Aryl-substituted BIAN complexes of iron dibromide: Synthesis, X-ray and electronic structure, and catalytic hydrosilylation activity

Article abstract of DOI:10.1016/j.poly.2016.02.020

Anhydrous iron dibromide complexes bearing bidentate α-diimine ligands ^ArN=C(Me)-(Me)C=N^Ar and ^ArBIAN (BIAN = bis(imino)acenaphthene; Ar = dpp and Mes; dpp = 2,6-diisopropylphenyl; Mes = 2,4,6-trimethylphenyl) have been pr

Full text of DOI:10.1016/j.poly.2016.02.020

Accepted Manuscript
Aryl-Substituted BIAN Complexes of Iron Dibromide: Synthesis, X-ray and
Electronic Structure, and Catalytic Hydrosilylation Activity
Michael J. Supej, Alexander Volkov, Louisa Darko, Ryan A. West, Jonathan
M. Darmon, Charles E. Schulz, Kraig A. Wheeler, Helen M. Hoyt
PII:
S0277-5387(16)00105-4
DOI:
http://dx.doi.org/10.1016/j.poly.2016.02.020
POLY 11835
Reference:
To appear in:
Polyhedron
Received Date:
Accepted Date:
26 October 2015
9 February 2016
Please cite this article as: M.J. Supej, A. Volkov, L. Darko, R.A. West, J.M. Darmon, C.E. Schulz, K.A. Wheeler,
H.M. Hoyt, Aryl-Substituted BIAN Complexes of Iron Dibromide: Synthesis, X-ray and Electronic Structure, and
Catalytic Hydrosilylation Activity, Polyhedron (2016), doi: http://dx.doi.org/10.1016/j.poly.2016.02.020
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Aryl-Substituted BIAN Complexes of Iron Dibromide: Synthesis, X-ray and Electronic
Structure, and Catalytic Hydrosilylation Activity
a
a
a
b
c
Michael J. Supej, Alexander Volkov, Louisa Darko, Ryan A. West, Jonathan M. Darmon,
b
d
a
Charles E. Schulz, Kraig A. Wheeler, Helen M. Hoyt*
a
Department of Chemistry, Knox College, 2 E South St, Galesburg, IL, 61401, United States
b
Department of Physics, Knox College, 2 E South St, Galesburg, IL, 61401, United States
c
Center for Molecular Electrocatalysis, Physical Sciences Division, Pacific Northwest National
Laboratory, P.O. Box 999, K2-57, Richland, Washington 99352, United States
d
Department of Chemistry, Eastern Illinois University, Charleston, Illinois 61920,United States.
*
Corresponding author. Address: Department of Chemistry, Knox College, 2 E South St,
Galesburg, IL, 61401, United States. Tel.: +1 309 341 7798.
E-mail address: hhoyt@knox.edu (H.M. Hoyt).
E-mail addresses: msupej@knox.edu (M.J. Supej); avolkov@knox.edu (A. Volkov);
lodarko@knox.edu (L. Darko); rwest@knox.edu (R.A. West), jonathan.darmon@pnnl.gov (J.M.
Darmon); cschulz@knox.edu (C.E. Schulz); kawheeler@eiu.edu (K.A. Wheeler);
hhoyt@knox.edu (H.M. Hoyt).
1

Ar
Abstract: Anhydrous iron dibromide complexes bearing bidentate α-diimine ligands N=C(Me)-
Ar
Ar
(
Me)C=N and BIAN (BIAN = bis(imino)acenaphthene; Ar = dpp and Mes; dpp = 2,6-
1
diisopropylphenyl; Mes = 2,4,6-trimethylphenyl) have been prepared and characterized by H
NMR spectroscopy. The aryl-substituted BIAN complexes were structurally characterized by
single-crystal X-ray diffraction, and their metrical parameters are consistent with a redox-
Ar
innocent chelating ligand. A high-spin iron(II) electronic structure description for the BIAN
iron complexes is supported by Mössbauer spectroscopy, solution state magnetic measurements,
and quantum-chemical calculations. Upon reduction, the iron complexes promote catalytic
o
hydrosilylation of 1-hexene with phenylsilane at 22 C.
Keywords: Iron; α-Diimine; Hydrosilylation; Catalysis; Bis(imino)acenaphthene
1
. Introduction
Transition-metal catalyzed alkene hydrosilylation is used commercially for fine chemical
synthesis and in the preparation of silicones for applications such as adhesives and sealing
materials [1,2]. Given that many commercial organosilicon (Si-C) products are derived from the
anti-Markovinikov addition of a tertiary silane (Si-H) to a terminal alkene (C=C), catalysts that
demonstrate high terminal selectivity are an important industrial concern [2]. Historically,
precious metal hydrosilylation catalysts such as Karstedt’s (Pt
2
{[(CH
2
=CH)SiMe
2
]
2
O}
3
) and
.
i
Speier’s (H
2
PtCl
6
6H
2
O/ PrOH) catalysts have been used industrially [2]. Recently, iron has
emerged as an attractive transition-metal catalyst for the hydrofunctionalization of terminal
olefins given that iron is abundant with low cost, low toxicity, and potential for sustainable long-
term commercial availability [3]. High selectivity and activity has been achieved by Chirik and
2

co-workers in the hydrosilylation of terminal alkenes with reduced iron catalysts bearing redox-
active ligands, such as the tridentate pyridinediimine (PDI, I) iron bis(dinitrogen) pre-catalysts
Ar
Ar
(
(
PDI)Fe(N
2
)
2
[4] and corresponding iron dialkyl pre-catalysts ( PDI)Fe(CH
2
SiMe
3
2
) [5]
Figure 1). Additionally, Nakazawa and co-workers have reported that activation of substituted
terpyridine (II) iron dibromide precatalysts selectively formed double hydrosilylation products
with phenylsilane and 1-octene [6], Ritter and co-workers have disclosed that bis(iminopyridine)
(
III) iron precatalysts promoted the 1,4-addition of tertiary silanes to dienes to produce
allylsilanes [7], and Huang and co-workers have demonstrated that phosphinite-iminopyridine
IV) iron complexes provided chemoselectivity in catalytic alkene hydrosilylation, tolerating
(
several carbonyl-containing functional groups [8] (Figure 1).
Figure 1. Representative ligands I-IV reported to support iron catalysts for the hydrosilylation of
terminal olefins.
In contrast to the primarily tridentate ligands shown in Figure 1, bidentate α-diimine
dpp
dpp
ligands, such as the parent N=C(Me)-(Me)C=N (dpp = 2,6-diisopropylphenyl) ligand [9,10]
Ar
and the aryl bis(imino)acenapthene ( BIAN) ligands (Figure 2), have so far been less well-
studied to support the iron-catalyzed hydrosilylation of terminal olefins. Metal systems
Ar
containing BIAN ligands are efficient catalysts in processes such as alkyne hydrogenation, C-C
coupling, and olefin polymerization [11]. Recently, Findlater and coworkers disclosed that a
dpp
6
reduced, toluene-stabilized iron catalyst,
7 8
BIANFe(η -C H ), successfully catalyzed the
chemo- and regioselective hydrosilylation of aldehydes and ketones with 1 mol% catalyst
3

o
loadings in excellent yields at 70 C [12]. In comparison to the parent α-diimine ligand, the
Ar
lower energy C=N π* orbitals for the conjugated BIAN ligands may be expected to act as
Ar
better π-accepting ligands to the iron metal center, allowing the BIAN ligands to better stabilize
higher oxidation states of the coordinated iron center [11]. In a traditional sense, these proposed
interactions can be considered as π-backbonding from the iron center, whereby electrons are
donated from the metal’s singly occupied orbital into an empty ligand π* orbital that is
delocalized throughout the conjugated π-system of the ligand backbone (Figure 2), effectively
stabilizing the complex and minimizing the formation of localized radical intermediates [13].
Ar
Figure 2. Delocalization of electrons in iron complexes supported by BIAN redox-active
ligands (R = aryl).
Building upon the previous work of Thomas and co-workers [14], we sought to develop
an operationally simple iron-catalyzed hydrosilylation of 1-hexene in which low-valent iron
catalysts were generated in-situ, with pre-catalysts prepared from relatively inexpensive
commercially available materials in two linear steps. To this end, new anhydrous iron dibromide
dpp
dpp Mes
Mes
complexes bearing bidentate α-diimine
N=C(Me)-(Me)C=N ,
N=C(Me)-(Me)C=N
,
dpp
Mes
BIAN, and
BIAN ligands (dpp = 2,6-diisopropylphenyl; Mes = 2,4,6-trimethylphenyl;
1
BIAN=bis(imino)acenaphthene) were prepared and characterized by H NMR spectroscopy. The
Ar
structures of the BIAN iron complexes were investigated by single crystal X-ray diffraction,
and the electronic structure of these pre-catalysts was elucidated by solution state magnetic
measurements, Mössbauer spectroscopy, and by DFT computational analysis. Subsequent in-situ
4

reduction of the iron complexes promoted the hydrosilylation of 1-hexene under mild and
solvent free conditions.
2
. Results and discussion
2
.1. Synthesis and spectroscopic characterization of iron complexes
Ar
The BIAN iron dibromide pre-catalysts were prepared in two linear steps from
1
commercially available starting materials and were characterized by IR and H NMR
Ar
spectroscopy. First, the substituted
BIAN ligands were prepared by acid-catalyzed
condensation of acenaphthenequinone (1) with two equivalents of the appropriate aniline (2a/2b)
Ar
following a similar protocol to that reported by Yuan et. al [15] (Figure 3). The BIANFeBr
2
pre-catalysts were then synthesized by addition of anhydrous iron(II) bromide to a solution of the
Ar
BIAN ligand in THF (3a) or CH
2
Cl
2
(3b) following a similar procedure described by Findlater
dpp
dpp
and coworkers for the analogous BIANFeCl
2
complex [12]. While BIANFeBr (4a) was
2
Mes
isolated from THF in high yield (95%), BIANFeBr (4b) could not be cleanly isolated when
2
prepared in this coordinating solvent, as THF was found to bind to the iron center and provided
.
an intractable mixture of 4b/4b THF based on broadened and shifted THF resonances at 5.93
1
and 4.09 ppm observed by H NMR spectroscopy (Appendix A, Figure A.4.1). Performing the
reaction in non-coordinating CH
2
Cl solvent allowed for isolation of purified 4b without a
2
coordinated Lewis base in reasonable yield (75%). The solid state IR C=N stretching frequencies
-1
-1
of 4a (1643, 1614 cm ) and 4b (1653, 1622 cm ) have lower C=N stretching frequencies than
Ar
the corresponding BIAN ligands (3a: 1668 cm ; 3b: 1674 cm ) and slightly higher frequencies
-1
-1
than are typical of α-diimine supported iron(II) dichloride species [RN=C(Ph)-C(Ph)=NR]FeCl
2
5

(
R = t-Bu, Cy) with non-conjugated backbones that display C=N stretches between 1600 and
-1
1
611 cm [16].
Ar
Figure 3. Synthesis of BIANFeBr
materials 1 and 2a/2b.
2
complexes 4a/4b from commercially available starting
dpp
Similarly, the parent α-diimine iron dibromide complexes [ N=C(Me)-
dpp
Mes
Mes
(
Me)C=N ]FeBr
2
(5a) and [ N=C(Me)-(Me)C=N ]FeBr (5b) (Figure 10, vide infra) have
2
been prepared in two linear steps from commercially available materials following an analogous
dpp
procedure to that reported by Chirik and coworkers for the preparation of [ N=C(Me)-
dpp
(
Me)C=N ]FeCl [9]. Upon addition of two equivalents of the appropriate aniline (2a/2b) to
2
2
,3-butanedione, each α-diimine ligand was isolated following an analogous procedure to that
reported by Yuan et. al [17], followed by subsequent addition of FeBr to the ligand in a solution
of either THF (5a) or CH Cl (5b) to form the desired parent α-diimine iron dibromide
2
2
2
complexes in excellent yields of 90% (5a) and 95% (5b). The structures of complexes 5a and 5b
1
were primarily characterized structurally by H NMR spectroscopy, and their spectra
demonstrated analogous features to those of 4a and 4b, respectively, as described below.
1
The contact-shifted resonances in the H NMR spectra for the paramagnetic iron
complexes are broadened and the chemical shifts are significantly more diffuse than those of the
corresponding diamagnetic ligands. Paramagnetic spectra are assigned primarily based on
6

integrated signal intensities and chemical shifts (Figure 4). For complex 4a, the diastereotopic,
magnetically inequivalent isopropyl methyl groups (due to hindered rotation around the N-Caryl
bonds) were found to integrate to 12H each at -4.02 and 2.77 ppm, and the p-Ar (2H) and m-Ar
(
4H) signals were observed at -18.53 and 5.64 ppm, respectively (Figure 4a). With the same N-
1
aryl diisopropylphenyl substituent, complex 5a demonstrates similar features in its H NMR
spectrum to complex 4a, such as resonances for the diasterotopic isopropyl methyl groups at -
3
.70 and -5.75 ppm, and the p-Ar and m-Ar signals were observed at -16.65 and 7.38 ppm,
respectively (Figure 4c). The isopropyl CH resonances for both 4a and 5a were not located under
these NMR experiment parameters as is typical for iron complexes with orthogonal ligands
bearing ortho i-Pr substituents [9]; these resonances are likely to be significantly broadened due
1
to their close proximity to the iron center. In the H NMR spectrum of complex 4b, the o- and p-
Me peaks of the N-mesityl substituent were located at 11.18 ppm (12H) and 16.18 ppm (6H),
respectively (Figure 4b), and the m-Ar (4H) signal was observed at 5.93 ppm. With the same N-
1
mesityl substituent, complex 5b demonstrates similar features in its H NMR spectrum to
complex 4b, such as the resonances for the o- and p-Me peaks located at 7.53 and 15.36 ppm,
respectively, and the m-Ar signal was observed at 2.55 ppm (Figure 4d). For both 5a and 5b, the
backbone imine methyl groups are located the farthest from their diamagnetic reference values,
with Cim-Me peaks at 74.37 ppm (5a) and 118.20 ppm (5b). For both 4a and 4b, the
acenaphthene backbone aryl resonances are shifted both upfield and downfield slightly from
their corresponding diamagnetic reference values. For all four iron complexes, the peaks
attributed to the ortho substituents are the most paramagnetically broadened, with peak-width at
half-height (400 MHz spectrometer) measurements greater than 100 Hz, compared to other peaks
7

with these values typically below 50 Hz; this is presumably due to the locations of these
hydrogen atoms with respect to the iron metal center (vide infra).
1
Figure 4. Comparison of H NMR spectra (400 MHz, CD
2
Cl
2
) for (a) 4a; (b) 4b; (c) 5a; (d) 5b;
peaks for 4a and 5a
(
C
im-Me peaks at 74.37 ppm (5a) and 118.20 ppm (5b) not shown; CHMe
2
not located).
2
.2 X-ray structures of iron complexes 4a and 4b
Single crystals of 4a and 4b suitable for X-ray diffraction analysis were obtained from
dichloromethane at -32 °C. Table 1 includes pertinent data collection and refinement parameters
for 4a and 4b, while selected bond distances and angles are provided in Table 2. Figure 5
provides molecular views and selected dihedral angles for 4a (left) and 4b (right), illustrating the
distorted pseudo-tetrahedral geometry of both complexes with τ
4
values (0.90 for 4a and 0.91 for
8

4
b) consistent with an idealized tetrahedral (τ
4
= 1.00) rather than an idealized square planar (τ
4
=
0.00) shape [18]. Both complexes exhibit smaller N-Fe-N’ angles than expected for a complex
dpp
with tetrahedral geometry (78.7(2)° (4a); 78.87(19)° (4b); 78.61(14)° ( BIANFeCl
2
) [11]),
while correspondingly larger Br-Fe-Br’ angles (117.03(6)° (4a); 115.39(5)°(4b)) are observed.
In part, these relatively smaller angles may be due to the constraints imposed by the tied-back,
planar BIAN ligand backbone, as a relatively larger N-Fe-N’ angle is provided to the 5-
membered chelate for a less constrained α-diimine ligand backbone such as that observed for the
[
CyN=C(Ph)-C(Ph)=NCy]FeCl
2
complex (N-Fe-N’: 119.24(9)°) [15]. In both complexes 4a and
4
b, the aryl substituents are oriented orthogonally to the plane in which the ligand backbone is
o
situated, as shown by the C-CArN-Cim dihedral angles near to 90 for both complexes (Figure 5).
Table 1. Selected crystallographic and refinement data for 4a and 4b.
4
a
4b
Empirical Formula
Formula Weight
Crystal color, Habit
Crystal dimensions (mm)
Crystal system
Space group
Z
C
36
H
40Br
2
FeN
2
C
30
H
28Br
2
FeN
2
716.38
green, plate
0.18 x 0.14 x 0.08
monoclinic
P21/
8
632.21
red, plate
0.38 x 0.12 x 0.10
orthorhombic
Pccn
c
8
a (Å)
b (Å)
27.1504(13)
12.0306(6)
22.8626(9)
113.511(3)
6847.8(6)
1.390
22.6321(7)
16.7681(5)
17.2755(6)
90
6556.0(4)
1.281
c (Å)
o
β ( )
V (Å )
3
-3
Dcalc (g cm )
-1
µ (CuKα) (cm )
T (K)
6.450
100(2)
6.670
100(2)
λ (Å)
1.54178
1.54178
Reflections with
I > 2.0σ(I)a
14377
5607
Residuals: R; R
w
; Rall
0.0619; 0.1394; 0.0845
1.062
0.0738; 0.1631; 0.0779
1.231
Ga oodness of Fit Indicator
Function minimized: Σ ω (|Fo| - |Fc|) ; least squares weights: 4a, ω = 1/[σ (F
2
2
2
2
o
) + (0.0262ρ) +
2
2
2
2
2
2
7.9032ρ] and 4b, ω = 1/[σ (F
o
) + (0.0256ρ) + 45.5069ρ]; ρ = (F
o
+ 2F )/3.
c
9

o
Table 2. Selected distances (Å) and angles ( ) for 4a and 4b with computed values included in
square brackets (vide infra).
4
a
4b
Fe-Br
Fe-Br’
Fe-N
2.3434(14) [2.399]
2.3708(15) [2.401]
2.140(6) [2.221]
2.3649(12) [2.400]
2.3582(12) [2.401]
2.137(5) [2.230]
Fe-N’
N-Cim
N’-Cim
2.142(6) [2.226]
1.277(9) [1.279]
1.287(9) [1.278]
2.142(5) [2.238]
1.273(8) [1.279]
1.275(8) [1.280]
’
Cim-Cim
’
1.508(10) [1.521]
1.518(9) [1.523]
N-Fe-N’
Br-Fe-Br’
N-Fe-Br
N-Fe-Br’
N’-Fe-Br
N’-Fe-Br’
78.7(2) [76.72]
78.87(19) [76.77]
115.39(5) [123.55]
113.74(14) [113.67]
113.96(14) [108.73]
116.78(14) [112.46]
112.99(14) [112.14]
117.03(6) [126.85]
113.30(17) [103.65]
113.69(19) [117.28]
115.79(18) [113.90]
112.47(18) [107.60]
Figure 5. Molecular views illustrating the geometry and atom labeling scheme with the
o
hydrogen atoms omitted for clarity for 4a (left); selected dihedral angles ( ) listed as absolute
values: C(14B)-C(13B)-N(1B)-C(1B): 80.5(9), C(18B)-C(13B)-N(1B)-C(1B): 96.8(9), C(20B)-
C(19B)-N(2B)-C(12B): 91.2(9), C(24B)-C(19B)-N(2B)-C(12B): 87.8(9); and 4b (right); 4b
o
(
right); selected dihedral angles ( ) listed as absolute values: C(14)-C(13)-N(1)-C(1): 82.7(7),
C(18)-C(13)-N(1)-C(1): 97.9(7), C(20)-C(19)-N(2)-C(12): 95.4(7), C(24)-C(19)-N(2)-C(12):
5.8(7).
8
The bond lengths of 4a and 4b are similar to other reported α-diimine iron dihalide
complexes, such as [CyN=C(Ph)-C(Ph)=NCy]FeCl [16]. In general, the Fe-N bond lengths are
longer in complexes 4a (2.140(6) Å; 2.142(6) Å), 4b (2.137(5) Å; 2.142(5) Å), and
2
10

dpp
BIANFeCl
2
(2.132(4) Å; 2.140(4) Å) [12] than in the [CyN=C(Ph)-C(Ph)=NCy]FeCl
2
complex (2.077(3) Å; 2.077(3) Å) [16]. This may result from decreased σ-donation of the
Ar
BIAN ligands due to lower overlap with the iron orbitals given the relatively smaller N-Fe-N’
angles required by the tied-back BIAN ligand backbone as compared to the a less constrained α-
diimine ligand backbone in [CyN=C(Ph)-C(Ph)=NCy]FeCl . The position of the aryl substituents
2
orthogonal to the N-Fe-N’ plane may play a role in the observed differences in catalytic
Mes
reactivity of 4a and 4b (vide infra). The
BIAN ligand for 4b allows for a more open
...
...
coordination sphere (non-bonded contacts: 4b: C(25) Br(1) 4.040 Å; 4a: C(32B) Br(2B) 3.893
Å) about the metal center due to the smaller methyl groups as compared to the larger isopropyl
dpp
Mes
groups on the BIAN ligand. Due to steric effects, the BIAN ligand may allow additional
ligands to bind to the metal center, such as THF, which was observed to occur on preparation of
Mes
.
.
the proposed BIANFeBr THF (4b THF) complex in THF solvent (vide supra).
2
Ar
While BIAN ligands have been reported to accept electrons when treated with reducing
agents such as Na, or when bound to metal centers of varied oxidation states [19,20], the bond
Ar
lengths for the BIAN ligands in complexes 4a and 4b are consistent with the redox-innocent
Ar
13] form of the ligand. When compared to other metal systems supported by BIAN ligands,
[
the Cim=N and Cim-Cim bond lengths found in 4a (C=N: 1.277(9) Å; 1.287(9) Å; C-C: 1.508(10)
dpp
Å), 4b (C=N: 1.273(8) Å; 1.275(8) Å; C-C: 1.518(9) Å), and BIANFeCl (C=N: 1.283(7) Å;
2
dpp
0
0
C-C: 1.506(6) Å) [12] show similarities to [ BIAN] Pd complexes (such as V in Figure 6)
dpp
0
bearing a non-reduced, redox-innocent [ BIAN] ligand (C=N: 1.276 Å; 1.297 Å; C-C: 1.506
Ar
Å) [18]. However, in [ BIAN] Mg complexes such as VII, the bond lengths (C-N: 1.401 Å;
2-
II
1
.408 Å; C=C: 1.389 Å) [19] indicate that the ligand can be described as an ene-diamide
Ar
structure with expanded C-N and contracted C-C bonds. Crystallographic data for [ BIAN]
1-
11

…
…
dpp
1- II
bound to ferrous iodide (C N: 1.328 Å; 1.330 Å; C C: 1.437 Å) in the [( BIAN) Fe (µ-I)]
2
Ar
dimer VI [20] has shown that the BIAN can stabilize a ligand radical anion through resonance
1-
…
…
as the C N and C C bond lengths are intermediate between the redox-innocent α-diimine
Ar
0
Ar
BIAN] and ene-diamide [ BIAN] forms of the ligand. The agreement of the Cim=N and Cim-
2-
[
dpp
C
im bond lengths for 4a and 4b with the non-chelated BIAN (C=N: 1.28 Å; C-C: 1.53 Å) [21]
Mes
and BIAN (C=N: 1.27 Å; C-C: 1.528 Å) [22] ligands, respectively, further indicates that these
Ar
complexes 4a and 4b contain redox-innocent BIAN ligands.
Ar
Figure 6. Metal systems V-VII supported by BIAN ligands of different oxidation states
[
19,20].
2
.3 Electronic structure of bis(imino)acenaphthene iron complexes
Ar
The electronic structures of the BIAN iron dibromide complexes 4a and 4b were
investigated using a combination of metrical parameters from X-ray diffraction (vide supra),
solid state Mössbauer spectroscopy, solution state magnetic measurements, and DFT
calculations. These studies were undertaken to compare the electronic structures of complexes 4a
dpp
6
and 4b to the reduced hydrosilylation catalyst
7 8
BIANFe(η -C H ). Full molecule DFT
dpp
6
computational analysis for pre-catalysts 4a and 4b, along with
7 8
BIANFe(η -C H ), is
12

dpp
presented below to elucidate the redox-activity of this BIAN ligand in the catalyzed in situ
hydrosilylation of 1-hexene with phenylsilane (vide infra).
2
.3.1 Spin state & magnetism of bis(imino)acenaphthene iron dibromide complexes
5
7
Zero-field Fe Mössbauer spectra were collected and analyzed at 25 K, 100 K, and 150
K for 4a and 25 K, 100 K, and 200 K for 4b (Figure 7; Appendix A). The experimentally
determined (25 K) isomer shifts (δ) and quadrupole splittings (ꢀ∆E ꢀ) are reported in Table 3.
Also reported in Table 3 are DFT-computed (vide infra) parameters for both compounds. It
Q
5
7
should be noted that varying the temperature for these zero-field Fe Mössbauer studies did not
produce any significant variation in isomer shifts or quadrupole splittings over the temperatures
-1
-1
investigated. The isomer shifts (0.88 mm s ) and quadrupole splittings (2.73 mm s (4a); 2.76
-1
mm s (4b)) are nearly identical for 4a and 4b, suggesting a similar electronic structure [23]
consistent with a high spin iron(II) metal center [16,24]. This data compares favorably to the
-1
-1
parameters of δ = 0.84 mm s and ꢀ∆E
Q
ꢀ = 2.67 mm s reported by Wieghardt and coworkers for
dpp
dpp
[
N=C(Me)-(Me)C=N ]FeCl [25]. Solution state magnetic moments were determined for
2
both 4a and 4b by Evans’ NMR method using a ferrocene standard and result from the average
value of at least two independent measurements. The resulting magnetic measurements taken at
2
95 K provided effective magnetic moments of 4.8 µ
B
(4a) and 5.1 µ (4b). These data are
B
consistent with an S = 2 ground state with a high spin Fe(II) center and a redox-neutral chelating
Ar
0
[
BIAN] ligand.
13

Figure 7. Mössbauer spectra including Lorentzian fits at 25 K for (a) 4a and (b) 4b. See
Appendix A for spectra recorded at higher temperatures.
5
7
Table 3. Comparison of experimental and computationally determined zero-field Fe
α
Mössbauer parameters
-1
compound
δ (mm s )
-1
ꢀ
∆E
Q
ꢀ (mm s )
dpp
BIANFeBr
2
(4a)
(4b)
0.88
2.73
2.52
2.76
2.64
0
.79
0.88
.80
Mes
BIANFeBr
2
0
α
The experimental values were measured at 25 K. Computed values are presented in italics.
2
.3.2 Quantum chemical computations
Full molecule density functional theory (DFT) calculations on BIANFeBr
BIANFeBr (4b) were performed at the B3LYP level of theory to provide additional insight
dpp
2
(4a) and
Mes
2
Ar
into the electronic structure description of these BIAN iron dibromide complexes. The
accuracy of the computations was assessed by comparison of the calculated and experimental
5
7
metrical parameters (Table 2) along with the experimental and computed Fe Mössbauer
14

parameters (Table 3) for each compound (vide supra). On the basis of the magnetic data, two
possible electronic structure models were computationally investigated that are consistent with
the experimentally determined S = 2 ground state for both complexes: (a) a high-spin Fe(II)
Ar
center with a redox-innocent ligand [ BIAN] , (d , SFe = 2, SBIAN = 0) and (b) a high spin Fe(III)
0
6
5
center (d , SFe = 5/2) antiferromagnetically coupled to a mono-reduced ligand bearing a
Ar
. 1-
delocalized radical anion [ BIAN ] (SBIAN = -1/2). These electronic structure models were
investigated with (a) a spin-unrestricted quintet (UKS) computational approach and (b) a broken
symmetry (BS) approach denoted as BS(5,1), respectively. Ground state calculations at the
B3LYP level of DFT were performed for both models, and in both complexes the computations
converged to the appropriate UKS solution. That both approaches converged to the same solution
Ar
0
indicates that a high spin Fe(II) center surrounded by a closed shell ligand, [ BIAN] , is the best
description for the electronic structure of both compounds. The experimentally obtained,
geometric metrical parameters were acceptably reproduced [26] by the spin-unrestricted
computational model (Table 2). The computed Mössbauer isomer shifts and quadrupole
splittings were also found to be within the generally accepted range established for these
comparisons [27] to the experimental parameters (Table 3). Additionally, numerical frequency
computations were performed at the same level of theory to confirm that a global energy
minimum had been located and that there were no imaginary frequencies for both complexes.
The agreement of the computed UKS parameters to those measured experimentally provides
further support for the description of these complexes as high-spin ferrous centers supported by
Ar
redox-innocent BIAN ligands.
dpp
Qualitative molecular orbital diagrams and spin density plots of BIANFeBr and
2
Mes
BIANFeBr derived from spin-unrestricted quintet model are depicted in Appendix A (4a)
2
15

and Figure 8 (4b), respectively. The molecular orbital diagrams and spin density plots are
remarkably similar, showing that the unpaired spins reside primarily on the high-spin ferrous
Ar
center rather than on the BIAN ligand. The spin density was calculated to be 3.68 a.u. for both
complexes and is consistent with the majority of the spin density residing on the high spin
iron(II) metal center with very minor contributions from the bromine and nitrogen atoms bound
Ar
to the iron. The combined experimental and computational data establish that the BIAN ligands
are redox-innocent in these iron dibromide complexes. Taken together, these data provide a basis
Ar
of comparison for other iron complexes supported by BIAN ligands that behave in a redox-
active fashion.
Figure 8. Qualitative molecular orbital diagram of 4b obtained from a spin-unrestricted (UKS)
DFT calculation at the B3LYP level (left); spin density plot obtained from a Mulliken population
analysis (right).
Two computational models were investigated for the experimentally determined singlet
dpp
6
7 8
ground state of BIANFe(η -C H ) [12] using full molecule DFT calculations. A closed-shell
16

spin-restricted (RKS) model, corresponding to a highly covalent, singlet ground state in which
the BIAN ligand acts as a traditional π-acid, was explored. Additionally, because the
.
1-
2-
experimentally determined ligand bond lengths are between those of [BIAN ] and [BIAN] ,
…
…
with a C C distance of 1.400(4) Å and a C N distance of 1.343(3) Å [12], a spin-unrestricted
dpp
2-
II
6
BS(2,2) model was also examined. This model describes the ( BIAN )Fe (η -C
7
8
H ) ground
II
6
state in which an intermediate-spin Fe ion (d , SFe = 1) antiferromagentically couples to a triplet
diradical dianion (SBIAN = -1).
Geometry optimizations were performed at the B3LYP level of theory on both models
and produced two solutions. The broken-symmetry calculation converged to a BS(1,1) solution,
dpp
1-
I
6
I
7
which corresponds to a ( BIAN )Fe (η -C
7
H ) ground state in which a low-spin Fe ion (d , SFe
8
.
1-
=
½) antiferromagenetically couples with a [BIAN ] radical anion (SBIAN = -½). Qualitative
molecular orbital diagrams for the broken-symmetry and RKS solutions are presented in Figure 9
and Figure A.3.2 (Appendix A), respectively. In the BS(1,1) solution, the overlap between the
two magnetic orbitals gives an overlap integral (S) of 0.54, and results in a 5.4 kcal/mol energy
stabilization relative to the RKS solution. The geometric parameters of the two solutions are
presented in Table 4, and both reasonably reproduce the experimental values [12]. Interestingly,
the RKS solution shows better agreement for the metal-ligand and intraligand distances.
5
7
To further distinguish between the two possible models, Fe Mössbauer parameters were
calculated for comparison with the experimental values (δ = 0.45 mm/s, ∣∆E ∣ = 0.41 mm/s)
12]. While the isomer shift obtained from the RKS solution (δ = 0.46 mm/s) is in excellent
Q
[
agreement with the experimental results, the value obtained from the BS(1,1) solution (δ = 0.61
mm/s) is less accurate and outside the generally accepted range [27]. The quadrupole splitting
(
Q
∆E ) is not acceptably reproduced by either the RKS or BS(1,1) solution, yielding values of
17

1
.51 and -1.52 mm/s, respectively. As the high isomer shift obtained from the BS(1,1) solution is
likely a result of the over estimation of the metal-ligand bond lengths, the broken-symmetry
calculation was repeated using the experimentally obtained geometric parameters [12]. The
isomer shift obtained from this calculation (δ = 0.43 mm/s) acceptably reproduces the
experimental value, while the quadrupole splitting (∆E = -1.08 mm/s) is still outside of the
Q
generally accepted range. Therefore, neither solution can be excluded on the basis of the
Mössbauer parameter calculations. However, it is important to note that the RKS and BS(1,1)
calculations yield asymmetry parameters of the electric field gradient, η, of 0.54 and 0.66,
respectively, indicating that applied field Mössbauer spectroscopy would allow for a more
definitive electronic structure assignment. On the basis of the significant energy stabilization
dpp
1-
I
6
obtained from symmetry breaking [28], we favor the redox-active, ( BIAN )Fe (η -C
7
8
H )
electronic structure description.
Table 4. Comparison of experimentally [12] and computationally determined geometric
dpp
6
7 8
parameters (Å) for BIANFe(η -C H ).
Experimental [12]
1.400(4)
RKS
1.415
1.347
1.346
1.918
1.917
2.156
2.137
2.142
2.124
2.131
2.115
BS(1,1)
1.433
1.338
1.336
1.971
1.973
2.190
2.151
2.145
2.136
2.141
2.156
imine
imine
imine
imine
C
C
–C
–N
1.343(3)
1
.344(3)
1.899(2)
.903(2)
imine
Fe–N
1
CH3
Fe–C
Fe–C
2.104(3)
2.083(3)
ortho
2
.080(3)
2.083(3)
.081(3)
2.082(3)
meta
Fe–C
Fe–C
2
para
18

dpp
6
Figure 9. Qualitative molecular orbital diagram of
BIANFe(η -C
7 8
H ) obtained from a
BS(2,2) DFT calculation at the B3LYP level; this calculation converged to a BS(1,1) solution.
2
.4 Catalytic hydrosilylation of 1-hexene with phenylsilane
While bidentate α-diimine ligands have yet to rival the tridentate PDI ligands for the
support of highly active iron catalysts for industrially viable applications in hydrosilylation
Ar
4,14], we hypothesized that the greater conjugation of BIAN ligands may provide a more
[
Ar
Ar
active iron catalyst than the parent α-diimine ligands N=C(Me)-(Me)C=N
in the
Ar
hydrosilylation of a terminal alkene. As monoreduced [ PDI] ligands have been proposed to
1-
play a role in stabilizing catalytic intermediates in related cycloisomerization reactions [29], we
Ar
reasoned that the greater conjugation provided by the BIAN ligand backbone may similarly
Ar
stabilize catalytic intermediates bearing monoreduced [ BIAN] as compared to parent
1-
Ar
Ar 1-
[
N=C(Me)-(Me)C=N ] ligands. Thomas and co-workers have reported that solvent-free, in
19

Et
situ activation of pre-catalyst PDIFeCl
2
(Et = 2,6-diethylphenyl) with n-BuLi produced gram-
at ambient temperature in high yield [14].
scale hydrosilylation of styrene with Et SiH
2
2
Building upon the previous work of Thomas and coworkers [14], we developed an
operationally simple iron-catalyzed hydrosilylation of 1-hexene in which reduced iron catalysts
were generated in situ. To this end, 2.5 mol% of iron pre-catalysts LFeBr
Figure 10) were activated in situ with 5 mol% LiCH SiMe for 24 h in THF. After removing
volatile materials in vacuo, neat substrates 1-hexene and PhSiH were added, and the 1-hexyl-
2
(4a, 4b, 5a, or 5b
2
3
3
phenylsilane product was isolated in high yield for pre-catalysts 4a (99%) and 5a (95%) and low
yield for pre-catalysts 4b (16%) and 5b (37%). The isolated yields reported in Figure 10 were
1
determined by H NMR spectroscopy upon comparison of product integration intensities to an
external standard of 1,3,5-trimethoxybenzene [14]. Of the four pre-catalysts investigated, 4a
produced the highest yield with a turnover number (TON) of 10.
Figure 10. Catalytic hydrosilylation of 1-hexene with phenylsilane at ambient temperature upon
in situ activation of pre-catalysts 4a/b and 5a/b with LiCH
2
SiMe .
3
20

Given the higher yields for pre-catalyst 4a, we sought to further optimize these conditions
by decreasing catalyst loading and exploring alternate reducing agents. To further optimize this
method, 1 mol% pre-catalyst 4a was reduced in situ with 2 mol% LiCH
2
SiMe activator in THF,
3
and this was found to catalyze the solvent-free hydrosilylation of 1-hexene (1 equiv) with
phenylsilane (1 equiv), producing 1-hexylphenylsilane in a 95% yield after 24 h at ambient
temperature with an improved TON of 48. Several different activators were investigated for this
1
method by measuring the H NMR yield of 1-hexylphenylsilane following the conditions
described in Figure 10, including EtMgBr (3%), NaHBEt (12%), and in situ-generated sodium
3
naphthalenide (24%). However, when the activation solvent used was toluene, the yield of 1-
hexylphenylsilane was 99% with in situ-generated sodium naphthalenide as the reducing agent,
notably identical to the best conditions presented in Figure 10 with the LiCH
2
SiMe reducing
3
agent. This suggests that both reducing agents may be forming the same active catalyst. The
crude activated catalyst mixture, resulting from the addition of 2 equiv of sodium napthalenide to
1
4
a in toluene for 24 h, was analyzed by H NMR spectroscopy in C
6
6
D , demonstrating that
dpp 6
BIANFe(η -C
7 8
H ) was predominantly formed [12] (Figure A.4.2, Appendix A). Findlater and
dpp
6
BIANFe(η -C
coworkers have demonstrated that isolated
7
H
8
) effectively catalyzes the
o
hydrosilylation of ketones and aldehydes at elevated temperature (70 C) [12], while Chirik and
coworkers have shown that isolated PDIFe(CH
2
SiMe
3
2
) complexes successfully catalyze the
o
dpp 6
BIANFe(η -
hydrosilylation of olefins at elevated temperatures of 60 C [5]. Presumably,
6
C
7
H
8
)
would undergo dissociation of the η -toluene ligand while proposed
dpp
BIANFe(CH
2
SiMe
3
2
) would likely proceed by reductive elimination to enter into a similar
catalytic cycle.
21

Comparison of the yields shown in Figure 10 may indicate certain features of the
bidentate α-diimine pre-catalysts that promote enhanced activity. The N-substituents appear to
play a significant role in the activity of these α-diimine iron catalysts, as the dpp substituents of
4
5
a and 5a presumably lead to significantly higher activity than the N-Mes substituents for 4b and
b. This observed lower activity for 4b in particular may be attributed to the in situ formation of
.
stable, ligated iron complexes such as 4b THF (vide supra), which may inhibit pre-catalyst
activation and/or catalytic turnover. As pre-catalyst 4a produced 1-hexyl-phenylsilane in a higher
yield than did pre-catalyst 5a under similar reaction conditions, the more conjugated backbone of
dpp
the BIAN ligand modestly enhanced the catalytic activity of the iron pre-catalyst as compared
dpp
dpp
dpp
1-
I
to the N=C(Me)-(Me)C=N ligand. This suggests that [ BIAN] Fe intermediates may be
dpp
dpp 1-
I
relatively stabilized compared to [ N=C(Me)-(Me)C=N ] Fe intermediates due to increased
dpp
dpp
2- II
conjugation of the
BIAN backbone. It should be noted that [ BIAN] Fe catalytic
2
-
intermediates would likely be relatively destabilized by the 10 π electrons (2n) of the BIAN
dpp
dpp 2-
ene-diamide form of the ligand (Figure 2) as compared to possible [ N=C(Me)-(Me)C=N ]
II
Fe intermediates, suggesting that the ene-diamide forms of these ligands are less likely to play a
significant role in stabilizing key catalytic intermediates. Overall, the N-dpp substituent leads to
dpp
more active catalysts than the N-Mes substituent, and the BIAN ligand leads to modestly
higher catalytic activity of iron complexes in the hydosilylation of 1-hexene with PhSiH as
3
compared to the parent α-diimine ligand.
3
. Conclusions
Anhydrous iron dibromide complexes (4a, 4b, 5a, 5b) bearing bidentate α-diimine
1
ligands have been prepared and characterized by H NMR spectroscopy. The aryl-substituted
22

BIAN complexes 4a and 4b were structurally characterized by single-crystal X-ray diffraction,
Ar
and their metrical parameters are consistent with a redox innocent [ BIAN] chelating ligand for
0
Ar
both complexes. High-spin iron(II) electronic structure descriptions for the 4a and 4b BIAN
iron pre-catalysts are supported by Mössbauer spectroscopy, solution state magnetic
measurements indicating an S = 2 ground state, and DFT quantum-chemical calculations. Upon
in situ reduction with LiCH
2
SiMe , the iron complexes promote catalytic hydrosilylation of 1-
3
hexene with phenylsilane at ambient temperature, with highest yields provided by pre-catalyst
dpp
BIANFeBr
2
(4a) followed closely by parent α-diimine pre-catalyst 5a. The smaller N-
Mes
substituents for
BIANFeBr (4b) and parent α-diimine pre-catalyst 5b may lead to the
2
formation of less active iron complexes with a higher coordination number. In situ reduction of
dpp
6
BIANFe(η -C
4
a with sodium naphthalenide produced
7
H
8
), and this also promoted
dpp 6
BIANFe(η -C
successful hydrosilylation. DFT computational analysis of
7
H
8
) suggests a
dpp
1-
I
6
redox-active, ( BIAN )Fe (η -C
7
H
8
) electronic structure description. Future studies focus on
Ar
defining the substrate scope for α-diimine- and BIAN-supported iron-catalyzed olefin
hydrosilylation reactions.
4
. Experimental
4
.1. General considerations
All air- and moisture-sensitive manipulations were carried out using standard vacuum
line and Schlenk techniques or in a Vacuum Atmospheres inert (purified nitrogen) atmosphere
o
glovebox. Glassware was dried overnight at 150 C or flame dried under vacuum immediately
o
prior to use. Celite, sieves, and other solid supports were dried in an oven for 1 week at 150 C
o
and were then further dried under vacuum at 150 C for 24 h. Solvents for air- and moisture-
23

sensitive manipulations were dried and deoxygenated either using a Vacuum Atmospheres
Solvent Purification System or were vacuum transferred from CaH and deoxygenated using
freeze-pump-thaw methods. Deuterated solvents (Cambridge Isotope Laboratories), 1-hexene
Sigma-Aldrich), and phenylsilane (Sigma-Aldrich) were either deoxygenated using freeze-
pump-thaw methods and dried over 4A molecular sieves, or were dried over CaH (Sigma-
2
(
2
Aldrich) and then vacuum transferred and deoxygenated using freeze-pump-thaw methods prior
to use [30]. Iron(II)bromide, acenaphthenquinone, 2,3-butanedione, 2,6-diisopropylaniline, and
2
,4,6-trimethylaniline were purchased from Sigma-Aldrich and were used as received. Formic
acid was purchased from Flinn Scientific and was used as received. Ferrocene (Sigma-Aldrich)
was sublimed under reduced pressure prior to use. A 1.0 M solution of LiCH
2
SiMe in pentane
3
o
(
Sigma-Aldrich) was crystallized at -10 C prior to use.
1
13
H and C NMR spectra were recorded on a Bruker Avance III 400 MHz spectrometer.
1
1
All H chemical shifts (δ) are reported relative to SiMe
4
using the H (residual) shift of the
solvent as a secondary standard. Peak width at half height is given for paramagnetically
broadened resonances. Solution magnetic moments were determined by Evans’ NMR method
[
31] using a ferrocene standard and result from the average value of at least two independent
1
3
1
measurements. C{ H} NMR chemical shifts (δ) are reported in ppm relative to the carbon
resonance of the deuterated solvent. Infrared (IR) spectra were taken as KBr films using a
PerkinElmer FTIR Spectrometer model FRONTIER MIR +SP10 STD. Elemental analyses were
performed by Robertson Microlit Laboratories, Inc., in Madison, NJ. Melting points were
measured in flame-sealed capillary tubes under a static vacuum using a Mel-Temp II.
24

4
.2. Preparation of α-diimine ligands
Ar
Preparation of the BIAN ligands followed analogous literature procedures to those
described in the literature [12,15] by acid-catalyzed condensation of the respective anilines with
1
13
acenaphthenequinone and were recrystallized from EtOH. The H and C NMR data for
dpp
Mes
BIAN (3a) [11] and BIAN (3b) [22] are consistent with those reported in the literature. The
dpp
dpp
Mes
Mes
α-diimine ligands N=C(Me)-C(Me)=N [9, 32] and
N=C(Me)-C(Me)=N
[17] were
prepared according to literature procedures.
dpp
.2.1. Preparation of bis[N,N’-(2,6-diisopropylphenyl)imino]acenaphthene ( BIAN, 3a)
4
To a slurry of acenapthenequinone (1) (9.386 g, 51.5 mmol) and formic acid (1.0 mL) in
ethanol (700 mL) was added 2,6-diisopropylaniline (2a) (19.75 mL, 104.7 mmol) in ethanol (200
o
mL) dropwise over 7 days at 65 C. The reaction mixture was cooled to ambient temperature and
the orange solid was collected by vacuum filtration. The remaining ethanol solution was
o
1
concentrated in vacuo and cooled to -12 C to give crystallized 3a (23.36 g, 91%); H NMR (400
MHz, CDCl ): δ 7.86 (d, J = 8.3 Hz, 2H, Aracen), 7.36 (t, J = 7.8 Hz, 2H, Aracen), 7.3-7.2 (m, 6H,
Ar), 6.64 (d, J = 7.3 Hz, 2H, Ar), 3.03 (sept, J = 6.9 Hz, 4H, CHMe ), 1.24 (d, J = 6.9 Hz, 12H,
MeCHMe), 0.98 (d, J = 6.9 Hz, 12H, MeCHMe); C{ H} NMR (100 MHz, CDCl ): δ 160.99,
47.51, 140.82, 135.44, 131.12, 129.51, 128.89, 127.89, 124.30, 123.48, 123.36, 28.63
3
2
1
3
1
3
1
-1
(
MeCHMe), 23.44 (MeCHMe), 23.15 (MeCHMe); IR, v/cm : 3065, 2961, 2925, 2867, 1668,
1
590, 1466, 1455, 1428.
Mes
.2.2. Preparation of bis[N,N’-(2,4,6-trimethylphenyl)imino]acenaphthene ( BIAN, 3b)
4
To a slurry of acenapthenequinone (1) (5.801 g, 31.84 mmol) and formic acid (1.0 mL) in
ethanol (400 mL) was added 2,4,6-trimethylaniline (2b) (9.00 mL, 64.1 mmol) in ethanol (200
25

o
mL) dropwise over 6 days at 65 C. The resulting mixture was cooled to ambient temperature
and the solid was collected by vacuum filtration. The remaining solution was concentrated in
o
vacuo, cooled to -12 C for crystallization, and then vacuum filtered to collect 3b (9.512 g, 72%)
1
as an red-orange solid; H NMR (400 MHz, CDCl
3
): δ 7.87 (d, J = 8.3 Hz, 2H), 7.4-7.3 (m, 2H),
6
.98 (s, 4H, m-Ar), 6.79 (d, J = 7.2 Hz, 2H), 2.38 (s, 6H, p-Me-Ar), 2.11 (s, 12H, o-Me-Ar);
1
3
1
C{ H} NMR (100 MHz, CDCl ): δ 161.12, 146.83, 140.58, 132.84, 131.04, 129.74, 128.97,
3
-1
1
1
28.82, 128.26, 124.63, 122.51, 20.98, 17.76; IR, v/cm : 2966, 2941, 2908, 2851, 1673, 1647,
601, 1593, 1472, 1420, 1437.
4
4
.3. Preparation of iron compounds
.3.1. Preparation of bis[N,N’-(2,6-diisopropylphenyl)imino]acenaphthene iron(II) bromide
dpp
(
2
BIANFeBr , 4a)
dpp
In an inert atmosphere glovebox, a solution of BIAN (3a) (3.426 g, 6.846 mmol) and
FeBr
2
(1.529 g, 7.090 mmol) in THF (100 mL) was stirred for 3 days at ambient temperature.
The solution was then filtered through celite to remove any insoluble material and the volatile
materials were removed under vacuum to afford 4a (4.654 g, 95%) as a dark green solid.
Crystals suitable for X-ray diffraction analysis were obtained by cooling a solution of 4a in
o
1
CH
54 Hz, 4H, m-Ar), 3.02 (27 Hz, 2H, Aracen), 2.77 (44 Hz, 12H, MeCHMe), 2.13 (83 Hz, 2H,
Aracen), -4.02 (118 Hz, 12H, MeCHMe), -16.65 (52 Hz, 2H, p-Ar), one peak not located;
2
Cl
2
to -32 C; mp 350-351°C; H NMR (400 MHz, CD
2
Cl
2
) δ 13.64 (42 Hz, 2H, Aracen), 7.38
(
-
1
Magnetic susceptibility (Evans’ Method, 295 K): µeff = 4.8(3) µ
B
; IR (KBr), v/cm : 3066, 2962,
927, 2868, 1643, 1614, 1597, 1576, 1436, 1420; Anal. Calcd for C36 FeN : C, 60.36; H,
.63; N, 3.91. Found: C, 59.95; H, 5.91; N, 4.11; Mössbauer data were collected at 25 K, 100 K,
2
5
H40Br
2
2
and 150 K (Appendix A, Figure A.2.1).
26

4
.3.2. Preparation of bis[N,N’-(2,4,6-trimethylphenyl)imino]acenaphthene iron(II) bromide
Mes
(
BIANFeBr
In an inert atmosphere glovebox, a solution of BIAN (3b) (1.000 g, 2.401 mmol) and
(0.518 g, 2.402 mmol) in CH Cl (20 mL) was stirred for 2 days at ambient temperature.
2
, 4b)
Mes
FeBr
2
2
2
The solution was then filtered through celite to remove any insoluble material and the volatile
Mes
materials were removed under vacuum to afford
BIANFeBr (4b) (1.136 g, 75%) as a dark
2
red/brown powder. Crystals suitable for X-ray diffraction analysis were obtained by cooling a
o
solution of 4b in CH
95 K): µeff = 5.0(6) µ
Hz, 2H, Aracen), 11.18 (189 Hz, 12H, o-Me-Ar), 5.93 (19 Hz, 4H, m-Ar), 0.46 (16 Hz, 2H,
2
Cl
2
to -32 C; mp 355-357 °C; Magnetic susceptibility (Evans’ Method,
1
2
B
; H NMR (400 MHz, CD
2
Cl ) δ 16.18 (4 Hz, 6H, p-Me-Ar), 14.92 (17
2
-1
Aracen), 0.06 (40 Hz, 2H, Aracen); IR (KBr), v/cm : 3014, 2945, 2915, 2860, 1653, 1622, 1602,
585, 1478, 1419, 1386, 1356; Anal. Calcd for C30 FeN : C, 56.99; H, 4.46; N, 4.43.
Found: C, 56.72; H, 4.49; N, 4.35; Mössbauer data were collected at 25 K, 100 K, and 200 K
Appendix A, Figure A.2.2).
1
H28Br
2
2
(
dpp
.3.3. Preparation of [ N=C(Me)-C(Me)=N ] iron(II) bromide ([ N=C(Me)-
dpp
dpp
4
dpp
C(Me)=N ]FeBr
2
, 5a)
This procedure follows that described in the literature for the analogous iron dichloride
dpp
dpp
complex [ N=C(Me)-C(Me)=N ]FeCl
2
[9]. In an inert atmosphere glovebox, a solution of
dpp
dpp
ligand N=C(Me)-C(Me)=N [9, 32] (1.617 g, 4.00 mmol) and FeBr (0.893 g, 4.14 mmol) in
2
THF (50 mL) was stirred for 3 days at ambient temperature. The solution was then filtered
through celite to remove any insoluble material and the volatile materials were removed under
1
vacuum to afford 5a (2.234 g, 90%) as a dark blue powder; mp 345-346 °C; H NMR (400 MHz,
27

CD
2
Cl
2
) δ 74.37 (165 Hz, 6H, CimMe), 5.64 (28 Hz, 4H, m-Ar), 3.70 (42 Hz, 12H, MeCHMe), -
-
5
.75 (100 Hz, 12H, MeCHMe), -18.53 (21 Hz, 2H, p-Ar), one peak not located; IR (KBr), v/cm
1
:
3062, 2964, 2924, 2867, 1631, 1587, 1542, 1464, 1442, 1374, 1324, 1291; Anal. Calcd for
FeN : C, 54.22; H, 6.50; N, 4.52. Found: C, 54.48; H, 6.51; N, 4.48.
C
28
H40Br
2
2
Mes
.3.4. Preparation of [ N=C(Me)-C(Me)=N ] iron(II) bromide ([ N=C(Me)-
Mes
Mes
4
Mes
2
C(Me)=N ]FeBr , 5b)
Mes
Mes
In an inert atmosphere glovebox, a solution of ligand
N=C(Me)-C(Me)=N
[17]
(
5.357 g, 16.72 mmol) and FeBr (3.605 g, 16.72 mmol) in CH Cl
2
2
2
(50 mL) was stirred for 1 day
at ambient temperature. The solution was then filtered to remove any insoluble material and the
volatile materials were removed under vacuum to afford 5b (8.466 g, 95%) as a dark green
1
powder; mp 286-287 °C; H NMR (400 MHz, CD
2
Cl ) δ 118.20 (131 Hz, 6H, CimMe), 15.36 (4
2
-1
Hz, 6H, p-Me-Ar), 7.53 (184 Hz, 12H, o-Me-Ar), 2.55 (20 Hz, 4H, m-Ar); IR (KBr), v/cm :
2
4
080, 1819, 1774, 1740, 1629, 1378, 1290, 1222, 1158; Anal. Calcd for C22
9.29; H, 5.26; N, 5.23. Found: C, 49.34; H, 5.27; N, 5.24.
H28Br
2
FeN : C,
2
4
.4. Representative procedure for catalytic hydrosilylation
dpp
To a mixture of pre-catalyst BIANFeBr (4a) (0.150 g, 0.209 mmol) in THF (3 mL), a
2
solution of LiCH
2
SiMe (0.039 g, 0.42 mmol) in THF (3 mL) was added dropwise with stirring,
3
and the reaction mixture was allowed to stir for 24 h at ambient temperature in an inert
atmosphere glovebox. After the volatile materials were removed in vacuo, phenylsilane (0.76
mL, 6.2 mmol) and 1-hexene (0.26 mL, 2.1 mmol) were added and the reaction mixture was
allowed to stir for 24 h at rt. The reaction mixture was then exposed to air, extracted with n-
28

pentane (10 mL), filtered through Celite (3 cm pipet column), and the volatile materials were
removed in vacuo. The organic residue was then dissolved in C
6
D
6
, and 1,3,5-trimethoxybenzene
1
1
(
0.0235 g, 0.140 mmol) was added as a H NMR integration standard. Comparison of the H
NMR integration values for a diagnostic 1-hexyl-phenylsilane peak in the product (4.46 ppm,
3
2
H, t, JHH = 3.7 Hz, CH
2
SiH Ph) [33] with the peaks for the integration standard [14] allowed
2
for calculation of a hydrosilylation yield of 97% (trial 1) and 100% (trial 2). Data reported in
Figure 10 represent an average of at least two independent trials. In each case, 2.5 mol% Fe was
used, given mol % = (mol Fe/total mol substrates)*100% [4].
4
.5. Single crystal X-ray diffraction
Single crystals suitable for X-ray diffraction were coated with degassed Paratone-N oil
(
Fisher Scientific) in an inert atmosphere glovebox and shipped under nitrogen to Eastern Illinois
University (EIU). At the EIU integrative applied crystallography center, the crystals were
quickly transferred to the Bruker APEXII CCD diffractometer and data were collected using phi
and omega scans with graphite monochromatic Cu Kα (λ = 1.54178 Å) radiation. Data sets were
corrected for Lorentz and polarization effects. The criterion for observed reflections is I > 2σ(I).
Lattice parameters were determined from least-squares analysis and reflection data. Empirical
absorption corrections were applied using TWINABS (4a) and SADABS (4b) [34]. Structures
2
were solved by direct methods and refined by full-matrix least-squares analysis on F using X-
SEED [35] equipped with SHELX-2014/7 [36]. All non-hydrogen atoms for 4a and 4b were
2
refined anisotropically unless specified by full-matrix least-squares on F by the use of the
SHELX-2014/7-XL [36] program. The diffuse electron density in the crystal lattice of 4b
29