A new and easy route to polysilanylpotassium compounds

Article abstract of DOI:10.1002/(SICI)1099-0682(199802)1998:2<221::AID-EJIC221>3.0.CO;2-G

A method is presented for the synthesis of tertiary, secondary, and primary polysilylpotassium compounds. Reaction of potassium tert-butoxide, in either DME or THF, with a suitable precursor molecule, proceeds by cleavage of a trimethylsilyl - polysilanyl bond, and formation of trimethylsilyl tert-butyl ether and a polysilanylpotassium compound. This route allows easy and flexible access to a number of novel polysilanylpotassium compounds, avoiding the hitherto common use of poisonous mercury compounds.

Full text of DOI:10.1002/(SICI)1099-0682(199802)1998:2<221::AID-EJIC221>3.0.CO;2-G

FULL PAPER
A New and Easy Route to Polysilanylpotassium CompoundsȠ[1]
Christoph Marschner
Institut für Anorganische Chemie, Technische Universität Graz
Stremayrgasse 16, 8010 Graz, Austria
Fax: (internat.) ϩ43(0)316/873-8701
E-Mail: marschner@anorg.tu-graz.ac.at
Received August 20, 1997
Keywords: Polysilanes / Polysilyl anions / Silylpotassium / Silicon
A method is presented for the synthesis of tertiary, secondary,
and primary polysilylpotassium compounds. Reaction of po-
tassium tert-butoxide, in either DME or THF, with a suitable
ether and a polysilanylpotassium compound. This route al-
lows easy and flexible access to a number of novel polysila-
nylpotassium compounds, avoiding the hitherto common use
precursor molecule, proceeds by cleavage of a trimethylsilylϪ of poisonous mercury compounds.
polysilanyl bond, and formation of trimethylsilyl tert-butyl
(Me3Si)6Si2 ϩ MeLi Ǟ (Me3Si)3SiLi ϩ MeSi(SiMe3)3
(2)
As is known from the pioneering work of Gilman and
coworkers tetrakis(trimethylsilyl)silane can react with meth-
yllithium to form tetramethylsilane and tris(trimethylsilyl)-
silyllithium.[2]
1
The results described in this paper show that this is not
a primary limitation, and that it can be easily overcome by
the correct choice of the transmetalating agent.
(Me3Si)4Si ϩ MeLi Ǟ (Me3Si)Me ϩ (Me3Si)3SiLi
(1)
1
Results and Discussion
The latter compound has been rapidly accepted as a
prototype of a bulky nucleophilic polysilyl anion and, only
recently, has been referred to as “hypersilyl” group.[3] It has
been used for the synthesis of transition metal silyl com-
pounds,[4] for the generation of silenes by means of sila-
Peterson reaction,[5] and for a number of other purposes.[6]
However, it is noticeable that this reagent, although so
widely accepted, has rarely been the subject of variation.
Only very recently have Klinkhammer[7] and Oehme[8] re-
ported on the synthesis of the potassium and magnesium
analogs, however the isotetrasilanyl moiety has almost
never been varied. Nevertheless, Apeloig et al. have recently
shown that variation of this moiety can be of some advan-
tage. They have been able to show that an increase in the
size of ligand gives rise to the formation of stable si-
lenes.[9][5g]
Although it would be interesting to have a set of bulky
polysilyl groups differing in reactivity and, especially, in
size,[10] in order to achieve different degrees of steric hin-
drance, this problem has not so far been systematically ad-
dressed. The reason for this becomes obvious if one looks
at a fact that was first pointed out by Gilman,[11] and then
recently studied in more detail by Apeloig et al.,[12] namely
that if polysilanes with “inner” SiϪSi bonds react with
methyllithium there is a pronounced tendency for these “in-
ner” SiϪSi bonds to undergo scission. This means that, for
example, hexakis(trimethylsilyl)disilane yields 1 and tris(tri-
methylsilyl)methylsilane in the reaction with methyllithium
(Eq. 2).
Our studies on the synthesis of dendritic and branched
polysilanes[13] led to the investigation of some aspects of
the chemistry of polysilyl anions. In reactions directed
towards the formation of polysilanes with monosilyl anions
and silyl halides we often found that the use of the potas-
sium compound was superior to the use of lithium anions.
So it was reasonable to also address the use of lithium vs.
potassium in the case of the “hypersilyl” anion. Although
this compound was already known from the work of Klink-
hammer,[7] transmetalation from the corresponding bis-“hy-
persilyl” mercury, zinc, or cadmium compounds with po-
tassium did not seem to be a real alternative to the use of
the readily available “hypersilyllithium” 1.[2] We felt that a
really competitive synthesis must introduce the potassium
in as easy a fashion as the lithium. So the obvious reagent
of choice was potassium tert-butylate, which very effectively
transforms dodecamethylcyclohexasilane into the corre-
sponding undecamethylcyclohexasilanylpotassium.[14]
Indeed initial experiments showed that the conversion of
tetrakis(trimethylsilyl)silane with tert-butylate in DME or
THF, in a clean and almost quantitative reaction, gives “hy-
persilylpotassium” 1a accompanied by the formation of tri-
methylsilyl tert-butyl ether (Eq. 3).
(Me3Si)4Si ϩ KOtBu Ǟ (Me3Si)3SiK ϩ (Me3Si)OtBu
(3)
1a
Changing the solvent to toluene or pentane gives 1a as a
crystalline compound, with one or two molecules of the
Eur. J. Inorg. Chem. 1998, 221Ϫ226
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1998
1434Ϫ1948/98/0202Ϫ0221 $ 17.50ϩ.50/0
221

C. Marschner
FULL PAPER
1
etheral solvent still present, as can be seen from H- and lane.[11] Similarly Apeloig et al. also found, in the reaction
13C-NMR data. This synthesis of 1a can be regarded as a of 9 with methyllithium, exclusive cleavage of one of the
significant improvement, avoiding the use of poisonous “inner” SiϪSi bonds.[12]
heavy metal compounds which are in turn prepared from 1
In contrast to these findings the reaction to give the po-
or tris(trimethylsilyl)silane.[7] It is an interesting alternative tassium polysilanyl, using potassium tert-butylate, exhibits
to use 1a instead of 1 because the reactivities are compa- the opposite selectivity, splitting exclusively one of the
rable, the potassium compound being slightly more reactive. “outer” SiϪSiMe3 bonds (10, 11).
We found that the stoichiometry of the potassium reaction
RSi(SiMe3)3 ϩ KOtBu Ǟ RSi(SiMe3)2K ϩ (Me3Si)OtBu
(5)
is easier to control dealing with potassium tert-butylate,
which is a weightable solid, compared to methyllithium
solutions which may have more or less well defined concen-
trations.
R ϭ (Me3Si)3Si
7
10
11
R ϭ (Me3Si)3SiSiMe2
9
These results open the door for the synthesis of higher
Treatment of 1a with a variety of electrophiles proceeded polysilyl anions. Both of the anions 10 and 11 are already
as smoothly as expected (Scheme 1). Reaction with aqueous known as lithium analogs. The lithium analog of 10 was
sulfuric acid gives the hydrosilane 2, reaction with ethyl bro- observed by Gilman in the reaction of 7 with 1.[11] The
mide yielded the ethylated product 3, and silylation with lithium derivative of 11 was found by Apeloig et al. in a
a number of phenylated silyl chlorides was easily achieved reaction that they describe as the addition of methyllithium
(4a,b,c). Syntheses of an acylsilane (5),[15] and of the known to a disilene.[12]
zirconocene hypersilyl chloride (6),[16] can be achieved in a
The polysilyl anions 10 and 11 are candidates for experi-
straightforward manner. The reaction of 1a with 1,2-dibro- ments in which the polysilyl moiety is required to be of
moethane at Ϫ78°C proceeds in an analogous manner to increased size compared to the “hypersilyl” unit in order to
that of the corresponding lithium compound 1, and yields fulfill its stabilizing function. It also should be mentioned
the dimerisation product 7.
that in both cases we attempted to obtain the dianionic
Scheme 1
4a can be converted quantitatively to the corresponding species by the addition of another equivalent of potassium
bromosilane 8 by a protodesilylation reaction in neat hydro- tert-butylate, but failed under normal reaction conditions.
gen bromide at Ϫ78°C. The reaction of 8 with another However, experiments to obtain such compounds are still
equivalent of 1a gives bis-“hypersilyl”-dimethylsilane 9 (Eq. in progress in our laboratory. The results reported so far
4). The latter compound was found first by Sakurai et al. as have demonstrated that the potassium polysilanyls are seri-
one of the products in the aluminum trichloride catalyzed ous competitors to their lithium analogs when higher polys-
isomerisation of linear permethylnonasilane,[17] and was re- ilyl anions are required. The following section will show
cently synthesized by Apeloig et al.[12][18]
that this chemistry is also able to yield smaller polysilyl
anions.
When ethylated 3 was subjected to standard treatment
(Me3Si)3SiSiMe2Ph ϩ HBr Ǟ (Me3Si)3SiSiMe2Br
4a
8
Ǟ [(Me3Si)3Si]2SiMe2 (4) with potassium tert-butylate the reaction proceeded as ex-
9
pected and secondary potassium trisilane 12 was formed.
Compounds 7 and 9 were chosen to test the reaction with
potassium tert-butylate for its ability to differentiate be-
tween “inner” and “outer” SiϪSi bonds. From Gilmans
work it is known that the reaction of 7 with methyllithium
(Me3Si)3SiEt ϩ KOtBu Ǟ (Me3Si)2Si(Et)K ϩ (Me3Si)OtBu
12
(6)
3
As 12 can as easily be treated with electrophiles, as was
leads to the formation of 1 and tris(trimethylsilyl)methylsi- demonstrated for 1a, the sequence: tetrakis(trimethylsilyl)si-
222
Eur. J. Inorg. Chem. 1998, 221Ϫ226

Polysilanylpotassium Compounds
FULL PAPER
lane Ǟ (Me3Si)3SiK (1a) Ǟ (Me3Si)3SiE (3) Ǟ (Me3Si)2Si- (entry 6) a 26 ppm higher field shift is observed, while the
(E)K (12) Ǟ (Me3Si)2Si(E)EЈ offers an interesting method molecule with one phenyl group shows only a 16 ppm shift.
of access to 2,2Ј-disubstituted trisilanes.[19]
Obviously some of the negative charge can be transferred
Even more interesting, but unfortunately less selective, is onto the phenyl group, resulting in a weaker deshielding
the conversion of 2. Under analogous conditions a second- effect. A very similar case is that of the monosilyl alkali
ary potassium trisilane 13 was formed with hydrogen and compounds. While trimethylsilylpotassium shows an even
potassium attached to the same silicon atom. The reaction weaker shift (only 14 ppm) compared to hexamethyldisilane
is accompanied by the formation of 1a, which seems to be (entry 8), dimethylphenylsilyllithium (entry 9) shows the
the result of a deprotonation reaction. The ratio of these weakest shift in the entire series (about 6 ppm) compared
two products is about 7:3 in favor of 13.
to the trimethylsilylated compound.
(Me3Si)3SiH ϩ KOtBu Ǟ 0.7 (Me3Si)2Si(H)K ϩ 0.3 (Me3Si)3SiK
Table 1. Comparison of 29Si-NMR shift values of polysilyl alkali
compounds with the respective trimethylsilylated molecules
2
13
1a
ϩ 0.7 (Me3Si)OtBu ϩ 0.3 HOtBu (7)
While it has been shown so far that tertiary and second-
ary potassium silanyls are easily available from the reaction
of suitable precursors with potassium tert-butylate, and it is
also known that monosilyl potassium compounds are ac-
cessible from disilanes by reaction with C8K,[20] primary
potassium silanyls are not so easy to obtain.
The reaction of octamethyltrisilane with potassium tert-
butylate in DME shows some rather complex chemistry,
which we are currently studying in more detail, but which
does not provide a suitable route to pentamethyldisilanylpo-
tassium. Since 1,1-diphenyltrimethyldisilanyl alkali com-
pounds are known to form from the reaction of 2,2-di-
phenylhexamethyltrisilane with methyllithium or sodium
potassium alloy,[21] it was decided to test our reaction con-
ditions on 2-phenylheptamethyltrisilane. However addition
of one equivalent of potassium tert-butylate to the latter
gave not only the desired disilanylpotassium compound 14,
Entry Oligosilane
δSiϪSiMe Alkali silanyl
δSiϪM ∆δSi
3
1
2
(Me3Si)4Si[a]
Ϫ135.6
Ϫ129.5
(Me3Si)3SiK[a]
(Me3Si)3SiSi(Si-
Me3)2K[a]
Ϫ195.8 60.2
Ϫ192.7 63.2
(Me3Si)3SiSi(Si-
[a]
Me3)3
3
(Me3Si)3SiSiMe2Si-
Ϫ117.8
(Me3Si)3SiSiMe2Si- Ϫ187.0 69.2
(SiMe3)3
(SiMe3)2K[a]
[a]
4
5
(Me3Si)3SiEt[a]
(Me3Si)3SiH[a]
Ϫ78.6
Ϫ115.4
Ϫ48.7
Ϫ46.0
Ϫ20.5
Ϫ21.7
Ϫ97.7
(Me3Si)2EtSiK[a] Ϫ111.7 33.1
(Me3Si)2HSiK[a]
Me3SiSiMe2Li[c]
Ϫ181.1 65.7
Ϫ74.9 26.2
[b]
6
7
8
Me3SiSiMe2SiMe3
Me3SiSiMePhSiMe3
Me3SiSiMe3
[d]
Me3SiSiMePhK[d] Ϫ62.9 16.9
Me3SiK[e]
PhMe2SiLi[b]
H3SiK[b]
Ϫ34.4 13.9
Ϫ27.3 5.6
Ϫ165.0 67.3
[b]
[b]
9
10
PhMe2SiSiMe3
Me3SiSiH3
[f]
[a]
[b]
[c]
Measured in DME. Ϫ
Value taken from ref.[26]. Ϫ Value
taken from ref.[27]. Ϫ measured in THF. Ϫ Value taken from
[d]
[e]
ref.[22]. Ϫ A sample of Me3SiSiH3 was kindly provided by Dr.
[f]
Robert Zink.
From the comparison of these data it is evident that the
but also a substantial amount of trimethylsilylpotassium, silicon atom bearing the negative charge can gain a release
which was detected by 29Si NMR. Changing the solvent to of charge which depends on the kind of substituents at-
THF slowed down the reaction rate and did improve the tached. While hydrogen and silyl groups, in accordance with
ratio, but only slightly. It was finally found that optimum their respective electropositive characters, do not take a
reaction conditions could be achieved by gradual addition great deal of the charge, charge can be transferred to methyl
of potassium tert-butylate to a THF solution of the start- and, to an even greater extent, to phenyl groups,[22] which
ing material.
in turn leads to smaller values of ∆δSi.
One should mention here that analysis of these NMR
Analysis of 29Si-NMR data from the experiments de-
scribed and from the literature (see Table 1) shows some data has to be undertaken with some caution with respect
interesting details. Comparison of the shift values of the to the solvents used in the measurement. While the shifts
silicon atom with a trimethylsilyl substituent and the analog of the uncharged compounds are relatively independent of
bearing a negative charge shows a high field shift of about solvent effects the silyl potassium compounds show a rather
60 ppm, for tertiary alkali silanyls (entries 1Ϫ3), and of 33 pronounced dependence on the solvent. Compared to the
ppm for the generation of a secondary potassium silanyl reported 29Si-NMR data of δ ϭ Ϫ185 for 1a in deuteroben-
compound with one alkyl group (entry 4). Interestingly the zene, we found that a complexing solvent such as DME
substitution of a trimethylsilyl group by hydrogen (cf. entry causes an additional upfield shift of 11 ppm, indicating a
5) also gives a ∆δSi value similar to the tertiary alkali silanyl less covalent character for the SiϪK bond. Formation of
(entry 1). Comparison of the known 29Si-NMR value of 1a in THF results in a signal at Ϫ194 ppm. Subsequent
H3SiK (entry 10) with that of the respective trimethyl- crystallization from pentane shows the presence of two mol-
1
silylated compound shows that even a trihydrogenated sili- ecules of THF in the H-NMR spectrum, and a value of
con atom shifts some 60 ppm to higher field when the tri- δ ϭ Ϫ189.5 in the 29Si-NMR spectrum. So we can observe,
methylsilyl substituent is replaced by potassium.
for one compound, a difference in the shift value of some
While the change from secondary alkali silanyls to pri- 10 ppm depending on the solvent. However, it was con-
mary ones (entry 4 vs. 6) is not so pronounced, the influ- sidered that, as the data compared in Table 1 always refer
ence of a phenyl group on the silicon atom bearing the to compounds in etheral solvents, this comparison was va-
negative charge is indicative, as can be seen by comparison lid. Also the tendencies discussed are so pronounced that
of entries 6 and 7. In the case of the permethylated molecule the solvent effects are relatively small.
Eur. J. Inorg. Chem. 1998, 221Ϫ226
223

C. Marschner
FULL PAPER
The author wants to express his gratitude to the late Prof.
Hengge for his encouragement and support, and for stimulating
discussions. The Wacker AG, Burghausen, kindly provided chloro-
silanes as starting materials. CHM is the recipient of an APART
(Austrian Program for Advanced Research and Technology) schol-
arship of the Austrian Academy of Science, Vienna which is grate-
fully acknowledged.
eral procedure. After complete conversion the solution of 1a in
THF was slowly poured into a mixture of ethyl ether (30 ml), aque-
ous H2SO4 (1  solution, 50 ml), and ice. The organic layer was
dried over Na2SO4 and evaporated in vacuo. The remaining oil (720
mg, 2.90 mmol, 93%) was pure by GC/MS. Ϫ 29Si NMR (59.6
MHz, pentane, D2O lock): δ Ϫ10.92, Ϫ115.36 (d, J ϭ 156 Hz). Ϫ
MS (70 eV): m/z (%): 248 (33) [Mϩ], 233 (27) [Mϩ Ϫ Me], 174 (82)
[Mϩ Ϫ HSiMe3], 160 (100) [Mϩ Ϫ SiMe4], 159 (88) [MH Ϫ SiMe4],
73 (80) [SiMe3]. Ϫ IR, 1H and 13C NMR data in accordance to
the literature.[25]
Experimental Section
All reactions were carried out in flame-dried glassware under an
inert atmosphere of dry argon or nitrogen. Solvents were distilled
prior to use from either sodium or sodium potassium alloy. Starting
materials tetrakis(trimethylsilyl)silane[23] and 2-phenylheptamethyl-
trisilane[24] were prepared according to literature procedures.
PhMe2SiCl, Ph2MeSiCl, Ph3SiCl, potassium tert-butylate, 1,2-di-
bromoethane, and ethyl bromide were used without further purifi-
cation.
Tris(trimethylsilyl)ethylsilane (3): Compund 3 was prepared
from tetrakis(trimethylsilyl)silane (2.00 g, 6.22 mmol) by essentially
the same method as for 2 with the difference that the solution of
1a was poured into a mixture of ethyl ether (30 ml) and ethyl bro-
mide (5 ml). After complete addition the solution was stirred for 5
min at room temp. and aqueous H2SO4 (1  solution, 50 ml) was
added. The workup was the same as for 2. The remaining white
solid (1.43 g, 5.17 mmol, 83%) was found, using GC/MS and
NMR, to be pure. An analytical sample was recrystallized from
ethanol (mp: 136Ϫ138°C). Ϫ 1H NMR (300 MHz, CDCl3): δ ϭ
1.10 (t, J ϭ 8.0 Hz, 3 H), 0.82 (q, J ϭ 7.9 Hz, 2 H), 0.17 (s, 27 H).
1H-, 13C-, and 29Si-NMR spectra were recorded on a Bruker
MSL 300 at the indicated frequencies. 29Si-NMR spectra were
measured either in C6D6 or directly from the reaction mixture using
a D2O capillary providing an external lock signal. In all cases, ex-
cept for 2 and 13, the INEPT pulse sequence was employed. Ϫ GC/
MS was performed on an HP 5890/II gas chromatgraph with an
HP 5971/A MSD, an HP1 (Cross-linked methyl silicone) column
and helium as a carrier gas. Ϫ Elemental analyses of the branched
polysilanes were done using a Heraeus elementar vario EL appar-
atus. For some reason the carbon values found in the elemental
analyses are too low when certain compounds were analyzed with
quarternary silicon atoms. While the hydrogen values correspond
nicely to the calculated numbers we attribute the low carbon values
to the formation of silicon carbide, which seems to be very favor-
able in the cases of the compounds mentioned.
Ϫ
13C NMR (75.4 MHz, CDCl3): δ ϭ 13.11, 1.39, 0.00. Ϫ 29Si
NMR (59.6 MHz, pentane, D2O lock): δ Ϫ12.35, Ϫ78.81. Ϫ MS
(70 eV): m/z (%): 276 (20) [Mϩ], 261 (4) [Mϩ Ϫ Me], 203 (19) [Mϩ
Ϫ SiMe3], 188 (30) [Mϩ Ϫ SiMe4], 175 (27) [Mϩ Ϫ EtSiMe3], 73
(100) [SiMe3]. Ϫ C11H32Si4 (276.72) calcd. C 47.75, H 11.66; found
C 47.46, H 11.71.
1,1,1-Tris(trimethylsilyl)dimethylphenyldisilane (4a): 1a was pre-
pared from tetrakis(trimethylsilyl)silane (3.10 g, 9.66 mmol) and
tert-BuOK (1.14 g, 10.1 mmol) in THF by the general procedure.
After complete reaction the solvent was removed in vacuo and tolu-
ene (15 ml) was added. The mixture was cooled to Ϫ78°C and
chlorodimethylphenylsilane (1.81 g, 10.6 mmol) was added in tolu-
ene (10 ml) gradually over 5 min. The reaction was stirred for
further 12 h at room temp.. The workup was done by the addition
of aqueous H2SO4 (1  solution, 20 ml), extraction of the aqueous
phase with two portions of ether, and drying of the combined or-
ganic phases over Na2SO4. After removal of the solvent by evapor-
ation the white residue was crystallized from ethanol to give 4a
(3.00 g, 7.83 mmol, 81%) as white crystals (mp: 156Ϫ157°C). Ϫ
1H NMR (300 MHz, CDCl3): δ ϭ 7.56 (m, 2 H), 7.36 (m, 3 H),
0.54 (s, 6 H), 0.18 (s, 27 H). Ϫ 13C NMR (75.4 MHz, CDCl3): δ ϭ
141.46, 134.32, 128.68, 127.86, 3.02, 1.42. Ϫ 29Si NMR (59.6 MHz,
pentane, D2O lock): δ ϭ Ϫ8.92, Ϫ12.48, Ϫ133.23. Ϫ MS (70 eV):
m/z (%): 382 (14) [Mϩ], 367 (7) [Mϩ Ϫ Me], 294 (4) [Mϩ Ϫ SiMe4],
232 (93) [Me2Si2(SiMe3)2], 135 (79) [SiMe2Ph], 73 (100) [SiMe3]. Ϫ
C17H38Si5 (382.92) calcd. C 53.32, H 10.00; found C, 53.04 H 9.93.
General Procedure: Starting material and one equivalent of po-
tassium tert-butylate were mixed together in a flask and either
DME or THF added. Almost immediately the solution turned yel-
low and, after some time, sometimes orange. Completion of the
reaction was detected either by 29Si-NMR spectroscopy (D2O capil-
lary as external lock) and/or by obtaining a derivative and sub-
sequent GC/MS or GC analysis. The derivative was obtained by
adding a 20-µl sample of the reaction mixture into a solution of
100 µl of ethyl bromide in 2 ml of ether, or by addition of the
sample into the etheral phase of a 2 ml ether/2 ml 2  H2SO4
mixture followed by mixing of the layers.
Tris(trimethylsilyl)silylpotassium (1a): In a typical experiment
tetrakis(trimethylsilyl)silane (3.10 g, 9.66 mmol) was treated with
tert-BuOK (1.14 g, 10.14 mmol) in 5 ml of either DME or THF
following the general procedure. While the conversion in DME was
usually complete within 10 min the use of THF sometimes required
reaction times up to 60 min. Ϫ 29Si NMR (59.6 MHz, DME, D2O
lock): δ ϭ Ϫ3.93, Ϫ195.83, (59.6 MHz, THF, D2O lock): δ ϭ
Ϫ4.55, Ϫ194.10.
1,1,1-Tris(trimethylsilyl)methyldiphenyldisilane (4b): 4b was pre-
pared following the procedure for the synthesis of 4a, and on the
same scale, using methyldiphenylchlorosilane (2.47 g, 10.6 mmol)
as an electrophile. White crystals of 4b (3.06 g, 6.88 mmol, 71%)
were isolated after crystallization from ethanol (mp: 161Ϫ164°C).
1
A sample of 1a prepared in THF on the same scale as described
above, which was isolated by removal of solvent in vacuo and crys-
tallization from toluene and pentane (3.20 g, 7.71 mmol, 80%), was
Ϫ H NMR (300 MHz, CDCl3): δ ϭ 7.56 (m, 4 H), 7.33 (m, 6 H),
0.79 (s, 3 H), 0.14 (s, 27 H). Ϫ 13C NMR (75.4 MHz, CDCl3): δ ϭ
138.90, 135.41, 128.97, 127.86, 3.10, 0.58. Ϫ 29Si NMR (59.6 MHz,
identified by 1H NMR and 13C NMR as a complex containing two pentane, D2O lock): δ ϭ Ϫ8.95, Ϫ11.95, Ϫ132.84. Ϫ MS (70 eV):
molecules of THF. Ϫ H NMR (300 MHz, C6D6): δ ϭ 3.42 (m, 8 m/z (%): 444 (4) [Mϩ], 429 (2) [Mϩ Ϫ Me], 294 (4) [Mϩ
Ϫ
1
H), 1.47 (m, 8 H), 0.47 (s, 27 H). Ϫ 13C NMR (75.4 MHz, C6D6):
PhSiMe3], 232 (100) [Me2Si2(SiMe3)2], 197 (75) [SiMePh2], 135 (34)
δ ϭ 68.33, 25.93, 7.67. Ϫ 29Si NMR (59.6 MHz, C6D6): δ ϭ [SiMe2Ph], 73 (47) [SiMe3]. Ϫ C22H40Si5 (444.99) calcd. C 59.38,
Ϫ5.35, Ϫ189.56.
H 9.09; found C 59.32, H 9.01.
Tris(trimethylsilyl)silane (2): 1a was generated in THF from
1,1,1-Tris(trimethylsilyl)triphenyldisilane (4c): 4c was prepared
tetrakis(trimethylsilyl)silane (1.00 g, 3.11 mmol) following the gen- following the procedure for the synthesis of 4a, and on the same
224 Eur. J. Inorg. Chem. 1998, 221Ϫ226

Polysilanylpotassium Compounds
FULL PAPER
scale, using triphenylchlorosilane (3.13 g, 10.5 mmol) as an elec- C6D6): δ ϭ 0.71 (s, 6 H), 0.26 (s, 27 H). Ϫ 13C NMR (75.4 MHz,
trophile. White crystals of 4c (3.74 g, 7.37 mmol, 76%) were yielded C6D6): δ ϭ 8.08, 2.90. Ϫ 29Si NMR (59.6 MHz, C6D6): δ ϭ 25.81,
1
after crystallization from ethanol (mp: 263Ϫ266°C). Ϫ H NMR
(300 MHz, CDCl3): δ ϭ 7.52 (m, 6 H), 7.37 (m, 9 H), 0.18 (s, 27
Ϫ9.61, Ϫ127.92; MS (70 eV): m/z (%): 369/371 (4/5) [Mϩ Ϫ Me],
311/313 (1/2) [Mϩ Ϫ SiMe3], 232 (45) [Mϩ Ϫ BrSiMe3], 73 (100)
H). Ϫ 13C NMR (75.4 MHz, CDCl3): δ ϭ 137.71, 136.71, 129.26, [SiMe3]. Ϫ C11H33BrSi5 (385.71) calcd. C 34.25, H 8.62; found C
127.99, 3.19. Ϫ 29Si NMR (59.6 MHz, pentane, D2O lock): δ ϭ
Ϫ9.17, Ϫ10.22, Ϫ130.84. Ϫ MS (70 eV): m/z (%): 297 (1) [Mϩ
33.97, H 8.57.
Ϫ
1,1,1,3,3,3-Hexakis(trimethylsilyl)dimethyltrisilane (9): A solu-
tion of 1a [prepared from tetrakis(trimethylsilyl)silane (1.66 g, 5.17
mmol) in DME] in toluene (5 ml) was added slowly at Ϫ50°C to
a solution of 8 (1.81 g, 4.70 mmol) in toluene (5 ml). The reaction
was allowed to reach room temp. and was followed by 29Si NMR.
After complete conversion a workup was performed by the ad-
dition of aqueous H2SO4 (1  solution, 20 ml), extraction of the
aqueous layer with two portions of ether, and drying of the com-
bined organic layers over Na2SO4. Solvent was removed in vacuo
and the residue was subjected to Kugelrohr distillation (1 mbar,
250°C) to remove volatile components. Finally crystallization of
the remaining solid was achieved from ethanol/ethyl acetate to yield
9 (1.62 g, 2.92 mmol, 62%) as white crystals (mp: 168Ϫ171; ref.[17]
SiPhMe2 Ϫ HSiMe3], 282 (8) [Mϩ Ϫ SiPhMe2 Ϫ SiMe4], 259 (93)
[SiPh3)], 232 (100) [Me2Si2(SiMe3)2], 135 (28) [SiMe2Ph], 73 (47)
[SiMe3]. Ϫ C27H42Si5 (507.06) calcd. C 63.96, H 8.35; found C
63.81, H 8.39.
Tris(trimethylsilyl)silylpivaloate (5): 1a was generated following
the general procedure from tetrakis(trimethylsilyl)silane (3.00 g,
9.35 mmol) in THF. After removal of the solvent toluene (10 ml)
was added and the solution was added slowly to an ice cold solu-
tion of pivaloyl chloride (1.16 g, 9.63 mmol) in toluene (10 ml).
After complete conversion, which was checked by GC/MS, the re-
action mixture was poured into an ice/sulfuric acid mixture (10%).
The aqueous layer was extracted with ether (2 ϫ 20 ml) and the
combined organic layer dried over Na2SO4. After removal of the
solvent final purification was achieved by chromatography on silica
with toluene/heptane, 10:1. Spectral data of the compound (2.17 g,
6.52 mmol, 70%) were in accordance with the literature.[15]
1
170Ϫ172°C). Ϫ H NMR (300 MHz, CDCl3): δ ϭ 0.50 (s, 6 H),
0.23 (s, 54 H). Ϫ 13C NMR (75.4 MHz, CDCl3): δ ϭ 5.81, 4.10. Ϫ
29Si NMR (59.6 MHz, pentane, D2O lock): δ ϭ Ϫ8.64, Ϫ25.19,
Ϫ117.70. Ϫ MS (70 eV): m/z (%): 537 (1) [Mϩ Ϫ Me], 305 (97)
[Mϩ Ϫ Si(SiMe3)3], 232 (100) [Me2Si2(SiMe3)2], 73 (95) [SiMe3]. Ϫ
C20H60Si9 (553.47) calcd. C 43.40, H 10.93; found C 41.94, H 10.92.
Tris(trimethylsilyl)silylchlorozirconocene (6): 1a was generated
following the general procedure from tetrakis(trimethylsilyl)silane
(2.00 g, 6.23 mmol) in DME. After complete conversion the solu-
tion was added, using a syringe pump, over a period of 40 min to
a suspension of Cp2ZrCl2 (1.82 g, 6.22 mmol) in pentane (10 ml).
Complete conversion was detected by 29Si NMR after another 60
min stirring at room temp. Crystallization was achieved by removal
of the solvent, addition of 20 ml of pentane and storage at Ϫ75°C.
Extremely air-sensitive red crystals (3.12 g, 5.78 mmol, 93%) were
obtained. Ϫ Spectral data in accordance to literature values.[16]
Pentakis(trimethylsilyl)disilanylpotassium (10): Compound 10
was prepared from 7 (200 mg, 0.40 mmol) and tert-BuOK (44 mg,
0.40mmol) in DME following the general procedure. The only
product after complete consumption of 7, as detected by NMR and
GC/MS of derivatives of the samples, was 10. Ϫ 29Si NMR (59.6
MHz, DME, D2O lock): δ ϭ Ϫ5.58, Ϫ10.68, Ϫ128.92, Ϫ192.62.
1
Ϫ H3Oϩ quench: H NMR (300 MHz, CDCl3): δ ϭ 2.81 (s, 1 H),
0.33 (s, 18 H), 0.32 (s, 27 H). Ϫ 13C NMR (75.4 MHz, CDCl3):
δ ϭ 3.74, 3.54. Ϫ 29Si NMR (59.6 MHz, pentane, D2O lock): δ ϭ
Ϫ11.84, Ϫ13.38, Ϫ118.02, Ϫ135.04. Ϫ MS (70 eV): m/z (%): 407
(3) [Mϩ Ϫ Me], 348 (18) [Mϩ Ϫ HSiMe3], 333 (2) [Mϩ Ϫ HSiMe4],
259 (4) [Mϩ Ϫ H2Si2Me7], 73 (100) [SiMe3]. Ϫ Ethyl bromide
Hexakis(trimethylsilyl)disilane (7): 1a was generated following
the general procedure from tetrakis(trimethylsilyl)silane (5.00 g,
15.6 mmol) in THF. After complete conversion the solution was
cooled to Ϫ78°C and 1,2-dibromoethane (1.63 g, 8.67 mmol) in
ethyl ether (30 ml) was added dropwise over a period of 30 min.
The solution then was allowed to come to room temp. and was
poured into an ice/aqueous H2SO4 mixture. The organic layer was
dried over Na2SO4 and the solvent removed in vacuo. The remain-
ing white crystals (3.80 g) were shown by 29Si NMR to be 90%
pure 7, with tetrakis(trimethylsilyl)silane being the only impurity.
Sublimation (120°C, 1 mbar) removed about 300 mg of tetrakis(tri-
methylsilyl)silane. Finally pure 7 was obtained by recrystallization
from acetone and ethanol which gave 7 (3.16 g 6.38 mmol, 82%)
1
quench: H NMR (300 MHz, CDCl3): δ ϭ 1.13 (t, J ϭ 7.0 Hz, 3
H), 0.97 (q, J ϭ 7.0 Hz, 2 H), 0.26 (s, 27 H), 0.24 (s, 18 H). Ϫ 13C
NMR (75.4 MHz, CDCl3): δ ϭ 13.52, 5.11, 4.18, 3.01. Ϫ 29Si
NMR (59.6 MHz, pentane, D2O lock): δ ϭ Ϫ8.94, Ϫ11.73, Ϫ68.92,
Ϫ130.43. Ϫ MS (70 eV): m/z (%): 377 (6) [Mϩ Ϫ SiMe3], 275 (4),
[Mϩ Ϫ 2 SiMe3, Ϫ Et], 232 (26) [Me2Si2(SiMe3)2], 73 (100) [SiMe3].
The experiment was repeated in THF with essentially the same re-
sults. Removal of solvent and crystallization from pentane gave
white crystals solvated with two molecules of THF in an almost
quantitative yield (220 mg, 0.37 mmol, 93%). Ϫ 1H NMR (300
MHz, C6D6): δ ϭ 3.41 (m, 8 H), 1.46 (m, 8 H), 0.51 (s, 18 H),
0.46 (s, 27 H). Ϫ 13C NMR (75.4 MHz, C6D6): δ ϭ 68.28, 25.94,
8.82, 4.78.
1
as white crystals. Ϫ H NMR (300 MHz, CDCl3): δ ϭ 0.20 (s). Ϫ
13C NMR (75.4 MHz, CDCl3): δ ϭ 4.60. Ϫ 29Si NMR (59.6 MHz,
pentane, D2O lock): δ ϭ Ϫ9.04, Ϫ129.49. Ϫ MS (70 eV): m/z (%):
406 (0.2) [Mϩ Ϫ SiMe4], 247 (6) [Mϩ Ϫ Si(SiMe3)3], 232 (29) [Mϩ
Ϫ MeSi(SiMe3)3], 73 (100) [SiMe3].
1,1,3,3,3-Pentakis(trimethylsilyl)-2,2-dimethyltrisilanylpotassium
(11): Compound 11 was prepared from 9 (95 mg, 0.17 mmol) and
tert-BuOK (19 mg, 0.17 mmol) in DME following the general pro-
cedure. The reaction was followed directly by 29Si NMR and GC/
MS of the quenched samples. Conversion was complete after 60
min. Addition of another equivalent of tert-BuOK (19 mg, 0.17
mmol) did not result in a second splitting reaction but led to the
slow decomposition of the product after two days. Ϫ 29Si NMR
(59.6 MHz, DME, D2O lock): δ ϭ Ϫ3.21, Ϫ8.80, Ϫ16.12, Ϫ127.09,
Ϫ186.99. Ϫ H3Oϩ quench: MS (70 eV): m/z (%): 480 (0.1) [Mϩ],
1,1,1-Tris(trimethylsilyl)dimethylbromodisilane (8):
A
flask
equipped with a reflux condenser cooled to Ϫ80°C and charged
with a sample of 4a (2.40 g, 6.27 mmol) was cooled to Ϫ80°C and
approx. 10 ml of hydrogen bromide was condensed on under a
pressure of 200Ϫ500 mbar. After 30 min the cooling bath was re-
moved and the solution of 4a in hydrogen bromide was refluxed
for 90 min. Then hydrogen bromide was removed in vacuo and
pentane was distilled onto the residue (10 ml). The solution was
filtered over a plug of glass wool and the solvent of the filtrate was
removed in vacuo. The remaining white powder of 8 (2.40 g, 6.22 406 (3) [Mϩ Ϫ HSiMe3], 305 (63) [Mϩ Ϫ HSi(SiMe3)3], 231 (25)
1
mmol, 99%) was pure by GC and NMR. Ϫ H NMR (300 MHz, [Me2Si2(SiMe3)2 Ϫ H], 73 (100) [SiMe3]. Ϫ Ethyl bromide quench:
Eur. J. Inorg. Chem. 1998, 221Ϫ226
225

C. Marschner
FULL PAPER
MS (70 eV): m/z (%): 493 (0.2) [Mϩ Ϫ Me], 305 [Mϩ Ϫ EtSi(Si-
Me3)2], 232 (30) [Me2Si2(SiMe3)2], 188 (24) [EtSi(SiMe3) Ϫ Me], 73
(100) [SiMe3)].
berg, made at the Xth International Symposium on Organosilicon
Chemistry in Posnan, Poland, 1993.
[4]
[5]
[4a] T. D. Tilley in The Chemistry of Organic Silicon Compounds,
Chapter 24, (Eds.: S. Patai, Z. Rappoport), Wiley, Chichester,
[4b]
1989, p. 1415. Ϫ
T. D. Tilley in The Silicon Heteroatom
Bis(trimethylsilyl)ethylsilylpotassium (12): Compound 12 was
prepared from 3 (89 mg, 0.36 mmol) and tert-BuOK (40 mg, 0.36
mmol) in DME following the general procedure. The reaction was
followed directly by 29Si NMR and GC/MS of the quenched
samples. It was found to be complete after 25 min. Ϫ 29Si NMR
(59.6 MHz, DME, D2O lock): δ ϭ Ϫ6.82, Ϫ111.68. Ϫ H3Oϩ
quench: MS (70 eV): m/z (%): 204 (14) [Mϩ], 189 (7) [Mϩ Ϫ Me],
Bond, (Eds.: S. Patai, Z. Rappoport), Wiley, Chichester, 1991,
p. 309.
[5a] D. Hoffmann, H. Reincke, H. Oehme, Z. Naturforsch. 1996,
[5b]
51b, 370. Ϫ
F. Luderer, H. Reincke, H. Oehme, Chem. Ber.
[5c]
1996, 129, 15. Ϫ
Ch. Wendler, H. Oehme, Z. Anorg. Allg.
[5d]
Chem. 1996, 622, 801. Ϫ
C. Krempner, H. Reincke, H.
[5e]
Oehme, Chem. Ber. 1995, 128, 1083. Ϫ
F. Luderer, H. Re-
incke, H. Oehme, J. Organomet. Chem. 1996, 510, 181. Ϫ [5f] C.
Krempner, H. Reincke, H. Oehme, Chem. Ber. 1995, 128, 143.
161 (8) [Mϩ Ϫ SiMe3], 130 (29) [Mϩ Ϫ HSiMe3], 102 (65) [Mϩ
EtSiMe3)], 73 (100) [SiMe3]. Ϫ Ethyl bromide quench: MS (70 eV):
m/z (%): 232 (12) [Mϩ], 159 (16) [Mϩ Ϫ SiMe3], 131 (27) [MHϩ
Ϫ
Ϫ
[5g] D. Bravo-Zhivotovskii, V. Braude, A. Stanger, M. Kapon,
Y. Apeloig, Organometallics 1992, 11, 2326.
[6] [6a]
E. Jeschke, T. Gross, H. Reincke, H. Oehme, Chem. Ber.
Ϫ
1996, 129, 841. Ϫ [6b] Y. Apeloig, I. Zharov, D. Bravo-Zhivotov-
EtSiMe3], 73 (100) [SiMe3]. Repeating the experiment the same
scale in THF led to the same results with some increased reaction
time. Removals of solvent and crystallization from pentane gave
white crystals (56 mg, 0.20 mmol, 56%), which were shown by
NMR to contain half a molecule of THF coordinated to the silyl
potassium compound. Ϫ 1H NMR (300 MHz, C6D6): δ ϭ 3.37 (m,
2 H), 1.37 (m, 2 H), 1.24 (t, J ϭ 7.8 Hz, 3 H), 0.86 (q, J ϭ 7.8 Hz,
2 H), 0.32 (s, 18 H). Ϫ 13C NMR (75.4 MHz, C6D6): δ ϭ 68.42,
25.82, 19.57, 5.03, 3.69.
skii, Y. Ochvinnikov, Y. Struchkov, J. Organomet. Chem. 1995,
[6c]
499, 73. Ϫ
D. Bravo-Zhivotovskii, Y. Apeloig, Y. Ochvinni-
kov, V. Igonin, Y. T. Struchkov, J. Organomet. Chem. 1993, 446,
123. Ϫ [6d] H. Oehme, R. Wustrack, A. Heine, G. M. Sheldrick,
D. Stalke, J. Organomet. Chem. 1993, 452, 33.
K. W. Klinkhammer, W. Schwarz, Z. Anorg. Allg. Chem. 1993,
619, 1777.
C. Krempner, H. Oehme, J. Organomet. Chem. 1994, 464, C7.
For the synthesis of (tert-BuMe2Si)(Me3Si)2SiLi and (tert-Bu-
Me2Si)2(Me3Si)SiLi see: Y. Apeloig, M. Bendikov, M. Yuzefov-
ich, M. Nakash, D. Bravo-Zhivotovskii, D. Bläser, R. Boese, J.
Am. Chem. Soc. 1996, 118, 12228.
For an interesting comparison of the size of ligands see: J. Frey,
E. Schottland, Z. Rappoport, D. Bravo-Zhivotovskii, M. Nak-
ash, M. Botoshansky, M. Kaftory, Y. Apeloig, J. Chem. Soc.
Perkin Trans. 2 1994, 2555.
[7]
[8]
[9]
Bis(trimethylsilyl)silylpotassium (13): Compound 13 was pre-
pared from 2 (775 mg, 3.11 mmol) and tert-BuOK (385 mg, 3.43
mmol) in DME following the general procedure. The reaction was
followed directly by 29Si NMR and GC/MS of the quenched
samples. Complete conversion was reached after 2 h and gave 13
[10]
[11]
[12]
H. Gilman, R. L. Harrel, J. Organomet. Chem. 1967, 9, 67.
Y. Apeloig, M. Yuzefovich, M. Bendikov, D. Bravo-Zhivotov-
skii, K. Klinkhammer, Organometallics 1997, 16, 1265.
Ch. Marschner, E. Hengge in: Organosilicon Chemistry III, N.
Auner, J. Weis , Eds., VCH, Weinheim, in press.
and 1a in a ratio of about 7:3. Ϫ IR ν: 1891 cmϪ1 (SiϪH). Ϫ 29Si
˜
[13]
[14]
NMR (59.6 MHz, DME, D2O lock): δ ϭ Ϫ4.02, Ϫ181.14 (d, J ϭ
82 Hz). Ϫ H3Oϩ quench: MS (70 eV): m/z (%):, 176 (11) [Mϩ], 161
(16) [Mϩ Ϫ Me], 101 (9) [Mϩ Ϫ H2SiMe3], 88 (21) [Mϩ Ϫ SiMe4],
73 (100) [SiMe3]. Ϫ Ethyl bromide quench: MS: vide supra. The
experiment was repeated in THF on larger scale and gave essen-
tially the same results. Removal of the solvent and crystallization
from pentane gave a mixture of 13 and 1a with THF coordinated
to the alkali silanyl compounds. Ϫ 1H NMR (300 MHz, C6D6):
δ ϭ 3.51 (m), 1.38 (m), 1.29 (s, 1 H), 0.47 (s, 18 H), 0.43 (s, 27 H).
F. Uhlig, P. Gspaltl, M. Trabi, E. Hengge, J. Organomet. Chem.
1995, 493, 33.
[15] [15a]
A. G. Brook, J. W. Harris, J. Lennon, M. El Sheik, J. Am.
[15b]
Chem. Soc. 1979, 101, 83. Ϫ
A. G. Brook, F. Abdesajken,
G. Gutekunst, N. Plavac, Organometallics 1982, 1, 994.
B. K. Campion, J. Falk, T. D. Tilley, J. Am. Chem. Soc. 1987,
109, 2049.
[16]
[17]
[18]
M. Ishikawa, J. Iyoda, H. Ikeda, K. Kotake, T. Hashimoto, M.
Kumada, J. Am. Chem. Soc. 1981, 103, 4845.
An X-ray structure of 9 reveals a bond angle between the two
hypersilyl groups and the central silicon atom of 125.7°. Ch.
Marschner, Ch. Kayser, P. Felde, to be published elsewhere.
The only other convenient access to 2-metallated trisilanes
which is known to the author employs transmetallation from
the respective mercury substituted compounds: A. Sekiguchi,
M. Nanjo, C. Kabuto, H. Sakurai, J. Am. Chem. Soc. 1995,
117, 4195.
1-Phenyltetramethyldisilanylpotassium (14): Compound 14 (200
mg, 0.75 mmol) was prepared from 2-phenylheptamethyltrisilane
in THF following the general procedure, with the variation that
tert-BuOK (85 mg, 0.75 mmol) was added in several portions over
two hours. Ϫ 29Si NMR (59.6 MHz, THF, D2O lock): δ ϭ Ϫ12.33,
Ϫ62.96. Ϫ H3Oϩ quench: MS (70 eV): m/z (%): 194 (21) [Mϩ], 179
(15) [Mϩ Ϫ Me], 135(83) [Mϩ Ϫ HSiMe2], 121 (21) [Mϩ Ϫ SiMe3],
105 (28) [Mϩ Ϫ HSiMe4], 73 (100) [SiMe3]. Ϫ Ethyl bromide
quench: MS (70 eV): m/z (%): 222 (20) [Mϩ], 207 (4) [Mϩ Ϫ Me],
193 (14) [Mϩ Ϫ Et], 149 (39) [Mϩ Ϫ SiMe3], 135 (44) [SiPhEt], 121
(100) [HSiPhMe], 105 (17) [SiPh], 73 (27) [SiMe3].
[19]
[20]
[21]
A. Fürstner, H. Weidmann, J. Organomet. Chem. 1988, 354, 15.
A. G. Brook, A. Baumegger, A. J. Lough, Organometallics 1992,
11, 310.
[22]
[23]
[24]
G. A. Olah, R. J. Hunadi, J. Am. Chem. Soc. 1980, 102, 6989
H. Gilman, C. L. Smith, J. Organomet. Chem. 1967, 8, 245.
M. Kumada, M. Ishikawa, S. Maeda, J. Organomet. Chem.
1964, 2, 478.
[25] [25a]
H. Bürger, W. Killian, J. Organomet. Chem. 1971, 26, 47.
H. C. Marsmann, W. Raml, E. Hengge, Z. Naturforsch.
[25b]
Ϫ
Ƞ In memoriam Prof. Edwin Hengge.
1981, 35b, 1541.
[1]
[26]
[27]
Presented in part at the XIth International Symposium on Or-
H. C. Marsmann in NMR Basic Principles and Progress 17,
29Si-NMR Spectroscopic Results, (Eds.: P. Diehl, E. Fluck, R.
Kosfeld), Springer-Verlag Berlin Heidelberg, 1981.
K. Krohn, K. Khanbabaee, Angew. Chem. 1994, 106, 100; An-
gew. Chem., Int. Ed. Engl. 1994, 33, 99.
[97197]
ganosilicon Chemistry, 1.Ϫ6. Sept. 1996, Montpellier, France.
[2]
[2a] H. Gilman, J. M. Holmes, C. L. Smith, Chem. Ind. (London)
[2b]
1965, 848. Ϫ
G. Gutekunst, A. G. Brook, J. Organomet.
Chem. 1982, 225, 1.
Using the term “hypersilyl” we follow the proposal of N. Wi-
[3]
226
Eur. J. Inorg. Chem. 1998, 221Ϫ226