconsistent field (CASSCF) followed by Nꢀelectron
36
valence perturbation theory up to second order (NEVPT2)
6
6
were performed in conjunction with def2ꢀTZVPP basis set.
For the reaction complexes (RCs), we chose CAS(12,14)
activeꢀspace consisting of five iron 3d based orbitals, the three
oxygen pꢀorbitals, the bonding counterpart of the Feꢀdx2ꢀy2
(1) (a) Krebs, C.; Fujimori, D. G.; Walsh, C. T.; Bollinger, J. M.
Acc. Chem. Res. 2007, 40, 484 – 492; (b) Nam, W. Acc. Chem. Res.
2
5
007, 40, 522 – 531. (c) Que, L. Jr. Acc. Chem. Res. 2007, 40, 493–
00.
*
orbital and five iron 4d orbitals. The σC–H and σ C–H orbitals of
(2) (a) McDonald, A. R.; Que L., Jr. Coord. Chem. Rev. 2013, 257,
14 – 428; (b) Nam, W.; Lee, YꢀM.; Fukuzumi, S. Acc.Chem. Res.
014, 47, 1146 – 1154; (c) Oloo, W. L.; Que, L., Jr., Acc. Chem. Res.
015, 48, 2612−2621; (d) Nam, W. Acc. Chem. Res. 2015, 48,
415−2423.
the substrate were not incorporated into the activeꢀspace beꢀ
cause of limited electron correlation effects between the two
orbitals that don’t need to be treated at the full configurationꢀ
interaction level. While for the transition states (TSs) the
4
2
2
2
doubly occupied σ
orbital and the corresponding antiꢀ
are necessary owing to the threeꢀcenter
(3) (a) Hong, S.; So, H.; Yoon, H.; Cho, K.ꢀB.; Lee, Y.ꢀM.; Fukuꢀ
CHO
*
bonding orbital σ
zumi, S.; Nam, W. Dalton Trans. 2013, 42, 7842–7845. (b) Kaizer, J.;
Klinker, E. J.; Oh, N. Y.; Rohde, J.ꢀU.; Song, W. J.; Stubna, A.; Kim,
J.; Münck, E.; Nam, W.; Que, L., Jr. J. Am. Chem. Soc. 2004, 126,
CHO
interaction (C–H–O) that develops in the TSs. A active space
of CAS(14,11) including five Fe 3d based orbitals, the three
4
72–473. (c) Comba, P.; Fukuzumi, S.; Kotani, H.; Wunderlich, S.
oxygen pꢀorbitals, the bonding counterpart of the Feꢀd
x2ꢀy2
*
Angew. Chem., Int. Ed. 2010, 49, 2622–2625. (d) England, J.;
Martinho, M.; Farquhar, E. R.; Frisch, J. R.; Bominaar, E. L.; Münck,
E.; Que, L., Jr. Angew. Chem., Int. Ed. 2009, 48, 3622–3626. (e)
Lacy, D. C.; Gupta, R.; Stone, K. L.; Greaves, J.; Ziller, J. W.;
Hendrich M. P.; Borovik, A. S. J. Am. Chem. Soc. 2010, 132, 12188–
12190.
(4) Janardanan, D.; Wang, Y.; Schyman, P.; Que, L., Jr.; Shaik, S.
Angew. Chem. Int. Ed. 2010, 49, 3342–3345.
(5) Xue, G.; De Hont, R.; Münck, E.; Que, L., Jr. Nat. Chem. 2010,
2, 400–405.
orbital, σCHO and σ
was therefore constructed. In addition,
CHO
to account for the doubleꢀshell effect, as was accounted for the
RCs, it is necessary to add the iron 4d shell on top of
CAS(14,11).
However,
the
resulting
activeꢀspace,
CAS(14,16), became too large to compute with the usual
CASSCF calculations. Therefore, to make a balance between
size and accuracy we added three iron 4d orbitals (d , d , d )
xy
xz
yz
to correlate the important t2g set of electrons and an active
space of CAS(14,14) results. The CASSCF/NEVPT2 reaction
(
6) England, J.; Guo, Y.; Van Heuvelen, K. M.; Cranswick, M. A.;
barriers
were
therefore
calculated
using
the
Rohde, G. T.; Bominaar, E. L.; Münck, E.; Que, L., Jr. J. Am. Chem.
Soc. 2011, 133, 11880–11883.
CASSCF(12,14)/NEVPT2 and CASSCF(14,14)/NEVPT2
energies for RCs and TSs, respectively. The solvent and therꢀ
mal corrections and the entropy terms to the Gibbs free energy
changes (∆G) were obtained from the B3LYP calculations.
These corrections are listed in Table S9 in the supporting
information.
(
7) (a) Seo, M. S.; Kim, N. H.; Cho, K.ꢀB.; So, J. E.; Park, S. K.;
Clémancey, M.; GarciaꢀSerres, R.; Latour, J.ꢀM.; Shaik, S.; Nam, W.
Chem. Sci. 2011, 2, 1039–1045. (b) Biswas, A. N.; Puri, M.; Meier,
K.K.; Oloo, W.N.; Rohde, G.T.; Bominaar, E. L.; Münck, E.; Que, L.,
Jr. J. Am. Chem. Soc. 2015, 137, 2428ꢀ2431. (c) Bae, S.H.; Seo, M.S.;
Lee, YꢀM.; Cho, K.ꢀB.; Kim, W.ꢀS.; Nam, Wonwoo Angew. Chem.
Int. Ed. 2016, 55, 8027ꢀ8031.
(
8) (a) Andris, E.; Jaꢀ
̌ík, J.; Gómez, L.; Costas, M.; Roithóva,
ASSOCIATED CONTENT
Supporting Information.
J. Angew. Chem. Int. Ed. 2016, 55, 3637ꢀ3641. (b) Andris, E.;
ACS Paragon Plus Environment

Page 11 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Navrátil, R.; Jasík, J.; Thibault, T.; Srnec, M.; Costas, M.; Roithóva,
J. J. Am. Chem. Soc. 2017, 139, 2757–2765.
9) (a) Shaik, S.; Hirao, H.; Kumar, D. Acc. Chem. Res. 2007, 40,
32–542. (b) Hirao, H.; Que, L., Jr.; Nam, W.; Shaik, S. Chem. Eur.
J. 2008, 14, 1740–1756.
10) (a) Mandal, D.; Ramanan, R.; Usharani, D.; Janardanan, D.;
Wu, X.; Seo, M. S.; Cho, K.ꢀB.; Yoon, H.; Park, Y. J.; Fukuzumi, S.;
Pushkar, Y. N.; Nam, W., J. Am. Chem. Soc. 2013, 135, 6388ꢀ6391.
(d) GarciaꢀBosch, I.; Company, A.; Cady, C. W.; Styring, S.; Browne,
W. R.; Ribas, X.; Costas, M., Angew. Chem. Int. Ed. 2011, 50, 5648–
5653. (e) Wang, D.; Ray, K.; Collins, M. J.; Farquhar, E. R.; Frisch, J.
R.; Gómez, L.; Jackson, T. A.; Kerscher, M.; Waleska, A.; Comba, P.;
Costas, M.; Que, L. Jr. Chem. Sci. 2013, 4, 282–291.
(30) Dunitz, J. D., Chem. Biol.1995, 2, 709–712.
( 31 ) Collins, T. J.; Ryabov, A. D. Chem. Rev. DOI:
10.1021/acs.chemrev.7b00034
(32) (a) Warren, J. J.; Tronic, T. A.; Mayer, J. M., Chem. Rev.
2010, 110, 6961–7001. (b) Geng, C.; Ye, S.; Neese, F. Dalton Trans.
2014, 43, 6079–6086.
(33) (a) Harvey, J. N. Annu. Rep. Prog. Chem., Sect. C: Phys.
Chem. 2006, 102, 203−226. (b) Ghosh, A. J. Biol. Inorg. Chem. 2006,
11, 712−724. (c) Pierloot, K.; Vancoillie, S. J. Chem. Phys. 2006,
125, 124303. (d) Pierloot, K.; Vancoillie, S. J. Chem. Phys. 2008,
128, 034104. (e) Vancoillie, S.; Zhao, H.; Radon, M.; Pierloot, K. J.
Chem. Theory Comput. 2010, 6, 576−582.
(34) Altun, A.; Breidung, J.; Neese, F.; Thiel, W. J. Chem. Theory
Comput. 2014, 10, 3807–3820.
(35) (a) Roos, B. O. Adv. Chem. Phys. 1987, 69, 399−445. (b)
Ruedenberg, K.; Cheung, L. M.; Elbert, S. T. Int. J. Quantum Chem.
1979, 16, 1069−1101. Ross, B. O.; Taylor, P. R.; Siegbahn, P. E. M.
Chem. Phys. 1980, 48, 157–173.
(36) Angeli, C.; Cimiraglia, R.; Evangelisti, S.; Leininger, T.;
Malrieu, J. P. J. Chem. Phys. 2001, 114, 10252–10264.
(37) Ye, S.; Xue, G.; Krivokapic, I.; Petrenko, T.; Bill, E.; Que, L.,
Jr; Neese, F. Chem. Sci. 2015, 6, 2909–2921.
(38) (a) Jannuzzi, S. A. V; Phung, Q. M.; Domingo, A.; Formiga,
A. L. B.; Pierloot, K. Inorg. Chem. 2016, 55, 5168–5179. (b) Alcoverꢀ
Fortuny, G.; Caballol, R.; Pierloot, K.; de Graaf, C. Inorg. Chem.
2016, 55, 5274–5280.
(39) Pierloot, K.; Phung, Q. M.; Domingo, A. J. Chem. Theory
Comput. 2017, 13, 537–553.
(40) (a) Mondal, B.; Roy, L.; Neese, F.; Ye, S. Isr. J. Chem. 2016,
56, 763–772. (b) Ye, S.; Neese, F. Proc. Natl. Acad. Sci. 2011, 108,
1228–1233.
(
5
(
Wang, B.; Shaik, S. J. Am. Chem. Soc. 2015, 137, 722–733. (b)
Mandal, D.; Shaik, S. J. Am. Chem. Soc. 2016, 138, 2094–2097.
(
11) (a) England, J.; Prakash, J.; Cranswick, M. A.; Mandal, D.;
Guo, Y.; Münck, E.; Shaik, S.; Que, L., Jr. Inorg. Chem. 2015, 54,
828−7839. (b) Kwon, Y. H.; Mai, B. K.; Lee, Y. M.; Dhuri, S. N.;
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
7
Mandal, D.; Cho, K. B.; Kim, Y.; Shaik, S.; Nam, W. J. Phys. Chem.
Lett. 2015, 6, 1472–1476.
(12) Meyer, S.; Klawitter, I.; Demeshko, S.; Bill, E.; Meyer, F.
Angew. Chem. Int. Ed. 2013, 52, 901–905.
(
13) Rohde, J.ꢀU.; In, J.ꢀH.; Lim, M. H.; Brennessel, W. W.;
Bukowski, M. R.; Stubna, A.; Münck, E.; Nam, W.; Que, L., Jr.
Science 2003, 299, 1037−1039.
(
14) Ye, S.; Kupper, C.; Meyer, S.; Andris, E.; Navrátil, R.; Krahe,
O.; Mondal, B.; Atanasov, M.; Bill, E.; Roithová, J.; Meyer, F.;
Neese, F. J. Am. Chem. Soc. 2016, 138, 14312ꢀ14325.
(
15) Jackson, T. A.; Rohde, J.ꢀU.; Seo, M. S.; Sastri, C. V.;
DeHont, R.; Ohta, T.; Kitagawa, T.; Münck, E.; Nam, W.; Que, L.,
Jr.; Stubna, A. J. Am. Chem. Soc. 2008, 130, 12394−12407.
(
16) Decker, A.; Rohde, J.ꢀU.; Klinker, E. J.; Wong, S. D.; Que, L.,
Jr.; Solomon, E. I. J. Am. Chem. Soc. 2007, 129, 15983−15996.
(17) Cho, K.ꢀB.; Hirao, H.; Shaik, S.; Nam, W. Chem. Soc. Rev.
2
016, 45, 1197–1210.
18) Kleespies, S. T.; Oloo, W. N.; Mukherjee, A.; Que, L., Jr.
Inorg. Chem. 2015, 54, 5053−5064.
19) Luo, Y.ꢀR. Handbook of Bond Dissociation Energies in Orꢀ
(
(
ganic Compounds; CRC Press LLC: Boca Raton, FL, 2003.
(20) Sastri, C. V.; Lee, J.; Oh, K.; Lee, Y. J.; Lee, J.; Jackson, T.
A.; Ray, K.; Hirao, H.; Shin, W.; Halfen, J. A.; Kim, J.; Que, L., Jr.;
Shaik, S.; Nam, W. Proc. Natl. Acad. Sci. 2007, 104, 19181–19186.
(
21) Planas, O.; Clémancey, M.; Latour, J.ꢀM.; Company, A.; Cosꢀ
tas, M. Chem. Commun. 2014, 50, 10887 – 10890.
22) (a) Bell, R. P. Proc. R. Soc. London, Ser. A 1936, 154, 414–
429. (b) Evans, M. G.; Polanyi, M. Trans. Faraday Soc. 1938, 34, 11–
3.
(
(41) (a) Ye, S.; Neese, F. Curr. Op. Chem. Biol. 2009, 13, 89–98.
(b) Shaik, S.; Chen, H.; Janardanan, D. Nat. Chem. 2010, 3, 19–27.
(42) Malrieu, J. P.; Guihéry, N.; Calzado, C. J.; Angeli, C. J. Comꢀ
put. Chem. 2006, 28, 35–50.
2
(
23) The error of this KIE value was derived from several experiꢀ
ments and error bars for each individual kobs value. Because of the
very slow reactions with deuterated substrates at low temperatures,
the differences in rate for the different concentrations of DHAꢀd are
4
relatively large compared to the experimental error in the determinaꢀ
(
(
43) Pierloot, K. Mol. Phys. 2003, 101, 2083–2094.
44) Grimme, S.; Ehrlich, S.; Goerigk, L. J. Comput. Chem.2011,
3
2, 1456–1465.
tion of kobs values, resulting in a rather high uncertainty for the KIE.
(45) Siegbahn, P. E. M. J. Biol. Inorg. Chem. 2006, 11, 695–701.
(46) Anneser, M. R.; Haslinger, S.; Pöthig, A.; Cokoja, M.; Basset,
(
24) Price, J. C.; Barr, E. W.; Tirupati, B.; Bollinger, M., Jr.;
Krebs, C. Biochemistry 2003, 42, 7497–7508.
25) Nesheim, J. C.; Lipscomb, J. D. Biochemistry 1996, 35,
J.ꢀM.; Kühn, F. E. Inorg. Chem. 2015, 54, 3797–3804.
47) Cantú Reinhard, F. G.; de Visser, S. P. Chem. Eur. J. 2017,
3, 2935–2944.
48) Bernasconi, L.; Baerends, E. J. Eur. J. Inorg. Chem. 2008,
(
(
1
0240–10247.
2
(26) (a) Kwart, H., Acc. Chem. Res. 1982, 15, 401–408. (b)
(
Knapp, M. J.; Klinman, J. P., Eur. J. Biochem. 2002, 269, 3113–3121.
c) Wang, Q.; Sheng, X.; Horner, J. H.; Newcomb, M. et al., J. Am.
1672–1681.
(
(49) (a) Andersson, K.; Malmqvist, P.ꢀÅ.; Roos, B. O.; Sadlej, A.
J.; Wolinski, K. J. Phys. Chem. 1990, 94, 5483−5488. (b) Andersson,
K.; Malmqvist, P.ꢀÅ.; Roos, B. O. J. Chem. Phys. 1992, 96,
1218−1226.
(50) Verma, P.; Vogiatzis, K. D.; Planas, N.; Borycz, J.; Xiao, D.
J.; Long, J. R.; Gagliardi, L.; Truhlar, D. G. J. Am. Chem. Soc. 2015,
137, 5770–5781.
(51) Johansson, A. J.; Blomberg, M. R. A.; Siegbahn, P. E. M. J.
Phys. Chem. C 2007, 111, 12397–12406.
(52) Lundberg, M.; Siegbahn, P. E. M. J. Chem. Phys. 2005, 122,
224103.
Chem. Soc. 2009, 131, 10629–10636. d) Klinman, J. P., Chem. Phys.
Lett. 2009, 471, 179–193.
(
27) Company, A.; Prat, I.; Frisch, J. R.; MasꢀBallesté, R.; Güell,
M.; Juhász, G.; Ribas, X.; Münck, E.; Luis, J. M.; Que, L. Jr.; Costas,
M. Chem. Eur. J. 2011, 17, 1622 – 1634.
(28) (a) Geng, C.; Ye S.; Neese, F. Angew. Chem., Int. Ed. 2010,
49, 5717; (b) Cho, K.ꢀB.; Wu, X.; Lee, Y.ꢀM.; Kwon, Y. H.; Shaik,S.;
Nam W. J. Am. Chem. Soc. 2012, 134, 20222; (c) Rana, S.; Dey, A.;
Maiti, D. Chem. Commun. 2015, 51, 14469–14472.
(
29) (a) Parsell, T. H.; Behan, R. K.; Green, M. T.; Hendrich, M.
P.; Borovik, A. S., J. Am. Chem. Soc. 2006, 128, 8728–8729. (b)
Parsell, T. H.; Yang, M.ꢀY.; Borovik, A. S., J. Am. Chem.
Soc. 2009, 131, 2762–2763. (c) Chen. J.; Lee, Y.ꢀM.; Davis, K. M.;
(53) Harvey, J. N. WIREs Comput. Mol. Sci. 2013, 4, 1–14.
(54) (a) Lehnert, N.; Ho, R. Y. N.; Que, L., Jr.; Solomon, E. I. J.
Am. Chem. Soc. 2001, 123, 8271–8290. (b) Decker, A.; Rohde, J.ꢀU.;
ACS Paragon Plus Environment

Journal of the American Chemical Society
Page 12 of 13
1
2
3
4
5
6
7
8
9
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
2
3
3
3
3
3
3
3
3
3
3
4
4
4
4
4
4
4
4
4
4
5
5
5
5
5
5
5
5
5
5
6
Que, L., Jr.; Solomon, E. I. J. Am. Chem. Soc. 2004, 126, 5378–
379. (c) Neese, F. J. Inorg. Biochem. 2006, 100, 716–726. (d)
Bernasconi, L.; Louwerse, M. J.; Baerends, E. J. Eur. J. Inorg. Chem.
007, 3023–3033. (e) Kumar, D.; Karamzadeh, B.; Sastry, G. N.; de
Visser, S. P. J. Am. Chem. Soc. 2010, 132, 7656–7667.
55) (a) Neidig, M. L.; Decker, A.; Choroba, O. W.; Huang, F.;
Kavana, M.; Moran, G. R.; Spencer, J. B.; Solomon, E. I. Proc. Natl.
Acad. Sci. 2006, 103, 12966–12973. (b) Srnec, M.; Wong, S. D.;
England, J.; Que, L., Jr.; Solomon, E. I. Proc. Natl. Acad. Sci. 2012,
109, 14326–14331. (c) Mai, B. K.; Kim, Y. Inorg. Chem. 2016, 55,
5
2
(
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
1
2
3
4
5
6
7
8
9
0
3
844–3852.
56) (a) Ye, S.; Geng, C.ꢀY.; Shaik, S.; Neese, F. Phys. Chem.
(
Chem. Phys. 2013, 15, 8017–8030. (b) Dietl, N.; Schlangen, M.;
Schwarz, H. Angew. Chem., Int. Ed. 2012, 51, 5544–5555.
(
57) (a) Kumar, D.; Hirao, H.; Que, L.; Shaik, S. J. Am. Chem. Soc.
2005, 127, 8026–8027. (b) Hirao, H.; Kumar, D.; Que, L., Jr.; Shaik,
S. J. Am. Chem. Soc. 2006, 128, 8590–8606.
(
58) Macikenas, D.; SkrzypczakꢀJankun, E.; Protasiewicz, J. D. J.
Am. Chem. Soc. 1999, 121, 7164ꢀ7165.
(
(
59) Neese, F. WIREs Comput. Mol. Sci. 2012, 2, 73–78.
60) (a) Becke, A. D. Phys. Rev. A 1988, 38, 3098–3100. (b) Lee,
C.; Yang, W.; Parr, R. G. Phys. Rev. B 1988, 37, 785–789.
(61) Schäfer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100,
5
2
829.
(
62) Schäfer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97,
571.
(
63) (a) van Lenthe, E.; Baerends, E. J.; Snijders, J. G. J. Chem.
Phys. 1993, 99, 4597−4610. (b) van Lenthe, E.; Baerends, E. J.;
Snijders, J. G. J. Chem. Phys. 1994, 101, 9783−9792. (c) van Lenthe,
E.; van Leeuwen, R.; Baerends, E. J.; Snijders, J. G. Int. J. Quantum
Chem. 1996, 57, 281−293.
(
64) Neese, F.; Wennmohs, F.; Hansen, A.; Becker, U. Chem.
Phys. 2009, 356, 98–109.
(65) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999–3094.
(
66) Weigend, F.; Ahlrichs, R. Phys. Chem. Chem. Phys. 2005, 7,
3
297.
ACS Paragon Plus Environment

Page 13 of 13
Journal of the American Chemical Society
1
2
3
4
5
6
7
8
9
Table of Contents artwork
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
ACS Paragon Plus Environment