Johann Hlina1, Henning Arp1, Małgorzata Walewska1, Ulrich Flörke2, Klaus Zangger3, Christoph Marschner1, Judith Baumgartner3. 1. Institut für Anorganische Chemie, Technische Universität Graz , Stremayrgasse 9, 8010 Graz, Austria. 2. Zentrale Analytik, Department Chemie, Universität Paderborn , Warburger Straße 100, 33098 Paderborn, Federal Republic of Germany. 3. Institut für Chemie, Karl Franzens Universität Graz , Stremayrgasse 9 and Heinrichstraße 28, 8010 Graz, Austria.
Abstract
Reactions of Et3P adducts of bissilylated germylenes and stannylenes with gold, silver, and copper cyanides led to cyanogermyl or -stannyl complexes of the respective metals. In the course of the reaction the phosphine moved to the metal, while the cyanide migrated to the low-coordinate group 14 element. The respective gold complexes were found to be monomeric, whereas the silver and copper complexes exhibited a tendency to dimerize in the solid state. Attempts to abstract the phosphine ligand with B(C6F5)3 led only to the formation of adducts with the borane coordinating to the cyanide nitrogen atom.
Reactions of Et3P adducts of bispan class="Chemical">silylated germylenes and stannylenes with gold, silver, and copper cyanides led to cyanogermyl or -stannyl complexes of the respective metals. In the course of the reaction the phosphine moved to the metal, while the cyanide migrated to the low-coordinate group 14 element. The respective gold complexes were found to be monomeric, whereas the silver and copper complexes exhibited a tendency to dimerize in the solid state. Attempts to abstract the phosphine ligand with B(C6F5)3 led only to the formation of adducts with the borane coordinating to the cyanide nitrogen atom.
The
organometallic chemistry of the group 10 metals Ni, Pd, and Pt has
been in the focus of chemists now for decades, but related research
on group 11 metals Ag and Au has been neglected for a long time. Of
the group 11 metals only copper chemistry was studied with much interest,
because its value as a synthetically useful metal had been established
early, in particular for a number of coupling reactions such as the
Ullmann,[1] Glaser–Hay,[2] and Cadiot–Chodkiewicz[3] reactions. The last some 20 years have brought a paradigm
shift, and silver and in particular gold have become very valuable
metals for a number of catalytic processes,[4,5] exhibiting
properties that complement chemistry developed earlier for palladium
and nickel.Compared to classical organometallic compounds with
carbon–metal bonds, related substances where carbon is replaced
by its heavier congeners silicon, germanium, tin, and lead have also
received much less attention. In 1962 Glockling and Hooten synthesized
the first germanium–group 11 compounds by metathesis reactions
of triphenylgermyl lithium and the corresponding group 11 chloride(I)
phosphine complexes.[6] The first group 11
stannyl complexes were obtained soon after by reaction of group 11
metal chloride(I) triphenylphosphine complexes with SnCl2.[7] While the reaction of metal halides
with anionic main-group compounds is rather straightforward, the second
reaction is more interesting, as it can be considered as the insertion
of a stannylene into a metal–halide bond. This synthetic strategy
has proven rather useful and turned out to be surprisingly general.[8,9] An interesting conceptual consequence of such reactivity is the
fact that the formed ligand can be regarded either as a stannyl group
or alternatively as a base-stabilized stannylene ligand with chloride
acting as base. Examples of gold complexes where both types of these
ligands (base-stabilized and nonstabilized germylenes) coordinate
to one gold atom were reported only recently.[10] Earlier studies by Klinkhammer and co-workers showed that reaction
of a bis[tris(trimethylsilyl)silyl]stannylene with an arylcopper
species led to the formation of a copper stannylene complex, where
the aryl group on copper was exchanged with one of the tris(trimethylsilyl)silyl
groups.[11] Tolman and co-workers later showed
that coordination of Lappert’s diaminogermylene [(Me3Si)2N]2Ge to a β-diketiminate complex
of Cu(I) gave a non-base-stabilized germylene copper complex.[12] The interaction between carbenes, silylenes,
or germylenes and group 11 chlorides has also been studied theoretically.[13] Anandhi and Sharp reported solution emission
spectra (excitation wavelength: 385 nm) of their diamidochlorogermyl
gold phosphine complexes in toluene to show bands between 540 and
600 depending on the type of phosphine.[9,14]
Results and Discussion
In the course of studies on the coordination
chemistry of disilylated pan class="Chemical">stannylenes and germylenes we recently found
that the reaction of stabilized examples of these compounds with group
10 d10metals complexes led to the isolation of silastannene
and silagermene complexes of platinum and palladium.[15−17] The formation of the initially expected stannylene and germylene
complexes was possible only in the case of nickel.[15] In order to study the coordination chemistry of these stannylenes[18] and germylenes[19] with
other d10 complexes, we decided to turn to group 11 M(I)
complexes.
Synthesis
Reactions of PEt3-stabilized germylenes 1 and 2 with gold cyanide in benzene gave the
respective complexes 3 and 4 (Scheme 1). In contrast to related reactions with group 4
metallocenes[20,21] and nickel,[15] which led to base-free tetrylene metal complexes, the reaction
with gold cyanide was found to involve the transfer of a metal ligand,
cyanide, to the germylene. The latter can thus be regarded either
as a base-stabilized germylene or alternatively, considering the fact
that a Ge–C bond is formed, as a cyanogermyl group. This behavior
is likely caused by the known reluctance of gold to engage in back-bonding,
making the germylene particularly Lewis acidic and thus more likely
to interact with the cyanide.
Scheme 1
Reactions of Germylene Phosphine Adducts
with Group 11 Cyanides
Reaction of gold complexes 3 and 4 with B(C6F5)3 was attempted to
abstract the phosphine ligand but only led to the formation of borane
adducts 5 and 6, where the Lewis acid coordinates
to the cyanide nitrogen without altering the bonding situation in
the complex significantly (Scheme 1). Additional
B(C6F5)3 caused no further effect,
and even heating to 100 °C for 18 h did not affect the Au–P
bond according to NMR spectroscopic analysis. One property that changed,
however, was light sensitivity. While solutions of complexes 3 and 4 decomposed under the influence of daylight
within a few hours, solutions of the borane adducts 5 and 6 were found to be stable for weeks.Reaction
of the PEt3-stabilized germylene 1 with silver
triflate in benzene or dichloromethane caused precipitation of metallic
silver and the formation of a number of decomposition products. The
same reaction with silver cyanide, however, proceeded like the reactions
with gold cyanide to yield the silver complex 7 as colorless
crystals (Scheme 1). Compared to the gold complex 3, the analogous silver complex was found to be more light-sensitive.
Although the brownish decomposition products could easily be removed
by filtration over Celite, only a few minutes in daylight produced
a color change, indicating further decomposition. The molecular structure
of 7 in the solid state is very similar to that of 3; however, dimerization was observed, which is caused by
coordination of the cyanide nitrogen to the silver atom of a second
complex. This leads to the formation of a dinuclear complex featuring
an eight-membered ring.Moving to copper required a chanpan class="Chemical">ge of
the reaction conditions. While the reaction of germylene adduct 1 with copper cyanide in benzene resulted in the formation
of the expected copper complex (8) analogous to 3 and 7, the reaction was not as clean and gave
a number of byproducts. Changing to THF as a solvent proved to be
a viable alternative for the clean formation of 8. Again
the reaction needed to be carried out under exclusion of light. While
complex 8 in the solid state is again colorless, the
reaction solution is green, which may indicate the presence of Cu(II)
ions. Exposure to light causes very fast color change to brown and
eventually the formation of a copper mirror. The structure of complex 8 resembles the silver complex 7, being dimeric
in the solid state. In analogy to the reactions of gold complexes 3 and 4 also the silver and copper complexes 7 and 8 were treated with B(C6F5)3 and yielded the respective borane adducts 9 and 10 (Scheme 1).
The obtained results with germylenes 1 and 2 encourapan class="Chemical">ged us to extend our efforts to stannylenes. Stannylene adduct 11(18) was thus reacted with gold
cyanide and subsequently with B(C6F5)3 (Scheme 2) to show analogous reactivity resulting
in the formation of complexes 12 and 13.
Scheme 2
Formation of Gold Stannylene Complexes 12 and 13
NMR Spectroscopy
Unfortunately, there are no germanium isotopes with favorable NMR
spectroscopic properties. However, the obtained germylene complexes
(3–10) contain a number of other
NMR-active nuclei (1H, 11B, 13C, 19F, 29Si, 31P, 107/109Ag)
(Table 1) for proper characterization and some
insight into the electronic nature of the interaction between the
germylene and the coinagemetals. As proton and carbon chemical shifts
of the backbone of the germylene and the phosphine ligand are not
very characteristic, the respective spectra contain mainly information
about symmetry properties.
Table 1
NMR Spectroscopic
Data
compound
Siq
SiMe3
SiMe2
31P
119Sn
11B
19F
1
–127.1 (15 Hz)
–7.9 (br)
–22.7 (10 Hz)
14.8
n.a.
n.a.
n.a.
2
n.a.
–2.0 (14 Hz)/ −4.1 (8 Hz)
–16.9 (8 Hz)
15.0
n.a.
n.a.
n.a.
3
–110.4 (7 Hz)
–5.6/–8.4
–21.4 (5 Hz)
49.0
n.a.
n.a.
n.a.
4
n.a.
0.2/–3.7
–15.3 (5 Hz)
48.8
n.a.
n.a.
n.a.
5
–110.0 (6 Hz)
–5.5/–7.9
–23.2 (3 Hz)
47.5
n.a.
–11.5
–132.3/–158.0/–164.3
6
n.a.
0.8 (3 Hz)/–2.6
–16.3 (3 Hz)
47.6
n.a.
–11.9
–132.3/–158.2/–164.4
7
–114.2
–5.8/–9.0
–21.4
4.7
n.a.
n.a.
n.a.
8
–115.6
–6.6/–8.8
–22.1
–16.5
n.a.
n.a.
n.a.
9
–114.6 (dd, 11 Hz, 4 Hz)
–5.9 (3 Hz)/ −8.4 (2 Hz)
–22.7 (dd, 3 Hz, 2 Hz))
7.4 (433/366 Hz)
n.a.
11.7
132.7/–158.2/–164.4
10
–115.3 (4 Hz)
–6.0/–8.6
–22.6
–7.6
n.a.
–11.7
–132.8/–158.1/–164.3
11
–137.9 (16 Hz)
–4.2/–7.4
–20.0
–1.0
–224.4 (2220 Hz)
n.a.
n.a.
12
–125.8
–4.4/–7.4
–19.1
49.5
–126.0 (1633 Hz)
n.a.
n.a.
13
–122.7 (9 Hz)
–3.9/–7.1
–20.0 (4 Hz)
50.1
31.1 (1545 Hz)
–11.2
–132.6/–158.1/–164.4
Comparison of the 29Si
NMR spectra of starting material 1(19) and the respective gold complex 3 features
a downfield shift of the 29Si resonance of the germylene-attached
silicon atoms from −127 ppm to −110 ppm. For compound 1 at ambient temperature the observed SiMe3 signals
are very broad (at −7.9 ppm), as they are in the coalescence
regime, indicating configurational flexibility of the germanium atom.[19] In contrast complex 3 shows face
differentiation of the cyclic germylene with chemical shifts for two
SiMe3 resonances of −5.6 and −8.4 ppm. The
chemical shifts of the SiMe2 groups of 1 and 3 reflect their more remote position, as they are very much
comparable (Table 1). If the 29Si
chemical shifts of 1 and 3 are considered
to be comparable, the differences between those of 3 and
its borane adduct 5 are almost negligible (Table 1). Supposedly, coordination of B(C6F5)3 to the nitrile substituent is not really affecting
the electronic situation of the germylene unit. At the same time 19F and 11B resonances of 5 are only
marginally different from the simple B(C6F5)3–acetonitrile complex,[22] again indicating minimal interaction of the borane with the metal
complex. This NMR-spectroscopic picture is in essence also true for
the other borane adducts 6, 9, 10, and 13.Compared to the starting materials 1 (δ = 14.8 ppm) and 2 (δ = 15.0
ppm) and also to Et3PAuCN (δ = 35.4 ppm)[23] the 31P resonances of gold complexes 3, 4, and 12 are shifted downfield
to values between 48.8 and 51.0 ppm. For complexes 7 and 9 it would be attractive to obtain 107Ag or 109Ag NMR spectra. Although both isotopes are spin 1/2 nuclei
with high abundance, silver NMR is not very common.[24] The main reasons for this are very low observation frequencies,
long relaxation times, and most importantly a receptivity that is
about 5 orders of magnitude lower than for 1H. Nevertheless,
indirect observation of 107Ag and 109Ag via
coupling is well known. For silver halidephosphine complexes a relationship
between the number of coordinating phosphines and the 107/109Ag–31P coupling constant was established.[25−27] Much less is known about the 107/109Ag–31P coupling of silver phosphine complexes with heavier group 14 ligands.
Older work by Sanghani et al. points out that the complex (Ph3P)3AgSnCl3 does not feature a Sn–Ag
bond in solution but is better described as Ag(PPh3)3+SnCl3–.[28] Gade and co-workers pointed out for an R3SnAg(PPh3) (n = 1,2) complex that the 107/109Ag–31P coupling for dicoordinate silver complexes is much larger
than for related tricoordinate ones.[29] For
complexes 7 and 9 of the current study the
situation seems to be similar. For complex 7 no 107/109Ag–31P coupling could be observed
at ambient temperature. At −30 °C the signal became a
broad doublet with a coupling constant of 360 Hz, not showing resolved
coupling to the two different silver nuclei. Such behavior might be
attributed to dynamic coordination change in solution.[30] Complex 9, on the other hand, where
the nitrilenitrogen is blocked by strong borane coordination, cannot
dimerize in solution, and accordingly, even at ambient temperature 107/109Ag–31P couplings of 366 and 433 Hz
were observed in the 31P NMR spectrum.Further information
about the molecular size of 7 in solution was obtained
from 2D DOSY diffusion measurements on a sample containing 7 and 9 in C6D6. An overlay of
a regular and a proton–proton-decoupled DOSY spectrum is shown
in Figure S1. Due to signal overlap, the
regular DOSY spectrum contains strong tailing artifacts, which are
almost completely removed in the decoupled DOSY. For both 7 and 9 similar diffusion coefficients D around −9.05 (log (m2/s)) corresponding to 9.1
× 10–10 m2/s were found. The hydrodynamic
radius can be calculated from D using the diffusion
coefficient and radius of another solution component for referencing.
For this purpose we used the residual benzenesignal, which is found
at D = −8.6 (log (m2/s)), corresponding
to 2.5 × 10–9 m2/s. On the basis
of the Stokes–Einstein equation, the relative hydrodynamic
radii of two components are related to the diffusion coefficients
by D1/D2 = r2/r1. The experimental
diffusion coefficients of benzene and 7 and a hydrodynamic
diameter of benzene of ∼7.2 Å yield a hydrodynamic diameter
for 7 (and also 9) around 19 Å. The
diameter of a dimer of 7 in the crystal structure is
roughly 16 Å, which fits quite nicely to the obtained hydrodynamic
diameter and is certainly above the size of the isostructural monomeric
gold complex 3, with a diameter of ∼12 Å.
When two aggregation states are in equilibrium, the NMR-derived diffusion
coefficient is more influenced by the smaller/faster diffusing component.
This indicates that in the monomer–dimer equilibrium of 7 the dimer is the predominant form.The picture of
the 29Si NMR spectra of pan class="Chemical">silver complexes 7 and 9 resembles that of the respective gold complexes 3 and 5 (Table 1). While
the chemical shifts of the germanium-attached silicon atoms of the
silver complexes (7, 9) are slightly moved
to higher field compared to the gold complexes (3, 5), the 29Si spectra of the silver complexes (7, 9) and their respective copper analogues (8, 10) are almost identical (Table 1).
As found for the germylene complexes, the 29Si, 13C, and 1H NMR spectra of the stannylene
gold complexes 12 and 13 indicated a symmetric
stannacyclopentasilane with face differentiation (i.e., two sets of
trimethylsilyl and methyl groups). The situation is very similar to
the germylene cases of 3 and 4. 29Si NMR resonances of the 12 and 13 are
shifted somewhat to lower field compared to the stannylene adduct 11 (Table 1). 29Si NMR chemical
shifts of the SiMe3, and SiMe2 resonances of 11, 12, and 13 are almost identical.
Not unexpected also 31P shifts of 12 and 13 are close to those of 3 and 5. Complex 12 exhibits a doublet at −126 ppm in
the 119Sn NMR spectrum with a P–Sn coupling constant
of 1581 Hz, suggesting that it possesses a very low degree of stannylene
character and might rather be considered as a stannyl complex. Again,
the fact that the cyanide ion is coordinated to the tin atom indicates
that the stannylene serves mainly as a σ-donor to gold and that
the extent of back-donation from the gold atom to tin is very small.
This behavior is certainly consistent with the rather common chemical
shifts for the tin atoms. The face differentiation visible in the
NMR spectra further indicates that the interaction between the stannylene
and the cyanide is not limited to the solid state.
X-ray Crystallography
Most of the complexes studied in this account could be characterized
by X-ray single-crystal structure diffraction analypan class="Chemical">sis. The thus obtained
structural data allow comparison of compounds containing gold, silver,
and copper complexes, germylene and stannylene ligands, and B(C6F5)3 adducts of these complexes. Basic
structural data are compiled in Table 2.
Table 2
Compilation of Structural Data Derived by Single-Crystal
XRD Analysis
compound
dM–E
dM–P
dE–C
DE–Si
dB–N
∠MEC
∠EMP
∠E′EE′
3
2.402(1)/2.428(1)
2.312(3)/2.314(4)
2.01(1)/1.98(1)
2.416(3)/2.416(3)/2.415(3)/2.413(2)
n.a.
104.3(3)/108.0(3)
177.28(8)/175.7(1)
106.47(9)/107.7(1)
4
2.4292(9)/2.3981(9)
2.320(3)/2.315(2)
1.980(6)/2.023(7)
n.a.
n.a.
108.2(2)/ 104.8(2)
175.70(7)/177.28(5)
107.31(3)/106.16(3)
5
2.4165(5)
2.2979(9)
2.005(3)
2.410(1)/2.406(1)
1.564(4)
97.09(8)
176.50(2)
110.29(3)
6
2.4145(8)
2.303(2)
2.018(5)
n.a.
1.582(7)
97.1(2)
176.61(4)
110.08(3)
7
2.4848(8)
2.427(3)
1.986(5)
2.423(2)/2.427(2)
n.a.
107.9(2)
156.65(8)
104.51(5)
8
2.379(2)
2.228(3)
1.989(9)
2.428(3)/2.425(3)
n.a.
100.9(3)
142.5(1)
102.56(9)
10
2.3166(5)
2.2011(8)
1.998(2)
2.4103(8)/2.4134(8)
1.575(3)
98.02(6)
176.98(2)
108.43(2)
12
2.5867(6)/2.5713(6)
2.314(2)/2.311(2)
2.179(4)/2.208(7)
2.592(2)/2.588(2)
n.a.
109.2(1)/105.9(2)
174.83(5)/176.91(4)
104.91(5)/103.71(5)
13
2.5755(7)
2.307(1)
2.241(3)
2.583(1)/2.589(1)
1.573(4)
103.27(9)
177.67(3)
105.32(3)
The number of reported solid-state structures containing
bonds between heavy group 14 atoms and coinagemetals is not very
high. If cluster compounds are excluded, a rather small number emerges.
A search in the Cambridge Crystallographic Database (CCDC)[31] reveals Ge–Au distances for complexes
with dicoordinate gold between 2.324 and 2.423 Å.[9,10,32−37] For Ge–Ag the distance range is 2.412–2.467 Å,[8,34,38−42] for Ge–Cu a range of 2.287–2.376 Å[42−47] was found for tetracoordinate Ge (i.e., germyl groups or base-stabilized
germylenes), and two values of 2.214 and 2.249 Å were reported
for bonds between non-base-stabilized germylene ligands and copper.[12] For the distance Sn–Au with dicoordinate
gold again only a few examples are known with a reported Sn–Au
range of 2.565–2.614 Å.[10,48] Compounds 3 (Figure 1) and 4 (Figure S2), which crystallize in the monoclinic
space group P2(1)/c, are isotypic
and feature two independent molecules in the asymmetric unit. The
Ge–Au distances, which are between 2.398 and 2.429 Å,
are at the longer end of the mentioned range. Complexes 5 (Figure 2) and 6 (Figure S3), being the B(C6F5)3 adducts of 3 and 4, crystallize
in the orthorhombic space group Pna2(1). They exhibit Ge–Au
distances of 2.417 and 2.415 Å. In accordance with the discussed
NMR data, these values show that the Ge–Au interaction is essentially
not affected by the coordination of the borane. The same is also true
for the Au–P distances of complexes 3–6, which range between 2.298 and 2.320 Å. Not even the
bond between the germanium and the nitrilecarbon displays much variation
comparing complexes 3 and 4 to 5 and 6 (Table 2). The only distinct
structural difference between the core structures of these two types
of molecules is the Au–Ge–CN angle. While rather typical
tetrahedral angle values of 104° and 108° were found for
complexes 3 and 4, this angle is somewhat
diminished to values close to 97° for complexes 5 and 6. Due to the higher steric demand of the B(C6F5)3 unit, the nitrile groups are bent
away to avoid interaction with the trimethylsilyl groups. All complexes 3–6 feature an almost linear Ge–Au–P
arrangement with respective angles between 176° and 178°.
Figure 1
Crystal
structure of 3. Displacement ellipsoids are represented
at the 30% level, and hydrogen atoms have been omitted for clarity
(bond lengths in Å, angles in deg): Au(3)–P(2) 2.314(3),
Au(3)–Ge(2) 2.4277(11), C(46)–N(2) 1.147(13), C(46)–Ge(2)
1.978(12), C(1)–Si(2) 1.882(10), C(2)–Si(2) 1.876(11),
C(42)–P(2) 1.813(11), Ge(2)–Si(9) 2.416(3), Si(9)–Si(10)
2.349(4), P(2)–Au(3)–Ge(2) 175.68(11), N(2)–C(46)–Ge(2)
176.6(10), C(46)–Ge(2)–Si(9) 102.6(3), Si(12)–Ge(2)–Si(9)
107.68(10), C(46)–Ge(2)–Au(3) 107.9(3), Si(9)–Ge(2)–Au(3)
115.60(7).
Figure 2
Crystal structure of 5. Displacement
ellipsoids are represented at the 30% level, and hydrogen atoms have
been omitted for clarity (bond lengths in Å, angles in deg):
Au(1)–P(1) 2.2979(9), Au(1)–Ge(1) 2.4165(5), Ge(1)–C(41)
2.005(3), Ge(1)–Si(4) 2.4064(11), Ge(1)–Si(1) 2.4100(11),
N(1)–C(41) 1.147(4), N(1)–B(1) 1.564(4), P(1)–C(17)
1.816(3), Si(1)–Si(2) 2.3407(13), Si(2)–C(1) 1.896(5),
B(1)–C(35) 1.640(4), F(1)–C(24) 1.357(4), P(1)–Au(1)–Ge(1)
176.50(2), C(41)–Ge(1)–Si(4) 106.54(8), C(41)–Ge(1)–Si(1)
106.74(9), Si(4)–Ge(1)–Si(1) 110.29(4), C(41)–Ge(1)–Au(1)
97.09(8), Si(4)–Ge(1)–Au(1) 118.44(3), Si(1)–Ge(1)–Au(1)
115.74(3).
Crystal
structure of 3. Displacement ellipsoids are represented
at the 30% level, and hydrogen atoms have been omitted for clarity
(bond lengths in Å, angles in deg): pan class="Chemical">Au(3)–P(2) 2.314(3),
Au(3)–Ge(2) 2.4277(11), C(46)–N(2) 1.147(13), C(46)–Ge(2)
1.978(12), C(1)–Si(2) 1.882(10), C(2)–Si(2) 1.876(11),
C(42)–P(2) 1.813(11), Ge(2)–Si(9) 2.416(3), Si(9)–Si(10)
2.349(4), P(2)–Au(3)–Ge(2) 175.68(11), N(2)–C(46)–Ge(2)
176.6(10), C(46)–Ge(2)–Si(9) 102.6(3), Si(12)–Ge(2)–Si(9)
107.68(10), C(46)–Ge(2)–Au(3) 107.9(3), Si(9)–Ge(2)–Au(3)
115.60(7).
Crystal structure of 5. Displacement
ellipsoids are represented at the 30% level, and hydrogen atoms have
been omitted for clarity (bond lengths in Å, angles in deg):
pan class="Chemical">Au(1)–P(1) 2.2979(9), Au(1)–Ge(1) 2.4165(5), Ge(1)–C(41)
2.005(3), Ge(1)–Si(4) 2.4064(11), Ge(1)–Si(1) 2.4100(11),
N(1)–C(41) 1.147(4), N(1)–B(1) 1.564(4), P(1)–C(17)
1.816(3), Si(1)–Si(2) 2.3407(13), Si(2)–C(1) 1.896(5),
B(1)–C(35) 1.640(4), F(1)–C(24) 1.357(4), P(1)–Au(1)–Ge(1)
176.50(2), C(41)–Ge(1)–Si(4) 106.54(8), C(41)–Ge(1)–Si(1)
106.74(9), Si(4)–Ge(1)–Si(1) 110.29(4), C(41)–Ge(1)–Au(1)
97.09(8), Si(4)–Ge(1)–Au(1) 118.44(3), Si(1)–Ge(1)–Au(1)
115.74(3).
In contrast to the gold complexes 3 and 4, the respective silver and copper complexes 7 (Figure S4) and 8 (Figure 3), which are again isotypic and
crystallize in the monoclinic space group P2(1)/n, are dimeric in the solid state. The dimerization occurs
by coordination of the nitrilenitrogen to the respective metal of
a second complex unit. The copper and silver atoms are thus tricoordinate
and the Ge–M–P angles are 158° and 142° for
M = Ag and Cu, respectively. Once the dimerization is inhibited by
coordination of B(C6F5)3 to the nitrilenitrogen atom, the Ge–Cu–P angle of the copper complex 10 (Figure S5) is restored to an
almost linear fashion of 178°. The dimeric structures of 7 and 8 represent eight-membered rings that are
almost completely flat. As the Ge–C–N unit is close
to linear, the appearance of the rings is that of distorted hexagons.
The Ag–Ge distance of 7 of 2.4848(8) Å is
comparably long. While the complex geometry of the gold complexes 3 and 4 and their respective borane adducts 5 and 6 was almost identical, there are quite
remarkable differences between the dimeric copper complex 8 and its borane adduct 10. The most striking difference
is seen in the Cu–Ge bond length, which at 2.379(2) Å
for the dimeric complex 8 is rather long but is shortened
to 2.3166(5) Å for the monomeric complex 10. Also
the Cu–P bond follows this trend; however, not to the same
extent, as the bond length shortens from 2.228(3) Å for 8 to 2.2011(8) C for 10. It seems likely that
the bond elongation of the dimeric complex is related to the tricoordinate
bonding situation at copper and the rather strong distortion of linearity
of the Ge–Cu–P unit.
Figure 3
Crystal structure of 8. Left
side: displacement ellipsoids are represented at the 30% level and
hydrogen atoms have been omitted for clarity (bond lengths in Å,
angles in deg). Cu(1)−N(1) 2.006(8), Cu(1)−P(1) 2.228(3),
Cu(1)−Ge(1) 2.3788(16), Ge(1)−C(17-1) 1.989(9), Ge(1)−Si(1)
2.428(3), N(1)−C(17) 1.141(10), P(1)−C(18) 1.839(9),
Si(1)−Si(2) 2.340(4), Si(2)−C(2) 1.848(14), N(1)−Cu(1)−Ge(1)
105.7(2), C(17-1)−Ge(1)−Cu(1) 100.9(3), C(17)−N(1)−Cu(1)
162.0(8), N(1)−C(17)−Ge(1-1) 171.2(8). Right side: hydrogen
and carbon atoms have been omitted for clarity.
Crystal structure of 8. Left
side: displacement ellipsoids are represented at the 30% level and
pan class="Chemical">hydrogen atoms have been omitted for clarity (bond lengths in Å,
angles in deg). Cu(1)−N(1) 2.006(8), Cu(1)−P(1) 2.228(3),
Cu(1)−Ge(1) 2.3788(16), Ge(1)−C(17-1) 1.989(9), Ge(1)−Si(1)
2.428(3), N(1)−C(17) 1.141(10), P(1)−C(18) 1.839(9),
Si(1)−Si(2) 2.340(4), Si(2)−C(2) 1.848(14), N(1)−Cu(1)−Ge(1)
105.7(2), C(17-1)−Ge(1)−Cu(1) 100.9(3), C(17)−N(1)−Cu(1)
162.0(8), N(1)−C(17)−Ge(1-1) 171.2(8). Right side: hydrogen
and carbon atoms have been omitted for clarity.
The structural situation of the pan class="Chemical">tin-containing gold complexes 12 (Figure 4) and 13 (Figure S6) is very similar to that of 3/4 and 5/6. Complex 12 is isotypic to 3/4 and therefore
has two independent complexes in the asymmetric unit. The Sn–Au
bond distance, at 2.587/2.571 Å, is in the expected region. In
the same way complex 13 is isotypic to 5/6. Again the only major structural difference of the
core structures of 12 and 13 is the Au–Sn–CN
angle, which is smaller for 13 for the same reasons as
outlined above.
Figure 4
Molecular structure of 12 (displacement ellipsoid
plot drawn at the 30% probability level). Hydrogen atoms are omitted
for clarity (bond lengths in Å, angles in deg): Au(1)–P(1)
2.314(3), Au(1)–Sn(1) 2.5876(8), Sn(1)–C(1) 2.181(10),
Sn(1)–Si(4) 2.588(3), Sn(1)–Si(1) 2.590(3), Si(1)–Si(6)
2.346(4), Si(2)–C(4) 1.886(11), P(1)–C(19) 1.813(12),
C(1)–N(1) 1.143(13), P(1)–Au(1)–Sn(1) 174.74(9),
C(1)–Sn(1)–Au(1) 109.3(2), Au(1)–Sn(1)–Si(1)
118.43(6), Si(4)–Sn(1)–Si(1) 104.99(8), N(1)–C(1)–Sn(1)
177.6(9).
Molecular structure of 12 (displacement ellipsoid
plot drawn at the 30% probability level). pan class="Chemical">Hydrogen atoms are omitted
for clarity (bond lengths in Å, angles in deg): Au(1)–P(1)
2.314(3), Au(1)–Sn(1) 2.5876(8), Sn(1)–C(1) 2.181(10),
Sn(1)–Si(4) 2.588(3), Sn(1)–Si(1) 2.590(3), Si(1)–Si(6)
2.346(4), Si(2)–C(4) 1.886(11), P(1)–C(19) 1.813(12),
C(1)–N(1) 1.143(13), P(1)–Au(1)–Sn(1) 174.74(9),
C(1)–Sn(1)–Au(1) 109.3(2), Au(1)–Sn(1)–Si(1)
118.43(6), Si(4)–Sn(1)–Si(1) 104.99(8), N(1)–C(1)–Sn(1)
177.6(9).
The question whether the complexes
described in this study can be considered as cyanide adducts of germylene
(Ge(II)) complexes or as cyanogermyl (Ge(IV)) complexes may be answered
by a comparison of already known structures containing the germacyclopentasilane
motif containing either Ge(II) or Ge(IV). Evaluation of the Si–Ge
bond distances reveals that this parameter is quite sensitive to the
oxidation state of the Ge atom. While the PEt3 and NHC
germylene adducts[19] feature distances of
2.477 and 2.471 Å, respectively, for the corresponding dimethylgermylene
compound,[49] only a distance of 2.408 Å
was found. The Si–Ge bond distances of 2.475, 2.465, and 2.455
Å found for the germylene complexes of titanocene, zirconocene,
and hafnocene[21] indicate these complexes
having germylene character, which diminishes in the shown order. If
we use these values as a basis for assigning the oxidation states
of complexes 3, 5, 7, 8, and 10, we find distances between 2.406 and
2.414 Å for the gold complexes 3 and 5, distances of 2.426 and 2.427 Å for the silver complex 7, and distances between 2.410 and 2.428 Å for the copper
complexes 9 and 10 (Table 2). All these values certainly assign the Ge atom an oxidation
number of IV. Therefore, it seems more appropriate to consider these
compounds as cyanogermyl complexes. The same type of analysis carried
out for structures containing the stannacyclopentasilane leads to
analogous conclusions. The PEt3 adduct of the cyclic five-membered
stannylene features Si–Sn bond distances of 2.653 Å,[18] whereas the related bond in the diphenylstannylene
compound[50] is as short as 2.594 Å.
Again the Si–Sn bond distances found for 12 (2.592(2)/2.588(2))
and 13 (2.583(1)/2.589(1)) (Table 2) are clearly indicative of cyanostannyl ligands.A similar
analypan class="Chemical">sis of the Ge–CN and Sn–CN bonds is not as straightforward.
The number of structurally characterized cyanogermanes[51−54] and cyanostannanes[54,55] is very small. However, given
a Ge–CN distance of 1.975(11) Å for Mes3GeCN,[52] the range of 1.98 to 2.02 Å found for the
Ge–-CN distance in complexes 3–8 and 10 is quite comparable and points at a regular
covalent bond. Although the structure of Me3SnCN is known
and a Sn–CN distance of 2.295(12) Å was found, the value
is likely not very representative, as Me3SnCN exists as
a coordination polymer with hypercoordinated tin atoms in the solid
state.[55] As the quality of the structure
of Me2Sn(CN)2 (with a Sn–CN distance
of 2.27(7) Å)[54] is not very good,
the values between 2.179(4) and 2.241(3) Å for complexes 12 and 13 also are indicative of regular covalent
Sn–C bonds.
Conclusion
Using
the Et3P adducts of bissilylated germylenes and stannylenes
as starting materials, reactions with gold, silver, and copper cyanides
led to cyanogermyl or -stannyl phosphine complexes of the respective
metals. In the reaction process the cyanide and phosphine ligands
were exchanged. While solid-state structures of gold complexes displayed
the well-known linear coordination geometry, the silver and copper
complexes were found to exist as dimers in the solid state with the
nitrile serving as the bridging unit, which coordinates with the nitrogen
atom to the metal of a neighboring complex. As a result of the dimerization,
the silver and copper atoms are tricoordinate, with the dimer forming
a hexagonally shaped eight-membered ring. The deviation from the linear
coordination mode characterized by the E–M–P angle is
much stronger for copper (142°) than for silver (158°).
Subsequent reactions of the obtained complexes with B(C6F5)3 led to the formation of adducts with the
borane coordinating to the cyanide nitrogen atom. The formation of
the borane adduct also causes a restoration of the linear coordination
mode. With respect to NMR spectroscopic and structural properties
there are almost negligible differences between the free complexes
and their borane adducts, with the exception of the copper complexes.
Supposedly, the interaction between the copper atom and the germylene
is stronger in the dimeric case.In principle all observed complexes
can be seen either as cyanide adducts of metal ylene complexes with
Ge(II) or Sn(II) ligands or as cyanogermyl or -stannyl complexes with
tetravalent Ge(IV) and Sn(IV). A thorough comparison of the structural
parameters of the five-membered rings showed explicit indication for
classifying these complexes as cyanogermyl and -stannyl complexes.
This assignment is further strengthened by 119Sn NMR spectroscopic
analysis.
Experimental Section
General
Remarks
All reactions involving air-sensitive compounds were
carried out under an atmosphere of dry nitrogen or argon using either
Schlenk techniques or a glovebox. All solvents were dried using a
column-based solvent purification system.[56] Phosphine-stabilized germylenes 1(19) and 2(19) and phosphine-stabilized
stannylene 11(18) were prepared
according to published procedures. All other chemicals were obtained
from different suppliers and used without further purification.1H (300 MHz), 13C (75.4 MHz), 29Si (59.3 MHz), 31P (124.4 MHz), 11B (96.0 MHz), 19F (282.2 MHz), and 119Sn (111.8 MHz) NMR spectra
were recorded on a Varian INOVA 300 spectrometer. If not noted otherwise
for all samples, benzene-d6 was used as
solvent, or in the case of reaction samples they were measured with
a water-d2 capillary in order to provide
an external lock frequency signal. To compensate for the low isotopic
abundance of 29Si, the INEPT pulse sequence was used for
the amplification of the signal.[57,58] To obtain
self-diffusion coefficients, two-dimensional diffusion ordered spectroscopy
(DOSY) was used.[59] The employed pulse sequence
was a bipolar pulse pair longitudinal eddy current delay (BPP-LED)
sequence, using 32 scans per increment, 60 ms diffusion delay time,
1 ms gradient pulses, and variation of the gradient strength in 32
increments, linearly varied between 2% and 95% of maximum (which is
53.5 G/cm). DOSY analysis was performed using the Bruker DOSY package
of TopSpin 3.1. Due to extensive signal overlap, also an instant homonuclear
broadband-decoupled 2D DOSY spectrum was acquired.[60] For this experiment 128 scans were recorded per increment,
and for proton–proton decoupling a 10 ms 180° Gaussian
pulse during a 0.5 G/cm slice-selection gradient was used. Fifty data
chunks of 27 ms were acquired, amounting to a total acquisition time
of 1.35 s. All other parameters were the same as used for the regular
DOSY. All DOSY measurements were carried out at 300 K on a Bruker
Avance III 500 MHz NMR spectrometer using a 5 mm TXI probe with z-axis gradients.Elementary analysis was carried
out upan class="Chemical">sing a Heraeus Vario Elementar. Due to SiF4 formation
in the combustion of B(C6F5)3 adducts,
no elemental analyses were determined for these compounds.
X-ray
Structure Determination
For X-ray structure analyses the
crystals were mounted onto the tip of glass fibers, and data collection
was performed with a Bruker-AXS SMART APEX CCD diffractometer using
graphite-monochromated Mo Kα radiation (0.710 73 Å).
The data were reduced to Fo2 and corrected for absorption
effects with SAINT[61] and SADABS,[62] respectively. The structures were solved by
direct methods and refined by full-matrix least-squares method (SHELXL97).[63] If not noted otherwise, all non-hydrogen atoms
were refined with anisotropic displacement parameters. All hydrogen
atoms were located in calculated positions to correspond to standard
bond lengths and angles. All diagrams were drawn with 30% probability
displacement ellipsoids, and all hydrogen atoms were omitted for clarity.
Crystallographic data (excluding structure factors) for the structures
of compounds 3, 4, 5, 6, 7, 8, 10, 12, and 13 reported in this paper have been deposited
with the Cambridge Crystallographic Data Center as supplementary publication
nos. CCDC-936049 (3), 936047, (4), 936048 (5), 936052 (6), 936050 (7), 936053 (8), 943915 (10), 936046 (12), and 971935 (13) and can be obtained free of charge at http://www.ccdc.cam.ac.uk/products/csd/request/.
A mixture of AuCN (65 mg, 0.29 mmol) and 11 (204 mg, 0.29 mmol) was suspended in pan class="Chemical">toluene (8 mL) and
stirred for 2 h at rt. During the stirring the suspension turned into
a clear brown solution. Half of the solvent was removed under reduced
pressure and stored for 36 h at −60 °C. Colorless crystals
of 12 (266 mg, 97%) could be isolated by decantation. 1H NMR (δ in ppm): 0.84 (dq, 3JHH = 7.7 Hz, 2JPH = 8.1 Hz, 6H, P(CH2CH3)3), 0.53 (dt, 3JHH =
7.7 Hz, 3JPH = 17.8 Hz, 9H,
P(CH2CH3)3), 0.42
(s, 18H, SiMe3), 0.33 (s, 6H, SiMe2), 0.27 (s,
18H, SiMe3), 0,22 (s, 6H, SiMe2). 13C NMR (δ in ppm): 17.8 (d, 2JPC = 25.2 Hz, P(CH2CH3)3), 8.2 (P(CH2CH3)3), 3.2 (SiMe3), 2.9 (SiMe3), −1.3
(SiMe2), −1.6 (SiMe2), CN could not be
detected. 29Si NMR (δ in ppm): −4.4, −7.4,
−19.1, −125.8. 31P NMR (δ in ppm):
49.5 (br). 119Sn NMR (δ in ppm): −126.0 (d, 2JPSn = 1633 Hz). Anal. Calcd for
C23H63AuNPSi8Sn (925.09): C 29.86,
H 6.86, N 1.51. Found: C 29.33, H 6.72, N 1.84.
Authors: Bernd Findeis; Maria Contel; Lutz H. Gade; Mariano Laguna; M. Concepción Gimeno; Ian J. Scowen; Mary McPartlin Journal: Inorg Chem Date: 1997-05-21 Impact factor: 5.165
Authors: Johann Hlina; Judith Baumgartner; Christoph Marschner; Lena Albers; Thomas Müller Journal: Organometallics Date: 2013-05-29 Impact factor: 3.876
Authors: Johann Hlina; Judith Baumgartner; Christoph Marschner; Patrick Zark; Thomas Müller Journal: Organometallics Date: 2013-05-17 Impact factor: 3.876
Authors: Małgorzata Walewska; Judith Baumgartner; Christoph Marschner; Lena Albers; Thomas Müller Journal: Dalton Trans Date: 2018-05-01 Impact factor: 4.390
Authors: Léon Witteman; Cody B van Beek; Oscar N van Veenhuizen; Martin Lutz; Marc-Etienne Moret Journal: Organometallics Date: 2019-01-09 Impact factor: 3.876
Authors: Ruth D Rittinghaus; Jakub Tremmel; Ales Růžička; Christian Conrads; Pascal Albrecht; Alexander Hoffmann; Agnieszka N Ksiazkiewicz; Andrij Pich; Roman Jambor; Sonja Herres-Pawlis Journal: Chemistry Date: 2019-11-15 Impact factor: 5.236