How Innocent are Potentially Redox Non-Innocent Ligands? Electronic Structure and Metal Oxidation States in Iron-PNN Complexes as a Representative Case Study

Article abstract of DOI:10.1021/acs.inorgchem.5b00509

Herein we present a series of new α-iminopyridine-based iron-PNN pincer complexes [FeBr₂L_PNN] (1), [Fe(CO)₂L_PNN] (2), [Fe(CO)₂L_PNN](BF₄) (3), [Fe(F)(CO)₂L_PNN

Full text of DOI:10.1021/acs.inorgchem.5b00509

Article
pubs.acs.org/IC
How Innocent are Potentially Redox Non-Innocent Ligands?
Electronic Structure and Metal Oxidation States in Iron-PNN
Complexes as a Representative Case Study
Burkhard Butschke,†,⊥ Kathlyn L. Fillman,§,⊥ Tatyana Bendikov,‡ Linda J. W. Shimon,‡
Yael Diskin-Posner,‡ Gregory Leitus,‡ Serge I. Gorelsky,¶ Michael L. Neidig,*,§ and David Milstein*,†
‡
†Department of Organic Chemistry and Department of Chemical Research Support, The Weizmann Institute of Science, P.O. Box
26, 76100 Rehovot, Israel
§Department of Chemistry, University of Rochester, Rochester, New York 14627, United States
¶Centre for Catalysis Research and Innovation, University of Ottawa, Ottawa, Ontario K1N 6N5, Canada
S
* Supporting Information
ABSTRACT: Herein we present a series of new α-
iminopyridine-based iron-PNN pincer complexes [FeBr2LPNN
]
(1), [Fe(CO)2LPNN] (2), [Fe(CO)2LPNN](BF4) (3), [Fe(F)-
(CO)2LPNN](BF4) (4), and [Fe(H)(CO)2LPNN](BF4) (5)
with formal oxidation states ranging from Fe(0) to Fe(II)
(LPNN = 2-[(di-tert-butylphosphino)methyl]-6-[1-(2,4,6-
mesitylimino)ethyl]pyridine). The complexes were character-
1
ized by a variety of methods including H, 13C, 15N, and 31P
NMR, IR, Mossbauer, and X-ray photoelectron spectroscopy
̈
(XPS) as well as electron paramagnetic resonance (EPR) and magnetic circular dichroism (MCD) spectroscopy, SQUID
magnetometry, and X-ray crystallography, focusing on the assignment of the metal oxidation states. Ligand structural features
suggest that the α-iminopyridine ligand behaves as a redox non-innocent ligand in some of these complexes, particularly in
[Fe(CO)2LPNN] (2), in which it appears to adopt the monoanionic form. In addition, the NMR spectroscopic features (13C, 15N)
indicate the accumulation of charge density on parts of the ligand for 2. However, a combination of spectroscopic measurements
that more directly probe the iron oxidation state (e.g., XPS), density functional theory (DFT) calculations, and electronic
absorption studies combined with time-dependent DFT calculations support the description of the metal atom in 2 as Fe(0). We
conclude from our studies that ligand structural features, while useful in many assignments of ligand redox non-innocence, may
not always accurately reflect the ligand charge state and, hence, the metal oxidation state. For complex 2, the ligand structural
changes are interpreted in terms of strong back-donation from the metal center to the ligand as opposed to electron transfer.
innocent nature of the NO ligand, which can adopt charge states
from −2 to +1.2 In general, the determination of a metal
oxidation state becomes ambiguous when redox non-innocent
ligands are involved, and typical examples of such ligands are
1,2-dioxolenes (or catecholates),3−5 1,2-dithiolenes,6 2-amido-
phenoxides,3,7−9 2,2′-bipyridines,10−16 α-iminopyridines,17−19
2,6-diiminopyridines,20−26 and numerous others.27−33 These
ligands are able to change their redox state and to participate in
electron transfer to or from the metal center; as a consequence,
unfavorable metal oxidation states can be avoided or seemingly
unusual metal oxidation states can be stabilized. In the past
decade, catalysis research has made significant progress in
utilizing the electronic flexibility possible with these
ligands.29,30,34−36 While the concept of formal oxidation states37
is a formalism that relies on counting covalent bonds and ionic
charges, a ligand set around a metal atom influences the
electron density at the metal center when compared with the
1. INTRODUCTION
It is the coordination chemist’s desire to classify metal
complexes in terms of their metal oxidation state. On the
basis of this designation, reactivity patterns and the tendency of
the metal center to bind certain ligands can be predicted.
However, the assignment of metal oxidation states is often
ambiguous, and the determination of the true electronic ground
state of organometallic complexes can be challenging. For
example, with the iron complex [Fe(CO)3(NO)]− it took more
than 50 years after its first synthesis to determine that this
complex is not isoelectronic to [Fe(CO)4]2−, the latter of which
is an Fe(−II) complex.1 [Fe(CO)3(NO)]− is rather an Fe(0)
complex that contains an anionic NO− ligand even though the
linear coordination of the ligand to the metal center seemingly
points to the presence of an NO+ ligand.1 Only a thorough
spectroscopic analysis combined with density functional theory
(DFT) and complete active space self-consistent field
(CASSCF) calculations revealed the unexpected electronic
structure of this iron complex. The challenge in the assignment
of the metal oxidation state in this example is the redox non-
Received: March 4, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
bare metal atom or ion. Consequently, it makes sense to think
about electronic structure descriptors that will pave the way to a
correct interpretation of physical and chemical properties of a
given complex. One of these electronic-structure descriptors
can be a parameter that reflects the electron density distribution at
and around a metal center in a complex as a consequence of
metal-ligation electronic interactions and, thus, the shift of
electron density between the different fragments of the
complex. Such a parameter is analogous to the oxidation state
except that it is more closely connected to the physical picture
(actual electronic distribution and metal−ligand interactions).
Thus, it can be called a physical oxidation state. Redox non-
innocent ligands can have a significant influence on the metal
center, but examples also exist where they serve as innocent
spectator ligands.30 The actual mode of interaction of the ligand
with the metal center is determined by the identity of the metal
center, the actual ligand set, and sometimes even by second-
sphere effects (counterions and interactions with solvent
molecules).38 To determine if a ligand participates in electron
transfer to or from the metal center (and, hence, if it is redox
non-innocent in the complex), its geometry within a given
metal complex can be used as a criterion, and this correlation of
the ligand geometry with its charge state is based on a variety of
thorough studies.9−13,15−19,21,22,24−26,39 The term metrical
[ligand] oxidation states was coined in this context.3 This
approach is very reasonable since electron transfer to or from a
ligand leads to structural changes within the ligand. However,
ligand structural changes could also be the consequence of
back-donation from the metal to the ligand as evidenced by a
recent computational study where it was shown that both
effects (electron transfer and backbonding) would indistin-
guishably cause the same type of ligand structural changes.40,41
Backbonding, however, which serves as the classical explanation
for geometric changes of various ligands such as CO or C2H4
upon binding to a metal center, would not be expected to result
in changes of the metal oxidation state. While structural
changes associated with classical backbonding are continuous,
rather discrete ligand structural changes would be expected for
integer electron transfer; and indeed, in studies concerning
redox non-innocent ligands, stepwise changes of the ligand
structure are usually described, which lead to characteristic
bond lengths independent of the metal center involved.3,9,15,18
Ultimately, it is this concept of stepwise ligand structural
changes with the ligand oxidation state that provides the basis
for the use of ligand geometries as a measure for the ligand
charge state. On the other hand, there are examples in which
the structural features of redox non-innocent ligands appear
intermediate between the discretely charged forms, and this
observation is usually explained in terms of delocalization
phenomena.18 No matter how ligand structural features are
interpreted, in terms of backbonding or as a consequence of
electron transfer, one must keep in mind that these
interpretations are fundamentally distinct since backbonding
is continuous, whereas electron transfer in these systems is
described in terms of integer behavior. Thus, the concept of a
physical oxidation state of a metal center appears attractive and
can be helpful in descriptions of the electronic structure. Since
such electronic-structure parameters are not physically
observable, they cannot be probed by a single physical/
spectroscopic method. Rather, only a multitechnique approach
allows a comprehensive understanding of a given complex and
provides means to determine the most suitable description of
the electronic structure.
A particularly challenging system for the identification of the
metal oxidation state is the 2,6-diiminopyridine-based iron
dicarbonyl complex [Fe(CO)2LNNN] depicted in Figure
1a.24,42,43 The structural features of this formal Fe(0) complex
Figure 1. Illustration of the concept of redox non-innocent ligands: (a)
two formulations for a 2,6-bisiminopyridine-based [Fe(CO)2LNNN
]
complex (N = NAr) as reported earlier24,42 and (b) possible
formulations for an α-iminopyridine-based [Fe(CO)2LPNN] complex
(N = NAr, P = PR2) as discussed in this study.
indicate the presence of a dianionic NNN ligand,24 which
would make this complex an Fe(II) system. The spectroscopic
features and broken-symmetry (BS) calculations, however, do
not unequivocally support an Fe(II) assignment, and an Fe(0)
formulation was also considered possible.24 An α-iminopyr-
idine-based iron-PNN dicarbonyl complex such as [Fe-
(CO)2LPNN] (Figure 1b) would presumably suffer from the
same dilemma concerning the oxidation-state assignment,
though Fe(0) and Fe(I) would be the anticipated limiting
oxidation states. A comparison of the spectroscopic features of
both types of complexes should provide insight toward the
actual physical oxidation states present.
Herein, we present the synthesis and characterization of a
series of α-iminopyridine-based iron-PNN pincer complexes
[FeBr2LPNN] (1), [Fe(CO)2LPNN] (2), [Fe(CO)2LPNN](BF4)
(3), [Fe(F)(CO)2LPNN](BF4) (4), and [Fe(H)(CO)2LPNN]-
(BF4) (5), which cover formal oxidation states of 0 to +2 (LPNN
= 2-[(di-tert-butylphosphino)methyl]-6-[1-(2,4,6-
mesitylimino)ethyl]pyridine, see Figure 2).44 All complexes
are characterized by means of a variety of spectroscopic
1
techniques and other methods including H, 13C, 15N, and 31P
NMR spectroscopy, IR, Mossbauer and X-ray photoelectron
̈
spectroscopy (XPS), electron paramagnetic resonance (EPR),
and magnetic circular dichroism (MCD) spectroscopy, as well
as SQUID magnetometry, cyclic voltammetry (CV), and X-ray
crystallography. Density functional theory (DFT) calculations
are employed to investigate the electronic structure of the
complexes in further detail. Also, the 2,6-bisiminopyridine-
based complexes [FeBr2LNNN] (6) and [Fe(CO)2LNNN] (7)
were synthesized and characterized for this study (LNNN = 2,6-
bis-[1-(mesitylimino)ethyl]pyridine, see Figure 2). While metal
oxidation state assignments in these complexes based on ligand
metrical parameters would suggest redox non-innocent ligand
behavior, we demonstrate that detailed electronic-structure
studies combining multiple spectroscopic methods (i.e.,
Mossbauer, EPR, MCD, XPS, electronic absorption spectros-
̈
copy) complemented by DFT and time-dependent DFT (TD-
DFT) studies support the presence of redox-innocent ligands.
Overall, these studies demonstrate the importance of a
multitechnique approach in elucidating the electronic structure
B
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 2. Overview of the PNN, NNN, PNP, and NN ligands discussed in this study.
Figure 3. Overview of the iron complexes [FeBr2LPNN] (1), [Fe(CO)2LPNN] (2), [Fe(CO)2LPNN](BF4) (3), [Fe(F)(CO)2LPNN](BF4) (4), and
[Fe(H)(CO)2LPNN](BF4) (5) as synthesized starting from the new PNN ligand LPNN (N = NMes (Mes = 2,4,6-trimethyl phenyl), P = PtBu2).
in complexes of potentially redox non-innocent ligands as well
as the contributions of backbonding effects in such complexes.
In addition, we employed XPS for the characterization of the
electron density at the metal center, which is related to the
physical metal oxidation state. XPS, as well as electron-
spectroscopic methods in general, has the advantage of being
applicable for any element (except hydrogen), in any possible
2. RESULTS AND ANALYSIS
2.1. [FeBr2LPNN] (1). The blue, paramagnetic iron PNN
complex [FeBr2LPNN] (1) is synthesized by the reaction of
electronic state (in contrast to NMR or Mossbauer spectros-
̈
copy, which are available only for a limited set of elements and
spin states). Several decades ago, this technique was applied for
the characterization of a wide range of organometallic
complexes and materials including iron-containing species,49−54
while it was within the past decade only sparsely used for
organometallic compounds.40,55,56 Certain spectral features, so-
called shake-up lines,49,51 also allow for the distinction of
diamagnetic and paramagnetic compounds (see Supporting
Information for details). With respect to the determination of
metal oxidation states using XPS, it is independently reported
that, irrespective of the metal, a formal oxidation-state change
of one is accompanied by a change in the electron binding
energy of ∼1 eV50 (or 1.2 eV57). Note, however, that this value
is merely empirical and that it applies only if the ligand sets are
comparable. For [FeBr2LPNN] (1), the maximum of the Fe
2p3/2 peak in the XPS spectrum is located at a binding energy of
709.4 eV (entry 1 in Table 3), and the comparison with the
corresponding chloro complex [FeCl2LPNN] (1′) (709.4 eV,
entry 2 in Table 3) indicates that the electronic influence of Br
versus Cl ligands on the Fe center is comparable. This behavior
is expected and in agreement with previous studies.50
L
PNN with anhydrous FeBr2 in tetrahydrofuran (THF) at room
temperature, and this complex serves as a starting point for the
syntheses of complexes 2−5 (Figure 3).
Crystals of suitable quality for X-ray diffraction were grown
from a benzene solution of 1 at room temperature (Figure 4A).
The Fe−N1 and Fe−N2 distances of 2.208(2) and 2.171(2)
(see Table 1), respectively, are characteristic for a high-spin
(hs) complex,45,46 and the magnetic-moment measurements
confirm this assignment (Evans’ method:47 μeff = 5.4 (at 298
K); SQUID: μeff = 4.77 (at 298 K, see Supporting Information
for further details)).
For α-iminopyridine complexes, Wieghardt and co-workers
described the CN, Cimine−Cipso, and Cipso−Npy bond lengths
to be indicative of the ligand charge state, and values of 1.28,
1.47, and 1.35 Å, respectively, were reported to be characteristic
for the neutral form.17−19 For [FeBr2LPNN] (1) we found values
of 1.282(4), 1.488(5), and 1.348(4) Å (see Table 1),
respectively, which are well in agreement with a neutral
formulation. Consequently, [FeBr2LPNN] (1) is a hs-Fe(II)
complex. This assignment is fully in agreement with the
Mossbauer spectrum of 1 at 80 K, which is well-fit as a doublet
̈
with parameters of δ = 0.94 mm s−1, ΔEQ = 2.59 mm s−1
(Figure 4 and Table 2).48 Moreover, saturation magnetization
data collected at 8810 cm−1 (Figure 4C, inset) are well-
described by an S = 2 negative zero-field split ground state
model with ground state spin Hamiltonian parameters of δ =
1.4 0.1 cm−1 and g∥ = 9.0 0.2, corresponding to the ligand-
field parameters Δ = −1400 200 cm−1 and |V/2Δ| = 0.18
0.02 (where Δ = E(dxz,yz) − E(dxy) and V = E(dxz) − E(dyz)).
However, in the series [FeBr2LNNN] (6), [FeBr2LPNN] (1),
and [FeBr2LPNP] (Table 3, entries 3, 1, and 4, respectively; for
the ligands, see Figure 2), in which the imine arms of
[FeBr2LNNN] (6) are successively exchanged for phosphine
donors, a systematic decrease of the Fe 2p3/2 binding energies
by 0.6 eV in 0.3 eV steps is observed. This finding is in
agreement with the stronger electron-donating ability of
phosphines when compared with imine ligands, and XPS is
C
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 2. Summary of the 80 K 57Fe Mossbauer Parameters
̈
for Iron LPNN Complexes
complex
[FeBr2LPNN] (1)
δ in mm s−1
ΔEQ in mm s−1
0.94
0.00
2.59
1.14
0.54
0.55
1.48
[Fe(CO)2LPNN] (2)
[Fe(CO)2LPNN](BF4) (3)
[Fe(F)(CO)2LPNN](BF4) (4)
[Fe(H)(CO)2LPNN](BF4) (5)
0.12
0.04
−0.04
Table 3. Fe 2p3/2 Electron Binding Energies for the Iron
Complexes Discussed in This Study
a
entry
complex
[FeBr2LPNN] (1)
Fe 2p3/2 BE in eV
1
709.4
709.4
709.7
709.1
708.1
708.4
709.5
708.0
708.8
709.9
2
[FeCl2LPNN] (1′)
3
[FeBr2LNNN] (6)
b
4
[FeBr2LPNP]
5
[Fe(CO)2LPNN] (2)
6
[Fe(CO)2LNNN] (7)
c
7
[FeBr2LPNN‑bp]
c
8
[Fe(CO)2LPNN‑bp]
9
[Fe(CO)2LPNN](BF4) (3)
10
11
[Fe(F)(CO)2LPNN](BF4) (4)
[Fe(H)(CO)2LPNN](BF4) (5)
709.2
a
b
c
Measured by XPS. Prepared according to ref 58. Data taken from
ref 40.
oxidation state: while the former adopts integer values by
definition (and all three dibromide complexes under discussion
are formal Fe(II) complexes),37 the latter is continuous.
Spin-unrestricted DFT calculations were used to further
evaluate the electronic structure of [FeBr2LPNN] (1). Geometry
optimization at the PBEPBE/TZVP level yielded overall
structural features, bond lengths, and angles in good agreement
with those observed by crystallography with only some minor
bond elongations and contractions of the metal−ligand bonds,
particularly the Fe−LPNN bonds (solvent model). The
optimized structure of 1 is best described as a distorted square
pyramidal complex with Fe−P, Fe−N1, and Fe−N2 bond
lengths of 2.591, 2.151, and 2.103 Å, respectively, Fe−Br bond
Figure 4. Structural and spectroscopic characterization of [FeBr2LPNN
]
(1). (A) X-ray molecular structure of 1 with 50% probability thermal
ellipsoids; the hydrogen atoms and a cocrystallized benzene molecule
are omitted for clarity; selected bond lengths are given in Table 1. (B)
80 K Mossbauer spectrum of 1. (C) 5 K, 7 T NIR MCD spectrum of
̈
1. (inset) Saturation-magnetization data (dots) and best fit (lines)
collected at 8810 cm−1.
sensitive to this effect. Note, however, that this observation also
demonstrates the difference between a formal and a physical
Table 1. Selected Bond Lengths [Å] of [FeBr2LPNN] (1), [Fe(CO)2LPNN] (2), [Fe(CO)2LPNN](BF4) (3),
[Fe(F)(CO)2LPNN](BF4) (4), and [Fe(H)(CO)2LPNN](BF4) (5)
1
2
3
4
5
Fe−N1
Fe−N2
Fe−P
P−C1
C1−C2
C2−C3
C3−C4
C4−C5
2.208(2)
2.171(2)
2.5147(8)
1.833(4)
1.508(5)
1.393(5)
1.381(5)
1.389(5)
1.385(4)
1.348(4)
1.488(5)
1.282(4)
1.924(2)
1.919(2)
2.2231(6)
1.847(2)
1.491(3)
1.366(3)
1.404(3)
1.363(3)
1.411(3)
1.377(3)
1.410(3)
1.339(3)
1.974(3)
1.990(4)
2.259(1)
1.837(4)
1.498(6)
1.394(5)
1.374(6)
1.388(6)
1.389(5)
1.361(5)
1.452(6)
1.299(5)
1.957(3)
2.008(3)
2.293(1)
1.838(4)
1.489(6)
1.398(6)
1.370(6)
1.382(6)
1.382(6)
1.360(5)
1.468(6)
1.295(5)
1.967(1)
1.985(1)
2.2329(6)
1.845(2)
1.500(2)
1.397(2)
1.386(2)
1.393(2)
1.386(2)
1.357(2)
1.469(2)
1.293(2)
C5−C6
C6−N1 (Cipso−Npy)
C6−C7 (Cimine−Cipso
C7−N2 (CN)
)
D
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
lengths of 2.458 and 2.483 Å, an N1−Fe−N2 bond angle of
75.38°, and an N1−Fe−P bond angle of 75.23° (compare with
the crystal details given in Figure 4A and Table 1). The
molecular orbitals and their corresponding energies as well as
excitation energies were calculated from the optimized structure
at the B3LYP/TZVP level. Importantly, TD-DFT calculations
provide excellent agreement between calculated and exper-
imentally determined d−d and charge-transfer (CT) transitions
(from MCD and UV−vis spectra), providing further validation
of the quality of the DFT model (see Supporting Information
for details). Both the experimental and computational studies of
1 are indicative of a hs-Fe(II) complex (S = 2). A molecular
orbital energy diagram for 1 is given in the Supporting
Information.
2.2. [Fe(CO)2LPNN] (2). Reduction of the Fe(II) dibromide
complex [FeBr2LPNN] (1) with an excess of sodium amalgam
(10%) under a CO atmosphere at room temperature gives rise
to [Fe(CO)2LPNN] (2). This formally Fe(0) complex is
obtained as an intensely purple, diamagnetic solid, which is
well-soluble in all common solvents including pentane, and
crystals were grown from a saturated pentane solution at −35
°C (Figure 5).
hydrogen atoms of the mesityl group at 2.25 ppm (s, 6H) and
6.95 ppm (s, 2H), respectively. For the CO ligands, only one
doublet is observed at 221.1 ppm (d, 2JCP = 11.3 Hz) in the 13
C
NMR spectrum. Similar discrepancies in the structural
assignment upon comparison of the crystal structure and
NMR data have been observed for iron-dicarbonyl complexes
before.24,40,59 The CN, Cimine−Cipso, and Cipso−Npy bond
lengths determined for [Fe(CO)2LPNN] (2) (1.339(3),
1.410(3), and 1.377(3) Å, respectively; see Table 1) are very
similar to those reported for the monoanionic form of α-
iminopyridine ligands according to Wieghardt and co-workers,
that is, 1.34, 1.41, and 1.39 Å, respectively.17−19 On the basis of
the ligand structural data alone, [Fe(CO)2LPNN] (2) would be
assigned as an Fe(I) complex. According to Figure 1b, electron
transfer to the PNN ligand should give rise to a special
signature with respect to the imine carbon atom C7 as well as
to the imine nitrogen atom N2. While for the free PNN ligand
the imine carbon atom C7 exhibits the highest 13C chemical
shift among the quaternary carbon atoms C2, C6, and C7 (and
even overall), in [Fe(CO)2LPNN] (2), the 13C NMR resonance
of C7 is shifted particularly to higher field when compared with
C2 and C6 (see Table 5).
Table 5. Selected NMR Spectroscopic Data of the Free PNN
Ligand as well as the Complexes [Fe(CO)2LPNN] (2),
[Fe(F)(CO)2LPNN](BF4) (4), and [Fe(H)(CO)2LPNN](BF4)
(5)
a
a
a
a
PNN ligand
38.5 (s)
2
4
5
2
δ(P)
[ppm]
138.4 (s)
258.1 (s)
259.2 (s)
120.3 (s)
258.6 (s)
274.8 (s)
106.9 (d, JPF
= 9.7 Hz)
Figure 5. Molecular structure of [Fe(CO)2LPNN] (2) with 50%
probability thermal ellipsoids. The hydrogen atoms and a cocrystal-
lized pentane molecule are omitted for clarity. Selected bond lengths
are given in Table 1.
δ(N1)
[ppm]
311.5 (s)
257.6 (s)
δ(N2)
[ppm]
334.0 (s)
274.1 (s)
2
2
2
2
The crystal structure of 2 indicates a distorted square
pyramidal coordination sphere around the Fe center (OC−Fe−
CO = 96.89(9)°, Npy−Fe−CO = 100.75(8)° and 162.34(8)°).
Also, the IR spectra in both the solid state and in solution
confirm an ∼90° angle between the two CO ligands as two
absorptions in a 1:1 intensity ratio are observed (see Table 4).
However, in the NMR spectra recorded at room temper-
ature, [Fe(CO)2LPNN] (2) appears C2v symmetric with a plane
δ(C2)
[ppm]
161.2 (d, JCP
159.2 (d, JCP 162.4 (d, JCP 164.5 (d, JCP
= 14.1 Hz)
= 7.8 Hz)
= 4.0 Hz)
= 2.8 Hz)
3
3
3
δ(C6)
[ppm]
155.8 (s)
168.0 (s)
144.7 (d, JCP 153.9 (d, JCP 156.6 (d, JCP
= 5.1 Hz)
= 4.3 Hz)
= 3.5 Hz)
δ(C7)
[ppm]
144.9 (s)
173.3 (s)
179.9 (s)
a
The 15N NMR chemical shifts were determined by 15N−1H HMQC
experiments.
1
of symmetry defined by the pyridine ring: in the H NMR
t
This observation might be in line with an accumulation of
charge density on this site as would be expected according to
Figure 1b. The 15N NMR chemical shift of N2 (when
compared with those obtained for [Fe(F)(CO)2LPNN](BF4)
(4) and [Fe(H)(CO)2LPNN](BF4) (5), vide infra) is also in line
with this interpretation, and we will come back to this aspect
further below. Moreover, according to the crystal structure data
of [Fe(CO)2LPNN] (2), the pyridine ring shows alternating
bond lengths (compare the C−C bonds connecting C2−C6 in
Table 1). Even though such a behavior would rather be
expected for the dianionic form of an α-iminopyridine
ligand,17−19 this observation is in line with the shift of electron
density from the metal center to the ligand.
spectrum only one doublet is observed for the Bu groups at
3
0.98 ppm (d, 18H, JHP = 12.6 Hz), the methylene group at
3.27 ppm (d, 2H, 2JHP = 8.3 Hz), and the o-methyl and the m-
Table 4. Infrared Stretches νCO for Complexes
[Fe(CO)2LPNN] (2), [Fe(CO)2LPNN](BF4) (3),
[Fe(F)(CO)2LPNN](BF4) (4), [Fe(H)(CO)2LPNN](BF4) (5),
and [Fe(CO)2LNNN] (7) as Determined in the Solid State on
NaCl Plates
a
2
3
4
5
7
ν(COsym) [cm−1
ν(COasym) [cm−1
ν(Fe−H) [cm−1
]
1934 (1952)
1875 (1897)
1987
1922
2052
2015
2022
1973
1929
1949
1878
]
In contrast to the NMR and crystal structure data, which
suggest an Fe(I) description of [Fe(CO)2LPNN] (2) but which
reflect only the ligand properties in a direct fashion, the
]
a
Values in parentheses were determined in a pentane solution.
E
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Mossbauer data (δ = 0.00 mm s−1, ΔE = 1.14 mm s−1, Table
ligand changes have a strong influence on the apparent binding
energy.50 Therefore, to further analyze the electronic structure
of 2, spin-unrestricted BS calculations at the B3LYP/TZVP
level were completed to test the possibility of LPNN acting as a
redox-active ligand in this complex. Taking into account that 2
has a diamagnetic ground state (S = 0), three different
computational models were explored (see Supporting In-
formation for details): (i) the neutral-ligand description
corresponding to [Fe0(CO)2(LPNN)], (ii) the BS(1,1) approach
to yield the electronic ground state of [FeI(CO)2(LPNN−)]
(considering the transfer of both an α- or a β-spin electron
from Fe (now, S = 1/2) to LPNN), and (iii) the BS(2,2) case,
producing [FeII(CO)2(LPNN2−)] (considering the transfer of
two electrons from Fe to generate either S = 0 or S = 1 with the
Fe site being antiferromagnetically coupled to LPNN2− to give an
̈
Q
2) for this complex are not unequivocally in agreement with an
Fe(I) designation, since an assignment as either Fe(0) or ls-
Fe(II) (ls = low spin) is possible based on the parameters,
demonstrating the challenge in correlating metal oxidation
states with Mossbauer parameters in low-valent iron com-
̈
plexes.48 Note that very similar parameters (δ = 0.03 mm s−1,
ΔEQ = 1.17 mm s−1) were found for Chirik’s 2,6-
bisiminopyridine-based complex [Fe(CO)2L′NNN], for which
both an Fe(0) and a ls-Fe(II) formulation was considered
possible.24 In Chirik’s study, the Fe(II) interpretation was
suggested by the ligand structural features, which were
interpreted as being indicative of a dianionic ligand, while the
Fe(0) assignment was based on the Mossbauer data. Also
̈
broken-symmetry DFT calculations did not provide a more
conclusive picture, and the authors concluded later that “the
electronic structure can be considered a hybrid of the Fe(0)
and Fe(II) limiting resonance forms.”42
S
total = 0 complex). Notably, the FeI/[LPNN−] models converged
to the same result: a system with a nonzero spin density on Fe
and an electronic energy that is 0.61 kcal/mol lower than for
the Fe0/[LPNN] model, the latter of which converged to an
electronic state with a zero spin density on the Fe atom. The
FeII/[LPNN2−] models were found to depend on the iron spin
state, and electron densities consistent with either the FeI/
[LPNN−] or Fe0/[LPNN] electronic descriptions could be
obtained. Importantly, however, previous computational studies
of 2,2′-bipyridine complexes of iron, cobalt, vanadium, and
titanium using a variety of exchange-correlation (XC) func-
tionals have demonstrated that B3LYP, a hybrid XC functional,
can yield unreliable results in BS calculations of pincer
complexes of this type and that nonhybrid XC functionals
(e.g., PBE, HCTC, BLYP) give more accurate predictions.41
This previous study and the small energy differences between
the two ground-state descriptions found in our B3LYP
calculations motivated the evaluation of our three computa-
tional models across a series of XC functionals. Analogous
calculations to those described using B3LYP were performed
using BLYP, PBEPBE, HCTC, and M06L with the same basis
set, TZVP, as before. Importantly, all calculations using these
four XC functionals were found to converge to an electronic
state with a zero atomic spin density on Fe, which is consistent
with the [Fe0(CO)2(LPNN)] description with no unpaired
XPS measurements on [Fe(CO)2LPNN] (2) give rise to an Fe
2p3/2 binding energy of 708.1 eV (entry 5 in Table 3), which is
1.3 eV lower than for the formal Fe(II) dibromide complex
[FeBr2LPNN] (1) (entry 1 in Table 3). Thus, the electron
density on the iron center of [Fe(CO)2LPNN] (2) appears
considerably increased when compared with that of
[FeBr2LPNN] (1). The same binding energy difference is
found for the couple [FeBr2LNNN] (6)/[Fe(CO)2LNNN] (7)
(entries 3 and 6 in Table 3), and consequently, the iron centers
in both [Fe(CO)2LPNN] (2) and [Fe(CO)2LNNN] (7) look very
similar electronically. Note that the Fe 2p3/2 binding-energy
values determined in this study for [FeBr2LPNN] (1) and
[Fe(CO)2LPNN] (2) very much resemble those obtained earlier
in an independent investigation from our group, in which the
bipy-analogous PNN pincer complexes [FeBr2LPNN‑bp] and
[Fe(CO)2LPNN‑bp] were investigated (entries 7 and 8 in Table
3; for the ligand see Figure 2; bipy = 2,2′-bipyridine).40 Thus,
all three iron-dicarbonyl complexes [Fe(CO)2LPNN] (2),
[Fe(CO)2LNNN] (7), and [Fe(CO)2LPNN‑bp] appear to be
electronically very similar based on the XPS data, while the
crystal structures of these complexes indicate oxidation states of
Fe(I) for the PNN complexes40 and Fe(II) for the NNN
complex.24
electrons on Fe or LPNN
.
To further evaluate the electronic structure of complex 2,
DFT studies were performed. Initial geometry optimization at
the PBEPBE/TZVP level yielded overall structural features,
bond lengths, and angles in good agreement with those
observed by crystallography. The optimized structure of 2 is
best described as a distorted square pyramidal complex with
Fe−P, Fe−N1, and Fe−N2 bond lengths of 2.258, 1.944, and
1.931 Å, respectively, Fe−CO bond lengths of 1.751 and 1.756
Å, an N1−Fe−N2 bond angle of 79.80°, and an N1−Fe−P
bond angle of 82.26°. This optimized geometry correlates well
with the crystal structure (see Figure 5 and Table 1). Note also
that the bond elongations and contractions of the N1−C6,
C6−C7, and C7−N2 bonds seen in the crystal structure of 2
are reproduced in the calculated structure to within 0.015 Å
(see Table 1). As previously discussed, according to the
structural parameters complex 2 should be assigned as an Fe(I)
complex. In addition, the XPS data seem to indicate that an
Fe(I) situation might apply for 2 if the 1.3 eV difference in the
electron binding energy compared with [FeBr2LPNN] (1) (see
Table 3) is interpreted in terms of an oxidation state difference
of one. However, it is important to recall that 1 carries two
bromide ligands, while 2 contains two CO ligands, and such
Further support for the [Fe0(CO)2(LPNN)] ground-state
electronic structure description derives from TD-DFT
calculations performed for 2 for both the Fe0/[LPNN] and
FeI/[LPNN−] electronic descriptions (note: B3LYP models were
utilized as both descriptions were obtained). Notably, the TD-
DFT calculations for FeI/[LPNN−] predict an additional low-
energy d−d transition at ∼8660 cm−1 compared to the Fe0/
[LPNN] solution. Therefore, near-IR (NIR) absorption studies
were completed for 2 (see Supporting Information for details)
that showed no additional low-energy d−d transition at
<10 000 cm−1 (even at very high complex concentrations).
This experimental evidence is consistent with the assignment of
2 as a ls-Fe(0) complex with no unpaired electron on either Fe
or LPNN. While it is likely that the [Fe0(CO)2(LPNN)] and
[FeI(CO)2(LPNN−)] electronic states are, in fact, relatively close
in energy, the combined results of XPS, electronic absorption
spectroscopy, and DFT/TD-DFT studies are more consistent
with LPNN acting as a redox-innocent ligand in complex 2.
Because of these findings, all further DFT calculations were
completed using the [Fe0(CO)2(LPNN)] electronic ground-state
description. The ground-state wave function of 2 is described
by the frontier molecular orbitals (FMOs), which are shown in
F
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 6. Calculated molecular orbital energy diagram for [Fe(CO)2LPNN] (2).
Figure 6. The Fe d-orbitals are significantly mixed with CO and
LPNN character, and listed in order of increasing energy those
2
2
2
are dxy(130), dxz (132), dyz (134), dz (135), and dx −y (138). In
addition, there are FMOs that represent an occupied MO of Fe
d-, N2 σ-, and aryl π-character (133), an unoccupied MO of Fe
d-, N1N2 π*-, aryl π*-, and CO σ*-character (136), and an
unoccupied MO of aryl π*-character (137).
TD-DFT calculations were also used to assign the observed
transitions in the absorption spectrum (see Supporting
Information for details). Overall, the energies of the absorption
transitions correlate well to the TD-DFT calculated energies.
However, the d−d and CT transitions are quite mixed,
involving multiple LPNN, CO, and Fe orbitals.
2.3. [Fe(CO)2LPNN](BF4) (3). Upon one-electron oxidation
of [Fe(CO)2LPNN] (2) with 1 equiv of ferrocenium
tetrafluoroborate in THF at room temperature, [Fe-
(CO)2LPNN](BF4) (3) was isolated as a paramagnetic, green
solid. Reduction of [Fe(CO)2LPNN](BF4) (3) with sodium
amalgam (10%) in THF at room temperature gives rise to the
regeneration of [Fe(CO)2LPNN] (2) (Figure 3). This kind of
reactivity has been indicated by the CV of [Fe(CO)2LPNN] (2)
(Figure 7a), which shows three redox processes between +1.0
and −3.5 V, and an analogous behavior was observed previously
for a similar set of PNN-based iron−dicarbonyl pincer
complexes.40
The waves at E1/2 = −0.65 and −2.76 V are reversible as
indicated by separate scans for these regions. We interpret these
two waves in terms of one-electron oxidation and reduction
processes relative to [Fe(CO)2LPNN] (2) as reported by Chirik
and co-workers for the NNN-iron dicarbonyl complex
[Fe(CO)2L′NNN].42 The oxidation wave at +0.64 V, however,
is not well-behaved reversible, and both reduction features at ca.
0.00 V and −1.48 V appear only if the cycling involves the
oxidation at +0.64 V. While we have no chemical explanation
for the two different reduction events, we speculate that the
respective oxidation event involves the generation of dicationic
[Fe(CO)2LPNN]2+, which decomposes prior to electrochemical
reduction. Higher scan rates of up to 600 mV s−1 do not make
Figure 7. CVs of (a) [Fe(CO)2LPNN] (2) and (b) [Fe(CO)2LPNN]-
(BF4) (3) in THF (0.1 M [nBu4N][BF4] supporting electrolyte) at
room temperature with scan rates of 100 mV s−1 (glassy-carbon
working-electrode). Potentials are referenced relative to the Fc+/Fc
couple.
this process reversible. The fact, however, that the CVs of
[Fe(CO)2LPNN] (2) and [Fe(CO)2LPNN](BF4) (3) (Figure 7)
exhibit similar features at comparable positions confirms the
electrochemical formation of [Fe(CO)2LPNN](BF4) (3) during
cyclic voltammetry of [Fe(CO)2LPNN] (2). Complex 3 exhibits
IR absorptions at 1922 and 1987 cm−1 in a 1:1 ratio, which are
indicative of the presence of two CO ligands with an ∼90°
angle between them. The higher wavenumbers compared with
[Fe(CO)2LPNN] (2) (see Table 4) are consistent with the
reduction of electron density on the metal center upon one-
electron oxidation, and the change of ca. 50 cm−1 is similar to
that reported for the analogous NNN-ligated system described
by Chirik and co-workers.42 Further, the magnetic measure-
ments (Evans’ method: μeff = 2.1 (at 298 K); SQUID: μeff
=
2.14 (at 298 K, see Supporting Information for further details))
are in agreement with a ls-Fe(I) assignment. The Mossbauer
̈
parameters of 3 at 80 K (δ = 0.12 mm s−1, ΔEQ = 0.54 mm s−1,
G
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Table 2)48 are also similar to those reported by Chirik and co-
electron-transfer processes. Rather, backbonding phenomena
might account for this observation (vide infra).41
To further characterize the electronic structure of 3, EPR and
MCD spectroscopy were utilized. The 10 K EPR spectrum of 3
shown in Figure 9A clearly depicts a rhombic g ≈ 2 signal (S =
workers for their [Fe(CO)2L′NNN](BArF ) complex, which was
4
assigned as ls-Fe(I).42 Further evidence for an Fe(I) assignment
of 3 is provided by the Fe 2p3/2 binding energy of 708.8 eV
(entry 9 in Table 3), as determined via XPS, which is 0.7 eV
higher than of the parent compound [Fe(CO)2LPNN] (2), the
latter of which we assigned to be an Fe(0) complex. Single
crystals of [Fe(CO)2LPNN](BF4) (3) were grown by diffusion
of pentane into a THF solution at room temperature (Figure
8).
Figure 8. Molecular structure of [Fe(CO)2LPNN](BF4) (3) with 50%
−
probability thermal ellipsoids. The hydrogen atoms, the BF4
counterion, and a cocrystallized THF molecule are omitted for clarity.
Selected bond lengths are given in Table 1.
While the CN, Cimine−Cipso, and Cipso−Npy parameters for
[FeBr2LPNN] (1) and [Fe(CO)2LPNN] (2) are indicative of the
presence of a neutral and a monoanionic α-iminopyridine
ligand, respectively, for [Fe(CO)2LPNN](BF4) (3) the values are
intermediate between the two descriptions (see Table 1). While
other complexes are known in which only one α-iminopyridine
ligand is bound to the metal center,60−62 [Fe(CO)2LPNN](BF4)
(3) is, to our knowledge, the first structurally characterized
complex with a single α-iminopyridine ligand bound to the
metal center, in which the structural parameters are between
the neutral and the anionic description. Wieghardt and co-
workers previously described an α-iminopyridine complex of
iron, [Fe(LNN)2]+ (L = 2,6-diisopropyl-N-(pyridin-2-
ylmethylene)aniline, Figure 2), in which the two bidentate
NN ligands exhibit intermediate structural features.18 This
observation, however, was interpreted in terms of electron
delocalization: the iron oxidation state in [Fe(LNN)2]+ was
assigned to be Fe(II), while one negative charge was assumed
to be distributed over the two LNN ligands. Such an
interpretation, however, can be clearly excluded for [Fe-
(CO)2LPNN](BF4) (3), which contains only one α-iminopyr-
idine unit. While a critical reader might argue that the present
PNN ligand is not a genuine α-iminopyridine ligand and that
the additional phosphorus arm might be redox non-innocent
itself, this option can be excluded since in all three complexes 1,
2, and 3, both the P−C1 and the C1−C2 bond lengths are
essentially unaltered (Table 1); the additional phosphorus
donor thus acts as a mere spectator. The fact that in
[Fe(CO)2LPNN](BF4) (3) the α-iminopyridine ligand adopts
a geometry that is between those expected for a neutral and a
monoanionic ligand rules out the possibility of the assignment
of an integer charge state to the ligand and, thus, questions the
strict interpretation of ligand structural changes in terms of
Figure 9. EPR and MCD spectroscopic characterization of [Fe-
(CO)2LPNN](BF4) (3). (A) 10 K EPR spectrum. (B) 5 K, 7 T NIR
MCD spectrum. (inset) Saturation-magnetization data (dots) and best
fit (lines) collected at 13 333 cm−1. C) 5 K, 7 T UV−vis MCD
spectrum. Peak fits are shown for the MCD spectra as dashed lines.
EPR parameters are given in the text.
1/2) that is hyperfine split by the 31P ligand (I = 1/2) of the
LPNN moiety. The spectrum was simulated using the following
parameters: g = [2.003, 2.044, 2.088] [gavg = 2.045] and A(P) =
[43, 40, 38] MHz [Aavg = 40.3 MHz] (see Supporting
Information). The simulation is in excellent agreement with the
experimental EPR spectrum and consistent with the assignment
of 3 as a ls-Fe(I) complex. In general, a qualitative indication of
the amount of spin density on a metal can be roughly estimated
from the deviation of the experimental Δg value (where Δg =
gmax − gmin) from the free-electron isotropic g value of 2.0023 in
a frozen solution.63 Typically, large deviations correspond to a
significant spin density on the metal.64 In this system, a Δg
value of 0.085 is obtained, which suggests significant metallo-
radical character being present, consistent with an unpaired
electron on Fe. In Figure 9, the 5 K, 7 T NIR (Figure 9B) and
UV−vis MCD spectra (Figure 9C) of 3 are shown. The
saturation magnetization data collected at 13 333 cm−1 (Figure
H
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
Figure 10. Calculated molecular orbital energy diagram for [Fe(CO)2LPNN](BF4) (3).
9B, inset) are well-described by an S = 1/2 ground-state model
with g = [2.003, 2.044, 2.088] and (Mxy, Mxz, Myz) = (0.420,
0.513, 1.022). The fit gives excellent agreement with the
variable-temperature, variable-field (VTVH) MCD data. The
transitions seen in both the NIR and UV−vis MCD spectra are
fully assigned using TD-DFT (see Supporting Information).
Thus, both the EPR and MCD studies support the assignment
of 3 as a ls-Fe(I) (S = 1/2) complex.
major contributions to bonding. The MOs and the
corresponding energy diagram are shown in Figure 10. In the
α-spin manifold, the highest occupied molecular orbital along
with the occupied α-130, α-131, and α-132, as well as the
unoccupied α-138, are comprised mostly of Fe d-character,
slightly mixed with LPNN and CO character. The Fe d-orbitals
listed in order of increasing energy are dxy(α-130), dxz (α-131),
dyz (α-132), dz (α-135), and dx −y (α-138). In addition, there
are FMOs that represent an occupied MO of aryl π-character
(α-133), an unoccupied MO of N1N2 π*-, CO σ*-, and aryl
π*-character (α-136), and an unoccupied MO of PN2 σ*- and
CO π*-character (α-140).
2
2
2
Spin-unrestricted DFT calculations were used to further
analyze the electronic structure of [Fe(CO)2LPNN](BF4) (3).
Initial geometry optimization with PBEPBE/TZVP yielded
overall structural features, bond lengths, and angles in good
agreement with those observed by crystallography (see
Supporting Information for details). BS calculations were
performed to test the possibility of LPNN acting as a redox-active
ligand in this complex. Three separate calculations on the
[FeI(CO)2(LPNN)], [FeII(CO)2(LPNN−)] (α-spin electron from
Fe → LPNN), and [FeII(CO)2(LPNN−)] (β-spin electron from Fe
→ LPNN) electron ground-state descriptions converged to an
Stotal = 1/2 spin as found experimentally (see Supporting
Information for details). Because of the sensitivity of these
calculations toward the employed DFT functional (as seen
previously for complex 2), the B3LYP, BLYP, and HCTC
functionals were utilized. Notably, all calculations, regardless of
the functional, converged to the same electronic state with the
unpaired electron density on Fe (MPA- and NPA-derived
values of 0.86 and 0.83, respectively, from the B3LYP
calculations) consistent with an [FeI(CO)2(LPNN)] description
of complex 3 with LPNN being redox-innocent. Furthermore,
TD-DFT studies using this electronic structure description
yield excellent agreement between the calculated and the
experimentally determined d−d and CT transitions as observed
in MCD and absorption spectroscopies, thus further supporting
the DFT results (see Supporting Information).
2.4. [Fe(F)(CO)2LPNN](BF4) (4). Upon treatment of [Fe-
(CO)2LPNN] (2) with 2 equiv of ferrocenium tetrafluoroborate,
the iron−fluoride complex [Fe(F)(CO)2LPNN](BF4) (4) was
obtained, and this result is likely the consequence of B−F bond
cleavage of the BF4− counterion.65,66 The diamagnetic complex
4 is characterized by a 19F NMR shift of −428.5 ppm (d, 2JFP
=
9.9 Hz), and the presence of the F− ligand is also evidenced by
the coupling observed in the 31P NMR spectrum (δ = 106.9
2
ppm, d, JPF = 9.7 Hz, Table 5). Moreover, in the XPS spectra
of this compound two different features for fluorine are found,
−
one centered at ca. 685.5 eV, which is common for all BF4
containing complexes under study, that is, 3, 4, and 5, and one
feature centered at ca. 682.9 eV, which is unique for this
complex and corresponds to the iron-bound F ligand. The IR
stretches of 2015 and 2052 cm−1 in an ∼1:1 ratio confirm the
presence of two CO ligands with an ∼90° angle between each
other. Moreover, in the 13C NMR spectrum two separate sets
of signals centered at 208.1 ppm (dd, 2JCP = 7.0 Hz, 2JCF = 22.4
2
2
Hz) and 209.9 ppm (dd, JCP = 18.5 Hz, JCF = 35.9 Hz) are
evidence for the presence of two distinguishable CO ligands,
which must be consequently coordinated in a cis arrangement,
1
and this interpretation is in agreement with the H and 13C
The ground-state wave function of 3 is described by the
FMOs with focus on the MOs in the α-spin manifold, in
conjunction with their β-spin counterparts, to describe the
NMR data as well as the crystal structure of [Fe(F)-
(CO)2LPNN](BF4) (4) (Figure 11a).
I
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(see Table 4) indicate a cis arrangement of two CO ligands.
The presence of two chemically inequivalent CO ligands is
clearly indicated by the two different sets of resonances in the
2
13C NMR spectrum at 205.0 ppm (d, JCP = 7.8 Hz, CO) and
2
1
212.0 ppm (d, JCP = 19.3 Hz, CO), and in the H NMR
spectrum the two sides of the molecule (as defined by the
pyridine-ring plane) can be distinguished based on the
t
assignment of the Bu groups, the methylene protons, and the
shifts of the mesitylene unit (see Experimental Section). The
crystal structure (Figure 11b) is further in agreement with the
hydride being in apical position. The Mossbauer data at 80 K (δ
̈
= −0.04 mm s−1, ΔEQ = 1.48 mm s−1, Table 2) are in
agreement with a ls-Fe(II) interpretation of this complex.48 In
the XPS experiments, however, [Fe(H)(CO)2LPNN](BF4) (5)
exhibits a 0.7 eV lower Fe 2p3/2 binding energy than
[Fe(F)(CO)2LPNN](BF4) (4), the latter of which is clearly an
Fe(II) compound. The difference in the iron electron-binding
energies of 4 and 5 is most likely the result of the hydride
ligand carrying less electron density compared to the fluoride
ligand in complex 4. This also makes sense chemically as the
hydride complex 5 readily reacts with various bases such as
KOtBu or KHMDS to regenerate the parent compound
[Fe(CO)2LPNN] (2); thus, the hydride ligand in 5 has quite a
protic character, and this observation is reflected in the XPS
data as well. Moreover, the CO stretches in the IR spectrum of
the hydride complex 5 are in agreement with this interpretation
(see Table 4) as they are by ca. 35 cm−1 lower than for the
corresponding fluoride complex 4, thus indicating increased
electron density at the metal center. As was observed for
[Fe(F)(CO)2LPNN](BF4) (4), the 13C NMR chemical shift of
the imine carbon atom C7 of 5 as well as the crystal structure
data are essentially in agreement with the presence of a neutral
PNN ligand. Consequently, the difference in electron density
on the Fe centers in 4 and 5 reflects fluoride versus hydride
ligation and is not a consequence of redox non-innocence of
the PNN ligand. Lastly, spin-restricted DFT calculations of 5
using B3LYP are indicative of a ls-Fe(II) complex (S = 0; see
Supporting Information for details).
Figure 11. Molecular structures of (a) [Fe(F)(CO)2LPNN](BF4) (4)
and (b) [Fe(H)(CO)2LPNN](BF4) (5) with 50% probability thermal
ellipsoids. The hydrogen atoms and the BF4 counter are omitted for
−
clarity. Selected bond lengths are given in Table 1
While the Mossbauer parameters of 4 at 80 K (δ = 0.04 mm
̈
s−1, ΔEQ = 0.55 mm s−1, Table 2) are in the range for a ls-
Fe(II) complex,48 they are also similar to the parameters
determined for 2, assigned as an Fe(0) complex. A ls-Fe(II)
assignment of 4, however, is consistent with the blue shift of the
IR stretches relative to 2 and 3, which indicates decreased
electron density at the Fe center (see Table 4). Moreover, the
XPS measurements yield an Fe 2p3/2 electron binding energy of
709.9 eV (entry 10 in Table 3), which is 1.1 eV higher than for
the Fe(I) complex [Fe(CO)2LPNN](BF4) (3). This observation
is consistent with the assignment of [Fe(F)(CO)2LPNN](BF4)
(4) as an Fe(II) complex with a neutral PNN ligand. Note that
the 13C NMR chemical shift of the imine carbon atom C7 is the
highest among all but the carbonyl carbon-atom shifts. This
observation is in line with the interpretation of a neutral PNN
ligand. The CN, Cimine−Cipso, and Cipso−Npy bond lengths in
complex 4 are intermediate between the values for the hs-
Fe(II) complex [FeBr2LPNN] (1) and the ls-Fe(I) complex
[Fe(CO)2LPNN](BF4) (3) (see Table 1). These parameters
cannot be interpreted within the framework of the “metrical
oxidation state” concept in a straightforward fashion. While a
minor shift of electron density from the metal to the ligand
might be indicated, the fact that the change does not come in
an integer fashion is strongly indicative of a backbonding effect.
In addition, DFT calculations using B3LYP, BLYP, PBEPBE,
and HCTC functionals on 5 are indicative of a ls-Fe(II)
complex (S = 0), consistent with the experimental data (see
Supporting Information for details).
3. DISCUSSION
The determination of ligand redox non-innocence has become
an important aspect of coordination chemistry due to the
importance of redox non-innocent complexes in catalysis.
Toward this goal, the use of ligand structural parameters to
identify systems that feature electron transfer to ligands (i.e.,
redox non-innocence), as demonstrated by Wieghardt and co-
workers, represents a simple, structure-based method for this
determination but is not unequivocal. Alternatively, studies
combining spectroscopic methods aimed at the direct
determination of the electronic structure of the complex as
well as the metal electron density combined with DFT studies
can permit the determination of ligand redox non-innocence,
though such approaches are more time-consuming and require
specialized instrumentation and expertise. In the present study,
both approaches were utilized to evaluate electronic structure
and ligand redox non-innocence in a series of PNN-ligated iron
complexes ranging in formal oxidation state from Fe(0) to
Fe(II), including possible backbonding versus electron-transfer
effects. While ligand reduction as a consequence of electron
transfer would result in changes of the ligand geometry, the
ligand would be negatively charged. By contrast, backbonding
from the metal center to a ligand can serve as a compensation
mechanism to avoid the accumulation of negative charge
2.5. [Fe(H)(CO)2LPNN](BF4) (5). The reaction of [Fe-
(CO)2LPNN] (2) with an excess of HBF4·Et2O in THF at
room temperature gives rise to the formation of [Fe(H)-
(CO)2LPNN](BF4) (5). This diamagnetic, formal Fe(II)
1
complex is characterized by a hydride resonance in the H
NMR spectrum centered at −4.59 ppm (d, 1H, 3JHP = 58.2 Hz,
Fe−H). The IR stretches of 1973 and 2022 cm−1 in a 1:1 ratio
J
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
density on the metal center when combined with donation
from a ligand to the metal. While in a completely balanced
situation the ligand would be neutral, its structure can be
influenced since donation and back-donation will involve
different ligand orbitals.
For the iron−dibromide complex [FeBr2LPNN] (1), ligand
structural parameters indicate the presence of a neutral α-
iminopyridine ligand (see Figure 12).17−19 Thus, 1 would be
significant Fe(I)-PNN back-donation in 3 might also result in
the observed ligand structural variations. Changes in the
populations of the occupied and unoccupied fragment orbitals
of the LPNN ligand, when the metal−ligand electronic
interactions are “turned on”, can be utilized to quantitate
donation and back-donation in this series of CO-ligated
complexes (Table 6).67
Table 6. Donation and Back Donation between the LPNN and
the Metal Fragments in [Fe(X)(CO)2LPNN]
0/+ (X = empty, F,
H) Complexes Calculated from Changes in Populations of
Occupied and Unoccupied Fragment Orbitals of the Ligand
total donation
(α + β)
total backdonation
complex
(α + β)
[Fe(CO)2LPNN] (2)
1.28 e−
1.34 e−
1.51 e−
1.46 e−
1.14 e−
0.43 e−
0.27 e−
0.34 e−
[Fe(CO)2LPNN]+ (3)
[Fe(F)(CO)2LPNN]+ (4)
[Fe(H)(CO)2LPNN]+ (5)
From this analysis, back-donation increases on going from
the ls-Fe(II) complexes 4 and 5 to the Fe(I) complex 3. This
trend is consistent with increased back-donation contributing to
ligand structural changes as the ligand acceptor orbital is a π*-
orbital of LPNN that is antibonding with respect to the N1−C6
and N2−C7 bond and bonding with respect to C6−C7 bond
(vide infra).
Figure 12. Comparison of the CN (blue), Cimine−Cipso (red), and
Cipso−Npy (green) bond lengths for complexes 1−5 as determined via
X-ray crystallography together with the expected values for the neutral
and the monoanionic form of α-iminopyridine ligands when bound to
a metal center according to Wieghardt and co-workers.17−19
assigned as an Fe(II) complex, and Mossbauer, MCD, XPS, and
Lastly, for [Fe(CO)2LPNN] (2) the ligand structural features
are very close to those of an anionic α-iminopyridine ligand
(Figure 12), and thus, 2 would be assigned as an Fe(I) complex
based on the ligand structural parameter method. The
accumulation of electron density on the ligand is further
indicated by NMR spectroscopy: While for the free PNN ligand
as well as for the ls-Fe(II) complexes 4 and 5 the imine carbon
atom C7 exhibits the highest 13C chemical shift among the
quaternary carbon atoms C2, C6, and C7, in complex 2, the 13C
NMR resonance of C7 is shifted particularly to higher field
(Table 5), which might be expected according to Figure 1b if
electron density is shifted to the ligand. Further, the 15N
chemical shift of N2 is nearly identical for complexes 4 and 5,
while it is shifted upfield by ca. 15 ppm, indicating that N2
might have some amide character68 as expected according to
Figure 1b. Notably, the 15N chemical shifts of N1 are within the
same range for 2, 4, and 5, and this observation indicates that, if
present, non-innocence does not influence the character of the
pyridine nitrogen atom.69 While both the crystallographic and
̈
DFT studies as well as magnetic measurements further support
the assignment of 1 as an hs Fe(II) complex with a redox-
innocent PNN ligand. Consistency between the structural and
spectroscopic/theoretical methods for the determination of
ligand redox non-innocence are also found for the formal
Fe(II) complexes [Fe(F)(CO)2LPNN](BF4) (4) and [Fe(H)-
(CO)2LPNN](BF4) (5). For these complexes, the ligand
structures are near to those expected for neutral α-
iminopyridine ligands (Figure 12), consistent with the
assignment of 4 and 5 as ls-Fe(II) complexes. Electronic
absorption spectroscopic studies (the latter assigned using TD-
DFT) combined with DFT studies also support the assign-
ments of 4 and 5 as ls-Fe(II) complexes carrying redox-
innocent PNN ligands, while according to IR spectroscopy and
XPS, the electron density on the iron center is considerably
lower for 4 than for 5 due to the highly electron-accepting
fluoride ligand. Still, for the hs- and ls-Fe(II) LPNN complexes,
clearly the ligand-structural parameter approach to the
determination of ligand redox non-innocence provides a simple
method consistent with more extensive electronic-structure
methods.
the NMR data support an Fe(I) formulation of [Fe(CO)2LPNN
]
(2), it is important to note that both techniques provide only
indirect evidence for the metal oxidation state. XPS offers the
possibility to estimate the oxidation state of 2 in a more direct
manner. Going from [Fe(CO)2LPNN] (2) via [Fe(CO)2LPNN]-
(BF4) (3) to [Fe(F)(CO)2LPNN](BF4) (4) (entries 5, 9, and 10
in Table 3), that is, from a formal Fe(0) via a formal Fe(I) to a
formal Fe(II) compound, steps of 0.7 and 1.1 eV separate the
respective species, and [Fe(CO)2LPNN] (2) corresponds to the
most reduced compound; this fact is also reflected in the IR
stretches of the three compounds (Table 4). Since it has been
demonstrated that [Fe(CO)2LPNN](BF4) (3) and [Fe(F)-
(CO)2LPNN](BF4) (4) are ls-Fe(I) and ls-Fe(II) complexes,
respectively, and considering that ∼1 eV steps separate integer
oxidation states for similar complexes,50 [Fe(CO)2LPNN] (2)
would be most consistent with having an Fe(0) oxidation state,
as supported by XPS and IR measurements. This assignment is
In contrast to the Fe(II)-PNN complexes, the ligand
structural features of [Fe(CO)2LPNN](BF4) (3) are between
those of the neutral and the anionic descriptions (see Figure
12), and consequently one might assign a charge state of −0.5
to the PNN ligand based solely on the ligand structural
parameters.3 This interpretation, however, is in contrast to the
expectations for electron transfer being responsible for the
structural changes as near integer charges would be observed.
Thus, for 3 the simple ligand structural parameter method for
the determination of ligand redox non-innocence is insufficient.
While detailed spectroscopic and theoretical studies support an
Fe(I) species (with a neutral PNN ligand), the origin of the
ligand structural changes remain undefined in the absence of
electron transfer (i.e., ligand redox non-innocence). However,
K
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
also consistent with TD-DFT studies combined with electronic-
absorption studies, which further support Fe(0) as the
oxidation state in this complex being the most consistent
with the experimental data. However, BS DFT calculations
clearly indicate that the Fe(I) state is close in energy to the
Fe(0) state. Note that a direct correlation between the XPS Fe
2p3/2 binding energy and ν(CO) in the IR spectra was observed
for the iron-dicarbonyl complexes under study (see Figure 13).
Figure 14. Fragment orbital analysis of back-donation in [Fe-
(CO)2LPNN] (2) including the lowest unoccupied fragment orbital
on LPNN that is the acceptor orbital for back-donation. The signed
changes (%) in the populations for these fragment orbitals are shown,
and the negative changes indicate the loss of the electron density from
a given orbital. 100% change in population corresponds to a complete
transfer of an electron from an orbital.
Figure 13. Correlation of ν(CO) and the XPS Fe 2p3/2 binding
energies as determined in this study for complexes 2−5 and 7; the
values for [Fe(CO)2LPNN‑bp] are taken from ref 40. The squares
correspond to ν(COsym) (trendline with R2 = 0.91), while the circles
indicate ν(COasym) (trendline with R2 = 0.95).
This graph also highlights the discrepancies between the formal
and the physical oxidation state description: the two formal
Fe(II) complexes 4 and 5, even though clearly identified as ls-
Fe(II) complexes by spectroscopic methods, DFT calculations,
and ligand structural parameters, bear a considerably different
electron density on the Fe centers according to IR and XPS
data.
As with complex 3, the significant changes in the ligand
structural parameters of 2 in the absence of electron transfer to
the ligand can be related to significant back-donation in 2.
is the erroneous assignment of ‘unusual’ oxidation states to
metals in coordination compounds”30 is certainly valid, the use
of ligand structural data alone for the determination of the
metal oxidation state can also lead to erroneous interpretation
of the electronic structure of transition-metal complexes. This is
demonstrated in this study on a series of α-iminopyridine-based
PNN-iron pincer complexes with formal oxidation states
ranging from Fe(0) to Fe(II). While the crystallographic data
as well as the NMR data suggest the formal Fe(0) complex
[Fe(CO)2LPNN] (2) to be an Fe(I) complex, physical
measurements on the metal center combined with DFT and
TD-DFT studies are more consistent with an Fe(0) center and
a redox-innocent PNN ligand. This interpretation is further in
line with an earlier statement that in complexes of the form
[FeI(CO)2(LPNN1−)] both the Fe(I) oxidation state and the
radical ligand represent two energetically unfavorable situations
assembled in one molecule.41 We interpret the observed
structural changes in terms of electron back-bonding rather
than being a consequence of electron transfer. The potential
importance of back-bonding contributions is particularly
influenced by the finding that in [Fe(CO)2LPNN](BF4) (3)
the PNN ligand adopts a geometry that is between the neutral
and monoanionic forms combined with the significant back-
donation of electron density in 2 and 3 as determined by charge
donation analysis. Importantly, the PNN π*-acceptor orbital for
back-donation is such that the bond elongations observed could
result from back-donation into this orbital. Overall, since a
variety of factors such as the identity of the metal center and
the actual ligand set determines if ligand structural changes are
the consequence of back-bonding or electron transfer, it is not
surprising that ligand structural parameters alone may be
insufficient in many complexes for the determination of the
ligand charge state (and, thus, the metal oxidation state). In
From the charge donation analyses (Table 6), [Fe(CO)2LPNN
(2) exhibits a large amount of Fe → LPNN back-donation (1.14
e−) compared to the [Fe(X)(CO)2LPNN 0/+ (X = empty, F, H)
]
]
species. Fragment molecular orbital (FO) analysis of back-
donation in 2 provides further insight into the origin of ligand
structural changes as a result of back-donation. While electronic
polarization effects are present, similar occupancy changes
(∼42%) are observed in both the α- and β-spin manifolds
(Figure 14), and the LPNN acceptor orbital for both is the
lowest unoccupied fragment orbital of LPNN (note that this is
also the ligand acceptor orbital for back-donation in 3).
Importantly, this π* orbital is antibonding with respect to the
N1−C6 and the N2−C7 bond and bonding with respect to C6−
C7. Hence, significant back-donation into this ligand orbital
contributes to elongations of the N1−C6 and N2−C7 bonds and
contraction of the C6−C7 bond as is observed crystallo-
graphically. Thus, significant back-donation can contribute to
similar ligand bond changes as occurs with electron transfer
(i.e., ligand redox non-innocence).
4. CONCLUSIONS
While Kaim’s statement that “an occasional consequence from
the presence of unrecognized non-innocently behaving ligands
L
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
measured in the temperature range of 5 ≤ T ≤ 300 K at external
magnetic field strength of H = 5000 Oe, along with sample cooling and
heating. Temperature steps of ΔT = 2 K were chosen. Two
measurements were performed at each (T, H) point to determine
the standard deviation of the resulting value. Field-dependent
measurements on M(H) were performed at T = 5 K and −1000 Oe
≤ H ≤ 10 000 Oe to check/confirm the paramagnetic behavior of the
samples. The measurements were performed from 0 to 10 000 Oe with
step ΔH = 500 Oe and from 10 000 Oe to −1000 Oe with ΔH = 1000
Oe. The molar magnetic susceptibility χmol was corrected by the sum
of the Pascal coefficients calculated according to composition of the
samples.73 In the following graphs the paramagnetic contribution to
the molar susceptibility is denoted as χ.
5.5. X-ray Photoelectron Spectroscopy. Powder samples were
loaded to the XPS instrument via a glovebox, which was purged for
several hours with N2. XPS measurements were performed with a
Kratos AXIS ULTRA system using a monochromatized Al Kα X-ray
source (hν = 1486.6 eV) at 75 W and a detection pass energy of 20 eV.
A low-energy electron-flood gun was applied for charge neutralization.
To define binding energies (BE) of different elements, the C 1s line at
284.8 eV was taken as a reference.74 Curve-fitting analysis was based
on linear or Shirley background subtraction and application of
Gaussian−Lorentzian line shapes.
5.6. Electron Paramagnetic Resonance Spectroscopy. All
samples for EPR spectroscopy were prepared in an inert atmosphere
glovebox equipped with a liquid nitrogen fill port to enable sample
freezing to 77 K within the glovebox. EPR samples were prepared in 4
mm OD Suprasil quartz EPR tubes from Wilmad Labglass. All samples
for EPR spectroscopy were 0.95 mM in iron. X-band EPR spectra were
recorded on a Bruker EMXplus spectrometer equipped with a 4119HS
cavity and an Oxford ESR-900 helium flow cryostat. The instrumental
parameters employed were as follows: power: 0.001 262 mW; time
constant: 40.96 ms; modulation amplitude: 8 G; experimental
frequency: 9.381 GHz; modulation frequency 100 kHz.
5.7. Magnetic Circular Dichroism Spectroscopy. All samples
for MCD spectroscopy were prepared in an inert atmosphere glovebox
equipped with a liquid nitrogen fill port to enable sample freezing to
77 K within the glovebox. MCD samples were prepared in 1:1 (v/v)
THF/2-methylTHF (to form low-temperature optical glasses) in
copper cells fitted with quartz disks and a 3 mm gasket. Low-
temperature MCD experiments were conducted using two Jasco
spectropolarimeters. Both instruments utilize a modified sample
compartment incorporating focusing optics and an Oxford Instru-
ments SM4000−7T superconducting magnet/cryostat. This setup
permits measurements from 1.6 to 290 K with magnetic fields up to 7
T. A calibrated Cernox sensor directly inserted in the copper sample
holder is used to measure the temperature at the sample to 0.001 K.
UV−visible (UV−vis) MCD spectra were collected using a Jasco J-715
spectropolarimeter and a shielded S-20 photomultiplier tube. Near-
infrared (NIR) MCD spectra were collected with a Jasco J-730
spectropolarimeter and a liquid nitrogen cooled InSb detector. The
range accessible with this NIR MCD setup is 2000−600 nm. All MCD
spectra were baseline-corrected against zero-field scans. VTVH-MCD
spectra were analyzed using previously reported fitting procedures.75,76
For VTVH-MCD fitting, both negative and positive zero-field splitting
models were evaluated. The reported error bars were determined via
evaluation of variations of the fit parameters on the quality of the
overall fit. D and |E/D| values were obtained directly from the fit
parameters using the relationships E = (δ/6)+1/3[(δ2/2) + δEs]1/2
and −D = E +(Es/3) − (δ/6) for S = 2 as previously described.75,76
5.8. Density Functional Theory Experimental Details. Spin-
unrestricted DFT calculations were performed with the Gaussian 09
package.77 All geometry-optimization calculations were performed
with the PBEPBE exchange-correlation functional with the TZVP78
basis set on all atoms with the inclusion of solvation effects using the
polarized continuum model with THF as the solvent.79 The
geometries of all complexes were fully optimized starting from X-ray
crystal structures with initial optimizations performed with CEP-4G
before optimizing with the TZVP basis set. All optimized geometries
have positive harmonic frequencies (confirming the calculated
these cases, a multitechnique approach combining physical
methods that directly probes the electronic structure of the
metal site (including XPS) combined with DFT and TD-DFT
studies provides a more rigorous method for the determination
of electronic structure and potential ligand redox non-
innocence.
5. EXPERIMENTAL SECTION
5.1. General Considerations. All reactions were performed under
a nitrogen atmosphere in a glovebox or using standard Schlenk
techniques. All solvents were reagent grade or better. THF, 1,4-
dioxane, diethyl ether, benzene, and pentane were refluxed over
sodium and distilled under a nitrogen atmosphere. Isopropyl alcohol
and triethylamine were refluxed over calcium hydride and distilled
under a nitrogen atmosphere. Methylene chloride (DCM) as well as
the deuterated solvents were purged with argon and stored in the
glovebox over 3 Å molecular sieves. All commercially available reagents
were used as received. NMR spectra were recorded using Bruker
Avance III 300, Avance III 400, and Avance 500 spectrometers.
Chemical shifts were referenced to the residual solvent peaks (1H)70
and to the generally accepted standards of CDCl3 at 77.0 ppm, C6D6 at
128.0 ppm, and CD2Cl2 at 53.8 ppm (13C), as well as to the external
standards of phosphoric acid (85% solution in D2O) at 0.0 ppm (31P)
and neat perfluorobenzene at −164.9 ppm (19F). Chemical shifts are
reported in parts per million, and coupling constants (J) are reported
1
1
in hertz. NMR assignments were assisted by H−1H-COSY, H−31P-
HMQC, 1H−13C-HSQC, 1H−13C-HMBC, and 13C-DEPTQ NMR
spectroscopy, as required. 15N NMR chemical shifts were identified by
1H−15N-HMQC NMR measurements and are reported downfield
from liquid ammonia (0.0 ppm). The effective magnetic moments in
solution were measured by Evans’ method at ambient temperature.47
IR spectra were recorded on a Nicolet FT-IR spectrophotometer.
UV−vis absorption measurements were performed on a Cary-5000
UV−vis−NIR spectrometer (Varian). Samples were measured in 1 × 1
cm quartz cuvettes. Electrospray ionization mass spectrometry (ESI-
MS) spectra were recorded on a Micromass ZQ V4.1 by the Chemical
Research Support Unit of the Weizmann Institute of Science.
Elemental analyses were performed on a Thermo Finnigan Italia
SpA (Flash EA 1112) C, H, N elemental analyzer by the Chemical
Research Support Unit of the Weizmann Institute of Science. Note
that some complexes gave unsatisfactory carbon analyses but
acceptable hydrogen and nitrogen content because of a combustion
problem due to the tetrafluoroborate anion,71 and discrepancies of
similar magnitude have been previously reported in such complexes.72
5.2. X-ray Structure Determinations. Crystal data were
measured at 100 K on a Bruker Kappa Apex-II CCD diffractometer
equipped with [λ(Mo Kα) = 0.710 73 Å] radiation, a graphite
monochromator, and MiraCol optics. The data were processed with
APEX2 collect package programs. Structures were solved by the
AUTOSTRUCTURE module and refined with full-matrix least-
squares refinement based on F2 with SHELXL-97 or SHELXL-2013.
Full details can be found in the CIF files and in the Supporting
Information.
57
5.3. Fe Mossbauer Spectroscopy. 57Fe Mossbauer spectro-
̈
̈
scopic data were collected on nonenriched samples of the as-isolated
complexes. All samples were prepared in an inert-atmosphere glovebox
equipped with a liquid nitrogen fill port to enable sample freezing to
77 K within the glovebox. Each sample was loaded into a Delrin
Mossbauer sample cup for measurements and loaded under liquid
̈
nitrogen. Low-temperature 57Fe Mossbauer measurements were
̈
performed using a SEE Co. MS4 Mossbauer spectrometer integrated
with a Janis SVT-400T He/N2 cryostat for measurements at 80 K with
̈
a 0.07 T applied magnetic field. Isomer shifts were determined relative
to α-Fe at 298 K. All Mossbauer spectra were fit using the program
̈
WMoss (SEE Co.).
5.4. SQUID Magnetometry. SQUID measurements were
performed on a Quantum Design MPMS XL-5 SQUID magneto-
meter. The powdered samples were placed in gelatin capsules. The
temperature dependence of the magnetic moment M(T) was
M
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
structures as electronic energy minima). Further calculations of
molecular orbitals (MOs) and TD-DFT used the B3LYP functional
with the TZVP basis set on all atoms. The analyses of the MO
compositions in terms of fragment orbitals, Mayer bond orders, and
the analysis of charge donation67 were performed using the AOMix
program.80,81 Atomic charges and spin densities were calculated using
Mulliken population analysis (MPA). Orbitals from the Gaussian 09
calculations were plotted with the ChemCraft program. TD-DFT was
used to calculate the electronic transition energies and intensities from
the ground state to the 60−80 lowest-energy excited states.
5.9. Broken Symmetry Calculations. The BS calculations of
complexes 2−5 were completed using the B3LYP, PBEPBE, BLYP,
HCTH, and M06L functionals with the TZVP basis set on all atoms
(see details of calculations in terms of the functionals investigated for
each complex as given in the text). In each case, two fragments were
defined: LPNN and Fe-(CO)2Xn, where separate MO calculations were
completed on both fragments. Using the AOMix-FO option, the wave
functions of the two separate fragments were combined to form the
wave function of the whole complex. An MO calculation was then
completed for the total complex (with the SCF-converged wave
function), and further electronic structure characterization was then
completed in a similar manner as typical charge donation and FO
analyses. Specific information regarding the BS calculations of 2 and 3
can be found in the text, while the corresponding information for 4
and 5 is as follows.
To further characterize the electronic structure of complexes 4 and
5, BS calculations were completed to test the possibility of LPNN being
a redox-active ligand in these complexes. Three separate calculations
were conducted for each complex to allow for the total spin (S = 0)
found experimentally to stay constant, while varying the spins on the
Fe and LPNN. The first calculation allowed a β-spin electron to be
removed from Fe (now, S = 1/2) and placed on LPNN, while the
second calculation allowed an α-spin electron to move from Fe (now,
S = 1/2) to LPNN. The final calculation allowed for a normal
description of a low-spin Fe(II) (S = 0) with no unpaired electron on
LPNN. Notably, all calculations converged to the same energy and spin
density on Fe, which is consistent with the normal description (Fe(II),
the mixture was stirred for 30 min at 0 °C, distilled water was added,
and the aqueous phase was washed several times with DCM. The
combined extracts were washed with a concentrated sodium
bicarbonate solution and filtered over a plug of deactivated neutral
alumina (stage II−III). After it was dried over magnesium sulfate, the
solvent was removed at 10 °C, and traces of solvent were removed
under high vacuum at room temperature. The crude product (5.5 g)
was obtained as a yellow oil and carried forward without further
purification.
1H NMR (500 MHz, CDCl3, 298 K): δ = 1.98 (s, 6H, o-CH3), 2.15
(s, 3H, (CN)−CH3), 2.29 (s, 3H, p-CH3), 3.12 (s, 3H, O−SO2−
CH3), 5.39 (s, 2H, CH2), 6.89 (s, 2H, m-H), 7.55 (d, 1H, J = 7.6 Hz,
β-H, P-arm side), 7.86 (vt, 1H, J = 7.8 Hz, γ-H), 8.35 (d, 1H, J = 7.9
Hz, β-H, imine side) ppm. 13C{1H} NMR (126 MHz, CDCl3, 298 K):
δ = 16.5 (s), 17.8 (s), 20.7 (s), 38.0 (s), 71.6 (s), 121.1 (s), 123.3 (s),
125.1 (s), 128.6 (s), 132.3 (s), 137.5 (s), 146.0 (s), 152.5 (s), 156.4
(s), 167.0 (s) ppm. MS (ESI, methanol, m/z+): 252.1 [M−
CH3SO3+H]+ = [C17H20N2]+, 347.3 [M + H]+ = [C18H23N2O3S]+,
369.2 [M + Na]+ = [C18H22N2NaO3S]+, 715.4 [2M+Na]+
=
[C36H44N4NaO6S2]+; (ESI, methanol, m/z−): 95.0 [CH3SO3]−, 80.0
[CH4SO2]−.
A solution of the crude (6-(1-(mesitylimino)ethyl)pyridin-2-yl)-
methylmethanesulfonate (E) and di-tert-butylphosphine (4.0 mL, 21.6
mmol) in 60 mL of isopropyl alcohol was stirred at 50 °C for 2 d. After
the addition of 10 mL of triethylamine (71.7 mmol) all volatiles were
removed, and the residue was thoroughly extracted with ca. 100 mL of
pentane. After filtration, the solvent was removed, and 4.3 g of a yellow
oil was obtained, which contained 60−70% of the desired ligand as
1
based on H NMR and complexation studies (43−50% yield). For
characterization purposes, ca. 100 mg of the ligand mixture was
purified by column chromatography on deactivated, neutral alumina
(stage II−III), employing a very thin column and pentane/diethyl
ether (1:1) as the eluent. Ca. 20 mg of the pure ligand was obtained as
a yellow oil. Note, however, that for the complexation reactions, the
crude mixture after the extraction with pentane was used, employing
appropriate amounts of the metal precursor.
S = 0) with no unpaired electrons on Fe or LPNN
.
5.10. Ligand Syntheses. The α-iminopyridine-based PNN ligand
2 - [ ( d i - t e r t - b u t y l p h o s p h i n o ) m e t h y l ] - 6 - [ 1 - ( 2 , 4 , 6 -
trimethylphenylimino)ethyl]pyridine (LPNN) was synthesized in five
steps (see Supporting Information, Figure S1) starting from
commercially available dimethylpyridine-2,6-dicarboxylate (A), which
was in the first two steps converted to methyl 6-(hydroxymethyl)-
picolinate82 (B) and 2-acetyl-6-(hydroxymethyl)pyridine83 (C)
according to known procedures.
1H NMR (400 MHz, CDCl3, 298 K): δ = 1.20 (d, 18H, 3JHP = 11.0
Hz, H10), 1.99 (s, 6H, H15), 2.14 (s, 3H, H8), 2.28 (s, 3H, H16), 3.11
3
(d, 2H, 2JHP = 7.8 Hz, H1), 6.87 (s, 2H, H13), 7.45 (d, 1H, JHH = 7.7
Hz, H3), 7.66 (vt, 1H, 3JHH = 7.75 Hz, H4), 8.09 (d, 1H, 3JHH = 7.8 Hz,
H5) ppm. 13C{1H} NMR (101 MHz, CDCl3, 298 K): δ = 16.7 (s, C8),
18.0 (s, C15), 20.9 (s, C16), 29.9 (d, 2JCP = 13.3 Hz, C10), 31.8 (d, 1JCP
2-(Hydroxymethyl)-6-[1-(mesitylimino)ethyl]pyridine (D). A
mixture of 2-acetyl-6-(hydroxymethyl)pyridine (3.1 g, 20.5 mmol),
freshly distilled 2,4,6-trimethylaniline (3.0 mL, 21.4 mmol), and a
catalytic amount p-toluenesulfonic acid in 250 mL of toluene was
heated for 3 h under reflux employing a Dean−Stark apparatus. The
solvent was afterward evaporated, and the dark brown residue was
purified by column chromatography on deactivated neutral alumina
(stage II−III) with pentane/ethyl acetate (ca. 4:1). The desired
product was obtained as a yellow oil. Yield: 75% (4.1 g).
1
5
= 24.2 Hz, C1), 32.1 (d, JCP = 21.9 Hz, C9), 118.1 (d, JCP = 1.7 Hz,
C5), 124.9 (d, 3JCP = 8.1 Hz, C3), 125.5 (s, C12), 128.6 (s, C13), 132.1
2
(s, C14), 136.6 (s, C4), 146.5 (s, C11), 155.8 (s, C6), 161.2 (d, JCP
=
14.1 Hz, C2), 168.0 (s, C7) ppm. 31P{1H} NMR (121 MHz, CDCl3,
298 K): δ = 38.5 (s) ppm. 15N−1H HMQC (41 MHz, CDCl3, 298 K):
δ = 311.5 (s, N1), 334.0 (s, N2) ppm. MS (ESI, methanol, m/z+):
397.37 [M + H]+ = [C25H38N2P]+.
1H NMR (400 MHz, CDCl3, 298 K): δ = 1.99 (s, 6H, o-CH3), 2.19
3
Note that 2-acetyl-6-[(di-tert-butylphosphino)methyl]-pyridine was
(s, 3H, (CN)−CH3), 2.29 (s, 3H, p-CH3), 3.97 (t, 1H, JHH = 4.6
3
identified by NMR and MS studies as one of the main impurities in the
Hz, OH), 4.82 (d, 2H, JHH = 4.2 Hz, CH2), 6.89 (s, 2H, m-H), 7.31
1
(d, 1H, 3JHH = 7.6 Hz, β-H, P-arm side), 7.80 (vt, 1H, 3JHH = 7.7 Hz, γ-
H), 8.28 (d, 1H, 3JHH = 7.8 Hz, β-H, imine side) ppm. 13C{1H} NMR
(101 MHz, CDCl3, 298 K): δ = 16.6 (s), 17.8 (s), 20.7 (s), 63.7 (s),
119.9 (s), 121.4 (s), 125.2 (s), 128.6 (s), 132.3 (s), 137.2 (s), 146.1
(s), 155.2 (s), 157.5 (s), 166.9 (s) ppm. MS (ESI, methanol, m/z+):
synthesis of LPNN: H NMR (300 MHz, CDCl3, 298 K): δ = 1.18 (d,
3
18H, JHP = 11.1 Hz, PC(CH3)3), 2.69 (s, 3H, (CN)−CH3), 3.10
(d, 2H, 3JHP = 3.1 Hz, CH2), 7.54 (d, 1H, J = 7.7 Hz, β-H, P-arm side),
7.69 (vt, 1H, J = 7.7 Hz, γ-H), 7.80 (d, 1H, J = 7.6 Hz, β-H, imine
side) ppm. 31P{1H} NMR (121 MHz, CDCl3, 298 K): δ = 39.2 (s)
ppm. MS (ESI, methanol, m/z+): 280.24 [M + H]+ = [C16H27NOP]+.
2,6-Bis-[1-(mesitylimino)ethyl]pyridine (LNNN). 2,6-Diacetylpyr-
idine (1.0 g, 6.2 mmol) and 2,4,6-trimethylaniline (7.0 mL, 49.9
mmol) were dissolved in 30 mL of methanol, and 10 drops of formic
acid were added. The mixture was heated for 45 h to 50 °C, and
precipitation of the product was completed in an ice bath. After
filtration, several washings with cold methanol, and drying under high
269.21 [M + H]+ = [C17H21N2O]+, 291.22 [M + Na]+
[C17H20N2NaO]+, 559.43 [2M+Na]+ = [C34H40N4NaO2]+.
=
2-[(Di-tert-butylphosphino)methyl]-6-[1-(2,4,6-
mesitylimino)ethyl]pyridine (LPNN). To a solution of 2-(hydrox-
ymethyl)-6-[1-(mesitylimino)ethyl]pyridine (4.1 g, 15.3 mmol) in 260
mL of DCM, cooled in an ice bath, were added triethylamine (8.7 mL,
62.8 mmol) and methanesulfonyl chloride (3.5 mL, 45.2 mmol). After
N
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
vacuum for several hours, the product was obtained as a yellow
powder. Yield: 78% (1.95 g).
Hz, H5) ppm. 13C{1H} NMR (101 MHz, C6D6, 298 K): δ = 14.7 (s,
2
C8), 18.8 (s, C15), 21.1 (s, C16), 29.3 (d, JCP = 3.8 Hz, C10), 35.7 (d,
1JCP = 17.2 Hz, C1), 38.0 (d, 1JCP = 11.3 Hz, C9), 108.9 (d, 3JCP = 9.3
Hz, C3), 121,1 (s, C5), 124.0 (s, C4), 129.2 (s, C13), 129.8 (s, C12),
3
133.8 (s, C14), 144.7 (d, JCP = 5.1 Hz, C6), 144.9 (s, C7), 153.1 (s,
C11), 159.2 (d, 2JCP = 7.8 Hz, C2), 221.1 (d, 2JCP = 11.3 Hz, CO) ppm.
31P{1H} NMR (121 MHz, C6D6, 298 K): δ = 138.4 (s) ppm. 15N−1H
HMQC (41 MHz, C6D6, 298 K): δ = 258.1 (s, N1), 259.2 (s, N2)
ppm. IR(NaCl): v = 1875 (νCO), 1934 cm−1 (νCO). IR(pentane): V =
̃
̃
1897 (νCO), 1952 cm−1 (νCO). Anal. Calcd for C27H37FeN2O2P: C,
63.78; H, 7.34; N, 5.51; Found: C, 63.16; H, 7.36; N, 5.10%. MS (ESI,
acetonitrile, m/z+): 480.3 [M−CO]+ = [C26H37FeN2OP]+. Magnetic
susceptibility (Evans): μeff = 0.0 (1,4-dioxane in C6D6, 298 K);
1H NMR (400 MHz, CDCl3, 298 K): δ = 2.04 (s, 12H, H10), 2.26
3
(s, 6H, H5), 2.32 (s, 6H, H11), 6.92 (s, 4H, H8), 7.92 (vt, 1H, JHH
=
7.8 Hz, H1), 8.49 (d, 2H, JHH = 7.8 Hz, H2) ppm. 13C{1H} NMR
(101 MHz, CDCl3, 298 K): δ = 16.4 (s, C5), 17.9 (s, C10), 20.7 (s,
C11), 122.1 (s, C2), 125.2 (s, C7), 128.6 (s, C8), 132.2 (s, C9), 136.8 (s,
C1), 146.3 (s, C6), 155.2 (s, C3), 167.4 (s, C4) ppm. 15N−1H HMQC
(41 MHz, CDCl3, 298 K): δ = 310.4 (s, N1), 336.0 (s, N2) ppm. MS
(ESI, methanol, m/z+): 398.3 [M + H]+ = [C27H32N3]+, 420.3 [M +
Na]+ = [C27H31N3Na]+, 817.6 [2M+Na]+ = [C54H62N6Na]+.
3
Mossbauer: 80 K: δ = 0.00 mm s−1, ΔE = 1.14 mm s−1; 5 K: δ = 0.01
̈
Q
mm s−1, ΔEQ = 1.13 mm s−1. XPS: Fe 708.1, N 398.7, C 284.8, P
130.5 eV.
[Fe(CO)2LPNN](BF4) (3). In a solution of [Fe(CO)2LPNN] (2) (107.4
mg, 0.211 mmol) in 15 mL of THF, ferrocenium tetrafluoroborate
(57.5 mg, 0.211 mmol) was suspended, and the mixture was stirred for
3 h, during which the color of the solution changed from purple to
green. The volume of the reaction mixture was reduced to 2 mL, and
precipitation was completed by the addition of pentane. The
supernatant was removed, and the dark green residue was washed
several times with pentane and dried in vacuum. Crystals suitable for
X-ray diffraction were grown at room temperature by slow diffusion of
pentane into a THF solution. Yield: 96% (120.4 mg).
5.11. Complex Syntheses. [FeBr2LPNN] (1). A mixture of the crude
PNN ligand (801.4 mg, contains ca. 560 mg of pure ligand, 1.412
mmol) and anhydrous FeBr2 (210 mg, 0.974 mmol) in 35 mL of THF
was stirred overnight. The solvent was evaporated completely, and the
residue was treated with diethyl ether (in which it is nearly insoluble)
under vigorous stirring for several hours. After the blue powder settled
completely, the supernatant was removed, and the residue was washed
1
several times with pentane and dried in high vacuum. Note that H
IR(NaCl): V = 1922 (νCO), 1987 cm−1 (νCO). Anal. Calcd for
and 31P NMR spectra of the supernatant indicate that only small
amounts of the PNN ligand remain unreacted. Elemental analysis and
̃
C27H37BF4FeN2O2P: C, 54.48; H, 6.27; N, 4.71; Found: C, 50.76; H,
6.37; N, 4.48%. MS (ESI, acetonitrile, m/z+): 480.3 [M−BF4−CO]+ =
[C26H37FeN2OP]+; (ESI, acetonitrile, m/z−): 87.0 [BF4]−. Magnetic
susceptibility (Evans): μeff = 2.1 (1,4-dioxane in CDCl3, 298 K).
SQUID magnetometry: μeff = 2.14 (at 298 K, see Supporting
Mossbauer spectroscopy unequivocally prove that no other iron-
̈
containing complexes than [FeBr2LPNN] (1) are present in the final
product. Crystals suitable for X-ray diffraction were obtained by
dissolving the complex in benzene; after a while, dark blue crystals
were formed. Yield: 96% (573.5 mg).
Information for further details). Mossbauer: 80 K: δ = 0.12 mm s−1,
̈
ΔEQ = 0.54 mm s−1; 5 K: δ = 0.12 mm s−1, ΔEQ = 0.54 mm s−1. XPS:
Fe 708.8, F 685.5, N 399.7, C 284.8, B 193.7, P 130.7 eV.
[Fe(F)(CO)2LPNN](BF4) (4). In a solution of [Fe(CO)2LPNN] (2)
(165.2 mg, 0.325 mmol) in 50 mL of THF, ferrocenium
tetrafluoroborate (173.9 mg, 0.637 mmol) was suspended, and the
mixture was stirred in the dark for 3 h. After the precipitate settled
completely, the supernatant was removed, and the dark green powder
was washed once with THF and three times with pentane and dried in
vacuum. Crystals suitable for X-ray diffraction were grown at room
temperature by slow diffusion of pentane into a DCM solution. Note
that the product is light-sensitive, particularly in solution (the identity
of the decomposition product(s) remains unknown). Yield: 92%
(179.5 mg).
Anal. Calcd for C25H37Br2FeN2P: C, 49.05; H, 6.09; N, 4.58;
Found: C, 48.53; H, 6.03; N, 4.24%. MS (ESI, acetonitrile, m/z+):
531.2 [M−Br]+ = [C25H37BrFeN2P]+; (ESI, acetonitrile, m/z−): 78.9
[Br]−. Magnetic susceptibility (Evans): μeff = 5.4 (1,4-dioxane in
CDCl3, 298 K). SQUID magnetometry: μeff = 4.77 (at 298 K, see
Supporting Information for further details). Mossbauer: 80 K: δ =
̈
0.94 mm s−1, ΔEQ = 2.59 mm s−1; 5 K: δ = 0.96 mm s−1, ΔEQ = 2.90
mm s−1. XPS: Fe 709.4, N 399.5, C 284.8, P 130.5, Br 68.1 eV.
[FeCl2LPNN] (1′). Synthesized in analogy to [FeBr2LPNN] (1). Crude
PNN ligand (100 mg, contains ca. 70 mg of pure ligand, 0.177 mmol);
FeCl2 (15.4 mg, 0.122 mmol); 5 mL of THF. Yield: 59% (37.3 mg).
MS (ESI, acetonitrile, m/z+): 487.3 [M−Cl]+ = [C25H37ClFeN2P]+.
XPS: Fe 709.4, N 399.5, C 284.8, Cl 197.8, P 130.4 eV.
[Fe(CO)2LPNN] (2). In a 100 mL pressure flask fitted with a Teflon
stopcock [FeBr2LPNN] (1) (438 mg, 0.715 mmol) was dissolved in 30
mL of THF, and sodium amalgam (10%) (464 mg, 2.018 mmol) was
added. After two freeze−pump−thaw cycles, the frozen mixture was
allowed to warm to room temperature under 1 bar of CO atmosphere
and was vigorously stirred overnight. After the solvent was evaporated
completely, the residue was extracted with pentane (ca. 200 mL), and
the solution was filtered through a syringe filter and evaporated to ca. 5
mL. After filtration through a frit, the product was obtained as a dark
purple, nearly black powder. Crystals suitable for X-ray diffraction were
obtained from a concentrated pentane solution at −30 °C. Yield: 90%
(328.6 mg).
1H NMR (400 MHz, CD2Cl2, 298 K): δ = 1.33 (d, 9H, 3JHP = 14.4
Hz, H10′), 1.46 (d, 9H, 3JHP = 14.4 Hz, H10), 2.11 (s, 3H, H15), 2.35 (s,
3H, H16), 2.38 (s, 3H, H15′), 2.39 (s, 3H, H8), 3.94 (dd, 1H, 2JHP = 7.2
Hz, 2JHH = 17.4 Hz, H1), 4.20 (dd, 1H, 2JHP = 13.7 Hz, 2JHH = 17.4 Hz,
H1′), 7.03 (s, 1H, H13), 7.08 (s, 1H, H13′), 8.16 (d, 1H, 3JHH = 7.5 Hz,
H5), 8.20 (d, 1H, 3JHH = 7.7 Hz, H3), 8.37 (vt, 1H, 3JHH = 7.6 Hz, H4)
ppm. 13C{1H} NMR (101 MHz, CD2Cl2, 298 K): δ = 18.3 (s, C15),
2
19.0 (s, C15′ and C8), 20.9 (s, C16), 29.7 (d, JCP = 9.9 Hz, C10), 30.4
2
1
1
(d, JCP = 1.3 Hz, C10′), 37.2 (d, JCP = 18.7 Hz, C1), 38.2 (d, JCP
=
1
16.1 Hz, C9′), 40.5 (d, JCP = 11.0 Hz, C9), 127.0 (s, C5), 127.7 (d,
3JCP = 8.4 Hz, C3), 127.8 (s, C12), 130.2 (s, C13), 130.8 (s, C12′), 130.9
(s, C13′), 138.2 (s, C14), 142.1 (s, C4), 145.3 (s, C11), 156.6 (d, JCP
=
3
1H NMR (400 MHz, C6D6, 298 K): δ = 0.98 (d, 18H, JHP = 12.6
3.5 Hz, C6), 164.5 (d, 2JCP = 2.8 Hz, C2), 179.9 (s, C7), 208.1 (dd, 2JCP
= 7.0 Hz, 2JCF = 22.4 Hz, CO), 209.9 (dd, 2JCP = 18.5 Hz, 2JCF = 35.9
Hz, CO) ppm. 31P{1H} NMR (121 MHz, CD2Cl2, 298 K): δ = 106.9
(d, 2JPF = 9.7 Hz) ppm. 19F{1H} NMR (282 MHz, CD2Cl2, 298 K): δ
Hz, H10), 1.74 (s, 3H, H8), 2.23 (s, 3H, H16), 2.25 (s, 6H, H15), 3.27
(d, 2H, 2JHP = 8.3 Hz, H1), 6.30 (d, 1H, 3JHH = 6.6 Hz, H3), 6.75 (dd,
1H, 3JHH = 6.6, 8.5 Hz, H4), 6.95 (s, 2H, H13), 7.04 (d, 1H, 3JHH = 8.5
O
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
−
= −154.5 (s, br, BF4 ), −428.5 (d, 2JFP = 9.9 Hz, Fe−F) ppm. 15N−1H
[Fe(CO)2LNNN] (7). In a 100 mL pressure flask fitted with a Teflon
stopcock, sodium amalgam (10%, 390 mg, 1.696 mmol) was added to
0.8 mL of mercury in 5 mL of THF, and the mixture was stirred until
the amalgam was dissolved in the mercury. To this mixture a
suspension of [FeBr2LNNN] (6) (506 mg, 0.825 mmol) in 25 mL of
THF was added. After two freeze−pump−thaw cycles, the frozen
mixture was allowed to warm to room temperature under 1 bar of CO
atmosphere and was vigorously stirred overnight. After the solvent was
evaporated completely, the residue was extracted with ether (ca. 500
mL), and the solution was filtrated through a syringe filter and
evaporated to ca. 5 mL. After filtration through a frit, the product was
obtained as a dark green powder. Crystals suitable for X-ray diffraction
were obtained by slow diffusion of pentane into a DCM solution.
Yield: 71% (296.3 mg).
HMQC (41 MHz, CD2Cl2, 298 K): δ = 257.6 (s, N1), 274.1 (s, N2)
ppm. IR(NaCl): V = 2015 (νCO), 2052 cm−1 (νCO). Anal. Calcd for
̃
C27H37BF5FeN2O2P: C, 52.80; H, 6.07; N, 4.56; Found: C, 51.44; H,
6.00; N, 4.27%. MS (ESI, acetonitrile, m/z+): 527.3 [M−BF4]+
=
;
[C27H37FFeN2O2P]+, 471.3 [M−BF4−2CO]+ = [C25H37FFeN2P]+
(ESI, acetonitrile, m/z−): 87.0 [BF ]−. Mossbauer: 80 K: δ = 0.04 mm
̈
4
s−1, ΔEQ = 0.55 mm s−1; 5 K: δ = 0.04 mm s−1, ΔEQ = 0.56 mm s−1.
−
XPS: Fe 709.9, F 682.9 (Fe-F) and 685.5 (BF4 ), N 399.8, C 284.8, B
193.9, P 130.8 eV.
[Fe(H)(CO)2LPNN](BF4) (5). To a solution of [Fe(CO)2LPNN] (2)
(196 mg, 0.386 mmol, 1 equiv) in 30 mL of THF, an excess of HBF4·
Et2O (ca. 50 drops, ca. 10 equiv) was added, and the mixture was
stirred at room temperature for 2 h, during which the color changed
from purple to green. The solvent volume was reduced to a third, and
the raw product was precipitated by the addition of pentane. The
supernatant was removed, and the residue was treated for 2 h with
diethyl ether under vigorous stirring. After the green powder settled
completely, the supernatant was removed, and the residue was washed
several times with pentane and dried in high vacuum. The residue was
taken up in DCM, filtered, and precipitated by dropping the solution
into diethyl ether while stirring. This procedure aided in removing
presumable paramagnetic impurities of unknown identity that caused
broadened lines in the NMR spectra. The precipitate was washed once
with diethyl ether and dried in vacuum. Crystals suitable for X-ray
diffraction were grown by slow diffusion of pentane into a DCM
solution. Yield: 89% (204.7 mg).
1H NMR (400 MHz, C6D6, 298 K): δ = 2.01 (s, 12H, H10), 2.01 (s,
6H, H5), 2.08 (s, 6H, H11), 6.78 (m, 4H, H8), 7.20 (vt, 1H, 3JHH = 7.7
3
Hz, H1), 7.77 (d, 2H, JHH = 7.7 Hz, H2) ppm. 13C{1H} NMR (101
MHz, C6D6, 298 K): δ = 15.0 (s, C5), 18.3 (s, C10), 20.9 (s, C11), 116.2
(s, C1), 120.6 (s, C2), 129.1 (s, C7), 129.3 (s, C8), 134.8 (s, C9), 145.0
(s, C3), 150.7 (s, C6), 154.9 (s, C4), 215.1 (s, CO) ppm. 15N−1H
HMQC (41 MHz, C6D6, 298 K): δ = 248.7 (s, N1), 255.2 (s, N2).
IR(NaCl): V = 1878 (νCO), 1949 cm−1 (νCO). Mossbauer: 80 K: δ =
̃
̈
0.00 mm s−1, ΔEQ = 1.44 mm s−1. XPS: Fe 708.4, N 398.9, C 284.8
eV.
ASSOCIATED CONTENT
* Supporting Information
Schematic illustration of synthesis including reaction con-
■
1H NMR (400 MHz, CDCl3, 298 K): δ = −4.59 (d, 1H, 2JHP = 58.2
S
3
3
Hz, Fe−H), 1.28 (d, 9H, JHP = 14.2 Hz, H10′), 1.40 (d, 9H, JHP
=
14.5 Hz, H10), 2.23 (s, 3H, H15), 2.25 (s, 3H, H15′), 2.28 (s, 3H, H8),
ditions, Mossbauer spectra, EPR simulations, electronic
̈
2
2
2.33 (s, 3H, H16), 3.73 (dd, 1H, JHP = 6.4 Hz, JHH = 18.0 Hz, H1),
4.24 (dd, 1H, 2JHP = 13.4 Hz, 2JHH = 18.1 Hz, H1′), 7.00 (s, 1H, H13′),
7.02 (s, 1H, H13), 8.13 (m, 1H, H5), 8.16 (m, 1H, H3), 8.16 (m, 1H,
H4) ppm. 13C{1H} NMR (101 MHz, CDCl3, 298 K): δ = 17.4 (s, C8),
absorption spectra, SQUID data, XPS spectra, X-ray structural
data; calculated MO diagrams, broken symmetry calculations,
DFT-optimized geometry coordinates, TD-DFT. This material
is available free of charge via the Internet at http://pubs.acs.org.
18.0 (s, C15), 18.3 (s, C15′), 20.8 (s, C16), 27.3 (s, C10′), 30.1 (d, 2JCP
=
2.3 Hz, C10), 37.0 (d, 1JCP = 19.0 Hz, C1), 37.5 (d, 1JCP = 16.2 Hz, C9),
1
3
AUTHOR INFORMATION
Corresponding Authors
*E-mail: neidig@chem.rochester.edu. (M.L.N.)
*E-mail: david.milstein@weizmann.ac.il. (D.M.)
38.4 (d, JCP = 22.0 Hz, C9′), 125.8 (s, C5), 126.5 (d, JCP = 8.6 Hz,
C3), 126.9 (s, C12), 127.3 (s, C12′), 129.9 (s, C13), 130.1 (s, C13′),
137.0 (s, C14), 139.5 (s, C4), 145.9 (s, C11), 153.9 (d, JCP = 4.3 Hz,
■
2
2
C6), 162.4 (d, JCP = 4.0 Hz, C2), 173.3 (s, C7), 205.0 (d, JCP = 7.8
Hz, CO), 212.0 (d, JCP = 19.3 Hz, CO) ppm. 31P{1H} NMR (121
2
Author Contributions
MHz, CDCl3, 298 K): δ = 120.3 (s) ppm. 19F{1H} NMR (282 MHz,
⊥These authors contributed equally.
CDCl3, 298 K): δ = −154.5 (s, br, BF4 ) ppm. 15N−1H HMQC (41
−
MHz, CDCl3, 298 K): δ = 258.6 (s, N1), 274.8 (s, N2) ppm.
Notes
IR(NaCl): V = 1929 (νFe−H), 1973 (νCO), 2022 cm−1 (νCO). Anal.
̃
The authors declare no competing financial interest.
Calcd for C27H38BF4FeN2O2P: C, 54.39; H, 6.42; N, 4.70; Found: C,
51.44; H, 6.20; N, 4.21%. MS (ESI, acetonitrile, m/z+): 509.3 [M-
ACKNOWLEDGMENTS
BF4 ]+
= =
[C2 7 H3 8 FeN2 O2 P]+ , 481.3 [M-BF4 −CO]+
■
[C26H38FeN2OP]+ ; (ESI, acetonitrile, m/z−): 87.0 [BF4]−.
We would like to thank Prof. B. Rybtchinsky and Prof. M. v. d.
Boom who allowed us to use their UV−vis spectrometry and
CV equipment, respectively. This research was supported by
the European Research Council under the FP7 framework
(ERC No. 246837), the MINERVA Foundation, and the Univ.
of Rochester (M.L.N.). B.B. received the Feodor-Lynen
postdoctoral fellowship from the Alexander von Humboldt
Foundation, and D.M. holds the Israel Matz Professorial Chair
of Organic Chemistry. The authors also acknowledge the
Center for Integrated Research Computing at the Univ. of
Rochester for providing the necessary computing systems and
support to enable the computational research presented in this
study.
Mossbauer: 80 K: δ = −0.04 mm s−1, ΔE = 1.48 mm s−1. XPS:
̈
Q
Fe 709.2, F 685.5, N 399.9, C 284.8, B 193.6, P 130.9 eV.
[FeBr2LNNN] (6). A mixture of 2,6-bis-[1-(mesitylimino)ethyl]-
pyridine (455 mg, 1.144 mmol) and anhydrous FeBr2 (233 mg,
1.080 mmol) in 50 mL of THF was stirred overnight. The volume was
reduced to one-third, the product was precipitated by the addition of
50 mL of diethyl ether, and the mixture was stirred for several hours.
After the blue powder settled completely, the supernatant was
removed, and the residue was washed several times with diethyl ether
and dried under vacuum. Yield: 96% (635.8 mg). MS (ESI,
acetonitrile, m/z+): 532.2 [M−Br]+ = [C27H31BrFeN3]+; (ESI,
acetonitrile, m/z−): 78.9 [Br]−. XPS: Fe 709.7, N 399.5, C 284.8, Br
68.2 eV.
P
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(34) Luca, O. R.; Crabtree, R. H. Chem. Soc. Rev. 2013, 42, 1440−
1459.
(35) Praneeth, V. K. K.; Ringenberg, M. R.; Ward, T. R. Angew.
Chem., Int. Ed. 2012, 51, 10228−10234.
(36) Chirik, P. J.; Wieghardt, K. Science 2010, 327, 794−795.
(37) http://goldbook.iupac.org/O04365.html.
(38) Leigh, V.; Carleton, D. J.; Olguin, J.; Mueller-Bunz, H.; Wright,
L. J.; Albrecht, M. Inorg. Chem. 2014, 53, 8054−8060.
(39) Gore-Randall, E.; Irwin, M.; Denning, M. S.; Goicoechea, J. M.
Inorg. Chem. 2009, 48, 8304−8316.
REFERENCES
■
(1) Klein, J. E. M. N.; Miehlich, B.; Holzwarth, M. S.; Bauer, M.;
Milek, M.; Khusniyarov, M. M.; Knizia, G.; Werner, H.-J.; Plietker, B.
Angew. Chem., Int. Ed. 2014, 53, 1790−1794.
(2) Evans, W. J.; Fang, M.; Bates, J. E.; Furche, F.; Ziller, J. W.; Kiesz,
M. D.; Zink, J. I. Nat. Chem. 2010, 2, 644−647.
(3) Brown, S. N. Inorg. Chem. 2012, 51, 1251−1260.
(4) Zanello, P.; Corsini, M. Coord. Chem. Rev. 2006, 250, 2000−
2022.
(5) Pierpont, C. G. Coord. Chem. Rev. 2001, 216−217, 99−125.
(6) Papavassiliou, G. C.; Anyfantis, G. C.; Mousdis, G. A. Crystals
2012, 2, 762−811.
(7) Broere, D. L. J.; de Bruin, B.; Reek, J. N. H.; Lutz, M.; Dechert,
S.; van der Vlugt, J. I. J. Am. Chem. Soc. 2014, 136, 11574−11577.
(8) Smith, A. L.; Hardcastle, K. I.; Soper, J. D. J. Am. Chem. Soc. 2010,
132, 14358−14360.
(9) Chaudhuri, P.; Verani, C. N.; Bill, E.; Bothe, E.; Weyhermuller,
T.; Wieghardt, K. J. Am. Chem. Soc. 2001, 123, 2213−2223.
(10) Bowman, A. C.; England, J.; Sproules, S.; Weyhermuller, T.;
Wieghardt, K. Inorg. Chem. 2013, 52, 2242−2256.
(11) Bowman, A. C.; Sproules, S.; Wieghardt, K. Inorg. Chem. 2012,
51, 3707−3717.
(12) England, J.; Scarborough, C. C.; Weyhermuller, T.; Sproules, S.;
Wieghardt, K. Eur. J. Inorg. Chem. 2012, 4605−4621.
(13) Irwin, M.; Doyle, L. R.; Kramer, T.; Herchel, R.; McGrady, J. E.;
Goicoechea, J. M. Inorg. Chem. 2012, 51, 12301−12312.
(14) Scarborough, C. C.; Wieghardt, K. Inorg. Chem. 2011, 50,
9773−9793.
(15) Scarborough, C. C.; Sproules, S.; Weyhermuller, T.; DeBeer, S.;
Wieghardt, K. Inorg. Chem. 2011, 50, 12446−12462.
(16) Irwin, M.; Jenkins, R. K.; Denning, M. S.; Kramer, T.;
Grandjean, F.; Long, G. J.; Herchel, R.; McGrady, J. E.; Goicoechea, J.
M. Inorg. Chem. 2010, 49, 6160−6171.
(17) Lu, C. C.; Weyhermuller, T.; Bill, E.; Wieghardt, K. Inorg. Chem.
2009, 48, 6055−6064.
(18) Lu, C. C.; Bill, E.; Weyhermuller, T.; Bothe, E.; Wieghardt, K. J.
Am. Chem. Soc. 2008, 130, 3181−3197.
(19) Lu, Connie C.; DeBeer George, S.; Weyhermuller, T.; Bill, E.;
Bothe, E.; Wieghardt, K. Angew. Chem., Int. Ed. 2008, 47, 6384−6387.
(20) Darmon, J. M.; Stieber, S. C. E.; Sylvester, K. T.; Fernandez, I.;
́
Lobkovsky, E.; Semproni, S. P.; Bill, E.; Wieghardt, K.; DeBeer, S.;
Chirik, P. J. J. Am. Chem. Soc. 2012, 134, 17125−17137.
(21) Stieber, S. C. E.; Milsmann, C.; Hoyt, J. M.; Turner, Z. R.;
Finkelstein, K. D.; Wieghardt, K.; DeBeer, S.; Chirik, P. J. Inorg. Chem.
2012, 51, 3770−3785.
(22) Bowman, A. C.; Milsmann, C.; Atienza, C. C. H.; Lobkovsky, E.;
Wieghardt, K.; Chirik, P. J. J. Am. Chem. Soc. 2010, 132, 1676−1684.
(23) Bart, S. C.; Lobkovsky, E.; Bill, E.; Wieghardt, K.; Chirik, P. J.
Inorg. Chem. 2007, 46, 7055−7063.
(40) Zell, T.; Milko, P.; Fillman, K. L.; Diskin-Posner, Y.; Bendikov,
T.; Iron, M. A.; Leitus, G.; Ben-David, Y.; Neidig, M. L.; Milstein, D.
Chem.Eur. J. 2014, 20, 4403−4413.
(41) Milko, P.; Iron, M. A. J. Chem. Theory Comput. 2013, 10, 220−
235.
(42) Tondreau, A. M.; Milsmann, C.; Lobkovsky, E.; Chirik, P. J.
Inorg. Chem. 2011, 50, 9888−9895.
(43) Bart, S. C.; Lobkovsky, E.; Chirik, P. J. J. Am. Chem. Soc. 2004,
126, 13794−13807.
̈
̈
(44) Note that a similar ligand bearing an O−P instead of a CH2−P
arm was synthesized before, and the corresponding FeBr2 and
Fe(CO)2 complexes were also described. The focus of the studies,
however, was on catalytic hydrosilylation reactions rather than on the
electronic nature of the complexes: Peng, D.; Zhang, Y.; Du, X.;
Zhang, L.; Leng, X.; Walter, M. D.; Huang, Z. J. Am. Chem. Soc. 2013,
135, 19154−19166.
̈
̈
(45) Nishida, Y.; Kino, K.; Kida, S. J. Chem. Soc., Dalton Trans. 1987,
1157−1161.
(46) Oliver, J. D.; Mullica, D. F.; Hutchinson, B. B.; Milligan, W. O.
Inorg. Chem. 1980, 19, 165−169.
̈
(47) Evans, D. F. J. Chem. Soc. 1959, 2003−2005.
̈
(48) Gutlich, P.; Ensling, J. Mossbauer Spectroscopy in Inorganic
̈
̈
Electronic Structure and Spectroscopy; Solomon, E. I., Lever, A. B. P.,
Eds.;Wiley: New York, 2006; Vol. I.
(49) Brant, P.; Feltham, R. D. J. Electron Spectrosc. Relat. Phenom.
1983, 32, 205−221.
(50) Feltham, R. D.; Brant, P. J. Am. Chem. Soc. 1982, 104, 641−645.
(51) Brant, P.; Feltham, R. D. Inorg. Chem. 1980, 19, 2673−2676.
(52) Enemark, J. H.; Feltham, R. D. Coord. Chem. Rev. 1974, 13,
339−406.
(53) Cahen, D.; Lester, J. E. Chem. Phys. Lett. 1973, 18, 108−111.
(54) Moddeman, W. E.; Blackburn, J. R.; Kumar, G.; Morgan, K. A.;
Albridge, R. G.; Jones, M. M. Inorg. Chem. 1972, 11, 1715−1717.
(55) Hu, L.; Liu, W.; Li, C.-H.; Zhou, X.-H.; Zuo, J.-L. Eur. J. Inorg.
Chem. 2013, 2013, 6037−6048.
̈
̈
̈
(56) Nielson, A. J.; Metson, J. B. Polyhedron 2012, 31, 143−149.
(57) Karpov, A.; Konuma, M.; Jansen, M. Chem. Commun. 2006,
838−840.
(58) Langer, R.; Leitus, G.; Ben-David, Y.; Milstein, D. Angew. Chem.,
Int. Ed. 2011, 50, 2120−2124.
(24) Bart, S. C.; Chłopek, K.; Bill, E.; Bouwkamp, M. W.; Lobkovsky,
E.; Neese, F.; Wieghardt, K.; Chirik, P. J. J. Am. Chem. Soc. 2006, 128,
13901−13912.
(25) Knijnenburg, Q.; Gambarotta, S.; Budzelaar, P. H. M. Dalton
Trans. 2006, 5442−5448.
(26) de Bruin, B.; Bill, E.; Bothe, E.; Weyhermuller, T.; Wieghardt, K.
(59) Pelczar, E. M.; Emge, T. J.; Krogh-Jespersen, K.; Goldman, A. S.
Organometallics 2008, 27, 5759−5767.
(60) Myers, T. W.; Berben, L. A. Inorg. Chem. 2012, 51, 1480−1488.
(61) Hong, S.; Huber, S. M.; Gagliardi, L.; Cramer, C. C.; Tolman,
W. B. J. Am. Chem. Soc. 2007, 129, 14190−14192.
(62) Bianchini, C.; Mantovani, G.; Meli, A.; Migliacci, F.; Laschi, F.
Organometallics 2003, 22, 2545−2547.
̈
Inorg. Chem. 2000, 39, 2936−2947.
(27) Chang, M.-C.; Dann, T.; Day, D. P.; Lutz, M.; Wildgoose, G. G.;
Otten, E. Angew. Chem., Int. Ed. 2014, 53, 4118−4122.
(28) Zhou, W.; Patrick, B. O.; Smith, K. M. Chem. Commun. 2014,
50, 9958−9960.
(63) Takaoka, A.; Peters, J. C. Inorg. Chem. 2011, 51, 16−18.
(64) De Bruin, B.; Hetterscheid, D. G. H.; Koekkoek, A. J. J.;
Grutzmacher, H. The Organometallic Chemistry of Rh-, Ir-, Pd-, and
̈
Pt-Based Radicals: Higher Valent Species. In Progress in Inorganic
Chemistry; John Wiley & Sons, Inc., 2008; pp 247−354.
(65) Kannan, S.; Moody, M. A.; Barnes, C. L.; Duval, P. B. Inorg.
Chem. 2006, 45, 9206−9212.
(29) Lyaskovskyy, V.; de Bruin, B. ACS Catal. 2012, 2, 270−279.
(30) Kaim, W. Inorg. Chem. 2011, 50, 9752−9765.
(31) Adhikari, D.; Mossin, S.; Basuli, F.; Huffman, J. C.; Szilagyi, R.
K.; Meyer, K.; Mindiola, D. J. J. Am. Chem. Soc. 2008, 130, 3676−3682.
(32) Butin, K. P.; Beloglazkina, E. K.; Zyk, N. V. Russ. Chem. Rev.
2005, 74, 531−553.
(66) Reger, D. L.; Watson, R. P.; Gardinier, J. R.; Smith, M. D.;
Pellechia, P. J. Inorg. Chem. 2006, 45, 10088−10097.
(67) Gorelsky, S. I.; Ghosh, S.; Solomon, E. I. J. Am. Chem. Soc. 2006,
128, 278−290.
(33) Evangelio, E.; Ruiz-Molina, D. Eur. J. Inorg. Chem. 2005, 2957−
2971.
Q
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX

Inorganic Chemistry
Article
(68) Zell, T.; Langer, R.; Iron, M. A.; Konstantinovski, L.; Shimon, L.
J. W.; Diskin-Posner, Y.; Leitus, G.; Balaraman, E.; Ben-David, Y.;
Milstein, D. Inorg. Chem. 2013, 52, 9636−9649.
(69) Note that a crucial influence on the 15N chemical shift of the
pyridine nitrogen atom is observed upon dearomatization of
[Fe(LPNN‑bp)2]2+ complexes (see ref 68).
(70) Fulmer, G. R.; Miller, A. J. M.; Sherden, N. H.; Gottlieb, H. E.;
Nudelman, A.; Stoltz, B. M.; Bercaw, J. E.; Goldberg, K. I.
Organometallics 2010, 29, 2176−2179.
(71) Marco,
2003, 142, 13−19.
́ ́
A.; Compano, R.; Rubio, R.; Casals, I. Microchim. Acta
̃
(72) Lagaditis, P. O.; Sues, P. E.; Sonnenberg, J. F.; Wan, K. Y.;
Lough, A. J.; Morris, R. H. J. Am. Chem. Soc. 2014, 136, 1367−1380.
(73) Bain, G. A.; Berry, J. F. J. Chem. Educ. 2008, 85, 532.
(74) Beamson, G.; Briggs, D. High-Resolution XPS of Organic
Polymers: The Scienta ESCA 300 Database; Wiley: Chichester, U.K.,
1992.
(75) Neese, F.; Solomon, E. I. Inorg. Chem. 1999, 38, 1847−1865.
(76) Pavel, E. G.; Kitajima, N.; Solomon, E. I. J. Am. Chem. Soc. 1998,
120, 3949−3962.
(77) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.;
Robb, M. A.; Cheeseman, J. R.; Scalmani, G.; Barone, V.; Mennucci,
B.; Petersson, G. A.; Nakatsuji, H.; Caricato, M.; Li, X.; Hratchian, H.
P.; Izmaylov, A. F.; Bloino, J.; Zheng, G.; Sonnenberg, J. L.; Hada, M.;
Ehara, M.; Toyota, K.; Fukuda, R.; Hasegawa, J.; Ishida, M.; Nakajima,
T.; Honda, Y.; Kitao, O.; Nakai, H.; Vreven, T.; Montgomery, J. A., Jr.;
Peralta, J. E.; Ogliaro, F.; Bearpark, M.; Heyd, J. J.; Brothers, E.; Kudin,
K. N.; Staroverov, V. N.; Kobayashi, R.; Normand, J.; Raghavachari, K.;
Rendell, A.; Burant, J. C.; Iyengar, S. S.; Tomasi, J.; Cossi, M.; Rega,
N.; Millam, M. J.; Klene, M.; Knox, J. E.; Cross, J. B.; Bakken, V.;
Adamo, C.; Jaramillo, J.; Gomperts, R.; Stratmann, R. E.; Yazyev, O.;
Austin, A. J.; Cammi, R.; Pomelli, C.; Ochterski, J. W.; Martin, R. L.;
Morokuma, K.; Zakrzewski, V. G.; Voth, G. A.; Salvador, P.;
̈
Dannenberg, J. J.; Dapprich, S.; Daniels, A. D.; Farkas, O.;
Foresman, J. B.; Ortiz, J. V.; Cioslowski, J.; Fox, D. J. Gaussian 09,
Revision D.01; Gaussian, Inc.: Wallingford, CT, 2013.
(78) Schafer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100,
̈
5829−5835.
(79) Tomasi, J.; Mennucci, B.; Cammi, R. Chem. Rev. 2005, 105,
2999−3094.
(80) Gorelsky, S. I. AOMix: Program for Molecular Orbital Analysis.
http://www.sg-chem.net/, version 6.85, 2014.
(81) Gorelsky, S. I.; Lever, A. B. P. J. Organomet. Chem. 2001, 635,
187−196.
̀ ́
(82) Zeng, X.; Coquiere, D.; Alenda, A.; Garrier, E.; Prange, T.; Li,
Y.; Reinaud, O.; Jabin, I. Chem.Eur. J. 2006, 12, 6393−6402.
(83) Zhang, W.; Liu, J.; Zhu, H.; Gao, W.; Sun, L. Synth. Commun.
2007, 37, 3393−3402.
R
DOI: 10.1021/acs.inorgchem.5b00509
Inorg. Chem. XXXX, XXX, XXX−XXX