Zhiyuan He1,2, Lingyu Liu1, Felix J de Zwart2, Xiaolian Xue1, Andreas W Ehlers2,3, KaKing Yan1, Serhiy Demeshko4, Jarl Ivar van der Vlugt2,5, Bas de Bruin2, Jeremy Krogman1. 1. School of Physical Science and Technology, ShanghaiTech University, Shanghai 201210, China. 2. van't Hoff Institute for Molecular Sciences, University of Amsterdam, Science Park 904, 1098 XH Amsterdam, The Netherlands. 3. Department of Chemistry, University of Johannesburg, Auckland Park, Johannesburg P.O. Box 254, ZA-2006, South Africa. 4. Department of Chemistry, Georg August University, Tammanstraße 4, 37077 Göttingen, Germany. 5. Institute of Chemistry, Carl von Ossietzky University, Carl-von-Ossietzky-Straße 9-11, 12629 Oldenburg, Germany.
Abstract
In this paper, we report the synthesis of a unique silicon(I)-based metalla-disilirane and report on its reactivity toward TMS-azide and benzophenone. Metal complexes containing disilylenes ((bis)silylenes with a Si-Si bond) are known, but direct ligation of the Si(I) centers to transition metals always generated dinuclear species. To overcome this problem, we targeted the formation of a mononuclear iron(0)-silicon(I)-based disilylene complex via templated synthesis, starting with ligation of two Si(II) centers to iron(II), followed by a two-step reduction. The DFT structure of the resulting η2-disilylene-iron complex reveals metal-to-silicon π-back donation and a delocalized three-center-two-electron (3c-2e) aromatic system. The Si(I)-Si(I) bond displays unusual but well-defined reactivity. With TMS-azide, both the initial azide adduct and the follow-up four-membered nitrene complex could be isolated. Reaction with benzophenone led to selective 1,4-addition into the Si-Si bond. This work reveals that selective reactions of Si(I)-Si(I) bonds are made possible by metal ligation.
In this paper, we report the synthesis of a unique silicon(I)-based metalla-disilirane and report on its reactivity toward TMS-azide and benzophenone. Metal complexes containing disilylenes ((bis)silylenes with a Si-Si bond) are known, but direct ligation of the Si(I) centers to transition metals always generated dinuclear species. To overcome this problem, we targeted the formation of a mononuclear iron(0)-silicon(I)-based disilylene complex via templated synthesis, starting with ligation of two Si(II) centers to iron(II), followed by a two-step reduction. The DFT structure of the resulting η2-disilylene-iron complex reveals metal-to-silicon π-back donation and a delocalized three-center-two-electron (3c-2e) aromatic system. The Si(I)-Si(I) bond displays unusual but well-defined reactivity. With TMS-azide, both the initial azide adduct and the follow-up four-membered nitrene complex could be isolated. Reaction with benzophenone led to selective 1,4-addition into the Si-Si bond. This work reveals that selective reactions of Si(I)-Si(I) bonds are made possible by metal ligation.
Transition metal silylene complexes (Scheme a) have attracted
significant interest as
they have shown interesting (electronic) structures and reactivity,
giving rise to synthetic and catalytic applications that differ significantly
from transition metal carbene complexes.[1] Several variants of transition metal silylene complexes have been
reported, varying in the oxidation state of the silicon atom and the
types of substituents, and among those variants, N-heterocyclic silylenes have been studied most extensively.[2]
Scheme 1
(a) Selection of Different Monosilylenes;
(b) Difference between
“Disilylene” and “Spacer-Separated Bis(silylene)”
and Examples of Metal Complexes of the Latter
Two types of the general class of bis(silylene)
compounds have
been reported: (1) “bis(silylenes) with a direct Si–Si
bond”, wherein the two divalent silicon atoms are adjacent
to each other and are connected by a central Si–Si bond (like
others in the field,[3] we term these “disilylenes”);
(2) “spacer-separated bis(silylenes)”, with the two
divalent silicon atoms separated by a spacer.[4] Driess and co-workers recently reported a ferrocene-separated bis(silylene)
acting as a bidentate ligand in catalytically active mono- and dicobalt
complexes (Scheme b).[4e]We consider “disilylenes”
to be particularly interesting
due to the potential reactivity associated with their Si–Si
bond. Compound I, with each Si center bearing an amidinate
ligand (Scheme ),
was synthesized by Roesky and co-workers,[5] while Jones et al. reported a bulkier derivative
thereof.[6] Additionally, “disilylenes”
that are stabilized by an N-heterocyclic carbene
(NHC) or an intramolecular phosphine are also reported.[4a,7] Targeting the synthesis of η2-coordinated Si(I)-based
“disilylene” complexes seems particularly useful because
the altered electronic structure and the lability of the Si–Si
bond induced by η2-coordination are expected to give
rise to unique reactivity.
Scheme 2
(a) Reported Di- and Multinuclear Disilylene
Transition Metal Complexes;
(b) Synthesis and Reactivity of a Mononuclear η2-Disilylene
Iron Complex (This Work)
Thus far, the reported reactivity of these platforms
predominantly
involves the Si-centered lone pairs rather than the Si–Si σ-bond.
For example, So et al. reported the formation of
the bimetallic disilylene Rh and Ir complexes II and III (Scheme a).[8] Also, the dinuclear iron(II)bromide
complex IV is formed cleanly, but the Si–Si bond
is cleaved by cobalt(II)bromide to form dinuclear cobalt(bromosilylene)(silyl)
complex V.[9] To the best of
our knowledge, no mononuclear metal adducts of ligand I are reported to date.To achieve η2-coordination
of I to
a mononuclear transition metal complex, we decided to explore the
reductive coupling of two Si(II)-chlorides in the coordination sphere
of a metal center (Scheme b). We herein describe iron complex 3 as the
first η2-disilylene complex prepared by this protocol
(Scheme ). This species
has an unusual electronic structure, with the Si–Si bond acting
as a π-acceptor moiety, while the three-membered Fe–Si–Si
ring shows 2π-aromaticity. It also displays selective Si–Si
bond-centered reactivity toward azide activation and ketone addition,
giving access to new “spacer-separated bis(silylenes)”.
Scheme 3
Synthetic Procedure for the Preparation of Complex 3
Results and Discussion
Synthesis of 3
The amidinate-stabilized
silylene iron-halide precursor 1, synthesized from Fe{N(SiMe3)2}2 using a literature procedure,[10] was characterized with zero-field 57Fe Mössbauer spectroscopy (Figure ). The isomer shift δ (0.61 m/s) and
the quadrupole splitting ΔEQ (2.81
mm/s) are in agreement with a four-coordinated Fe(II) center.[11] Iron-centered reduction of this well-defined
Fe(II) species with a slight excess of KC8 in a benzene-THF
mixture (1:1 v/v%) generated bis-silylene compound 2,
featuring an η6-benzene fragment bound to Fe(0).
Figure 1
Mössbauer
spectra of compounds 1 (left) and 2 (right)
at 80 K.
Mössbauer
spectra of compounds 1 (left) and 2 (right)
at 80 K.The 1H NMR spectrum of diamagnetic 2 in
deuterated benzene exhibits a chemical shift for the η6-benzene fragment at δ 5.15 ppm, which is similar to that seen
for the previously reported zero-valent iron complex [(SiFcSi)Fe-η6(C6H6)] bearing a ferrocene-bridged
bis(silylene) ligand (δ 5.16 ppm).[12] The 29Si NMR spectrum of 2 exhibits two
sharp singlets at δ 45.12 and 42.45 ppm (suggesting chemically
inequivalent Si centers), which are close to the reported chemical
shift (δ 43.10 ppm) for an N-heterocyclic silylene
iron(0) complex.[13] Species 2 was also examined using zero-field 57Fe Mössbauer
spectroscopy to confirm the oxidation state of Fe (Figure ). Indeed, both the isomer
shift δ (0.38 m/s) and quadrupole splitting ΔEQ (1.53 mm/s) support reduction to Fe(0).[14]Single crystals of 2 that were obtained
from pentane
at −30 °C proved suitable for crystallographic analysis
by X-ray diffraction at room temperature; loss of crystallinity was
observed at 150 K. The molecular structure (Figure ) shows two different orientations for the
Si–Cl bonds with respect to the Fe(η6-benzene)
fragment (one “up” and one “down”), which
explains the presence of two chemically inequivalent Si nuclei in
the 29Si NMR spectrum. The Si–Fe–Si angle
is nearly 90° (91.99°), and the intramolecular Si1···Si2
distance is 3.100(2) Å. The Si–Fe1 bond lengths are 2.162(2)
Å (Si1) and 2.148(1) Å (Si2), which are significantly shorter
than the corresponding bond lengths in precursor 1 (∼2.445
Å) and slightly shorter than those found in the zero-valent iron
silylene complexes ([(NHSi)Fe(dmpe)2] (2.184(2) Å)
and Fe(0)[SiNSi] (2.164(15) and 2.170(13) Å)).[15]
Figure 2
Molecular structure of 2 with thermal ellipsoids drawn
at 30% probability. Hydrogen atoms are omitted for clarity. Selected
bond lengths (Å) and angles (°): Si1–Fe1, 2.162(2);
Si2–Fe1, 2.148(1); Si1–N1, 1.860(5); Si1–N2,
1.866(5); Si1–Si2, 3.100(2); Si1–Fe1–Si2, 91.99(6).
Molecular structure of 2 with thermal ellipsoids drawn
at 30% probability. Hydrogen atoms are omitted for clarity. Selected
bond lengths (Å) and angles (°): Si1–Fe1, 2.162(2);
Si2–Fe1, 2.148(1); Si1–N1, 1.860(5); Si1–N2,
1.866(5); Si1–Si2, 3.100(2); Si1–Fe1–Si2, 91.99(6).Further reduction of this zero-valent Fe species 2 by an excess of KC8 at room temperature yielded
the three-membered
metallacyclic complex 3, featuring a direct Si–Si
bond (Scheme ). This
reaction was also monitored using 1H NMR spectroscopy,
revealing quantitative conversion (see also Figure S26). The complex was characterized by multinuclear NMR spectroscopy
(1H, 13C, and 29Si). The 1H NMR signal of the Fe-coordinated η6-benzene shifts
from δ 5.15 ppm in 2 to δ 5.34 ppm in 3. The 29Si NMR spectrum of 3 contains
only one singlet at δ 34.49 ppm, indicating formation of a symmetric
molecule.Dark red single crystals of 3 were collected
after
storing a pentane solution at −30 °C for 1 month. Compound 3 crystallizes in the monoclinic space group C2/c, which is in line with the high degree
of symmetry observed in the 1H NMR spectrum. The Si(I)–Fe–Si(I)
three-membered heterocycle forms an equilateral triangle with interatomic
distances of 2.217(6) Å (Figure ); for the Si–Si bond, this value falls in the
range observed for Si=Si double bonds (2.120–2.250 Å).[16] The isosceles Si–B–Si disilaborirane
compound reported by Roesky and co-workers shows a Si–Si bond
length of 2.188(5) Å,[17] while the
equilateral triangular (di-t-butyl(methyl)silyl)bis(tri-t-butylsilyl)cyclotrisilenylium cation has an average Si–Si
bond length of 2.217(3) Å.[16]
Figure 3
Molecular structure
of 3 with thermal ellipsoids drawn
at 50% probability. Hydrogen atoms are omitted for clarity. Selected
bond lengths (Å) and angles (°): Si1–Si1 and Si1–Fe1,
2.217(6); Si1–N1, 1.865(2); Si1–N2, 1.887(2); Si1–Fe1–Si1,
59.99; Fe1–Si1–Si1, 60.00.
Molecular structure
of 3 with thermal ellipsoids drawn
at 50% probability. Hydrogen atoms are omitted for clarity. Selected
bond lengths (Å) and angles (°): Si1–Si1 and Si1–Fe1,
2.217(6); Si1–N1, 1.865(2); Si1–N2, 1.887(2); Si1–Fe1–Si1,
59.99; Fe1–Si1–Si1, 60.00.
Electronic Structure of 3
Given the unprecedented
nature of the ferracyclic motif found in this species, the bonding
of the Si(I)–Si(I) fragment to iron(0) was theoretically investigated
at the TZ2P/OPBE[18] level of theory using
energy decomposition analysis (EDA).[19] The
lowest energy structure with C2 symmetry
is analogous to that derived from X-ray diffraction. To facilitate
the EDA analysis (for symmetry reasons), the bonding of the disilylene
moiety to iron has been analyzed using a C2v optimized structure, which is only 4.4 kcal·mol–1 higher in energy. This small energy difference is consistent with
the dynamic behavior observed in solution-state NMR spectroscopy,
resulting in a single signal for the Bu substituents at nitrogen.The Fe(η6-benzene)
fragment has been computed in the singlet state (Fe(0)-d8). It has a two-electron occupied d orbital (perpendicular to the Si–Fe–Si plane)
and an empty d orbital. The electronic
structure of the Si(I)–Si(I) fragment is best described as
having a Si–Si single bond with a lone pair on each silicon
atom and an empty π-orbital perpendicular to that plane (Figure S31). The symmetry-adapted linear combinations
of the two lone pairs on silicon donate electron density toward the
Fe(η6-benzene) fragment, forming two (delocalized)
σ-bonds of −30.3 kcal·mol–1 (A1)
and −74.4 kcal·mol–1 (B1). The iron
itself donates electron density back from the occupied d orbital into an empty π-orbital formed by
the two Si p orbitals (B2: −67.8
kcal·mol–1), giving rise to considerable π-backbonding
(Figure ). To probe
the resulting orbital (which is delocalized over the Si–Fe–Si
three-membered ring) for aromaticity, we employed the nucleus-independent
chemical shift (NICS(0/1)) approach at the B3LYP/6-11+G(d,p) level
(negative value of the isotropic magnetic shielding at the center
of the Fe–Si–Si three-membered ring and 1 Å perpendicularly
above and below, respectively). These are found to be −47.8
and −21.4 ppm, respectively, corroborating the magnetic aromaticity.
For comparison, the NICS(0/1) values at the center of the benzene
ring in 3 are calculated to be −42.8 and −18.4
ppm. Similar values have been reported for other metallacycles.[20] A more refined method by Stanger[21] following the out-of-plane component to the
shielding tensor along a trajectory orthogonal to the plane of the
ring (NICS) has been used by Roesky et al. to assign 2π-aromaticity to their disilaborirane
species.[17] However, instead of showing
the typical off-center minimum for 2π-aromatic systems, the
NICS scan of 3 shows a steady and continuous increase
to less negative NICS values over a
value of 3 Å (see Figure S33), which
may be caused by anisotropy of the metal center at the ring and/or
by σ-aromaticity.[17,19,22]
Figure 4
Graphical
representation of the Fe-to-(Si–Si) π-back donation in 3.
Graphical
representation of the Fe-to-(Si–Si) π-back donation in 3.Therefore, we resorted to the canonical molecular
orbital (CMO)
analysis of the NICS(0), which separates the total shieldings into
contributions from canonical molecular orbitals.[23] Indeed, there is a sizable contribution of −13.8
ppm from the Fe d orbital (HOMO–5,
see Figure S32), which is part of the σ-framework.
More importantly, the major contribution of −16.6 ppm originates
from the delocalized π-orbital shown in Figure , substantiating the 2π-aromaticity
of the Si2Fe three-membered ring in complex 3.
Reactivity of 3
The unligated Si(I)-based
disilylene species I shows stoichiometry-dependent reactivity
toward trimethylsilyl azide, forming either a silaazatriene or a silaimine
product.[24,25] Roesky and co-workers described ring opening
of their disilaborirane species VI with TMSN3, forming a 1-aza-2,3-disila-4-boretidine derivative (Scheme ).[17] We observed selective conversion of 3 with 1 equiv
of TMSN3 to give the unique azide adduct 4 after 1 h at 40 °C in benzene solution. This species was obtained
as a single crystalline material by recrystallization from pentane
at −30 °C. The molecular structure, as determined by X-ray
diffraction, is displayed in Figure . It contains a planar four-membered Si1–Fe1–Si2–N5
heterocycle resulting from the insertion of the azide terminal N-atom
in the Si–Si bond, with Fe–Si–N5 angles of roughly
100°. The two Si–Fe distances are slightly different,
Si1/2–Fe1 (2.1687(6) and 2.1746(7) Å). The intramolecular
Si···Si distance is long at 2.504(8) Å.
Scheme 4
Synthetic Scheme for the Preparation of Azide Adduct 4 and Final Product 5 upon Reaction of 3 with Trimethylsilyl Azide
Figure 5
X-ray crystal
structures of 4 (left) and 5 (right) with
thermal ellipsoids drawn at 50% probability. Hydrogen
atoms are omitted for clarity. Selected bond lengths (Å) and
angles (°) of 4: Si1···Si2, 2.5040
(8); Si1–Fe1, 2.1687(6); Si2–Fe1, 2.1746(7); Si1–N5,
1.784(2); Si2–N5, 1.791(2); Si1–Fe1–Si1, 70.41(2);
Fe1–Si1–N5, 100.57(6); Fe1–Si2–N5, 100.11(6);
Si1–N5–Si2, 88.91(8); 5: Si1···Si2,
2.412(2); Si1–Fe1, 2.174(2); Si2–Fe1, 2.177(2); Si1–N5,
1.770(5); Si2–N5, 1.767(4); Si1–Fe1–Si1, 67.33(6);
Fe1–Si1–N5, 103.30(1); Fe1–Si2–N5, 103.30(1);
Si1–N5–Si2, 86.00(2).
X-ray crystal
structures of 4 (left) and 5 (right) with
thermal ellipsoids drawn at 50% probability. Hydrogen
atoms are omitted for clarity. Selected bond lengths (Å) and
angles (°) of 4: Si1···Si2, 2.5040
(8); Si1–Fe1, 2.1687(6); Si2–Fe1, 2.1746(7); Si1–N5,
1.784(2); Si2–N5, 1.791(2); Si1–Fe1–Si1, 70.41(2);
Fe1–Si1–N5, 100.57(6); Fe1–Si2–N5, 100.11(6);
Si1–N5–Si2, 88.91(8); 5: Si1···Si2,
2.412(2); Si1–Fe1, 2.174(2); Si2–Fe1, 2.177(2); Si1–N5,
1.770(5); Si2–N5, 1.767(4); Si1–Fe1–Si1, 67.33(6);
Fe1–Si1–N5, 103.30(1); Fe1–Si2–N5, 103.30(1);
Si1–N5–Si2, 86.00(2).Prolonged heating of a mixture of 3 with TMSN3 or isolated 4 at 80 °C generated
the follow-up
trimethylsilylnitrene complex 5 as the major product,
together with small amounts of intermediate 4 (see Figure S17). Single crystals of 5, obtained by recrystallization from benzene at room temperature,
were analyzed by X-ray diffraction (Figure ). Compound 5 also shows a planar
four-membered Si1–Fe1–Si2–N5 heterocycle with
Fe–Si–N5 angles of 103(1)° but with a different
substitution pattern at N5, i.e., only a TMS group resulting from
the insertion of trimethylsilylnitrene in the Si–Si bond. The
two Si–Fe distances are nearly identical, Si1/2–Fe1
(2.174(2) and 2.177(2) Å), and so are the Si1/2–N5 (1.770(5)
and 1.767(4) Å) bond lengths. The Si–Fe σ-bonds
are on average 0.2547 Å longer than the Si–B distance
in the N-insertion product obtained from disilaborirane,[18] while the Si–N5 bond length is similar.
The intramolecular Si···Si distance (2.412(2) Å)
in 5 is shortened relative to that in 4.
Apart from the X-ray crystallographic analysis, reaction monitoring
and product identification have also been achieved using NMR spectroscopy.
The 1H NMR signal of the Fe-bound η6-benzene
fragment shifts from δ 5.34 ppm (3) to δ
5.26 ppm (4) and 5.12 ppm (5), and the trimethylsilyl
group in azide adduct 4 appears at δ 0.47 ppm (for 5, δ 0.39 ppm) compared to −0.08 ppm for TMSN3. In 29Si NMR, the NSiN signals
for both 4 and 5 are strongly upfield-shifted
(Δδ ∼ 27 ppm) with respect to 3 (4: δ 7.05 ppm; 5: δ 7.42 ppm), in
line with the rupture of the Si–Si bond, which also disrupts
the 2π-aromaticity and leads to loss of ring current. The Si(CH3)3 signal appears at δ
15.41 ppm for 4 and at −20.71 ppm for 5.To understand the observed two-step reaction between 3 and trimethylsilylazide, we used DFT calculations to explore
the
reaction mechanism. The initial step involves the nucleophilic attack
of the azide onto one of the Si centers, which results in induced
nucleophilic character at the second Si center that subsequently attacks
back onto the azide in TS1 (Figure ). As a result, this insertion can be considered
to involve induced FLP reactivity. The energy barrier to afford Int1 (15.4 kcal·mol–1) is consistent
with the experimentally determined barrier (15.1 kcal·mol–1) for formation of 4 (Arrhenius plot, Figure ). After forming
this symmetric azide adduct, the energy barrier for the subsequent
dinitrogen release (29.4 kcal·mol–1) is high
enough to rationalize the successful isolation of 4.
Finally, a Staudinger pathway, involving four-membered ring transition
state TS3, releases dinitrogen from 4 to
afford product 5 (−110 kcal·mol–1).
Figure 6
Computed free energy profile for nitrene formation from trimethylsilyl
azide and complex 3 (B3LYP-TZVP and def2-TZVP). Transient
bonds in transition states are drawn as dashed lines. The inset shows
the Arrhenius plot for the consumption of 3 (for details,
see the Supporting Information).
Computed free energy profile for nitrene formation from trimethylsilyl
azide and complex 3 (B3LYP-TZVP and def2-TZVP). Transient
bonds in transition states are drawn as dashed lines. The inset shows
the Arrhenius plot for the consumption of 3 (for details,
see the Supporting Information).The ambivalent reactivity of the Si(I) centers
in the Si–Si
bond of 3, able to act either as nucleophile or electrophile,
was also apparent during the conversion of species 3 conversion
with benzophenone. Disilylene I was previously reported
to react with benzophenone to furnish selective C–O cleavage
with formation of a cyclodisiloxane (Scheme ).[26] Strikingly
different reactivity was observed when complex 3, featuring
a “protected” Si–Si bond, was exposed to benzophenone
for 1 h at 80 °C. Selective formation of seven-membered ring
product 6 was obtained via a formal 1,4-addition of benzophenone,
displaying strongly attenuated and controlled reactivity of the Si–Si
fragment in complex 3. Complex 6 was characterized
in the solid state using single-crystal X-ray diffraction (Figure ). The structure
consists of a nonplanar seven-membered O–Si–Fe–Si–C–C–C
ring (angle between the planes O1–C37–C44–C49
and O1–Si1–Fe1–Si2 is 7.20°). The C37–C44
bond length (1.348(5) Å) lies in between that of a typical single
carbon bond and a double carbon bond (1.510–1.317 Å).[27] The Si–Fe–Si angle (∠88.21(4)°)
is larger than in species 4 and 5 because
of increased steric hindrance. One of the phenyl rings has undergone ortho-silylation, concomitant with ring de-aromatization
and formation of an enolate-type fragment bound via the oxygen to
the second Si center. As a result, overall 1,4-addition of benzophenone
has occurred, with no sign of the 1,2-C,O addition product. The 1H NMR signal for the η6-benzene fragment
shifts from δ 5.34 ppm (3) to δ 5.07 ppm
(6). The C–H hydrogen at the ortho-silylated position resonates at around δ 6.99 ppm according
to 2D-COSY NMR spectroscopy, while the other four hydrogens of the
de-aromatized ring appear in the range of 4.90–6.20 ppm. The 29Si NMR spectrum of 6 exhibits two sharp singlets
at δ 65.57 and δ 33.66 ppm, in accordance with the two
different bonding features.
Scheme 5
Synthetic Scheme for the Preparation of 6
Figure 7
X-ray crystal structure of 6 with
thermal ellipsoids
drawn at 50% probability. Hydrogen atoms are omitted for clarity.
Selected bond lengths (Å) and angles (°) of 6: Si1···Si2, 3.101(1); Si1–Fe1, 2.146(9); Si2–Fe1,
2.179(9); Si1–O1, 1.688(2); O1–C37, 1.384(3); C37–C44,
1.348(5); C44–C49, 1.503(5); C49–Si2, 1.983(4); Si1–Fe1–Si1,
88.21(4).
X-ray crystal structure of 6 with
thermal ellipsoids
drawn at 50% probability. Hydrogen atoms are omitted for clarity.
Selected bond lengths (Å) and angles (°) of 6: Si1···Si2, 3.101(1); Si1–Fe1, 2.146(9); Si2–Fe1,
2.179(9); Si1–O1, 1.688(2); O1–C37, 1.384(3); C37–C44,
1.348(5); C44–C49, 1.503(5); C49–Si2, 1.983(4); Si1–Fe1–Si1,
88.21(4).
Conclusions
In summary, we have developed an efficient
method to synthesize
the first example of a mononuclear transition metal complex bearing
a ligated Si(I)–Si(I) disilylene fragment, ferracyclic η2-disilylene complex 3. The electronic structure
of this species shows that the Si–Si fragment acts as a four-electron
σ-donor to iron, while significant π-back donation from
the iron(0) center to the silicon atoms of the disilylene moiety leads
to further stabilization of the overall structure. Complex 3 shows well-defined nucleophile-induced FLP reactivity toward TMS-azide
and benzophenone, leading to Si–Si bond cleavage by addition
of the reagents to the Si2 fragment, generating unexpected
four- and seven-membered ring structures. Expanding this unique reactivity
to other small molecules is currently being explored within our groups.
Experimental Section
Materials and Methods
Unless otherwise stated, all
manipulations were performed under a nitrogen atmosphere using Schlenk
techniques or in a Vigor glovebox maintained at or below 1 ppm of
O2 and H2O. All new metal complexes were prepared
and handled in the glovebox under a N2 atmosphere. Anhydrous
FeCl2 (98%) was purchased from Strem Chemicals. PhC(NBu)2SiHCl2,[28] LiN(SiMe3)2(Et2O),[28] Fe(N(SiMe3)2)2,[29] and complex 1 (FeCl2{PhC(NBu)2SiCl}2)[10] were synthesized
according to reported procedures. Other reagents were purchased from
J&K Chemical and SCRC. Glassware was dried at 150 °C overnight.
Celite and molecular sieves were dried at 200 °C under vacuum.
Benzene, pentane, hexanes, and diethyl ether were degassed with nitrogen,
dried over activated molecular sieves, and kept over 4 Å molecular
sieves in a N2-filled glovebox. NMR data were recorded
either on a Bruker 400 or a 500 MHz spectrometer and were internally
referenced to residual proton solvent signals in C6D6 (7.16 ppm). Data for 1H NMR are reported as follows:
chemical shift (δ ppm) and multiplicity (s = singlet, d = doublet,
t = triplet, m = multiplet, br = broad). IR data were recorded on
a Thermo Scientific Nicolet iS5 FTIR, and signal strength is represented
as follows: VS = very strong, W = weak, S = strong, VW = very weak,
m = middle, w = wide. The UV–Vis spectra were recorded using
a StellarNet BLACK Comet C-SR diode array miniature spectrophotometer
connected to deuterium and halogen lamp by optical fiber using 1 cm
matched quartz cuvettes at room temperature. Elemental analysis was
performed by the Analytical Laboratory of Shanghai Institute of Organic
Chemistry (CAS).
X-ray Crystallography
Crystals were coated with Paratone-N
oil and mounted on a Bruker D8 Venture diffractometer equipped with
an APEX-II CCD diffractometer. The crystal was kept at 150 K during
data collection. Using Olex2,[30] the structure
was solved with the ShelXT[31] structure
solution program using Intrinsic Phasing and refined with the XL[32] refinement package using least squares minimization.
CCDC 2157512–2157516 contain the supplementary crystallographic data
for this paper. These data can be obtained free of charge from The
Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.uk/data_request/cif.
Mössbauer Spectroscopy
Mössbauer spectra
were recorded with a 57Co source in a Rh matrix using an
alternating constant acceleration Wissel Mössbauer spectrometer
operated in transmission mode and equipped with a Janis closed-cycle
helium cryostat. Isomer shifts are given relative to the iron metal
at ambient temperature. Simulation of the experimental data was performed
with the Mfit program (developed by Dr. E. Bill, Max-Planck Institute
for Chemical Energy Conversion, Mülheim/Ruhr, Germany) using
Lorentzian line doublets.
Computational Details
Geometries were fully optimized
as minima or transition states using the Turbomole program package,[33] coupled to the PQS Baker optimizer[34] via the BOpt package.[35] We used unrestricted ri-DFT-D3 calculations at the B3LYP level,[36] in combination with the def2-TZVP basis set[37] and a small (m4) grid size. Grimme’s
dispersion corrections[38] (version 3, disp3,
“zero damping”) were used to include van der Waals interactions.
The energy decomposition analysis (EDA)[19] was performed on the TZ2P/OPBE[18] optimized
geometry constrained to C2v symmetry (+3.9
kcal·mol–1). The nucleus-independent chemical
shift (NICS)[39] was used as a diagnostic
probe for quantitative measure for aromaticity at the B3LYP/6-11+G(d,p)
level.[40,41] For further details, see the Supporting Information.
{PhC(NBu)2SiCl}2Fe(C6H6) (2)
A
solution of 1 (110 mg, 0.153 mmol) in benzene (5 mL)
was added dropwise to a solution of KC8 (53.7 mg, 0.398
mmol) in THF (5 mL) in a vial while stirring. The color of the reaction
mixture turned from yellow to dark red-brown. After stirring for 12
h, volatile materials were removed under vacuum and compound 2 was extracted with pentane solution. The solid was crystallized
in pentane solution in a −30 °C freezer for 2 days, and
only crystalline material was used for subsequent reactions (yield:
40 mg, 40%). 1H NMR (500 MHz, benzene-d6, ppm) δ 8.04 (m, 1H, Ar-H), 7.26 (d, 1H, Ar-H),
7.09 (d, 1H, Ar-H), 7.00–6.9 (m, 6H, Ar-H), 6.86 (t, 1H, Ar-H),
5.15 (s, 6H, benzene-H), 1.52 (s, 18H, NBu-H), 1.31 (s, 18H, NBu-H). 13C NMR (126 MHz, benzene-d6, ppm) δ
171.84 (NCN), 170.75 (NCN), 132.55,
132.44, 129.80, 129.70, 129.47, 129.24, 129.02, 128.84, 128.55, 127.63,
127.36 (132.55–127.36: Ph), 80.23 (Fe-benzene), 54.24 (CMe3), 53.64 (CMe3), 31.86 (CH3), 31.39 (CH3). 29Si NMR (99 MHz, benzene-d6, ppm) δ 45.12, 42.45. UV–Vis (THF, λ
(nm) (ε, M–1·cm–1)):
410 (1604). IR-ATR (cm–1): 3059 (VW), 2970 (w),
2928 (VW), 2868 (VW), 1640 (VW), 1577 (VW), 1519 (VW), 1472 (m), 1443
(m), 1415.63 (S), 1389 (S), 1361 (S), 1272 (m), 1203 (S), 1085 (m),
1022 (m), 972 (VW), 926 (W), 882 (W), 789 (W), 753 (S), 726 (W), 708
(S), 636 (m), 617 (S). Anal. calcd for C36H52Cl2FeN4Si2: C, 59.74; H, 7.24; N,
7.74. Found: C, 57.43; H, 7.29; N, 7.77. Note: Due to the formation
of silicon carbide, the carbon values in the elemental analyses were
consistently too low for all the disilylene Fe compounds reported
in this paper.
{PhC(NBu)2Si}2Fe(C6H6) (3)
A solution
of 2 (37 mg, 0.051 mmol) in benzene (10 mL) was added
dropwise to KC8 (42 mg, 0.311 mmol) in a vial with stirring.
About 20 mg of KC8 was added every 3 h until all material
was converted to {PhC(NBu)2Si}2Fe(C6H6), which can be determined
by 1H NMR monitoring. The color of the reaction mixture
turned from red-brown to black. Compound 3 (yield: 29
mg, 89.3%) was collected by removing solvents and volatile materials
under vacuum. The solid was stored in pentane solution in a −30
°C freezer for 1 month to give X-ray quality crystals. 1H NMR (500 MHz, benzene-d6) δ 7.14
(m, 2H, Ar-H), 7.00–6.9 (m, 6H, Ar-H), 6.80 (td, 2H, Ar-H),
5.34 (s, 6H, benzene-H), 1.47 (s, 36H, NBu-H). 13C NMR (126 MHz, benzene-d6, ppm) δ 162.70 (NCN), 135.72, 129.91,
128.98, 128.93, 127.77, 127.56 (135.72–127.56: Ph), 76.74 (Fe-benzene),
54.62 (CMe3), 32.78 (CH3). 29Si NMR (99 MHz, benzene-d6, ppm) δ 34.49. UV–Vis (THF, λ (nm)
(ε, M–1·cm–1)): 390
(4280). IR-ATR (cm–1): 3047 (W), 2961 (m), 2922
(m), 2855 (W), 1957 (VW), 1598 (W), 1442 (VW), 1403 (W), 1387 (VS),
1356 (S), 1266 (m), 1203 (S), 1071 (m), 1029 (W), 965 (W), 925 (W),
889 (VW), 836 (VW), 791 (W), 752 (S), 704 (VS), 654 (VW), 610 (m),
560 (VW). Anal. calcd for C36H52FeN4Si2: C, 66.23; H, 8.03; N, 8.58. Found: C, 65.11; H, 8.17;
N, 8.57.
{PhC(NBu)2Si}2Fe(C6H6)(N3SiMe3) (4)
Compound 4, which is also formed
during the formation of species 5 (vide infra), can be
obtained as an isolable species upon reaction of 3 with
TMSN3 (7.5 μL, 0.153 mmol) for 1 h at 40 °C
in benzene solution in a J-Young tube (quantitative conversion). After
40 min of reaction, the solution was evaporated to dryness under vacuum
and redissolved in pentane. X-ray quality crystals were grown from
a pentane solution stored in a −30 °C freezer. 1H NMR (500 MHz, benzene-d6) δ 7.32
(dt, 2H, Ar-H), 7.22 (m, 1H, Ar-H), 7.11 (m, 2H, Ar-H), 7.04 (t, 2H,
Ar-H), 7.00 (t, 2H, Ar-H), 6.97 (d, 1H, Ar-H), 5.26 (s, 6H, benzene-H),
1.33 (s, 36H, NBu-H), 0.47 (s, 9H, Si(CH3)3). 13C NMR (126 MHz, benzene-d6, ppm) δ 159.33–127.76 (5 signals;
2 Ph), 76.66 (Fe-benzene), 53.68 (CMe3), 32.04 (CH3), −0.08 (Si(CH3)3), −1.23 (N3Si(CH3)3). 29Si
NMR (99 MHz, benzene-d6, ppm) δ
15.41, 7.05. UV–Vis (THF, λ (nm) (ε, M–1·cm–1)): 485 (2343). IR-ATR (cm–1): 3048 (VW), 2962 (W), 2033 (W), 1523 (VW), 1472 (W), 1422 (S),
1391 (W), 1358 (W), 1274 (W), 1237 (W), 1207 (S), 1148 (m), 1078 (W),
1021 (W), 989 (m), 966 (W), 923 (W), 832 (S), 789 (W), 751 (S), 724
(W), 704 (S), 641 (W), 609 (W). Elemental analysis of this species
did not yield satisfactory results, which is attributed to the demonstrated
thermal instability of this complex, leading to “decomposition”
to form species 5.
{PhC(NBu)2Si}2Fe(C6H6)(NSiMe3) (5)
TMSN3 (7.5 μL, 0.153 mmol) was added
by a pipette to a solution of 3 (36 mg, 0.055 mmol) in
a J-Young tube or Schlenk tube while stirring at 80 °C. After
heating the reaction mixture for 90 min, the product was isolated
(yield: 20.3 mg, 50%) as a purple-red solid by removing all solvents
and other volatile materials under vacuum and washing with cold pentane.
At shorter reaction times, this product coexists with complex 4 according to NMR spectroscopy. X-ray quality crystals were
collected by dissolving the obtained solid in benzene solution and
slow evaporation of the solution at room temperature. 1H NMR (500 MHz, benzene-d6) δ 7.36
(s, 1H, Ar-H), 7.31 (d, 3H, Ar-H), 7.01–6.99 (m, 4H, Ar-H),
6.96 (m, 2H, Ar-H), 5.12 (s, 6H, benzene-H), 1.43 (s, 36H, NBu-H), 0.39 (s, 9H, Si(CH3)3). 13C NMR (126 MHz, benzene-d6, ppm) δ 169.61 (NCN), 133.69, 129.94, 128.47,
127.67 (133.69–127.67: Ph), 75.31 (Fe-benzene), 53.61 (CMe3), 31.75 (CH3), 3.79 (NSi(CH3)3). 29Si NMR (99 MHz, benzene-d6, ppm)
δ 7.42, −20.71. UV–Vis (THF, λ (nm) (ε,
M–1·cm–1)): 530 (592). IR-ATR
(cm–1): 3357 (VW), 2961 (w), 2029 (S), 1597 (W),
1473 (VW), 1419 (S), 1392 (W), 1358 (m), 1237 (W), 1204 (S), 1074
(W), 1019 (S), 924 (VW), 832 (S), 787 (VW), 746 (S), 704 (S), 649
(VW), 641 (VW), 609 (W). Anal. calcd for C39H61FeN5Si3: C, 63.30; H, 8.31; N, 9.46. Found:
C, 60.19; H, 7.99; N, 8.82.
{PhC(NtBu)2Si}2Fe(C6H6)(Ph2CO) (6)
Benzophenone (437 μL
of a 100 mg/10 mL stock solution in benzene) was added dropwise to
a solution of 3 (15.6 mg) in benzene (5 mL) in a J-Young
tube or Schlenk tube at 80 °C with stirring. Product 6 was isolated as a dark greenish-black solid by washing with cold
pentane after removing all solvents and other volatile materials.
Yield: 12 mg, 60%. X-ray quality crystals were collected by dissolving
the solid in diethyl ether solution and then placing the sample in
a −30 °C freezer. 1H NMR (500 MHz, benzene-d6) δ 8.36 (d, 0.5H, Ar-H), 7.76 (d, 2H,
Ar-H), 7.52 (d, 1H, Ar-H), 7.31 (d, 0.5H, Ar-H), 7.25 (t, 2H, Ar-H),
7.06–6.95 (m, 9H, Ar-H), 6.98 (d, 1H, de-ArCH), 6.17 (dd, 1H,
de-ArCH), 5.86 (m, 1H, de-ArCH), 5.65 (dd, 1H, de-ArCH), 4.90 (d,
1H, de-ArCH), 5.07 (s, 6H, benzene-H), 1.59/1.45/1.31/0.97 (s, 9H,
NBu-H). 13C NMR (126 MHz,
benzene-d6, ppm) δ 170.46 (NCN), 167.56 (NCN), 159.33–122.22
(18 signals; 6 C + 18 C; CCCHCHCHCH + C of 3 Ph), 78.46 (Fe-benzene),
53.23 (CMe3), 53.14 (CMe3), 52.67 (CMe3), 52.38
(CMe3), 32.12 (CH3), 32.01 (CH3), 31.15 (CH3), 31.09 (CH3). 29Si NMR (99 MHz, benzene-d6, ppm)
δ 65.57, 33.66. UV–Vis (THF, λ(nm) (ε, M–1·cm–1)): 350 (4785), 435 (2782).
IR-ATR (cm–1): 3052 (VW), 2965 (W), 2962 (VW). 1647
(VW), 1596 (VW), 1472 (W), 1421 (S), 1390 (m), 1358 (m), 1264 (m),
1204 (S), 1106 (VW), 1072 (W), 1009 (W), 970 (W), 923 (W), 865 (W),
790 (m), 745 (S), 722 (VW), 699 (VS), 607 (S), 544 (S). Anal. calcd
for C49H62FeN4OSi2: C,
70.48; H, 7.48; N, 6.71. Found: C, 65.07; H, 7.09; N, 6.15.
Authors: Yi Ding; Samir Kumar Sarkar; Mohd Nazish; Shahila Muhammed; Daniel Lüert; Paul Niklas Ruth; Christina M Legendre; Regine Herbst-Irmer; Pattiyil Parameswaran; Dietmar Stalke; Zhi Yang; Herbert W Roesky Journal: Angew Chem Int Ed Engl Date: 2021-11-10 Impact factor: 16.823