Catalytic
Reduction
of Cyanide
to Ammonia
and Methane
at a
Mononuclear
Fe Site
Christian
M. Johansen
and Jonas
C. Peters
*
Cite This:
J. Am. Chem.
Soc.
2024,
146, 5343−5354
Read
Online
ACCESS
Metrics
& More
Article
Recommendations
*
sı
Supporting
Information
ABSTRACT:
Nitrogenase
enzymes
catalyze
nitrogen
reduction
(N
2
R) to
ammonia
and also the reduction
of non-native
substrates,
including
the 7H
+
/
6e
−
reduction
of cyanide
to CH
4
and NH
3
. CN
−
and N
2
are isoelectronic,
and
it is hence
fascinating
to compare
the mechanisms
of synthetic
Fe catalysts
capable
of both
CN
−
and
N
2
reduction.
Here,
we describe
the catalytic
reduction
of CN
−
to NH
3
and CH
4
by a highly
selective
(P
3
Si
)Fe(CN)
catalyst
(P
3
Si
represents
a tris(phosphine)silyl
ligand).
Catalysis
is driven
in the
presence
of excess
acid
([Ph
2
NH
2
]OTf)
and reductant
((C
6
H
6
)
2
Cr),
with
turnover
as high
as 73 demonstrated.
This
catalyst
system
is also modestly
competent
for N
2
R and structurally
related
to other
tris(phosphine)Fe-based
N
2
R catalysts.
The choice
of catalyst
and reductant
is important
to observe
high
yields.
Mechanistic
studies
elucidate
several
intermediates
of CN
−
reduction,
including
iron
isocyanides
(P
3
Si
FeCNH
+/0
) and
terminal
iron aminocarbynes
(P
3
Si
FeCNH
2
+/0
). Aminocarbynes
are isoelectronic
to iron hydrazidos
(Fe
�
N
−
NH
2
+/0
), which
have
been
invoked
as selectivity-determining
intermediates
of N
2
R (NH
3
versus
N
2
H
4
products).
For the present
CN
−
reduction
catalysis,
reduction
of aminocarbyne
P
3
Si
FeCNH
2
+
is proposed
to be rate but not selectivity
contributing.
Instead,
by comparison
with
the
reactivity
of a methylated
aminocarbyne
analogue
(P
3
Si
FeCNMe
2
), and associated
computational
studies,
formation
of a Fischer
carbene
(P
3
Si
FeC(H)(NH
2
)
+
) intermediate
that is on path
for either
CH
4
and NH
3
(6 e
−
) or CH
3
NH
2
(4 e
−
) products
is proposed.
From
this carbene
intermediate,
pathways
to the observed
CH
4
and NH
3
products
(distinct
from
CH
3
NH
2
formation)
are
considered
to compare
and contrast
the (likely)
mechanism/s
of CN
−
and N
2
reduction.
■
INTRODUCTION
Nitrogenases
catalyze
nitrogen
reduction
to ammonia
(N
2
R) as
well as the reductive
protonation
of non-native
substrates,
1
−
4
including
cyanide
(CN
−
).
5
−
13
These
are mechanistically
fascinating
bioorganometallic
transformations
which,
for the
case
of CN
−
(and
CO/CO
2
as well),
may
involve
metal-to-
carbon
intermediates
such
as alkyls,
carbenes,
and carbynes/
carbides
that are conceptually
related
to posited
intermediates
of N
2
R (e.g.,
NNH,
NNH
2
, NH).
Whereas
substantial
attention
from
the synthetic
community
has
been
directed
toward
functional
N
2
R models
with
associated
mechanistic
studies,
14
−
16
there
has
been
only
limited
attention
paid
to catalytic
cyanide
reduction
by
comparison.
17
−
22
Given
potential
mechanistic
parallels
be-
tween
catalytic
N
2
and CN
−
reduction,
including
an isolobal
relationship
between
aminocarbynes
(e.g.,
M
�
CNR
2
)
23
−
26
and their
hydrazido
(M
�
NNR
2
) counterparts,
27
−
29
mecha-
nistically
well-defined
CN
−
reduction
catalysts
present
an
attractive
target
for further
study.
In contrast
to terminal
hydrazido
systems,
the reductive
protonation
of terminal
carbynes
to liberate
products
(e.g.,
CH
4
/NH
3
) has rarely
been
observed.
25,30
−
33
Indeed,
catalytic
transformations
involving
bona
fide
carbyne
intermediates,
outside
of the scope
of
metathesis
reactions,
34,35
are essentially
without
precedent.
Toward
these
objectives,
our lab reported
in 2016
a single-
site
iron
model
system
capable
of mediating
the (sub)-
stoichiometric
reductive
protonation
of CN
−
to CH
4
and
NH
3
.
25
We also characterized
a number
of species
as plausible
intermediates
of the overall
transformation,
most
notably
the
carbyne
complex
(P
3
Si
)Fe(CNH
2
)
+
(P
3
Si
represents
a tris-
(phosphino)silyl
ligand;
Figure
1, top).
25
The
product
distribution
observed
mimics
that of ATP-dependent
cyanide
reduction
by nitrogenases
(Figure
1, middle),
where
the major
observed
products
under
most
conditions
studied
are methane
and ammonia
(6 e
−
reduction);
methylamine
(H
3
CNH
2
; 4 e
−
reduction)
and methylenimine
(H
2
C
�
NH;
2 e
−
reduction)
can also
be observed
as minor
products,
along
with
trace
ethane
and ethylene.
6
−
9
Several
synthetic
Fe
−
S
clusters
have
also
been
shown
to
catalyze
cyanide
reduction
and
exhibit
substantially
higher
selectivities
for C
−
C
coupled
products
than
has been
observed
Received:
November
7, 2023
Revised:
January
25, 2024
Accepted:
January
26, 2024
Published:
February
16,
2024
Article
pubs.acs.org/JACS
© 2024
The Authors.
Published
by
American
Chemical
Society
5343
https://doi.org/10.1021/jacs.3c12395
J. Am. Chem.
Soc.
2024,
146, 5343
−
5354
This article is licensed under CC-BY 4.0
with
nitrogenase
enzymes
as the
catalysts
(Figure
1,
middle).
10
−
12,20
−
22
Catalytically
relevant
species
in trans-
formations
employing
such
clusters
as precatalysts
are ill-
defined,
and to date,
associated
mechanistic
information
has
not been
forthcoming.
Against
this
backdrop,
we have
sought
conditions
for
catalytic
cyanide
reduction
via our
well-defined
(P
3
Si
)Fe
system,
ideally
manifesting
product
distributions
akin
to
nitrogenases
(chiefly
favoring
the C
1
products
CH
4
and
CH
3
NH
2
) and amenable
to mechanistic
scrutiny.
This
study
presents
our findings
(Figure
1, bottom).
Guided
by measured
and
estimated
thermochemical
parameters
(Figure
2a),
36
we show
herein
that
the iron
complex
(P
3
Si
)Fe(CN)
(abbreviated
as [FeCN])
efficiently
catalyzes
cyanide
reduction
in the presence
of acids
and
reductants.
By employing
a combination
of synthetic,
57
Fe
Mo
̈
ssbauer,
optical,
and
theoretical
studies,
we outline
a
mechanistic
scheme
for the catalytic
cycle,
which
can be
juxtaposed
with
that of catalytic
nitrogen
fixation
mediated
by
analogous
iron complexes.
■
RESULTS
Canvassing
Conditions
for More
Efficient
[FeCN]
Reduction.
To target
the catalytic
reduction
of CN
−
, we
sought
conditions
for the proton-coupled
reduction
of [FeCN]
to produce
NH
3
/CH
4
(or CH
3
NH
2
) and an [Fe]
byproduct
that might
re-enter
a catalytic
cycle.
In our original
report,
25
we
described
the proton-coupled
reduction
of [FeCN]
using
excess
[
2,5
‑
Cl
PhNH
3
]OTf
and
Cp
*
2
Co
(p
K
a
4.5
for
[
2,5
‑
Cl
PhNH
3
]OTf
in THF;
all p
K
a
’s reported
in THF;
36
E
°
=
−
1.9
V for Cp
*
2
Co; all redox
potentials
are reported
in THF
and referenced
to Fc
+/0
).
37
Such
reaction
mixtures
invariably
afforded
low yields
of NH
3
/CH
4
(Figure
1a) despite
being
effective
for catalytic
N
2
R.
38
Curiously,
in our original
study,
we had observed
that the
cationic
aminocarbyne,
[FeCNH
2
]OTf,
prepared
via double
protonation
of [FeCN][Na(12-c-4)
2
],
25
decays
upon
warming
to liberate
0.09
equiv
NH
3
/Fe and 0.07
CH
4
/Fe (Figure
2),
with
[FeCNH]OTf
and [FeOTf]
as the major
Fe products.
This
NH
3
yield
represents
∼
50%
of that theoretically
possible
for a disproportionation
reaction
assuming
a stoichiometry
of
five
equivalents
of [FeCNH
2
]OTf
providing
five
H atom
equivalents
([FeCNH
2
]
+
→
[FeCNH]
+
+ H
+
/e
−
) to reduce
one equivalent
[FeCNH
2
]OTf
to NH
3
and CH
4
(eq 1).
[
]
[
]
+
[
] +
+
6
FeCNH
OTf
5
FeCNH
OTf
FeOTf
NH
CH
2
3
4
(1)
Based
on thermochemical
data
(Figure
2a),
36
removal
of a
H
+
/e
−
pair from
[FeCNH
2
]
+
is equivalent
to removal
of 1H
+
/
1e
−
from
an acid/reductant
pair with
p
K
a
∼
7 and
E
°
∼ −
1.3
V. Reagents
suiting
these
values
would
be significantly
milder
than
[
2,5
‑
Cl
PhNH
3
]OTf
and Cp
*
2
Co. Hence,
once
[FeCNH
2
]
+
is formed
in situ
via reductive
protonation
of [FeCN]
(p
K
a
=
5.6),
comparatively
mild
reagents
should
drive
net CN
−
reduction.
Because
[FeCN]
can be converted
to [FeCNH
2
]-
OTf
with
a reductant
strength
of
E
°
≈
−
1.3
V, we deduced
that the 7H
+
/6e
−
reduction
of [FeCN]
should
be accessible
with
reductants
at
E
°
≈
−
1.3
V.
Gratifyingly,
[FeCN]
was
stirred
with
(C
6
H
6
)
2
Cr (
E
°
=
−
1.2
V, Figure
S23)
39
and [Ph
2
NH
2
]OTf
(p
K
a
3.2)
40
in Et
2
O
at
−
78
°
C and then
the reaction
mixture
was allowed
to warm
to RT overnight
yielding
0.75
equiv
NH
3
/Fe (75%
yield
per
reductant)
with
[FeOTf]
as the major
Fe product
(Figure
2b
and Figure
S9).
Moreover,
it was established
that
[FeOTf]
reacts
cleanly
with
excess
[TBA][CN]
to reform
[FeCN],
setting
the stage
for catalysis
(Figure
S10).
Figure
1.
Summary
of prior
studies
on stoichiometric
and catalytic
cyanide
reductions
mediated
by iron complexes
as a context
for this
study.
Figure
2.
(A) Relevant,
previously
measured,
thermochemical
data
(values
in THF
at 25
°
C;
E
°
in referenced
to Fc
+/0
). (B) Exploring
new conditions
for the reductive
protonation
of [FeCN].
Journal
of the American
Chemical
Society
pubs.acs.org/JACS
Article
https://doi.org/10.1021/jacs.3c12395
J. Am. Chem.
Soc.
2024,
146, 5343
−
5354
5344
Catalytic
CN
−
Reduction.
Thus,
using
[FeCN]
(0.72
mM)
as a precatalyst,
in a reaction
mixture
containing
140
equiv
[TBA][CN]
(100
mM),
480 equiv
[Ph
2
NH
2
]OTf,
and
360 equiv
(C
6
H
6
)
2
Cr in Et
2
O at 25
°
C, yielded
28
±
5 equiv
NH
3
/Fe after
80 min (Figure
2c and Table
1, entry
1).
Using
these
same
catalytic
conditions,
we also analyzed
the
gaseous
products.
CH
4
was observed
as the major
reduced
carbon
product,
with
a yield
of 25
±
4 equiv
of CH
4
/Fe,
consistent
with
a net 7H
+
/6e
−
reduction
of CN
−
(eq 2; yield
based
on consumed
(C
6
H
6
)
2
Cr is 47
±
8%).
+
+
+
+
CN
7H
6e
CH
N
H
4
3
(2)
Under
these
conditions,
trace
C
2
products
ethylene
and
ethane
were
also
identified
(0.4
equiv
C
2
H
4
and 0.3 equiv
C
2
H
6
per Fe).
These
products
correspond
to 10H
+
/8e
−
and
12H
+
/10e
−
reductions
of CN
−
. Combined,
these
C
2
evolving
reactions
accounted
for less
than
2% of the consumed
reductant.
Hence,
the [FeCN]
catalyst
is nearly
quantitatively
selective
for C
1
products,
as is observed
via the nitrogenase
enzyme.
6,7
H
2
accounts
for most
of the remaining
reducing
equivalents
(yield
based
on (C
6
H
6
)
2
Cr: 29
±
11%).
Neither
CH
3
NH
2
(4e
−
product)
or CH
2
NH
(2e
−
) was
detected,
regardless
of initial
temperature,
using
[FeCN]
as a catalyst.
Curiously,
whereas
synthetic
iron
catalysts
for N
2
R have
shown
highest
efficiency
at low temperatures
due to mitigated
HER
(hydrogen
evolution
reaction)
and entropically
favored
N
2
binding,
41
no such
advantage
is observed
for catalytic
cyanide
reduction
by [FeCN]
(entries
2
−
4).
Instead,
back-
ground
HER
via the combination
of this reductant
and acid is
comparatively
slow
(
vide
infra
). Also,
CN
−
binds
favorably
to
[Fe(II)]
at RT.
For reactions
started
at
−
78
°
C, catalytic
turnover
is slow,
reflecting
a slow
OTf
−
for CN
−
metathesis
step needed
to turn
the system
over
(
vide
infra
); most
of the
observed
catalysis
occurs
as the reaction
is warmed.
For a
catalytic
reaction
run at
−
20
°
C and quenched
after
20 min,
1.7 equiv
of NH
3
was detected,
demonstrating
that catalytic
turnover
occurs
at this temperature
but is relatively
slow.
Increasing
the catalyst
loading
to 2.9 mM (entry
5) modestly
increased
the NH
3
yield
relative
to reductant
present
(65
±
1%).
Lowering
the catalyst
loading
(0.15
mM;
entry
6)
improved
the TON
for produced
NH
3
(73
±
4 equiv)
but led
to a corresponding
drop
in yield
per (C
6
H
6
)
2
Cr (24
±
1%).
A catalyst-free
reaction
yielded
no detectable
NH
3
, CH
4
, or
other
gaseous
carbon
products
(entry
7). This
conclusion
is
further
supported
by experiments
with
[TBA][
13
CN]
as the
cyanide
source.
13
C NMR
spectroscopy
of catalytic
runs
using
[TBA][
13
CN]
confirmed
the
formation
of
13
CH
4
and
consumption
of
13
CN
−
(Figures
S5 and S6).
By contrast,
a
corresponding
catalyst-free
reaction
(under
otherwise
identical
conditions)
showed
negligible
consumption
of
13
CN
−
and no
observable
13
CH
4
. These
observations
collectively
establish
that the Fe catalyst
is required
for consumption
of substrate
and responsible
for the NH
3
and CH
4
products.
Catalysis
run in the absence
of [TBA][CN]
produced
0.7
equiv
of NH
3
, with
[FeC
15
N] used
to demonstrate
that this
NH
3
arose
solely
from
precatalyst
reduction
and
not N
2
R
(entry
8, Figure
S4).
The nature
of the phosphine-ligated
iron catalyst
appears
to
be critical.
FeCl
2
instead
of [FeCN]
produced
only 0.3 equiv
of
NH
3
under
the standard
conditions
(entry
9), and
CrCl
2
instead
of [FeCN]
produced
no detectable
NH
3
(entry
10).
The
tris(phosphino)iron
complexes
(P
3
B
)Fe[BAr
F
4
]
16
and
(PhBP
iPr
3
)FeBr
42
(P
3
B
represents
a trisphosphine
borane
ligand;
PhBP
iPr
3
represents
a trisphosphine
borate
ligand)
showed
very
moderate
activity
as (pre)catalysts
compared
to
[FeCN]
(entries
11
−
12).
Curiously,
for (PhBP
iPr
3
)FeBr,
a
small
amount
of methylamine,
CH
3
NH
2
(0.6
±
0.2),
was
detected
as a product.
These
iron
phosphine
precatalysts
produce
CH
4
as the major
hydrocarbon
product
but with
a
lower
selectivity.
The ratio
of C
2
/C
1
products
produced
is 0.16
and 0.11
for (P
3
B
)Fe[BAr
F
4
] and (PhBP
iPr
3
)FeBr,
respectively
(Table
S3),
compared
to 0.02
for [FeCN].
While
(C
6
H
6
)
2
Cr is the favored
reductant
for CN
−
reduction,
other
reductants
including
Cp
2
Co (
E
°
=
−
1.3
V,
Figure
S25;
entry
13) and Cp
*
2
Cr (
E
°
=
−
1.5
V, Figure
S27;
entry
14) were
modestly
competent.
The low yields
for these
reductants
do not appear
to correlate
with
the reduction
potential
of the chemical
reductant.
Instead,
we attribute
the
strong
attenuation
in yield
to enhanced
background
HER.
Accordingly,
we find that the rate of reaction
of each
reductant
independently
with
[Ph
2
NH
2
]OTf
(to
produce
H
2
), as
measured
by cyclic
voltammetry,
inversely
correlates
with
the
NH
3
TON
observed
in a catalytic
run when
CN
−
is present
under
the standard
conditions
(see
Section
S8.2
for details).
To summarize,
we have
established
a highly
selective
catalytic
system
for NH
3
and CH
4
production
via reductive
protonation
of CN
−
; the choice
of catalyst
([FeCN])
and
reductant
((C
6
H
6
)
2
Cr) is crucial
for observing
high
turnover
and significant
yields.
Table 1. Results
for the Catalytic
Reduction
of CN
−
to
Ammonia
entry
change
from
standard
conditions
NH
3
(equiv/Fe)
CH
4
(equiv/Fe)
yield
per
reductant
(%)
a
1
none
28
±
5
25
±
4
47
±
8
2
−
78
°
C
→
25
°
C
33
±
6
33
±
3
55
±
10
3
−
20
°
C
→
25
°
C
35
±
8
58
±
13
4
0
°
C
→
25
°
C
26
±
2
43
±
3
5
2.9 mM
[FeCN]
b
9.7
±
0.2
65
±
1
6
0.15
mM
[FeCN]
b
73
±
4
24
±
1
7
no [FeCN]
b
<0.4
<0.3
<1
8
no [TBA][CN]
b
0.7
1.2
9
8.0 mM
FeCl
2
as cat.
b
0.3
5.5
10
8.0 mM
CrCl
2
as cat.
b
<0.05
<1
11
2.9 mM
(PhB
iPr
3
)
FeBr
as cat.
b
1.4
±
0.7
1.0
±
0.1
12
±
3
c
12
2.9 mM
(P
3
B
)
Fe[BAr
F
4
] as cat.
b
2.3
±
0.3
1.6
15.6
±
0.6
13
Cp
2
Co instead
of
(C
6
H
6
)
2
Cr
2.8
±
0.8
12
±
4
14
Cp
*
2
Cr instead
of
(C
6
H
6
)
2
Cr
13.5
±
3
32
±
6
15
[FeOTf]
as cat.
32
53
16
reloaded
catalysis
d
4.1
±
1.0
6.8
±
1.7
a
Yields
assume
net 6 e
−
reduction
to generate
NH
3
.
b
Catalysis
initiated
at
−
78
°
C and then
allowed
to warm
gradually
to 25
°
C
(total
reaction
time
of 12 h).
c
This
yield
includes
0.6
±
0.2 equiv
CH
3
NH
2
.
d
After
80 min of catalysis
under
standard
conditions
(entry
1), soluble
Fe species
were
extracted
into Et
2
O and then
re-exposed
to
the standard
catalytic
conditions.
Journal
of the American
Chemical
Society
pubs.acs.org/JACS
Article
https://doi.org/10.1021/jacs.3c12395
J. Am. Chem.
Soc.
2024,
146, 5343
−
5354
5345
Mechanistic
Studies.
Scheme
1 provides
a working
outline
for the catalytic
CN
−
reduction
cycle
starting
from
[FeCN],
emphasizing
the early
intermediates
of the cycle.
To
guide
the following
discussion,
summary
remarks
concerning
a
plausible
pathway
are as follows:
[FeCN]
is first protonated
(step
a) to form
independently
characterized
[FeCNH]
+
,
25
which
is then
reduced
(to [FeCNH];
step b) and protonated
(step
c) to afford
the independently
characterized
amino-
carbyne,
[FeCNH
2
]
+
.
25
[FeCNH
2
]
+
is in redox
equilibrium
with
[FeCNH
2
] in the presence
of (C
6
H
6
)
2
Cr (step
d). These
carbyne
intermediates
are suggested
to be rate-contributing
to
overall
CN
−
reduction
(see below).
Along
the ET-PT
pathway,
[FeCNH
2
] can be protonated
to form
a posited
carbene
intermediate
(step
e), [FeC(H)(NH
2
)]
+
. This
carbene
is
modeled
via the independent
generation
of its methylated
analogue,
[FeC(H)(NMe
2
)]
+
, via the
protonation
of
[FeCNMe
2
] (see below).
This
observation
and computational
evidence
each
lends
support
to C
−
H
bond
formation
to
produce
[FeC(H)(NH
2
)]
+
during
catalysis.
A direct
PCET
pathway
from
[FeCNH
2
]
+
to [FeC(H)(NH
2
)]
+
is also
plausible
(step
f). Finally,
a series
of downstream
(as yet
undefined),
facile
reductive
protonation
steps
of [FeC(H)-
(NH
2
)]
+
are proposed
to release
NH
3
and CH
4
along
with
[FeOTf]
(step
g); the latter
is returned
to [FeCN]
via
metathesis
with
[TBA][CN],
a step (step
(h))
that is turnover
limiting.
Probing
Catalyst
Resting
State
and Deactivation.
To
probe
speciation
during
catalysis,
we prepared
[
57
FeCN]
to
facilitate
monitoring
the catalysis
by
57
Fe Mo
̈
ssbauer
spectros-
copy
via low-temperature
quenching
of catalytic
runs
initiated
at 25
°
C. Related
studies
proved
insightful
for N
2
R catalysis
by
related
Fe systems.
40,43
Freeze-quenching
(77 K) the catalysis
after
1 min at 25
°
C,
we found
[
57
FeOTf]
as the sole
iron
species
present
(see
Section
S6.2).
44
This
result
points
to [FeOTf]
as the catalyst
resting
state
with
metathesis
step (h) being
turnover
limiting.
Consistent
with
this observation,
[FeOTf]
performs
analogous
to [FeCN]
as a catalyst
(Table
1, entry
15). Freeze-quenched
snapshots
at later
reaction
times
show
attenuation
in the signal
for [
57
FeOTf]
and the growth
of unknown
iron species.
After
80 min,
all of the [
57
FeOTf]
has
been
consumed;
the
remaining
iron
species
showed
poor
activity
following
extraction
and (re)subjection
to catalytic
conditions
at 25
°
C
with
fresh
acid,
reductant
and [TBA][CN],
yielding
only
an
additional
4.1
±
1 equiv
NH
3
(entry
16). NMR
analysis
of the
postcatalysis
mixture
revealed
evidence
of a diamagnetic
iron
hydride
(possibly
[Fe(H)(NHPh
2
)])
with
the
(P
3
Si
)Fe
platform
intact,
as well as free Si(H)P
3
(Figure
S13).
Relatedly,
iron hydrides
(e.g.,
[Fe(H)(N
2
)]) have
been
shown
to be off-
cycle
sinks
during
catalytic
N
2
R.
43,45,46
While
our mechanistic
studies
have
focused
on the most
efficient
catalyst,
[FeCN],
initial
studies
of the reactivity
of
(P
3
B
)Fe[BAr
F
4
] and
(PhBP
iPr
3
)FeBr
suggest
that
these
less
efficient
precatalysts
are also less stable
to excess
CN
−
. When
reacted
with
20 equiv
[TBA][CN]
in Et
2
O (in the absence
of
acid
or reductant),
free
phosphine
is observed,
indicating
partial
demetalation
as a pathway
for deactivation,
offering
a
plausible
reason
for the lower
turnover
numbers
(Figures
S14
and S15).
Early
N
−
H
Bond
Forming
Steps.
Since
metathesis
to
produce
[FeCN]
from
[FeOTf]
appeared
to be turnover
limiting,
we turned
to stoichiometric
experiments
to probe
the
role(s)
of early
intermediates
of reductive
protonation
in this
catalysis.
Exposing
a solution
containing
a mixture
of [FeCN]
and a
large
excess
(20
equiv)
of (C
6
H
6
)
2
Cr (unreactive
in the
absence
of acid)
to [Ph
2
NH
2
]OTf
(20 equiv,
added
via a
syringe)
caused
distinct
color
changes
that could
be monitored
by UV
−
vis
spectroscopy
(Figure
3). While
higher
energy
absorptions
(
λ
< 600 nm)
are complicated
by the absorption
spectrum
of (C
6
H
6
)
2
Cr
+/0
, the lower
energy
transitions
provide
a useful
handle
for the iron
speciation.
Upon
addition
of the
acid (all 20 equiv)
at RT, the characteristic
near-IR
absorption
of [FeCN]
(
λ
max
= 905 nm, FWHM
≈
100
nm;
pink
trace)
decayed
rapidly
(
τ
1/2
< 6 s) and a new,
broader
absorbance
characteristic
of [FeCNH
2
]
+
(
λ
max
= 929 nm, FWHM
≈
250
nm;
blue
trace)
was observed.
25
This
feature
decayed
more
slowly
(
τ
1/2
∼
40 s under
the conditions
studied)
with
Scheme
1. Proposed
Mechanism
for CN
−
Reduction
to NH
3
and CH
4
as Catalyzed
by [FeCN]
Figure
3.
UV
−
vis
data
showing
the
in situ
formation
of [FeCNH
2
]-
OTf
(blue
trace,
6 s after
acid addition)
in the reduction
of [FeCN]
(pink
trace)
to [FeOTf]
(orange
trace,
240 s after
acid addition).
Journal
of the American
Chemical
Society
pubs.acs.org/JACS
Article
https://doi.org/10.1021/jacs.3c12395
J. Am. Chem.
Soc.
2024,
146, 5343
−
5354
5346
simultaneous
growth
of a shoulder
that extends
further
into the
near-IR,
and a strong
absorption
centered
at 485 nm (orange
trace).
These
latter
features
are consistent
with
the formation
of [FeOTf],
as was also confirmed
by
1
H NMR
spectroscopy
(Figure
S12).
Products
were
also analyzed
(1.0 equiv
of NH
3
/
Fe and
0.9 equiv
CH
4
/Fe).
These
data
demonstrate
the
aminocarbyne
[FeCNH
2
]
+
as an observable
on-path
inter-
mediate
in the conversion
of [FeCN]
to [FeOTf].
Isosbestic
points
at 520 and 900 nm establish
that no further
downstream
intermediates
buildup
as [FeCNH
2
]
+
is converted
to [FeOTf]
during
liberation
of NH
3
and CH
4
.
To interrogate
shorter-lived
intermediates,
we next
studied
the consumption
of [FeCN]
at lower
temperature,
using
fewer
equivalents
of reductant
and
acid.
Mixing
[
57
FeCN]
with
[Ph
2
NH
2
]BAr
F
4
and (C
6
H
6
)
2
Cr (2.5
equiv
each)
in Et
2
O at
−
78
°
C for 1 min,
followed
by freeze-quench
(77 K) and
analysis
by Mo
̈
ssbauer
spectroscopy,
showed
primarily
[
57
FeCNH
2
]BAr
F
4
(Figure
4a;
δ
= 0.13
mm
s
−
1
and
Δ
E
Q
=
1.47
mm s
−
1
), again
supporting
the proposed
intermediacy
of
[FeCNH
2
]
+
during
catalysis.
25
These
low-temperature
con-
ditions
also allowed
identification
of the first intermediate
of
protonation,
[
57
FeCNH]BAr
F
4
(
δ
= 0.407
mm s
−
1
and
Δ
E
Q
=
3.20
mm s
−
1
),
25
as a minor
component,
consistent
with
step
(a) (Scheme
1).
We also
obtained
evidence
for a facile
redox
equilibrium
between
[FeCNH
2
]
+/0
and (C
6
H
6
)
2
Cr
+/0
, as can be expected
based
on the estimated
difference
in their
reduction
potentials
(
E
°
(Fe
+/0
)
∼ −
1.2
V,
E
°
(Cr
+/0
) =
−
1.2
V; see Section
S8.3 for
data).
Relatedly,
the available
data
imply
that single
electron
reduction
of [FeCNH
2
]
+
to [FeCNH
2
] (step
d) is feasible
under
conditions
relevant
to the
catalysis.
Accordingly,
addition
of 2.5 equiv
of [Ph
2
NH
2
]BAr
F
4
to a THF
solution
of [FeCN]
at
−
80
°
C resulted
in the immediate
formation
of
[FeCNH]BAr
F
4
(Figure
4b, blue trace).
25
,
47
Following
this, the
solution
was titrated
with
0
−
6
equiv
of (C
6
H
6
)
2
Cr to study
its
response
(Figure
4b).
During
the
addition
of the
first
equivalent
of (C
6
H
6
)
2
Cr, UV
−
vis
maxima
for [FeCNH]
+
(800
and 610 nm)
decreased
in intensity
and new
maxima
appeared,
reflecting
the growth
of [FeCNH
2
]
+
(929
and 570
nm; pink
trace).
Isosbestic
points
at 570 and 860 nm establish
no other
intermediate
buildup.
Upon
addition
of further
equivalents
of (C
6
H
6
)
2
Cr, the signals
for [FeCNH
2
]
+
attenuate
with
corresponding
growth
of a strong
absorbance
with
a
shoulder
at around
560 nm (orange
trace).
These
changes
are
consistent
with
the reduction
of [FeCNH
2
]
+
to [FeCNH
2
].
Still,
even
after
addition
of 6 equiv
(C
6
H
6
)
2
Cr, a large
fraction
of [FeCNH
2
]
+
remained.
48
These
results
confirm
a redox
equilibrium
between
[FeCNH
2
]
+/0
and
(C
6
H
6
)
2
Cr
+/0
(step
(d)).
As expected
for such
a redox
equilibrium,
cobaltocene,
a
stronger
reductant
than
(C
6
H
6
)
2
Cr (
E
°
(Cp
2
Co
3+/2+
) =
−
1.3
V), completely
reduces
[FeCNH
2
]
+
to [FeCNH
2
]. Accord-
ingly,
57
Fe Mo
̈
ssbauer
spectra
of the
reaction
between
[
57
FeCN]
with
[Ph
2
NH
2
]BAr
F
4
and Cp
2
Co (2.5
equiv
each)
at
−
78
°
C in Et
2
O reveal
the formation
of a single
new species
(
δ
= 0.02
mm s
−
1
and
Δ
E
Q
= 0.99
mm s
−
1
; Figure
4a). These
parameters
closely
resemble
those
of [FeCNMe
2
] (
δ
= 0.06
mm s
−
1
and
Δ
E
Q
= 1.12
mm s
−
1
),
25
consistent
with
formation
of [FeCNH
2
]. Complete
formation
of [FeCNH
2
] with only 2.5
equiv
of reductant
differs
markedly
from
conditions
using
(C
6
H
6
)
2
Cr. Titrations
monitored
by UV
−
vis
spectroscopy
using
Cp
2
Co still showed
[FeCNH
2
]
+
as an intermediate
upon
addition
of just 1 equiv
of Cp
2
Co to a mixture
of [FeCN]
and
[Ph
2
NH
2
]BAr
F
4
(Figure
4c, pink
trace).
Upon
addition
of a
second
equiv
of Cp
2
Co, [FeCNH
2
]
+
is fully
consumed
with
concomitant
formation
of [FeCNH
2
] (orange
trace).
Evidence
for C
−
H
Bond
Formation
via Fe
�
C(H)NH
2
+
.
The
intermediacy
of iron
carbynes
[FeCNH
2
]
+/0
in catalytic
CN
−
reduction
corresponds
to the intermediacy
of isolobal
hydrazidos
[FeNNH
2
]
+/0/
−
, during
Fe-catalyzed
N
2
R.
27
−
29,49,50
With
this
analogy
in mind,
we wondered
whether
iron carbynes
might
be selectivity
determining
in CN
−
reduction,
with
final
N
−
H
bond
formation
releasing
NH
3
(analogous
to N
β
−
H bond
formation
in N
2
R via hydrazido
intermediates),
resulting
in the observed
6 e
−
products
(CH
4
and
NH
3
), possibly
via a transient
carbide
[Fe(C)]
intermediate.
Computational
evidence
and
a study
of the
reactivity
of the methylated
carbyne
[FeCNMe
2
] complex
instead
support
the formation
of a C
−
H
bond
via a carbene
intermediate,
[Fe
�
C(H)(NH
2
)]
+
, as the next
step
of the
Figure 4.
(A)
57
Fe Mo
̈
ssbauer
spectra
of reaction
of [FeCN]
with
[Ph
2
NH
2
]BAr
F
4
and (C
6
H
6
)
2
Cr or Cp
2
Co. (B, C) UV
−
vis
data for reaction
of
[FeCN]
with
[Ph
2
NH
2
]BAr
F
4
and (C
6
H
6
)
2
Cr (B) or Cp
2
Co (C) to form
early
intermediates
[FeCNH]BAr
F
4
(blue
trace),
[FeCNH
2
]BAr
F
4
(pink
trace),
and [FeCNH
2
]
0
(orange
trace).
Journal
of the American
Chemical
Society
pubs.acs.org/JACS
Article
https://doi.org/10.1021/jacs.3c12395
J. Am. Chem.
Soc.
2024,
146, 5343
−
5354
5347