Intermediates and products of the hexachlorodisilane cleavage of group 14 element phosphanes and amines - Molecular structure of di-tert- butyl(trichlorosilyl)phosphane in the gas phase determined by electron diffraction and ab Initio calculations

Article abstract of DOI:10.1002/(sici)1099-0682(199908)1999:8<1381::aid-ejic1381>3.0.co;2-h

Reactions of dialkyl(trimethylsilyl)phosphanes RR'PSiMe₃ (1: R, R' = tBu; 3: R, R' = iPr; 5: R = iPr, R' = tBu) with Si₂Cl₆ provide stable trichlorosilylphosphanes RR'PSiCl₃ (2, 4, 6); the reactions of silyl- and stannylamines of iPr₂NMMe₃ (M = Si: 11; M = Sn: 12) with Si₂Cl₆, however, provide the stable pentachlorodisilanylamine iPr₂NSi₂Cl₅ (13). Heating of 1 with the technical mixture Me₂(Cl)SiSiCl₂Me/(MeCl₂Si)₂ yields the stable silylphosphane tBu₂PSiMe₂Cl (8) and the disilanylphosphane tBu₂PSi(Me)(Cl)Si(Me)Cl₂ (9). Methylation of 9 with MeLi gave tBu₂PSi₂Me₅ 10, which was isolated in a pure state. Reactions of tBu(iPr)PSiMe₃ (5) and of organometal phosphanes tBu(iPr)PMR₃ (14: M = Ge, R = Me; 17a-c: M = Sn; R = Me, Et, nBu) with Si₂Cl₆ were monitored by ³¹P, ²⁹Si, and ¹¹⁹Sn NMR. - In the first step of these reactions, new tBu(iPr)PSi₂Cl₅ (7) is formed. 7 is accompanied by increasing amounts of tBu(iPr)PSiCl₃ (6) and Me₃GeSiCl₃ (15)/(Me₃Ge)₂Si(SiCl₃)₂ (16) or traces of compounds R₃SnSiCl₃ (19a-c) that decompose providing (R₃Sn)₂Si(SiCl₃)₂ (18a-c) and nBu₃SnSi(SiCl₃)₃ (20c). Subsequently, compounds 19a-c decompose providing increasing amounts of 18a-c. Stannylphosphane 17b is also cleaved by SiCl₄ leading to 6 with liberation of Et₃SnCl, whereas 17b is formed from the reaction of 5 with Et₃SnCl under liberation of Me₃SiCl. The suggestion of an extra stabilisation of P-Si bonds of trichlorosilylphosphanes was subjected to direct evidence through the structure determination of the trichlorosilylphosphane tBu₂PSiCl₃ (2) in the gas phase by electron diffraction. This crowded molecule has a 'normal' P-Si bond length of 225.0(12) pm; its C₁ symmetric conformation with both tBu groups and the SiCl₃ group twisted about 17°from the perfectly staggered positions, and with each of the three groups tilted about 6°away from each other, allows to reduce steric strain.

Full text of DOI:10.1002/(sici)1099-0682(199908)1999:8<1381::aid-ejic1381>3.0.co;2-h

FULL PAPER
Intermediates and Products of the Hexachlorodisilane Cleavage of Group 14
Element Phosphanes and Amines ؊ Molecular Structure of
Di-tert-butyl(trichlorosilyl)phosphane in the
Gas Phase Determined by Electron Diffraction and ab Initio Calculations
Wolf-W. du Mont,*[a] Lars Müller,[a] Reiner Martens,[a]
Paul M. Papathomas,[b] Bruce A. Smart,[b] Heather E. Robertson,[b]
and David W. H. Rankin*[b]
Keywords: Phosphanes / Disilanes / Disilanylamines / Germanes / Stannanes / Stannanes
Reactions of dialkyl(trimethylsilyl)phosphanes RRЈPSiMe3 (1:
compounds R3SnSiCl3 (19a–c) that decompose providing
R, RЈ = tBu; 3: R, RЈ = iPr; 5: R = iPr, RЈ = tBu) with Si2Cl6 (R3Sn)2Si(SiCl3)2 (18a–c) and nBu3SnSi(SiCl3)3 (20c).
provide stable trichlorosilylphosphanes RRЈPSiCl3 (2, 4, 6);
the reactions of silyl- and stannylamines of iPr2NMMe3 (M =
Si: 11; M = Sn: 12) with Si2Cl6, however, provide the stable
pentachlorodisilanylamine iPr2NSi2Cl5 (13). Heating of 1
Subsequently, compounds 19a–c decompose providing
increasing amounts of 18a–c. Stannylphosphane 17b is also
cleaved by SiCl4 leading to 6 with liberation of Et3SnCl,
whereas 17b is formed from the reaction of 5 with Et3SnCl
with the technical mixture Me2(Cl)SiSiCl2Me/(MeCl2Si)2 under liberation of Me3SiCl. The suggestion of an extra
yields the stable silylphosphane tBu2PSiMe2Cl (8) and the
disilanylphosphane tBu2PSi(Me)(Cl)Si(Me)Cl2 (9). Meth-
ylation of 9 with MeLi gave tBu2PSi2Me5 10, which was
isolated in a pure state. Reactions of tBu(iPr)PSiMe3 (5) and
stabilisation of P–Si bonds of trichlorosilylphosphanes was
subjected to direct evidence through the structure
determination of the trichlorosilylphosphane tBu2PSiCl3 (2)
in the gas phase by electron diffraction. This crowded
of organometal phosphanes tBu(iPr)PMR3 (14: M = Ge, R = molecule has a “normal” P–Si bond length of 225.0(12) pm;
Me; 17a–c: M = Sn; R = Me, Et, nBu) with Si2Cl6 were
monitored by 31P, 29Si, and 119Sn NMR. – In the first step of
its C1 symmetric conformation with both tBu groups and the
SiCl3 group twisted about 17° from the perfectly staggered
these reactions, new tBu(iPr)PSi2Cl5 (7) is formed. 7 is positions, and with each of the three groups tilted about 6°
accompanied by increasing amounts of tBu(iPr)PSiCl3 (6) and
Me3GeSiCl3 (15)/(Me3Ge)2Si(SiCl3)2 (16) or traces of
away from each other, allows to reduce steric strain.
products of reductive C-trichlorosilylations,[3,4] we recently
used the “Benkeser” type of reaction of chlorotrimethyl-
stannane with the trichlorosilane/triethylamine reagent as
an alternative route to trichlorosilyltrimethylstannane.[5] In
the course of the extension of this type of reductive trichlo-
rosilylation for the preparation of trihalosilyl derivatives of
main group elements, we were surprised by the nature of
the products from the trichlorosilane/triethylamine reagent
with chlorotrimethylgermane[6] and dichlorodimethylger-
Introduction
Trihalosilyl compounds are of importance as trifunc-
tional precursors for the synthesis of highly functionalised
silicon compounds. The cleavage of trimethylgermyl and
-stannyl phosphanes with hexachlorodisilane was reported
to furnish trichlorosilylphosphanes as well as trichlorosilyl-
trimethylgermane and trichlorosilyltrimethylstannane.[1]
This novel reductive trichlorosilylation of Me3Ge and
Me3Sn groups attached to phosphorus was thought to be
due to nucleophilic attack of Me3Ge and Me3Sn groups by
latent trichlorosilyl anions. Such anions are generated from
one SiCl3 group of Si2Cl6 when the other silicon atom is
attacked by a phosphorus nucleophile leading to an ad-
duct.[2] Because in situ-generated trichlorosilyl anions are
also the key precursors in reactions of trichlorosilane/tri-
ethylamine mixtures with various organic halides leading to
mane.[6,7] Contradictory data concerning liquid Me3GeϪ
[1,6]
SiCl3 and solid (Me3Ge)2Si(SiCl3)2
led us to study the
“Benkeser” type of reaction of organotin and organoger-
manium halides with the trichlorosilane/triethylamine re-
agent in detail by heteronuclear NMR.[6,7]
[a]
Institut für Anorganische und Analytische Chemie der Techni-
schen Universität Braunschweig,
Hagenring 30, DϪ38106 Braunschweig, Germany
Fax: (internat.) ϩ49 (0)531/ 391 5387
E mail: w.du-mont@tu-bs.de
The observation of unexpected products of the consti-
tution (Me3M)2Si(SiCl3)2 (M ϭ Ge, Sn) from these reac-
tions led us to reinvestigate also, by heteronuclear NMR-
spectroscopy, the course of the hexachlorodisilane cleavage
of PϪGe and PϪSn bonds which yields trichlorosilylphos-
[b]
Department of Chemistry, The University of Edinburgh,
West Mains Road, Edinburgh, EH9 3JJ, U. K.
Fax: (internat.) ϩ44 (0)131/ 650 6452
E mail: d.w.h.rankin@ed.ac.uk
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392
WILEY-VCH Verlag GmbH, D-69451 Weinheim, 1999
1434Ϫ1948/99/0808Ϫ1381 $ 17.50ϩ.50/0
1381

W.-W. du Mont et al.
FULL PAPER
phanes accompanied by new trichlorosilyl compounds con- to oxygen and moisture. The constitution of 9 was con-
taining SiϪGe and SiϪSn bonds.[1] These reactions are ex- firmed by 1H-, 13C-, 29Si-, and 31P-NMR spectra. Disilanyl-
pected to be related to the known cleavage of trimethyl- phosphane 9 (about 50% yield) is obviously the product of
silylphosphanes tBu2PSiMe3 (1) and tBuP(SiMe3)2 with a transsilylation of 1 with (MeCl2Si)2. The formation of 8
hexachlorodisilane, which allowed the isolation of the re- from 1 (24%) with Me2(Cl)SiSiCl2Me is more surprising.
lated trichlorosilylphosphanes tBu2PSiCl3 (2) and tBuP- The formation of 8 by transsilylation of 1 with Me2SiCl2 is
(SiMe3)(SiCl3) in high yields.[8] As “model” phosphanes we known to proceed significantly more slowly than the reac-
chose silyl-, germyl-, and stannylphosphane derivatives con- tion of 1 with MeSiCl3.[10] Adding 9 to an excess of meth-
taining the tBu(iPr)P group, which are moderately bulky yllithium led to permethylation giving tBu2PSi2Me5 (10),
and fairly accessible from the corresponding chlorophos- which was then isolated in a pure state.
phane on the HSiCl3/NEt3 pathway[5] followed by meth-
2. Reactions of Si2Cl6 with iPr2NMMe3 (11: M ؍ Si;
ylation and transmetalation reactions.[9]
12: M ؍ Sn)
Heating the mixture of Si2Cl6 and iPr2NSiMe3 (11) for 6
h at 60°C led to complete consumption of 11. Distillation
provided iPr2NSi2Cl5 (13) in fair yield. The stannylamine
12 is more reactive: after mixing 12 and Si2Cl6 at 0°C and
warming to room temperature, the formation of crude 13
Reactions of Hexachlorodisilane with Silyl-, Germyl-,
and Stannylphosphanes Followed by NMR
Spectroscopy
and Me3SnCl was quantitative.
1. Reactions of Chlorodisilanes with Silylphosphanes
The reaction of Si2Cl6 with tBu2PSiMe3 (1) was reported
to provide trichlorosilylphosphane tBu2PSiCl3 (2) in high
yield under mild conditions.[8] Similarly, iPr2PSiMe3 (3) was
easily cleaved to give iPr2PSiCl3 (4) in fair yield. After distil-
lation of 2 and 4, the residues contained orange oils that
are expected to contain yet undefined polysilanes, possibly
with a certain degree of SiPR2 functions. Addition of Si2Cl6
to tBu(iPr)PSiMe3 (5) at 0°C provided a red-brown solu-
tion, which turned brown on warming to room temperature.
31P-NMR spectra showed that 5 min after mixing the com-
pounds about 75% 5 was accompanied by about 8% of tBu-
(iPr)PSiCl3 (6), 8% of tBu(iPr)PSi2Cl5 (7), and about 3Ϫ4%
each of two other new phosphorus compounds. After
80 min compound 7 appeared as the main product. After
13 h, 5 was completely consumed and 6 became the main
product (about 55%), accompanied by 7 (about 40%) and
two other phosphanes. After 10 d at room temp. more than
90% of 6 and less than 10% of other phosphanes were pre-
sent in solution.
3. Reaction of Si2Cl6 with tBu(iPr)PGeMe3 (14)
Five minutes after mixing the starting materials, about
30% of 14 was consumed in favour of five new phosphorus
compounds, among them 6 and 7. After 3 h, 50% of 14 had
been consumed, 6 was the main product and small amounts
of 7 (among other P-containing compounds) were also pre-
sent. A 29Si-NMR spectrum confirmed the presence of
Me3GeSiCl3 (15) (a rather broad signal, 29Si ϭ ϩ17.7)[7]
and (Me3Ge)2Si(SiCl3)2 (16) (two signals, at δ 29Si ϭ ϩ16.8
and Ϫ84.9, the upfield one being less intense)[7] as well as
SiCl4. One peak of the 29Si doublet, doublet pattern of 6
(from Si,P and Si,H couplings) shows “enhanced intensity”,
possibly due to accidental overlap with the singlet signal of
another new germylsilane [δ 29Si ϭ ϩ10.1; probably Me3-
GeSi(SiCl3)3[9]]. This signal is quite intense in the earliest
stage of the reaction. After several days, when 14 had been
completely consumed and the crude yield of 6 was close to
90%, the 29Si-NMR pattern of 6 exhibited the expected four
lines of equal intensity. Evidence for the formation of Me3-
GeCl was provided by 13C-NMR spectra of the reaction
mixture.
4. Reaction of Si2Cl6 with Stannylphosphanes
The reaction of tBu2PSiMe3 (1) with an excess of the in- tBu(iPr)PSnR3 (17a؊c: R ؍ Me, Et, nBu)
dustrial disilane fraction Me2(Cl)SiSiCl2Me/(MeCl2Si)2 re-
tBu(iPr)PSnMe3 (17a): Twenty minutes after mixing the
quired heating for several hours. Distillation provided satis- starting materials 17a had been completely consumed. Silyl-
factory yields of the silylphosphane tBu2PSiMe2Cl (8)[10] as phosphane 6 and the branched bis(stannyl)silane 18a were
first fraction and the new disilanylphosphane tBu2PSi- the main products; small amounts of disilanylphosphane 7
(Me)(Cl)Si(Me)Cl2 9 as second fraction. Both compounds (less than 10% of 6, by 31P-NMR), stannylsilane 19a (about
are thermally stable at room temperature, but very sensitive 25% of 18a), and a further stannylsilane [δ 29Si ϭ ϩ11.3
1382
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392

Intermediates and Products of the Hexachlorodisilane Cleavage
FULL PAPER
and Ϫ88; probably Me3SnSi(SiCl3)3[9]] were also present.
These compounds can be detected by NMR in solution,
After 15 h, the crude yields (by NMR) of 6 and of 18a were but they are not persistent under the reaction conditions
both more than 90%. (presence of phosphorus nucleophiles). Their decompo-
tBu(iPr)PSnEt3 (17b): After 5 min, 17b had been con- sition furnishes new branched stannylsilanes (R3Sn)2Si-
sumed and large amounts of the disilanylphosphane 7 were (SiCl3)2 18 and R3SnSi(SiCl3)3 20 (Equations 11, 12). The
present. Tin-NMR spectra showed the presence of large previous assignments of NMR data of compounds 15 and
amounts of Et3SnSiCl3 (19b) accompanied by 18b[7] and 19a made in ref.[1] have to be withdrawn (see the following
Et3SnCl. From such a sample, mass spectrometric evidence section, Table 1). The isolated solids were not 15 and 19a,[1]
for 19b (Mϩ Ϫ Et) was provided. Within one hour, however, but 16 and 18a.[7] The formation of compounds 20 from 19
trichlorosilylphosphane 6 and the branched bis(stannyl)sil- (which is favoured by bulkier alkyl groups at tin) can be
ane 18b became the main products at the expense of 7 and explained as being reaction steps analogous to those that
19b. Besides 18b, the presence of a further branched silyl- occur in the course of the base-catalysed disproportionation
Ϫ
stannane was indicated by 29Si-NMR (δ 29Si ϭ ϩ11.7 and of Si2Cl6, i.e. ClϪ/SiCl3 exchange or SiCl2 insertion at the
Ϫ88).
most branched Si atom of an oligochlorosilane.[11Ϫ13] Disil-
Ϫ
In an additional experiment, 17b was mixed with SiCl4 anylphosphane 7 can also be a source of SiCl3 or SiCl2
at room temperature. NMR spectra revealed that 17b was (Equation 13). The predominant formation of bis(stannyl)-
consumed and that trichlorosilylphosphane 6 and Et3SnCl silanes 18a, b and the related germanium compound 16[7]
were formed, accompanied by some tBu(iPr)PH.
deserves further investigation; the formation of a second
SiϪSn or SiϪGe bond leading from 15/19 to intermediates
(R3M)2SiCl2 (M ϭ Ge, Sn) as precursors for 16/18 is appa-
rently affected by bulkier alkyl groups at Ge or Sn.
2. Cleavage of Trimethylgermylphosphane 14 by Si2Cl6
This reaction proceeds much more slowly than the reac-
tion of stannylphosphanes 17. Disilanylphosphane 7 is al-
ready being consumed by its reaction with Me3GeCl before
its formation by the cleavage of 14 can be quantitative. For-
mation of trichlorosilylphosphane 6 is slower than in the
stannylphosphane reactions, but finally it is also close to
quantitative. Contribution of a direct GeMe3/SiCl3 ex-
change (that was previously proposed[1]) to the formation
of Me3GeSiCl3 (15) and 6 cannot be ruled out. Even at an
early phase of the reaction, approximately equal amounts
of 15 and the branched bis(germyl)silane 16 are present in
solution. Experiments with pure 15 show that the presence
of Me3GeCl (its decomposition product !) inhibits the base-
catalysed decomposition.[7] The formation of deeply col-
oured precipitates on addition of Si2Cl6 to the phosphanes
at 0°C deserves further attention. Phosphanes are known
to catalyse the disproportionation,[14] they can coordinate
with Si2Cl6,[2] and they can activate disilanes for radical re-
actions with nBu3SnH.[15]
tBu(iPr)PSn(nBu)3 (17c): as in the cases described above,
consumption of the stannylphosphane 17c was very fast.
After 2 h, 7 and 19c[7] were the main products and after 12
d, 6 and nBu3SnSi(SiCl3)3 (20c) were the predominant spec-
ies. After an even longer reaction time (2 months), stannyl-
silane 19c was completely decomposed and in addition to
20c, the branched bis(stannyl)silane 18c could also be de-
tected by 29Si- and 119Sn-NMR.[7]
Discussion of the Reactions
1. Cleavage of Sn؊N and Sn؊P Bonds with Si2Cl6
Si2Cl6 cleaves stannylamines and stannylphosphanes
quantitatively within a few minutes under mild conditions
by transmetalation-like SnR3/Si2Cl5 exchange reactions
providing large amounts of the corresponding pentachloro-
disilanylamines and -phosphanes and the trialkylchloro-
stannanes (Equation 6, Equation 10a). From such reac-
tions, diisopropyl(pentachlorodisilanyl)amine 13 can be iso-
lated in pure state; pentachlorodisilanylphosphane 7, how-
3. Cleavage of Trimethylsilylphosphanes by Si2Cl6 and by
the Disilane Fraction
Silylphosphane 5 reacts more slowly than the stannyl-
phosphanes but faster than the germylphosphane 14 with
Si2Cl6 (silylphosphane 5 requires more than 2 h, germyl-
phosphane 14 about a day, and stannylphosphanes 17a-c
less than 20 min to be consumed completely). In the course
of the decomposition of disilanylphosphane 7 in the pres-
ence of Me3SiCl, the compound Me3SiSiCl3 was not de-
tected by 29Si NMR. This compound is known to dispro-
portionate in the presence of N- or P-nucleophiles.[16] Silyl-
phosphane 1 reacts with (MeCl2Si)2 by transsilylation yield-
Ϫ
ever, serves as a source of SiCl2 or SiCl3 moieties towards
Ϫ
the trialkylchlorostannanes, leading by ClϪ/SiCl3 ex-
change to trichlorosilylstannanes R3SnSiCl3 19 (Equation
10b).
ing tBu2PSi(Me)(Cl)Si(Me)Cl2 (9),
a stable disilanyl-
phosphane. Me2(Cl)SiSiCl2Me, however, suffered from
SiϪSi bond cleavage by 1 by attaching the chlorodimethylsi-
lyl group to phosphorus, i.e. “the more electrophilic sily-
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392
1383

W.-W. du Mont et al.
FULL PAPER
lene” MeSiCl (that allows better base-stabilisation) was lost Table 2. 29Si- and 31P-NMR data of disilanylphosphanes and
hexachlorodisilane adducts (δ in ppm, magitudes of J in Hz)
when tBu2PSiMe2Cl (8) was formed. Whether 8 is in fact
the kinetic product, i.e. whether the higher chlorinated silyl-
δ
31P δ 29Si(ϪP)
δ
29Si(ϪSiϪP)
phosphanes (like R2PSiCl2Me, R2PSiCl3) profit from extra
stabilisation of their PϪSi bonds compared with di- and tri-
alkylsilylphosphanes, deserves further investigation.
Խ(1J 29Si, 31P)Խ Խ(2J 29Si, 31P)Խ
tBu(iPr)PSiCl2SiCl3 7
tBu2PSiMeClSiMeCl2 9
tBu2PSiMe2SiMe3 10
8.3 13.2 (125.5)
8.6 12.1 (100.2)
Ϫ0.4 (24.0)
26.8 (21.9)
5.6 Ϫ25.2 (63.4) Ϫ14.9 (12.1)
[2]
iPr(iPr2N)PSiCl2SiCl3
42.7 8.9 (132.0)
1.1 (16.6)
Characterisation of New Germyl- and Stannylsilanes
by NMR Spectroscopy
[2]
iPr(iPr2N)(Cl)PSiCl3SiCl3 57.1 Ϫ69.6 (160.7) 9.8 (24.8)
[2]
tBu(Et2N)PSiCl2SiCl3
77.4 * (134.2)
* (not resolved)
* (24.3)
[2]
tBu(Et2N)(Cl)PSiCl3SiCl3 98.0 * (166.2)
Assignments were supported by measurements on pure
samples of 15 and 18a from independent syntheses using
the trichlorosilane/triethylamine route.[7]
In the series of compounds Me3EϪSiCl3 (E ϭ C, Si, Ge) Characterisation of New Disilanylphosphanes by
the 29Si resonances of the trichlorosilyl groups appear in NMR
the range 18 ± 1 ppm (Table 1). The 29Si nuclei of the SiCl3
Characteristic coupling patterns allowed safe assign-
ments of the 31P-NMR and 29Si-NMR spectra of transient
and persistent disilanylphosphanes (Table 2). Compared
with organic subsituents or hydrogen, chlorine atoms at sili-
groups of the branched compounds (Me3E)2Si(SiCl3)2 exhi-
bit rather similar resonances. Our previous assignments of
15 and 19a, however, were wrong[1]. Contributions from the
electric field gradient (from the p-electron imbalance at Si)
to σpara will overcome inductive effects on δ29Si (electrone-
gativity range C >> Ge > Si, Sn). In the case of compounds
Me3ESiCl3, however, the 13C resonances of the methyl
groups adjacent to C (δ ϭ ϩ24.4), Ge (δ ϭ Ϫ2.8), and Sn
(δ ϭ Ϫ8.1) follow the inductive trend. Again, the related
resonances of the branched compounds 16 and 18a appear
not far from those of 15 and 19a.
The central silicon atoms of the branched bis(germyl)-
and bis(stannyl)silanes (R3E)2Si(SiCl3)2 appear at rather
low frequency, the tin compounds 18a؊c at even lower fre-
quency than the germanium derivative 16. The 29Si shift of
the central atom of stannylsilane 20c appears at signifi-
cantly higher frequqency than that of the bis(stannyl)silane
18a. The 119Sn resonances of compounds 18a-c, 19a-c, and
20c appear in a narrow range from δ ϭ Ϫ45 to Ϫ70.[7]
con lead to increased couplings J and J (31P, 29Si).[10,18]
29Si-NMR shifts allow distinction between disilanylphos-
phanes (CN 4 at both Si) and phosphane-hexachlorodisi-
lane adducts (one Si with CN 5 [low frequency], one Si with
CN 4).[2]
1
2
Experimental and Calculated Structure of 2
Ab initio Calculations
Preliminary calculations on tBu2PSiCl3 were carried out
at the 3-21G*/SCF level, assuming Cs symmetry. Vi-
brational-frequency calculations revealed the presence of a
single imaginary frequency (74i cmϪ1), indicating that the
Cs structure represents a transition state connecting two
equivalent C1 minima. This structure can be regarded as
Table 1. 13C-, 29Si-, and 119Sn NMR shifts of trichlorosilyl derivati-
ves of C, Si, Ge, and Sn and reference compounds
Table 3. Selected theoretical bond lengths, bond angles, and dihe-
dral angles[a]
δ13C
δ29Si
δ119Sn
Parameter
3-21G*/
SCF
6-31G*/
SCF
6-31G*/
MP2
Me3CSiCl3
Me3SiSiCl3
ϩ24.37 [CH3], ϩ17.3
ϩ26.14 [CC3]
[12]
*
ϩ17.5 [SiCl3],
Ϫ7.2 [SiMe3]
ϩ17.8
PϪSi
225.0
204.8
189.6
189.8
155.0
108.3
118.2
108.8
111.9
105.5
104.6
111.4
106.4
108.5
110.7
227.0
206.1
191.8
192.0
154.0
108.4
118.2
109.0
112.0
106.3
105.4
111.8
106.2
108.2
111.2
Ϫ166.1
Ϫ162.4
Ϫ164.0
224.9
205.8
190.7
191.0
153.4
109.4
118.4
108.5
111.5
104.9
103.9
111.0
105.7
107.6
110.8
Ϫ164.8
Ϫ160.8
Ϫ162.6
Me3GeSiCl3 15
Me3SnSiCl3 19a
Et3SnSiCl3 19b
nBu3SnSiCl3 19c
Ϫ2.8
Ϫ8.6
*
SiϪCl(mean)
*
Ϫ70[7]
Ϫ59
P(1)ϪC(6)
*
*
P(1)ϪC(19)
*
Ϫ72
CϪC(mean)
[17]
Si(SiCl3)4
ϩ3.5 [SiCl3],
Ϫ80.0 [SiSi4]
ϩ17.2 [SiCl3],
Ϫ84.2 [Ge2SiSi2]
ϩ19.6 [SiCl3],
Ϫ106.4 [Sn2SiSi2]
ϩ21.0 [SiCl3],
CϪH(mean)
P(1)ϪSi(2)ϪCl(3)
P(1)ϪSi(2)ϪCl(4)
P(1)ϪSi(2)ϪCl(5)
Si(2)ϪP(1)ϪC(6)
Si(2)ϪP(1)ϪC(19)
C(6)ϪP(1)ϪC(19)
P(1)ϪC(6)ϪC(7)
P(1)ϪC(19)ϪC(20)
CϪCϪH(mean)
(Me3Ge)2Si(SiCl3)2 16 ϩ0.8
(Me3Sn)2Si(SiCl3)2 18a Ϫ8.0
(Et3Sn)2Si(SiCl3)2 18b ϩ3.8 [SnC]
Ϫ55.2
Ϫ45.0
Ϫ52.7
ϩ12.1 [SnCC] Ϫ108.2 [Sn2SiSi2]
(nBu3Sn)2Si(SiCl3)2 18c ϩ12.2 [SnC],
ϩ21.3 [SiCl3],
ϩ14.0 [SnC3C], Ϫ107.4 [Sn2SiSi2]
ϩ27.8 [SnC2C],
Cl(5)ϪSi(2)ϪP(1)ϪC(19) Ϫ164.7
C(20)ϪC(19)ϪP(1)ϪC(6) Ϫ161.3
C(8)ϪC(6)ϪP(1)ϪSi(2)
ϩ30.3 [SnCC]
nBu3SnSi(SiCl3)3 20c
ϩ12.0 [SiCl3],]
Ϫ88.3 [SnSiSi3
Ϫ46.8
Ϫ163.4
[a]
All distances (ra) in pm, all angles in°. See Figure 1 for atom
* Not available.
1384
numbering.
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392

Intermediates and Products of the Hexachlorodisilane Cleavage
FULL PAPER
Molecular Model
The large number of geometric parameters needed to de-
fine the structure of tBu2PSiCl3 in C1 symmetry made it
necessary to introduce a number of assumptions into the
GED refinement. At first the tBu and SiCl3 groups had no
symmetry and 9 parameters were used to define the C-
(methyl) and Cl positions in each groups. The number of
parameters was reduced only when (a) they were shown to
be defined effectively entirely by restraints and (b) they were
uncorrelated with the refining parameters. This procedure
allows deviations from local symmetry to be fully explored
rather than applying local symmetry without testing the val-
idity of the assumptions. In this case the refined structures
of tert-butyl and SiCl3 groups showed insignificant varia-
tion from C3 local symmetry. Therefore all further refine-
ments were carried out assuming this symmetry. Moreover,
within the levels of experimental uncertainty all six CϪC
bond lengths were also found to be indistinguishable and
therefore were constrained to be equal. Since hydrogen
Figure 1. The molecular structure of tBu2PSiCl3 as determined by
gas-phase electron diffraction
being derived from a fully staggered structure with tert-bu-
tyl and SiCl3 groups twisted in the same sense by around
15Ϫ20°, increasing the minimum distance between H atoms
on neighbouring tert-butyl groups from 195.4 pm (Cs) to
211.1 pm (C1) and reducing the molecular energy by 16.8 kJ
molϪ1 at the 3-21G*/SCF level of theory. All subsequent
calculations were carried out in C1 symmetry. Selected geo-
metric parameters are shown in Table 3 and the atom num-
bering is shown in Figure 1.
Table 4. Refined and calculated geometric parameters (distances in
pm, angles in°) from the GED study[a]
No. Parameter
GED(ra)
6؊31G*/MP2 (re)
a) independent parameters
p1 CϪH
p2 CϪC
p3 PϪC (mean)
p4 PϪC (diff)
p5 SiϪCl
113.6(4)
153.6(2)
192.5(10)
Ϫ0.1(5)
204.7(3)
225.0(12)
109.2(9)
109.2(4)
0.05(51)
112.9(2)
194.1(15)
6.0(6)
109.4[b]
153.4[b]
190.9
Ϫ0.3
205.8[b]
224.9
The computational facilities available to us restricted ge- p6 PϪSi
p7 CϪCϪH
110.0[b]
110.6
ometry optimisations to the 3-21G*/SCF, 6-31G*/SCF, and
6-31G*/MP2 levels, which should give an acceptably accu-
rate geometry, since tBu2PSiCl3 is not expected to contain
any significant multiple-bond character and there are few
electronegative atoms.
With the 6-31G* basis set, raising the level of theory
from SCF to MP2 produced little change (Table 3). Bond
lengths were generally within ca. 1 pm; for example, the p17 Butyl torsion (diff)
p8 PϪCϪC (mean)
p9 PϪCϪC (diff)
p10 PϪSiϪCl (mean)
p11 SiCl3 torsion
Ϫ0.07
113.1
p12 SiCl3 axial tilt
p13 SiCl3 equat. tilt
p14 Methyl twist
p15 Methyl tilt
p16 Butyl torsion (mean)
Ϫ0.5(19)
6.2(19)
0.0(fixed)
18.1(21)
288.3(10)
5.9(6)
p18 Butyl axial tilt (mean)
p19 Butyl axial tilt (diff)
Si(2)ϪCl(3) bond length is predicted to be 206.6 and 206.5
Ϫ1.2(14)
Ϫ0.7(4)
Ϫ4.0(7)
Ϫ31.5(8)
4.4(27)
pm at the SCF and MP2 levels, respectively, and similarly
the P(1)ϪC(19) bond length to be 192.0 and 191.0 pm at
the same levels. Only the PϪSi bond length changed appre-
ciably, shortening from 227.0 pm to 224.9 pm. Bond angles
generally varied by little more then 1°, although angles
around the central phosphorus atom differed by up to 2°.
p20 Butyl equat. tilt (mean)
p21 Butyl equat. tilt (diff)
p22 Butyl dip (mean)
p23 Butyl dip (diff)
p24 Butyl dihedral
105.4(7)
b) dependent parameters
p25 C(6)ϪPϪC(19)
110.6(13)
103.4(8)
102.8(6)
118.8(7)
108.9(16)
110.6(16)
104.9(10)
107.6(10)
114.9(10)
106.3(8)
105.3(8)
115.7(9)
111.0
104.9
103.9
118.4
108.5
111.5
105.7
108.5
116.2
107.6
106.4
116.8
At the SCF level the SiϪPϪC(6), SiϪPϪC(19), and p26 SiϪPϪC(6)
C(6)ϪPϪC(19) bond angles were predicted to be 106.3°,
p27 SiϪPϪC(19)
p28 PϪSiϪCl(3)
105.4°, and 111.8° and to 104.9°, 103.9°, and 111.0° at the
higher level.
Although the overall symmetry of tBu2PSiCl3 is pre-
dicted to be C1, the SiCl3 and two tert-butyl groups were
all found to exhibit approximate local C3 symmetry. At the
6-31G*/MP2 level, the calculated SiϪCl bond lengths lie in
the range 205.4 to 206.5 pm, with internal ClϪSiϪCl angles p37 Cl(5)ϪSiϪPϪC(19)
ranging from 105.3° to 106.5°. Similarly, for the tert-butyl
groups the CϪC bond lengths are calculated to range from
p29 PϪSiϪCl(4)
p30 PϪSiϪCl(5)
p31 PϪC(6)ϪC(7)
p32 PϪC(6)ϪC(8)
p33 PϪC(6)ϪC(9)
p34 PϪC(19)ϪC(20)
p35 PϪC(19)ϪC(21)
p36 PϪC(19)ϪC(22)
Ϫ163.4(16) Ϫ164.8
Ϫ156.5(23) Ϫ160.8
Ϫ161.8(21) Ϫ162.6
p38 C(20)ϪC(19)ϪPϪC(6)
p39 C(8)ϪC(6)ϪPϪSi
153.0 to 153.7 pm, with internal PϪCϪC angles from
106.9° to 109.9°.
[a]
[b]
See text for atom numbering and parameter definitions. Ϫ ab
initio values quoted are mean values.
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392
1385

W.-W. du Mont et al.
FULL PAPER
atom positions are poorly defined by the experimental data, Electron Diffraction Refinement
all CH3 groups were assumed to be identical and to have
local C3v symmetry.
The radial distribution curve for tBu2PSiCl3 (Figure 2)
consists of five distinct peaks at distances of ca. 110, 155,
205, 280, and 340 pm together with shoulders and weaker
peaks above 400 pm. The peaks at 110 and 155 pm corre-
spond to CϪH and CϪC scattering, respectively, and the
intense peak at 205 pm to scattering from SiϪCl bonds,
broadened by contributions associated with the PϪSi and
PϪC bonds. The peak at 280 pm consists mainly of scat-
tering from P···C non-bonded pairs, while the intense broad
peak at 340 pm is attributed to scattering from a number
of nonbonded atom pairs, with major contributions from
Cl···Cl and P···Cl pairs.
In total 24 independent geometrical parameters were
used to describe the structure of tBu2PSiCl3 in C1 symmetry
(Table 4). The bond length parameters are CϪH, p1, CϪC,
p2, the mean and difference of P(1)ϪC(6) and P(1)ϪC(19),
p3 and p4, SiϪCl, p5, and PϪSi, p6. The bond angles are
CϪCϪH, p7, the average of the PϪCϪC bond angles, p8,
the difference between the averages of the PϪCϪC bond
angles found in each of the tert-butyl groups, p9, and the
average PϪSiϪCl angle, p10.
The remaining parameters are best described with a coor-
dinate system in which the PϪSi bond defines the z axis,
with P(1) at the origin and Si(2) in the positive z direction.
The Cl atoms are arranged initially so that the z axis is an
axis of local 3-fold symmetry, with Cl(3) lying in the xz
plane in the positive x direction. By assuming a right-
handed set of coordinate axes the y axis is also defined.
The SiCl3 torsion angle is a rotation about the z axis, p11,
anticlockwise when viewed down the axis from Si to P. The
SiCl3 axial and equatorial tilts are anticlockwise rotations
at Si(2) about the y axis, p12, and the x axis, p13, respectively.
All subsequent torsion and tilt angle directions are antic-
lockwise when viewed down the rotation axis.
The tert-butyl groups were generated by initially placing
a methyl group carbon at the origin [i.e. same position as
P(1)] so that its three H atoms were arranged with local C3v
symmetry about the z axis and one H in the xz plane in the
positive x direction. The methyl torsion, p14, and tilt, p15, Figure 2. Experimental and difference (experimental Ϫ theoretical)
radial distribution curves, P(r)/r, for tBu2PSiCl3. Before Fourier
are rotations about the z axis and x axis, respectively. The
inversion the data were multiplied by sиexp(Ϫ0.00002s2)/(ZSi
fSi)(ZCl Ϫ fCl).
Ϫ
methyl group is then translated along the positive z axis by
the average CϪC bond length and the central carbon of the
tert-butyl group is placed at the origin. The correct PϪCϪC
The starting parameters for the ra structure refinement
bond angle is generated by rotating the methyl group about were taken from the theoretical geometry optimised at the
the y axis. The two remaining methyl groups are defined by 6-31G*/MP2 level. The program ASYM40[19] was used to
replicating the first methyl group and then rotating about convert the theoretical (6-31G*/SCF) Cartesian force field
the z axis by 120° and Ϫ120°, respectively. The tert-butyl to one described by symmetry coordinates and this was
torsion angle is a rotation of the group about the z axis, scaled to obtain root-mean-square amplitudes of vibration
while the axial and equatorial tilts are rotations about the (u). Scaling factors were chosen to be 0.90, 0.85, and 0.80
y axis and x axis, respectively. Positive axial tilts move tert- for bond stretches, angle bends and torsions, respectively.
butyl or SiCl3 groups towards the phosphorus lone pair and The presence of a large number of low-frequency vi-
equatorial tilts move these groups around the 3-fold belt. brational modes led to overestimated predictions of the per-
Parameters introduced here are the mean and difference of pendicular amplitudes of vibration (k). Since these values
the butyl torsion angles, p16 and p17, mean and difference were considered to be unreliable, corrections for shrinkage
of the butyl axial tilts, p18 and p19, and the mean and differ- effects were not included.
ence of the butyl equatorial tilts, p20 and p21. The tert-butyl
groups are then translated along the positive z axis by their of similar interatomic distances, the SARACEN method[20]
respective PϪC bond lengths. was employed in the GED refinements. Twelve geometrical
Since the molecular structure contained a large number
The three groups that were initially placed along the z and twelve vibrational amplitude restraints were used
axis are moved into their final positions by rotating both (Table 5). Values for geometrical restraints were taken to be
tert-butyl groups about the y axis by their respective “dip” those predicted at the 6-31G*/MP2 level, while amplitude
angles to allow a nonplanar PC2Si fragment, and then restraints were based upon the force-field calculations at the
about the x axis by an equal amount in opposite directions 6-31G*/SCF level. Uncertainties were 0.5 pm for bond
to give the correct angles about phosphorus. This intro- length differences, 1° for bond angles and 0.5° for bond
duces three final parameters: the mean and difference of the angle differences. Torsion angles were assigned uncertain-
butyl dip angles, p22 and p23, and the tert-butyl rotation ties of 2°, and 1° for differences. Uncertainties of 10% were
angle, p24.
assigned to vibrational amplitude restraints. The methyl tilt
1386
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392

Intermediates and Products of the Hexachlorodisilane Cleavage
Table 5. Flexible restraints
FULL PAPER
Number
Description
Parameter definition
Calculated value
(Uncertainty)
Refined value
Geometrical restraints:
1
PϪC bond length difference
Mean CϪCϪH angle
p4
Ϫ0.2(5)
Ϫ0.14
109.2
0.05
2
p7
109.5(10)
Ϫ0.07(50)
4.0(20)
3
PϪCϪC angle difference
Methyl torsion angle
p9
4
p11
6.2
5
SiCl3 torsion angle
p13
194.0(20)
288.0(10)
111.0(10)
1.03(50)
105.7(10)
Ϫ7.7(10)
1.1(5)
194.1
288.3
110.6
0.6
6
Butyl torsion angle difference
C(6)ϪPϪC(19) angle
p17
7
p25
8
[SiϪPϪC(6)] Ϫ [SiϪPϪC(19)]
PϪC(6)ϪC(7) angle
p26 Ϫ p27
p31
9
104.9
Ϫ7.3
1.0
10
11
12
[PϪC(6)ϪC(8)] Ϫ [PϪC(6)ϪC(9)]
[PϪC(19)ϪC(20)] Ϫ [PϪC(19)ϪC(21)]
PϪC(19)ϪC(22) angle
p32 Ϫ p33
p34 Ϫ p35
p36
116.9(10)
115.7
Vibrational Restraints:
13
14
15
16
17
18
19
20
21
22
23
24
P(1)ϪSi(2)
P(1)···C(6)
P(1)···Cl(3)
Si(2)···C(6)
Si(2)···C(7)
Si(2)···C(8)
Si(2)···C(9)
Cl(3)···C(6)
Cl(3)···C(7)
Cl(3)···C(8)
Cl(3)···C(9)
Cl(4)···C(8)
u1
5.9(6)
5.5
u2
5.8(6)
5.6
u7
10.1(10)
9.2(10)
12.5(10)
9.5(10)
15.3(10)
18.9(20)
20.5(20)
20.0(20)
27.8(30)
12.3(10)
8.9
9.3
u16
u18
u19
u20
u25
u27
u28
u29
u36
12.4
10.7
15.4
19.8
22.7
21.5
24.8
12.6
parameter, p12, which helped define the position of the hy-
drogen atoms, was fixed, since little information is con-
tained in the experimental data because of the poor scat-
tering ability of hydrogen.
The use of flexible restraints allowed the refinement of
41 independent parameters, comprising 23 geometrical pa-
rameters and 18 amplitudes of vibration.
In the final refinement the R factors RG and RD were
0.091 and 0.077 respectively. Experimental and calculated
radial distribution curves are shown in Figure 2, while Fig-
ure 3 displays the experimental and calculated molecular
scattering intensity curves. Final refined parameters are
listed in Table 4, selected interatomic distances and the cor-
responding amplitudes of vibration in Table 6 and the least-
squares correlation matrix in Table 7.
Discussion
The SARACEN method[20] has allowed the refinement
of all significant structural parameters for tBu2PSiCl3. The
highest possible molecular symmetry is Cs, but the perfectly
staggered arrangement of all tert-butyl and SiCl3 groups
leads to parallel 1:3 interactions between two methyl groups
or between a methyl group and a chlorine atom. The steric
interactions of Y(MX3)3 groups are typically minimised by
twisting the three groups by 15Ϫ20°, and this is what hap-
pens in this case, reducing the symmetry to C1.
Overall, good agreement between theory and experiment
was found; theoretical bond-length, bond-angle, and di-
hedral-angle predictions were in general found to be within
one or two standard deviations of the experimental value.
The introduction of electron correlation at the MP2 level is
seen to be important in obtaining an accurate prediction of
the PϪSi distance; the 6-31G*/MP2 value at 224.9 pm is
essentially identical to the experimental value of 225.0(12)
pm, whilst the 6-31G*/SCF uncorrelated computation over-
estimates the distance at 227.0 pm. The PϪSi bond length
is also effectively identical to those found by GED in other
compounds such as silylphosphane[21] [224.9(3) pm], silyl-
methylphosphane[21] [224.8(3) pm] and silyldimethylphos-
phane[21] [224.5(3) pm]. In H2PϪSiF3, PϪSi has been
found[22] to be 220.7(3) pm, consistent with the fact that
fluorine substituents at silicon result in a significant short-
ening of the PϪSi bond length relative to that of H-substi-
tuted compounds. Substitution of the less electronegative
chlorine for the silyl hydrogens does not seem to have a
significant effect.[2]
Figure 3. Experimental and final weighted difference (experimental
Ϫ theoretical) molecular scattering intensities for tBu2PSiCl3. Ca-
mera distances were (a) 259 mm and (b) 95 mm
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392
1387

W.-W. du Mont et al.
FULL PAPER
[a]
Table 6. Selected interatomic distances (ra) and root-mean-square
amplitudes of vibration for tBu2PSiCl3 from the GED study[a]
Table 7. Least-squares correlation matrix (ϫ100) for tBu2PSiCl3
p3 p5 p6 p11 p16 p18 u1
u5
u6 u24 k1
Number
Atom pair
ra/pm
u/pm
p6
74
1
PϪSi
225.0(12)
192.4(10)
192.5(10)
204.7(3)
6.9(5)
p8
Ϫ64
Ϫ53
2
PϪC(6)
5.6(1)
p12
p13
p20
p24
u2
Ϫ58
3
PϪC(19)
5.6(tied to u2)
4.2(5)
Ϫ50
Ϫ53
4
SiϪCl
5
CϪC
153.6(2)
4.2(6)
Ϫ54 Ϫ62
Ϫ53
6
CϪH
113.6(2)
6.9(5)
7
P···Cl(3)
370.1(20)
350.0(40)
353.5(33)
275.4(18)
280.1(16)
292.4(16)
277.9(13)
276.1(13)
293.8(18)
328.2(40)
329.9(38)
375.4(62)
463.6(29)
340.2(53)
327.1(72)
460.8(34)
384.0(36)
326.7(6)
8.9(9)
u4
60
8
P···Cl(4)
9.8(tied to u7)
9.8(tied to u7)
9.8(10)
u10
u49
k1
68
9
P···Cl(5)
Ϫ50
50
59
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
P···C(7)
Ϫ50
73 52 57
64 51 58 83
P···C(8)
10.1(tied to u10)
9.3(tied to u10)
10.0(tied to u10)
9.7(tied to u10)
9.4(tied to u10)
9.2(10)
9.0(tied to u16)
12.4(10)
10.7(9)
15.4(10)
k2
P···C(9)
P···C(20)
[a]
Only elements with absolute values Ն 50% are shown; k1 and k2
P···C(21)
are scale factors.
P···C(22)
Si···C(6)
Si···C(19)
Si···C(7)
sidual steric interactions between two tert-butyl groups
(after widening of the angle) are less than those between
tert-butyl and SiCl3 groups
Si···C(8)
Si···C(9)
Si···C(20)
Si···C(21)
Si···C(22)
Cl(3)···Cl(4)
Cl(3)···C(6)
Cl(3)···C(19)
Cl(3)···C(7)
Cl(3)···C(8)
Cl(3)···C(9)
Cl(3)···C(20)
Cl(3)···C(21)
Cl(3)···C(6)
Cl(4)···C(6)
Cl(4)···C(19)
Cl(4)···C(7)
Cl(4)···C(8)
Cl(4)···C(9)
Cl(4)···C(20)
Cl(4)···C(21)
Cl(4)···C(22)
Cl(5)···C(6)
Cl(5)···C(19)
Cl(5)···C(7)
Cl(5)···C(8)
Cl(5)···C(9)
Cl(5)···C(20)
Cl(5)···C(21)
Cl(5)···C(22)
C(7)···C(8)
13.4(tied to u20)
10.9(tied to u19)
13.5(tied to u18)
10.1(6)
Experimental Section
429.9(68)
388.0(70)
511.9(88)
551.2(62)
369.0(106)
375.1(124)
541.2(66)
369.1(68)
505.2(42)
415.3(64)
536.4(61)
625.7(38)
541.6(50)
340.0(93)
529.7(80)
507.0(59)
375.2(50)
506.3(25)
345.1(90)
528.1(50)
372.3(66)
528.1(62)
625.0(29)
553.3(29)
251.3(7)
19.8(19)
General: 1H-, 13C-, 29Si-, and 31P-NMR spectra: Bruker AC-200
spectrometer (200 MHz for 1H, 50.3 MHz for 13C, 81 MHz for 31P,
39.8 MHz for 29Si); solvent [D6]benzene; shifts are given relative to
TMS (1H, 13C, 29Si) and 85% H3PO4 (31P). Ϫ MS: Finnigan Mat
8430. Ϫ Elemental analyses: Carlo Erba analytical gas chromatog-
raph. Ϫ All experiments were carried out under deoxygenated dry
nitrogen as inert gas; solvents were dried according to standard
procedures.
15.5(tied to u28)
22.7(19)
21.5(13)
24.8(27)
23.8(tied to u28)
18.5(tied to u27)
19.3(tied to u28)
12.4(tied to u26)
15.3(tied to u24)
19.4(tied to u27)
12.6(9)
Ab initio Calculations: All calculations were performed on Dec Al-
pha 1000 4/200 and 8400 3/500 computers using the GAUSSIAN94
program.[23] Geometry optimisations were performed using the
standard 3-21G*[24Ϫ26] and 6-31G*[27Ϫ29] basis sets at the SCF level
and at the correlated MP2 level with the 6-31G* basis. Vibrational
frequencies were calculated from analytic second derivatives at the
3-21G*/SCF and 6-31G*/SCF levels to determine the nature of
stationary points and to provide estimates of amplitudes of vi-
bration (u) for use in the GED refinements.
15.8(tied to u27)
16.5(tied to u28)
18.9(tied to u27)
20.9(tied to u26)
13.3(tied to u28)
11.5(tied to u26)
15.0(tied to u28)
15.9(tied to u27)
21.6(tied to u28)
14.1(tied to u27)
12.2(tied to u35)
17.4(tied to u27)
4.5(9)
Electron Diffraction Measurements: Electron scattering intensities
were recorded on Kodak Electron Image plates with the Edinburgh
gas-diffraction apparatus operating at ca. 44.5 kV (electron wave-
length ca. 5.7 pm).[30] Nozzle-to-plate distances for the metal inlet
nozzle were 94.7 and 259.2 mm yielding data in the range s
20Ϫ360 nmϪ1; four and three plates were exposed at the short and
long distances, respectively. The sample and nozzle temperatures
were maintained at ca. 423 and 448 K respectively during the expo-
sure periods.
[a]
See Figure 1 for atom numbering; all other distances were inc-
luded in the refinement, but are not listed here.
All PϪSiϪCl angle predictions were within one standard
deviation of the experimentally determined values, and
angles about the central phosphorus within two standard
deviations. The angles about the central phosphorus
[CϪPϪC 110.6(3)° and SiϪPϪC 103.6(3)°] are consider-
ably larger than for the corresponding angles in silyldimeth-
ylphosphane[20] [CϪPϪC 100.8(12)° and SiϪPϪC
99.0(5)°], but can be attributed to the strong electron with-
drawing character of the SiCl3 group and to steric interac-
tions between the tert-butyl groups. Axial tilts are all close
to 6°, implying that SiCl3 and tert-butyl groups are all tilted
away from each other and towards the phosphorus lone
pair. The equatorial tilts of the tert-butyl groups are smaller
The scattering pattern of benzene was also recorded for the purpose
of calibration; this was analysed in exactly the same way as for
tBu2PSiCl3 so as to minimise systematic errors in the wavelengths
and camera distances. Nozzle-to-plate distances, weighting func-
tions used to set up the off diagonal weight matrix, correlation
parameters, final scale factors and electron wavelengths for the
measurements are collected in Table 8.
The electron scattering patterns were converted into digital form
with a computer-controlled Joyce Loebl MDM6 microdensitometer
and a scanning program described elsewhere.[31] The programs used
(ca. 2°) and towards one another. This implies that the re- for data reduction[31] and least squares refinement[32] have been
1388
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392

Intermediates and Products of the Hexachlorodisilane Cleavage
FULL PAPER
Table 8. Nozzle-to-plate distances [mm], weighting functions [nmϪ1], correlation parameters and electron wavelengths [pm] used in the
electron diffraction study
Nozzle-to-plate distance[a] ∆s
smin
sw1
sw2
smax
Correlation Parameter Scale Factor[b]
Electron Wavelength
94.67
259.17
4
2
80
20
100
40
210
140
248
164
0.391
0.411
0.432(19)
0.638(24)
5.653
5.708
[a]
[b]
Determined by reference to the scattering pattern of benzene. Ϫ Values in parentheses are estimated standard deviations.
Table 9. 31P-NMR data of tBu(iPr)PSiMe3 (5), tBu(iPr)PSi2Cl5 (7), and tBu(iPr)PSiCl3 (6)
time
5
7
6
other
signals
[ppm]
relative
intensity
relative
intensity
relative
intensity
relative
intensity
[ppm]
[ppm]
[ppm]
5 min
Ϫ17.4
Ϫ17.4
Ϫ17.5
100
100
51
7.9
8.0
9
6.4
6.5
10
64
9.6
4
5
10.8
30 min
78
9.7
6
5
10.1
10.9
1.8
12
14
5
80 min
13 h
8.1
8.3
8.3
100
73
6.5
6.6
6.7
82
10.9
2.2
100
100
11
3
9.9
43 h
46
2.3
1
10.0
10.0
10.0
10.0
3
61 h
85 h
10 d
8.3
8.3
8.4
27
20
7
6.7
6.7
6.7
100
100
100
3
4
3
1
described previously; the complex scattering factors were those
listed by Ross et al.[33]
SiSiCH3, 3 H). Ϫ 13C NMR: δ ϭ 34.2 [d, J(31P, 13C) ± 26.5 Hz,
PC], 33.1 [d, J(31P, 13C) ± 12.4 Hz, C-CH3], 7.1 (s, SiSi-C), 4.9 [d,
2
2J(31P, 13C) ± 8.3 Hz, PSi-C]. Ϫ 29Si NMR: δ ϭ 12.1 [d, 1J(31P,
29Si) ± 100.2 Hz, P-Si], 26.8 [d, 2J(31P, 29Si) ± 21.9 Hz, PSi-Si].
Reaction of Diisopropyl(trimethylsilyl)phosphane 3 with Si2Cl6:
Hexachlorodisilane (3.15 g, 11.7 mmol) was added dropwise to ice-
cooled 1 (2.2 g, 11.6 mmol). A strongly exothermic reaction oc-
curred. After 2 h stirring at room temperature, trimethylchlorosil-
Ϫ
31P NMR: δ ϭ 8.6 [s, 1J (31P, 29Si) ± 100.8 Hz, 2J (31P, 29Si)
± 22.0 Hz].
ane was distilled off from the red liquid; subsequent distillation at Synthesis of Di-tert-butyl(pentamethyldisilanyl)phosphane (10):
2.5 mbar provided 2.3 g (78%) 4 as a colourless liquid, b.p.
Methyllithium dissolved in ether (28.5 mL of a 1.6  solution,
45.6 mmol) was added slowly to a solution of 4.52 g (13.8 mmol)
19 in ether at Ϫ20°C. Stirring was continued for 2 h at room tem-
pature, and subsequently the mixture was heated under reflux for
30 min. After separation of the solution from lithium chloride, dis-
65 °C.[34]
Reaction of tert-Butyl(isopropyl)(trimethylsilyl)phosphane 5[9] with
Si2Cl6 Followed by Heteronuclear NMR: Hexachlorodisilane
(0.43 g, 1.6 mmol) is added dropwise to ice-cooled 5 (0.33 g,
1.6 mmol). After warming up to room temperature, the brown
liquid was transferred into an NMR tube. For results see Table 9.
Table 10. 31P-NMR data of tBu(iPr)PGeMe3 (14) and tBu(iPr)P-
SiCl3 (6)
NMR data of disilanylphosphane 7: δ31P 8.3 [s, 1J(31P, 29Si) ±
126.0 Hz; 2J(31P, 29Si) ± 22.2 Hz]; δ29Si 13.2 [d, d, 1J(31P, 29Si) ±
time
14
relative
6
relative other
relative
intensity
3
1
2
125.5 Hz; J(29Si, H) ± 8.7 Hz], Ϫ0.4 [d, J(31P, 29Si) ± 24.0 Hz].
[ppm] intensity [ppm] intensity signals
[ppm]
Reaction of Di-tert-butyl(trimethylsilyl)phosphane (1) with the Disi-
lane Fraction: A mixture of 6.21 g (28.43 mmol) of silylphosphane
1 and 12.52 g of a freshly distilled WACKER AG. technical disilane
fraction sample [Me2(Cl)SiSiCl2Me/(MeCl2Si)2 ratio about 1:1;
corresponding to about 28 mmol of each of the two disilanes] was
heated to 100°C. After 16 h, a 31P-NMR spectrum confirms com-
plete consumption of silylphosphane 1. Distillation at 1 mbar al-
lowed separation of two fractions: b.p. 55Ϫ60°C, 1.6 g (24%)
tBu2PSiMe2Cl (8),[10] and b.p. 110Ϫ114°C, 4.4 g (48%) tBu2PSi(-
Me(Cl)Si(Me)Cl2 (9).
5 min
3 h
Ϫ1.5 100
Ϫ1.6 100
Ϫ1.6 23
6.0
6.1
6.2
6.5
10
1.4
7
3.0
7
7.5 (*)
11.8
1.4
12
4
53
13
14
9
3.1
7.5 (*)
11.7
1.4
4
15 h
6 d
100
100
12
22
3
3.5
11.4
11.9
1.5
8: 1H NMR: δ ϭ 1.28 (d, J ϭ ± 11.4 Hz, 18 H), 0.6 (d, J ϭ ±
3
8
1
3.5 Hz); 31P NMR: δ ϭ 0.8 (s, J (31P, 29Si ϭ ± 67.5 Hz).[10]
3.9
6
NMR data of 9: 1H NMR: δ ϭ 1.26 [d, 3J(31P, 1H) ± 12 Hz,
1
PCCH3, 18 H], 0.87 [d, 3J(31P, H) ± 3.7 H, PSiCH3,, 3 H], 0.78 (s, (*) Probably tBu(iPr)PSiCl2SiCl3 (7).
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392
1389

W.-W. du Mont et al.
FULL PAPER
Table 11. 29Si-NMR: of Me3GeSiCl3 15, (Me3Ge)2Si(SiCl3)2 16, SiCl4
time
15
[ppm]
relative
intensity
16
[ppm]
relative
intensity
SiCl4
[ppm]
relative
intensity
other signals
[ppm]
relative
intensity
3 h
17.5
17.7
11
19
16.9
11
Ϫ19.0
Ϫ19.0
100
100
10.1
43 (*)
15 h
16.8Ϫ84.9
307
3 (d,d):
38 (ϩ),
10.1 (ϩ), 10.4, 13.1, 13.3 27, 23, 25
7 d
17.9
89
71
16.9
16.9
76
74
Ϫ19.0
Ϫ19.0
89
64
3: 11.8 (d,d)
3: 11.8
(d,d)
100 (ϩϩ)
(broad)
17.8
26 d
100 (ϩϩ)
(broad)
(*) Reasonable assignment: Me3GeSi(SiCl3)3; (ϩ) intensity of this peak of the d,d pattern enhanced due to overlap with the singlet at
δ ϭ 10.1 of (*); (ϩϩ) normal intensity d,d pattern.
Table 12. 31P-NMR data of tBu(iPr)PSiCl3 (6), and tBu(iPr)PSi2Cl5 (7)
time
7
relative
intensity
6
relative
intensity
other signals
[ppm]
relative
intensity
[ppm]
[ppm]
20 min
15 h
8.1
6
6.3
6.2
100
100
1.2[a] 4.1
1.1[a]
35
2
[a]
tBu(iPr)PH.
Table 13. 29Si-NMR data of (Me3Sn)2Si(SiCl3)2 (19a), and tBu(iPr)PSiCl3 (3)
time
19a [ppm]
relative
intensity
SiCl4 [ppm]
relative
intensity
other signals
[ppm]
relative
intensity
60 min
15 h
19.4
58
Ϫ19.2
Ϫ19.2
100
100
11.2 (*)
36
3: 11.8 (d,d)
3: 11.8 (d,d)
11.3 (*),
38, 38, 36, 25
19.4
Ϫ106.6
64
11
35, 35, 33, 33
21,
3
Ϫ88 (*)
(*) Me3SnSi(SiCl3)3 (20a).
tillation yielded 2.5 g (66%) of disilanylphosphane 10 as a colour-
276 (14) [Mϩ Ϫ CH3, Ϫ C3H6]; 176 (12) [Mϩ Ϫ CH3, Ϫ C3H6, Ϫ
less liquid. Ϫ 1H NMR: δ ϭ 1.29 [d, J (31P, 1H) ± 11.1 Hz, SiCl2]; 133 (18) [SiCl3ϩ]; 98 (4) [SiCl2ϩ], 44 (100) [C3H8ϩ]. Ϫ IR ν˜
3
3
1
PCCH3, 18 H], 0.38 [d, J(31P, H) ± 4.1 Hz, PSiCH3, 6 H], 0.18 (liquid between KBr) [cmϪ1]: 2970 (s), 2939 (s), 2880 (s), 2850 (m)
2
[s, SiSiCH3), 9H]. 13C NMR: δ ϭ 33.5 [d, J(31P, 13C) ± 12.1 Hz,
1460 (m), 1385 (m), 1370 (s), 1305 (w), 1205 (w), 1175 (m), 1155
(w), 1000 (w), 885 (m), 840 (m), 815 (w), 800 (w), 645 (m), 570 (vs,
1
2
CCH3], 32.8 [d, J(31P, 13C) ± 24.9 Hz, PC], 4.9 [d, J(31P, 13C) ±
11 Hz, PSi-C]. Ϫ0.6 [d, 3J(31P, 13C) ± 1.2 Hz, SiSi-C]. Ϫ 29Si NMR: broad), 510 (s), 485 (m), 445 (s), 375 (m). Ϫ C6H14Cl5NSi2 (336.62)
δ ϭ Ϫ25.2 [d, 1J(31P, 29Si) ± 63.4 Hz, P-Si], Ϫ14.9 [d, J(31P, 29Si)
found (calcd.) C 21.56 (21.41), H 4.14 (4.19), Cl 53.03 (52.66).
2
± 12.1 Hz, PSi-Si]. Ϫ 31P NMR: δ ϭ 5.6 [s, 1 J (31P, 29Si) ± 63.6 Hz,
Reaction of Diisopropyl(trimethylstannyl)amine 12 with Si2Cl6:
Hexachlorodisilane (2.7 g, 10 mmol) was added slowly at 0°C to
2.45 g (9.3 mmol) diisopropyl(trimethylstannyl)amine. After warm-
ing up to room temperature, complete consumption of the stannyl-
amine was indicated by NMR. After removal of Me3SnCl by crys-
tallisation from pentane, distillation at 0.8 mbar gave 1.3 g (40%)
diisopropyl(pentachlorodisilanyl)amine (13) as a moisture-sensitive
colourless liquid,[35] b.p. 65 °C (see above).
J (31P, 29Si) ± 12.5 Hz]. Ϫ MS (EI, 70 eV) m/z(%): 276 (3, Mϩ),
2
131 (22, Me5Si2ϩ), 73 (60, Me3Siϩ), 58 (100, C4H10ϩ). Ϫ IR (liquid
between KBr) [cmϪ1] 2970 (s), 2950 (vs), 2900 (s), 2870 (s) 1475
(m), 1395 (m), 1370 (m), 1250 (s), 1200 (w), 1165 (m), 1045 (m),
1005 (m), 945 (m), 920 (m), 895 (w), 835 (s), 810 (s), 780 (m), 740
(vw), 690 (vw), 675 (w), 620 (w), 540 (vw), 495 (w), 440 (vw). Ϫ
C13H33PSi2 (276.55), found (calc.) C 56.19 (56.46), H 11.83 (12.02),
P 11.13 (11.20).
Reaction of Diisopropyl(trimethylsilyl)amine (11) with Si2Cl6:
Hexachlorodisilane (7.35 g, 27.3 mmol) was added to 4.5 g
(26 mmol) diisopropyl(trimethylsilyl)amine. After heating the mix-
ture for 6 h at 60°C complete consumption of the silylamine was
indicated by NMR. Distillation at 0.5 mbar gave 6.24 g (72%) di-
Table 14. 119Sn-NMR data of Me3SnSiCl3 (18a) and (Me3Sn)2Si-
(SiCl3)2 (19a)
time
18a [ppm]
relative
intensity
19a [ppm]
relative
intensity
isopropyl(pentachlorodisilanyl)amine as
a
moisture-sensitive
colourless liquid, b.p. 61°C.[35] Ϫ 1H NMR: δ ϭ 1.02 [d, 3J(1 H, 1H)
3
± 6.7 Hz, CH3,, 12 H], 3.26 [sept, J(1 H, 1H) ± 6.7 Hz, CH, 2 H].
15 min
15 h
Ϫ66.3
Ϫ66.5
26
4
Ϫ53.5
Ϫ53.5
100
100
Ϫ
13C NMR: δ ϭ 23.55 (s, CH3), 47.25 (s, CH). 29Si NMR: δ ϭ
Ϫ1.8 (s), Ϫ26 (s). Ϫ MS (EI, 70 eV): m/z (%) 318 (14) [Mϩ Ϫ CH3];
1390
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392

Intermediates and Products of the Hexachlorodisilane Cleavage
FULL PAPER
Table 15. 31P-NMR data of tBu(iPr)PSiCl3 (6), and tBu(iPr)PSi2Cl5 (7)
time
7
6
other
signals
[ppm]
relative
intensity
relative
intensity
relative
intensity
[ppm]
[ppm]
5 min
8.5
8.4
8.4
100
63
6.7
6.6
6.6
19
1.1
4.3
1.2
4.3
1.2
4.2
1.1
1.2
27
2
30 min
70 min
100
100
26
2
16
8
1
5 h
3 d
6.5
6.6
100
100
6
1
Table 16. 119Sn-NMR data of Et3SnSiCl3 (18b), (Et3Sn)2Si(SiCl3)2
(19b), and Et3SnCl
to room temperature, the yellow liquid was transferred into an
NMR tube. For NMR results see Tables 12, 13, and 14.
Reaction of tert-Butyl(isopropyl)(triethylstannyl)phosphane (17b)
with Si2Cl6 Followed by Heteronuclear NMR: Hexachlorodisilane
(0.24 g, 0.89 mmol) was added dropwise to ice-cooled 17b (0.29 g,
0.86 mmol). The liquid turned red, then yellow and, at room tem-
perature, green. The green liquid was transferred into an NMR
tube. For NMR results see Table 15.
29Si-NMR after 5 h: [SiCl4: relative intensity ϭ 100%] δ29Si ϭ 21.0/
Ϫ108.1 (50%/9%) [(Et3Sn)2Si(SiCl3)2 (19b)], 11.8 (d,d) (4 lines, 20%
each) [tBu(iPr)PSiCl3 (6)], 12.1 (35%)[20b]; in a separate experiment,
a spectrum with about 12 h acquisition time allowed the resolution
of two signals for 20b: δ29Si ϭ 11.7 (SiCl3), Ϫ88.0 [Si(SiCl3)3]. For
NMR results see Table 16.
time
18b
relative 19b
relative Et3SnCl relative
[ppm] intensity [ppm] intensity [ppm]
intensity
25 min Ϫ58.4 100
60 min Ϫ58.5 87
330 min Ϫ58.7 12
24 h
3 d
9 d
Ϫ44.6 17
Ϫ44.7 100
Ϫ44.8 100
Ϫ44.8 100
Ϫ44.8 100
Ϫ44.8 100
150 (*)
145 (*)
136 (*)
135 (*)
135 (*)
135 (*)
6
7
7
7
8
8
Ϫ58.6
Ϫ58.6
Ϫ58.5
9
7
8
(*) Broad.
Reaction of tert-Butyl(isopropyl)(trimethylgermyl)phosphane 14 with
Si2Cl6 Followed by Heteronuclear NMR: Hexachlorodisilane
(0.35 g, 1.3 mmol) was added dropwise to ice-cooled 14 (0.33 g,
1.3 mmol). A violet precipitate could be observed. After warming
up to room temperature, the brown liquid was transferred into an
NMR tube. For NMR results see Tables 10 and 11.
Reaction of tert-Butyl(isopropyl)(tri-n-butylstannyl)phosphane (17c)
with Si2Cl6 Followed by Heteronuclear NMR: Hexachlorodisilane
(0.27 g, 1.0 mmol) was added dropwise to ice-cooled 17c (0.4 g,
0.95 mmol). After warming up to room temperature, the greenish-
yellow solution was transferred into an NMR tube (see the follow-
ing tables).
Reaction of tert-butyl(isopropyl)(trimethylstannyl)phosphane 17a
with Si2Cl6 followed by heteronuclear NMR: Hexachlorodisilane
(0.4 g, 1.5 mmol) was added dropwise to ice-cooled 17a (0.41 g,
Adding excess of the disilane (0.4 g, 1.5 mmol) at Ϫ20 °C to 17c
(0.27 g, 1.0 mmol) led qualitatively to the same products, but in a
1.4 mmol). A red precipitate could be observed. After warming up quite different distribution (faster consumption of 7 and of 18c;
Table 17. 31P-NMR data of tBu(iPr)PSiCl3 (6) and tBu(iPr)PSiCl2SiCl3 (7)
time
7
6
other
signals
[ppm]
relative
intensity
relative
intensity
relative
intensity
[ppm]
[ppm]
2 h
7.3
7.3
7.0
100
100
3
5.9
5.9
5.9
9
1.4
3
3
3
.
10.3
1.4
2.5 h
12 d
21
100
10.3
Table 18. 119Sn-NMR data of nBu3SnSiCl3 (18c), (nBu3Sn)2Si(SiCl3)2 (19c) and nBu3SnSi(SiCl3)3 (20c)
time
18c [ppm]
relative intensity
20c [ppm]
relative intensity
other signals [ppm]
45.5
relative intensity
57
2.5 h
12 d
75 d
Ϫ72.3
Ϫ72.7
100
21
Ϫ48.1
Ϫ47.8
100
100
Ϫ53.8 (*)
142.0 (**)
64
59
(*) 19c; (**) nBu3SnCl.
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392
1391

W.-W. du Mont et al.
FULL PAPER
about 5:4 intensity ratio of the 119Sn-NMR signals of 19c and 20c
[17]
K. G. Sharp, P. A. Sutor, E. A. Williams, J. D. Cargioli, T. C.
Farrar, K. Ishibitsu, J. Amer. Chem. Soc. 1976, 98, 1977.
G. Fritz, H. Schäfer, Z. Anorg. Allg. Chem. 1974, 407, 295.
ASYM 40, version 3.0, update to program ASYM 20: L. Hed-
berg, I. M. Mills, J. Mol. Spectrosc. 1993, 160, 117.
A. J. Blake, P. T. Brain, H. McNab, J. Miller, C. A. Morrison,
S. Parsons, D. W. H. Rankin, H. E. Robertson, B. A. Smart, J.
Phys. Chem., 1996, 100, 12280; P. T. Brain, C. A. Morrison, S.
Parsons, D. W. H. Rankin, J. Chem. Soc., Dalton Trans., 1996,
4589.
after 30 days). For NMR results see Tables 17 and 18.
[18]
[19]
[20]
Acknowledgments
We thank the Deutsche Forschungsgemeinschaft, Bonn-Bad Go-
desberg, and the Fonds der Chemischen Industrie, Frankfurt, for
financial support, the WACKER AG, Burghausen, for a sample of
the disilane fraction, and Professor Uwe Klingebiel (Universität
Göttingen) for a 29Si-NMR measurement. We thank the U. K.
Engineering and Physical Sciences Research Council for a research
studentship (P. M. P.), for support of the Edinburgh gas electron
diffraction Service (grant GR/K 44411) and ab initio calculations
(grant GR/K 04194) and for provision of microdensitometer facili-
ties at the Daresbury Laboratory.
[21]
C. Glidewell, P.M. Pinder, A.G. Robiette, G.M. Sheldrick, J.
Chem. Soc., Dalton Trans., 1972, 1402.
[22]
[23]
R. Demuth, H. Oberhammer, Z. Naturforsch. A, 1973, 28, 1862.
Gaussian 94 (Revision C.2), M. J. Frisch, G. W. Trucks, H. B.
Schlegel, P. M. W. Gill, B. G. Johnson, M. A. Robb, J. R. Chees-
man, T. A. Keith, G. A. Petersson, J. A. Montgomery, K.
Raghavachari, M. A. Al-Laham, V. G. Zakrzewski, J. V. Ortiz,
J. B. Foresman, J. Cioslowski, B. B. Stefanov, A. Nanayakkara,
M. Challacombe, C. Y. Peng, P. Y. Ayala, W. Chen, M. W.
Wong, J. L. Andres, E. S. Replogle, R. Gomperts, R. L. Martin,
D. J. Fox, J. S. Binkley, D. J. Defrees, J. Baker, J. P. Stewart, M.
Head-Gordon, C. Gonzalez, J. A. Pople, Gaussian Inc., Pitts-
burgh, PA, 1995.
[1]
R. Martens, W.-W. du Mont, Chem. Ber. 1993, 126, 1115Ϫ1117.
[2]
L.-P. Müller, W.-W. du Mont, J. Jeske, P. G. Jones, Chem. Ber.
[24]
[25]
[26]
[27]
J. S. Binkley, J. A. Pople, W. J. Hehre, J. Am. Chem Soc., 1980,
102, 939.
M. S. Gordon, J. S. Binkley, J. A. Pople, W. J. Pietro, W. J.
Hehre, J. Am. Chem. Soc., 1982, 104, 2797.
W. J. Pietro, M. M. Francl, W. J. Hehre, D. J. Defrees, J. A.
Pople, J. S. Binkley, J. Am. Chem. Soc., 1982, 104, 5039.
W. J. Hehre, R. Ditchfield, J. A. Pople, J. Chem. Phys., 1972,
56, 2257.
1995, 128, 615; L.-P. Müller, A. Zanin, W.-W. du Mont, J. Jeske,
P. G. Jones, Chem. Ber. 1997, 130, 377.
[3]
R. A. Benkeser, Acc. Chem. Res. 1971, 4, 94.
[4]
R. A. Benkeser, K. M. Foley, J. B. Grutzner, W. E. Smith, J.
Am. Chem. Soc. 1970, 92, 697.
[5]
W.-W. du Mont, L.-P. Müller, L. Müller, S. Vollbrecht, A.
Zanin, J. Organomet. Chem. 1996, 521, 417.
[6]
W.-W. du Mont, L.-P. Müller, L. Müller, S. Vollbrecht, A.
[28]
[29]
[30]
P. C. Hariharan, J. A. Pople, Theor. Chim. Acta, 1973, 28, 213.
M. S. Gordon, Chem. Phys. Lett. 1980, 76, 163.
C. M. Huntley, G. S. Laurenson, D. W. H. Rankin, J. Chem.
Soc., Dalton Trans. 1980, 954.
Zanin, XI. Int. Symp. on Organosilicon Chemistry, Montpellier
1996, Abstract.
[7]
L. Müller, W.-W. du Mont, F. Ruthe, P. G. Jones, H. C.
Marsmann, J. Organomet. Chem., in print
[31]
[32]
[33]
[8]
S. Cradock, J. Koprowski, D. W. H. Rankin, J. Mol. Struct.
1981, 77, 113.
R. Martens, W.-W. du Mont, Chem. Ber. 1992, 125, 657.
[9]
L. Müller, W.-W. du Mont, F. Ruthe, P. G. Jones, to be pub-
A. S. F. Boyd, G. S. Laurenson, D. W. H. Rankin, J. Mol.
Struct. 1981, 71, 217.
lished; L. Müller, Ph.D. thesis, TU Braunschweig, 1998.
[10]
H. Schumann, W.-W. du Mont, Z. Anorg. Allg. Chem. 1975,
A. W. Ross, M. Fink, R. Hildebrandt, in: International Tables
for Crystallography, Vol. C, (Ed.: A. J. C. Wilson), Kluwer Aca-
demic Publishers, Dordrecht, 1992, p. 245.
418, 259.
[11]
G. D. Cooper, A. R. Gilbert, J. Am. Chem. Soc. 1960, 82, 5042.
[12]
G. Urry, Acc. Chem. Res. 1970, 3, 306.
[34]
[35]
[13]
R. Martens, W.-W. du Mont, L. Lange, Z. Naturforsch. B 1991,
46, 1609.
U. Herzog, R. Richter, E. Brendler, G. Roewer, J. Organomet.
Chem. 1996, 507, 221.
[14]
H. Schuh, T. Schlosser, P. Bissinger, H. Schmidbaur, Z. Anorg.
Allg. Chem. 1993, 619, 1347.
H. J. Emeleus, M. Tufail, J. Inorg. Nucl. Chem. 1967, 29, 2081.
[15]
U. Herzog, G. Roewer, U. Pätzold, J. Organomet. Chem. 1996,
494, 143.
Received December 28, 1998
[16]
U. Herzog, G. Roewer, J. Organomet. Chem. 1997, 544, 217.
[I98451]
1392
Eur. J. Inorg. Chem. 1999, 1381Ϫ1392