From a stable silylene to a mixed-valent disiloxane and an isolable silaformamide-borane complex with considerable silicon-oxygen double-bond character

Article abstract of DOI:10.1002/anie.200700398

(Chemical Equation Presented) Kipping's dream has (almost) become reality with the first isolable silanone-like complex 2, prepared by borane-assisted addition of water to the stable silylene 1. The short Si-O bond in 2 points to considerable Si=O character. In contrast, treatment of 1 with water in the absence of a borane leads to the crystalline siloxy silylene 3, an unprecedented mixed-valent disiloxane with di- and tetravalent silicon atoms (R = 2,6-iPr₂C₆H₃).

Full text of DOI:10.1002/anie.200700398

Angewandte
Chemie
DOI: 10.1002/anie.200700398
Silicon Chemistry
From a Stable Silylene to a Mixed-Valent Disiloxane and an Isolable
Silaformamide–Borane Complex with Considerable Silicon–Oxygen
Double-Bond Character**
Shenglai Yao, Markus Brym, Christoph van Wüllen, and Matthias Driess*
Dedicated to Professor Walter Siebert on the occasion of his 70th birthday
For a long time, unsaturated silicon compounds with multiple
bonds to silicon have been thought to be unisolable at room
temperature.[1] This situation changed profoundly in 1981.[2,3]
During the past 25 years, intriguing progress has been
achieved to generate a wealth of isolable compounds with
silicon–heteroatom double bonds (heteroatom: elements
from Groups 13,[4] 14,[5] and 15,[6–8] and sulfur[9]), and more
recently even isolable compounds with a silicon–silicon triple
bond have been reported[10] which represent unique and
indispensable building blocks in organosilicon chemistry.
=
However, the isolation of silanones (R2Si O) that are stable
at room temperature (“Kippingꢀs dream”)[9d,e] has hitherto
been unsuccessful. The absence of isolable silanones is
probably a result of the lack of suitable synthetic methods
as well as the difficulty in preventing oligomerization of the
silicon–oxygen double bond.[11] Clearly, the pronounced
polarity of the silicon–oxygen p bond, estimated by theoret-
ical calculations,[12] accounts for the extraordinarily high
tendency of silanones to undergo dimerization and trimeriza-
tion. This proceeds with no barrier, in contrast to their carbon
analogues or any other related silicon–heteroatom p system.
Thus, taming the high polarity of the silicon–oxygen double
bond is pivotal for the generation of an isolable silanone
derivative.
Figure 1. a) Model study of the relative stability [kJmolꢀ1] of silaformyl
compounds A versus hydroxo silylenes B. The hypothetical silanones A
result through hydrogen migration from the OH group to silicon in B.
b) Generation and relative energies [kJmolꢀ1] of the hypothetical N!Si
silaformyl-amine adducts
[A(NMe3)] through the corresponding hydroxo silylene amine adducts
[B(NMe3)]. c) Generation and relative energies [kJmolꢀ1] of the hypo-
thetical N!Si and O!B donor–acceptor-supported silaformyl amine/
boron trifluoride complexes [A(NMe3)(BF3)] through the corresponding
hydroxo silylene amine/boron trifluoride complexes [B(NMe3)(BF3)].
Calculations were performed at the B3LYP/TZVP level of DFT.[13a]
=
Isolable silaformyl compounds R(H)Si O (A) could be
prepared by taking advantage of the tautomerization (hydro-
gen-atom migration) of suitable hydroxo silylenes RSi(OH)
[12d]
=
(B; i.e. RSi(OH)!R(H)Si O)
in the presence of donor
and acceptor groups attached to the silicon and oxygen atom,
respectively, which at the same time may reduce the polar
nature and provide additional steric protection of the silicon–
oxygen double bond (Figure 1). In fact, our calculations of the
model series B, [B(NMe3)], and [B(NMe3)(BF3)] (R = H, Me,
SiH3, NH2; see Figure 1)[13a] revealed that the desired hydro-
gen migration from the OH group to the divalent silicon atom
is adverse for B but strongly favored in the presence of
suitable donor (e.g. amines) and acceptor groups (e.g.
boranes) bound to silicon and oxygen, respectively.
These predictions prompted us to challenge the discour-
aging attempts to synthesize an isolable silaformamide
complex by using the concept of donor–acceptor stabilization.
Although hydroxo silylenes are promising starting mate-
rials, isolable derivatives are currently unknown. Recently, we
prepared the stable zwitterion-like silylene [LSiD] 1,[14] which
could be a suitable starting material for the generation of the
desired hydroxo silylene (HL)Si(OH) (2) by 1,4-addition of
H2O (Figure 2); alternatively, 1,1-addition of water at the
divalent Si atom could lead to the tautomer LSiH(OH) (2’;
Figure 2).
[*] Dr. S. Yao, Dr. M. Brym, Prof. Dr. C. van Wüllen, Prof. Dr. M. Driess
Technische Universität Berlin
Institute of Chemistry: Metalorganics and Inorganic Materials
Sekr. C2
Strasse des 17. Juni 135, 10623 Berlin (Germany)
Fax: (+49)30-314-22168
E-mail: matthias.driess@tu-berlin.de
[**] Financal support by the Deutsche Forschungsgemeinschaft and the
Fonds der Chemischen Industrie is gratefully acknowledged.
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 4159 –4162
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
4159

Communications
rotational isomers as a result of the presence of a tetracoor-
dinate stereogenic Si atom and hindered rotation around the
Si O bonds. Thus, the 29Si NMR spectrum shows two singlet
ꢀ
resonance signals at d = ꢀ7.9 and ꢀ9.6 ppm, which can be
unequivocally assigned to the silylene, and two doublets at
d = ꢀ53.7 and ꢀ54.2 ppm for the siloxy 29Si nuclei.[14]
One of the two isomeric forms could be enriched by
fractional crystallization and was characterized by X-ray
diffraction analysis. In the solid state (Figure 3), the two
almost-planar C3N2Si rings in 4 prefer a gauche conformation
Figure 2. Proposed access to the hydroxo silylene (HL)Si(OH) (2)
(1,4 adduct) and/or its 1,1 tautomer hydroxo silane LSiH(OH) (2’) by
addition of water to the mesomeric forms of the potentially zwitter-
ionic silylene 1.
ꢀ
relative to each other. The Si O bonds are somewhat short
compared to silicon–oxygen single bonds observed in
common disiloxanes;[15] in the present case, the Si2–O1
bond (165.6(1) pm) with the coordinatively unsaturated
divalent silicon is slightly longer than the Si1–O1 bond
(163.2(1) pm) involving the tetravalent silicon atom. The Si1-
O1-Si2 bond angle of 137.08(5)8 is similar to values observed
for other disiloxanes.
Thus, solutions of 1 in hexane at 48C were exposed to
oxygen-free water vapor. Notably, the reaction of 1 with water
proceeds in the molar ratio of 2:1, leading solely to the
unexpected siloxy derivative of the desired hydroxo silylene 2,
that is, the siloxy silylene 4 (Figure 3). The latter was isolated
The mechanism is still unknown, how-
ever, the formation of 4 suggests that the
proton migration from a terminal methyl
group in the proposed reactive intermediate 2
seems to be much faster than the proton
transfer from the OH group to the divalent
silicon atom, preventing the formation of the
desired silaformamide 3. The migration of the
proton from OH to the divalent silicon in 2
should be on a competitive basis if the acidity
of the OH group is drastically increased by
the presence of a strong Lewis acid bound to
oxygen. In fact, addition of the water–borane
adduct H2O·B(C6F5)3 to 1 in a molar ratio of
1:1 affords exclusively the desired silaforma-
mide-borane
complex
[3{B(C6F5)3}]
(Figure 3). The latter complex was isolated
in the form of colorless, air-stable crystals in
67% yield. Their composition was confirmed
by correct elemental analysis and electron-
impact mass spectrometry (EI-MS). The
structure of [3{B(C6F5)3}] was established by
means of multinuclear NMR spectroscopy
and confirmed by single-crystal X-ray crys-
tallography (Figure 3). The compound con-
sists of a puckered six-membered C3N2Si ring
with an exocyclic OB(C6F5)3 group attached
to silicon. The silicon atom of the silaformyl
group (Si(H)O) is tetrahedrally coordinated
because of an intramolecularly dative N!Si
bond. Because of the relatively large intrinsic
Si–O and B–O s-bond polarity, the Si-O-B
Figure 3. Generation of the silaformamide-borane complex [3{B(C6F5)3}] and the mixed-
valent disiloxane (siloxy silylene) 4 by addition of H2O·B(C6F5)3 and H2O to the silylene 1,
respectively. The compounds 2, [2{B(C6F5)3}], as well as 2’ are proposed as reactive
intermediates for the formation of [3{B(C6F5)3}] and 4, respectively. Molecular structure
representations of [3{B(C6F5)3}] and 4 derived from X-ray diffraction studies are inclu-
ded.[13a]
in the form of brown-red crystals in 52% yield. In contrast,
reactions of 1 in the presence of more water molecules (e.g.
1:1 molar ratio) afford merely unidentified hydrolysis prod-
ucts. The siloxy silylene 4 represents an unprecedented type of
mixed-valent disiloxane, which contains di- and tetravalent
silicon atoms that are bridged by an oxygen atom. Its
composition was proven by multinuclear NMR spectroscopy,
bond angle is widened to 163.7(1)8, which minimizes the
dipole moment, releases steric congestion, and supports Si–O
p interaction.
The structure is most notable for its unique short silicon–
oxygen interatomic distance of 155.2(2) pm, about 7%
shorter than those observed in silyl ethers (Si-O-C) and
common disiloxanes (Si-O-Si)[15] and only marginally longer
than the value of 153.7 pm calculated for the parent silafor-
1
mass spectrometry, and elemental (C,H,N) analysis. The H
and 29Si NMR spectra revealed that 4 is a mixture of two
mamide, H2N(H)Si O, by density functional theory
=
4160 www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 4159 –4162

Angewandte
Chemie
(DFT).[13a,b] This observation indicates that multiple bonding
between silicon and oxygen in [3{B(C6F5)3}] is not signifi-
cantly disturbed. In line with this suggestion, the B–O
distance of 150.3(3) pm is significantly longer than a covalent
B–O bond (ca. 131 pm) but close to the value of the
coordinative O!B bond in H2O·B(C6F5)3 (159.7(2) pm).[16a]
Likewise, and in accord with DFT calculations of the related
silaformyl derivatives A and their donor–acceptor adducts
[A(NMe3)] and [A(NMe3)(BF3)], respectively, the Si–O bond
reveals a small elongation (ca. 4%) upon additional N!Si
donor and O!B acceptor coordination.[15] The electronic
features of the silicon–oxygen functionality in [3{B(C6F5)3}]
are similar to that described for related Lewis acid stabilized
monomeric boron–oxygen[16b] and aluminum–oxygen com-
plexes,[16c] which reveal also unusually short B–O and Al–O
bond lengths, respectively, suggesting considerable double-
bond character. However, a relatively short element–oxygen
distance as in the aforementioned cases and in [3{B(C6F5)3}] is
a unreliable criterion for the assessment of the bond order.
Alternatively, the short distances could be explained by the
intrinsically high bond polarity and the contribution of polar
resonance structures as previously suggested.[16b]
ecules with Si–O single and double bonds and for silanones
supported by N donors as well as stabilized both by amine
donors and borane acceptors, including compounds which
model the bonding situation in [3{B(C6F5)3}]. From a plot of
calculated Si–O force constant versus the square of the
vibrational frequency,[13a] one can estimate that a Si = 16O
stretching frequency of 1165 cmꢀ1 corresponds to a Si–O force
constant of about 800 Nmꢀ1. The linear interpolation between
the prototypes Me3SiOMe (k = 438 Nmꢀ1, bond order set to
ꢀ1
=
1.0) and H2Si O (k = 874 Nm , bond order set to 2.0) reveals
ꢀ
1.83for the Si O bond order in [3{B(C6F5)3}]. In this
interpolation, the bond order has been defined through the
bond strength, but a claim for a partial silicon–oxygen double-
bond character also calls for a rationalization in terms of
molecular orbitals. Because the silicon atom is tetracoordi-
nated, there is no classical Lewis structure featuring a double
bond. However, a population analysis for [3{B(C6F5)3}] in
terms of natural atomic orbitals, performed with the NBO
module of the Gaussian program,[13a] reveals a significant
ꢀ
population for the two antibonding Si N bonds (s*p acceptor
orbitals mainly located at silicon) and a decreased population
for the two oxygen lone pairs (np donor orbitals) which
indicates a substantial p-bonding interaction between these
orbitals. Together with the large force constant (and bond
order derived therefrom), this provides enough evidence that
[3{B(C6F5)3}] can faithfully be viewed as a silaformamide
equivalent. Investigations on its reactivity are currently
underway.
To support whether the silicon–oxygen functionality in
[3{B(C6F5)3}] retains considerable double-bond character, we
performed IR measurements. Previous experiments on low-
temperature matrix-isolated silanones revealed characteristic
=
=
Si O stretching vibration modes (n(Si O)) in the region of
1200 cmꢀ1,[17] whereas Si–O single bonds in organosilanols
(R3SiOH; R = alkyl, aryl) and siloxanes (R3Si-O-X; R =
alkyl, aryl; X = Me3Si, organic group) exhibit peaks at much
Received: January 29, 2007
Revised: March 2, 2007
Published online: April 13, 2007
lower wavenumbers in the range of 800 to 900 cmꢀ1 [17b]
The
.
vibrational modes in the IR spectrum of [3{B(C6F5)3}] were
unambiguously assigned by means of isotope labeling experi-
ments and respective DFT calculations.[13b] Accordingly, 18O-
labeled [3{B(C6F5)3}] was prepared from 1 and H218O·B-
(C6F5)3. Comparison of the observed spectra of these two
isotopomers with the calculated IR spectra of the two
respective slightly smaller substituted isotopomer models
are nearly identical except for one band, which originates
Keywords: boranes · density functional calculations ·
.
multiple bonds · silanones · silicon
[1] Reviews: a) “Multiple Bonds to Silicon”: G. Raabe, J. Michl in
The Chemistry of Organic Silicon Compounds (Eds.: S. Patai, Z.
Rappoport), Wiley, New York, 1989, part 2, chap. 17, p. 1015 –
1142; b) “Multiply Bonded Main Group Metals and Metalloids”:
Advances in Organometallic Chemistry, Vol. 39 (Eds.: R. West,
F. G. A. Stone), Academic Press, San Diego, 1996; c) P. P. Power,
Chem. Rev. 1999, 99, 3463; d) R. West, Polyhedron 2002, 21, 467.
[2] R. West, M. J. Fink, J. Michl, Science 1981, 214, 1343.
[3] A. G. Brook, F. Abdesaken, B. Gutekunst, G. Gutekunst, R. K.
Kallury, J. Chem. Soc. Chem. Commun. 1981, 191.
[4] N. Nakata, A. Sekiguchi, J. Am. Chem. Soc. 2006, 128, 422.
[5] a) M. Weidenbruch, Organometallics 2003, 22, 43 48; b) N.
Tokitoh, R. Okazaki, Adv. Organomet. Chem. 2001, 47, 121.
[6] M. Driess, Adv. Organomet. Chem. 1996, 39, 193.
[7] M. Driess, S. Block, M. Brym, M. T. Gamer, Angew. Chem. 2006,
118, 2351; Angew. Chem. Int. Ed. 2006, 45, 2293.
[8] I. Hemme, U. Klingebiel, Adv. Organomet. Chem. 1996, 39, 159.
[9] a) R. Arya, J. Boyer, F. CarrØ, R. Corriu, G. Lanneau, J.
Lappasset, M. Perrot, C. Priou, Angew. Chem. 1989, 101, 1069;
Angew. Chem. Int. Ed. Engl. 1989, 28, 1016; b) H. Suzuki, N.
Tokitoh, R. Okazaki, J. Am. Chem. Soc. 1994, 116, 11578; c) R.
Okazaki, N. Tokitoh, Acc. Chem. Res. 2000, 33, 625; d) S.
Kipping, L. L. Lloyd, J. Chem. Soc. 1901, 449; e) “Historical
Overview and Comparison of Silicon with Carbon”: J. Y. Corey
from Si O stretching.[17c] This comparison allows us to assign
=
the observed band at 1112 cmꢀ1 for 18O-labeled [3{B(C6F5)3}]
to a Si 18O stretching mode. On the basis of the calculation as
=
well as a wealth of experimental data,[13a] we conclude that the
corresponding vibration of the 16O isotopomer is shifted about
40 cmꢀ1 towards higher wavenumbers and is part of the broad
band at 1165 cmꢀ1, whose intensity decreases considerably
upon 18O labeling.
The calculated Si–O stretching frequencies (1122 cmꢀ1 for
Si 16O, 1082 cmꢀ1 for Si 18O) are smaller than those found
experimentally, most likely because the N-aryl substituents in
the calculated model are smaller and thus lead to less steric
repulsion with B(C6F5)3. However, this does not question our
assignment, which is primarily based on the changes in the
spectra upon 18O labeling. The observed Si–O stretching
frequency (1165 cmꢀ1 for the 16O isotopomer) is somewhat
=
=
smaller than for matrix-isolated silanones (about 1200 cmꢀ1
)
but clearly far above frequencies typical for Si–O single
ꢀ
bonds. To derive a Si O bond order from the frequency, we
calculated Si–O stretching frequencies for prototypical mol-
Angew. Chem. Int. Ed. 2007, 46, 4159 –4162
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
4161

Communications
in The Chemistry of Organic Silicon Compounds, Vol. 1 (Eds.: S.
[14] M. Driess, S. Yao, M. Brym, C. van Wꢁllen, D. Lentz, J. Am.
Chem. Soc. 2006, 128, 9628.
[15] “Structural Chemistry of Organic Silicon Compounds”: W. S.
Sheldrick in The Chemistry of Organic Silicon Compounds,
Vol. 1 (Eds.: S. Patai, Z. Rappoport), Wiley, New York, 1989,
chap. 3, p. 227, and references therein.
Patai, Z. Rappoport), Wiley, New York, 1989, chap. 1, pp. 1 – 56.
[10] a) N. Wiberg, K. S. Vasisht, G. Fischer, P. Meyer, Z. Anorg. Allg.
Chem. 2004, 630, 1823; b) A. Sekiguchi, R. Kinjyo, M. Ichinohe,
Science 2004, 305, 1755.
[11] “Recent Advances in the Chemistry of Siloxane Polymers and
Copolymers”: R. Drake, I. Mackinnon, R. Taylor in The
Chemistry of Organic Silicon Compounds, Vol. 2 (Eds.: S.
Patai, Z. Rappoport), Wiley, New York, 1998, Part 3, chap. 38,
p. 2217, and references therein.
[12] a) “Theoretical Aspects of Organosilicon Compounds”: Y.
Apeloig in The Chemistry of Organic Silicon Compounds,
Vol. 1 (Eds.: S. Patai, Z. Rappoport), Wiley, New York, 1989,
chap. 2, p. 57, and references therein; b) J. Kapp, M. Remko,
P. v. R. Schleyer, J. Am. Chem. Soc. 1996, 118, 5745; c) M. S.
Gordon, C. George, J. Am. Chem. Soc. 1984, 106, 609; d) M.
Kimura, S. Nagase, Chem. Lett. 2001, 1098.
[16] a) L. D. Doerrer, M. L. H. Green, J. Chem. Soc. Dalton Trans.
1999, 4325; b) D. Vidovic, J. A. Moore, J. N. Jones, A. H. Cowley,
J. Am. Chem. Soc. 2005, 127, 4566; c) D. Nuculai, H. W. Roesky,
A. M. Neculai, J. Magull, B. Walfort, D. Stalke, Angew. Chem.
2002, 114, 4470; Angew. Chem. Int. Ed. 2002, 41, 4294.
[17] a) “Matrix Isolation Studies of Silicon Compounds”: G. Maier,
A. Meudt, J. Jung, H. Pacl in The Chemistry of Organic Silicon
Compounds, Vol. 2 (Eds.: S. Patai, Z. Rappoport), Wiley, New
York, 1998, Part 2, chap. 19, p. 1143, and references therein;
b) V. N. Khabashesku, Z. A. Kerzina, K. N. Kudin, O. M.
Nefedov, J. Organomet. Chem. 1998, 566, 45; c) S. Bailleux, M.
Bogey, C. Demuynck, J. L. Destomes, A. Walters, J. Chem. Phys.
1994, 101, 2729; d) H. Schnꢂckel, Z. Anorg. Allg. Chem. 1980,
460, 37; e) C. A. Arrington, R. West, J. Michl, J. Am. Chem. Soc.
1983, 105, 6176; f) R. Withnall, L. Andrews, J. Am. Chem. Soc.
1985, 107, 2567.
[13] a) Experimental and computational details as well as X-ray
diffraction data for [3{B(C6F5)3}] and 4 are provided in the
=
Supporting Information. b) The CCSD/TZVP value for the Si
=
O bond length in H2N(H)Si O is 153.2 pm.
4162
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 4159 –4162