of 26
N–H Bond Dissociation Enthalpies and Facile H Atom Transfers
for Early Intermediates of Fe–N
2
and Fe–CN Reductions
Jonathan Rittle
and
Jonas C. Peters
Division of Chemistry and Chemical Engineering, California Institute of Technology (Caltech),
Pasadena, California 91125, United States
Abstract
Fe-mediated biological nitrogen fixation is thought to proceed via either a sequence of proton and
electron transfer steps, concerted H atom transfer steps, or some combination thereof. Regardless
of the specifics and whether the intimate mechanism for N
2
-to-NH
3
conversion involves a distal
pathway, an alternating pathway, or some hybrid of these limiting scenarios, Fe–N
x
H
y
intermediates are implicated that feature reactive N–H bonds. Thermodynamic knowledge of the
N–H bond strengths of such species is scant, and is especially difficult to obtain for the most
reactive early stage candidate intermediates (e.g., Fe–N=NH, Fe=N–NH
2
, Fe–NH=NH). Such
knowledge is essential to considering various mechanistic hypotheses for biological (and
synthetic) nitrogen fixation and to the rational design of improved synthetic N
2
fixation catalysts.
We recently reported several reactive complexes derived from the direct protonation of Fe–N
2
and
Fe–CN species at the terminal N atom (e.g., Fe=N–NH
2
, Fe–C
NH, Fe
C–NH
2
). These same
Fe–N
2
and Fe–CN systems are functionally active for N
2
-to-NH
3
and CN-to-CH
4
/NH
3
conversion, respectively, when subjected to protons and electrons, and hence provide an excellent
opportunity for obtaining meaningful N–H bond strength data. We report here a combined
synthetic, structural, and spectroscopic/analytic study to estimate the N–H bond strengths of
several species of interest. We assess the reactivity profiles of species featuring reactive N–H
bonds and estimate their homolytic N–H bond enthalpies via redox and acidity titrations. Very low
N–H bond dissociation enthalpies (BDE
N–H
), ranging from 65 (e.g., Fe–C
NH) to ≤37 kcal/mol
(Fe–N=NH), are determined. The collective data presented herein provides insight into the facile
reactivity profiles of early stage protonated Fe–N
2
and Fe–CN species.
Graphical abstract
Correspondence to: Jonas C. Peters.
Supporting Information: The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/
jacs.6b12861., Crystallographic data (CIF), Detailed experimental procedures and spectroscopic and theoretical data (PDF),
The authors declare no competing financial interest.
HHS Public Access
Author manuscript
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Published in final edited form as:
J Am Chem Soc
. 2017 March 01; 139(8): 3161–3170. doi:10.1021/jacs.6b12861.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Introduction
The conversion of N
2
to NH
3
via reductive protonation by six electrons and protons is a
fascinating transformation that is distinct from the Haber–Bosch process
1
and may offer a
conceptual pathway for direct, distributed solar-driven ammonia synthesis.
2
Free N
2
exhibits
a low electron affinity (−1.9 eV) and proton affinity (118 kcal/mol),
3
and activation by a
suitable catalyst is hence required. While growing in number,
4
few well-defined molecular
systems mediate catalytic N
2
-to-NH
3
conversion; those that do use a combination of
inorganic reductant and acid rather than H
2
. Nitrogenase enzymes represent the most
efficient and well-studied non-Haber–Bosch catalysts and are collectively responsible for the
generation of
50% of the global fixed nitrogen pool.
5
These enzymes sustain life and
provide proof-of-concept that reductive protonation of N
2
can be carried-out under ambient
conditions on a grand scale. But we do not as yet have a clear mechanistic picture of this
biological transformation.
Well-defined inorganic model complexes that catalyze N
2
-to-NH
3
conversion are
particularly well suited to mechanistic investigations and a number of synthetic systems
featuring early transition metals (especially Mo) have been studied in the context of N
2
activation
6
and catalytic N
2
-to-NH
3
conversion.
4
An early hallmark of the Mo model
systems, noteworthy owing to the presence of Mo in the FeMo-cofactor of MoFe-
nitrogenase, has been the ability to independently prepare many of the intermediate Mo–
N
x
H
y
species thought to be relevant to N
2
-to-NH
3
conversion.
7
Fe is an essential element in all known nitrogenases, and there is growing evidence that it
plays a critical mechanistic role in the key bond-making and breaking events of overall N
2
-
to-NH
3
conversion.
5
,
8
For example, structural, spectroscopic, and theoretical evidence
implicate the involvement of one (or more) of the iron atoms of the FeMo-cofactor in
substrate binding/activation.
8
This state of affairs has motivated the study of well-defined
Fe–N
x
H
y
species relevant to N
2
activation chemistry. In this context, Holland and Murray
have for instance independently demonstrated that
β
-diketiminate-supported iron systems
enable the delivery of electron equivalents to N
2
to generate bridging Fe-nitrido and/or
imido species.
9
,
10
N
2
-derived NH
3
can then be released from these systems upon subsequent
protonation.
9b
Our lab has reported several Fe systems that catalyze N
2
-to-NH
3
conversion,
11
and we have
very recently reported on a related (but as yet only stoichiometric) reductive protonation of
Rittle and Peters
Page 2
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Fe–CN to release CH
4
and NH
3
.
12
Interest in this subfield is growing. Very recently,
Nishibayashi and co-workers disclosed a phosphine–pyrrole Fe–N
2
system that is a
competent catalyst for NH
3
generation (along with some N
2
H
4
),
4a
and Ashley and co-
workers reported a low-valent phosphine-only Fe–N
2
system that mediates catalytic N
2
-to-
N
2
H
4
conversion.
4b
Fe-mediated, catalytic N
2
activation in these Fe systems is thought to
involve an initial end-on binding of N
2
to a single Fe atom, followed (in our P
3
E
Fe-systems
at least (E = B, C, Si)) by reductive protonation at the terminal N atom to afford Fe=N–NH
2
species (Scheme 1).
13
Fe-mediated reduction of P
3
Si
Fe–CN is thought to proceed via an
isostructural Fe
C–NH
2
species.
12
These species are characteristic of a “distal” or “Chatt-
type” pathway for substrate reduction that is commonly thought to operate in certain N
2
reduction systems.
3
,
4
,
7
Recent mechanistic studies of one of our reported Fe systems
(P
3
Si
Fe) also hints at competing downstream pathways following the generation of the
doubly protonated distal intermediate P
3
Si
Fe=N–NH
2
. These Fe=N–NH
2
(and related
Fe
C–NH
2
) species display a high degree of instability in comparison to analogous Mo=N–
NH
2
species; for instance, they decay rapidly at ambient and even lower temperatures in
solution and furnish mixtures of H
2
and NH
3
. Understanding the factors that govern their
reactivity patterns is critical from a mechanistic perspective and will aid in the development
of improved catalysts that show better selectivity for NH
3
versus H
2
. We anticipate that the
high degree of reactivity observed in certain Fe–N
x
H
y
(and Fe–CNH
x
) species must be
correlated with unusually weak N–H bonds; this should be especially true of the first
protonated species, Fe–N=NH (and Fe-C
NH).
14
These weak N–H bonds might then serve
a dual role: they could be susceptible to undesirable H
2
formation pathways that short circuit
overall substrate reduction, while concomitantly facilitating downstream H atom-transfer
pathways that productively lead to reduced product.
Whereas the thermochemistry of N–H bond formation in certain Ti-, Zr- and Mo-containing
scaffolds operative in stoichiometric NH
3
formation has been recently studied,
15
such
information is generally scant, and we saw an opportunity to map the N–H bond strengths of
the most highly reactive species relevant to Fe-mediated N
2
-to-NH
3
(and CN
) conversion
catalysis. In this report, we investigate H
+
/e
transfer to several Fe–CN- and Fe–N
2
-derived
species using the tripodal tris(phosphino)silyl ligand, P
3
Si
(P
3
Si
= [(2-
i
Pr
2
P–C
6
H
4
)
3
Si]
;
Scheme 1). While the P
3
Si
Fe–N
2
system is not as active as its P
3
B
Fe-cousin for N
2
-to-NH
3
conversion, it does mediate NH
3
generation and is technically better suited to the types of
studies detailed here. Similarly, the P
3
Si
Fe–CN system is competent for NH
3
and CH
4
generation
12
and offers a number of advantages in terms of measuring N–H bond strengths,
including experimental access to the first protonated intermediate (P
3
Si
Fe–C
NH
+/0
) and
well-behaved redox and acid–base properties.
Results and Discussion
N–H Bond Strength of P
3
Si
Fe–C
NH
0/+
While the conversion of Fe–N
2
to an Fe–N
x
H
y
species could in principle be mediated by
HAT steps, most commonly used H atom donor reagents are not sufficiently reactive to
donate H atoms to Fe–N
2
precursors.
16
The same is true of Fe–CN species.
12
Consistent
with this statement, there is no reaction between either P
3
Si
Fe–N
2
or P
3
Si
Fe–CN with 9,10-
Rittle and Peters
Page 3
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
dihydroanthracene (BDE
CH
= 81 kcal/mol) or TEMPOH (BDE
CH
= 71 kcal/mol) within 1 h
at room temperature in THF solution (Scheme 2).
17
This observation suggests (but does not
require) that the BDE
N–H
values for the corresponding P
3
Si
Fe=C=NH and P
3
Si
Fe–N=NH
species are <70 kcal/mol. For comparison, the experimentally determined BDE
N–H
of
diazene, HN=NH, is
61 kcal/mol.
18
Interestingly, an immediate reaction is observed upon
combining [P
3
Si
Fe–C
NH][BAr
F
24
]
12
with 1 equiv of TEMPO
to form [P
3
Si
Fe–CN]
[BAr
F
24
] and TEMPOH (Scheme 2). This latter observation indicates that the homolytic
bond dissociation free energy (BDFE) or bond dissociation enthalpy (BDE) of the N–H
bond in [P
3
Si
Fe–C
NH][BAr
F
24
] is significantly lower than that of the BDE
O–H
or
BDFE
O–H
in TEMPOH.
The prolonged THF solution stability of P
3
Si
Fe–CN, P
3
Si
Fe–CN
+
, P
3
Si
Fe–C
NH
+
, and
mixtures thereof permits the collection of data that can then be used to quantify the BDE
N–H
of P
3
Si
Fe–C
NH
+
. The solution BDE
N–H
is proportional to the sum of the acidity of the
relevant N–H bond (p
K
a
) and the oxidation potential
and may be estimated by the
application of eq 1 or 2,
(1)
(2)
where
C
H
is a solvent-dependent constant reported as 66 kcal/mol in THF.
17
,
19
,
20
,
Cyclic
voltammetry of P
3
Si
Fe–CN in a THF electrolyte (0.1 M [TBA][PF
6
]) reveals a one-electron
oxidation event at −0.38 versus the ferrocene/ferrocenium couple. [P
3
Si
Fe–C
NH][BAr
F
24
]
displays a reduction event at −1.27 V under similar conditions (Figure 1). To determine the
apparent acidity of [P
3
Si
Fe–C
NH]
+
in THF solution, titrations of P
3
Si
Fe–CN were
performed with a series of acids and monitored by UV–visible spectroscopy. P
3
Si
Fe–CN
does not detectably react with the acids [HNEt
3
][OTf] (p
K
a
13.7) or [LutH][OTf] (p
K
a
9.5);
stoichiometric addition of the chlorinated anilinium salts [2-Cl-PhNH
3
][OTf] (p
K
a
= 6.0 in
THF) or [2,5-Cl
2
-PhNH
3
][OTf] (p
K
a
= 4.5 in THF) results in the formation of equilibrium
mixtures of P
3
Si
Fe–CN and [P
3
Si
Fe–C
NH][OTf] at room temperature (Figure 1).
21
Spectrophotometric titrations performed with these acids and P
3
Si
Fe–CN furnish an average
p
K
a
= 5.6(1) for the acidic N–H proton of [P
3
Si
Fe–C=NH][OTf].
With these data in hand, the BDE
N–H
of [P
3
Si
Fe–C
NH][OTf] and its one-electron reduced
relative P
3
Si
Fe=C=NH may be estimated using eqs 1 and 2, respectively. These values are
found to be 65(1) and 44(1) kcal/mol, respectively (Table 1), and are consistent with the
complete and irreversible hydrogen atom abstraction reaction between [P
3
Si
Fe–C
NH]
[BAr
F
24
] and TEMPO
(BDE
O−H
= 71 kcal/mol), and the inability of P
3
Si
Fe–CN to react
with TEMPOH or dihydroanthracene via hydrogen atom abstraction. The decrease by more
than 20 kcal/mol in the N–H bond strength on going from cationic [P
3
Si
Fe–C
NH]
+
to
neutral P
3
Si
Fe=C=NH is striking.
Rittle and Peters
Page 4
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
We think this dramatic difference is largely a consequence of varying degrees of C–N bond
weakening that occur concomitant with H atom addition to the Fe–CN unit. Crystallographic
studies indicate that in cationic P
3
Si
Fe–C
NH
+
a short C–N bond is observed (1.130(3) Å)
and the N atom is largely sp-hybridized (
CNH 176(3)°) (Table 2).
12
Density functional
theory (DFT) calculations on the gas-phase optimized geometry of P
3
Si
Fe=C=NH predict an
sp
2
-hybridized N atom (
CNH 126.6°) and a long C–N bond (1.233 Å) (Supporting
Information). Comparison of these C–N bond lengths with the corresponding P
3
Si
Fe–CN
+/0
species reveal that formal H atom addition to P
3
Si
Fe–CN
+
results in minimal alteration to the
C–N bond length (0.014 Å) whereas H atom addition to neutral P
3
Si
Fe–CN results in
significant C–N bond lengthening (0.066 Å). Accordingly, the measured BDE
N–H
values
reflect the extent to which the C
N bond is weakened; weak N–H bonds are correlated with
a HAT process that causes significant weakening of the C
N (or N
N) bond. Generally
speaking, weak N–H bonds may be signatures of early stage intermediates of N
2
/CN
activation that are preceded by the most challenging H atom additions. Identifying these
intermediates is critical to understanding the thermodynamic challenges associated with
these difficult bond activation processes.
Monoprotonation of Fe–N
2
Terminally bound iron diazenido complexes (Fe–N=NH) have been invoked as the first
intermediates of biological N
2
reduction
5b
and are likewise presumed to be reactive
intermediates in stoichiometric and catalytic N
2
-to-NH
3
reactions effected by P
3
E
Fe–N
2
(E
= B, C, Si) and related systems.
11
,
13
Direct characterization of Fe–N=NH species has proven
to be a significant challenge, and circumstantial observations suggest that the BDFE
N–H
or
BDE
N–H
values for P
3
E
Fe–N=NH
0/+
intermediates are significantly lower than that of gas
phase HN=NH (
61 kcal/mol)
18
but can be anticipated to be similar to, or perhaps less than,
that estimated for P
3
Si
Fe=C=NH (44 kcal/mol). Given the comparatively high stability of the
related monoprotonated cyanide complex, [P
3
Si
Fe–C
NH][BAr
F
24
], we wondered whether
the N–H bond strength in isostructural P
3
Si
Fe–N=NH
+
could be estimated. Although the
oxidation potential of P
3
Si
Fe–N
2
(
E
° = −1.0 V) has been previously reported,
16
the weakly
basic properties of P
3
Si
Fe–N
2
prevents access to the needed p
K
a
data to directly estimate the
BDE
N–H
for P
3
Si
Fe–N=NH
+
in ethereal solvents. Nonetheless, several observations that
place a firm upper limit on its BDE
N–H
are as follows.
The addition of stoichiometric quantities of chlorinated anilinium triflate salts or [H(OEt
2
)
2
]
[BAr
F
24
] to P
3
Si
Fe–N
2
in THF at 298 K results in immediate oxidation to form P
3
Si
Fe–OTf
and [P
3
Si
Fe–N
2
][BAr
F
24
], respectively, prohibiting further investigations at this
temperature.
13
,
16
Combining solutions of P
3
Si
Fe–N
2
with stoichiometric [H(OEt
2
)
2
]
[BAr
F
24
] (p
K
a(THF)
0)
22
at 173 K leads to the formation of EPR-silent [P
3
Si
Fe–N
2
]
[BAr
F
24] over 15 min, as determined by UV–visible spectroscopy. EPR analysis of freeze-
quenched solutions of P
3
Si
Fe–N
2
rapidly combined with 1–5 equiv of [H(OEt
2
)
2
]-[BAr
F
24
]
shows unaltered EPR signatures for P
3
Si
–N
2
remaining in the frozen reaction mixtures
(Supporting Information). In contrast, the analogous protonation of isoelectronic
P
3
Si
Fe=C=NMe to form the carbyne [P
3
Si
Fe
C–N(Me)H][BAr
F
24
] results in a significant
change in the observed EPR signatures (vide infra). The aforementioned observations
collectively establish that P
3
Si
Fe–N
2
is not detectably protonated by [H(OEt
2
)
2
][BAr
F
24
] (at
Rittle and Peters
Page 5
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
the N atom or otherwise) at 173 K, and hence the p
K
a
of the putative P
3
Si
Fe–N=NH
+
is ≤0
in THF solution. According to eq 1, an upper limit for the BDE
N–H
value of P
3
Si
Fe–N=NH
+
can thereby be estimated at ≤43 kcal/mol. This value is significantly less than that
experimentally determined for P
3
Si
Fe–C
NH
+
(65 kcal/mol). The DFT-optimized geometry
of P
3
Si
Fe–N=NH
+
indicates significant N–N bond weakening (by 0.12 Å) on H atom
addition to P
3
Si
Fe–N
2
+
that is likely correlated with a very weak N–H bond. (Table 2,
Supporting Information).
In a similar context, our attempts to generate neutral P
3
Si
Fe–N=NH resulted in apparent
disproportionation reactions. The addition of stoichiometric [H(OEt
2
)
2
][BAr
24
] to P
3
Si
Fe–
N
2
results in immediate one-electron oxidation to form P
3
Si
Fe–N
2
when performed at
ambient temperatures, with simultaneous formation of 0.5 mol equiv of H
2
.
13b
,
16
We have
re-examined this reaction at 138 K in thawing 2-MeTHF, and while an instantaneous color
change is observed on combining P
3
Si
Fe–N
2
with 1 equiv of either [H(OEt
2
)
2
][BAr
F
24
] or
[PhNH
3
]-[OTf], we do not observe the multinuclear NMR resonances we expect for P
3
Si
Fe–
N=NH, a species that is likely to be diamagnetic by comparison to its isoelectronic relatives
P
3
Si
Fe–N=NMe (vide infra) and P
3
Si
Fe–N=NSiMe
3
16
and as corroborated by DFT
calculations that reveal an estimated singlet–triplet gap of 18 kcal/mol for P
3
Si
Fe–N=NH.
EPR signatures for the previously characterized
S
= 1/2 P
3
Si
Fe=N–NH
2
and P
3
Si
Fe–N
2
complexes are instead observed (eq 3; Supporting Information).
13b
Similar EPR features are
observed when [P
3
Si
Fe=N–NH
2
][OTf] is mixed with NEt
3
at 138 K (eq 3; Supporting
Information). To explain these observations, we suggest that a rapid disproportionation of
transient P
3
Si
Fe–N=NH may occur via intermolecular H atom transfer to furnish a 1:1
mixture of P
3
Si
Fe–N
2
and P
3
Si
Fe=N–NH
2
(Scheme 3).
23
For this disproportionation process
to proceed, the BDE
N–H
or BDFE
N–H
of P
3
Si
Fe–N=NH must be less than that of P
3
Si
Fe=N–
NH
2
. We note that P
3
Si
Fe=N–NH
2
is itself highly reactive and can only be prepared in
thawing 2-MeTHF (138 K). Collectively, these observations hint at very reactive N–H bonds
in P
3
Si
Fe–N=NH
+/0
and P
3
Si
Fe=N–NH
2
+/0
. The magnitude of these values are refined in
the following section.
The N–H Bond Enthalpy of the Iron Hydrazido Complex P
3
Si
Fe=N–N(Me)H
+
and Estimates
of the N–H Enthalpies of P
3
Si
Fe=N–NH
2
and P
3
Si
Fe–N=NH
To gain more quantitative insight into the reactive N–H bonds of certain protonated P
3
Si
Fe–
N
x
H
y
species, we examined the reactivity of the quasi-stable methyldiazenido complex,
P
3
Si
Fe–N=NMe, and its more stable conjugate acid, [P
3
Si
Fe=N–N(Me)H][OTf]. P
3
Si
Fe–
N=NMe is generated via the reaction of [K(Et
2
O)
n
][P
3
Si
Fe–N
2
] with MeOTf at 195 K
(Scheme 4) and is stable for hours at this temperature but rapidly decays at ambient
temperatures (
τ
(273k)
= 14 min) to a mixture of [P
3
Si
Fe=N–NMe
2
][OTf] and P
3
Si
Fe–OTf. At
193 K, NMR analysis reveals the presence of two distinct structural isomers of diamagnetic
P
3
Si
Fe–N=NMe that give rise to independent sets of multinuclear NMR parameters (refer to
the Supporting Information for further discussion and spectra). Nonetheless, sharp
resonances in the
15
N{
1
H} NMR spectrum at
δ
471 and 274 ppm characterize the N
atoms of the diazenido ligand; these resonance are distinct from those of [P
3
Si
Fe=N–NMe
2
]
[OTf] (
δ
= 530, 185 ppm) and [K-(Et
2
O)
n
][P
3
Si
Fe–N
2
] (
δ
= 340, 310 ppm).
13b
,
16
The
position of the [P
3
Si
Fe–N=NMe]-derived
15
N resonances are reasonably similar to those of
Rittle and Peters
Page 6
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
the isoelectronic and structurally characterized iron-silyldiazenido, P
3
Si
Fe–N=NSiMe
3
(419
and 271 ppm).
16
In situ IR spectra of THF solutions containing P
3
Si
Fe–N=NMe and
15
N-
enriched-P
3
Si
Fe–N=NMe reveal broad and intense features at 1616 and 1568 cm
−1
,
respectively, that are diagnostic of terminally bound diazenido ligands (Figure 2). These
v
(N=N) values are substantially lower than those of P
3
Si
Fe–N=NSiMe
3
(1748 and 1694
cm
−1
for the
14
N and
15
N isotopomers, respectively).
16
These spectroscopic differences
indicate a higher degree of N=N bond activation in P
3
Si
Fe–N=NMe as compared to P
3
Si
Fe–
N=NSiMe
3
. Accordingly, P
3
Si
Fe–N=NMe reacts productively with electrophiles (such as
Me
+
or H
+
, vide infra), whereas P
3
Si
Fe–N=NSiMe
3
does not.
Protonation of P
3
Si
Fe–N=NMe generated in situ with HOTf at 195 K results in an immediate
color change to lavender, and subsequent addition of pentane enables the isolation of
[P
3
Si
Fe=N–N(Me)H][OTf] (Scheme 4) as a crystalline pink powder. NMR data for
[P
3
Si
Fe=N–N(Me)-H][OTf] are similar to those of [P
3
Si
Fe=N–NH
2
][OTf]
13b
and include
two resonances in the
15
N{
1
H} NMR spectrum (
δ
= 518 and 199 ppm) with the upfield
resonance coupled to a single H atom (
1
J
NH
= 101 Hz), an N–
H
resonance at 9.96 ppm in
the
1
H NMR spectrum and a single
31
P resonance at 97 ppm. Broad features ascribed to N–
H vibrations are observed in the solid IR spectrum of [P
3
Si
Fe=N–N(Me)H][OTf] at =2980
cm
−1
that shift to
2295 cm
−1
in [P
3
Si
Fe=N–N(Me)D][OTf]; the low frequency of these
features are suggestive of strong hydrogen bonding interactions in the solid state. The solid-
state crystal structure of [P
3
Si
Fe=N–N(Me)-H][OTf] (Figure 3) corroborates the
spectroscopic data and reveals a short Fe–N distance of 1.671(3) Å, a long N–N distance
(1.284(4) Å), and a strong hydrogen bond (
d
(N···O) 2.846(4),
(N–H···O) 164(4)°) between
the OTf anion and the hydrazido ligand.
Analytical data collected on P
3
Si
Fe–N=NMe and [P
3
Si
Fe=N–N(Me)H][OTf] allow for the
estimation of the BDE
N–H
of the latter species. Cyclic voltammetry studies of P
3
Si
Fe–
N=NMe at 195 K in a 50 mM NaBAr
F
24
THF electrolyte reveal a single well-behaved
oxidation event centered at −1.22 V versus Fc/Fc
+
(Figure 2). The apparent acidity of
[P
3
Si
Fe=N–N(Me)H][OTf] was determined via its reaction with organic bases at 193 K in
THF to regenerate P
3
Si
Fe–N=NMe. While 1.1 equiv of NEt
3
(p
K
b(THF)
= 13.7)
quantitatively converts [P
3
Si
Fe=N–N(Me)H][OTf] to P
3
Si
Fe–N=NMe, the addition of 9.6
equiv of 2-Cl-PhNH
2
(p
K
b(THF)
= 6.0) to [P
3
Si
Fe=N–N(Me)H][OTf] results in little
detectable P
3
Si
Fe–N=NMe (Figure 4A). The addition of PhNH
2
(p
K
b(THF)
= 8.0) or 4-Cl-
PhNH
2
(p
K
b(THF)
= 7.0) to solutions of [P
3
Si
Fe=N–N(Me)H][OTf] results in immediate
deprotonation to afford mixtures of P
3
Si
Fe–N=NMe and [P
3
Si
Fe=N–N(Me)H]-[OTf] that
persist at 193 K for 30 min (Figure 4B,C). Subsequent analysis of the resulting Fe speciation
reveals an average p
K
a
= 8.0(2) for the hydrazido ligand of [P
3
Si
Fe=N–N(Me)H][OTf] in
THF solution. Combined with the oxidation potential for P
3
Si
Fe–N=NMe (vida supra), the
application of eq 1 provides an estimate of BDE
N–H
= 49(1) kcal/mol for [P
3
Si
Fe=N–
N(Me)H][OTf] (Table 1).
The thermal instability exhibited by the species more mechanistically relevant to N
2
-to-NH
3
conversion, P
3
Si
Fe=N–NH
2
0/+
, precludes the collection of similar experimental data.
Nonetheless, knowledge of the redox properties of P
3
Si
Fe–N=NMe (
E
ox
= −1.22 V) and
P
3
Si
Fe=N–NMe
2
(
E
ox
= −1.73 V) coupled with the experimental acidity of P
3
Si
Fe=N–
Rittle and Peters
Page 7
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
N(Me)H
+
(p
K
a
= 8.0) allow for rough estimations of the N–H bond strengths of P
3
Si
Fe=N–
NH
2
0/+
. If we plausibly assume that P
3
Si
Fe–N=NMe displays a similar basicity to P
3
Si
Fe–
N=NH, then we approximate the BDE
N–H
for P
3
Si
Fe=N–NH
2
+
as 49 kcal/mol. Additionally,
if we assume that P
3
Si
Fe=N–NMe
2
+
and P
3
Si
Fe=N–NH
2
+
are reduced at similar potentials,
the BDE
N–H
of P
3
Si
Fe=N–NH
2
may be estimated as 37 kcal/mol. DFT-predicted BDE
N–H
values for these species in the gas-phase are in excellent agreement with these estimations:
35 and 49 kcal/mol are predicted for the neutral and cationic P
3
Si
Fe=N–NH
2
species,
respectively (Supporting Information). In this context, the formation of P
3
Si
Fe=N–NH
2
(and
P
3
Si
Fe–N
2
) via disproportionation of P
3
Si
Fe–N=NH (Scheme 3) allows us to place a
reasonable upper bound on the BDE
N–H
of P
3
Si
Fe–N
NH at 37 kcal/mol. The magnitude of
this value speaks to a high degree of N=N bond activation that occurs concomitant with the
delivery of the first and second hydrogen atoms.
Thermoneutral H
+
/e
Equilibria between P
3
Si
Fe–C
NMe
0/+
and P
3
Si
Fe
C–N(Me)H
0/+
Attempts to determine the BDE
N–H
of the aminocarbyne [P
3
Si
Fe=C–NH
2
][OTf] are in part
hampered by the tendency of this compound to spontaneously decay with formation of H
2
and NH
3
.
12
Moreover, attempts to isolate neutral P
3
Si
Fe=C=NH via the reaction of [P
3
Si
Fe–
C
NH][BArF
24
] with stoichiometric Cp*
2
Co result only in intractable mixtures of Fe
species, presumably reflecting the low BDE
N–H
estimated for P
3
Si
Fe=C=NH (44 kcal/mol).
We therefore examined the reactivity of the isoelectronic alkylisocyanide complex,
P
3
Si
Fe=C=NMe (Scheme 5). This complex facilitates access to an iron-aminocarbyne-
derived N–H bond suitable for analytical study and provides calibration for the BDE
N–H
of
P
3
Si
Fe
C–NH
2
+
, the latter species (presumably) being a mechanistically relevant early stage
intermediate in Fe–mediated CN
reduction to NH
3
and CH
4
.
12
In contrast to P
3
Si
Fe=C=NH, both neutral P
3
Si
Fe=C=NMe and one-electron oxidized
[P
3
Si
Fe–C
NMe][OTf] are stable in the solid state (Figure 3) and as THF solutions at room
temperature. The cyclic voltammogram of a THF electrolyte solution of P
3
Si
Fe=C=NMe
displays a one-electron oxidation event at −1.31 V (Figure 5). Exposure of P
3
Si
Fe=C=NMe
to 1 equiv of [H(OEt
2
)
2
][BAr
F
24
] in Et
2
O and subsequent workup furnishes the secondary
aminocarbyne complex, [P
3
Si
Fe
C–N(Me)H][BAr
F
24
], as a purple solid that slowly
converts to [P
3
Si
Fe–C
NMe][BAr
F
24
] overnight in room temperature THF solution with the
concomitant loss of H
2
(0.4 equiv/Fe measured by GC; eq 3).
(3)
P
3
Si
Fe=C=NMe and [P
3
Si
Fe=C–N(Me)H][BAr
F
24
] are distinct by EPR (P
3
Si
Fe=C=NMe,
g
1
= 2.37,
g
2
= 2.02; [P
3
Si
Fe
C–N(Me)H][BAr
F
24
],
g
1
= 2.54,
g
2
= 1.97) and
57
Fe Mossbauer
spectroscopy (P
3
Si
Fe=C=NMe,
δ
= 0.25, Δ
E
Q
= 1.47; [P
3
Si
Fe
C–N(Me)H][BAr
F
24
],
δ
=
0.16, Δ
E
Q
= 1.51) (Figure 5). [P
3
Si
Fe
C–N(Me)H][BAr
F
24
] additionally displays a sharp
feature at 3337 cm
−1
in its solid-state IR spectrum that shifts to 2478 cm
−1
in [P
3
Si
Fe
C–
N(Me)D][BAr
F
24
], assigned as the N–H and N-D stretching frequencies, respectively. The
solid-state crystal structure (Figure 3) of [P
3
Si
Fe
C–N(Me)H][BAr
F
24
] reveals a short Fe–C
Rittle and Peters
Page 8
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
bond (1.751(9) Å) and a long C–N bond (1.28(1) Å) relative to those of P
3
Si
Fe=C=NMe
(
d
(
Fe–
C), 1.821(2) Å,
d
(C–N), 1.186(2) Å), consistent with a terminally bound
aminocarbyne ligand (Table 2).
24
The cyclic voltammogram of freshly prepared [P
3
Si
Fe=C–
N(Me)H][BAr
F
24
] in THF displays a reversible one-electron reduction event at −1.27 V
(Figure 5, Table 1). [P
3
Si
Fe=C–N(Me)H] [BAr
F
24
] is quantitatively deprotonated upon
exposure to 1.1 equiv of NEt
3
in THF solution, reforming P
3
Si
Fe=C=NMe and establishing
reversible proton transfer under these conditions.
Unlike the titration experiments performed with P
3
Si
Fe–CN and P
3
Si
Fe=N–N(Me)H
+
, the
reaction of P
3
Si
Fe=C=NMe with substoichiometric proton equivalents is complicated by
competing electron or H atom transfer processes (Figure 6A–C). While the exposure of
P
3
Si
Fe=C=NMe to 4 equiv of [2-Cl-PhNH
3
][OTf] results in the quantitative formation of
[P
3
Si
Fe
C–N(Me)H][OTf]; features assigned to those of the one-electron oxidized iron
isocyanide complex, [P
3
Si
Fe–C
NMe][OTf], and the neutral iron aminocarbyne P
3
Si
Fe
C–
N(Me)H (vide infra) are observed at earlier stages of the titration (Figure 6A,B). On further
addition of proton equivalents, the expected features of [P
3
Si
Fe
C–N(Me)H]-[OTf] grow in
with the concomitant loss of features assigned to the neutral P
3
Si
Fe
C–N(Me)H and
[P
3
Si
Fe–C=NMe]-[OTf] (Figure 6C). This series of spectral changes has been observed with
a variety of chlorinated anilinium salts and with subsubstoichiometric amounts of
[H(OEt
2
)2][BAr
F
24
]. Inspection of the available cyclic voltammetry data (Figure 5A)
indicates that the
oxidation
of P
3
Si
Fe=C=NMe occurs at nearly the same potential as the
reduction of [P
3
Si
Fe
C–N(Me)H][BAr
F
24
]. Solutions generated during the acid titration
experiments contain mixtures of these two species and are hence prone to spontaneously
disproportionate to an equilibrium mixture of P
3
Si
Fe–CNMe
+/0
and P
3
Si
Fe=C–N(Me)H
+/0
(illustrated schematically in Figure 6D).
Multinuclear NMR spectroscopy was employed to provide direct evidence for the competing
H
+
/e
transfer processes observed in solutions containing both P
3
Si
Fe=C=NMe and
P
3
Si
Fe
C–N(Me)H
+
. The presented model (Figure 6D) predicts the accumulation of neutral
and diamagnetic P
3
Si
Fe
C–N(Me)H via an intermolecular electron or H atom transfer
between P
3
Si
Fe=C=NMe and P
3
Si
Fe=C–N(Me)H
+
. To mimic the conditions at the midpoint
of an acidity titration, P
3
Si
Fe=
13
C=NMe, P
3
Si
Fe=C=
15
NMe, and [H(OEt
2
)
2
]-[BAr
F
24
] can
be dissolved in THF-
d
8
in a 1:1:1 ratio at 193 K, and the resulting solution can be analyzed
by multinuclear NMR spectroscopy. At this temperature, a sharp doublet (
1
J
NH
93 Hz) in
the
15
N NMR spectrum appears at
δ
= 135.2 ppm that is attributed to diamagnetic
P
3
Si
Fe
C–N(Me)H.
25
The chemical shift of this resonance is comparable to that of
structurally characterized and isoelectronic P
3
Si
Fe
C–NMe
2
(129.1 ppm), and the
magnitude of
1
J
NH
unambiguously confirms the presence of an N–H unit.
12
Two resonances
are observed in the
13
C{
1
H} NMR spectrum at
δ
= 279.7 and 282.7 ppm that may be
assigned to aminocarbyne carbons of two slightly different structural isomers of P
3
Si
Fe
C–
N(Me)H.
26
These NMR data confirm the spontaneous generation of P
3
Si
Fe
C–N(Me)H in
these solutions despite the absence of an exogenous reductant or H atom source and support
the hypothesis that P
3
Si
Fe=C=NMe and P
3
Si
Fe=C–N(Me)H
(+)
are subject to rapid,
reversible, and nearly thermoneutral H
+
/e
transfer in THF solutions at temperatures as low
as 193 K.
Rittle and Peters
Page 9
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Despite the complex speciation resulting from acid titrations of P
3
Si
Fe
C=NMe, the p
K
a
for
P
3
Si
Fe
C–N(Me)H
(+)
can be determined by the deconvolution of the spectral data at a
known acid concentration to define the relative proportions of the individual Fe species (see
the Supporting Information for details). A pure spectrum of neutral P
3
Si
Fe
C–N(Me)H is
unavailable, however, as attempts to isolate this compound in pure form have been
unsuccessful (
τ
(1/2)
< 60 min in THF at 273 K). Available NMR data indicates that neutral
P
3
Si
Fe
C–NMe
2
expectedly bears a similar electronic structure to P
3
Si
Fe
C–N(Me)H,
hence providing a model UV–visible absorption spectrum of the latter species.
27
Indeed, the
spectra shown in Figure 6 are satisfactorily simulated by a linear combination of the pure
spectra of [P
3
Si
Fe–C
NMe][OTf], [P
3
Si
Fe
CN(Me)H][OTf], P
3
Si
Fe=C=NMe, and
P
3
Si
Fe
C–NMe
2
(as a surrogate for P
3
Si
Fe
C–N(Me)H). Analysis of the spectral data
obtained by titrating P
3
Si
Fe=C=NMe with [PhNH
3
][OTf] (p
K
a
= 8.0), [4-Cl-PhNH
3
][OTf]
(p
K
a
= 7.0), and [2-Cl-PhNH
3
][OTf] (p
K
a
= 6.0) reveals an average p
K
a
of 7.1(3) for the
acidic proton of [P
3
Si
Fe=C–N(Me)H][OTf] in THF. Using this value, the relevant BDE
N–H
values for [P
3
Si
Fe
C–N(Me)H][OTf] and P
3
Si
Fe
C–N(Me)H are estimated with eqs 1 and 2
and found to be 47(1) and 46(1) kcal/mol, respectively, corroborating the plausibility of
intermolecular H atom transfer between P
3
Si
Fe=C=NMe and [P
3
Si
Fe
C–N(Me)H][OTf] in
solution (Table 1, Figure 6).
The invariance of the BDE
N–H
within the P
3
Si
Fe
C– N(Me)H
+/0
redox series is unusual and
starkly contrasts that of the P
3
Si
Fe–CNH
+/0
redox series. In general, the one-electron
oxidation of transition metal complexes that bear X–H (X = N, O) bonds results in a noted
increase in BDE
X–H
.
15d
,
28
The BDE
M–H
values measured in redox-active metal hydride
species also display significant oxidation state dependence, but in this case the trend is
reversed.
29
The equivalent BDE
N–H
values observed for P
3
Si
Fe
C–N(Me)H underscores
their similar hybridizations at N(sp) and also likely reflects similar degrees of C–N bond
weakening upon H atom addition to P
3
Si
Fe– CNMe
+/0
. Strong backbonding from Fe to C in
the isocyanide P
3
Si
Fe=C=NMe causes bending at N, favoring a resonance contributor with
significant Fe to C
π
-bonding (as opposed to P
3
Si
Fe–C
NMe). This is born out in its crystal
structure (Figure 3) and may serve to minimize structural reorganization associated with H
atom transfer between P
3
Si
Fe=C=NMe and [P
3
Si
Fe=C–N(Me)H][OTf]. Collectively, these
facile H atom transfer reactions provide a model for the various disproportionation reactions
invoked for the P
3
Si
Fe–N= NH
0/+
and P
3
Si
Fe=N–NH
2
0/+
species discussed above.
Concluding Remarks
The N
2
-to-NH
3
conversion mediated by biological and synthetic systems requires the
delivery of multiple proton/ electron equivalents to N
2
via various intermediates. We have
speculated that P
3
X
Fe–based catalysts are likely to operate via mechanisms that initiate
along a “distal” pathway, and previous studies indicate similar mechanisms for a
conceptually related stoichiometric conversion of CN
to CH
4
and NH
3
by the P
3
Si
Fe
system. Synthetic access to rare examples of Fe– N=NR, Fe=N–NR
2
, Fe–CNR, and Fe=C–
NR
2
species (R = H, Me) derived from P
3
Si
Fe–N
2
/CN has enabled the in-depth study of
relevant intermediates in these proposed mechanisms. The measurements of pertinent redox
and acidity values for many of these species reported herein indicate that that these early
intermediates display weak N–H bonds, ranging from <37 to 65 kcal/mol. This range of
Rittle and Peters
Page 10
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
values is likely to be linked to the varying degrees of N
N or C
N bond weakening that
occurs concomitant with H atom addition.
A noteworthy exception to the preceding manifold ofcharacterized compounds is the initial
intermediate of N
2
-to-NH
3
conversion, P
3
Si
Fe–N=NH
+/0
, which has eluded
directspectroscopic detection. P
3
Si
Fe–N=NH
+
is apparently tooacidic to observe in ethereal
solvents and neutral P
3
Si
Fe–N=NH appears to undergo rapid disproportionation to
P
3
Si
Fe=N
NH
2
and P
3
Si
Fe–N
2
at temperatures as low as 138 K.Experimentally, we can
assign an upper limit of 37 kcal/mol forthe BDE
N–H
of P
3
Si
Fe–N=NH by comparison
toappropriate model complexes. This suggests that the initial Hatom addition to P
3
Si
Fe–N
2
is likely the most thermody-namically uphill step in the overall N
2
-to-NH
3
conversion.Theoretical studies suggests that this step is also challenging inthe biological
system
8b
,
30
and highlight the possibility of cooperative activation of N
2
by two or more of
the metal ions in the active-site cofactor. Indeed, numerous bimetallic complexes have been
shown to stabilize diazenido and diazene ligands,
6d
,
31
and the stability of these complexes
may be ascribed to cooperative N=N bond activation by the two metal centers. These
considerations certainly motivate further development of multimetallic platforms capable of
reductive N
2
protonation.
The reactions of the various protonated P
3
Si
Fe–N
2
/CN species studied herein have been
conducted in relative isolation under well-defined conditions. While these individual
reactions are likely mechanistically relevant to catalytic and stoichiometric N
2
/CN
reductions, these conditions contrast those employed in catalytic NH
3
synthesis, namely, the
simultaneous
inclusion of excess inorganic acid and reductant sources in Et
2
O. Under these
conditions, both H
2
evolution and N
H bond formation reactions will proceed at unique
rates that will depend on the choice of Fe precatalysts and the nature of these reagents and
will hence compete for the H
+
/e
equivalents present in solution. Both of these processes are
undoubtedly facilitated by the low BDE
NH
values exhibited by these species, but these
thermodynamic values are only indirectly linked to the rates of the relevant N–H bond
formation and N–N or C– N cleavage reactions. Detailed kinetic investigations of these
processes should shed light on additional factors that govern proton-coupled N
2
fixation by
molecular Fe systems. The complexes described herein will provide useful synthons for such
investigations.
Supplementary Material
Refer to Web version on PubMed Central for supplementary material.
Acknowledgments
This work was supported by the NIH (Grant GM 070757) and the Gordon and Betty Moore Foundation. J.R. was
additionally supported by a fellowship from the Caltech Center for Environmental Microbial Interactions (CEMI).
References
1. Schlögl R. Angew Chem, Int Ed. 2003; 42:2004–2008.
2. (a) Brown KA, Harris DF, Wilker MB, Rasmussen A, Khadka N, Hamby H, Keable S, Dukovic G,
Peters JW, Seefeldt LC, King PW. Science. 2016; 352:448–450. [PubMed: 27102481] (b) Liu J,
Rittle and Peters
Page 11
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Kelley MS, Wu W, Banerjee A, Douvalis AP, Wu J, Zhang Y, Schatz GC, Kanatzidis MG. Proc Natl
Acad Sci U S A. 2016; 113:5530–5535. [PubMed: 27140630] (c) Lehnert N, Peters JC. Inorg
Chem. 2015; 54:9229–9223. [PubMed: 26434648] (d) Michalsky R, Pfromm PH, Steinfeld A.
Interface Focus. 2015; 5:20140084. [PubMed: 26052421]
3. Jia HP, Quadrelli EA. Chem Soc Rev. 2014; 43:547–564. [PubMed: 24108246]
4. (a) Kuriyama S, Arashiba A, Nakajima K, Matsuo Y, Tanaka H, Ishii K, Yoshizawa K, Nishibayashi
T. Nat Commun. 2016; 7:12181. [PubMed: 27435503] (b) Hill PJ, Doyle LR, Crawford AD, Myers
WK, Ashley AE. J Am Chem Soc. 2016; 138:13521–13524. [PubMed: 27700079] (c) Yandulov DV,
Schrock RR. Science. 2003; 301:76–78. [PubMed: 12843387] (d) Ritleng V, Yandulov DV, Weare
WW, Schrock RR, Hock AS, Davis WM. J Am Chem Soc. 2004; 126:6150–6163. [PubMed:
15137780] (e) Arashiba K, Kinoshita E, Kuriyama S, Eizawa A, Nakajima K, Tanaka H, Yoshizawa
K, Nishibayashi Y. J Am Chem Soc. 2015; 137:5666–5669. [PubMed: 25879994] (f) Kuriyama S,
Arashiba K, Tanaka H, Matsuo Y, Nakajima K, Yoshizawa K, Nishibayashi Y. Angew Chem, Int Ed.
2016; 55:14291–14295.
5. (a) Howard JB, Rees DC. Chem Rev. 1996; 96:2965–2982. [PubMed: 11848848] (b) Seefeldt LC,
Hoffman BM, Dean DR. Annu Rev Biochem. 2009; 78:701–722. [PubMed: 19489731] (c) Burgess
BK, Lowe DJ. Chem Rev. 1996; 96:2983–2011. [PubMed: 11848849]
6. Representative examples include Chatt J, Pearman AJ, Richards RL. J Chem Soc, Dalton Trans.
1977:1852–1860.Laplaza CE, Cummins CC. Science. 1995; 268:861–863. [PubMed: 17792182]
Shima T, Hu SW, Luo G, Kang XH, Luo Y, Hou ZM. Science. 2013; 340:1549–1552. [PubMed:
23812710] Fryzuk MD, Love JB, Rettig SJ, Young VG. Science. 1997; 275:1445–1447.
7. (a) Yandulov DV, Schrock RR. J Am Chem Soc. 2002; 124:6252–6253. [PubMed: 12033849] (b)
Yandulov DV, Schrock RR. Inorg Chem. 2005; 44:1103–1117. [PubMed: 15859292] (c) Weare
WW, Dai X, Byrnes MJ, Chin JM, Schrock RR, Muller P. Proc Natl Acad Sci U S A. 2006;
103:17099–17106. [PubMed: 17085586] (d) Munisamy T, Schrock RR. Dalton Trans. 2012;
41:130–137. [PubMed: 22031021] (e) Chatt J, Dilworth J, Richards RL. Chem Rev. 1978; 78:589–
625.
8. (a) Spatzal T, Perez KA, Einsle O, Howard JB, Rees DC. Science. 2014; 345:1620–1623. [PubMed:
25258081] (b) Lukoyanov D, Khadka N, Yang ZY, Dean DR, Seefeldt LC, Hoffman BM. J Am
Chem Soc. 2016; 138:10674. [PubMed: 27529724] (c) Hinnemann B, Norskov JK. Top Catal. 2006;
37:55–70.(d) Schneider K, Gollan E, Dröttboom M, Selsemeier-Voigt S, Müller A. Eur J Biochem.
1997; 244:789–800. [PubMed: 9108249]
9. (a) Rodriguez MM, Bill E, Brennessel WW, Holland PL. Science. 2011; 334:780–783. [PubMed:
22076372] (b) Macleod KC, McWilliams SF, Mercado BQ, Holland PL. Chem Sci. 2016; 7:5736–
5746. [PubMed: 28066537]
10. Lee Y, Sloane FT, Blondin G, Abboud KA, Garcia-Serres R, Murray LJ. Angew Chem, Int Ed.
2015; 54:1499–1503.
11. (a) Anderson JS, Rittle J, Peters JC. Nature. 2013; 501:84–87. [PubMed: 24005414] (b) Creutz SE,
Peters JC. J Am Chem Soc. 2014; 136:1105–1115. [PubMed: 24350667] (c) Ung G, Peters JC.
Angew Chem, Int Ed. 2015; 54:532–535.(d) Del Castillo TJ, Thompson NB, Peters JC. J Am
Chem Soc. 2016; 138:5341–5350. [PubMed: 27026402]
12. Rittle J, Peters JC. Angew Chem, Int Ed. 2016; 55:12262–12265.
13. (a) Anderson JS, Cutsail GE, Rittle J, Connor BA, Gunderson WA, Zhang LM, Hoffman BM,
Peters JC. J Am Chem Soc. 2015; 137:7803–7809. [PubMed: 26000443] (b) Rittle J, Peters JC. J
Am Chem Soc. 2016; 138:4243–4248. [PubMed: 26937584]
14. The Lewis structures we adopt provide simple representations that emphasize aspects of the
bonding but are not reliable representations as such. For example, we draw P
3
Si
Fe(CNH)
+/0
as
P
3
Si
Fe–C
NH
+
or P
3
Si
Fe=C=NH
0
, and these representations are informed by available structural,
spectroscopic, or theoretical data. Additionally we draw P
3
Si
Fe(NNH) as P
3
Si
Fe–N=NH but
realize that in all cases there is varying degrees of
π
-bonding between Fe and the connected C- or
N-atom.
15. (a) Bezdek MJ, Chirik PJ. Dalton Trans. 2016; 45:15922–15930. [PubMed: 27378665] (b) Pappas
I, Chirik PJ. J Am Chem Soc. 2015; 137:3498–3501. [PubMed: 25719966] (c) Pappas I, Chirik PJ.
J Am Chem Soc. 2016; 138:13379–13389.(d) Bezdek MJ, Guo S, Chirik PJ. Science. 2016;
354:730–733. [PubMed: 27846601]
Rittle and Peters
Page 12
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
16. Lee Y, Mankad NP, Peters JC. Nat Chem. 2010; 2:558–565. [PubMed: 20571574]
17. Warren JJ, Tronic TA, Mayer JM. Chem Rev. 2010; 110:6961–7001. [PubMed: 20925411]
18. Luo, YR. Comprehensive Handbook of Chemical Bond Energies. CRC Press; Boca Raton: 2007.
19. The estimation of BDFE
NH
values by this approach is not currently possible in THF as the
corresponding C
G
value is unavailable. We therefore report BDE
NH
values, for which a C
H
constant in THF is known.
20
While entropic effects may influence these estimated BDE
NH
values,
the general trends shown for these complexes are unlikely to be significantly affected.
20. Cappellani EP, Drouin SD, Jia G, Maltby PA, Morris RH, Schweitzer CT. J Am Chem Soc. 1994;
116:3375–3388.
21. Garrido G, Roses M, Rafols C, Bosch E. J Solution Chem. 2008; 37:689–700.
22. Ding FZ, Smith JM, Wang HB. J Org Chem. 2009; 74:2679–2691. [PubMed: 19275192]
23. The reduction potential of P
3
Si
Fe=N–NH
2
+
is unknown but is likely similar to that measured for
P
3
Si
Fe=N–NMe
2
+
(−1.7 V), and thus outer sphere electron transfer from P
3
Si
Fe–N
2
(
E
ox
= −2.1
V) to P
3
Si
Fe=N–NH
2
+
is an alternative mechanism by which P
3
Si
Fe=N–NH
2
and P
3
Si
Fe–N
2
are
generated in these reaction mixtures.
24. The aminocarbyne ligand is disordered over two well-defined positions, but in both instances, a
short Fe–C bond (1.751(9) Å and 1.72(2) Å) and long C–N bond (1.28(1) Å and 1.27(2) Å)
corroborate the aminocarbyne assignment.
25. Apparently, chemical exchange of the various P
3
Si
Fe species is slow at this temperature, relative to
the NMR time scale. Experiments performed at higher temperatures are complicated by the
thermal decomposition of P
3
Si
Fe
C–N(Me)H.
26. The presence of two resonances likely arises from slightly different geometric isomers of
P
3
Si
Fe
C–N(Me)H that are trapped out at low temperatures. Indeed, P
3
Si
Fe
CN–Me
2
similarly
displays two features in its
13
C{
1
H} NMR spectrum at this temperature (
δ
= 280.8 and 277.2
ppm), but only one feature is observed at 293 K (
δ
= 279.1 ppm).
12
27. Rittle J, Younker JM, Green MT. Inorg Chem. 2010; 49:3610–3617. [PubMed: 20380463]
28. Taguchi T, Stone KL, Gupta R, Kaiser-Lassalle B, Yano J, Hendrich MP, Borovik AS. Chem Sci.
2014; 5:3064–3071. [PubMed: 25580212]
29. Hu Y, Shaw AP, Estes DP, Norton JR. Chem Rev. 2016; 116:8427–8462. [PubMed: 26828562]
30. Dance I. Dalton Trans. 2008:5977–5991. [PubMed: 19082054]
31. (a) Saouma CT, Muller P, Peters JC. J Am Chem Soc. 2009; 131:10358–10359. [PubMed:
19722612] (b) Hirotsu M, Fontaine PP, Zavalij PY, Sita LR. J Am Chem Soc. 2007; 129:12690–
12692. [PubMed: 17902670] (c) Pool JA, Lobkovsky E, Chirik PJ. Nature. 2004; 427:527–530.
[PubMed: 14765191]
Rittle and Peters
Page 13
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Figure 1.
(A) UV–visible absorption spectra of P
3
Si
Fe–CN (dark red) and [P
3
Si
Fe–C
NH][OTf]
(black) in THF at 25 °C. The other traces were obtained following the sequential addition of
0.33 mol equiv of [2-Cl-PhNH
3
][OTf] to P
3
Si
Fe–CN. The dark blue trace was obtained
following the combined addition of 2.66 equiv [2-Cl-PhNH
3
][OTf] to P
3
Si
Fe–CN. Arrows
indicate features whose intensity increases or decreases during the experiment. Cyclic
voltammograms of (B) P
3
Si
Fe–CN and (C) [P
3
Si
Fe–C=NH]-[BAr
F
24
] scanned at 200 mV/s
in 0.1 M [TBA][PF
6
] THF electrolytes. Solid lines indicate voltammograms initially
scanned in the anodic direction, and dashed lines indicate voltammograms initially scanned
in the cathodic direction.
Rittle and Peters
Page 14
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Figure 2.
(A) Overlaid solution IR spectra of THF solutions of in situ-generated P
3
Si
Fe–N=NMe
(black) and N-enriched P
3
Si
Fe–N=NMe (red) collected at 193 K. (B) Cyclic voltammograms
obtained on a 195 K THF solution of P
3
Si
Fe–N=NMe at the listed scan rate (0.1 M
NaBAr
F
24
electrolyte).
Rittle and Peters
Page 15
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript
Figure 3.
X-ray diffraction crystal structures of [P
3
Si
Fe–C
NMe]-[OTf], P
3
Si
Fe=C=NMe, [P
3
Si
Fe=C–
N(Me)H][BAr
F
24
], and [P
3
Si
Fe=N–N(Me)H][OTf] with thermal ellipsoids drawn at 50%
probability. Hydrogen atoms (except the N–H's), the BAr
F
24
counteranion of [P
3
Si
Fe=C–
N(Me)H][BAr
F
24
], and the OTf counteranion of [P
3
Si
Fe–CNMe][OTf] have been removed
for clarity.
Rittle and Peters
Page 16
J Am Chem Soc
. Author manuscript; available in PMC 2018 March 01.
Author Manuscript
Author Manuscript
Author Manuscript
Author Manuscript