The preparation of hexasilsesquioxane (T6) cages by "non aqueous" hydrolysis of trichlorosilanes

Article abstract of DOI:10.1039/b302556j

Hexasilsesquioxane cages (T₆) have been prepared from a range of alkyl and aryl trichlorosilanes using a "non aqueous" hydrolysis with dimethyl sulfoxide.

Full text of DOI:10.1039/b302556j

The preparation of hexasilsesquioxane (T6) cages by “non aqueous”
hydrolysis of trichlorosilanes
Alan R. Bassindale,*a Iain A. MacKinnon,b Maria G. Maesanoa and Peter G. Taylor*a
a The Chemistry Department, The Open University, Walton Hall, Milton Keynes, UK MK7 6AA.
E-mail: P.G.Taylor@open.ac.uk; Fax: (+44) 1908 858327; Tel: (+44) 1908 652512
b Dow Corning Ltd, Cardiff Road, Barry, South Glamorgan, UK CF63 2YL
Received (in Cambridge, UK) 6th March 2003, Accepted 15th April 2003
First published as an Advance Article on the web 13th May 2003
Hexasilsesquioxane cages (T6) have been prepared from a
range of alkyl and aryl trichlorosilanes using a “non
aqueous” hydrolysis with dimethyl sulfoxide.
196512 and has subsequently been revisited by Voronkov13,
Weber14 and Brook.15
Brook’s most recent study has shown that early in the
reaction the six membered cyclotrisiloxane, D3, is formed in
good yields. If six membered rings are preferred with
trichlorosilanes this should favour T6 formation through the ring
acting as a template or through the coupling of two such T3
rings.
Table 1 gives the results of our study†. We found that
trichlorosilanes could be reacted with dimethyl sulfoxide (2
equivalents) in chloroform at room temperature to give T6
T6 silsesquioxane cages (RSiO1.5)6 have received a great deal of
attention as cores for building octopus molecules1 and den-
drimers2 and as models of resins. Such cages have also been
used for the preparation of partially opened cages that
themselves act as models for silica surfaces and supports for
transition metal catalysts.3–6 However, only a limited range of
spherical organosilsesquioxanes have been prepared even
though a large variety of trichlorosilanes are available.7 This is
because conventional aqueous hydrolysis, including “scarce
water hydrolysis” leads to mainly resin and a large variety of
cage and open cage compounds. Only with specific alkyl groups
is particular cage formation favoured to such an extent and/or
solubility differences large enough to enable isolation of pure
silsesquioxane cages. Nevertheless the yields obtained by these
methods are very poor and the octasilsesquioxanes, (RSiO1.5)8,
T8, seem to be the preferred outcome. An alternative route to
such species is the formation of the spherical hydrogen
silsesquioxane followed by hydrosilylation using alkene.1 We
have used this approach to build T8 molecules with a range of
functionalities such as dendrimers, liquid crystals and surfac-
tants. We are interested in the preparation of functionalised
hexasilsesquioxanes, T6, and the mechanism of their conversion
into T8 and larger cage structures.8 However, we have been
unable to prepare hydrogen hexasilsesquioxane for use in our
hydrosilylation methodology and therefore we examined alter-
native methods of T6 formation.
Hexa(cyclohexylsilsesquioxane) has been prepared by
Feher3 using a modification of Brown’s method9, involving
aqueous hydrolysis in a water/acetone mixture. Reasonable
yields are obtained, but only after long periods. For example,
after four months of hydrolysis, Molloy obtained sufficient
crystals to carry out an X-ray crystal structure analysis.10 We
have obtained reasonable yields of this T6 molecule in a shorter
time using a similar aqueous/organic hydrolysis, however, this
route is limited to a few compact substituents with secondary or
tertiary groups adjacent to the silicon. Matsumoto has devel-
oped an alternative approach to T6 based on the dicyclohex-
ylcarbodiimide coupling of the silane triol or 1,1,3,3-tetra-
hydroxydisiloxane.11 Unfortunately this route seems only
possible with bulky substituents such as tert-butyl or 1,1,2-tri-
methylpropyl.
Table 1 Yields and properties of T6 cages prepared by the “non aqueous”
hydrolysis of trichlorosilanes using DMSO
Physical
state
of T6
product
29Si-
NMR/
ppm
Yield/
%
Starting trichlorosilane
Octyltrichlorosilane
254.2
256.6
254.4
255.4
266.9
Gel
25
11
9
11
7
Cyclohexyltrichlorosilane
Cyclopentyltrichlorosilane
2-Methylpropyltrichlorosilane
Phenyltrichlorosilane
Solid
Solid
Solid
Solid
Solid
3-(p-Methoxyphenyl)propyl trichlorosilane 254.4
6
cages. All yields quoted are isolated yields after purification.
The T6 cage is formed together with resin by-products, which
are sometimes difficult to separate. Silsesquioxane cages are
often purified by column chromatography, however, because of
the T3 rings we found that the T6 cages were particularly
susceptible to decomposition in the presence of nucleophilic
species. Thus the T6 cages decomposed on the silica or alumina
stationary phases used in column chromatography. However,
we found that treating the silica gel with 5–10% (by mass)
trimethylsilyl chloride before elution of the reaction mixture
gave an ideal chromatography medium that was sufficiently
active to maintain separation but not too active to cause
degradation of the T6. Although the yields are low in absolute
terms, they are perfectly acceptable when compared to conven-
tional cage syntheses. For example, scarce water hydrolysis of
trichlorosilanes gives T8 cages in about 15% yield. We have
carried out the “non aqueous” hydrolysis of dichlorosilanes to
give D3, D4, D5 and D6 rings using a range of other high
oxidation state oxygen donors such as triphenylphosphine
oxide, pyridine N-oxide, iodosobenzene and sulfur trioxide/
pyridine. However, DMSO was the only reagent that gave
specific cage compounds with organotrichlorosilanes.
We thus looked for an alternative method of constructing the
T6 cage that was easy to carry out, used readily available
starting materials and was applicable to a wide range of
substituents on silicon. The “non-aqueous” hydrolysis of
dichlorosilane using dimethyl sulfoxide was first reported in
1382
CHEM. COMMUN., 2003, 1382–1383
This journal is © The Royal Society of Chemistry 2003

Table 1 shows that, as expected, compact substituents such as
cyclopentyl, cyclohexyl and phenyl give T6 cages in acceptable
yields. Surprisingly, primary alkyltrichlorosilanes such as octyl,
2-methylpropyl and 3-(p-methoxyphenyl)propyl gave T6 cages
in comparatively good yields and this represents the first
synthesis of this important class of compounds. In contrast to
the Matsumoto route, we were unable to isolate T6 cages when
bulky substituents such as tert-butyl were employed. In this case
a range of products were observed which were probably D
polyol rings. Similarly cage structures were not produced with
small substituents, for example, vinyltrichlorosilane gave no
identifiable products. Chan has shown that HSiCl3 reduces
DMSO to dimethyl sulfide.16 In contrast we have been able to
form D3 and D4 rings from CH3SiHCl2 using DMSO, however,
reaction of DMSO with trichlorosilane led to a complex mixture
of products.
Our previous strategy for preparing organooctasilsesquiox-
anes relied on the synthesis of the alkene arm followed by
attachment to hydrogen silsesquioxane by hydrosilylation.1 We
have been able to transfer this methodology to T6 synthesis
through hydrosilylation of the alkene with trichlorosilane
followed by reaction with DMSO. For example, hydrosilylation
of methyl 3,3-dimethylpent-4-enoate using chloroplatinic acid
gives the corresponding trichlorosilane derivative, 1, in 77%
yield and thence the T6 derivative in 6.3% yield.
starting materials for the preparation of partially opened
cages.
Notes and references
†
Preparation of T6 cages: A solution of DMSO (4 mL, 54 mmol) in
chloroform (35 mL) was added to a solution of the trichlorosilane (27 mmol)
in chloroform (40 mL) and the mixture stirred at room temperature for 24 h.
The mixture was washed with water (4 3 20 mL), dried with MgSO4 and
the solvent removed under reduced pressure. Column chromatography of
the resulting gel using chloroform as the eluant gave a pure sample of the T6
cage compound.
Hexa(octylsilsesquioxane); yield 25%; dH (300 MHz, CDCl3, Me4Si)
0.10 (2H, br, CH2–Si), 0.36 (3H, t, CH3), 0.74 (12H, br, CH2); dC (75 MHz,
CDCl3, Me4Si) 11.36 (CH2–Si), 14.07 (CH3), 22.28 (CH2), 22.65 (CH2),
29.23 (CH2), 30.85 (CH2), 31.91 (CH2), 32.54 (CH2); dSi (79.3 MHz,
CDCl3, Me4Si) 254.2; m/z (MALDI) 1013 [M + Na]+; EA: Found: C,
58.10; H, 10.51. C48H102Si6O9 requires C, 58.18; H, 10.30%.
Hexa(cyclohexylsilsesquioxane)10; yield 11%; dH (300 MHz, CDCl3,
Me4Si) 0.78 (1H, br, CH–Si), 1.18 (5H, br, CH(ax)), 1.71(5H, br, CH(eq)); dC
(75 MHz, CDCl3, Me4Si) 22.6 (CH), 26.1 (2CH2), 26.6 (2CH2), 27.2 (CH2);
dSi (79.3 MHz, CDCl3, Me4Si) 256.6; m/z (EI) 811 [MH]+, 730.
Hexa(cyclopentylsilsesquioxane); yield 9%; dH (300 MHz, CDCl3,
Me4Si) 0.97 (1H, t, CH–Si), 1.45 (6H, br, CH2), 1.71 (2H, br, CH2); dC (75
MHz, CDCl3, Me4Si) 21.68, 26.80, 26.97; dSi (79.3 MHz, CDCl3, Me4Si)
254.4; m/z (EI) 726 [M]+. This was successfully analysed by X-ray
crystallography to confirm the structure. This will be discussed in detail in
a subsequent publication.
Hexa(2-methylpropylsilsesquioxane); yield 11%; dH (300 MHz, CDCl3,
Me4Si) 0.63 (2H, d, CH2), 0.96 (2H, d, 2CH3), 1.78 (1H, m, CH); dC (75
MHz, CDCl3, Me4Si) 22.0, 23.8, 25.6; dSi (79.3 MHz, CDCl3, Me4Si)
254.4; m/z (EI) 654 [M]+; EA: Found: C, 44.25; H, 8.49. C24H54Si6O9
requires C, 44.00; H, 8.31%).
Hexa(phenylsilsesquioxane); yield 7%; dH (300 MHz, CDCl3, Me4Si)
6.98 (5H, m, Ph) dC (75 MHz, CDCl3, Me4Si) 127.5, 130.3, 131.8, 134.2;
dSi (79.3 MHz, CDCl3, Me4Si) 266.9; m/z (EI) 774 [M]+; EA: Found: C,
55.16; H, 4.14. C36H30Si6O9 requires C, 55.88; H, 3.87 %).
Hexa(3-(p-methoxyphenyl)propylsilsesquioxane); yield 6%; dH (300
MHz, CDCl3, Me4Si) 0.65 (2H, t, CH2), 1.68 (2H, p, CH2), 2.53 (2H, t,
CH2), 3.73 (3H, s, CH3), 6.7 (2H, d, Ph), 6.9 (2H, d, Ph); dC (75 MHz,
CDCl3, Me4Si) 11.0 (CH2–Si), 24.4 (CH2), 37.6 (CH2–Ph), 55.3 (CH3),
113.7 (Ph), 129.4 (Ph), 134.2 (Ph), 157.8 (Ph); dSi (79.3 MHz, CDCl3,
Me4Si) 254.4; m/z (EI) 1205 [M]+ This was successfully analysed by X-ray
crystallography to confirm the structure. This will be discussed in detail in
a subsequent publication.
As with the T8 cages we are interested in preparing
polyfunctional cages and dendrimers. This requires the synthe-
sis of cages with functionality on the arms. Reaction of the
methyl dimethylpentanoate derivative with ethanediamine fol-
lowed by methyl acrylate will give a dendrimer with twelve
ester groups.17
The 29Si NMR chemical shifts of the T6 cage silicons are
usually in the region 254 ppm to 257 ppm. This is in
agreement with the chemical shifts of previously prepared T6
cages.4,11 The exception to this is phenyl hexasilsesquioxane
which has a 29Si NMR chemical shift of 266.89 ppm. This
upfield shift is typical of a phenyl substituent, as observed in the
D ring series. The corresponding range for T8 cages is 265 ppm
to 267 ppm. Marsmann and co-workers have developed a
method to estimate the 29Si NMR chemical shift of an unknown
1 A. R. Bassindale and T. E. Gentle, J. Mater. Chem., 1993, 3, 1319.
2 F. J. Feher and K. D. Wyndham, Chem. Commun, 1998, 323.
3 F. J. Feher, R. Terroba and J. W. Ziller, Chem. Commun., 1999,
2153.
4 F. J. Feher, R. Terroba and J. W. Ziller, Chem. Commun., 1999,
2309.
5 F. J. Feher, R. Terroba and R.-Z. Jin, Chem. Commun., 1999, 2513.
6 F. J. Feher, D. A. Newman and J. F. Walzer, J. Am. Chem. Soc., 1989,
111, 1741.
7 R. H. Baney, M. Itoh, A. Sakakibara and T. Suzuki, Chem. Rev., 1995,
95, 1409.
8 A. R. Bassindale, T. E. Gentle, J. Hardy, I. MacKinnon, M. Maesano, P.
G. Taylor, A. Watt and Y. Yang, 12th Int. Symp. Organosilicon Chem.,
Sendai, Japan, May 1999.
9 J. F. Brown and L. H. Vogt, J. Am. Chem. Soc., 1965, 87, 4313.
10 H. Behbehani, B. J. Brisdon, M. F. Mahon and K. C. Molloy, J.
Organomet. Chem., 1994, 469, 19.
11 M. Unno, S. B. Alias, H. Saito and H. Matsumoto, Organometallics,
1996, 15, 2413.
12 J. C. Gossens, French patent, 1 456 981, 1964; Chem. Abstr., 1967, 67,
, 54259.
13 S. V. Basenko and M. G. Voronkov, Dokl. Akad. Nauk SSSR, 1994, 339,
486; M. G. Voronkov and S. V. Basenko, J. Organomet. Chem., 1995,
500, 325.
14 P. Lu, J. K. Paulasaari and W. P. Weber, Organometallics, 1996, 15,
4649.
15 C. Le Roux, H. Yang, S. Wenzel, S. Grigoras and M. A. Brook,
Organometallics, 1998, 17, 556.
16 T. H. Chan, A. Melnyk and D. H. Harpp, Tetrahedron Lett., 1969,
201.
T
10 or T12 cage based on the chemical shift of the corresponding
T8 cage.18 For example, for T10 cages the equation dT10 = 1.028
3 dT8 holds.
A plot of the 29Si NMR chemical shift of the T6 cages in
Table 1 versus the chemical shift of the corresponding T8 cages
has a slope of 0.82 (R2 = 0.94 (forced to go through the origin)),
suggesting the equation dT6 = 0.82 3 dT8
.
The chemical shifts estimated using this equation correlate
well with the measured chemical shifts in Table 1, however,
differ most significantly for the phenyl and benzyl substituents.
The size of the constant in the equations above reflects the size
of the rings associated with the silicon in question. A silicon in
a T8 is associated with three T4 rings, whereas a silicon in a T10
is associated with two T4 rings and a T5 ring, the relief of angle
strain leading to a slightly higher factor. A silicon in a T6 cage
is associated with one T3 ring and two T4 rings, the increase in
ring strain leading to lower factor and the decrease is relatively
large. This matches the behaviour in D5, D4 and D3 rings.
In conclusion we have developed a robust method for the
synthesis of T6 cages in reasonable yield with a wide range of
substituents. This should lead to T6 structures being used in
exactly the same way that T8 cages are employed for building
octopus molecules and dendrimers, as models of resins and as
17 D. A. Tomalia and D. M. Hedstrand, Actual. Chim., 1992, 5, 347.
18 E. Rikowski and H. C. Marsmann, Polyhedron, 1997, 16, 3357.
CHEM. COMMUN., 2003, 1382–1383
1383