Roles of Iron Complexes in Catalytic Radical Alkene Cross-Coupling: A Computational and Mechanistic Study

Article abstract of DOI:10.1021/jacs.9b02117

A growing and useful class of alkene coupling reactions involve hydrogen atom transfer (HAT) from a metal-hydride species to an alkene to form a free radical, which is responsible for subsequent bond formation. Here, we use a combination of experimental and computational investigations to map out the mechanistic details of iron-catalyzed reductive alkene cross-coupling, an important representative of the HAT alkene reactions. We are able to explain several observations that were previously mysterious. First, the rate-limiting step in the catalytic cycle is the formation of the reactive Fe-H intermediate, elucidating the importance of the choice of reductant. Second, the success of the catalytic system is attributable to the exceptionally weak (17 kcal/mol) Fe-H bond, which performs irreversible HAT to alkenes in contrast to previous studies on isolable hydride complexes where this addition was reversible. Third, the organic radical intermediates can reversibly form organometallic species, which helps to protect the free radicals from side reactions. Fourth, the previously accepted quenching of the postcoupling radical through stepwise electron transfer/proton transfer is not as favorable as alternative mechanisms. We find that there are two feasible pathways. One uses concerted proton-coupled electron transfer (PCET) from an iron(II) ethanol complex, which is facilitated because the O-H bond dissociation free energy is lowered by 30 kcal/mol upon metal binding. In an alternative pathway, an O-bound enolate-iron(III) complex undergoes proton shuttling from an iron-bound alcohol. These kinetic, spectroscopic, and computational studies identify key organometallic species and PCET steps that control selectivity and reactivity in metal-catalyzed HAT alkene coupling, and create a firm basis for elucidation of mechanisms in the growing class of HAT alkene cross-coupling reactions.

Full text of DOI:10.1021/jacs.9b02117

Article
Cite This: J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
pubs.acs.org/JACS
Roles of Iron Complexes in Catalytic Radical Alkene Cross-Coupling:
A Computational and Mechanistic Study
†
‡
†
,‡
Dongyoung Kim, S. M. Wahidur Rahaman, Brandon Q. Mercado, Rinaldo Poli,*
,
†
and Patrick L. Holland*
†Department of Chemistry, Yale University, New Haven, Connecticut 06511, United States
‡LCC−CNRS, Universite
de Toulouse, INPT, 205 Route de Narbonne, BP 44099, Toulouse CEDEX 4 F-31077, France
S Supporting Information
́
*
ABSTRACT: A growing and useful class of alkene coupling
reactions involve hydrogen atom transfer (HAT) from a
metal-hydride species to an alkene to form a free radical,
which is responsible for subsequent bond formation. Here, we
use a combination of experimental and computational
investigations to map out the mechanistic details of iron-
catalyzed reductive alkene cross-coupling, an important
representative of the HAT alkene reactions. We are able to
explain several observations that were previously mysterious.
First, the rate-limiting step in the catalytic cycle is the
formation of the reactive Fe−H intermediate, elucidating the
importance of the choice of reductant. Second, the success of the catalytic system is attributable to the exceptionally weak (17
kcal/mol) Fe−H bond, which performs irreversible HAT to alkenes in contrast to previous studies on isolable hydride
complexes where this addition was reversible. Third, the organic radical intermediates can reversibly form organometallic
species, which helps to protect the free radicals from side reactions. Fourth, the previously accepted quenching of the
postcoupling radical through stepwise electron transfer/proton transfer is not as favorable as alternative mechanisms. We find
that there are two feasible pathways. One uses concerted proton-coupled electron transfer (PCET) from an iron(II) ethanol
complex, which is facilitated because the O−H bond dissociation free energy is lowered by 30 kcal/mol upon metal binding. In
an alternative pathway, an O-bound enolate-iron(III) complex undergoes proton shuttling from an iron-bound alcohol. These
kinetic, spectroscopic, and computational studies identify key organometallic species and PCET steps that control selectivity
and reactivity in metal-catalyzed HAT alkene coupling, and create a firm basis for elucidation of mechanisms in the growing
class of HAT alkene cross-coupling reactions.
INTRODUCTION
Scheme 1. General Mechanism Proposed for Catalytic HAT
Alkene Coupling Reactions
■
The alkene functional group is an inexpensive and versatile
building block for biologically active molecular structures.
However, the chemoselective functionalization of unactivated
alkenes to generate highly substituted carbon centers is
challenging, because traditional routes based on carbocations
1
often suffer from side reactions. As a milder alternative,
chemists have developed a class of transition-metal-catalyzed
Markovnikov-selective hydrofunctionalization reactions for
forming C−C, C−O, and C−N bonds from unactivated
1
b,2
alkenes.
These reactions have excellent functional-group
tolerance, predictable regioselectivity, and experimental sim-
3
plicity. The precatalysts are typically simple Mn, Fe, or Co
When this radical reacts with a trapping agent, the product is
the Markovnikov product with a new C−C, C−N, or C−O
bond. In many HAT alkene reactions, such as Co-catalyzed
hydrohydrazination, Fe-catalyzed hydroamination, hydrome-
thylation, and alkene coupling, the product of trapping is
salts of halide, acac, or oxalate, and the reducing equivalents
2
,4
come from borohydrides or silanes.
A general mechanism has been proposed for these reactions
Scheme 1), which features hydrogen atom transfer (HAT) to
(
the alkene from a putative metal-hydride complex, which is
1b
generated in situ from a metal salt and a hydride source.
HAT would generate a reduced metal species and the alkyl
radical at the more substituted side of the original alkene.
Received: February 24, 2019
©
XXXX American Chemical Society
A
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
1
b
1
1
another radical. According to the proposed mechanism, this
coupled radical undergoes an electron transfer reaction with
the metal byproduct, which oxidizes the metal to a form that
reacts with borohydride or silane to regenerate the active
metal-hydride.
observe by H NMR spectroscopy. Broad peaks in H NMR
spectra of a cobalt-catalyzed silylperoxidation reaction were
tentatively assigned to a metal-bound hydride, but the
8
interpretation of the data is ambiguous. In the iron-catalyzed
HAT alkene cross-coupling, we took first steps toward
In this report, we focus on one specific HAT alkene reaction,
the intermolecular cross-coupling of two alkenes. The catalytic
cycle is proposed to follow the general pathway of Scheme 1,
with an electron-poor alkene serving as the trap. This reaction
identifying relevant metal species, by discovering that during
catalysis both iron(II) and iron(III) acac complexes gave H
1
NMR resonances that could be assigned by comparison with
6b
independently synthesized complexes. These measurements
5
57
9
was reported by Baran in 2014, and later publications
were complemented by Fe Mossbauer spectroscopy studies
̈
demonstrated the reaction scope and utility in C−C bond
that showed that Fe(acac) is the reduced iron species and that
[Fe(acac) (μ-OEt)] is catalytically competent. However,
2
6
6b
constructions of synthetic relevance for natural products. This
2
2
reaction uses Fe(acac) as a precatalyst and PhSiH as a
reductant, and HAT catalysis was proposed to proceed
through the iron(III)/iron(II) cycle shown in Scheme 2. It
there have been no systematic computational studies showing
that the catalytic cycle as proposed in Scheme 2 is energetically
feasible or would give the observed selectivity.
3
3
Another issue that has been proposed in the context of HAT
1
b,3
10
Scheme 2. Previously Proposed Mechanism for Iron-
Catalyzed Intermolecular Cross-Coupling of Alkenes
alkene reactions
is the persistent radical effect (PRE). In
the PRE, the life of a reactive radical (normally susceptible to
bimolecular self-termination) is extended by its engagement in
an equilibrium with a high-concentration trapping species,
usually having radical character (the “persistent radical”), thus
lowering its concentration and hence the rate of bimolecular
termination (Scheme 3). The PRE often leads to cross-
Scheme 3. Persistent Radical Effect in Organometallic-
Mediated Processes
11
selectivity of radical addition reactions, and similarly leads to
10b
the controlled polymerization of alkenes.
The PRE can
influence metal-catalyzed reactions through reversible trapping
x
of a radical by a reduced metal (M ) to form a metal−alkyl
x+1
complex (M −R). If subsequent reactions take place only
was presumed that the hypothetical iron(III) hydride
selectively attacks the electron-rich “donor” alkene (blue in
Scheme 2), giving an iron(II) species and a nucleophilic alkyl
radical. This radical then is trapped by the electron-poor
through the free radical, then the formation of the M−C bond
“protects” most of the radical, lowering the concentration of
the free radical. This effect has been used for controlling free-
radical polymerization reactions by the approach termed
organometallic-mediated radical polymerization (OMRP), in
“acceptor” alkene (green in Scheme 2), and the formation of
1
2
the new C−C bond generates a product with a radical adjacent
to the electron-withdrawing group. In the final steps, this
radical accepts an electron and a proton to generate product.
Lo et al. concluded that this is a stepwise process proceeding
through a carbanion for two reasons: (a) The use of EtOD
solvent gives deuteration α to the electron-withdrawing group
which metal−alkyl species are well-documented.
In the alkene cross-coupling reaction, we used NMR and
Mossbauer spectroscopies to show that the vast majority of the
̈
iron (>95%) is present as reduced Fe(acac) , which could in
2
principle react with transient alkyl radicals to form an iron(III)
6b
alkyl species. Though these putative iron(III) alkyls were not
observed by Mossbauer spectroscopy, their concentration
(
shown in purple in Scheme 2), even though the ethanol O−H
̈
7
bond (104 kcal/mol) is stronger than the C−H bond that is
formed. (b) It was possible to perform a three-component
coupling that adds an aldehyde α to the electron-withdrawing
group. Below, we systematically evaluate a number of
alternative mechanisms for this proton-coupled electron
transfer (PCET) pathway.
could be too low to detect. In order to understand and control
the HAT alkene reactions, it is important to evaluate the
reactivity of such metal species, even when they are not
experimentally observed. Here, we describe the results of
mechanistic experiments and DFT computations that increase
our understanding of several key steps in the HAT alkene
cross-coupling by iron-acac catalysts. They illuminate unan-
ticipated mechanistic features that are likely to be applicable to
the broader class of HAT alkene reactions with Mn, Fe, and
Co. Therefore, the results are likely to be useful for analysis
and improvement of related reactions such as hydro-
pyridylation (Minisci reaction), hydroamination, hydroazida-
tion, hydrohydrazination, hydrocyanation, hydration, and
6
b
So far, this proposed mechanism lacks key tests of its
veracity. In the alkene cross-coupling, as in all the catalytic
HAT alkene reactions, no organometallic species or radicals
have been observed. Detection of the metal-containing species
in HAT alkene reactions in general is complicated by the lack
of chelating supporting ligands to stabilize the intermediate
metal complexes, and by the weak-field nature of halides, acac,
and oxalate that render the metal sites high-spin and difficult to
1b
hydrogenation.
B
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
RESULTS
and high-spin states have trigonal bipyramidal iron with the
hydride in the axial and equatorial positions, respectively. All
three spin states have energies within 1.2 kcal/mol of one
■
Formation of the Reactive Hydride Species. In an
6
b
earlier paper some of us developed a new catalyst,
III
another. Dimeric species [(acac) Fe (μ-H)] , in each spin
2
2
[
(acac) Fe(μ-OEt)] , which is much more active than the
2 2
state and with parallel or antiparallel spins on the two sites,
were found to be higher in energy than the corresponding
monomers by at least 5 kcal/mol. At catalytic concentrations of
iron that are much lower than the standard state of 1 M, the
dimer would be even more unfavorable. We also considered
the potential for solvent ethanol to coordinate to a monomeric
iron(III) hydride complex. However, the calculations show
that binding of ethanol is weak: the octahedral EtOH adducts
are not lower in energy in either the doublet or quartet state
Fe(acac) precatalyst. Its higher activity was attributed to the
3
ability of the ethoxide to react directly with silanes to give an
iron hydride. It was interesting that the rate of the catalytic
reaction depended on the hydride donor used: it was faster
when using the monoalkoxysilane PhSi(OEt)H , but slower
2
using PhSiH or PhSi(OEt) H. Obradors and Shenvi have
3
2
13
described an analogous rate increase using PhSi(OiPr)H2.
However, it was not clear whether this silane is a more active
hydride donor or whether the monoalkoxysilane somehow
modifies the iron catalyst. To resolve this question, we
performed the catalytic reaction with equal amounts of PhSiD3
(
Supporting Information section 2.C).
With the alkoxide and the hydride complexes modeled, we
compared the thermodynamics and kinetics for hydride
and PhSi(OEt)H . Product analysis by mass spectrometry
2
transfer from Si to Fe (details in Supporting Information
showed that this reaction gave less than 5% deuteration of the
product (Scheme 4 and Figure S25), which contrasts with the
III
section 2.D). The reaction of monomeric [Fe (acac) (OMe)]
2
(
[
S = 5/2) with PhSiH to form the aforementioned
3
III
Fe (acac) H] (S = 3/2) along with PhSi(OMe)H is uphill
2
2
Scheme 4. Competition Experiment Indicates that
i
by 5.4 kcal/mol. Adding the energy of splitting the alkoxide
dimer, the hydride product lies a total of 8.6 kcal/mol above
the starting materials. The transition state for H transfer from
PhSi(O Pr)H Is More Reactive than PhSiH
2
3
PhSiH to Fe (TS1) on the sextet PES lies 24.9 kcal/mol
3
above the starting materials (see Figure 1) in free energy. The
complete deuteration observed when PhSiD is used as the
3
1
4
sole silane. This result demonstrates that the monoalkox-
ysilane more rapidly generates the reactive hydride that ends
up in the product.
The reaction trajectory for hydride transfer with a slightly
truncated model system [(acac) Fe(μ-OMe)] was explored
2
2
using density-functional theory (DFT) calculations with the
BPW91* functional, which has been shown to be suitable for
15
variable spin states for 3d transition metals. The geometries
were optimized in the gas phase, and then the electronic
energies were corrected for thermal (PV) and entropy (TS)
effects at 298 K, dispersion forces, solvation in ethanol, and
solution standard state (see Computational Details in the
Supporting Information), to give standard-state free energies.
First, we resolved the electronic structure of the iron(III)
alkoxide starting material. We considered all possible spin
states for iron(III) (S = 1/2, 3/2, and 5/2, which are
conventionally termed low-, intermediate-, and high-spin,
respectively). Details are in the Supporting Information
(
section 2.A). The calculations show that the dimeric starting
material is most stable as an open-shell singlet (Stotal = 0), from
antiferromagnetically coupled ions with local spins SFe = 5/2.
The Stotal = 5 state lies only 0.3 kcal/mol higher, consistent
with the weak antiferromagnetic coupling observed exper-
Figure 1. Relative free energies of the starting materials (1/
III
2[Fe (OMe)(acac) ] + PhSiH ), of the corresponding monomeric
2
2
3
III
[Fe (OMe)(acac) ] intermediate, and the reaction coordinate over
TS1 to give [Fe H(acac) ] and PhSi(OMe)H .
2
III
2
2
1
6
III
imentally. Splitting into monomeric [Fe (acac) (OMe)] is
2
feasible, since the two high spin monomer molecules are only
corresponding TS along the quartet PES lies at a higher ΔG
value (34.4 kcal/mol, see Supporting Information section 2.D).
In light of the small barriers for other steps (see below), this
alkoxide/hydride exchange is predicted to be the turnover-
limiting step in the catalytic cycle. We measured the initial rate
6
.5 kcal/mol higher in energy than the dimer.
For the hydride product, we also considered both
monomeric and dimeric structures, since hydrides are often
17
found to be bridging in iron compounds. In this case, the
three monomeric spin isomers are calculated to be more stable
than the corresponding dimers (Supporting Information
−
5
for the catalytic reaction at 40 °C to be 6.5 ± 0.7 × 10 M/s,
⧧
which leads to an estimate of the experimental ΔG as 22.8 ±
section 2.B). The optimized structure of monomeric
0.2 kcal/mol. Thus, the computed barrier of 24.9 kcal/mol is
similar to that observed experimentally.
III
[
Fe (acac) H] is spin-state-dependent. The low-spin state
2
has a square pyramidal geometry at iron with the strong trans-
influence hydride in the axial position, while the intermediate-
The calculated trajectory of this step is interesting (Figure
2). First, the iron-bound alkoxide approaches the Si atom
C
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 5. Deuterium Transfer to Product Indicates that
HAT Is Not Reversible
reaction, the metal-bound hydrogen in the putative hydride is
derived from phenylsilane; thus, its fate can be probed by using
1
13
PhSiD in place of PhSiH . H and C NMR analysis of the
3
3
product (Figures S1 and S2) show that the cross-coupled
product has regiospecific D incorporation (Scheme 5), which
6a
agrees with previous work. In addition, it is important to
consider whether the H atom transfer is reversible, because M-
H additions from isolated hydride species to alkenes were
Figure 2. Partially optimized structures (at fixed Si···O distances) of
three representative points along the reaction pathway of precatalyst
activation by PhSiH3.
19
shown to be reversible by Halpern, Bullock, and Norton.
Since the radical generated from the M-D compound would be
singly deuterated, reversible HAT would lead to partially
deuterated alkene starting material at low conversion. There-
fore, a catalytic reaction run under normal conditions was
stopped at 3 min (30% conversion), and analysis by NMR
spectroscopy and mass spectrometry (Figures S1−S5) showed
that no D was incorporated into the remaining starting
material. Therefore, the initial hydrogen atom addition to the
alkene is irreversible, which agrees with our qualitative
rationale for the normal isotope effect previously observed
opposite from the phenyl group (pseudotrigonal bipyramidal
with the incoming O considered as a bond). Then, the
geometry at silicon undergoes a Berry pseudorotation to yield,
at the transition state, a geometry close to an ideal square
pyramid with a Si−H bond placed axially and then eventually
forms a new trigonal bipyramidal geometry with the trans-
ferring H atom and another H atom in the pseudoaxial
positions and the Ph and OMe substituents together with the
third H atom in the pseudoequatorial plane. From here, there
is simultaneous transfer of hydride to Fe and OMe to Si to
smoothly yield the products.
6b
upon substitution of PhSiD for PhSiH .
3 3
The pathway for HAT from the transient iron(III) hydride
to the donor alkene was evaluated by DFT using isobutene as
model H atom acceptor (Figure 3; see details in Supporting
The nature of this transition state helps to explain the
greater activity when moving from PhSiH to PhSi(OR)-
3
6
b,13
H .
2
The calculations, carried out with R = Me for
simplicity, reveal a greater driving force for the alkoxysilane, as
hydride transfer from the monoalkoxysilane is 4.1 kcal/mol less
1
8
endergonic (Supporting Information section 2.E). This
driving force comes from the greater hydridicity of the H
atom (calculated Mulliken charges on the H atoms are −0.087
in PhSiH , −0.118 in PhSi(OMe)H and −0.153 in PhSi-
3
2
(
OMe) H). The geometry of TS1 shows that one additional
2
alkoxide group may be accommodated in the axial position of
the square pyramidal geometry of the TS. Indeed, reoptimiza-
tion of TS1 after replacement of one silane H atom with OMe
gave a transition state corresponding to a barrier that is 3.6
kcal/mol lower in energy, which compares reasonably well to
the ΔΔG = 1.4 kcal/mol derived from the competition
experiment in Scheme 4. However, use of the dialkoxysilane
⧧
14
III
PhSi(OR) H as hydride donor, even though presumably
Figure 3. HAT pathway for ground-state [Fe H(acac)
2
] reaction
2
II
leading to a thermodynamically even more favorable trans-
formation, would place an additional OR group in an
equatorial position of the square pyramid where it would be
destabilized by a steric clash with the other equatorial
substituents. In this way, the computational investigation of
the hydride transfer to the Fe atom elucidates the empirically
observed silane trends. In light of this result, it seems likely that
appropriate silanes can be designed in the future to accomplish
faster hydride transfer for catalytic reactions.
with isobutene to give [Fe (acac) ] and tert-butyl radical.
2
Information section 2.F). This reaction yields a tert-butyl
radical and [Fe (acac) ], which was treated as the
II
2
experimentally known quintet ground state. Hence, the
reaction can take place on a single potential energy surface
(PES), i.e., is spin-allowed, starting from either the
intermediate-spin (S = 2 − 1/2 = 3/2) or high-spin (S = 2
+ 1/2 = 5/2) hydride complex, whereas the HAT from the
low-spin system would be spin-forbidden. The barrier for the
spin-allowed reaction along the quartet surface is extremely
low: we were able to optimize a transition state (TS2) that lies
only 4.7 kcal/mol higher in free energy, making this step
Hydrogen Atom Transfer from Fe−H to Alkene. In the
HAT alkene reactions, a key step is the transfer of a hydrogen
1b
atom from a metal-hydride complex to an alkene; therefore,
we sought ways to experimentally probe this step for the
intermolecular alkene cross-coupling. Thus, we studied the
HAT alkene cross-coupling of TBS-protected 3-methylbut-3-
en-1-ol with benzyl acrylate under the conditions of Lo et al.,
9
−1
−1
extremely rapid. A rate constant of 2.2 × 10 s
M
can be
calculated from the Eyring equation, though one should note
that the geometry optimization uses the electronic energy;
therefore, the saddle point on the free energy surface might be
III
6a
using Fe (acac) as the precatalyst (Scheme 5). In this
3
D
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
slightly different. In any case, this finding supports our
persistent radical effect that “protects” the intermediate alkyl
radical from bimolecular radical coupling.
conclusion above that HAT to the alkene is unlikely to be
6b
the turnover-limiting step as proposed earlier. HAT from Fe
to the alkene was calculated to be exergonic by 26 kcal/mol,
which agrees with the experimental finding that HAT occurs
The cross-selectivity of the alkene cross-coupling reaction
arises because the nucleophilic alkyl radical (in the
computations, modeled as a tert-butyl radical) reacts most
rapidly with electron-poor alkenes. The relative rates for
reaction of such radicals with alkenes are well-known in the
irreversibly under the cross-coupling conditions.
II
The iron product from HAT, [Fe (acac) ], is known to
2
21
form weak adducts with monodentate donor molecules,
literature. In this specific case, a radical is expected to react
2
0
II
3
including EtOH. Indeed, trans-[Fe (acac) (EtOH) ] has
more than 10 times more rapidly with an acrylate ester than
2
2
6
b
22
been isolated and crystallographically characterized. We
measured the variation of the chemical shift of the H NMR
resonance for [Fe (acac) ] in benzene as a function of
with a simple alkene. According to our calculations, the
1
addition of tert-butyl radical to methyl acrylate has a barrier
(TS3) of only 8.4 kcal/mol, and has a favorable ΔG = −15
kcal/mol (Figure 4 and Supporting Information section 2.I) as
expected for the formation of a stabilized radical. Therefore,
this C−C bond forming reaction is also expected to be
irreversible. Because the acrylate radical is more stable, it is
expected to form a weaker bond to the iron atom. This
II
2
[
EtOH], and the curve (Figure S20) yielded an association
−
1
constant of 3.2 ± 0.2 M at 80 °C, corresponding to ΔG° =
−
0.8 ± 0.1 kcal/mol. However, in this weak-binding regime,
the stoichiometry is not evident from the binding curve. DFT
calculations on the EtOH coordination process (Supporting
Information section 2.G) gave a ΔG of +1.3 kcal/mol for
proposition was explored using the simpler propionate radical
II
•
formation of [Fe (acac) (EtOH)] and −0.6 kcal/mol for
CH CH (COOMe) as a model for the produced radical. The
2
3
II
[
Fe (acac) (EtOH) ]. The excellent agreement validates the
formation of an Fe−C bond between this radical and iron(II)
2
2
III
DFT calculations and explains why a single binding curve was
observed (the second EtOH molecule binds more tightly than
to give [Fe (acac) -(CHMeCOOMe)] in the preferred
2
quartet state is calculated to be endergonic by 0.6 kcal/mol
(Supporting Information section 2.J). The O-bound isomer
was calculated to be 4 kcal/mol higher in energy than the C-
bound isomer. Therefore, the acrylate radical should not
benefit from the persistent radical effect at catalytic
the first).
II
These energetics suggest that the speciation of [Fe (acac) ]
2
in pure EtOH solvent (17 M) at 298 K is ca. 99% 6-coordinate
II
II
trans-[Fe (acac) (EtOH) ] and 1% 4-coordinate [Fe (acac) ],
2
2
2
II
with only trace amounts of 5-coordinate [Fe (acac) (EtOH)].
Hence, both the unsolvated [Fe (acac) ] and the alcohol
complex trans-[Fe (acac) (EtOH) ] are readily accessible
concentrations of iron. Additional calculations of the Fe−C
2
II
III
2
bond strength in [Fe (acac) (CHMeCOOMe)] were also
2
II
2
2
conducted with other functionals to validate our computational
method (see Supporting Information section K).
reactive species for the ethanolic reactions to be explored in
the remainder of this study.
Trapping of the Acrylate Radical through the
Previously Proposed Stepwise ET/PT Mechanism. The
acrylate radical, which derives from the attack of the initial
alkyl radical on the acceptor alkene, is a key species in the
catalytic cycle. Its conversion into the product requires one
proton and one electron. The regioselectivity of deuterium
Behavior of the Initial Alkyl Radical. The product of
HAT to the alkene is an alkyl radical. Though this alkyl radical
may add to the acceptor alkene directly, our experience with
OMRP and the previous suggestions of assistance by the
3
persistent radical effect encouraged us to evaluate the
II
6a
possibility of its reversible binding to [Fe (acac) ], which
transfer from EtOD to the product indicates that the proton
2
would give an alkyliron(III) complex (eq 1).
derives from ethanol. In the previously proposed mechanism
for the cross-coupling reactions, the trapping takes place
through a two-step mechanism: initial electron transfer from
the iron(II) species to the acrylate radical to yield an acrylate
anion, which accepts proton transfer from ethanol to generate
the product and the iron(III) ethoxide (this stepwise
mechanism “ET/PT” is shown as (a) in Scheme 6). We
tested this idea using both computations and experiments.
Since mechanism (a) is initiated by electron transfer (ET)
from iron(II) to the acrylate radical, we tested its feasibility by
II
•
III
[
Fe (acac) ] + R ⇋ [Fe (acac) R]
(1)
2
2
We calculate that the formation of the quartet state of the
iron(III) tert-butyl complex is exergonic by 1.5 kcal/mol
(Figure 4 and Supporting Information section 2.H), whereas
the low- and high-spin states are found at higher energy. For
this reason, the formation of the iron(III) alkyl complex can
serve to decrease the concentration of the radical in solution.
Thus, our computations indicate the potential for a weak
determining the thermodynamics of ET from iron(II) to the
II
acrylate radical. Cyclic voltammetry of [Fe (acac) ] in ethanol
2
(
the solvent used in the catalytic reaction) showed an
irreversible oxidation wave at moderate scan rates (≤500
mV/s), with the cathodic peak current being much smaller
than the anodic peak current (Figure 5, black trace). We
attribute the irreversibility to the limited stability of
III
+
[
Fe (acac) ] in solution. The instability of related iron-
2
23
acetylacetonate species has been reported previously, and we
II
have observed that oxidation of [Fe (acac) ] in air gives
2
III
6b
20c
[
Fe (acac) ] with ligand redistribution.
Fortunately,
3
increasing the scan rate to 2 V/s improved chemical
reversibility (ipc ∼ i ), enabling the calculation of a half-
pa
+/0
wave potential (E1/2) of −0.48 V vs Fc (Figure 5, blue
trace). DFT calculations (SI, section 2.L) predicted this
potential to be −0.42 V, which supports this value (and
Figure 4. Reaction coordinates of two tert-butyl radical reactions:
II
(
left) trapping by [Fe (acac) ]; (right) addition to methyl acrylate.
2
E
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 6. Mechanisms for Quenching the Product Radical
examined this concerted PCET process in a truncated
substrate using DFT calculations (Figure 6 and Supporting
Information section 2.M). Approach of the model propionate
•
radical CH CH (COOMe) to iron(II) compound trans-
3
II
[
Fe (acac) (EtOH) ] gave transfer of the proton and electron
2
2
to yield acrylate and the iron(III) compound trans-
III
[
Fe (acac) (OEt)(EtOH)], through transition state TS4
2
⧧
lying only ΔG
= 9.4 kcal/mol above the energy of the
PCET
starting materials. This low barrier corresponds to a very rapid
rate at room temperature and importantly is lower than the
barrier for the ET/PT mechanism above. In the concerted
PCET transition state (TS4), the O−H distance is 1.14 Å,
much longer than 0.97 Å in the starting material, and the
incipient C---H bond distance is 1.456 Å. The long distances,
as well as the low spin density (−0.021 vs 0.000 in the starting
II
Figure 5. Cyclic voltammetry of [Fe (acac) ] (3 mM) in EtOH with
2
III/II
NBu BF (0.1 M) as supporting electrolyte, showing the Fe
couple at −0.48 V. The greater reversibility at faster scan rate suggests
that the iron(III) species is unstable.
redox
4
4
−
complex) and Mulliken charge (+0.38 e , identical to that in
provides another successful validation of the computational
protocol).
II
the starting Fe complex) on the transferring proton are
27
characteristic of a synchronous proton−electron transfer.
II
•
trans‐[Fe (acac) (EtOH) ] + MeCH (COOMe)
2
2
III
+
−
→
trans‐[Fe (acac) (EtOH) ] + MeCH(COOMe)
2
2
(2)
Etotal = Eenolate radical/anion − EFe(III)/Fe(II)
Judging the thermodynamics of ET requires one to compare
this value to the potential for the acrylate radical/anion couple.
•
/−
The redox potential of CH(CH )CO CH , a close relative
3
2
3
of the product radical, has been measured to be −1.04 V vs
+
/0
24
Fc in acetonitrile. With the caveat that this literature value
is in a different solvent, this analysis suggests that ET from
iron(II) to the radical is unfavorable, with E of roughly −0.56
V (uphill by 13 kcal/mol). Our DFT calculations predict that
it is even more uphill (35.0 kcal/mol) in ethanol (Supporting
Information section 2.L). Thus, ET would be expected to
introduce a substantial barrier to the ET/PT two-step
mechanism that potentially keeps it from being kinetically
competent.
Concerted Proton-Coupled Electron Transfer. In
mechanism (b) of Scheme 6, we consider the hypothesis
that there could be inner-sphere electron transfer from iron to
the radical during proton transfer from ethanol. This process
has been termed concerted proton−electron transfer (CPET),
Figure 6. Reaction coordinate for concerted PCET from
II
•
[
Fe (acac) (EtOH) ] to CH CH (COOMe).
2 2 3
25
which is a subset of PCET reactions that are concerted. In
these concerted PCET reactions, the proton and the electron
are transferred in a single elementary step, with a barrier that is
In addition to the productive concerted PCET described
above, we also used DFT to evaluate concerted PCET from
[Fe(acac) (EtOH) ] to the tert-butyl radical, which models a
2
2
2
6
lower than that for either proton transfer or electron transfer.
side reaction that would give the net hydrogenation of the
donor alkene without C−C bond formation. The calculated
transition state energy is 7.3 kcal/mol, which is slightly lower
25b
This situation is termed multisite concerted PCET, because
the proton and electron come from different locations. We
F
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
than that for the productive concerted PCET (9.4 kcal/mol).
Scheme 7. Kinetic Isotope Effect Experiments
•
-
However, the rate law for the side reaction is kPCET,tBu[tBu ]
II
[
(acac) Fe (EtOH) ], while the rate law for the C−C
2
2
•
coupling is k [tBu ][alkene]. Since the concentration of the
CC
trapping alkene is much higher than the Fe concentration, the
rate of the reaction with the alkene would be similar despite
the lower rate constant. Using the Eyring equation, we
7
calculate a rate constant for the CPET side reaction of 3 × 10
−
1
−1
•
-
M
s
at 298 K for a rate law kPCET,tBu[tBu ]
II
[
(acac) Fe (EtOH) ]. The rate constant for the C−C bond
2
2
6 −1 −1
formation, in contrast, would be 4 × 10 M
CC
s
for a rate law
•
k [tBu ][alkene]. Importantly, this model predicts that the
C−C coupling of tert-butyl radical to acceptor alkene should
become less favorable as the reaction progresses due to the
decrease in the relative concentration of trapping alkene.
Monitoring the product distribution in the standard coupling
reaction at 60 °C indicates that indeed the reduction of donor
alkene is a competing side reaction and occurs only later in the
reaction as the acceptor olefin is consumed (Figure S21). This
explains why the reaction yield for cross-coupling is often
enhanced by using superstoichiometric amounts of the
forms a bond to iron(II) to give an intermediate that is
susceptible to protonation. As noted above, a putative
alkyliron(III) intermediate lies at higher energy than the free
radical + iron(II), but it could be a kinetically competent
intermediate if the subsequent protonation occurs with a low
barrier. For the sake of computational efficiency, in addition of
keeping propionate as a model alkyl, MeOH was used as model
alcohol in these calculations (Supporting Information section
2.O). Relaxed PES scans, for both quartet and sextet spin states
and considering up to two MeOH molecules, identified
reasonable pathways leading from this alkyl complex to
products. Protonation of the C-bound enolate (Supporting
Information section 2.P) required a transition state with an
electronic energy more than 20 kcal/mol higher than
6
b
acceptor alkene, because then the concentration of acceptor
alkene does not drop precipitously in the later part of the
reaction.
A second question concerns why the acrylate product radical
undergoes concerted PCET rather than continuing to add to
further acrylate substrate. In this case, we must compare
•
II
•
•
kPCET,R [R ][Fe ] and k [R ][alkene], where R is the acrylate
p
radical and k is the propagation rate constant of the acrylate
p
monomer polymerization. The calculations for the
•
⧧
CH CH (COOMe) model predict ΔG
of 9.4 kcal/
3
PCET,R
mol (see above), while the calculation of the model product
radical addition to methyl acrylate predicts a barrier ΔG
0.9 kcal/mol (Supporting Information section 2.N). Due to
⧧
=
p
•
II
1
CH CH (COOMe), MeOH, and [Fe (acac) ] or led to Fe−
3
2
the skewed relative concentrations, the calculated rates are
again close, and in agreement, Lo et al. observed side products
C homolysis and subsequent concerted PCET through a
transition state indistinguishable from that in mechanism (b).
Therefore, we rule out this mechanistic possibility.
6b
from multiple additions of acceptor alkene. Clearly, there is a
delicate balance between the concerted PCET and alkene
addition rates, and the success of the catalytic cross-coupling
rests on the tuning of these two processes for the two different
radicals.
In contrast, the higher-energy O-bound enolate (Supporting
Information section 2.Q) revealed feasible pathways (Scheme
8). A first possibility consists of the direct proton transfer to
the enolate C atom, generating the product as shown as (c1).
Kinetic Isotope Effects. If PCET to the acrylate radical
either concerted or stepwise with slow PT) were turnover-
This was calculated to be difficult, with the optimized TS lying
II
(
at nearly 20 kcal/mol above [Fe (acac) ], MeOH, and
2
•
limiting, then there would be a kinetic isotope effect between
reactions in EtOH versus EtOD. The kinetic isotope effect for
the overall catalysis was measured by comparing separate
standard catalytic reactions in EtOH and EtOD (Scheme 7,
CH CH (COOMe). The high activation barrier is attributed
3
Scheme 8. Variants of Mechanism (c) Intramolecular
Proton Transfer to O-Bound Ester Enolate
−
5
top). The initial rates at 40 °C were 6.5 ± 0.7 × 10 M/s and
−
5
6
.2 ± 1.2 × 10 M/s, respectively; thus, there is no kinetic
isotope effect. This suggests that PCET is not the turnover-
limiting step in the cycle. We also used a competition
experiment in order to isolate the KIE of the step involving
transfer of the H from ethanol (which we call the product
isotope effect “PIE”). Performing the catalytic reaction in a 1:1
mixture of EtOH and EtOD (Scheme 7, bottom) gave a
product with 21 ± 1% D incorporation, which indicates a PIE
28
of 3.8 ± 0.1. The difference in the KIE and PIE values
indicates that the PCET step is not turnover-limiting, but
rather lies after the turnover-limiting step. This is consistent
with a low-energy transition state for transfer of H to the
propionate radical.
Protonation of Iron-Bound Intermediates. In addition,
we considered mechanism (c), in which the product radical
G
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
to the unfavorable hybridization change of the C atom from
DISCUSSION
■
2
3
sp to sp , with loss of the enolate π delocalization. Direct
proton transfer to the enolate O atom, leading to the ester enol
tautomer of the desired product, is expected to be unfavorable
because it requires a strained 4-membered ring, as shown in
Hydride Species in HAT Alkene Reactions. A number
of stoichiometric hydrogenation and cyclization reactions are
mediated by isolable transition metal hydride species. The
seminal observation of chemically induced dynamic nuclear
polarization (CIDNP) and an inverse kinetic isotope effect
from the reduction of styrene with HMn(CO) established
hydrogen atom transfer (HAT) from a metal hydride to an
alkene and indicated that HAT is reversible.
observation of inverse isotope effects from 2-cyclopropylpro-
pene reacting with HCr(CO) Cp and the observation of
CIDNP from HFe(CO) Cp in diene hydrogenation demon-
(
c2) in Scheme 8.
However, the computed activation barriers drop dramati-
5
cally by including an additional methanol (mechanism (c3) in
29
19a
Scheme 8) as a “proton shuttle” as shown for other systems.
Further
Despite the assistance from additional alcohol molecules, the
II
product of protonation lies 11.0 kcal/mol above [Fe (acac) ],
2
3
•
MeOH and CH CH (COOMe). The pathway would then be
3
2
completed by the exergonic ester enol tautomerization. So, the
protonation pathway requires an intermediate that is slightly
higher in energy than the PCET barrier (see Supporting
Information section 2.Q), but it is close enough for
consideration.
Experimentally Testing Concerted versus Stepwise
PCET to a Model Alkyl Radical. In order to experimentally
test the conditions under which the putative active species can
achieve multisite concerted PCET or stepwise PCET to an
alkyl radical, we used azobis(isobutyronitrile) (AIBN) as a
source of the cyanoisopropyl radical. Transfer of a proton and
strated reversible hydrogen atom transfer from first row
19b,33
transition metal hydrides to alkenes.
Two characteristics
of these literature reactions of isolable hydride complexes were
as follows: (1) HAT was reversible. (2) HAT was the rate-
limiting step. This is presumably related to the fact that these
reactions use stable hydride complexes with relatively high
metal−hydrogen bond energies. For example, the metal-
hydride BDE values of HCr(CO) Cp, HMn(CO) , cis-
3
5
HMn(CO) PPh , and HFe(CO) Cp have been reported to
4
3
2
19b,34
be 62, 68, 69, and 68 kcal/mol, respectively.
Our studies indicate that the unobserved iron(III) hydride
species that are involved in room-temperature catalytic alkene
coupling have amazingly weak M−H bonds that are quite
different than the previously studied systems. The DFT
an electron would lead to isobutyronitrile (Me C(CN)-H), by
2
30
forming a new C−H bond of 92 kcal/mol. In control
II
experiments without [Fe (acac) ], heating AIBN with 20 equiv
2
calculated gas-phase bond dissociation enthalpy of
of EtOH at 80 °C in acetonitrile yields undetectable amount of
III
(
acac) Fe −H at 298 K from the quartet ground state is
2
1
Me C(CN)-H by H NMR spectroscopy, and the main
2
only 17.3 kcal/mol (Supporting Information section 2.R). As a
result of this very weak M−H bond, hydrogen atom addition
from the hydride complex to the alkene is irreversible, which
contrasts with the reversible hydrogen atom additions in the
above systems with M−H bond energies greater than 60 kcal/
product is the radical homocoupling product (Me C(CN)−
2
C(CN)Me ) (Figure S15). In contrast, the same reaction with
2
II
2
equiv of [Fe (acac) ] per AIBN (one per radical) gave an 86
2
±
3% yield of Me C(CN)-H. and substituting ethanol-OD and
2
1
9a−c
ethanol-d result in Me C(CN)-D with 88 ± 2% and 89 ± 2%
mol.
This difference in reversibility was previously
6
2
1
6b
deuterium incorporation (quantified by H NMR, Figures
proposed on the basis of kinetic isotope effect studies and
is supported here by isotopic labeling experiments that rule out
reversibility and by computational studies that show the
hydrogen atom transfer (HAT in Scheme 9) to be
thermodynamically favorable by more than 25 kcal/mol in
S12−S14). Thus, the combination of Fe(acac) and EtOH
2
indeed generates a competent system for PCET.
In order to distinguish whether it is a coordinated acid that
performs PCET, we repeated these experiments with
+
−
Et NH BF , where the acid cannot coordinate to Fe.
3
4
Importantly, the triethylammonium cation is a stronger acid
Scheme 9. Mechanism for Iron-Catalyzed Intermolecular
Cross-Coupling of Alkenes, Supported by Experiments and
Computations
10
31
than ethanol (K is more than 10 higher). Addition of
a
+
−
Et NH BF to a heated acetonitrile solution of AIBN as a
3
4
control experiment gives only 1% of Me C(CN)-H (Figure
2
S22). Repeating this experiment with 1 and 20 equiv of
Fe(acac) gave 25 ± 2% and 54 ± 5% yield of Me C(CN)-H,
2
2
respectively. These yields are lower than the ones observed
with ethanol, indicating that there is a pathway in ethanol that
requires coordination: we attribute this to concerted PCET
from [Fe(acac) (EtOH) ]. However, PCET is still observed
2
2
+
when using the Et NH salt, which indicates that there is a
3
second pathway: we attribute this to a stepwise PCET
pathway. In the present case, this second pathway cannot
correspond to one of the mechanism (c) variants, because the
proton donor cannot coordinate the metal center. However,
there may be an intermolecular variant consisting of proton
+
III
transfer from outer sphere Et NH to the C- or N-bound Fe
3
cyanoisopropyl complex. An intermolecular proton transfer
+
from outer sphere Et NH to an cyanopropyl radical that is N-
3
bound to copper(I) has been documented recently using a
3
2
similar strategy.
H
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
free energy. The barrier for this exergonic HAT is much lower
than the barriers for the previously studied reversible systems
with isolable hydrides, as expected from the Hammond
principle. This low barrier explains the rapid rate of the iron-
catalyzed cross-coupling reaction despite the undetectably low
concentration of the transient hydride species during catalysis.
The observation that the choice of silane and iron alkoxide
drastically influence the catalytic reaction rate, combined with
the agreement between calculated and observed barries of 22−
proposed ET step in the ET/PT mechanism (mechanism (a))
may not be kinetically competent and that other mechanisms
should be considered, in which ET and PT are concerted or
38
involve binding of substrates.
Here, we use “PCET” as a general term for movement of
proton and electron between reagents, “HAT” to indicate
movement of proton and electron from the same site (e.g.,
transfer from the iron(III)-hydride to the donor alkene), and
“concerted PCET” to indicate the concerted transfer of proton
and electron from different sites (e.g., transfer of the electron
from iron(II) and the proton from coordinated alcohol in the
2
4 kcal/mol, indicate that the turnover determining transition
state (TDTS in Scheme 9) is the generation of the key acac-
supported iron(III) hydride complex. Its steady-state concen-
tration would be too small for spectroscopic detection (see
calculation in Supporting Information section 2.S). Because
the transient Fe−H species has not been possible to observe,
we used DFT to computationally define the most likely
geometry, energy, and reactivity. The computations indicate
that the mononuclear hydride is lower in energy than
multimetallic hydrides and is highly reactive. As noted by a
reviewer, a sufficiently reactive polynuclear hydride is a
conceivable intermediate, but bridging hydrides typically
25b
quenching of the product radical).
PCET has long been
discussed in reactions that oxidize organic compounds,
26,39
particularly in bioinorganic chemistry,
but the involvement
of concerted PCET in organic reduction reactions has been
40
studied more recently. Initial attention in the synthetic
community came from the realization that coordination to a
metal can lower the BDFE of the O−H bond, enabling the use
•
of abundant H sources like water and alcohols in place of
•
traditional H sources like tin hydrides. Car
́
denas, Cuerva,
Flowers, and Mayer have shown that H O complexes of
titanium(III) and samarium(II) are particularly effective H
2
35
•
have lower reactivity than terminal hydrides. Interestingly,
all three possible spin states of the mononuclear hydride
complex have very similar energies and geometries, which
suggests that crossover from one state to the other may be
facile. We have suggested that electronic flexibility of this type
may facilitate catalytic reactions, by providing multiple
donors due to their strong reducing ability and formation of
strong M−O bonds, and computations indicate that the proton
41
and electron transfer to the radical are concerted. Chirik and
others have expanded this concept to include N−H bonds as
42
reductants. However, we are not aware of previous studies
where concerted PCET from metal−alcohol complexes has
been linked to catalytic HAT alkene cross-coupling reactions.
In the HAT alkene cross-coupling reactions studied here, the
incorporation of deuterium from EtOD shows that the
alcoholic OH proton is the source of the new C−H bond
36
opportunities for low barriers.
The low barrier for this hydride addition conflicts with our
earlier conclusion that this is the turnover-limiting step in the
6
b
catalytic cycle. Other data also conflict with this idea: here
we show that the overall catalytic reaction rate depends on the
choice of silane and the presence of ethoxide, implicating the
transfer of hydride from the silane to iron as the turnover-
limiting step. This concords with the high-energy hydride
species and is supported by calculations which indicate that the
exchange of hydride and ethoxide has a relatively high
activation barrier.
43
(92 or 90 kcal/mol for the model H−CH(Me)(COOMe)
system according to our DFT calculations (Supporting
Information section 2.P2). This new bond is much weaker
6
a,b
than the O−H bond in ethanol (104 kcal/mol).
However,
our computations indicate that coordination of the ethanol to
Fe (acac) lowers the O−H bond dissociation free energy
II
2
Proton-Coupled Electron Transfer (PCET) from an
Iron(II)-Ethanol Complex. There are a variety of different
HAT alkene reactions, but despite the different products, all
involve HAT from a metal hydride that should result in
formation of a metal center in which the formal oxidation state
from 104 to 74 kcal/mol (see Supporting Information section
2.P3). We experimentally confirmed the PCET ability of the
iron(II) ethanol complex by showing that it reacts with the
cyanoisopropyl radical formed by AIBN, forming a C−H bond
that is 92 kcal/mol. Therefore, the predominant iron(II)
during the catalytic HAT alkene cross-coupling reaction binds
EtOH, enabling the alcohol to become a competent PCET
reductant. We calculate a low-lying transition state for a
synchronous multisite PCET. The normal kinetic isotope effect
of 3.8 for this step (from competition experiments that target
this step of the mechanism) is consistent with concerted
PCET.
30
1b
is reduced by one. In order to reoxidize the metal, it is
natural that chemists have postulated electron transfer from the
1b
reduced metal to the radical. However, alkyl radicals are
37
weak oxidizing agents. Therefore, we explored the feasibility
of several potential pathways.
Earlier work on the HAT alkene cross-coupling reaction
proposed that electron transfer (ET) is followed by a separate
proton transfer (PT) step. However, in the FeIII/II case studied
here, the iron(II) species has too little reducing power to
accomplish the proposed ET step. Of course, the ET step does
not need to be favorable to be part of the catalytic cycle: if the
barrier is small, then an uphill reaction can be followed by a
more exergonic step that consumes the high-energy
intermediate. However, the reorganization energy of the ester
Mayer has noted that two-step PT/ET or ET/PT sequences
tend to occur when there are similar acidities for the two sites
25b
or similar redox potentials for the two sites, respectively.
Here, the coordinated ethanol is expected to be much more
III/II
acidic than the ester product, and the Fe
potential is
substantially more positive than that of the ester-based radical/
anion couple, consistent with the preference for concerted
24
radical to anion was determined to be 26 kcal/mol, which
would add a substantial kinetic barrier to the unfavorable
thermodynamics of the proposed electron transfer reaction.
The experimentally determined overall barrier from our
experiments on the catalytic reaction described in Scheme 5
at 40 °C is 22.8 ± 0.2 kcal/mol. These results suggest that the
PCET. Similar considerations have been made by Car
́
denas
III
and Cuerva for Ti -water complexes, leading to the conclusion
4
1a,c
that proton and electron transfers are concerted.
We
speculate that other HAT alkene reactions catalyzed by
different metals may similarly use concerted PCET as a
“shortcut” that avoids unfavorable ET in a range of HAT
I
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
alkene reactions. For example, formal hydroamination by nitro
concentration of free radicals as well as Fe−O bonds that make
4
k
4a−c
•
compounds, hydrohydrazidation of alkenes,
coupling of
alcohols into good H donors.
44
diazoalkanes to alkenes, and radical cyclization of alkenes and
Our results show many of the characteristics of the metal
center that are required during catalysis, which should be
useful for rational design of catalysts. One is the high reactivity
of the metal-hydride species, which gives rapid and irreversible
transfer of a hydrogen atom to the donor alkene. The use of
weak-field ligands in the acac-iron system gives a very weak
4
5
ketones all involve postulated ET/PT sequences that might
be concerted instead, and thus these mechanisms deserve study
in this context. However, related HAT additions to quinone
methides would give a radical that is much easier to reduce,
and therefore a stepwise ET/PT mechanism could be more
4
6
III
•
likely.
Fe −H bond that transfers H rapidly to the alkene, and the
lack of reversibility prevents chain transfer or β-hydride
elimination. Second, our calculations and experiments have
showed that there are concurrent multiple low-energy
pathways for quenching the radical with alcohol and iron(II).
The large normal PIE is consistent with concerted PCET from
an iron(II)−alcohol complex and is also consistent with iron
binding of the radical to iron(II) to form an O-bound enolate−
iron(III) complex that has formally transferred an electron to
the radical. In the latter case, subsequent irreversible
protonation yields the product.
Consideration of Fe(III) Alkyl and Enolate Intermedi-
ates. The formation of bonds between metals and radicals is
well-established in related OMRP reactions. For example, the
4
-coordinate diaminebis(phenolate)iron(II) system, which is
isoelectronic with Fe(acac) , reversibly traps polystyrene and
2
poly(methyl methacrylate) chains and controls polymer
47
growth by an OMRP mechanism. In the case studied here,
there is weak binding of the initial alkyl radical to iron(II),
which lowers its concentration and prevents extensive
homocoupling. In contrast, the acrylate radical (formed after
reaction with the alkene) is more stable and forms an
accordingly weaker bond to iron(II). Our calculations indicate
that the BDFE of the iron(III)-acrylate radical adduct is
negative; therefore, this radical is not protected through the
PRE and could be preferentially quenched through concerted
We show that the iron(III) hydride is important, but after it
•
loses H , the iron(II) product is also influential. It is a reducing
agent but not in the traditional sense. It gives an alcohol
complex that is an excellent H• donor through PCET.
Alternatively, it can bind radicals to form a variety of species
including an iron(III)-enolate complex which can be
protonated to release the product. The alcohol, which was
previously viewed as merely the proton source for the
intermediate enolate, is now also implicated as a central player
in key mechanistic steps.
II
PCET from [Fe (acac) (EtOH) ]. Thus, the differential
2
2
metal−carbon bond strength in the transient organometallic
complex potentially represents another control strategy for
HAT alkene reactions.
Though the iron(III) complex that comes from iron(II) and
the acrylate product radical has a very low concentration, it is
susceptible to protonation by the alcohol present, but only as
the O-bound enolate. This iron(III) adduct with the radical
proceeds to product with very low barriers using a proton-
shuttle pathway, though it has a relatively high barrier
mandated by the high energy of the enol tautomer of this
We highlight the new insights into the mechanism above in
Scheme 9, which we propose as a more accurate picture of the
mechanism for alkene cross-coupling by Fe-acac catalysts.
These insights into the energetics of various mechanistic steps
may be applicable to various HAT alkene coupling reactions
such as hydropyridylation, hydroamination, hydroazidation,
hydrohydrazination, hydrocyanation, hydration, and hydro-
ester. Our experimental comparison of coordinating (EtOH)
+
1b
and noncoordinating (Et NH ) acids is most consistent with
genation that were recently reviewed and therefore provide
3
the idea that both the protonation of the enolate and the
concerted PCET can occur, and this is compatible with the
similar calculated barriers for these two pathways.
motivation and methods for mechanistic work on other HAT
alkene reactions.
Our results also help to analyze the selectivity of the catalytic
reaction and how it avoids undesired pathways. For example,
the product acrylate radical could add to a second acceptor
alkene to give polymerized products. The two likely pathways
to quench the product radical presented in this work
ASSOCIATED CONTENT
■
*
S
Supporting Information
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/jacs.9b02117.
(
concerted PCET and intermolecular protonolysis of O-
Experimental, spectroscopic, and computational details
bound enolate complex) are in competition with the
polymerization process. Therefore, understanding the factors
governing the rate of the product radical quenching step will be
a key to outcompete the undesired side reaction in the iron-
catalyzed alkene cross-coupling reaction. Future studies will
explore a range of alcohols that could provide higher selectivity
and yield.
(PDF)
Crystallographic information file for [Fe(acac) (μ-
2
OEt)] (CIF)
2
AUTHOR INFORMATION
■
Corresponding Authors
*E-mail: patrick.holland@yale.edu (P.L.H.).
*E-mail: rinaldo.poli@lcc-toulouse.fr (R.P.).
CONCLUSIONS
■
Though HAT alkene reactions are typically viewed as free
radical reactions in which the metal’s main role is radical
generation, our experimental and computational studies on the
iron-catalyzed alkene cross-coupling demonstrate the intimate
involvement of iron species throughout the catalytic cycle.
ORCID
Rinaldo Poli: 0000-0002-5220-2515
Patrick L. Holland: 0000-0002-2883-2031
•
These function not only to donate H from an iron-hydride
Notes
complex but also to form Fe−C bonds that decrease the
The authors declare no competing financial interest.
J
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
2
017, 139, 5998−6007. (p) Desage-El Murr, M.; Fensterbank, L.;
ACKNOWLEDGMENTS
■
Ollivier, C. Iron and Single Electron Transfer: All is in the Ligand. Isr.
J. Chem. 2017, 57, 1160−1169. (q) Inoki, S.; Kato, K.; Isayama, S.;
Mukaiyama, T. A New and Facile Method for the Direct Preparation
of α-Hydroxycarboxylic Acid Esters from α,β-Unsaturated Carboxylic
Acid Esters with Molecular Oxygen and Phenylsilane Catalyzed by
Bis(dipivaloylmethanato)manganese(II) Complex. Chem. Lett. 1990,
19, 1869−1872. (r) Magnus, P.; Payne, A. H.; Waring, M. J.; Scott, D.
A.; Lynch, V. Conversion of α,β-unsaturated ketones into α-hydroxy
P.L.H. thanks the National Science Foundation (CHE-
465017), National Institutes of Health (GM129081) and
the Humboldt Foundation for support. We thank Prof. James
Mayer, Dr. Kazimer Skubi, Dr. Julian Lo, and Prof. Phil Baran
for helpful discussions. R.P. thanks the CNRS (Centre
National de la Recherche Scientifique) for support. This
work used the HPC resources of IDRIS under the allocation
1
́
III
2
016-086343 made by GENCI (Grand Equipement National
ketones using an Mn catalyst, phenylsilane and dioxygen:
de Calcul Intensif) and the resources of the CICT (Centre
Interuniversitaire de Calcul de Toulouse, project CALMIP).
acceleration of conjugate hydride reduction by dioxygen. Tetrahedron
Lett. 2000, 41, 9725−9730.
(
5) Lo, J. C.; Yabe, Y.; Baran, P. S. A Practical and Catalytic
Reductive Olefin Coupling. J. Am. Chem. Soc. 2014, 136, 1304−1307.
(6) (a) Lo, J. C.; Gui, J.; Yabe, Y.; Pan, C.-M.; Baran, P. S.
Functionalized olefin cross-coupling to construct carbon−carbon
bonds. Nature 2014, 516, 343−348. (b) Lo, J. C.; Kim, D.; Pan, C.-
REFERENCES
■
(
1) (a) Olah, G. A.; Prakash, G. K. S. Carbocation Chemistry; Wiley:
Hoboken, NJ, 2004. (b) Crossley, S. W. M.; Obradors, C.; Martinez,
R. M.; Shenvi, R. A. Mn-, Fe-, and Co-Catalyzed Radical
Hydrofunctionalizations of Olefins. Chem. Rev. 2016, 116, 8912−
M.; Edwards, J. T.; Yabe, Y.; Gui, J.; Qin, T.; Gutierrez, S.; Giacoboni,
́
J.; Smith, M. W.; Holland, P. L.; Baran, P. S. Fe-Catalyzed C−C Bond
Construction from Olefins via Radicals. J. Am. Chem. Soc. 2017, 139,
9
(
000.
2) Leggans, E. K.; Barker, T. J.; Duncan, K. K.; Boger, D. L.
2
484−2503. (c) Shen, Q.; Hartwig, J. F. [(CyPF-tBu)PdCl ]: An Air-
2
Iron(III)/NaBH -Mediated Additions to Unactivated Alkenes: Syn-
4
Stable, One-Component, Highly Efficient Catalyst for Amination of
thesis of Novel 20′-Vinblastine Analogues. Org. Lett. 2012, 14, 1428−
Heteroaryl and Aryl Halides. Org. Org. Lett. 2008, 10, 4109−4112.
1
(
431.
(d) George, D. T.; Kuenstner, E. J.; Pronin, S. V. Synthesis of
3) (a) Yan, M.; Lo, J. C.; Edwards, J. T.; Baran, P. S. Radicals:
Emindole SB. Synlett 2016, 28, 12−18. (e) Qi, J.; Zheng, J.; Cui, S.
Fe(III)-Catalyzed Hydroallylation of Unactivated Alkenes with
Morita−Baylis−Hillman Adducts. Org. Lett. 2018, 20, 1355−1358.
Reactive Intermediates with Translational Potential. J. Am. Chem. Soc.
2
016, 138, 12692−12714. (b) Recent evidence for reversible radical
formation: Shevick, S. L.; Obradors, C.; Shenvi, R. A. Mechanistic
Interrogation of Co/Ni-Dual Catalyzed Hydroarylation. J. Am. Chem.
Soc. 2018, 140, 12056−12068.
(
f) Amatov, T.; Pohl, R.; Cisarova, I.; Jahn, U. Sequential Oxidative
̌ ́
and Reductive Radical Cyclization Approach toward Asperparaline C
and Synthesis of Its 8-Oxo Analogue. Org. Lett. 2017, 19, 1152−1155.
(
4) (a) Waser, J.; Carreira, E. M. Convenient Synthesis of
(7) McMillen, D. F.; Golden, D. M. Hydrocarbon Bond Dissociation
Alkylhydrazides by the Cobalt-Catalyzed Hydrohydrazination Re-
action of Olefins and Azodicarboxylates. J. Am. Chem. Soc. 2004, 126,
676−5677. (b) Waser, J.; Carreira, E. M. Catalytic Hydro-
Energies. Annu. Rev. Phys. Chem. 1982, 33, 493−532.
(8) Tokuyasu, T.; Kunikawa, S.; Masuyama, A.; Nojima, M.
5
Co(III)−Alkyl Complex- and Co(III)−Alkylperoxo Complex-Cata-
lyzed Triethylsilylperoxidation of Alkenes with Molecular Oxygen and
Triethylsilane. Org. Lett. 2002, 4, 3595−3598.
hydrazination of a Wide Range of Alkenes with a Simple Mn
Complex. Angew. Chem., Int. Ed. 2004, 43, 4099−4102. (c) Waser, J.;
Gaspar, B.; Nambu, H.; Carreira, E. M. Hydrazines and Azides via the
Metal-Catalyzed Hydrohydrazination and Hydroazidation of Olefins.
J. Am. Chem. Soc. 2006, 128, 11693−11712. (d) Gaspar, B.; Carreira,
E. M. Mild Cobalt-Catalyzed Hydrocyanation of Olefins with Tosyl
Cyanide. Angew. Chem., Int. Ed. 2007, 46, 4519−4522. (e) Gaspar, B.;
Carreira, E. M. Catalytic Hydrochlorination of Unactivated Olefins
with para-Toluenesulfonyl Chloride. Angew. Chem., Int. Ed. 2008, 47,
(9) Munck, E. Aspects of Iron-57 Mossbauer Spectroscopy. In
̈
Physical Methods in Bioinorganic Chemistry; Que, L., Ed.; University
Science Books: New York, 2000; pp 287−319.
(10) (a) Bachmann, W. E.; Wiselogle, F. Y. The Reversible
Dissociation of Pentaarylethanes. J. Org. Chem. 1936, 1, 354−382.
(b) Fischer, H. The Persistent Radical Effect: A Principle for Selective
Radical Reactions and Living Radical Polymerizations. Chem. Rev.
2001, 101, 3581−3610. (c) Studer, A. The Persistent Radical Effect in
Organic Synthesis. Chem. - Eur. J. 2001, 7, 1159−1164.
5
758−5760. (f) Gaspar, B.; Carreira, E. M. Cobalt Catalyzed
Functionalization of Unactivated Alkenes: Regioselective Reductive
C−C Bond Forming Reactions. J. Am. Chem. Soc. 2009, 131, 13214−
3215. (g) Barker, T. J.; Boger, D. L. Fe(III)/NaBH -Mediated Free
Radical Hydrofluorination of Unactivated Alkenes. J. Am. Chem. Soc.
012, 134, 13588−13591. (h) King, S. M.; Ma, X.; Herzon, S. B. A
Method for the Selective Hydrogenation of Alkenyl Halides to Alkyl
1
(11) Daikh, B. E.; Finke, R. G. The persistent radical effect: a
prototype example of extreme, 105 to 1, product selectivity in a free-
4
2
II
radical reaction involving persistent Co [macrocycle] and alkyl free
radicals. J. Am. Chem. Soc. 1992, 114, 2938−2943.
Halides. J. Am. Chem. Soc. 2014, 136, 6884−6887. (i) Crossley, S. W.
́
M.; Barabe, F.; Shenvi, R. A. Simple, Chemoselective, Catalytic Olefin
(12) (a) Poli, R. Relationship between one-electron transition metal
reactivity and radical polymerization processes. Angew. Chem., Int. Ed.
006, 45, 5058−5070. (b) Smith, K. M.; McNeil, W. S.; Abd-El-Aziz,
Isomerization. J. Am. Chem. Soc. 2014, 136, 16788−16791. (j) Iwasaki,
K.; Wan, K. K.; Oppedisano, A.; Crossley, S. W. M.; Shenvi, R. A.
Simple, Chemoselective Hydrogenation with Thermodynamic Stereo-
control. J. Am. Chem. Soc. 2014, 136, 1300−1303. (k) Gui, J.; Pan, C.-
M.; Jin, Y.; Qin, T.; Lo, J. C.; Lee, B. J.; Spergel, S. H.; Mertzman, M.
E.; Pitts, W. J.; La Cruz, T. E.; Schmidt, M. A.; Darvatkar, N.;
Natarajan, S. R.; Baran, P. S. Practical olefin hydroamination with
nitroarenes. Science 2015, 348, 886−891. (l) Dao, H. T.; Li, C.;
Michaudel, Q.; Maxwell, B. D.; Baran, P. S. Hydromethylation of
Unactivated Olefins. J. Am. Chem. Soc. 2015, 137, 8046−8049.
2
A. S. Organometallic-Mediated Radical Polymerization: Developing
Well-Defined Complexes for Reversible Transition Metal-Alkyl Bond
Homolysis. Macromol. Chem. Phys. 2010, 211, 10−16. (c) Hurtgen,
M.; Detrembleur, C.; Jerome, C.; Debuigne, A. Insight into
Organometallic-Mediated Radical Polymerization. Polym. Rev. 2011,
51, 188−213. (d) Poli, R. Radical Coordination Chemistry and Its
Relevance to Metal-Mediated Radical Polymerization. Eur. J. Inorg.
Chem. 2011, 2011, 1513−1530. (e) Allan, L. E. N.; Perry, M. R.;
Shaver, M. P. Organometallic mediated radical polymerization. Prog.
Polym. Sci. 2012, 37, 127−156. (f) Poli, R. Organometallic Mediated
Radical Polymerization. In Reference Module in Materials Science and
Materials Engineering; Elsevier: Amsterdam, 2016.
(
m) Ma, X.; Herzon, S. B. Intermolecular Hydropyridylation of
Unactivated Alkenes. J. Am. Chem. Soc. 2016, 138, 8718−8721.
n) Green, S. A.; Matos, J. L. M.; Yagi, A.; Shenvi, R. A. Branch-
Selective Hydroarylation: Iodoarene−Olefin Cross-Coupling. J. Am.
Chem. Soc. 2016, 138, 12779−12782. (o) Ma, X.; Dang, H.; Rose, J.
A.; Rablen, P.; Herzon, S. B. Hydroheteroarylation of Unactivated
Alkenes Using N-Methoxyheteroarenium Salts. J. Am. Chem. Soc.
(
(13) Obradors, C.; Martinez, R. M.; Shenvi, R. A. Ph(i-PrO)SiH :
2
An Exceptional Reductant for Metal-Catalyzed Hydrogen Atom
Transfers. J. Am. Chem. Soc. 2016, 138, 4962−4971.
K
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
6
b
(
14) We showed earlier that PhSiH /PhSiD gives a kinetic
(24) Bortolamei, N.; Isse, A. A.; Gennaro, A. Estimation of standard
reduction potentials of alkyl radicals involved in atom transfer radical
polymerization. Electrochim. Acta 2010, 55, 8312−8318.
3
3
isotope effect of 1.5, which is too small to have a major influence on
the outcome of this labeling experiment.
15) (a) Salomon, O.; Reiher, M.; Hess, B. A. Assertion and
(
(25) (a) Costentin, C.; Evans, D. H.; Robert, M.; Saveant, J.-M.;
́
validation of the performance of the B3LYP* functional for the first
transition metal row and the G2 test set. J. Chem. Phys. 2002, 117,
Singh, P. S. Electrochemical Approach to Concerted Proton and
Electron Transfers. Reduction of the Water−Superoxide Ion
Complex. J. Am. Chem. Soc. 2005, 127, 12490−12491. (b) Warren,
J. J.; Tronic, T. A.; Mayer, J. M. Thermochemistry of Proton-Coupled
Electron Transfer Reagents and its Implications. Chem. Rev. 2010,
110, 6961−7001.
4
729−4737. (b) Harvey, J.; Aschi, M. Modelling spin-forbidden
reactions: recombination of carbon monoxide with iron tetracarbonyl.
Faraday Discuss. 2003, 124, 129−143. (c) Harvey, J. N.; Poli, R.
Computational study of the spin-forbidden H oxidative addition to
2
(
26) Mayer, J. M. Proton-coupled electron transfer: A reaction
chemist’s view. Annu. Rev. Phys. Chem. 2004, 55, 363−390.
27) Reece, S. Y.; Nocera, D. G. Proton-Coupled Electron Transfer
1
(
6-electron Fe(0) Complexes. Dalton Trans. 2003, 4100−4106.
d) Carreon-Macedo, J.-L.; Harvey, J. N. Do spin state changes matter
in organometallic chemistry? A computational study. J. Am. Chem. Soc.
004, 126, 5789−5797. (e) Besora, M.; Carreon-Macedo, J. L.;
́
(
in Biology: Results from Synergistic Studies in Natural and Model
2
Systems. Annu. Rev. Biochem. 2009, 78, 673−699.
Cowan, A. J.; George, M. W.; Harvey, J. N.; Portius, P.; Ronayne, K.
L.; Sun, X. Z.; Towrie, M. A Combined Theoretical and Experimental
Study on the Role of Spin States in the Chemistry of Fe(CO)5
Photoproducts. J. Am. Chem. Soc. 2009, 131, 3583−3592.
(28) Simmons, E. M.; Hartwig, J. F. On the Interpretation of
Deuterium Kinetic Isotope Effects in C-H Bond Functionalizations by
Transition-Metal Complexes. Angew. Chem., Int. Ed. 2012, 51, 3066−
3
(
072.
(16) Wu, C.-H. S.; Rossman, G. R.; Gray, H. B.; Hammond, G. S.;
29) (a) Sambrano, J. R.; Andres
́
, J.; Gracia, L.; Safont, V. S.;
Schugar, H. J. Chelates of β-diketones. VI. Synthesis and character-
ization of dimeric dialkoxo-bridged iron(III) complexes with
acetylacetone and 2,2,6,6-tetramethylheptane-3,5-dione (HDPM).
Inorg. Chem. 1972, 11, 990−994.
Beltran, A. DFT study of the water-assisted tautomerization process
between hydrated oxide, MO(H O) , and dihydroxide, M(OH) ,
́
+
2+
2
cations (M = V, Nb and Ta). Chem. Phys. Lett. 2004, 384, 56−62.
(b) Hratchian, H. P.; Sonnenberg, J. L.; Hay, P. J.; Martin, R. L.;
(17) (a) Dugan, T. R.; Bill, E.; MacLeod, K. C.; Brennessel, W. W.;
Bursten, B. E.; Schlegel, H. B. Theoretical Investigation of Uranyl
Dihydroxide: Oxo Ligand Exchange, Water Catalysis, and Vibrational
Spectra. J. Phys. Chem. A 2005, 109, 8579−8586. (c) Jee, J.-E.;
Holland, P. L. Synthesis, Spectroscopy, and Hydrogen/Deuterium
Exchange in High-Spin Iron(II) Hydride Complexes. Inorg. Chem.
2
014, 53, 2370−2380. (b) Gieshoff, T. N.; Chakraborty, U.; Villa, M.;
Comas-Vives, A.; Dinoi, C.; Ujaque, G.; van Eldik, R.; Lledos, A.; Poli,
́
Jacobi von Wangelin, A. Alkene Hydrogenations by Soluble Iron
+
R. Nature of Cp*MoO2 in Water and Intramolecular Proton-
Transfer Mechanism by Stopped-Flow Kinetics and Density Func-
tional Theory Calculations. Inorg. Chem. 2007, 46, 4103−4113.
Nanocluster Catalysts. Angew. Chem., Int. Ed. 2017, 56, 3585−3589.
(18) The same trend has been observed for borohydrides; addition
of OR substituents renders them more powerful hydride donors:
Heiden, Z. M.; Lathem, A. P. Establishing the Hydride Donor
Abilities of Main Group Hydrides. Organometallics 2015, 34, 1818−
(d) Hayes, J. M.; Deydier, E.; Ujaque, G.; Lledos, A.; Malacea-
́
Kabbara, R.; Manoury, E.; Vincendeau, S.; Poli, R. Ketone
Hydrogenation with Iridium Complexes with “non N−H” Ligands:
The Key Role of the Strong Base. ACS Catal. 2015, 5, 4368−4376.
1827. Hammond’s postulate suggests that this would lead to a lower
barrier: Alherz, A.; Lim, C.-H.; Hynes, J. T.; Musgrave, C. B.
Predicting Hydride Donor Strength via Quantum Chemical
Calculations of Hydride Transfer Activation Free Energy. J. Phys.
Chem. B 2018, 122, 1278−1288.
(
30) Brocks, J. J.; Beckhaus, H.-D.; Beckwith, A. L. J.; Ruchardt, C.
̈
Estimation of Bond Dissociation Energies and Radical Stabilization
Energies by ESR Spectroscopy. J. Org. Chem. 1998, 63, 1935−1943.
(
31) Kutt, A.; Movchun, V.; Rodima, T.; Dansauer, T.; Rusanov, E.
̈
(19) (a) Sweany, R. L.; Halpern, J. Hydrogenation of alpha-
B.; Leito, I.; Kaljurand, I.; Koppel, J.; Pihl, V.; Koppel, I.; Ovsjannikov,
G.; Toom, L.; Mishima, M.; Medebielle, M.; Lork, E.; Roschenthaler,
methylstyrene by hydridopentacarbonylmanganese. Evidence for a
̈
free-radical mechanism. J. Am. Chem. Soc. 1977, 99, 8335−8337.
G.-V.; Koppel, I. A.; Kolomeitsev, A. A. Pentakis(trifluoromethyl)-
phenyl, a Sterically Crowded and Electron-withdrawing Group:
Synthesis and Acidity of Pentakis(trifluoromethyl)benzene, -toluene,
(b) Bullock, R. M.; Samsel, E. G. Hydrogen atom transfer reactions of
transition-metal hydrides. Kinetics and mechanism of the hydro-
genation of α-cyclopropylstyrene by metal carbonyl hydrides. J. Am.
Chem. Soc. 1990, 112, 6886−6898. (c) Kuo, J. L.; Hartung, J.; Han,
A.; Norton, J. R. Direct Generation of Oxygen-Stabilized Radicals by
H• Transfer from Transition Metal Hydrides. J. Am. Chem. Soc. 2015,
-
phenol, and -aniline. J. Org. Chem. 2008, 73, 2607−2620.
(
32) Thevenin, L.; Fliedel, C.; Fantin, M.; Ribelli, T. G.;
Matyjaszewski, K.; Poli, R. Reductive Termination of Cyanoisopropyl
Radicals by Copper(I) Complexes and Proton Donors: Organo-
metallic Intermediates or Coupled Proton−Electron Transfer?. Inorg.
Chem. [Online early access]. DOI: 10.1021/acs.inorgchem.9b00660.
Published online: April 16, 2019. https://pubs.acs.org/doi/10.1021/
acs.inorgchem.9b00660.
1
37, 1036−1039. (d) Li, G.; Han, A.; Pulling, M. E.; Estes, D. P.;
Norton, J. R. Evidence for Formation of a Co−H Bond from
H O) Co(dmgBF ) under H : Application to Radical Cyclizations.
(
2
2
2
2
2
J. Am. Chem. Soc. 2012, 134, 14662−14665.
(
20) (a) Laugier, J.; Mathieu, J. P. Structure of
(33) Binstead, R. A.; Meyer, T. J. Hydrogen-atom transfer between
bisacetylacetonatoiron(II) dihydrate. Acta Crystallogr., Sect. B: Struct.
Crystallogr. Cryst. Chem. 1975, B31, 631. (b) Buckingham, D. A.;
Gorges, R. C.; Henry, J. T. Reaction of Pyridine with Iron(II)
Acetylacetonate in Benzene Solutions. Aust. J. Chem. 1967, 20, 497−
metal complex ions in solution. J. Am. Chem. Soc. 1987, 109, 3287−
3297.
(34) (a) Tilset, M.; Parker, V. D. Solution homolytic bond
dissociation energies of organotransition-metal hydrides. J. Am.
Chem. Soc. 1989, 111, 6711−6717. (b) Estes, D. P.; Vannucci, A.
K.; Hall, A. R.; Lichtenberger, D. L.; Norton, J. R. Thermodynamics
5
02. (c) Xue, Z.; Daran, J.-C.; Champouret, Y.; Poli, R. Ligand
1
adducts of bis(acetylacetonato)iron(II): a H NMR study. Inorg.
Chem. 2011, 50, 11543−11551.
5
of the Metal−Hydrogen Bonds in (η -C H )M(CO) H (M = Fe, Ru,
5
5
2
(21) Curran, D. P.; Trost, B. M. Radical Addition Reactions. Compr.
Os). Organometallics 2011, 30, 3444−3447.
Org. Synth. 1991, 4, 715−777.
22) Giese, B. Formation of CC Bonds by Addition of Free Radicals
to Alkenes. Angew. Chem., Int. Ed. Engl. 1983, 22, 753−764.
23) Murray, R. W.; Hiller, L. K. Supporting electrolyte effects in
(35) (a) Smith, J. M.; Lachicotte, R. J.; Holland, P. L. N = N Bond
Cleavage by a Low-Coordinate Iron(II) Hydride Complex. J. Am.
Chem. Soc. 2003, 125, 15752−15753. (b) Barton, B. E.; Rauchfuss, T.
B. Terminal hydride in [FeFe]-hydrogenase model has lower potential
(
(
nonaqueous electrochemistry. Coordinative relaxation reactions of
for H production than the isomeric bridging hydride. Inorg. Chem.
2
reduced metal acetylacetonates in acetonitrile. Anal. Chem. 1967, 39,
2008, 47, 2261−2263. (c) Schilter, D.; Camara, J. M.; Huynh, M. T.;
Hammes-Schiffer, S.; Rauchfuss, T. B. Hydrogenase Enzymes and
1221−1229.
L
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Their Synthetic Models: The Role of Metal Hydrides. Chem. Rev.
(45) Saladrigas, M.; Bosch, C.; Saborit, G. V.; Bonjoch, J.; Bradshaw,
B. Radical Cyclization of Alkene-Tethered Ketones Initiated by
Hydrogen-Atom Transfer. Angew. Chem., Int. Ed. 2018, 57, 182−186.
(46) Shen, Y.; Qi, J.; Mao, Z.; Cui, S. Fe-Catalyzed Hydroalkylation
of Olefins with para-Quinone Methides. Org. Lett. 2016, 18, 2722−
2725.
(47) (a) Allan, L. E. N.; MacDonald, J. P.; Nichol, G. S.; Shaver, M.
P. Single Component Iron Catalysts for Atom Transfer and
Organometallic Mediated Radical Polymerizations: Mechanistic
Studies and Reaction Scope. Macromolecules 2014, 47, 1249−1257.
(b) Poli, R.; Shaver, M. P. Atom Transfer Radical Polymerization
(ATRP) and Organometallic Mediated Radical Polymerization
(OMRP) of Styrene Mediated by Diaminobis(phenolato)iron(II)
Complexes: A DFT Study. Inorg. Chem. 2014, 53, 7580−7590.
2
(
016, 116, 8693−8749.
36) (a) Poli, R. Open shell organometallics: a general analysis of
their electronic structure and reactivity. J. Organomet. Chem. 2004,
89, 4291−4304. (b) Poli, R. Spin Crossover Reactivity. In
6
Comprehensive Inorganic Chemistry II, Reedijk, J.; Poeppelmeier, K.
Eds.; Elsevier: Amsterdam, 2013; Vol. 9, pp 481−500. (c) Holland, P.
L. Distinctive Reaction Pathways at Base Metals in High-Spin
Organometallic Catalysts. Acc. Chem. Res. 2015, 48, 1696−1702.
(
37) (a) Rao, P. S.; Hayon, E. Redox potentials of free radicals. I.
Simple organic radicals. J. Am. Chem. Soc. 1974, 96, 1287−1294.
b) Fu, Y.; Liu, L.; Yu, H.-Z.; Wang, Y.-M.; Guo, Q.-X. Quantum-
(
Chemical Predictions of Absolute Standard Redox Potentials of
Diverse Organic Molecules and Free Radicals in Acetonitrile. J. Am.
Chem. Soc. 2005, 127, 7227−7234.
(38) A piece of evidence that has been proposed in favor of the ET/
PT mechanism is the ability to perform three-component couplings
with aldehydes, which are more easily rationalized through the
6b
formation of a carbanion than a radical. Similarly, Cui has reported
HAT alkene reactions in which acetyl or bromide displacement is
6e
proposed to proceed from an intermediate carbanion. However, the
formation of these products can be explained without formation of an
enolate. One possibility is that the electrophiles can coordinate to
iron(II), activating them toward attack by a radical. This would give
inner-sphere electron transfer that is concerted with C−C bond
formation and iron abstraction of the leaving group.
(39) Waidmann, C. R.; Miller, A. J. M.; Ng, C.-W. A.; Scheuermann,
M. L.; Porter, T. R.; Tronic, T. A.; Mayer, J. M. Using combinations
of oxidants and bases as PCET reactants: thermochemical and
practical considerations. Energy Environ. Sci. 2012, 5, 7771−7780.
(40) (a) Gentry, E. C.; Knowles, R. R. Synthetic Applications of
Proton-Coupled Electron Transfer. Acc. Chem. Res. 2016, 49, 1546−
1556. (b) Chciuk, T. V.; Anderson, W. R.; Flowers, R. A. Interplay
between Substrate and Proton Donor Coordination in Reductions of
Carbonyls by SmI −Water Through Proton-Coupled Electron-
2
Transfer. J. Am. Chem. Soc. 2018, 140, 15342−15352.
(
41) (a) Paradas, M.; Campan
̃
a, A. G.; Jimen
́
ez, T.; Robles, R.;
Oltra, J. E.; Bunuel, E.; Justicia, J.; Cardenas, D. J.; Cuerva, J. M.
̃
́
Understanding the Exceptional Hydrogen-Atom Donor Character-
istics of Water in Ti -Mediated Free-Radical Chemistry. J. Am. Chem.
III
Soc. 2010, 132, 12748−12756. (b) Kolmar, S. S.; Mayer, J. M.
SmI (H O) Reduction of Electron Rich Enamines by Proton-
2
2
n
Coupled Electron Transfer. J. Am. Chem. Soc. 2017, 139, 10687−
1
0692. (c) Gansauer, A.; Behlendorf, M.; Cangonul, A.; Kube, C.;
̈
̈
̈
2
Cuerva, J. M.; Friedrich, J.; van Gastel, M. H O Activation for
Hydrogen-Atom Transfer: Correct Structures and Revised Mecha-
nisms. Angew. Chem., Int. Ed. 2012, 51, 3266−3270. (d) Chciuk, T.
V.; Anderson, W. R.; Flowers, R. A. High-Affinity Proton Donors
Promote Proton-Coupled Electron Transfer by Samarium Diiodide.
Angew. Chem., Int. Ed. 2016, 55, 6033−6036. (e) Chciuk, T. V.;
Flowers, R. A. Proton-Coupled Electron Transfer in the Reduction of
Arenes by SmI −Water Complexes. J. Am. Chem. Soc. 2015, 137,
2
1
1526−11531. (f) Chciuk, T. V.; Anderson, W. R.; Flowers, R. A.
Proton-Coupled Electron Transfer in the Reduction of Carbonyls by
Samarium Diiodide−Water Complexes. J. Am. Chem. Soc. 2016, 138,
8
(
738−8741.
42) (a) Bezdek, M. J.; Guo, S.; Chirik, P. J. Coordination-induced
weakening of ammonia, water, and hydrazine X−H bonds in a
molybdenum complex. Science 2016, 354, 730. (b) Tarantino, K. T.;
Miller, D. C.; Callon, T. A.; Knowles, R. R. Bond-Weakening
Catalysis: Conjugate Aminations Enabled by the Soft Homolysis of
Strong N−H Bonds. J. Am. Chem. Soc. 2015, 137, 6440−6443.
(43) Oyeyemi, V. B.; Dieterich, J. M.; Krisiloff, D. B.; Tan, T.;
Carter, E. A. Bond Dissociation Energies of C10 and C18 Methyl Esters
from Local Multireference Averaged-Coupled Pair Functional Theory.
J. Phys. Chem. A 2015, 119, 3429−3439.
(44) Zheng, J.; Qi, J.; Cui, S. Fe-Catalyzed Olefin Hydroamination
with Diazo Compounds for Hydrazone Synthesis. Org. Lett. 2016, 18,
128−131.
M
DOI: 10.1021/jacs.9b02117
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX