A higher yielding route to octasilsesquioxane cages using tetrabutylammonium fluoride, Part 2: Further synthetic advances, mechanistic investigations and X-ray crystal structure studies into the factors that determine cage geometry in the solid state

Article abstract of DOI:10.1016/j.jorganchem.2004.06.063

The recently-reported reaction of triethoxysilanes with tetrabutylammonium fluoride in the presence of scarce water provides a higher yielding route to substituted octasilsesquioxane cages. Further cages have now been prepared in good yield using this route and their X-ray crystal structures demonstrate that their cage geometries in the solid state are very similar with the packing constraints of the crystal lattices appearing to have the greatest influence on the geometry of the cage core. The mechanism of the reaction has been examined and the proposed mechanism of cage formation validated, though its exclusivity has not been categorically proven. By careful choice of reaction conditions, dramatic increases in rate and yield have been demonstrated.

Full text of DOI:10.1016/j.jorganchem.2004.06.063

Journal of Organometallic Chemistry 689 (2004) 3287–3300
www.elsevier.com/locate/jorganchem
A higher yielding route to octasilsesquioxane cages using
tetrabutylammonium fluoride, Part 2: further synthetic
advances, mechanistic investigations and X-ray crystal structure
studies into the factors that determine cage geometry in the solid state
a,
Alan R. Bassindale , Huiping Chen , Zhihua Liu , Iain A. MacKinnon ,
b
a
c
*
a
David J. Parker , Peter G. Taylor , Yuxing Yang , Mark E. Light ,
a,
a
d
*
d
Peter N. Horton , Michael B. Hursthouse
d
a
Department of Chemistry, Open University, Walton Hall, Milton Keynes, MK7 6AA, UK
b
Dow Corning Corporation, Midland, MI 48686, USA
c
Dow Corning Ltd., Barry, CF63 2YL, UK
EPSRC National Crystallography Service, University of Southampton, Highfield, Southampton, SO17 1BJ, UK
d
Received 26 January 2004; accepted 25 June 2004
Available online 28 August 2004
Abstract
The recently-reported reaction of triethoxysilanes with tetrabutylammonium fluoride in the presence of scarce water provides a
higher yielding route to substituted octasilsesquioxane cages. Further cages have now been prepared in good yield using this route
and their X-ray crystal structures demonstrate that their cage geometries in the solid state are very similar with the packing con-
straints of the crystal lattices appearing to have the greatest influence on the geometry of the cage core. The mechanism of the reac-
tion has been examined and the proposed mechanism of cage formation validated, though its exclusivity has not been categorically
proven. By careful choice of reaction conditions, dramatic increases in rate and yield have been demonstrated.
ꢀ
2004 Elsevier B.V. All rights reserved.
Keywords: T8 silsequioxane cage; Alkoxysilane; X-ray study; Mechanism
1
. Introduction
Initial routes to functionalised T cages focussed on
8
the hydrosilylation of 1 [6] but were limited by the poor
yields of the starting cage prepared from the hydrolysis
of trichlorosilane [7,8]. Other routes involving the
controlled hydrolysis of trichloro- or trialkoxysilanes
[9–11] or acid catalysed hydrolysis of 3-aminopropyltri-
ethoxysilane [6], allowing functionalised cages to be con-
structed in a single synthetic step, have also been
reported but similarly suffer from yields of no more than
T spherosilicate silsesquioxane cages, and in particu-
8
lar octahydro-octasilsesquioxane 1, have been used
extensively as scaffolds for the development of liquid
crystals [1,2], biocompatible materials [3], catalysts
[
4,5] and dendrimers [6]. The notation T denotes that
8
the closed cage comprises eight T-silicon atoms (i.e., sil-
icon atoms bonded to three oxygens).
30% (Scheme 1).
We have recently published a higher yielding route to
T cages using trialkoxysilanes, catalysed by tetrabutyl-
ammonium fluoride (TBAF), for a series of functional-
*
8
Corresponding authors. Tel.: +44 1908 652512; fax: +44 1908
8
58327.
E-mail address: p.g.taylor@open.ac.uk (P.G. Taylor).
ised T Õs and reported improved yields of up to 95%
8
0
022-328X/$ - see front matter ꢀ 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.jorganchem.2004.06.063

3288
A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
H
Cy
new cages in a similar way with respectable yields. The
H
Cy
Cy
reaction of 3-para-methoxyphenylpropyltriethoxysilane
with TBAF in chloroform for 24 h gave the corresponding
pure octa-3-para-methoxyphenylpropyl-octasilsesquiox-
ane 5 in 17% yield as the only isolated cage compound
and with no further purification or chromatographic
separation of components required. Iso-octyltri-
methoxysilane was similarly reacted with TBAF in a
dichloromethane solution to give octa-iso-octyl-octa-
silsesquioxane 6 as the major component which could
be purified by column chromatography with an isolated
yield of 43% (Scheme 3).
Si
O
H
Si
Si
O
Si
O
O
H
O
O
O
O
O
O Si
O
Si
O
Si
O
Si
O
Si
Si
O
Si
O
O
O
O
O
Cy
Cy
H
H
Si
Si
O
Si
1
2
Cy
H
H
Scheme 1.
In order to probe further our proposed mechanism of
cage formation [12] we have conducted a series of exper-
iments on the reactions of cyclopentyltriethoxysilane or
cyclohexyltriethoxysilane with TBAF to give the corre-
[
12]. We suggested that because our source of TBAF
contained 5% water, complexation of the basic fluoride
ion to water enabled hydroxide ion to attack the alkox-
ysilane to form the corresponding silanol. Further inter-
action of fluoride ion with the hydrogen of this silanol
would similarly lead to an increase in the nucleophilicity
of the oxygen leading to further reaction and Si–O–Si
bond formation. We have also proposed that such a
mechanism is present in the ring expansion reactions
of hexacyclohexyl-hexasilsesquioxane 2 by dialkyl/aryl
diethoxysilanes in the presence of scarce water and
TBAF that leads to isolable and fully characterised
T D 3 and T D cages (for example regioisomer 4)
sponding T cages 7 or 8, respectively. By preparing 7
8
firstly using stock TBAF solution containing ꢀ5% water
and then using two TBAF solutions containing different
1
8
16
proportions of O and O-labelled water, we have been
able to use a Fourier transform ion cyclotron resonance
mass spectrometer (FTICR-MS) to look at how the
mean mass of a selected cage-comprising ion changes
from experiment to experiment. This has allowed us to
1
8
see whether O is incorporated into the cage from the
1
water, determine an average number of O atoms per
8
1
8
cage and see how that number changes with O content
in the water. As discussed later, we were unable to deter-
6
1
6
2
1
8
[
13] (Scheme 2).
In this paper we examine further functionalised T8
mine the exact proportion of O-labelled water in our
stock solutions and thus comparisons with the propor-
1
8
cages made using our scarce water, TBAF catalysed
route and consider in more detail the structural and geo-
metric implications of different pendant functionalities
attached to the cageÕs Si O core. We also further probe
tion of O in the sislesquioxane cages is qualitative
rather than quantitative.
Fig. 1 shows the FTICR mass spectrum of the
+
8
12
[M + NH ] ion of 7 prepared (a) in the absence of
4
1
8
the mechanism of cage formation.
O-labelled water in the TBAF solution, (b) with a
1
8
low proportion of O-labelled water-containing TBAF
1
8
solution and (c) with a higher proportion of O-labelled
water-containing TBAF solution. Our stock supply of
TBAF was a THF solution containing 5% water and
was used to conduct reaction (a). Reaction (b) was per-
formed similarly but using a solution of TBAF which
2
. Results and discussion
Following our previously reported synthesis of T8
cages with ÔwetÕ TBAF we have prepared a number of
Cy
R
Cy
Cy
Si
O
Si
Si
O
O
Si
Cy
Me
Me
O
O
R
Cy
O
O
O
O
O
R
R
Cy
Si
O
Si
Si
O
Si
Si
O
Si
Si
Si
Cy
Cy
O
O
R=Me, Et, Ph
O
O
O
O
O
Cy
Si
O
Si
Si
Cy
3
Cy
Cy
4
Scheme 2.

A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
3289
R
R
CH CH CH
OMe
5
: R =
2
2
2
Si
O
Si
R
R
O
O
O
O
O
O
H
Si
O
Si
Si
O
Si
6: R =
CH2
C
CH2
C(CH3)3
O
O
R
CH3
R
7
8
: R = Cp
: R = Cy
Si
O
Si
R
R
Scheme 3.
+
18
ion of 7 prepared from TBAF solutions containing (a) no O-labelled water, (b) a low proportion of
Fig. 1. FTICR mass spectra of the [M + NH
4
]
18
18
O-labelled water and (c) a higher proportion of O-labelled water-containing TBAF solution.
1
8
had been dried over molecular sieves to remove most of
its water content. Removing all the moisture was not a
practical possibility because of the strong hydrogen
bonding by fluoride ion. O enriched water was then
added to make the final water content 5%. Reaction
O atoms into the cage. In reaction (c) the isotopic dis-
tribution is even more complex with a higher mean mass
+
for the [M + NH ] ion. This demonstrates that the ex-
4
1
8
18
tent of O atom incorporation is more extensive in the
products of reaction (c) than in reaction (b) and is con-
8
1
(
tion but the added water had a higher O-label
c) used a similar portion of pre-dried batch TBAF solu-
1
sistent with the higher proportion of O labelled water
present at the start of reaction (c).
8
1
The extent of O atom incorporation can be quanti-
8
enrichment.
The isotope pattern for the [M + NH ] ion of reac-
tion (a) is relatively simple as expected with no incorpo-
+
18
fied in terms of the average number of O atoms per cage
4
(AO18) from the isotopic distribution of a given pseudo-
molecular ion of T using the following expression:
1
8
ration of O isotope, however for (b) there is greater
complexity and a slightly higher mean mass for the pseu-
domolecular ion due to the incorporation of up to three
8
AO18 ¼ ðMmeas ꢁ MO16Þ=ðmO18 ꢁ mO16Þ:

3290
A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
1
8
18
O inclusion in the cage is proportional to the O con-
Mmeas, MO16, mO18 and mO16 are the average mass
measured for the pseudomolecular ion of a T cage con-
tent of the water in Table 1 together compliment our
proposed reaction mechanism whereby the oxygens that
form the cage structure are derived from the water of the
TBAF and not from the alkoxy-groups of the starting
silane.
8
1
taining both O and O isotopes, the average mass
8
16
measured for the pseudomolecular ion of a T cage con-
8
1
6
18
taining all (i.e., 12) OÕs, the mass of an O atom, and
1
6
the mass of an O atom, respectively. Mmeas and MO16
are mathematically derived by calculating each isotopic
peakÕs fractional intensity contribution to the overall
intensity, multiplying that fraction by its mass and then
summing the results. The values of Mmeas and AO18 for
reactions (a), (b) and (c) are given in Table 1. The value
of AO18 obtained is shown to be dependent on the
Our investigations to determine accurately the pro-
1
8
portion of O in our starting TBAF solutions by
studying the hydrolysis of monochlorosilanes (chlorotri-
methylsilane and chloromethyldiphenylsilane) to their
corresponding disiloxanes with O-labelled water were
inconclusive despite the water in the TBAF solutions
being the only oxygen atom source available for the
reactions. The O/ O ratio of the disiloxanes prepared
from the same TBAF solution that was used in reaction
(c) were calculated from the EI mass spectral data from
GC–MS analyses. The theoretical AO18 values for the
octasilsesquioxane cages was then determined from
these ratios. Corresponding AO18 values ranging from
7.7 to 8.5 were obtained and were found to vary
from day to day and sample to sample. This means that
though the corresponding AO18 values calculated for the
disiloxane reactions are similar to that measured for the
1
8
1
8
amount of O present in the water.
1
8
16
A similar statistical treatment of the FTICR-MS data
was carried out on samples of 7 and 8 prepared under
identical reaction conditions using fractions of the same
1
8
TBAF solution containing O-labelled water (reactions
(
(
e) and (g), respectively) as was used to conduct reaction
c). Values of A
+
based on both the [M + H] and
O18
+
[
M + NH ] ions were calculated and compared to sam-
4
ples of 7 and 8 prepared without labelled water (reac-
tions (d) and (f), respectively) as shown in Table 2.
Confirmation that the statistical analysis results are
reproducible and independent of which pseudomolecu-
lar ion is chosen, is provided by the mean AO18 values
calculated from the isotopic distributions of the
1
8
16
T cage in reaction (c), the O/ O ratios measured in
8
the disiloxanes are not accurate representations of
16
the O/ O ratio in the water content of the TBAF
solution.
1
8
+
+
[
RT + NH ] and [RT + H] ions which are, within
8 4 8
experimental error, the same as each other. Also, as
reaction (e) is effectively a repeat of reaction (c) we find
as expected that their AO18 values are essentially the
same as each other within experimental error. The vari-
ation that we do see between these experiments we pre-
sume reflects the different reaction times we employed
and the hygroscopic nature of our TBAF solution.
The very similar AO18 values of the different cages pre-
pared in reactions (e) and (g) plus the observation that
We attribute the reason why the AO18 values for the
disiloxanes were consistently higher than that calculated
for the cage in reaction (c) to the fact that the disiloxane
reactions were performed first and the hygroscopic nat-
ure of our TBAF solutions may have resulted in a de-
1
8
16
crease in the O/ O ratio between experiments. Our
findings bring us to conclude that while oxygen incorpo-
ration into T cages from the ethoxy-groups of the tri-
8
ethoxysilane cannot be ruled out, this mechanism
requires the fluoride ion promoted breaking of an O–
Csp bond under mild conditions. These limitations
3
Table 1
Values of Mmeas and AO18 calculated for 7 in reactions (a), (b) and (c)
make such an alternative mechanism less plausible and
point to our fluoride/water mechanism as being far more
probable. Our attempts to prepare 7 by replacing TBAF
with tetrabutylammonium chloride were unsuccessful
and only starting materials were retrieved from the reac-
tion over similar time scales. This points to fluoride ion
as having a critical role in the mechanism of cage
formation.
+
from the corresponding isotopic distribution of the [M + NH
4
]
ion
shown in Fig. 1
+
Reaction
M
meas ([CyPT
8
+ NH
4
]
ions)
A
O18
a
(a)
(b)
(c)
988.0612
990.1377
1001.7673
0
1.0
6.8
a
Mmeas = MO16.
Table 2
+
+
4
Values of Mmeas and AO18 calculated for 7 or 8 in reactions (d), (e), (f) and (g) from the isotopic distributions of the [M + H] and [M + NH ] ions
observed by FTICR-MS
Reaction
Compound
R
M
meas for
Mean number of
18
O from [RT + H] ions
M
meas for
Mean number of
18
A
O18
+
+ H] ions
+
+
+
[RT
8
8
[RT
8
+ NH
4
]
ions
8 4
O from [RT + NH ] ions
(
(
(
(
d)
e)
f)
7
7
8
8
CyP
CyP
CyH
CyH
970.4385
983.1270
0
987.6062
1000.2182
1099.8050
1112.6543
0
0
6.3
0
6.3
0
6.3
0
1082.7616
1095.9170
g)
6.6
6.4
6.5

A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
3291
Table 3
The isolated yield of 7 obtained from reactions of cyclopentyltrieth-
oxysilane with TBAF in different solvents
deuteriochloroform solution containing TBAF. Deute-
riochloroform was chosen as it is a relatively good sol-
vent for solublising 7. After stirring a 2:1 ratio
solution of cyclopentyltriethoxysilane and TBAF for
24 h the Si NMR spectrum of the reaction mixture
showed a sharp peak determined from standard samples
to arise from 7 and an unresolved array of smaller sharp
peaks between ꢁ60 and ꢁ70 ppm. The peaks in the ꢁ60
to ꢁ64 ppm region particularly suggested the presence
of silanol groups in partial cages. After a further 24 h
the peak for 7 had grown in relation to others and pre-
cipitation from acetone afforded 7 in 13% yield.
Solvent
Reaction time (h)
8
Yield of T (%)
2
9
Chloroform
THF
Acetone
48
12
4
13
23
95
Table 3 shows the effect of different solvents on the
yield and reaction times for the preparation of 7 using
the same TBAF solution. The low solubility of 7 in
many organic solvents, particularly acetone, may ex-
plain why it is formed in such a high yield since it precip-
itates readily from solution upon formation and
therefore drives the reaction towards completion.
To test whether these uncharacterised silanol-con-
taining species were formed from the breakdown of 7
and so were end products in their own right or whether
they were intermediates in the formation of 7 we under-
took the synthesis of 7 in deuteriochloroform with a
twofold excess of TBAF to cyclopentyltriethoxysilane.
The synthesis of 7 has also been studied using NMR
spectroscopy where the TBAF solution contained differ-
ent levels of moisture. Under dried conditions the
2
9
After 24 h the Si NMR spectrum showed only a quin-
tet at ꢁ113.99 ppm corresponding to 9 and a single peak
at ꢁ66.55 ppm for 7. By stopping and working-up the
reaction, 7 was isolated with an improved yield of
23%, consistent with the greater conversion of the sila-
nol-containing partial cages into 7 by the excess of fluo-
+
formation of the pentacoordinate species [NBu ]
4
ꢁ
29
cyclopentylSiF4] 9 was observed in the Si NMR
[
spectrum of the reaction mixture as a pentet at
ꢁ
113.99 ppm alongside that of 7 at ꢁ66.55 ppm. The
formation of 9 is consistent with cage formation taking
place using the remaining trace water in the TBAF until
its supply is exhausted and then fluoride ion directly at-
tacks the remaining cyclopentyltriethoxysilane (Fig. 2).
2
9
ride ion and their consequent absence from the Si
NMR spectrum. If these silanol-containing compounds
were formed from the breakdown of 7 then the yield
of 7 would have been expected to be less due to the ex-
cess of fluoride ion promoting more extensive cage
breakdown and consequently increasing the proportion
of silanol-containing partial cage compounds (Scheme
4).
2
9
Under conditions of excess water the Si NMR spectra
of the reaction mixture showed no pentacoordinate spe-
cies but a mixture containing a reduced amount of 7 and
an array of silanol-containing partial cage structures.
This observation underlines the importance of careful
control of the TBAFÕs water content in order to obtain
the highest yield of cage product.
We have also studied the X-ray crystal structures of
many of the octasilsesquioxane cages we have synthe-
sised and have recently reported the structures of 7, 14
and 15 [12]. This paper supplements these with new or
further structures for 5, 8 and 10.
To understand the role of these partial cage struc-
tures as potential intermediates in cage synthesis we used
2
9
Si NMR to study the formation of 7 with time in a
Cp
Cp
EtO
Si
OEt
F- +NBu
EtO
Si
F
F- +NBu
4
4
OEt
OEt
_
Cp
Si
Cp
Si
Cp
Si
F
F
F
F
EtO
F
F-
F-
F
F
F
F
+NBu
+NBu
4
+NBu
4
4
9
Fig. 2. The proposed mechanism for the formation of 9 from cyclopentyltriethoxysilane and TBAF in the absence of water.

3292
A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
nC H
shows that the octyl arms of 10 do not interdigitate with
8
17
nC H
8
17
those of adjacent molecules in the lattice. This arrange-
ment of arms is strongly reminiscent of the phase align-
ing of mesogenic arms in liquid crystal silsesquioxane
cages proposed by Maurer et al. [14] and supporting
the work of Goodby and Mehl [15] on the liquid crystal-
line properties of the species formed by coupling of the
cyanobiphenyl mesogen 11 with various siloxanes or 1
[15–17] Fig. 6 (Scheme 5).
Si
O
Si
nC H
8
17
nC H
8
17
O
O
O
O
Si
O
Si
Si
O
Si
O
O
O
O
H17
nC8
nC H
8
17
Si
O
Si
Though a crystal structure for 8 was first obtained by
Barry et al. [18] we have obtained three further structures
which vary in both unit cell dimensions and space group.
These structures 8a, 8b and 8c were obtained as the crys-
tallised products of rearrangement reactions of 2 in the
presence of various alkylethoxysilanes and TBAF. While
the full discussion of these rearrangements will be the
subject of a future paper, their key crystal and cage dimen-
sional data are summarised in Table 4 together with the
corresponding data of selected octasilsesquioxanes from
nC H
nC H
8 17
8
17
1
0
Scheme 4.
Figs. 3 and 4 show how the octyl arms of 10 and to a
lesser extent the 3-para-methoxyphenylpropyl arms of 5
align themselves along one axis. Fig. 5 additionally
Fig. 3. The isotropic crystal structure of 5. All eight arms are disordered over two positions (50/50 ratio) with only one orientation being shown here.
Arm orientations alternate between adjacent unit cells. Hydrogen atoms are omitted for clarity.
Fig. 4. ORTEP-representation of 10 with envelopes drawn at the 30% probability level. The arm beginning C9 is disordered over two positions (60/
40 ratio) but only the major orientation is shown here. The hydrogen atoms are omitted for clarity.

A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
3293
Fig. 5. Isotropic structure view down the b axis of the unit cell of 10 showing the packing arrangement of the molecules. The hydrogen atoms are not
shown for clarity.
in larger than ideal thermal parameters we can still rea-
sonably conclude from our data that at least a percent-
age of the molecules in the crystals 8a and 8b contain
cyclohexyl rings in a real boat conformation, occurring
in the positions suggested Figs. 8 and 9.
Table 4 shows that despite the wide variety and size of
carbon-bond pendant groups considered, the variation
in cage geometry is surprisingly limited given the known
flexibility of the Si–O–Si and O–Si–O unit and the varia-
tion in hybridisation state and electronic environment of
the binding carbon atom. While the variation in mean
ortho, meta and para inter-silicon distances are closely
linked and changes to the mean Si–O–Si and O–Si–O
bond angles are also strongly related, there is little over-
all correspondence between them.
In particular the mean RO–Si–O bond angles of
structures 8a, 8b and 8c are different enough to span
over half the total range of values observed. It is hard
therefore to explain the relatively large structural differ-
ences between them by an argument based either on
Fig. 6. The isotropic crystal structure 8a. The rings in the boat
conformation are disordered over three positions (only one orientation
is shown here) around the ꢁ3 axis in a manner that only boat
conformations can exist. Hydrogen atoms omitted for clarity.
electronic or steric control of the geometry of the
Si O cage core by the pendant group. Our observa-
8
12
tions are comparable with the conclusions of a similar
crystal structure study of octasilsesquioxanes by Podbe-
rezskaya et al. [25]. We postulate therefore that the pri-
mary influencing factor on cage geometry in the solid
state is the moleculeÕs crystal lattice and that the influ-
ence of a carbon-bound pendant group on the geometry
of a silicon to which it is attached is relatively unpro-
nounced. This conclusion contrasts with the important
influence we believe the steric bulk of a pendant group
has in determining which size silsesquioxane cages are
preferentially formed [12].
(
CH ) O
CN
2
n
n =2, 4 or 9
11
Scheme 5.
the literature. Structure 8b contains three independent
molecules 8b.1, 8b.2 and 8b.3 each of which is shown in
Fig. 7.
Close inspection of the cyclohexyl rings of 8a and 8b
shows that the constraint of crystal packing sometimes
causes them to adopt a boat conformation. Though
the often poor internal order of these crystals made
obtaining high quality X-ray data difficult and resulted
We are currently examining other high-yielding
synthetic routes to substituted T , T , T10 and T12 sils-
6 8
esquioxane cages including studies of their interconver-
sion and of their liquid crystalline and other material
properties.

3294
A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
Fig. 7. The isotropic crystal structures of the three independent molecules 8b.1, 8b.2 and 8b.3 in the asymmetric unit that comprise structure 8b.
While there is disorder in a number of the cyclohexyl rings in each molecule (only one orientation of each is shown here) the boat conformation does
occur as shown in at least a percentage of the molecules in the crystals. The hydrogen atoms omitted for clarity.
Fig. 8. ORTEP-representation of 8c with envelopes drawn at the 30% probability level. The hydrogen atoms are omitted for clarity.

A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
3295
Table 4
A summary of key cage dimensional data of various substituted octasilsesquioxanes determined from the X-ray crystal structures of our own
compounds and analogues the literature
Pendant group
Number Reference Mean
Mean
Mean
Mean
Ortho Si–Si Meta
˚
Para
˚
˚
Si–Si (B) (A) Si–Si (C) (A)
˚
Si–O (A) Si–O–Si (ꢁ) O–Si–O (ꢁ) RO–Si–O (ꢁ) (A) (A)
Allyl
CH CH
11
12
13
8a
14
5
[19]
[20]
[21]
–
1.62
1.60
1.61
1.61
1.62
1.62
1.61
1.62
1.62
1.61
1.62
1.62
1.62
1.62
1.61
1.61
150.7
150.3
150.3
150.1
149.8
149.5
149.4
149.2
149.2
148.8
148.6
149.3
149.3
148.9
149.3
149.2
108.2
108.4
108.5
108.6
108.7
108.8
108.8
108.8
108.8
108.9
108.9
109.0
109.0
109.0
109.0
109.1
324.7
325.3
325.6
325.7
326.1
326.4
326.4
326.5
326.5
326.6
326.8
326.9
327.0
327.1
327.1
327.2
3.13
3.10
3.10
3.10
3.12
3.12
3.11
3.12
3.12
3.11
3.12
3.11
3.12
3.11
3.10
3.10
4.43
4.38
4.39
4.39
4.41
4.41
4.39
4.41
4.41
4.40
4.41
4.40
4.41
4.40
4.39
4.39
5.42
5.37
5.37
5.38
5.40
5.40
5.38
5.40
5.40
5.38
5.40
5.39
5.40
5.39
5.37
5.38
2
2
2 3
Si(CH@CH )
Vinyl
Cyclohexyl (1)
i-Butyl
p-Methoxyphenylpropyl
[21]
–
a
Cyclohexyl (2)
CH CH CMe
8b
15
10
16
7
–
2
2
2
CO Me
2
[12]
–
n-Octyl
Me-piperidone
Cyclopentyl
[22]
[12]
[23]
–
3-Iodopropyl
Cyclohexyl (3)
17
8c
18
19
a
Benzyl
Ethyl
[24]
[25]
[26]
Ph (w. acetone solvate) 20
a
Average values from two or more independent molecules in the crystal.
R
tions mechanism and that strong donor solvents such as
acetone dramatically improve the reaction rate and there-
fore the isolated yield of T cages. Our findings provide
considerable supporting evidence for the mechanism of
cage formation that we have previously proposed.
R
R
R
Si
Si
Si
8
Si
Si
Si
Si
B
C
Si
4. Experimental
A
R
R
4.1. General
R
R
IR. Infrared spectra were obtained as Nujol mulls
using sodium chloride plates with a Nicolet 205 FT-IR
spectrometer.
Fig. 9. A schematic outline of an octasilsesquioxane cage to show the
three internal Si–Si cage distances measured. The mean values of each
in the X-ray crystal structures of various examples are reported in
Table 4.
NMR. All NMR spectra were obtained using either a
JEOL EX 400 NMR or a JEOL Lambda 300 NMR spec-
2
9
3
. Conclusions
trometer with the pulse delay for Si NMR spectra stand-
ardised at 20 s. All spectra were recorded at room
The scarce-water reaction of tetrabutylammonium flu-
temperature (25 ꢁC) using deuterated chloroform (CDCl )
3
˚
dried over 4 A molecular sieves as solvent unless stated
oride with substituted triethoxysilanes has been shown to
be a higher yielding route to octasilsesquioxane cages. We
have prepared further T cages using this route and dem-
otherwise. The NMR external reference compound was
1
13
29
tetramethylsilane (TMS) for H, C and Si NMR spec-
tra. The spectral data point position of this compound
8
onstrated by the comparison of the X-ray crystal struc-
tures of our cages with those in the literature that the
geometry of the cages in the solid state are very similar
and what differences there are between them is primarily
the result of packing constraints of the crystal lattices
on whole molecules rather than the influence of pendant
groups on the individual cages to which they are attached.
Our investigations into the mechanism of cage forma-
tion showed that it can occur by incorporation of the oxy-
gen atoms from the water contained in the TBAF solution
and that by optimising the waterÕs concentration im-
1
was accurately located before data acquisition. H– H
1
1
13
and H– C COSY and DEPT programs were frequently
1
13
used to help the assigning of H and C NMR spectra.
GC–MS. GC–MS analyses were performed using an
Agilent 5973 series instrument with a 30 m · 0.25
mm · 0.25 lm Rtx-1 column connected to the liquid
injection port and a 30 m · 0.25 mm · 0.50 lm Rtx-1
column connected to the head space injection port. Con-
ventional 70 eV electron impact (EI) ionisation was used
for structural identification.
proved yields of T cages can be obtained. We have also
shown that fluoride ion is a critical component of the reac-
FTICR-MS. All experiments were performed on a
Bruker Apex II FTICR-MS equipped with a 4.7 T
8

3296
A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
superconducting magnet and an external Analytica elec-
trospray ionisation (ESI) source. A Cole–Parmer series
stirred at room temperature for 24 h then the mixture
3
washed with water (3 · 50 cm ). The organic layer was
separated and a combination of white solid and yellow
oil obtained after removal of the solvent. This residue
7
4900 syringe pump was used to continuously infuse
samples into the ESI source. The instrument was oper-
ated in the positive-ion mode and data was acquired in
the broadband mode.
3
was extracted with acetone (5 cm ) to give a white solid
after filtration. Crystals suitable for X-ray structure
analysis were prepared by re-crystallising the solid from
a 1:1 dichloromethane/acetone solution. Yield: 0.13 g
4
.2. Preparation of TBAF solutions
ꢁ
1
(
17%); m.p. 144 ꢁC; mmax (/cm ) 2730, 2671, 1610,
The TBAF came from Aldrich Chemical Company
1580, 1513, 1305, 1253, 1194, 1150, 1023, 823, 771, 734
and 689; NMR: dH (300 MHz, CDCl ) 0.55 (16H, t,
Ltd. as a 1 M THF solution and contained ꢀ5% water.
From this a series of ÔstandardÕ solutions were prepared
in THF for use in our experiments:
3
SiCH ), 1.78 (16H, m, SiCH CH CH Ar), 2.47 (16H, t,
2
2
2
2
CH Ar), 3.69 (24H, s, CH ), 6.71 (16H, d, CH of Ar)
3
2
Solution A. The TBAF solution as purchased.
Solution B. A sample of Solution A was dried for 48 h
˚
over type 4 A molecular sieves pre-dried for 5 days at
and 6.93 (16H, d, CH of Ar); dC (75.5 MHz, CDCl3)
12.30 (SiCH2), 25.07 (SiCH CH CH Ar), 37.97
(CH Ar), 55.28 (CH ), 113.76 (CH of Ar), 129.36 (CH
2 3
2
2
2
3
00 ꢁC.
Solution C. To a sample of Solution B was added
of Ar), 134.50 (C of Ar) and 158.70 (C of Ar); dSi
1
8
O
O
(79.3 MHz, CDCl ) ꢁ66.77; m/z (MALDI-TOF):
3
+
1610.55 (100%, [M + H] ).
enriched water so as to make its content 5%.
Solution D. To a sample of Solution B was added
enriched water so as to make its content 20%.
1
8
4.3.3. Octa-iso-octyl-octasilsesquioxane 6
Iso-octyltrimethoxysilane (0.58 g, 2.48 mmol) and
Solution E. To a sample of Solution A was added a
0-fold excess of water.
3
TBAF Solution A (1.2 cm ) were dissolved in dichloro-
2
4
4
5
3
methane (50 cm ) and stirred for two days. The solution
3
was extracted with water (20 cm ) and further dichloro-
.3. Synthesis of octasilsesquioxane cages using TBAF
3
methane (200 cm ) and the organic phase separated and
.3.1. 3-para-Methoxyphenylpropyltriethoxysilane
Ethanol (2.59 g, 56.4 mmol) and triethylamine (5.7 g,
6.4 mmol) were dissolved in THF (100 ml), and a solu-
dried over magnesium sulfate. The solvent was then re-
moved under vacuum and the residue washed with ace-
3
tone (20 cm ) to give a colourless oil. The oil was
tion of 3-para-methoxyphenylpropyltrichlorosilane (5 g,
7.6 mmol) in THF (100 ml) added to the mixture very
purified by column chromatography (SiO /hexane) giv-
2
1
ing a single fraction of 6. Yield: 0.18 g (43%); NMR:
1
slowly. After 1 h stirring, a white solid precipitated out.
The mixture was filtered by vacuum and the filtrate col-
lected. A yellow liquid was obtained after removal of
solvent and was used at the next stage without further
dH (300 MHz, CDCl ) 0.58 (16H, d, J = 8.25 Hz,
3
HH
SiCH ), 0.89 (72H, s, C(CH ) ), 1.00 (24H, d,
3 3
2
1
1
JHH = 6.33 Hz, CHCH ), 1.14 (16H, d, J = 6.42
3
HH
Hz, CH C(CH ) ) and 1.30 (8H, m, CH); d (300
2
3 3
C
ꢁ
1
purification. Yield: 4.74 g (86.0%); mmax (Neat/cm
)
MHz, CDCl ) 12.36 (SiCH ), 23.54 (CHCH ), 24.97
3 2 3
2
1
1
983, 2931, 2886, 2842, 2768, 2723, 1610, 1573, 1513,
461, 1432, 1387, 1298, 1253, 1246, 1172, 1112, 1090,
(CH C(CH ) ), 30.13 (C(CH ) ), 30.93 (CH) and 53.90
2 3 3 3 3
(C(CH ) ); d (400 MHz, CDCl ) ꢁ68.20; m/z (MAL-
3
3
Si
3
+
DI-TOF): 1323.42 [M + H] .
038, 9499, 801 and 749; dH (300 MHz; CDCl ) 7.09
3
(
OCH ), 3.54 (6H, q, CH CH ), 2.37 (2H, quintet,
2H, d, CH of Ar), 6.81 (2H, d, CH of Ar), 3.82 (1H, s,
4.3.4. Octa-cyclopentyl-octasilsesquioxane 8
3
2
3
CH CH CH Si), 1.21 (3H, t, CH CH ) and 0.64 (2H, t,
2
This is a known compound and was first character-
ised by Barry et al. [18]. Therefore the experimental de-
tails presented here are only to demonstrate the various
reaction and workup conditions that resulted in the
crystals of the new solid state structures for this com-
pound that we have obtained.
2
2
2
3
CH Si); d (75.5 MHz; CDCl ) 157.61 (COCH ),
2
C
3
3
1
Ar), 58.23 (OCH CH ), 55.12 (CH O), 38.20 (CH Ar),
34.41 (CCH ), 129.30 (s, CH of Ar), 113.55 (CH of
2
2
3
3
2
2
4.95 (s, CH CH CH Si), 18.21 (CH CH ) and 9.97
2 2 2 2 3
(
(
CH Si); d (79.3 MHz; CDCl ) ꢁ45.42; m/z (EI) 313
2
Si
3
+
M ), 270, 256, 238, 227, 210, 196, 165, 121, 107, 91
and 77; Found: C, 61.53; H, 9.20%. C H SiO requires
4
4.3.4.1.
structure 8a. 2 (0.04 g, 0.05 mmol), alkoxysilane (0.03
Octa-cyclopentyl-octasilsesquioxane
crystal
1
6
28
C, 61.50; H, 9.20.
3
g, 0.16 mmol) and TBAF Solution A (0.03 cm ) were dis-
3
solved in dichloromethane (20 cm ) and stirred for 24 h.
4
.3.2. Octa-3-para-methoxyphenylpropyl-octasilsesquiox-
ane 5
-para-Methoxyphenylpropyltriethoxysilane (1.11 g,
3
The solution was then extracted with water (10 cm ) and
3
further dichloromethane (100 cm ) and the organic
3
3
3
TBAF Solution A (1.77 cm ) added. The mixture was
.54 mmol) was dissolved in chloroform (50 cm ) and
3
phase dried over magnesium sulfate. The solvent was
then removed under vacuum and the residue washed

A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
3297
3
with acetone (20 cm ) to give 8 as a white solid. The so-
30 min and was characterised from its FTICR mass
spectrum.
lid was recrystallised from a dichloromethane/acetone
mixture to give colourless crystals of 8a suitable for X-
ray structure analysis.
4.4.4. Reaction (d)
Cyclopentyltriethoxysilane (0.092 g/0.40 mmol) was
3
dissolved in dry acetone (2 cm ) and TBAF Solution A
4
structure 8b. 2 (0.16 g, 0.19 mmol), alkoxysilane (0.10
.3.4.2.
Octa-cyclopentyl-octasilsesquioxane crystal
(
100 ll) added. The mixture was stirred to give a white
solid product of 7 which precipitated from solution after
0 min and was characterised from its FTICR mass
3
g, 0.40 mmol) and TBAF Solution A (0.10 cm ) were dis-
3
solved in dichloromethane (50 cm ) and stirred for 24 h.
3
3
The solution was then extracted with water (20 cm ) and
spectrum.
3
dichloromethane (200 cm ) and the organic phase dried
over magnesium sulfate. The solvent was then removed
under vacuum and the residue washed with acetone (20
3
cm ) to give a colourless gel. Crystals of 8b suitable for
4.4.5. Reaction (e)
Cyclopentyltriethoxysilane (0.090 g/0.39 mmol) was
3
dissolved in dry acetone (2 cm ) and TBAF Solution D
X-ray structure analysis were prepared by crystallising
the material from a 1:1 dichloromethane/acetone
solution.
(100 ll) added. The mixture was stirred to give a white
solid product of 7 which precipitated from solution after
30 min and was characterised from its FTICR mass
spectrum.
4
structure 8c. 2 (0.20 g, 0.25 mmol), alkoxysilane (0.12
.3.4.3.
Octa-cyclopentyl-octasilsesquioxane crystal
4.4.6. Reaction (f)
Cyclohexyltriethoxysilane (0.096 g/0.39 mmol) was
3
g, 1.02 mmol) and TBAF Solution A (0.13 cm ) were dis-
3
solved in dichloromethane (50 cm ) and stirred for 24 h.
3
dissolved in dry acetone (2 cm ) and TBAF Solution A
3
The solution was then extracted with water (20 cm ) and
(100 ll) added. The mixture was stirred to give a white
solid product of 7 which precipitated from solution after
30 min and was characterised from its FTICR mass
spectrum.
3
dichloromethane (200 cm ) and the organic layer dried
over magnesium sulfate. The solvent was then removed
under vacuum leaving a residue containing predomi-
2
9
nantly 8. Following Si NMR analysis of this residue
dissolved in deuteriochloroform, crystals of 8c suitable
for X-ray crystal structure analysis were formed inside
the NMR tube and were separated by decanting the
solution.
4
.4.7. Reaction (g)
Cyclohexyltriethoxysilane (0.099 g/0.40 mmol) was
3
dissolved in dry acetone (2 cm ) and TBAF Solution D
(100 ll) added. The mixture was stirred to give a white
solid product of 7 which precipitated from solution after
30 min and was characterised from its FTICR mass
spectrum.
1
8
4
.4. O labelling experiments
4
.4.1. Reaction (a)
Cyclopentyltriethoxysilane (0.089 g/0.38 mmol) was
4.5. Preparation of disiloxanes from monochlorosilanes by
hydrolysis using the water contained in a TBAF solution
3
dissolved in dry acetone (2 cm ) and TBAF Solution A
(
100 ll) added. The mixture was stirred to give a white
solid product of 7 which precipitated from solution after
0 min and was characterised from its FTICR mass
spectrum.
The siloxanes prepared are known compounds and
characterised only from the EI mass spectra of species
of interest immediately following GC separation.
3
4
.4.2. Reaction (b)
Cyclopentyltriethoxysilane (0.087 g/0.37 mmol) was
4.5.1. Hydrolysis of (chloro)methyldiphenylsilane
(Chloro)methyldiphenylsilane (0.13 g/0.52 mmol) was
3
dissolved in dry acetone (2 cm ) and after TBAF Solu-
3
dissolved in dry acetone (2 cm ) and TBAF Solution C
(
188 ll) added. The mixture was stirred to give a white
solid product of 7 which precipitated from solution after
0 min and was characterised from its FTICR mass
spectrum.
tion D (200 ll) was added the solution was stirred for
24 h. The solvent was then removed under vacuum
and the remaining oily residue characterised without
further purification. Isolated:
3
a: 1,3-Dimethyl-1,1,3,3-tetraphenyldisiloxane: m/z (EI):
+
+
4
.4.3. Reaction (c)
Cyclopentyltriethoxysilane (0.089 g/0.38 mmol) was
397 (100%, [MO18 ꢁ Me] ), 395 (34%, [M
ꢁ Me] ),
O16
+
319 (64%, [M ꢁ Me ꢁ Ph] ), 317 (23%, [M
ꢁ
O18
O16
3
dissolved in dry acetone (2 cm ) and TBAF Solution D
+
Me ꢁ Ph] ).
(100 ll) added. The mixture was stirred to give a white
solid product of 7 which precipitated from solution after
b: 1-Fluoro-1,3-dimethyl-1,3,3-triphenyldisiloxane: m/z
+
(EI): 339 (100%, [MO18 ꢁ Me] ), 337 (65%, [M
ꢁ
O16

Table 5
Crystal data, data collection and refinement parameters for compounds 5, 8a, 8b, 8c and 10
5
8a
8b
8c
10
Formula
C
80
H
104
O
20Si
8
C
48
H
88
O
12Si
8
C
H
48 88
O12Si
8
C
48
H
88
O
12Si
8
64 136 8
C H O12Si
Molecular weight
Colour
Morphology
1610.35
Colourless
Plate
1081.90
Colourless
Block
1081.90
Colourless
Block
1081.90
Colourless
Block
1322.45
Colourless
Plate
Crystal size (mm)
Crystal system
Space group
0.07 · 0.05 · 0.02
Triclinic
0.25 · 0.20 · 0.15
Rhombohedral
0.25 · 0.20 · 0.15
Triclinic
0.20 · 0.20 · 0.15
Tetragonal
P4/n
0.20 · 0.20 · 0.15
Triclinic
ꢀ
ꢀ
ꢀ
ꢀ
P1
R3
P1
P1
Unit cell dimensions
˚
a (A)
˚
9.2155(2)
20.0378(4)
22.6830(5)
81.137(2)
83.650(2)
81.934(2)
4080.66(15)
2
16.6048(8)
16.6048(8)
17.3069(7)
90
17.0886(8)
17.8454(8)
25.4512(15)
77.133(2)
71.658(2)
88.819(2)
7171.85(16)
5
15.7365(2)
15.7365(2)
12.5643(4)
90
8.5186(4)
9.1418(4)
25.8700(13)
80.073(2)
82.611(2)
84.185(3)
1961.52(16)
1
b (A)
˚
c (A)
a (ꢁ)
b (ꢁ)
c (ꢁ)
90
120
90
90
˚
3
V (A )
4132.57(13)
3
1.304
3111.39(11)
2
1.155
Z
3
D
calc (g cm )
1.311
1712
1.252
2920
1.120
728
F(0 0 0)
1752
1168
Temperature (K)
Reflections collected
Independent reflections
150(2)
25 601
12 287
120(2)
9148
1631
120(2)
72 647
22 578
120(2)
15 246
2737
120(2)
15 392
6614
R
int
0.0764
2.95–23.75
ꢁ9, 10
0.1264
3.07–25.02
ꢁ19, 15
ꢁ16, 19
ꢁ19, 20
0.3726
0.0611
3.05–25.02
ꢁ18, 18
ꢁ18, 18
ꢁ11, 14
0.0733
2.99–25.03
ꢁ9, 10
h Range (ꢁ)
h Range
k Range
l Range
2.99–25.03
ꢁ20, 20
ꢁ20, 21
ꢁ27, 28
ꢁ20, 22
ꢁ25, 25
ꢁ10, 10
ꢁ30, 30
2
2
Final R indices [F > 2r(F )]
R indices (all data)
2
S (goodness-of-fit on F )
R
R
1
= 0.0978, wR
= 0.2173, wR
2
= 0.1750
= 0.2006
R
R
1
= 0.1217, wR
= 0.2528, wR
2
= 0.3100
= 0.3948
R
R
1
= 0.1680, wR
= 0.4744, wR
2
2
= 0.3719
= 0.4975
R
R
1
= 0.0553, wR
= 0.0553, wR
2
= 0.1653
= 0.1653
R
R
1
= 0.0705, wR
= 0.1468, wR
2
= 0.1792
= 0.2230
1
2
1
2
1
1
2
1
2
1.165
12 287
900
1.111
1631
118
1.046
22 578
1531
0.793
2737
154
0.955
6614
419
N
N
ref
par

A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
3299
+
Me] ), 261 (38%, [M
+
ꢁ Me ꢁ Ph] ), 259 (23%,
5. Supplementary data
O18
+
[
MO16 ꢁ Me ꢁ Ph] ).
Crystallographic data for the structures reported in
this paper has been deposited with the Cambridge Crys-
tallographic Data Centre as Supplementary Publication
Nos. CCDC 216990–216994. Copies of the data can be
obtained free of charge on application to CCDC, 12 Un-
ion Road, Cambridge, CB2 1EZ, UK, fax: (+44) 1223
4
.5.2. Hydrolysis of chlorotrimethylsilane
Chlorotrimethylsilane (0.05 g/0.46 mmol) was dis-
3
solved in dry acetone (2 cm ) and after TBAF Solution
D (200 ll) was added the solution was stirred for 24 h.
The solvent was then removed under vacuum and the
remaining oily residue characterised without further
purification. Isolated:
336 033. Email: data_request@ccdc.cam.ac.uk.
Hexamethyldisiloxane: m/z (EI): 149 (100%,
+
+
Acknowledgement
[
MO18 ꢁ Me] ), 147 (49%, [M
ꢁ Me] ).
O16
We thank Dow Corning for their support of Y.Y.
4
.6. Synthesis of 7 under ÔdryÕ and ÔwetÕ TBAF conditions
4
.6.1. ÔDryÕ conditions
Cyclopentyltriethoxysilane (0.05 g/0.19 mmol) was
References
3
dissolved in deuterioacetone (1.5 cm ) in a septa-sealed
NMR tube and TBAF Solution B (50 ll) added. The
mixture was briefly shaken, allowed to stand for 24 h
and then analysed by Si NMR spectroscopy. The sig-
nificant chemical shift results are reported in Section 2.
[1] I.M. Saez, J.W. Goodby, Liquid Cryst. 26 (1999) 1101.
[2] R. Elsaber, G.H. Mehl, J.W. Goodby, D.J. Photinos, J. Chem.
Soc., Chem. Commun. (2000) 851.
[3] F.J. Feher, K.D. Wydham, M.A. Sciadone, Y. Hamuro, J. Chem.
Soc., Chem. Commun. (1998) 1469.
2
9
[
[
4] F.J. Feher, J. Am. Chem. Soc. 108 (1986) 3850.
5] F.J. Feher, R.L. Blanski, J. Chem. Soc., Chem. Commun. (1990)
1
614.
6] F.J. Feher, K.D. Wyndham, J. Chem. Soc., Chem. Commun.
1998) 323.
4
.6.2. ÔWetÕ conditions
The reaction was performed in the same manner and
[
(
with the same analysis method as for the ÔdryÕ reaction
but using TBAF Solution A (50 ll) supplemented by
water (100 ll). The significant chemical shift results
are reported in Section 2.
[7] A.R. Bassindale, T.E. Gentle, J. Mater. Chem. 12 (1993) 1319.
[8] P.A. Agaskar, Inorg. Chem. 30 (1991) 2707.
[9] J.F. Brown, L.H. Vogt, J. Am. Chem. Soc. 87 (1965) 4313.
[
[
[
10] H. Behbehani, B.J. Brisdon, M.F. Mahon, K.C. Molloy, J.
Organometal. Chem. 469 (1994) 19.
11] F.J. Feher, D.A. Newman, J.F. Walzer, J. Am. Chem. Soc. 111
(
1989) 174.
4
.7. Synthesis of 7 using TBAF in different solvents
12] A.R. Bassindale, Z. Liu, I.A. MacKinnon, P.G. Taylor, Y. Yang,
M.E. Light, P.N. Horton, M.B. Hursthouse, J. Chem. Soc.,
Dalton Trans. (2003) 2945.
These reactions were carried out by the standard
method [12] using TBAF Solution A and the appropriate
3
solvent (30 cm ).
[
[
13] A.R. Bassindale, Z. Liu, D.J. Parker, P.G. Taylor, P.N. Horton,
M.B. Hursthouse, M.E. Light, J. Organometal. Chem. 687
(
2003) 1.
14] R. Maurer, F.H. Kreuzer, P. Spes, Makromol. Chem. Macromol.
Symp. 30 (1991) 215.
4
.8. X-ray crystallography
[
[
15] J.W. Goodby, G.H. Mehl, Chem. Ber. 129 (1996) 521.
16] G.H. Mehl, J.W. Goodby, Angew. Chem., Int. Ed. Engl. 35
(1996) 2641.
All structures were measured on a Nonius Kappa
CCD diffractometer (/ scans and x scans to fill asym-
metric unit or Ewald sphere) using graphite monochro-
[
[
17] G.H. Mehl, J.W. Goodby, I.M. Saez, R.P. Tuffin, G.H. Macken-
zie, R. Auzely-Velty, T. Benvegru, D. Plusquellec, J. Chem. Soc.,
Chem. Commun. (1998) 2057.
˚
mated Mo Ka radiation (k = 0.71073 A). The data
collection was performed in a w rotation mode. A list
of crystal data, data collection, structure solution and
refinement parameters is provided in Table 5. The raw
intensity data was corrected for Lorentz and polarisa-
18] A.J. Barry, W.H. Daudt, J.J. Domicons, J.W. Gilkey, J. Am.
Chem. Soc. 77 (1955) 4248.
[19] N.V. Podberezskaya, I.A. Baidina, V.I. Alekseev, S.V. Borisov,
T.N. Martynova, J. Struct. Chem. (Engl. Transl.) 22 (1981)
7
37.
20] P.-A. Jaffr e` s, R.E. Morris, J. Chem. Soc., Dalton Trans. (1998)
767.
tion effects as well as absorption correction (SORTAV)
27]. Structure solutions were searched and refined by
[
[
[
[
[
[
Direct Methods [28] (SHELXS 97) and Difference Fourier
2
21] C. Bonhomme, P. Toledano, J. Maquet, J. Livage, L. Bon-
homme-Coury, J. Chem. Soc., Dalton Trans. (1997) 1617.
22] I. Richter, C. Burschka, R. Tacke, J. Organomet. Chem. 646
2
Analyses (SHELXL 97) [29], respectively, based on F .
Atom form factors for neutral atoms were taken from
the literature [30]. Second and third row elements were
allowed to refine anisotropically with hydrogen atoms
assumed in idealised positions riding on their pivot
atoms.
(
2002) 200.
23] U. Dittmar, B.J. Hendan, U. Fl o¨ rke, H.C. Marsmann, J.
Organomet. Chem. 489 (1995) 185.
24] F.J. Feher, T.A. Budzichowski, J. Organomet. Chem. 373 (1989)
153.

3300
A.R. Bassindale et al. / Journal of Organometallic Chemistry 689 (2004) 3287–3300
[
25] N.V. Podberezskaya, S.A. Magarill, I.A. Baidina, S.V. Borisov,
´
L.E. Gorsh, A.N. Kanev, T.N. Martynova, J. Struct. Chem.
[28] G.M. Sheldrick, SHELXS 97_2. A Program for Crystal Structure
Solution, University of G o¨ ttingen, FRG, 1998.
(
Engl. Transl.) 23 (1982) 422.
[29] G.M. Sheldrick, SHELXL 97_2. A Program for Crystal Structure
Refinement, University of G o¨ ttingen, FRG, 1998.
[30] International Tables for Crystallography, vol. C, Tables 4.2.6.8
and 6.1.1.4, Kluwer Academic Publishers, Dordrecht, The Neth-
erlands, 1995.
[
[
26] M.A. Hossain, M.B. Hursthouse, K.M.A. Malik, Acta Crystal-
logr. B35 (1979) 2258.
27] (a) R.H. Blessing, Acta Crystallogr. A51 (1995) 33;
(b) R.H. Blessing, J. Appl. Crystallogr. 30 (1997) 421.