From elusive thio- and selenosilanoic acids to copper(i) complexes with intermolecular Si=E → Cu-O-Si coordination modes (E = S, Se)

Article abstract of DOI:10.1039/c3cc41965g

The facile synthesis of the first stable selenosilanoic acid-base adduct LSi(=Se)OH(dmap) 3 (L = CH[C(Me)NAr]₂, Ar = 2,6-iPr₂C ₆H₃, dmap = 4-dimethylaminopyridine), the heavier analogue of the thiosilanoic acid adduct LSi(=S)OH(dmap) 1, is reported. Both adducts 1 and 3 react readily with MesCu (Mes = 2,4,6-trimethylphenyl) to form the novel dimeric Cu(i) complexes [LSi(=E)OCu]₂ (4: E = S; 5: E = Se) with unprecedented intermolecular Si=E → Cu-O-Si coordination modes. The latter are efficient pre-catalysts for the Cu(i)-mediated aziridination of styrene with Ph=IN(Ts) (Ts = tosyl).

Full text of DOI:10.1039/c3cc41965g

ChemComm
View Article Online
View Journal | View Issue
COMMUNICATION
From elusive thio- and selenosilanoic acids to copper(I)
Q
complexes with intermolecular Si E - Cu–O–Si
coordination modes (E = S, Se)†
Cite this: Chem. Commun., 2013,
49, 5595
Received 16th March 2013,
Accepted 29th April 2013
Gengwen Tan, Yun Xiong, Shigeyoshi Inoue, Stephan Enthaler, Burgert Blom,
Jan D. Epping and Matthias Driess*
DOI: 10.1039/c3cc41965g
www.rsc.org/chemcomm
The facile synthesis of the first stable selenosilanoic acid–base adduct
Q
LSi( Se)OH(dmap) 3 (L = CH[C(Me)NAr]2, Ar = 2,6-iPr2C6H3, dmap =
4-dimethylaminopyridine), the heavier analogue of the thiosilanoic
Q
acid adduct LSi( S)OH(dmap) 1, is reported. Both adducts 1 and 3
react readily with MesCu (Mes = 2,4,6-trimethylphenyl) to form the
Q
novel dimeric Cu(I) complexes [LSi( E)OCu]2 (4: E = S; 5: E = Se) with
Q
unprecedented intermolecular Si
E - Cu–O–Si coordination
modes. The latter are efficient pre-catalysts for the Cu(I)-mediated
Q
aziridination of styrene with PhI N(Ts) (Ts = tosyl).
Using organic molecular defined species containing M–O–Si moieties
as structural models for silica-supported metal/metal oxide catalysts
has been proved to be a facile method to unravel the structure of a
pre-catalyst as well as the mechanism of a catalytic reaction.1 For
example, the globular 56-membered copper(I) siloxane containing
core made up of Cu–O–Si moieties (Chart 1), which is even soluble in
organic solvents, could be prepared by the reaction of silanetriol with
(MesCu)4 (Mes = 2,4,6-Me3C6H2). This copper siloxane cluster is
active for the Ullmann–Goldberg-type C–N coupling reaction,2 and
can serve as a structural and functional model for silica-supported
copper (pre)catalysts. In order to better understand the structure–
reactivity relationships of Cu–O–Si systems for various Cu-based
chemical transformations, the synthesis of other types of Cu–O–Si
containing compounds as structural and functional models for
metallated terminal OH groups of silica surfaces is desired.
Organic silanols have been reported to form well defined M–O–Si
type compounds including silanediols, disiloxane-1,3-diols, silane-
triols, trisiloxane-diol and silsesquioxane-triols (Chart 1).1a,b,3,4 As
possible intermediates involved in silica synthesis through hydrolysis
of suitable silicon(IV) starting materials under acidic conditions,
silicic acids, SiOx(OH)4ꢀ2x, have not received much attention as
precursors for the selective synthesis of M–O–Si compounds owing
to their elusive nature. Likewise, organic silanoic acids RSi(QO)OH,
Chart 1
the silicon analogues of carboxylic acids, represent elusive species
which could only be studied at liquid nitrogen temperatures because
of the presence of the highly polarized SiQO subunit which can
undergo facile isomerisation or intermolecular head-to-tail poly-
merisation.5 Recently, we reported the synthesis of the first isolable
silanoic acid–base adducts, including the thiosilanoic system
LSi(QS)OH(dmap) 1 (L = CH[C(Me)NAr]2, Ar = 2,6-iPr2C6H3, dmap =
4-dimethylaminopyridine) which is stabilised via hydrogen bonding
to the pyridine N atom of dmap (Chart 1).6 With these compounds in
hand it became possible to employ 1 as a building block for Si(QS)–
O–M formation. Accordingly, we reported the formation and reactivity
of an isolable monomeric Si(QS)–O–Mn(II) complex.7 Herein, we
report the unexpectedly facile synthesis of the first isolable seleno-
silanoic acid–base adduct LSi(QSe)OH(dmap) 3 (Chart 1), and the
Cu-metallation reactions of 1 and 3 with (MesCu)4 to give the novel
dimeric LSi(QE)OCu complexes 4 (E = S) and 5 (E = Se) with
unprecedented intermolecular SiQE - Cu–O–Si coordination
modes. In addition, their ability to serve as efficient pre-catalysts
¨
Technische Universitat Berlin, Department of Chemistry, Metalorganics and
Inorganic Materials, Sekr. C2, Strasse des 17. Juni 135, 10623 Berlin, Germany.
E-mail: Matthias.driess@tu-berlin.de; Fax: +49 30-314-29732
† Electronic supplementary information (ESI) available: Experimental details and
crystal data and structure refinement for 3–5. CCDC 921657–921659. For crystal-
lographic data in CIF or other electronic format see DOI: 10.1039/c3cc41965g
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 5595--5597 5595

View Article Online
Communication
ChemComm
Fig. 1 Molecular structure of compound 3 in the solid state with the 50%
probability level for the core structure. Hydrogen atoms (except for H1) and the
n-hexane molecules are omitted for the sake of clarity.
Scheme 1 Synthesis of compound 3.
in the aziridation of a CQC bond with PhIQN(Ts) has been
demonstrated.
The thiosilanoic acid–base complex 1 is accessible from the
reaction of the corresponding stable silanone complex L0Si(QO)
(dmap) 2 (L0 = CH[C(Me)(CQCH2)](NAr2), Ar = 2,6-iPr2C6H3) with
H2S at ambient temperature. Considering the toxicity of H2Se, we
attempted to synthesize the selenium congener of 1 with dilithium
selenide as a selenium source, which can be obtained from the
reaction of elemental selenium with lithium triethylhydridoborate in
THF.8 Treatment of the in situ prepared dilithium selenide with one
equivalent of 2 in THF at ꢀ78 1C, followed by protonation with two
molar equivalents of trimethylammonium chloride, leads to a clear
yellow solution, from which compound 3 could be isolated in the
form of yellow crystals in 75% yield (Scheme 1).
Scheme 2 Syntheses of the Cu(I) complexes 4 and 5.
known to serve as a smooth reagent for preparing Cu(I) complexes
by deprotonation of amines and silanols.2,3,10
Accordingly, the reaction of compound 1 with 14 molar equivalent
of (MesCu)4 was carried out in THF at ꢀ20 1C. The 1H NMR
spectrum of the resulting reaction mixture already shows the
absence of the resonance signals of the OH group of 1 and those
of (MesCu)4 with concomitant liberation of dmap, indicating that
the metallation of 1 was successful. Indeed, the desired compound
4 could be isolated from the reaction mixture as yellow crystals in
85% yield (Scheme 2). Single crystals of 4 suitable for X-ray diffrac-
tion analysis could be obtained in toluene solutions at 0 1C. The
structure analysis revealed that the compound is a dimer with
intermolecular SiQS - Cu–O–Si interactions (Fig. 2). The com-
pound possesses C2 symmetry and consists of a planar eight-
membered Si2O2S2Cu2 ring with two trans-oriented b-diketiminato
ligands L. The Cu centres are linearly coordinated by one oxygen
and a sulfur atom of the neighbouring SiQS subunit. The
Cu(1)ꢁꢁꢁCu(1A) distance of 2.8135(7) Å suggests a weak d10ꢁꢁꢁd10
interaction in this molecule.11 The Si(1)–O(1) distance of 1.568(2) Å
is significantly shorter than that in the precursor 1 (1.620(2) Å),
whereas the Si(1)–S(1) bond length (2.0609(9)) is longer than that
The strikingly stable compound 3 has very similar solubility
properties to 1. It is soluble in toluene, THF, chloroform, dichloro-
methane, marginally soluble in benzene but insoluble in n-hexane.
It was fully characterised using multinuclear NMR spectroscopy,
elemental analysis, mass spectrometry and IR spectroscopy, as well
as single-crystal X-ray diffraction analysis. The high resolution
electrospray ionization mass spectrometry (ESI-MS) shows the
molecular ion peak at m/z 665.3138 [LSi(QSe)OH(DMAP) + H]+ as
1
the base signal. In the H NMR spectrum in CDCl3, the chemical
shift of the ring proton in the g-position of the b-diketiminate L is
observed at d = 5.73 ppm. This chemical shift is comparable to that
observed for 1 (d = 5.71 ppm). The proton of the OH group resonates
at d = 6.42 ppm in the same region as the proton resonance signals
for the pyridine moiety of dmap. The 29Si{1H} resonance signal of 3
in CDCl3 at d = ꢀ25.5 ppm is low-field shifted compared to that of
1 (d = –30.0 ppm), while its 77Se{1H} NMR spectrum exhibits a
singlet at d = –545.2 ppm, which is up-field shifted compared
to those of the related selenosilanoic silylester diastereomers
(d = ꢀ384.8 and ꢀ401.3 ppm).9
The molecular structure of compound 3 is depicted in Fig. 1.
%
It crystallizes in the triclinic space group P1 with one n-hexane
molecule lying about an inversion center. It is isotypic with 1, with
one dmap ligand connected to the selenosilanoic acid moiety
through an O–HꢁꢁꢁN hydrogen bond. The Si1–O1 bond length of
1.619(2) Å is comparable to that of 1 (1.620(2) Å), whereas the
Si(1)–Se(1) bond distance of 2.1348(7) Å is close to the SiQSe
distance of a related selenosilanoic silyl ester (2.117(1) Å).9 As
expected, other metric parameters of 3 are akin to those of 1.
Both compounds 1 and 3 were allowed to react with (MesCu)4
(Mes = 2,4,6-trimethylphenyl) in the hope of producing the
corresponding Cu(I) complexes 4 and 5, respectively. (MesCu)4 is
Fig. 2 (a) Molecular structures of compounds 4 (EQS) and 5 (E = Se) in the solid
state with the 50% probability level for the core structure. Hydrogen atoms and
the toluene molecules are omitted for the sake of clarity. Operation symmetry for
all atoms labelled ‘‘A’’: ꢀx + 3/2, ꢀy + 1/2, ꢀz (4); ꢀx + 2, ꢀy, ꢀz (5);
(b) representation of the eight-membered planar Si2O2S2Cu2 ring without
substituents at silicon atoms.
c
5596 Chem. Commun., 2013, 49, 5595--5597
This journal is The Royal Society of Chemistry 2013

View Article Online
ChemComm
Communication
the longer Si–Se distance (2.2011(9) Å) compared to that of the Si–S
bond (2.0609(9) Å) in 4.
It has been shown that Cu(I) complexes can be efficiently applied
as pre-catalysts in metal-catalyzed nitrene-transfer reactions if the
Cu(I) centre is efficiently chelate coordinated.13 In a preliminary
study the catalytic ability of compounds 4 and 5 in the nitrene-
transfer reaction (aziridation) of a CQC bond has been evaluated.
The catalytic reactions were carried out by using styrene and
PhIQN(Ts) (Ts = tosyl) as a nitrene source in the presence of
2.5 mol% of 4 and 5, respectively, in CH2Cl2 at ambient temperature.
The resulting yields of the N-tosyl-2-phenylaziridine product [85% (4)
and 87% (5)] are similar to a reported result.13a It is generally
considered that a Cu(I)–nitrene species is the active component for
the aziridination reaction.13b Thus, it is reasonable to assume
that the dimeric Cu(I) complexes react initially with PhIQN(Ts) to
form the corresponding Cu(I)–nitrene intermediates, which are
capable of facile nitrene-transfer to the CQC bond of styrene.
In summary, using dilithium selenide as a selenium source, the
first isolable selenosilanoic acid–base adduct 3 has been synthesized
which is isostructural with the thiosilanoic acid–base adduct 1.
Facile reaction of 1 and 3 with 14 molar equivalents of (MesCu)4 led
to the unprecedented dimeric copper(I) complexes 4 and 5. Both
complexes exist in two stereoisomeric forms in chloroform solutions
and verify the novel SiQE - Cu–O–Si structural motif in Cu(I)
siloxane chemistry. In addition, compounds 4 and 5 can act as
reliable pre-catalysts in aziridination of styrene with PhIQN(Ts).
Fig. 3 Optimized structure for the proposed stereoisomer 4b. Hydrogen atoms
are omitted for the sake of clarity.
observed in 1 (1.993(1) Å), suggesting carboxylate-like p-conjugation
in the SiSO moiety.12
The composition of 4 has been confirmed by multinuclear NMR
spectroscopy, elemental analysis and IR spectroscopy. The solubility
of 4 is similar to that of precursor 1. Unexpectedly, two sets of
1
resonances for the b-diketiminato ligand L appear in the H NMR
spectrum with a ratio of 1 : 0.62 as indicated by the integrals of the
resonances from the ring proton in the g-position of L at d = 5.55 and
5.49 ppm, respectively. Accordingly, the 29Si{1H} NMR spectrum
reveals two close signals at d = ꢀ38.0 and ꢀ39.1 ppm, respectively.
The two sets of resonances suggest the presence of two stereo-
isomers of 4 (4a and 4b) in chloroform solutions. Indeed this is
substantiated by the results obtained from diffusion ordered
spectroscopy (DOSY) experiments, revealing identical diffusion
coefficients for these two species (see ESI†) and thus the same
molecular size and composition, respectively. In other words, dis-
sociation of 4 in chloroform solutions can be excluded. As expected,
the two stereoisomers can be interconverted as shown using variable
1H NMR spectroscopy. Cooling of a CDCl3 solution of 4 to 230 K
changes the ratio of signal sets of 4a and 4b from 1 : 0.62 at ambient
temperature to 1 : 0.42 (see ESI†). In the solid state 29Si NMR
spectrum of 4 in crystalline form, there is only one resonance signal
at d = ꢀ41.6 ppm, whereas in the spectrum of the fine powder two
signals are observed (d = ꢀ41.6 ppm and ꢀ38.7 ppm) (see ESI†), this
suggests that 4a holds a structure as shown in Fig. 2 with two ligands
L in the trans-position. We inferred that the other stereoisomer 4b
preserves the C2 symmetry of 4a but the two ligands L are now
cis-oriented. However, this is ruled out by the results of DFT
calculations, which revealed that the proposed cis isomer is least
favoured. Instead the DFT calculations suggest a twisted Si2O2S2Cu2
core structure also with C2 symmetry as a stereoisomer which is only
5.5 kJ molꢀ1 less stable than 4a (Fig. 3 and ESI†).
Notes and references
1 (a) R. Murugavel, A. Voigt, M. G. Walawalkar and H. W. Roesky, Chem.
Rev., 1996, 96, 2205–2236; (b) R. Murugavel, V. Chandrasekhar and
H. W. Roesky, Acc. Chem. Res., 1996, 29, 183–189; (c) N. Winkhofer,
A. Voigt, H. Dorn, H. W. Roesky, A. Steiner, D. Stalke and A. Reller,
Angew. Chem., Int. Ed. Engl., 1994, 33, 1352–1354; (d) H. C. L. Abbenhuis,
Chem.–Eur. J., 2000, 6, 25–32; (e) K. L. Fujdala, I. J. Drake, A. T. Bell and
T. D. Tilley, J. Am. Chem. Soc., 2004, 126, 10864–10866.
2 G. Tan, Y. Yang, C. Chu, H. Zhu and H. W. Roesky, J. Am. Chem. Soc.,
2010, 132, 12231–12233.
3 F. Schax, C. Limberg and C. Mu¨gge, Eur. J. Inorg. Chem., 2012, 4661–4668.
4 (a) V. Lorenz, A. Fischer, K. Jacob, W. Brueser and F. T. Edelmann,
Chem.–Eur. J., 2001, 7, 848–857; (b) V. Lorenz, S. Gießmann,
Y. K. Gun’ko, A. K. Fischer, J. W. Gilje and F. T. Edelmann, Angew.
Chem., Int. Ed., 2004, 43, 4603–4606; (c) V. Lorenz, A. Fischer,
K. Jacob, W. Brueser, T. Gelbrich, P. G. Jones and F. T. Edelmann,
Chem. Commun., 1998, 2217–2218.
5 (a) C. A. Arrington, R. West and J. Michl, J. Am. Chem. Soc., 1983, 105,
6176–6177; (b) R. Withnall and L. Andrews, J. Am. Chem. Soc., 1985,
107, 2567–2568.
6 Y. Xiong, S. Yao and M. Driess, Angew. Chem., Int. Ed., 2010, 49,
6642–6645.
7 S. Yao, Y. Xiong and M. Driess, Chem.–Eur. J., 2012, 18, 11356–11361.
8 U. Herzog, H. Lange, H. Borrmann, B. Walfort and H. Lang,
J. Organomet. Chem., 2004, 689, 4909–4916.
9 S. Yao, Y. Xiong, M. Brym and M. Driess, Chem.–Asian. J., 2008, 3,
113–118.
Similar to the synthesis of 4, compound 5 could be obtained by
the reaction of 3 with (MesCu)4 in THF (Scheme 2). Compound 5
has been fully characterized by multinuclear NMR spectroscopy,
elemental analysis, IR and mass spectroscopy, and X-ray crystallo-
graphy. Akin to the situation of 4, there are two stereoisomers
1
present in CDCl3 solutions as shown by the H, 13C, 29Si NMR
1
10 (a) G. Tan and H. Zhu, Inorg. Chem., 2011, 50, 6979–6986;
(b) M. Stollenz and F. Meyer, Organometallics, 2012, 31, 7708–7727.
11 (a) P. K. Mehrotra and R. Hoffmann, Inorg. Chem., 1978, 17,
2187–2189; (b) K. M. Merz Jr. and R. Hoffmann, Inorg. Chem.,
1988, 27, 2120–2127.
12 (a) R. D. Mounts, T. Ogura and Q. Fernando, Inorg. Chem., 1974, 13,
802–805; (b) A. S. Filatov, O. Hietsoi, Y. Sevryugina,
N. N. Gerasimchuk and M. A. Petrukhina, Inorg. Chem., 2010, 49,
1626–1633.
spectra as well as H-DOSY experiments (see ESI†).
%
Single-crystals of 5 in the triclinic space group P1 could be
obtained in toluene solutions. The X-ray diffraction analysis
revealed that 5 and 4 are isotypic (Fig. 2). Akin to the structure
of 4, compound 5 is a dinuclear copper(I) complex with both
copper centers coordinated by O and Se atoms. A linear geometry
of the O–Cu–Se connection with an angle of 172.07(7)1 is observed.
The Cu(1)ꢁꢁꢁCu(1A) bond distance of 2.9271(8) Å is slightly longer
than that in compound 4 (2.8135(7) Å). This can be explained by
´
13 (a) P. J. Perez, M. Brookhart and J. L. Templeton, Organometallics,
1993, 12, 261–262; (b) Y. M. Badiei, A. Krishnaswamy, M. M. Melzer
and T. H. Warren, J. Am. Chem. Soc., 2006, 128, 15056–15057.
c
This journal is The Royal Society of Chemistry 2013
Chem. Commun., 2013, 49, 5595--5597 5597