Reaction of decamethylsilicocene with group 13 element halides: Insertions, rearrangements, and eliminations

Article abstract of DOI:10.1021/om9906114

In the reaction of decamethylsilicocene (1; (Me₅C₅)₂Si) with halides or organohalides of trivalent boron, aluminum, gallium, and indium, quite different and sometimes very complex pathways are observed which include adduct formation, 1,2-halide or -alkyl shifts, 1,2-dyotropic rearrangements, and reductive elimination and oxidative addition reactions. Cp*BCl₂ (Cp* = pentamethylcyclopentadienyl), BCl₃, and BBr₃ reacted with 1 to form the pentacarba-nido-hexaboronium salts [Cp*SiCl₂BCp*]⁺[Cp*BCl₃]- (3a) and [Cp*SiX₂BCp*]⁺-[BX₄]^- (3b, X = Cl; 4, X = Br). A second product (5b) of composition Cp*₃Si₂Br₄B with an arachno-cluster framework was isolated from the reaction with BBr₃. With AlCl₃ and AlBr₃ metathesis reactions gave the ionic compounds [Cp*₂Al]^-[AlX₄]^- (6a,b, X = Cl, Br), respectively. The formation of 6a,b is the result of a Lewis-base-catalyzed dismutation of Cp*AlX₂ in which 1 is the base, as proved by separate experiments. The compound Cp*Al-(Me)Cl (7) was formed in the reaction of 1 with Me₂AlCl. Silicocene 1 functioned as a dehalogenating agent in its reactions with GaCl₃, GaBr₃, Cp*GaBr₂, InCl₃, and InBr₃, giving the corresponding monovalent, metastable ( GaCl , GaBr ) or stable (Cp*Ga, InCl, InBr) species. All new compounds were characterized by NMR (¹H, ¹³C, ¹¹B, ²⁹Si) and mass spectrometry. The solid-state structures of 3a and 5b were determined by X-ray diffraction analysis.

Full text of DOI:10.1021/om9906114

Organometallics 1999, 18, 5531-5538
5531
Rea ction of Deca m eth ylsilicocen e w ith Gr ou p 13
Elem en t Ha lid es: In ser tion s, Rea r r a n gem en ts, a n d
Elim in a tion s
Udo Holtmann, Peter J utzi,* Thorsten Ku¨hler, Beate Neumann, and
Hans-Georg Stammler
Faculty of Chemistry, University of Bielefeld, D-33615 Bielefeld, Germany
Received August 2, 1999
In the reaction of decamethylsilicocene (1; (Me5C5)2Si) with halides or organohalides of
trivalent boron, aluminum, gallium, and indium, quite different and sometimes very complex
pathways are observed which include adduct formation, 1,2-halide or -alkyl shifts, 1,2-
dyotropic rearrangements, and reductive elimination and oxidative addition reactions.
Cp*BCl2 (Cp* ) pentamethylcyclopentadienyl), BCl3, and BBr3 reacted with 1 to form the
pentacarba-nido-hexaboronium salts [Cp*SiCl2BCp*]+[Cp*BCl3]- (3a ) and [Cp*SiX2BCp*]+-
[BX4]- (3b, X ) Cl; 4, X ) Br). A second product (5b) of composition Cp*3Si2Br4B with an
arachno-cluster framework was isolated from the reaction with BBr3. With AlCl3 and AlBr3
metathesis reactions gave the ionic compounds [Cp*2Al]+[AlX4]- (6a ,b, X ) Cl, Br),
respectively. The formation of 6a ,b is the result of a Lewis-base-catalyzed dismutation of
Cp*AlX2 in which 1 is the base, as proved by separate experiments. The compound Cp*Al-
(Me)Cl (7) was formed in the reaction of 1 with Me2AlCl. Silicocene 1 functioned as a
dehalogenating agent in its reactions with GaCl3, GaBr3, Cp*GaBr2, InCl3, and InBr3, giving
the corresponding monovalent, metastable (“GaCl”, “GaBr”) or stable (Cp*Ga, InCl, InBr)
species. All new compounds were characterized by NMR (1H, 13C, 11B, 29Si) and mass
spectrometry. The solid-state structures of 3a and 5b were determined by X-ray diffraction
analysis.
In tr od u ction
observed. Only very few reactions of carbene analogue
species of silicon, germanium, and tin with compounds
of group 13 elements have been described. In 1996 Denk
and Metzler reported on the reaction of 2 with tris-
(pentafluorophenyl)borane in which a Lewis-acid-Lewis-
base complex was formed which rearranged by insertion
of the silylene into a B-C bond.6 Very recently, Belzner
et al. reported the insertion of a diarylsilylene into an
Al-C bond of trimethylaluminum.7 In 1992 No¨th and
co-workers described the reaction of the transient dim-
ethylgermylene with chloro- and hydridoboranes,8 ear-
lier on Zuckerman and Harrison discussed the reaction
of stannocene with boron trifluoride,9 and recently the
In 1986 we synthesized decamethylsilicocene (1), the
first stable monomeric compound with silicon in the
formal oxidation state +2.1 In this species the central
silicon atom is π-bound to two pentamethylcyclopenta-
dienyl (Cp*) ligands. The bonding is best described as
an interaction of the Cp* π-orbitals of appropriate
symmetry with empty p orbitals on silicon. As a
consequence, the lone electron pair on silicon has nearly
exclusive s-character. The reactivity of 1 can best be
described as that of a nucleophilic silylene.2 In conjunc-
tion with the synthesis of 1 and with the preparation
of 1,3-di-tert-butyl-1,3,2-diazasilol-2-ylidene (2), the first
stable two-coordinate silylene, in 1994 by Denk et al.3
there has been increased interest in the reactions of
nucleophilic silylenes.4
(4) See for example: (a) Denk, M.; Hayashi, R. K.; West, R. J . Chem.
Soc., Chem. Commun. 1994, 33. (b) Gehrhus, B.; Hitchcock, P. B.;
Lappert, M. F.; Maciejewski, H. Organometallics 1998, 17, 5599. (c)
Denk, M. K.; Hatano, K.; Lough, A. J . Eur. J . Inorg. Chem. 1998, 1067.
(d) Boesfeld, W. M.; Gehrhus, B.; Hitchcock, P. B.; Lappert, M. F.;
Schleyer, P. v. R. J . Chem. Soc., Chem. Commun. 1999, 755. (e) Scha¨fer,
A.; Saak, W.; Weidenbruch, M.; Marsmann, H.; Henkel, G. Chem. Ber./
Recl. 1997, 130, 1733. (f) Gehrhus, B.; Hitchcock, P. B.; Lappert, M.
F.; Heinicke, J .; Boese, R.; Bla¨ser, D. J . Organomet. Chem. 1996, 521,
211. (g) Denk, M.; West, R. Pure Appl. Chem. 1996, 68(4), 785. (h)
Denk, M.; Hayashi, R. K.; West, R. J . Am. Chem. Soc. 1994, 116, 10813.
(i) Drost, C.; Gehrhus, B.; Hitchcock, P. B.; Lappert, M. F. J . Chem.
Soc., Chem. Commun. 1997, 1845.
After earlier preliminary investigations5 we recently
have concentrated on the reactivity of 1 toward com-
pounds of group 13 elements, where a variety of reduc-
tion, insertion, and rearrangement processes has been
(1) (a) J utzi, P.; Kanne, D.; Kru¨ger, C. Angew. Chem. 1986, 98, 163.
(b) J utzi, P.; Holtmann, U.; Kanne, D.; Kru¨ger, C.; Blom, R.; Gleiter,
R.; Hyla-Kryspin, I. Chem. Ber. 1989, 122, 1629.
(2) See for example: J utzi, P.; Eikenberg, D.; Bunte, E.-A.; Mo¨hrke,
A.; Neumann, B.; Stammler, H.-G. Organometallics 1996, 15, 1930 and
references therein.
(3) (a) Denk, M.; Lennon, R.; Hayashi, R.; West, R.; Belyakov, A.
V.; Verne, H. P.; Haaland, A.; Wagner, M.; Metzler, N. J . Am. Chem.
Soc. 1994, 116, 2691. (b) Haaf, M.; Scmiedl, A.; Schmedake, T. A.;
Powell, D. R.; Millevolte, A. J .; Denk, M.; West, R. J . Am. Chem. Soc.
1998, 120, 12714.
(5) J utzi, P. In Frontiers of Organosilicon Chemistry; Bassindale,
A. R., Gaspar, P. P., Eds.; Royal Society of Chemistry: Cambridge,
U.K., 1991.
(6) Denk, M.; Metzler, N. Chem. Commun. 1996, 2657-2658.
(7) Belzner, J .; Ronneberger, V.; Scha¨r, D.; Bro¨nnecke, C.; Herbst-
Irmer, R.; Noltemeyer, M. J . Organomet. Chem. 1999, 577, 330-336.
(8) Mayer, E. P.; No¨th, H.; Rattay, W.; Wietelmann, U. Chem. Ber.
1992, 125, 401-405.
10.1021/om9906114 CCC: $18.00 © 1999 American Chemical Society
Publication on Web 11/30/1999

5532 Organometallics, Vol. 18, No. 26, 1999
Holtmann et al.
Sch em e 1
F igu r e 1. ORTEP plot of 3a . Thermal ellipsoids are shown
at the 50% probability level.
Ta ble 1. Selected Bon d Len gth s (Å) a n d An gles
(d eg) of 3a
group of Lappert reported the characterization of borane
(BH3) adducts of a germylene and of a stannylene.10
Here we report the reactions of decamethylsilicocene
(1) with trihalides of boron, aluminum, gallium, and
indium and with some organo-substituted derivatives
of these elements.
Bond Lengths
B(1)-C(1)
B(1)-C(2)
B(1)-C(3)
B(1)-C(4)
B(1)-C(5)
1.738(6)
1.747(6)
1.749(6)
1.762(6)
1.768(6)
B(1)-XCp
1.262
B(1)-Si(1)
Si(1)-C(11)
Si(1)-Cl(1)
Si(1)-Cl(2)
1.973(5)
1.865(4)
2.0666(16)
2.0671(16)
Bond Angles
102.50(8)
118.4(2)
136.0(3)
144.6(3)
Cl(1)-Si(1)-Cl(1)
C(11)-Si(1)-B(1)
C(1)-B(1)-Si(1)
C(2)-B(1)-Si(1)
C(3)-B(1)-Si(1)
C(4)-B(1)-Si(1)
C(5)-B(1)-Si(1)
XCp-B(1)-Si(1)
139.9(3)
127.7(3)
130.0(3)
170.8
Resu lts a n d Discu ssion
The reaction of decamethylsilicocene (1) with Cp*BCl2,
BX3, AlX3, Cp*AlCl2, Me2AlCl, GaX3, Cp*GaBr2, and
InX3 (X ) Cl, Br) proceeded cleanly under mild condi-
tions. In all cases the reaction pathway can be described
as a sequence of adduct formation followed by subse-
quent rearrangement steps.
shifts. No signal in the 29Si NMR spectrum could be
detected, probably due to fast relaxation processes
caused by the quadrupole moment of the neighboring
11/10B nucleus. The 13C NMR resonances were consistent
with the interpretation of one slowly rearranging σ-bound
and one π-bound Cp* ring. In the 11B NMR spectrum,
signals were observed at 6.1 and -54.6 ppm. The latter
value is in a range characteristic for pentacarba-nido-
hexaborane cations;14 the former corresponds to the
tetrachloroborate anion.
Crystals of 3a suitable for an X-ray structure analysis
were obtained by recrystallization from hot benzene or
toluene. An ORTEP plot of the molecular structure of
3a is shown in Figure 1. Selected bond lengths and
angles are given in Table 1, and crystallographic data
are collected in Table 2.
In the cation of 3a one Cp* ring is bound in an η1-
fashion to the silicon atom, while the other one is η5-
bound to the boron atom with an average B-C bond
length of 1.75 Å. This compares to 1.68 Å in [Cp*BBr]+-
[AlBr4]-,15 to 1.70 Å in 2,3,4,5-tetracarba-nido-hexabo-
rane (C4B2H6),16 and to 1.81 Å in [(Cp*B)Fe(CO)4].17 The
angle Si-B-XCp (where XCp is the in-plane center of the
Cp* ring) is 171°; the deviation from the expected linear
arrangement can be explained by crystal-packing ef-
fects. As predicted by theoretical calculations,18 the
methyl groups of the η5-Cp* ring are bent toward the
boron center with an average deviation from the ring
Silicocene 1 reacted slowly with Cp*BCl2 in benzene
or toluene at room temperature. Three equivalents of
the boron compound was needed for complete consump-
tion of 1 (Scheme 1), and 3a was formed as a colorless
solid that is sparingly soluble in aliphatic or nonpolar,
aromatic solvents. 3a is thermally stable at room
temperature but very reactive toward air and mois-
ture.11 Besides 3a , 1 equiv of the known Cp*2BCl12 was
formed, which was unambiguously identified by means
1
of H and 11B NMR spectroscopic examination of the
reaction mixture or of the mother liquor after crystal-
lization of 3a .
Due to the low solubility of 3a it was not possible to
obtain NMR spectra in solvents other than CD2Cl2 or
-
CDCl3. Under these conditions the Cp*BCl3 anion
decomposes with formation of Cp*H and BCl4-. Accord-
ingly, 3b was formed in these solutions (Scheme 1).
Similar behavior has been observed in the case of base
adducts of Cp*BCl2.13 The spectroscopic data of the
cation in 3a are not affected by the change in anion. In
the 1H NMR spectrum, one sharp resonance, in ac-
cordance with a π-bound Cp* ring, was found as well
as broadened signals due to another Cp* ring that is
σ-bound and undergoes relatively slow sigmatropic
(9) (a) Harrison, P. G.; Zuckerman, J . J . J . Am. Chem. Soc. 1970,
92, 2577. (b) Dory, T. S.; Zuckerman, J . J .; Barnes, C. L. J . Organomet.
Chem. 1985, 281, C1.
(10) Drost, C.; Hitchcock, P. B.; Lappert, M. F. Organometallics
1998, 17, 3838 and literature cited therein.
(11) No satisfactory microanalytical data could be obtained for
compounds 3-5. This is most probably due to (a) the high reactivity
of the compounds toward even trace amounts of air or moisture and
(b) partial formation of B4C3 on combustion.
(12) J utzi, P.; Krato, B.; Hursthouse, M.; Howes, A. J . Chem. Ber.
1987, 120, 565.
(14) (a) J utzi, P.; Seufert, A. Angew. Chem. 1977, 89, 339. (b) J utzi,
P.; Seufert, A.; Buchner, W. Chem. Ber. 1979, 112, 2488. (c) J utzi, P.;
Seufert, A. J . Organomet. Chem. 1978, 161, C5.
(15) Dohmeier, C.; Ko¨ppe, R.; Robl, C.; Schno¨ckel, H. J . Organomet.
Chem. 1995, 487, 127.
(16) Pasinski, J . P.; Beaudet, R. A. J . Chem. Soc., Chem. Commun.
1973, 928.
(17) Cowley, A. H.; Lomelf, V.; Voigt, A. J . Am. Chem. Soc. 1998,
120, 6401.
(18) J emmis, E. D.; Schleyer, P. v. R. J . Am. Chem. Soc. 1982, 104,
4781.
(13) J utzi, P.; Krato, B. Unpublished results.

Reaction of Silicocene with Group 13 Element Halides
Organometallics, Vol. 18, No. 26, 1999 5533
Ta ble 2. Cr ysta llogr a p h ic Da ta
3a
5b
empirical formula
C
30H45B2Cl5Si
C30H45BBr4Si2
0.5 × 0.4 × 0.3
792.3
cryst size, mm3
fw
1 × 0.4 × 0.2
632.62
cryst syst
space group
lattice params
a, Å
b, Å
c, Å
monoclinic
Cc
triclinic
P1
8.757(3)
15.192(5)
25.281(7)
90
98.85(3)
90
3323.3(18)
4
1.264
8.386(4)
8.530(5)
13.970(6)
104.97(4)
92.23(4)
117.36(4)
842.8(7)
1
R, deg
â, deg
γ, deg
V, Å3
Z
d
calcd, g/cm3
1.561
F igu r e 2. ORTEP plot of 5b. Thermal ellipsoids are shown
at the 50% probability level.
diffractometer
F(000)
Siemens R3
1336
Siemens P2(1)
398
µ(Mo KR), Å
0.710 73
-100
0.710 73
-100
benzene or aliphatic hydrocarbons, with 4 being slightly
more soluble in benzene than 3b, but they dissolve
readily in dichloromethane or chloroform. The NMR
spectroscopic examination of the reaction mixtures
showed that 5a ,b and Cp*BX2 were formed as byprod-
ucts (vide infra).
temp, °C
2θmax, deg
57
54
no. of data collected
no. of params refined
no. of obsd data (F > 4σ(F))
4634
358
3976
3931
195
3275
2
residuals: RF, RwF
0.0457, 0.1101 0.0526, 0.1267
for obsd data
largest peak in final diff
map, e/Å3
program used in refinement
abs cor
1
Compounds 3b and 4 were characterized by H, 13C,
0.397
1.243
and 11B NMR spectroscopy.11 The spectra of 3b were
identical with those described above. The 1H and 13C
NMR spectra of 4 show resonances that were assigned
to two different Cp* rings. In the 11B NMR spectrum
SHELX97
empirical
SHELX97
empirical
Sch em e 2
-
signals at -3.8 ppm for BBr4 and at -55.1 ppm for
the boron center in the cluster cation were detected.
As pointed out above, the product ratio depends
strongly on the polarity of the solvent. In the reaction
with BBr3 in dichloromethane/hexane (2/1) 5b was
isolated as the main product, although in moderate
yield. After storage for 2 weeks at -30 °C, colorless
crystals of 5b had precipitated from the yellow solution.
In the mother liquor, 4 and Cp*BBr2 were identified by
NMR spectroscopy.14b
plane of 4.2°. This behavior is documented in the
literature thus far only for [Cp*BBr]+[AlBr4]- (3.9°).15
In addition to the η5-Cp* moiety, the boron center bears
a Cp*SiBr2 substituent. The boron-silicon distance of
1.973(5) Å is comparable to that of 1.976(4) Å in t-Bu-
The 1H and 13C NMR spectra of 5b are consistent with
the presence of three different σ-Cp* moieties. The 13C
NMR data show that one Cp* ring additionally engages
in some kind of π-coordination. The 11B NMR spectrum
shows one resonance at - 39.2 ppm; in the 29Si NMR
spectrum one signal at 20.0 ppm is observed.
19
NdB-Si(SiMe3)3 and is well in the range of the sum
of the covalent radii of boron (0.81 Å) and silicon (1.17
Å).20 In the Cp*BCl3- anion of 3a , the Cl-B-Cl angles
are slightly smaller than expected for ideal tetrahedral
symmetry (∼106°), while the angles between the chlo-
rine atoms and the Cp* carbon atom are slightly
widened (∼113°). This is most probably due to the steric
bulk of the Cp* ligand.
The reaction of 1 with boron trihalides is more
complex. At least three different types of products are
formed, and the product ratio was found to depend on
the chosen solvent (Scheme 2). On reaction of 1 with
boron trichloride or tribromide in toluene at low tem-
perature, a deep red reaction mixture was obtained
initially. The red color vanished nearly immediately to
yield a yellow solution, from which 3b or 4 precipitated
during the reaction. The compounds were isolated in
moderate yield as a colorless powder and a crystalline
solid, respectively. Both are very sensitive toward air
and moisture. Both compounds are sparingly soluble in
The quality of the crystalline material of 5b was
sufficient to prove the connectivity by an X-ray structure
analysis. Figure 2 shows an ORTEP plot of the molec-
ular structure of 5b. The boron center is η4-bonded to a
pentamethylcyclopentadiene unit. One Cp*SiBr2 group
is σ-bonded to the boron atom, while another one is
σ-bonded at an allylic position of the π-bonded pentam-
ethylcyclopentadiene. Difficulties occurred in the X-ray
diffraction experiment. The collected data did not allow
us to surely distinguish between a displacement of the
molecule at a center of inversion or a racemic twinning
of the crystal. The plot shows the results of a calculation
assuming a twinned crystal (space group P1) with a
domain distribution of 57:43. Therefore, no detailed
discussion of the structural data is possible. Crystal-
lographic data are collected in Table 2.
Compound 3a is best described as a pentacarba-nido-
hexaboronium salt. Figure 3 shows the two possible
descriptions of the central unit in the cation, as a
π-complex of a borylene fragment or as a nido cluster
according to the rules established by Wade and Min-
(19) Haase, M.; Klingebiel, U.; Boese, R.; Polk, M. Chem. Ber. 1986,
119, 1117.
(20) Holleman-Wiberg. Textbook of Inorganic Chemistry, 91-100th
ed.;, Walter de Gruyter: Berlin/New York, 1985.

5534 Organometallics, Vol. 18, No. 26, 1999
Holtmann et al.
Sch em e 3
F igu r e 3.
gos.21 In analogy to the considerations concerning the
structure of 3a , the structure of the central fragment
in 5b can be described as a π complex or as an arachno-
carborane derived from pentaborane(11), as shown in
Figure 3. To our knowledge 5b is the first example of a
tetracarba-arachno-pentaborane that is stable at ambi-
ent temperature.
The arachno structure of 5b can be formally con-
-
structed by attaching an anionic Cp*SiBr2 fragment
to the π-bonded Cp* ring in the cation of 4. Presumably
4 is an intermediate in the formation of 5b. This
suggestion was supported by NMR monitoring experi-
ments. The 11B NMR spectra showed that, when the
starting materials were combined, at first only the
signal due to 4 at -55 ppm was present in the reaction
mixture. After a short period of time a signal at -39
ppm for 5b became apparent.
Sch em e 4
Although we have not as yet carried out detailed
mechanistic studies, we wish to propose possible path-
ways for the reactions described above as summarized
in Scheme 3. In the reaction of 1 with Cp*BCl2, the first
step is a nucleophilic attack of 1 on the Lewis-acidic
boron center. The adduct formation is followed by a
halide shift from boron to silicon. This process corre-
sponds to a formal insertion of a silylene into a boron-
halogen bond. The insertion product is a reactive
intermediate and rearranges by exchange of a Cp*
ligand from silicon to boron and of a halide ligand from
boron to silicon. This double migration corresponds to
a [1,2]2-dyotropic rearrangement,22,23 a concerted mech-
anism anticipated. The intermediate reacts further by
reductive elimination of Cp*SiCl and formation of Cp*2-
BCl. This kind of metathesis reaction has also been
observed for 1 in reactions with aluminum compounds
(see below). The silylene formed can be assumed to be
stable enough against rapid polymerization24 to allow
its insertion into a B-Cl bond of a Cp*BCl2 molecule.
Abstraction of a chloro ligand by attack of a further
Cp*BCl2 molecule leads to the final product. This
proposed mechanism is in accordance with the fact that
3 equiv of the boron compound is needed for complete
consumption of 1, and it explains the formation of 1
equiv of Cp*2BCl as a byproduct.
The situation is obviously more complicated for the
reaction of 1 with boron trihalides. At least three
different products are found (3b/4, 5a ,b, Cp*BX2) in
different amounts, depending on the solvent. The mech-
anisms leading to the final products are not fully
understood, but it is reasonable to assume pathways
which are in part comparable to those described in
Scheme 3.
The reaction of 1 with AlCl3 or AlBr3 in 1:1 stoichi-
ometry led to a mixture of unidentified products. When
1 was reacted with 2 equiv of the aluminum trihalides,
the known decamethylaluminocenium salts 6a ,b25 were
formed in high yield (Scheme 4). These were isolated
as colorless, crystalline materials which are very sensi-
tive toward air and moisture and virtually insoluble in
nonpolar solvents but easily soluble in dichloromethane.
6a ,b were identified by 27Al NMR spectroscopy. As a
(21) (a) Wade, K. Adv. Inorg. Chem. Radiochem. 1976, 18, 1. (b)
Mingos, D. M. P.; J ohnston, R. L. Struct. Bonding 1987, 68, 29.
(22) This designation of the rearrangement as [1,2]2 was introduced
by Paetzold in: Paetzold, P. I.; Grundke, H. Synthesis 1973, 635.
(23) Reetz, M. T. Adv. Organomet. Chem. 1976, 16, 33.
(24) Apeloig, Y.; Maxka, J . J . Chem. Soc., Chem. Commun. 1990,
737.
(25) Dohmeier, C.; Schno¨ckel, H.; Robl, C.; Schneider, U.; Ahlrichs,
R. Angew. Chem. 1993, 105, 1714.

Reaction of Silicocene with Group 13 Element Halides
Organometallics, Vol. 18, No. 26, 1999 5535
Sch em e 5
Sch em e 6
further reaction product, small amounts of the known
corresponding (pentamethylcyclopentadienyl)aluminum
dihalide26 were found. It was not possible to characterize
a silicon-containing product in the reaction mixture.
It was surprising that the reaction proceeded under
2-fold Cp* transfer from a silicon to an aluminum
center. A better understanding of the process described
in Scheme 4 could be achieved after investigating the
reaction of 1 with Cp*AlCl2. When 1 was added to an
equimolar amount of Cp*AlCl2 in toluene, a colorless
solid precipitated within a short period of time. The
product was unambiguously identified as the alumi-
nocenium salt 6a .25 Even with a 10-fold excess of
Cp*AlCl2, complete conversion to the aluminocenium
salt was observed without any consumption of 1 (Scheme
5).
Since no stoichiometric relationship between the two
components was found and 1 could be recovered quan-
titatively from the reaction mixture, the process is best
described as a decamethylsilicocene-catalyzed dismu-
tation of the aluminum compound (the dimeric Cp*AlCl2
and the aluminocenium salt 6a are structural isomers
with the molecular formula Cp*2Al2Cl4). Base-induced
dismutation reactions of cyclopentadienylaluminum
alkyls have been reported, but not for Cp*-substituted
compounds and not as quantitative processes.27 The fact
that 6a precipitates from solution is probably the
driving force for the quantitative conversion. This is the
first time that the function of 1 as a mere Lewis-base
catalyst has been observed.
Considering these results, we suggest that the
aluminocenium salts obtained from the reaction of 1
with aluminum trihalides are formed via the corre-
sponding (pentamethylcyclopentadienyl)aluminum di-
halides (Cp*AlX2). The proposed reaction pattern is
shown in Scheme 6. Nucleophilic attack of 1 on the
aluminum center and formation of a Lewis-acid-Lewis-
base complex is followed by halide transfer from alu-
minum to silicon. The formed Si(IV) species rearranges
by Cp* transfer to aluminum and halide transfer to
silicon. Subsequently Cp*SiX is formed by reductive
elimination. As pointed out above, this silylene can be
considered stable enough against polymerization to
allow further reaction with AlX3 to form another 1 equiv
of Cp*AlX2 and SiX2. Since 1 is present in the solution,
it enforces the dismutation of Cp*AlX2 to 6a ,b. At a
certain point 1 has been completely consumed in the
metathesis step. Afterward no further dismutation
occurs, and a small amount of Cp*AlX2 is left in the
reaction mixture.
Sch em e 7
a complex mixture of oligomers. This assumption is in
accordance with the fact that no defined silicon-contain-
ing product was detected. The reaction pathway de-
scribed is also in accordance with the observation of
Cp*AlX2 as a “byproduct”.
As the reactions of 1 with aluminum trihalides show,
halide ligands on aluminum are very easily substituted
by Cp* ligands. Use of dimethylaluminum chloride
should allow the substitution of only one halide ligand
and stop the reaction at an earlier state. Therefore, we
studied the reaction of 1 with Me2AlCl. Surprisingly,
the aluminum compound Cp*Al(Me)Cl28 (7) was formed
in good yield (Scheme 7). The product was identified by
1H and 27Al NMR and mass spectrometry. All attempts
to isolate or to trap the extruded [Cp*SiMe] have been
unsuccessful to date. It is remarkable that in this
reaction metathesis occurs with transfer of a methyl
rather than a chloro ligand. This surprising result has
initiated further exchange experiments with homoleptic
group 13 element alkyls.29
On going from the aluminum trihalides to the triha-
lides of the heavier homologues gallium and indium, an
interesting change in reactivity was observed. Instead
of reacting by Cp* transfer, 1 now served as a reducing
agent. In the reaction of 1 with equimolar amounts of
GaCl3 or GaBr3 in toluene as solvent, at low tempera-
ture (-80 °C), an orange solution was obtained. On
warming to room temperature, the color changed to
yellow; the quantitative conversion of 1 to the corre-
In the reaction described in Schemes 4 and 6, a silicon
atom which is introduced in the form of 1 leaves the
reaction sequence as the dihalo compound SiX2, which
is a highly reactive species that can be assumed to form
30
sponding dihalosilane Cp*2SiX2 was proved by NMR
spectroscopy. Within 3 to 5 days at room temperature,
gallium metal precipitated as a gray powder; Cp*2SiCl2
(26) Koch, H.-J .; Schulz, S.; Roesky, H. W.; Noltemeyer, M.; Schmidt,
H.-G.; Heine, A.; Herbst-Irmer, R.; Stalke, D.; Sheldrick, G. M. Chem.
Ber. 1992, 125, 1107.
(27) (a) Kroll, W. R. J . Chem. Soc,. Chem. Commun. 1969, 844. (b)
Fisher, J . D.; Wie, M.; Willett, R.; Shapiro, P. J . Organometallics 1994,
13, 3324.
(28) (a) Schonberg, P. R.; Paine, R. T.; Campana, C. F. J . Am. Chem.
Soc. 1979, 101, 7726. (b) Schonberg, P. R.; Paine, R. T.; Campana, C.
F.; Duesler, E. N. Organometallics 1982, 1, 799.
(29) J utzi, P.; Ku¨hler, T. Manuscript in preparation.
(30) J utzi, P.; Kanne, D.; Hursthouse, M.; Howes, A. J . Chem. Ber.
1988, 121, 1299.

5536 Organometallics, Vol. 18, No. 26, 1999
Holtmann et al.
Sch em e 8
Sch em e 9
Sch em e 10
filtration. From the separated solution, Cp*2SiCl2 and
Cp*2SiBr2, respectively, were isolated in nearly quan-
titative yield (Scheme 9).
Con clu sion
and Cp*2SiBr2, respectively, as well as the correspond-
ing gallium trihalide could be isolated from solution. The
silicon compounds were identified by NMR spectros-
copy;30 the gallium trihalides were identified in form of
their pyridine adducts. In these reactions, 1 obviously
serves as a dehalogenating agent, and not yet charac-
terized species of the composition “GaX” are formed,
which slowly disproportionate to yield GaX3 and Ga0
(Scheme 8). This behavior is similar to that of meta-
stable solutions of Ga(I) halides as described by Schno¨ck-
el and co-workers.31 The compounds described here
disproportionate at comparably higher temperatures;
thus, a different structure (perhaps comparable to “GaI”
as described by Green et al.32) and/or a stabilization by
Cp*2SiX2 molecules has to be taken into account. When
the yellow solutions were not allowed to stand for a few
days, and when all volatiles were removed immediately
after warming to room temperature, a yellow residue
was obtained, consisting of “GaX” and Cp*2SiX2. In the
solid state as well as in C6D6 solution the residue
decomposed with precipitation of gallium metal within
a few days.
If the proposed formation of “GaX” correctly describes
the course of events, then suitable substituents at the
gallium center should facilitate the formation of stable
Ga(I) compounds. In this context, use of the Cp*
substituent occurred to us as particularly promising,
since Cp*Ga is a well-known, stable compound.33 In an
NMR experiment, 1 was treated with a stoichiometric
amount of Cp*GaBr2. The NMR spectroscopic examina-
tion showed 1H, 29Si, and 71Ga resonances that could
unambiguously be assigned to Cp*Ga(I) (8) and to Cp*2-
SiBr2 (Scheme 8). The reaction proceeded quantitatively,
as no other products were identified.
The reactions of decamethylsilicocene (1) with halides
or organohalides of the trivalent group 13 elements
boron, aluminum, gallium, and indium can be explained
by multistep processes that include insertion reactions
of 1, rearrangements, and reductive-elimination steps.
Two different general pathways can be distinguished.
With boron and aluminum compounds, 1 engages in
Cp* transfer reactions. Here, the pronounced leaving
group character of the Cp* substituent on silicon comes
to the fore. In the reactions with boron compounds, the
silylboranes 3a ,b, 4, and 5a ,b are isolated, while in the
reactions with aluminum compounds the Cp*-Al spe-
cies 6a ,b and 7 are formed. In the latter no defined
silicon-containing products can be characterized; un-
stable silylenes are formed which are subject to oligo-
merization reactions. As shown in Schemes 3 and 6,
reaction sequences often include double-migration pro-
cesses classified as [1,2]2-dyotropic rearrangements.
Generally, the presence of a π-system in one of the
migrating groups facilitates such processes due to easy
interactions with the involved element center. In this
context another advantage of the Cp* ligand comes to
the fore, as indicated in Scheme 10. The ligand can
easily interact with and migrate to a neighboring
element center.
Toward gallium and indium compounds, 1 reacts as
a reducing agent. In all cases the corresponding Cp*2-
SiX2 (X ) Cl, Br) is formed in nearly quantitative yield.
The low-valent gallium halides that are formed in the
reactions with GaX3 so far could only be characterized
by their disproportionation products (Ga0 + GaX3), while
in the reaction with Cp*GaBr2 the stable Cp*Ga (8) is
formed. From the reactions with indium trihalides the
stable monohalides can be isolated.
In contrast to the Ga(I) halides, the monovalent
halides of In are stable at room temperature. Thus, we
expected a clean redox reaction of indium trihalides with
1. Indeed, 1 reacted with indium trichloride or tribro-
mide by complete reduction of the In(III) to the In(I)
species. In toluene slow dissolution of the indium
trihalide was observed on warming from -30 °C to room
temperature, followed by precipitation of the corre-
sponding indium(I) halide, which was removed by
The different results obtained within a series of group
13 element halides once more demonstrate the wide
reactivity spectrum of decamethylsilicocene.
Exp er im en ta l Section
Gen er a l Con sid er a t ion s. Standard Schlenk techniques
were used for all syntheses and sample manipulations. The
solvents were dried by standard methods and distilled under
argon prior to use. Mass spectra were run on a Varian 311 A
mass spectrometer (70 eV, 300 µA emission), and NMR spectra
were obtained using a Bruker Avance 500 spectrometer (1H,
500.132 MHz; 13C, 125.771 MHz; 29Si, 99.354 MHz; 11B,
(31) Tacke, M.; Plaggenborg, L.; Schno¨ckel, H. Z. Anorg. Allg. Chem.
1991, 604, 35.
(32) Green, M. L. H.; Mountford, P.; Smout, G. J .; Speel, S. R.
Polyhedron 1990, 9, 2763.
(33) Loos, D.; Schno¨ckel, H. J . Organomet. Chem. 1993, 463, 37.

Reaction of Silicocene with Group 13 Element Halides
Organometallics, Vol. 18, No. 26, 1999 5537
1
160.462 MHz; 27Al, 130.321 MHz). Proton NMR spectra were
referenced to the residual protic impurities of the deuterated
solvents, carbon NMR spectra were referenced to the solvent
signals, and hetero nuclei (29Si, 11B, 27Al) NMR spectra were
referenced to external standards. Melting points (uncorrected)
were measured with a Bu¨chi 510 melting point apparatus
using sealed capillary tubes. Unfortunately, no satisfactory
microanalytical data could be obtained for compounds 3-5,
probably due to their distinct sensitivity to even trace amounts
of air or moisture and perhaps also due to partial formation
of B4C3 on combustion. Products that were already known in
the literature were identified by comparison of the NMR
spectroscopic data.
198 mg (50% based on 1) of 5b. H NMR (CD2Cl2): δ 0.98 (s,
6H), 1.05 (s, 3H), 1.21 (s, 3H), 1.24 (br s, 3H), 1.83 (s, 12H),
1.84 (s, 6H), 1.89 (s, 12H). 29Si NMR (CD2Cl2): δ 19.99 ppm.
11B NMR (CD2Cl2): δ -39.2 ppm. 13C NMR (CD2Cl2): δ 9.68,
9.89, 10.11, 11.72, 12.03, 13.15, 13.40, 17.08, 18.46 (Cp* Me);
51.78, 56.56, 59.01 (Cp* ring/allylic), 109.34, 114.81, 136.96
(+ sh), 139.54, 139.98 (Cp* ring/vinylic). MS (EI): m/z (%) 792
(<1), M+; 657 (<1), M+ - Cp*; 469 (9.7), M+ - Cp*SiBr2; 415
(39.1), M+ - Cp*SiBr2 - Cp*; 163 (100), Cp*Si+. Mp: >150
°C dec.
Rea ction of 1 w ith Alu m in u m Tr ich lor id e.38 A slurry
of 508 mg (4 mmol) of AlCl3 in toluene (30 mL) was stirred at
-80 °C. A solution of 596 mg (2 mmol) of 1 in toluene (10 mL)
was added slowly, and the mixture was warmed to room
temperature while the color changed from orange to light
yellow. The upper solution phase was separated from the
resulting oily residue, which was dried in vacuo to yield
decamethylaluminocenium tetrachloroaluminate25 (6a ) in 80%
yield (742 mg). The compound was recrystallized from dichlo-
romethane/pentane. In the upper solution phase trace amounts
of (pentamethylcyclopentadienyl)aluminum dichloride26 were
identified by NMR spectroscopy. Yield: 742 mg (80%) of 6a .25
1H NMR (CD2Cl2): δ 2.20 s. 13C NMR (CD2Cl2): δ 9.40 (Me5C5),
122.13 (Me5C5). 27Al NMR (CD2Cl2): δ -115 (Cp*2Al+), 103
(AlCl4-).
Rea ction of 1 w ith Cp *BCl2. A solution of 0.52 g of 1 (1.7
mmol) in toluene (12 mL) was added to a solution of 1.12 g
(5.2 mmol) of Cp*BCl2 in toluene (17 mL). The resulting yellow
reaction mixture was stirred for 24 h and then concentrated
to 7 mL. Pentane (7 mL) was added, and the solution was
placed in
a freezer at -30 °C for 72 h. Afterward the
precipitated colorless crystals were filtered at low temperature.
The procedure was repeated once more to give a total yield of
997 mg (93%) of 3a as colorless crystals. The solution was
concentrated to yield 600 mg of a yellow residue which was
shown by NMR spectroscopy to consist nearly exclusively of
Cp*2BCl.12 Crystals of 3a which were suitable for X-ray
structure analysis were obtained by recrystallization from
benzene. Yield: 997 mg (93% with respect to 1). 1H NMR (CD2-
Cl2):37 δ 1.26, 1.93 (br, 15 H, SiCp*), 2.09 (s, 15H, BCp*). 13C
NMR (CD2Cl2): δ 9.41 (BMe5C5), 12.48 (br. SiMe5C5), 114.59
(BMe5C5), 136.21, 141.71 (SiMe5C5). 11B NMR (CD2Cl2): δ
-54.6 (BCp*), 6.1 (BCl4-). Mp: 110 °C dec.
Rea ction of 1 w ith Bor on Tr ich lor id e. A solution of 0.78
g (2.61 mmol) of 1 in toluene (25 mL) was cooled to -80 °C.
On dropwise addition of 11 mL of a 0.23 M solution of BCl3
(2.53 mmol) in toluene a red color appeared, which im-
mediately changed to deep yellow. When the reaction mixture
was warmed to room temperature, a colorless solid precipi-
tated. The solvent was removed in vacuo. The yellow residue
was recrystallized from methylcyclohexane (80 mL). Yield:
0.46 g (34%) of 3b. Mp: 151 °C dec. 1H NMR (CD2Cl2): δ 1.26,
1.93 (br, 15 H, SiCp*), 2.09 (s, 15H, BCp*). 13C NMR (CD2-
Cl2): δ 9.41 (BMe5C5), 12.48 (br, SiMe5C5), 114.59 (BMe5C5),
136.21, 141.71 (SiMe5C5). 11B NMR (CD2Cl2): δ -54.6 (BCp*),
6.1 (BCl4-). MS (EI): m/z (%) 414 (3.9), Cp*2SiBCl3+; 379 (67.5),
Cp*2SiBCl2+; 281 (100), Cp*SiBCl3+; 181 (75.6), Cp*BCl+; 163
(73.2), Cp*Si+.
Rea ction of 1 w ith Cp *AlCl2. A solution of 696 mg (3
mmol) of Cp*AlCl2 in 10 mL of toluene was cooled to 0 °C. A
solution of 100 mg (330 µmol) of 1 in 1 mL of toluene was
added, and the mixture was warmed to room temperature. A
colorless precipitate started to form after a few minutes. The
solution was separated from the precipitate after 2 h. The
colorless solid was identified as decamethylaluminocenium
tetrachloroaluminate (6a ). From the solution 1 could be
isolated in 90% yield. Yield: 670 mg (96%) of 6a .25 1H NMR
(CD2Cl2): δ 2.20 s. 13C NMR (CD2Cl2): δ 9.40 (Me5C5), 122.13
(Me5C5). 27Al NMR (CD2Cl2): δ -115 (Cp*2Al+), 103 (AlCl4-).25
Rea ction of 1 w ith Dim eth yla lu m in u m Ch lor id e. A
solution of 596 mg (2 mmol) of 1 in hexane (20 mL) was cooled
to - 80 °C. A 2 mL portion of a 1 M solution of dimethylalu-
minum chloride in hexane was added, and the mixture was
warmed to room temperature. The solvent was removed to give
a light yellow residue that was washed three times with
hexane (10 mL) to yield (pentamethylcyclopentadienyl)-
methylaluminum chloride (7)28 as a light yellow powder which
was recrystallized from toluene or benzene. The pentane
extracts were combined and volatiles removed in vacuo to give
380 mg of a yellow residue. The 29Si NMR spectroscopic
examination allowed no identification of defined products.
Yield: 0.28 g (65%) of 7.28 1H NMR (C6D6): δ 1.83 (s, 15H,
Cp*Al), -0.61 (s, 3H, MeAl). 27Al NMR (C6D6): δ -3 ppm (br).
MS (EI): m/z (%) 232 (10), Cp*AlCl2+; 212 (27), Cp*Al(Me)-
Cl+; 197 (73), Cp*AlCl+; 177 (23), Cp*AlMe+.
Rea ction of 1 w ith Bor on Tr ibr om id e. To a solution of
2.07 g (6.93 mmol) of 1 in toluene (30 mL) was added dropwise
a solution of 1.73 g of BBr3 (6.91 mmol) in toluene (10 mL)
while the reaction mixture was cooled to -80 °C. When the
mixture was warmed to room temperature, the color changed
from red to deep yellow, and a white precipitate was formed.
The slurry was concentrated, and pentane (30 mL) was added.
The precipitate was filtered, washed with 5 mL of pentane,
and recrystallized from dichloromethane. Yield: 2.03 g (37%)
of 4. Mp: >215 °C dec. 1H NMR (C6D6): δ 1.54 (br s, 15H,
SiCp*), 1.65 (s, 15H, Cp*B). 13C NMR (CDCl3): δ 9.7 (Me5C5B),
12.9 (Me5C5Si), 114.5 (Me5C5B), 141.7, 144.1 (Me5C5Si). 11B
NMR (CDCl3): δ -55.1 (Cp*B), -23.8 (BBr4-). MS (EI): m/z
(%) 469 (42.7), Cp*2SiBBr2+; 415 (60.2), Cp*SiBBr3+; 306 (8.5),
Cp*SiBBr2+; 225 (59.4), Cp*BBr+; 163 (100), Cp*Si+.
Rea ction of 1 w ith Ga lliu m (III) Ha lid es (Ga Cl3/Ga Br 3).
In a typical experiment 596 mg (2 mmol) of 1 was dissolved in
20 mL of toluene and the solution cooled to -80 °C. An
equimolar amount of the gallium halide was dissolved in 10
mL of toluene and added dropwise. On addition the reaction
mixture turned orange. After the mixture was warmed to room
temperature, the color had changed to light yellow. Within 3-5
days gallium metal precipitated from solution. The gallium
metal was separated by filtration. Pyridine was added to the
filtrate, and the corresponding gallium trihalide pyridine
adduct (GaX3‚C5H5N) precipitated, which was also separated
by filtration. Volatile components of the filtrate were removed
in vacuo, and the residue was examined by NMR spectroscopy.
Rea ction of 1 w ith Bor on Tr ibr om id e in Dich lor o-
m eth a n e/Hexa n e. A 298 mg portion of 1 (1 mmol) was
dissolved in 2 mL of dichloromethane, and the solution was
cooled to -80 °C. One milliliter of a 1 M solution of BBr3 in
hexane was layered on top, and the mixture was warmed to
-30 °C. On addition a dark red color occurred that slowly
changed to light yellow. The reaction flask was kept at -30
°C for 2 weeks. From the yellow solution colorless crystals of
5b did form, which were isolated and dried in vacuo. Yield:
(a) GaCl3: formation of Cp*2SiCl2.30 1H NMR (CDCl3): δ 1.15
(br s, 6H), 1.70 (br s, 12H), 1.80 (br s, 12H). 13C NMR (CDCl3):
δ 11.47, 12.16, 17.32 (Me5C5), 55.10, 136.12, 137.88 (Me5C5).
29Si NMR (CDCl3): δ 21.2. Yield: >90% (by 1H NMR and

5538 Organometallics, Vol. 18, No. 26, 1999
Holtmann et al.
weight); Ga0:34 yield 98% (by weight). GaCl3‚C5H5N: mp 126
°C (lit.35mp 126 °C). 1H NMR (CDCl3): δ 7.63 (m, 2H), 8.05
(m, 1H), 8.77 (m, 2H), Yield: 87%.
(a) InCl3 f InCl: light yellow powder. Yield: >90% (by
weight). Mp: 225 °C (lit.36 mp 225 °C). Cp*2SiCl2:30 1H NMR
(CDCl3) δ 1.15 (br s, 6H), 1.70 (br s, 12H), 1.80 (br s, 12H); 13
C
(b) GaBr3: formation of Cp*2SiBr2.30 1H NMR (CDCl3): δ 1.29
NMR (CDCl3) δ 11.47, 12.16, 17.32 (Me5C5), 55.10, 136.12,
1
137.88 (Me5C5); 29Si NMR (CDCl3) δ 21.2; yield: >90% (by H
(br s, 6H), 1.59 (br s, 12H), 1.76 (br s, 12H). 29Si NMR
1
(CDCl3): δ 19.3. Yield: >90% (by H NMR and weight). Ga0:
NMR).
yield 96% (by weight). GaBr3‚C5H5N: mp: 126 °C (lit.35 mp
(b) InBr3 f InBr: orange powder. Yield: >90% (by weight);
Mp: 220 °C (lit.36 mp 220 °C). Cp*2SiBr2:30 1H NMR (CDCl3)
δ 1.29 (br s, 6H), 1.59 (br s, 12H), 1.76 (br s, 12H); 29Si NMR
1
126 °C). H NMR (CDCl3): δ 7.63 (m, 2H), 8.05 (m, 1H), 8.81
(m, 2H). Yield: 79%.
1
(CDCl3) δ 19.3; yield >90% (by H NMR).
Rea ction of 1 w ith Cp *Ga Br 2. A 49 mg portion of 1 (160
µmol) and 58 mg of Cp*GaBr2 (160 µmol) were charged into
an NMR tube. A 0.6 mL amount of benzene-d6 was added to
form a yellow solution. Cp*2SiBr2 and Cp*Ga (8) were detected
as the only species present by NMR spectroscopy. 1H NMR
(C6D6): δ 1.29 (br s, 6H), 1.59 (br s, 12 H), 1.77 (br s, 12H,
Cp*Si), 1.90 (s, 15H, Cp*Ga). 29Si NMR (C6D6): δ 19.19 (Cp*2-
SiBr2).30 71Ga NMR (C6D6): δ -639.5 (br, Cp*Ga).33
Rea ction of 1 w ith In d iu m (III) Ha lid es. The indium
trihalide was suspended in toluene and cooled to -30 °C. A
solution of an equimolar amount of 1 in toluene was added
slowly, and the cooling bath was removed. The reaction
mixture was then stirred for 12 h. Afterward the precipitated
indium(I) halide was filtered. Volatile components of the
solution were removed in vacuo, and the residue was examined
by NMR spectroscopy without further purification.
Ack n ow led gm en t. This work was generously sup-
ported by the Fonds der chemischen Industrie and the
Deutsche Forschungsgemeinschaft.
Su p p or tin g In for m a tion Ava ila ble: Tables giving crys-
tal data and structure refinement details, atomic coordinates
and isotropic thermal parameters, bond distances and angles,
anisotropic displacement parameters, and hydrogen coordi-
nates for 3a and 5b. This material is available free of charge
via the Internet at http://pubs.acs.org.
OM9906114
(36) CRC Handbook of Chemistry and Physics, 61st ed.; Weast, R.
C., Ed.; CRC Press: Boca Raton, FL, 1980.
(37) In C6D6 or toluene-d8 the solubility of 3a was too low to obtain
(34) The yields of Ga0 and the pyridine adducts are given on the
basis of the disproportionation process 3“GaX” f 2Ga0 + GaX3. The
1H NMR spectra of the pyridine adducts correspond to spectra obtained
from commercially available gallium trihalides and pyridine.
(35) Greenwood, N. N. Adv. Inorg. Chem. Radiochem. 1963, 5, 91.
NMR spectra. In CD2Cl2 or CDCl3 the Cp*BCl3- anion decomposes to
-
BCl4 (vide supra).
(38) The reaction with aluminum tribromide was carried out in a
similar fashion. The product 6b was also characterized by NMR
spectroscopy (27Al NMR (CDCl3): δ -115, 80 (AlBr4-)).