Studies of octameric vinylsilasesquioxane by carbon-13 and siIicon-29 cross polarization magic angle spinning and inversion recovery cross polarization nuclear magnetic resonance spectroscopy

Article abstract of DOI:10.1039/a700700k

The acidic hydrolysis of triethoxyvinylsilane Si(C2H3)(OEt)3 led to octavinyloctasilasesquioxane (C2H3SiO,j)81 . and to a transparent derived gel 2. Compound 1 was obtained as single crystals, suitable for X-ray analysis: trigonal, space group β3, Z = 3, a = 13.533(2), c = 14.222(2) A, y = 120°. Disorder of the methylene groups was observed. Compound 1 was carefully studied by IR spectroscopy and 29Si/13C cross polarization magic angle spinning (CP MAS) NMR spectroscopy. The CP dynamics of the various I3C sites were analysed by inversion recovery cross polarization (IRCP). Very different behaviours were observed for CH and CH₂ signals and could be attributed to anisotropic reorientations of the vinyl groups. The IRCP sequence enabled a complete editing of the spectra and definite assignments of the observed lines. Owing to its highly resolved spectra and small intrinsic linewidths, it appears that compound 1 could act as a secondary reference for the set-up of the Hartmann-Hahn condition in 29Si CP NMR spectroscopy as well as a test for the set-up of the magic angle. All these results were applied to the study of the amorphous gel 2. Dipolar coupling constants were also strongly affected by local motions. .

Full text of DOI:10.1039/a700700k

View Online / Journal Homepage / Table of Contents for this issue
Studies of octameric vinylsilasesquioxane by carbon-13 and silicon-29
cross polarization magic angle spinning and inversion recovery cross
polarization nuclear magnetic resonance spectroscopy
Christian Bonhomme,*,a Paul Tolédano,a Jocelyne Maquet,a Jacques Livagea and
Laure Bonhomme-Coury†,b
a
Laboratoire Chimie de la Matière Condensée, Université Paris 6, 4 Place Jussieu,
75252 Paris Cedex 05, France
b Laboratoire Céramiques et Matériaux Minéraux, Ecole Supérieure de Physique et Chimie
Industrielles de la Ville de Paris, 10 rue Vauquelin, 75231 Paris Cedex 05, France
The acidic hydrolysis of triethoxyvinylsilane Si(C2H3)(OEt)3 led to octavinyloctasilasesquioxane (C2H3SiO1.5)8 1
and to a transparent derived gel 2. Compound 1 was obtained as single crystals, suitable for X-ray analysis:
¯
trigonal, space group R3, Z = 3, a = 13.533(2), c = 14.222(2) Å, γ = 120Њ. Disorder of the methylene groups was
observed. Compound 1 was carefully studied by IR spectroscopy and 29Si/13C cross polarization magic angle
spinning (CP MAS) NMR spectroscopy. The CP dynamics of the various 13C sites were analysed by inversion
recovery cross polarization (IRCP). Very different behaviours were observed for CH and CH2 signals and could be
attributed to anisotropic reorientations of the vinyl groups. The IRCP sequence enabled a complete editing of the
spectra and definite assignments of the observed lines. Owing to its highly resolved spectra and small intrinsic
linewidths, it appears that compound 1 could act as a secondary reference for the set-up of the Hartmann–Hahn
condition in 29Si CP NMR spectroscopy as well as a test for the set-up of the magic angle. All these results were
applied to the study of the amorphous gel 2. Dipolar coupling constants were also strongly affected by local
motions.
During the last few years, spherosiloxanes, including octameric
silasesquioxanes (RSiO1.5)8, have been extensively studied. Sev-
eral review articles have been published.1–4 Special attention was
paid to syntheses of new silasesquioxane structures5,6 including
heterometallic silasesquioxanes.2,7,8 This class of compounds
has been used as models in different fields such as silicate/
zeolite chemistry, catalytic and analytical chemistry, etc. There-
fore, silasesquioxanes have been studied by different spectro-
scopic techniques, i.e. X-ray diffraction,9–16 neutron diffrac-
tion,17 vibrational spectroscopies14,18 including normal
coordinate analysis, liquid-state (1H, 29Si, 13C and 17O)
NMR,6,19,20 etc. A few crystalline silasesquioxanes were studied
by 29Si and 13C solid-state NMR spectroscopy, including the
‘Q8M8’ derivative (Me3SiOSiO1.5)8, which is now widely used
for the set-up of 29Si cross polarization magic angle spinning
(CP MAS) NMR.21,22 Moreover, interest in the use of these
compounds as precursors for ceramics and macromolecular
materials, by bridging organofunctional silasesquioxanes, has
grown significantly very recently.4,23–25
cross polarization (IRCP) sequence can be compared to the well
known editing sequences in liquid-state NMR, e.g. distortion-
less enhancements by polarization transfer (DEPT) and insensi-
tive nuclei enhanced by polarization transfer (INEPT).29 Such
editing techniques could act as invaluable tools of investigation
for bridged-silasesquioxane amorphous materials. Indeed, the
determination of the proton multiplicities of 13C sites is crucial
for the complete characterization of these compounds. More-
over, the study of the CP dynamics allowed us to investigate
local motion of the vinyl groups and to propose simple geo-
metrical explanations for the astonishing CP behaviour of
adjacent CH and CH2 groups. The results were applied to the
study of an amorphous gel (compound 2) containing mostly
3
᎐
vinyl T units(*), i.e. fully condensed C H Si*(OSi᎐) entities.
᎐
2
3
3
Results and Discussion
Syntheses
Vinyl derivatives and heterofunctional compounds have been
widely investigated by Martynova and co-workers26,27 including
X-ray diffraction studies.10 In this paper we present a complete
study of the ‘cubane shaped’ octavinyloctasilasesquioxane
(C2H3SiO1.5)8 1 by X-ray diffraction, infrared and 29Si/31C solid-
state CP MAS NMR spectroscopy; 29Si CP MAS NMR spec-
troscopy was used to extract isotropic chemical shifts as well as
chemical shift anisotropies (CSAs). Few CSA values related to
small organosilicon molecules are available in the literature.
Such CSAs may give information on the electronic environment
around 29Si nuclei, local structure and molecular motion if
present. The CP MAS NMR technique was not only used for
enhancing the nuclear spin magnetization,28 but was also suc-
cessfully used to edit the spectra, i.e. to assign without ambi-
guity the observed 13C resonances. The inversion recovery
Synthesis of (C2H3SiO1.5)8 1 and the derived gel 2 are described
in the Experimental section. Hydrolysis under acidic conditions
and subsequent condensation of an organically modified silicon
alkoxide, triethoxyvinylsilane Si(C2H3)(OEt)3 led to: (i) crystals
(compound 1) with only fully condensed T3 units and (ii) gels
(compound 2) containing a priori T0–3 units(*), C2H3Si*X3Ϫx
-
᎐
(OSi᎐) (X = OH or OEt, x = 0–3).
᎐
x
Hydrolytic polycondensation of organotrichloro- (or
methoxy-) silanes seems to be the main route of synthesis: Mar-
tynova and Chupakhina 27 obtained octaorganooctasilasesqui-
oxanes containing both methyl and vinyl groups by hydrolysis of
SiMeCl3 and Si(C2H3)Cl3; Hendan and Marsmann 19 presented
the synthesis of mixed ethyl and vinyl derivatives starting from
SiEtCl3 and Si(C2H3)Cl3.
The use of modified silicon alkoxides SiR(ORЈ)3 (R = H or
organic group, RЈ = organic group) may open new synthetic
routes to hydrido- and/or organo-silasesquioxanes. In particular,
† E-Mail: laure.bonhomme@espci.fr
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626
1617

View Online
Table 1 Selected bond lengths (Å) and angles (Њ) with estimated
standard deviations in parentheses for C16H24Si8O12 1. Symmetry codes
as in Fig. 1
Si(1)᎐O(1)
Si(2)᎐O(1)
Si(2)᎐O(2)
Si(2)᎐O(2V)
Si(1)᎐C(1)
Si(2)᎐C(2)
1.606(6)
1.596(6)
1.596(6)
1.616(6)
1.81(2)
C(1)᎐C(11)
C(2)᎐C(21)
C(2)᎐C(22)
C(2)᎐C(23)
C(2)᎐C(24)
1.25(3)
1.22(3)
1.26(3)
1.23(4)
1.27(4)
1.82(1)
O(1)᎐Si(1)᎐O(1II)
O(1)᎐Si(2)᎐O(2)
O(1)᎐Si(2)᎐O(2V)
O(2)᎐Si(2)᎐O(2V)
O(1)᎐Si(1)᎐C(1)
O(1)᎐Si(2)᎐C(2)
O(2)᎐Si(2)᎐C(2)
O(2V)᎐Si(2)᎐C(2)
108.5(3)
108.5(3)
108.7(4)
108.6(3)
110.5(3)
110.0(5)
111.4(5)
109.7(5)
Si(1)᎐O(1)᎐Si(2)
Si(2)᎐O(2)᎐Si(2IV)
Si(1)᎐C(1)᎐C(11)
Si(2)᎐C(2)᎐C(21)
Si(2)᎐C(2)᎐C(22)
Si(2)᎐C(2)᎐C(23)
Si(2)᎐C(2)᎐C(24)
150.5(4)
150.0(4)
127(2)
128(2)
121(2)
132(3)
125(3)
the choice of the experimental parameters (pH, temperature,
solvent, concentration, nature of RЈ) allows a partial control of
the hydrolysis–condensation process.30 Hydrolysis of such
alkoxides leads to polymeric gels, which can act as ceramic pre-
cursors.4 In this context, the formation of crystalline phases
(such as compound 1) from media which usually lead to
amorphous gels is a very interesting starting point for spectro-
scopic studies of such gels. This could be especially valuable for
NMR studies of amorphous or poorly crystallized materials,
which are extensively investigated by 13C/29Si/1H solid-state
NMR techniques.31,32 In the case of compound 1 one unique
crystalline phase is obtained, in contrast with the synthesis pro-
posed by Baidina et al.10 [i.e. direct hydrolysis of Si(C2H3)Cl3]
leading to octa- and deca-silasesquioxanes (C2H3SiO1.5)n with
n = 8 or 10. However, in our case, the reaction yield for com-
pound 1 remains poor.
Fig. 1 An ORTEP 33 view of the structure of C16H24Si8O12 1, showing
the atom labelling scheme. Of the second vinyl carbons only C(11) and
C(21) are represented as well as their bonded hydrogens (calculated
positions). The CH2 positions are disordered (see text). Symmetry
codes: I x¯, y¯, z¯; II y¯, x Ϫ y, z; III y Ϫ x, x¯, z; IV y, y Ϫ x, z¯; V x Ϫ y, x, z¯
C(2)᎐C angles have a mean value of 127(3)Њ, in agreement with
the Si(1)᎐C(1)᎐C(11) angle [127(2)Њ]. All C᎐C bond lengths are
᎐
close to 1.25 Å (significantly shorter than usual, i.e. 1.33 Å).
These observations justify the fact that positions C(21)–C(24)
correspond to methylene sites but they are surely estimations.
One unique C(21) position (occupancy = 1) cannot be defined.
There are two non-equivalent distances between centro-
symmetrically related silicon atoms within the molecule.
Opposite Si atoms in general positions are 5.371(5) Å apart
and those on the C3 axis are 5.360(1) Å apart. Such a slightly
distorted cubane structure is accompanied by the absence of
Si᎐O intermolecular contacts less than 3.7 Å and a low crystal
density (Dc = 1.398 g cmϪ3). An increase in the number of
short intermolecular Si᎐O contacts led to more important dis-
tortions of the cubane cores as well as to an increase in crystal
density.14 Larsson 9 showed also that the crystal density
decreased from 1.51 g cmϪ3 for (MeSiO1.5)8 to 1.09 g cmϪ3 in
(PrnSiO1.5)8.
Crystal structure of compound 1
The molecular structure of compound 1, C16H24Si8O12, is
shown in Fig. 1. Only the second vinyl carbons C(11) and
C(21) are represented as the structure is disordered (see
below). Selected bond lengths and angles are presented in
Table 1. The main feature of the octameric cluster is the
presence of a C3 axis containing atoms C(1) and Si(1). Dis-
order at the CH2 sites is present. Atom C(11) is bonded to
C(1) (located on the C3 axis). Four positions for carbon
atoms bonded to C(2) were found, corresponding to C(21)–
C(24). Fractional occupancies of C(11) and C(21)–C(24) were
refined for the structure solution. The occupancy of C(11)
was found to be ¹. Such a characteristic has been previously
The main characteristic of compound 1 is the non-
equivalence of the vinyl groups. The C(11) positions are related
by C3 symmetry, whereas C(21)–C(24) are non-symmetry-
related. As X-ray diffraction is unable to distinguish between
static and dynamic disorder the vinyl groups are either dis-
ordered or in motion.
¯
³
clearly evidenced by Larsson,9 in the case of isomorphous
homologues belonging to the hexagonal system and showing
space group R3 [e.g. (RSiO1.5)8, R = Me, Et, Prn and Pri]. This
¯
has been also observed by Baidina et al.10 in the case of the
Vibrational spectroscopy
¯
octavinyl derivative (trigonal, space group R3, a = 13.584,
c = 14.364 Å, final R = 0.11). However, several differences
between the molecular structure of the octavinyl derivative
described in ref. 10 and compound 1 have to be noted. First,
the Si᎐C bond lengths are shorter in our work than in ref. 10
to a rather important extent: Si(1)᎐C(1) and Si(2)᎐C(2) 2.03 Å
in ref. 10, whereas we found 1.81(2) and 1.82(1) Å, respect-
ively. Secondly, two distinct positions (with site occu-
pancy ≈ ¹) seem to be sufficient to describe the methylene car-
bon atom in the general position in ref. 10, whereas, in the
case of compound 1 four disordered methylene positions
C(21)–C(24) were found with site occupancies varying from
0.15 to 0.39 as estimations.
Compounds 1 and 2 were analysed by infrared spectroscopy.
Assignments of the main bands are given in Table 2, based on
data from the literature.14,31,34 The IR spectrum of neat
Si(C2H3)(OEt)3 was also used as a reference.
The IR spectrum of compound 1 is characterized by a very
strong band centred at 1113 cmϪ1, assigned to νasym(Si᎐O᎐Si).
This band is broad, compared to the other vibrations
(∆w ≈ 100 cmϪ1). It is interesting that the other bands assigned
₁
₂

¯
²
to specific vibrations of the cubic siloxane cage (ν р 780 cmϪ1
)
are also rather broad, when compared to bands involving only
the organic part of the molecule. This may reflect the small
distortions of the cubane cage, lowering the symmetry of the
entity from Oh. The band centred at 780 cmϪ1 has a shoulder
(756 cmϪ1) which could indicate the presence of two Si᎐C
bonding types, as already seen by X-ray analysis.
The Si᎐O and Si᎐C bond lengths as well as O᎐Si᎐O, Si᎐O᎐Si
and O᎐Si᎐C angles are in accord with values observed for well
characterized octameric silasesquioxanes.11,12,14 The Si(2)᎐
1618
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626

View Online
Table 2 Selected infrared absorptions (cmϪ1) of C16H24Si8O12 1 and the
gel 2
1
2
Assignment*
3067w
3026w
1604m
1409m
3063w
3024w
1603m
1409m
1278m
νasym(CH2)
ν(CH)
ν(C᎐C)
᎐
δ(CH2) in plane
δ(CH) in plane
1277m
1152 (sh)
1113vs (br)
1129vs (br)
1043s (br)
1005m
963m
762m (br)
νasym(Si᎐O᎐Si)
1005m
970m
780m
δ(CH) out of plane
δ(CH2) out of plane
ν(Si᎐C)
756 (sh)
585s
Fig. 3 The 13C CP MAS NMR spectrum of compound 1 (a), with



587m (br)
546m
428m (br)
δ(O᎐Si᎐O)
ϩδ(Si᎐C᎐C)
simulations of CH (b) and CH2 (c) patterns. Isotropic resonances are
labelled by arrows: νrot = 1200 Hz, Ξ (13C) = 75.46 MHz, φ = 7 mm,
Ns = 104, tCP = 5 ms, r.d. = 15 s, l.b. = 3 Hz. * Satellites corresponding to
1J(13C᎐29Si)
᎐
465w
νsym(Si᎐O᎐Si)
* According to refs. 14, 31 and 34. w = Weak, m = medium, s = strong,
br = broad, sh = shoulder, v = very, sym = symmetric, asym =
asymmetric.
densed (i.e. it contains mainly T3 units) as has been shown by
solid-state 13C NMR spectroscopy (see below).
Solid-state NMR spectroscopy
Compound 1. Standard 29Si and 13C CP MAS experiments.
Spectra of compound 1 obtained by 29Si and 13C CP MAS
experiments (at various rotation speeds) are presented in Figs. 2
and 3, respectively. All NMR data, including isotropic chemical
shift (29Si and 13C), shielding tensor principal components,
linewidths, relative intensities and assignments are presented in
Table 3.
Few data concerning solid-state 29Si NMR results for silases-
quioxanes are available. Most of the 29Si chemical shift data were
obtained in solution including sophisticated two-dimensional
experiments [i.e. incredible natural abundance double quantum
transfer experiment (INADEQUATE)19] and leading to correl-
ations between chemical shifts and geometrical parameters.6
Hoebbel and co-workers22,35,36 studied several octa- and deca-
silasesquioxanes by means of solid-state techniques but pub-
lished only isotropic chemical shift values. The determination
of the principal components of the tensor σ (chemical shift
tensor) was mainly devoted to the structure elucidation of silic-
ate minerals.37 To our knowledge, few papers dealing with shield-
ing tensors of molecular organosilicon compounds have been
published so far. Harris et al.38 presented 29Si NMR studies of
various polysilanes, exhibiting different silicon environments.
Compound 1 is characterized by two isotropic 29Si peaks
[Fig. 2(a) and Table 3] characteristic of T3 units. The intensities
of the lines are in the ratio 3:1 in good agreement with the
presence of three equivalent Si(2) atoms to one Si(1) located on
the C3 axis. This allows definite assignments for both reson-
ances. It should be noted that a small shielding of Si atoms
located on the C3 axis was also observed in the case of the
octamethyl cubane derivative (MeSiO1.5)8:22 δSi(1) Ϫ66.5 and
δSi(2) Ϫ65.9. These line intensities were obtained for a contact
time of 10 ms, corresponding to an optimized value. An
important experimental detail can be added: the Hartmann–
Hahn condition was set up using compound 1, maximizing the
intensities of both lines. In addition, another test for the
Hartmann–Hahn set-up was the resolution of both lines (with
small intrinsic linewidths, ≈8–10 Hz), which was maximal at
exact Hartmann–Hahn condition. In other words, compound
1 can be safely used as a standard for the set-up of CP
experiments, a well resolved free induction decay (FID) being
observed after eight scans (recycle delay = 15 s). At intermed-
iate MAS rate (νrot = 1000 Hz, not shown here) two sets of spin-
ning sidebands (with low intensities) were observed. Definitions
of the shielding anisotropy (∆σ) and asymmetry (η) are given in
Table 3. The anisotropy was estimated to be ϩ35 ppm for both
Fig. 2 The 29Si CP MAS NMR spectra of compound 1: (a) ‘high’
rotation speed, νrot (rotation speed) = 4000 Hz, Ξ (29Si) (Larmor fre-
quency) = 59.62 MHz, φ (diameter of rotor) = 4 mm, Ns (number of
scans) = 8, tCP (contact time) = 10 ms, r.d. (recycle delay) = 15 s, l.b. (line
broadening) = 0 Hz; (b) ‘intermediate’ rotation speed, νrot = 500 Hz, Ξ
(29Si) = 79.50 MHz, φ = 7 mm, Ns = 200, tCP = 10 ms, r.d. = 15 s, l.b. = 3
Hz, isotropic resonances labelled by arrows [(d) simulation of sideband
patterns]; (c) static experiment, νrot = 0 Hz, Ξ (29Si) = 59.62 MHz, φ = 7
mm, Ns = 400, tCP = 10 ms, r.d. = 15 s, l.b. = 20 Hz [(e) simulations of
Si(1) and Si(2) powder patterns convoluted by a gaussian broadening]
The IR spectrum of the xerogel 2 is very similar to that for 1.
All the bands (ν < 2000 cmϪ1) already assigned to vinyl vibra-
tions for 1 are observed as five sharp signals centred at 1603,
1409, 1278, 1005 and 963 cmϪ1. However, bands involving Si᎐O
and Si᎐C vibrations are broadened, leading to a less resolved
spectrum at low frequency. Such a broadening can be safely
related to more disordered and/or distorted environments
around Si atoms (compound 2 is amorphous to X-rays). Vibra-
tions related to vinyl groups are much less sensitive to this dis-
order. Although not quantitative, this spectrum suggests that
the xerogel contains very few residual ethoxy groups: no clear
vibration is observed in the range 1350–1500 cmϪ1 (relative to
aliphatic CH2 and CH3 deformations). This gel is highly con-
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626
1619

View Online
Table 3 Silicon-29 and 13C CP NMR data for compounds 1 and 2 and cadmium acetylacetonate including isotropic shielding (σiso), shielding-tensor
components (σii), shielding anisotropy (∆σ), asymmetry (η), linewidth, relative intensities and assignmentsa
Compound
σiso
σ33
σ22
σ11
∆σ
δA
η
Linewidth/Hz Relative intensity (%)
Assignment
1 (29Si)
ϩ80.4
ϩ79.9
Ϫ138.0
Ϫ138.5
Ϫ128.8
Ϫ135
ϩ103.1
ϩ102.7
Ϫ216.6
Ϫ214.6
Ϫ186.2
Ϫ205
Ϫ184
Ϫ196.4
Ϫ194.8
ϩ68.9
ϩ71.8
Ϫ110.4
Ϫ119.4
Ϫ116.0
Ϫ118
Ϫ124
Ϫ55.1
Ϫ53.5
ϩ68.9
ϩ65.1
Ϫ92.0
Ϫ81.9
Ϫ84.5
Ϫ83
Ϫ80
Ϫ55.1
Ϫ53.5
ϩ34.2
ϩ34.2
Ϫ114.4
Ϫ114.0
Ϫ85.9
Ϫ105
Ϫ83
Ϫ141.3
Ϫ141.3
ϩ22.8
ϩ22.8
Ϫ74.3
Ϫ75.9
Ϫ57.3
Ϫ70
Ϫ55
Ϫ94.2
Ϫ94.4
0.00
0.30
0.20
0.50
0.55
0.5
0.8
0.00
0.00
8
10
18
18
15
280
225
31
25
75
13
32
55
≈50
≈50
50
Si(1)
Si(2)
C(11)
C(21)b
C(1)/C(2)
CH2
CH
CH
CH
(13C)
2 (13C)
Ϫ129
[Cd(acac)2] (13C) Ϫ102.1
Ϫ100.6
34
50
a
σ = Ϫδ (in ppm); σiso = ¹(σ11 ϩ σ22 ϩ σ33); |σ33 Ϫ σiso| у |σ11 Ϫ σiso| у |σ22 Ϫ σiso|; the errors in σii values are typically ±2 ppm; ∆σ = σ33 Ϫ ¹(σ11
ϩ
¯
³
¯
²
σ22) = ³δA; δA = σ33 Ϫ σiso; η = (σ22 Ϫ σ11)/δA; typical error in η is ±0.1. b Corresponding to the average value of σ[C(21)]–σ[C(24)].
¯
²
29Si sites. In order to get an accurate determination of (∆σ, η)
for each 29Si site, an experiment at much lower MAS rate
(νrot = 500 Hz) and at higher field [Ξ(29Si) = 79.5 MHz] was used
[Fig. 2(b)]. Assuming that the shielding is the only magnetic
influence on the intensities of the spinning sidebands, both pat-
terns were analysed by a standard graphical procedure39,40 [Fig.
2(d)]. Sets of (∆σ, η) values were (ϩ34.2 ppm, ≈0) for Si(1) and
(ϩ34.2 ppm, ≈0.2) for Si(2) leading to axial or nearly axial
shielding tensors. However, the exact determination of η for
axial or nearly axial symmetry using graphical methods is very
difficult, whereas the shielding anisotropy is reasonably well
defined.41 Therefore, the static powder spectrum [Fig. 2(c)] was
simulated with a unique set of parameters (η1, η2) and fixed
anisotropy (∆σ = ϩ34.2 ppm) [Fig. 2(e)]. The best simulation
was obtained with η1 = 0.0 for Si(1) and η2 = 0.3 for Si(2). The
static pattern could not be described by two tensors having the
same η parameter. The shielding tensor relative to Si(2) is non-
axial whereas the Si(1) tensor can be considered as axial. How-
ever, the deviation from axiality for Si(2) remains rather low.
The value of the anisotropy is in accord with values obtained
for various organosilicon derivatives.38
Fig. 4 An IRCP sequence based on a standard 13C CP MAS experi-
ment:48 tCP = contact time (fixed) and ti = inversion time (variable)
The 13C NMR spectrum of compound 1 is characterized by
three isotropic resonances centred at δ 138.0, 138.5 and 128.8
[Fig. 3(a)]. The group of lines at δ 138.0/138.5 and the line at δ
128.8 are roughly in the ratio 1:1. The lines at δ 138.0 and 138.5
are roughly in the ratio 1:3. This intensity analysis was done for
a contact time tCP = 5 ms, corresponding to an optimized value.
Indeed, all magnetizations reached their maximum for tCP у 5
ms as shown by a variable-contact-time experiment [T1ρ-
(1H) ≈ 200 ms, i.e. T1ρ(1H) ӷ TC and TD , see below]. Both
groups of lines are distinguished not only by isotropic chemical
shifts but also by shielding anisotropies [Fig. 3(b), 3(c) and
Table 3]. These observations indicate that each group of reson-
ances may be assigned to CH or CH2 sites (direct definite
assignment is not straightforward at this stage). The sideband
pattern corresponding to δ 128.8 [Fig. 3(b)] consists of narrow
lines of equal width: this ascertains the set-up of the magic
angle. It has been shown that off-axis sample spinning in the
slow-spinning regime led to distorted sideband patterns and
could result in misinterpretations of the spectra 42 (especially in
the case of high 13C shielding anisotropies). Therefore, the
resonances closely centred at δ 138.0 and 138.5 do correspond
to isotropic peaks and are not artifacts. In this sense compound
1 could be used as a secondary reference for precise magic
angle set-up (at least on the 13C chemical shift scale).
Investigation by IRCP techniques. A few sequences based on
CP NMR allow one to edit 13C resonance lines in connection
with thenumber of directlybonded protons. Themost commonly
used sequence is referred to in the literature as NQS (non-
quaternary suppression). First introduced by Alla and Lipp-
maa 46 and developed by Opella and Frey,47 it allows one to
distinguish between rigid protonated groups (such as CH and
CH2) and weakly coupled sites (such as carbonyl or quaternary
C) by dipolar dephasing. Methyl has an intermediate behaviour
due to fast reorientation, leading to a substantial reduction in the
dipolar coupling strength. However, this sequence fails to dis-
tingish between ‘rigid’ CH and CH2 groups.48 Indeed, CH and
CH2 have similar dipolar dephasing constants; moreover, these
constants can largely be modulated by molecular motion. A
second approach to spectral editing is related to solid-state J
1
spectroscopy combined with multipulse on the H channel.49
However, so far, this has been applied mainly to species with
high intrinsic mobilities and very rarely to rigid solid com-
pounds. More recently, sequences based on polarization/
inversion of polarization have been widely developed in the
framework of spectral editing.48,50
The IRCP sequence used in this work is presented in Fig. 4. It
is derived from a standard CP sequence: after a contact time tCP
(long enough to polarize all 13C nuclei), a 180Њ phase shift is
operated on the 13C channel. Owing to spin–temperature inver-
sion, the obtained carbon-13 magnetization will invert during
the inversion time ti. Then, the remaining 13C signal is acquired,
including high-power decoupling. This experiment is also called
CPPI (cross polarization with polarization inversion).48 It has
been demonstrated that cross-polarization dynamics and polar-
ization inversion dynamics are identical.51 In the case of
strongly coupled sites (such as CH and CH2), analytical expres-
Shielding anisotropies (∆σ) for C(1)/C(2), C(11) and C(21)
(see Table 3) are small when compared to common values
observed for rigid entities R1R2C᎐CR3R4.43,44 This may suggest
᎐
motional averaging at the vinyl sites leading to a drastic reduc-
tion in CSA values for every 13C site.45 This point will be care-
fully analysed by using the IRCP technique. Such a NMR
sequence will allow also definite assignments for the observed
lines.
1620
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626

View Online
sions of the magnetizations during the inversion process have
been derived for powdered samples at moderate MAS (Ӷ10
kHz), neglecting relaxation,48 equation (1) where M0 represents
2
ti
M(13CHn) = M0
ͫ
exp
exp
ͩ
Ϫ
ͪ
ϩ
n ϩ 1
TD
2n
3 ti
ti2
(1)
ͩ
Ϫ
ͪ
exp
ͩ
Ϫ
ͪ ͬ
Ϫ 1
2
n ϩ 1
2 TD
TC
the maximal magnetization reached after tCP and n corre-
sponds to the number of directly bonded protons in the CHn
group. The inversion of polarization is characterized by two
time constants which are related to different types of magnet-
ization transfer: TC accounts for the coherent transfer of mag-
netization, involving the 13C nucleus and the directly bonded
protons; it can be considered as an inverse measure of the
strength of the heteronuclear dipolar coupling for CHn groups.
It can be estimated for rigid CHn groups. Rapid molecular
motion will lead to a reduction in dipolar coupling and will be
related therefore to an increase in TC. Parameter TD is related to
spin-diffusion processes, which involve all the remaining pro-
tons. Equation (1) is a generalization of that first derived by
Müller et al.,52 accounting for magnetization oscillations in CP
transfer for ferrocene single crystals. Generally, TC Ӷ TD and
the polarization inversion can be described by two regimes: the
first (first tens of microseconds for ti) is dominated by a rapid
decay of M(13CHn), characterized by TC. Then, a much slower
regime of inversion follows (TD ). As the dynamics of the two
regimes are very different, a sharp turning point is observed. It
corresponds to the end of the fast regime of inversion. Then,
the magnetization reaches (1 Ϫ n)(1 ϩ n)Ϫ1M0 [equation (1)].
The CH and CH2 groups can be distinguished by the value
of their turning point: 0 and Ϫ¹M0, respectively. In this
Fig. 5 The 13C IRCP MAS NMR spectra of compound 1 (tCP = 5 ms);
ti = 0 µs corresponds to a standard CP MAS experiment (νrot = 4000 Hz,
Ns = 48, l.b. = 10)
¯
³
sense spectral editing is achieved. Moreover, it has been demon-
strated that the criterion of the turning point is much less sensi-
tive to molecular motions than are methods which rely only on
the relative strengths of dipolar couplings (NQS sequence). In
the case of weakly coupled sites (such as C᎐O, quaternary C or
᎐
rapidly rotating Me), the inversion of polarization is well
described by an exponential process and one unique time
constant, equation (2), again neglecting relaxation; TCH is the
ti
M(13C) = M0
ͫ
2 exp
ͩ
Ϫ
ͪ ͬ
Ϫ 1
(2)
TCH
standard cross-polarization constant. Equation (2) corresponds
to the standard thermodynamic approach of cross-polarization
dynamics.28
The IRCP sequence (13C) has been applied to the study of
compound 1 and to gel 2. Our goal was to propose definite
assignments for the 13C lines and possibly to obtain information
about molecular motions. Spectra obtained for compound 1
(rotation speed = 4 kHz) with different increments of ti are pre-
sented in Fig. 5. At this rotation speed only two sets of spinning
sidebands (with low intensities) are observed. The inversion of
the normalized magnetization intensities is presented in Fig. 6.
The total magnetization of the group of resonances centred at
δ 138.0 and 138.5 was considered. The IRCP sequence was
also applied to a sample of cadmium acetylacetonate [Cd-
(MeCOCHCOMe)2], i.e. [Cd(acac)2] (see Experimental sec-
tion). In thiscompound theacetylacetonateligandsarestabilized
in the enol form showing CH spin pairs with sp2 carbon atoms.
This metalloorganic compound can act as a reference for ‘rigid’
Fig. 6 Evolution of the magnetizations (compound 1) versus inversion
time ti. Fits are related to equation (1) (with n = 1 for CH and n = 2 for
CH2). ᭹, δ 128.8; ᭿, 〈δ〉 ≈ 138
higher values of ti. However, the turning points between the two
regimes are obviously different. For δ 128.8 it corresponds
roughly to 0; for 〈δ〉 ≈ 138 it corresponds roughly to Ϫ¹. Inver-
¯
³
sion curves were fitted using equation (1). Best fits were
obtained with n = 1 (δ 128.8) and 2 (〈δ〉 ≈ 138). Fits obtained
with equation (2) involving a unique time constant were very
poor. The extracted values of TC and TD are presented in Table
4. It is then possible to confirm assignments on the basis of the
inversion curves. The resonance centred at δ 128.8 is assigned to
CH groups (n = 1); both resonances at δ 138.0 and 138.5 are
assigned to CH2 groups (n = 2). As indicated above, it is also
obvious that TD ӷ TC. We should mention at this stage that the
resonance at δ 128.8 and its spinning sidebands are flanked by
2
Csp H groups. Actually, two CH resonances are observed due to
acetylacetonate moieties consistent with crystallographic data.53
Shift parameters are presented in Table 3. It is clearly seen that
both inversions for 〈δ〉 ≈ 138 and δ 128.8 present two regimes: a
rapid one during the first 100 µs and a much slower one for
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626
1621

View Online
Table 4 Extracted TC and TD values from 13C IRCP MAS NMR
experiments according to equation (1)
Compound
13C site
TC/µs
TD /ms
3 ± 0.7
1.3 ± 0.2
3.5*
2.2 ± 0.3
0.43 ± 0.02
1
CH[C(1)/C(2)]
CH2[C(11)/C(21)]
CH
CH2
CH
75 ± 4
30 ± 2
80*
30 ± 2
25 ± 0.6
2
[Cd(acac)2]
* Estimation, see text.
two satellites [see the insert in Fig. 3(a)]. The satellites are
assigned to 13C nuclei coupled to 29Si [natural abundance 4.7%
and |1J(13C᎐29Si)| = 136 ± 5 Hz]. The 13C INEPT study (see
Experimental section) of the molecular precursor Si(C2H3)-
(OEt)3 confirmed these results: (i) the CH2 signal was less
shielded than the CH signal, (ii) |1J(13C᎐29Si)| = 119 Hz, in
agreement with the value observed in the solid state. The signal
centred at δ 128.8 corresponds obviously to C(1) and C(2)
resonances (see crystallographic section above). These reson-
ances are not resolved as the isotropic chemical shifts must be
very similar. The extracted value of TC for CH2 (Table 4) is
clearly higher than those observed for ‘rigid’ aliphatic CH2
groups.48,54 Moreover, the magnetization related to CH2 in
compound 1 is still positive for ti = 25 µs (Fig. 5), whereas the
magnetization of CH2 in glycine (NH2CH2CO2H) is slightly
negative for the same inversion time (see Experimental section).
Fig. 7 Approximate geometry of the vinyl group corresponding to
atoms Si(1), C(1) and C(11) and to related H atoms (calculated posi-
tions): β1 ≈ 63, β2 ≈ 6 and β3 ≈ 54Њ
unique axis. Then, the dipolar coupling ‘strength’ is modulated
by the factor R = (3 cos2 β Ϫ 1)/2, where β corresponds to the
angle between the C᎐H bond and the axis of rotation. This
result is general45 and can be transposed for any second-rank
tensor such as CSA and J coupling (when considering first-
order effects on the lineshapes). When the tensor is not axially
symmetric the reduction factor for the interaction is a more
complex function of the principal values of the tensor and the
Euler angles between the principal axes and the reorientation
direction. In the case of CH, β1 ≈ 63Њ so that |R| ≈ 0.2. In other
words, this particular geometry of the CH pair leads to a very
strong reduction in dipolar coupling under rapid reorientation.
This can explain the high value of TC observed for the CH
group in the IRCP experiment. The IRCP curves of [Cd(acac)2]
(containing ‘rigid’ CH groups), compound 1 and ferrocene
[Fe(η-C5H5)2] (adapted from ref. 56 and neglecting rotational
echoes, see below) are presented in Fig. 8. We assume for this
comparison that the C᎐H bond lengths are nearly equal for the
three compounds. The rigid CH in [Cd(acac)2] experiences
strong dipolar coupling leading to a low value of TC and rapid
inversion during the first tens of µs. Ferrocene is characterized
by rapid rotation around the internal axis with β = 90Њ (|R| =
0.5);56 its evolution is intermediate between the rigid case of
[Cd(acac)2] and the behaviour of the vinyl groups in compound
1 as expected from inspection of the R values. In the case of
methylene groups (CH2), β2 ≈ 6 and β3 ≈ 54Њ. This means that
the C᎐H(2) bond is nearly collinear with the rotation axis
whereas the C᎐H(3) bond is nearly at the magic angle. The
C᎐H(2) dipolar coupling ‘strength’ is therefore almost
unaffected by the motion (R ≈ 1). The removal of dipolar coup-
ling from H(3) is almost complete. Therefore, the inversion of
the CH2 magnetization is rapid during the first tens of micro-
seconds (TC = 30 µs) but less than for a ‘rigid’ CH2 group. This
TC value is comparable to values observed for rigid CH groups.
We have shown that simple geometrical arguments may explain
the different evolutions of CH and CH2 magnetizations. These
results are at best qualitative since βi angles are estimated
through the calculated positions of the H atoms. Conversely,
theoretical analysis of the IRCP curves may lead to the
determination of particular angles and bond lengths in the
molecule.
2
Owing to slightly shorter C᎐H bond lengths, a ‘rigid’ Csp H2
3
group should invert more rapidly than a ‘rigid’ aliphatic Csp H2
group. For compound 1 such experimental observations suggest
motional reduction of the dipolar coupling between 13C and the
directly bonded protons. As the dipolar constant TC is signifi-
cantly affected, the frequency of the motion can be estimated to
be ν у 105 Hz.55 At this stage we can conclude that the apparent
disorder observed for C(11) and C(21) by X-ray diffraction
cannot be related to a static one. Indeed, in this case the
behaviour of the rigid CH2 groups should be characterized by
TC(CH2) < TC(CH) for [Cd(acac)2], which is obviously not the
case (Table 4).
Fast anisotropic motion of the vinyl groups (considered as
rigid entities) around the Si(1)᎐C(1) and Si(2)᎐C(2) bonds may
then be considered. The crystallographic refinements show that
the local motion of the vinyl groups may consist of rapid jumps
between discrete positions: three symmetry-related positions
for C(11) and (at least) four for C(21). It is known that a con-
tinuous rotation of vinyl groups is not necessary to obtain aver-
aging of dipolar coupling under reorientation. Rapid jumps
between p positions (p у 3) located on a regular polygon lead
exactly to the same averaging. By symmetry, the three C(11)
positions (with equal occupancy) describe an equilateral tri-
angle. The four C(21) positions describe a distorted square but
refinement in this case is much less safe. If the approach of
rapid motion of the vinyl groups is correct two points remain
open: (i) TC for CH is much higher than TC observed for CH2 in
compound 1 (see Table 4), which indicates a much stronger
reduction of the dipolar coupling in the case of the CH spin
pair; (ii) the fits of the IRCP curves (Fig. 6) obtained with
equation (1) are less reliable for higher values of ti. Let us
answer these two questions, considering reorientation of the
vinyl groups around the Si᎐C bonds.
(i) The approximate geometry of one vinyl group is repre-
sented in Fig. 7. Angles βi (i = 1–3) are those related to the
atoms Si(1), C(1), C(11), H(1), H(2) and H(3) (see Fig. 1). Let
us first consider the strong dipolar coupling between the CH
spin pair. In a static configuration the dipolar tensor corre-
sponding to CH is axially symmetric with the C᎐H bond direc-
tion as a unique axis. If we consider rapid anisotropic reorien-
tation around the Si᎐C direction (rapid jumps) the dipolar ten-
sor remains axially symmetric but with the Si᎐C bond as a new
(ii) In the case of ferrocene the second stage of polarization
inversion was modulated by strong rotational echoes.56 Indeed,
the interactions between CH pairs are periodically refocused by
MAS at times tn = (n2π/ωrot) = (n/νrot) with n = 1, 2 etc., νrot
being the spinning frequency. In our experiments, νrot = 4000
Hz and t1 = 250 µs, t2 = 500 µs, etc. Such echoes could explain the
fact that the calculated curves in Fig. 6 are less reliable in the
second stage of polarization inversion. Nevertheless, the order
of magnitude for extracted TD values is correct.
1622
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626

View Online
2
Fig. 8 Evolution of Csp H magnetization versus inversion time ti
(IRCP sequence) for: ᭡, cadmium acetyloacetonate (‘rigid’ CH); ——,
ferrocene, adapted from ref. 56 with TC = 50 µs and TD = 1.7 ms, |R| = 0.5;
᭹, compound 1 (expansion of Fig. 6), |R| ≈ 0.2
The 13C CSA should also be modulated by vinyl reorien-
tations. As stated above, CSA values for all carbon-13 sites are
unusually small. However, the interpretation for the reduc-
tion of ∆σ is difficult. The CSA can be represented by a second-
rank tensor, characterized by three initial principal values.
Upon rapid reorientation, the reduction in CSA can be repre-
sented by a new tensor which is axially symmetric. The new
principal values are functions of the initial values and Euler
angles between the reorientation axis and the initial principal
axes.45 In the case of compound 1 the orientations of the differ-
ent CSA principal axes systems in the ‘static’ configuration of
the cubane are not known. In other words, the Euler angles
which orientate the initial tensor from the Si᎐C bond are not
known. The IRCP results, concerning the reduction of dipolar
tensors, are easier to analyse than the ∆σ reduction, as the CH
direction corresponds to a principal direction of the initial
dipolar tensor. Therefore, the IRCP results and ∆σ reduction
cannot be directly related.
As a conclusion the IRCP dynamics study allows a deep
insight into the dipolar local fields. The 13C spectra can be edited
completely (CH and CH2 distinction) though strong reduction
of the dipolar coupling was present. Turning points in the
IRCP curves are safe criteria for spectral editing. Reorientation
of the vinyl groups could explain the different magnetization
behaviours during the IRCP sequence. Moreover, the reorient-
ation of the vinyl groups could be considered independently
from the cubane core of the molecule as will be seen below.
Finally, one should note that such different CP dynamics for
closely spaced CH and CH2 may have a considerable influence
on quantification by 13C CP MAS NMR spectroscopy and may
lead to misinterpretations in spectral editing. One must be
aware of specific molecular motions leading to strongly reduced
dipolar coupling.
Fig. 9 The 13C CP and IRCP MAS NMR spectra of compound 2:
νrot = 4000 Hz, Ξ (13C) = 75.46 MHz, φ = 4 mm, Ns = 600, tCP = 5 ms,
r.d. = 6 s, l.b. = 30 Hz
and 17: CH2 and CH3, respectively). Low-speed sideband pat-
terns related to vinylic 13C can be well simulated by two sets of
average (∆σ, η) CSA values (see Table 3 and Fig. 10). The ∆σ
values are in agreement with those observed for compound 1,
suggesting again rapid reorientation of vinyl groups around
Si᎐C bonds.
Investigation by 13C IRCP MAS NMR. The IRCP spectra
obtained for ti = 30 and 50 µs are presented in Fig. 9. Evolution
of both magnetizations (vinylic CH and CH2) is presented in
Fig. 11. Simulations were obtained using two peaks with fixed
isotropic chemical shift and linewidth. Only amplitudes were
allowed to vary. The inversion of magnetization shows clearly
two regimes for both resonances and equation (1) was used to
extract TC and TD values. Once again, spectral editing is easy
and the less shielded peak (δ 135) can be safely assigned to
CH2 [n = 2 in equation (1)]. The fit obtained for δ 129 (using
n = 1) is less safe because the turning point between the two
regimes of inversion is higher than 0; TC and TD values are
given as estimations. However, the CH IRCP behaviours for
compounds 1 and 2 are identical during the first tens of µs.
The TC values are in very close agreement with those
extracted for CH and CH2 in compound 1. This indicates
that the motion of the vinyl groups localized around Si᎐C
bonds is independent of the rest of the structure (i.e. the
cubane cage in the case of compound 1). The TD values are
of the same order of magnitude but show some variations
between compounds 1 and 2. This is not surprising as TD is
strongly influenced by the 1H᎐1H interactions, which may
vary considerably from one compound to another.
Compound 2 (xerogel). Standard 29Si and 13C CP MAS
experiments. The 29Si CP MAS spectrum (not shown) of com-
pound 2 is characterized by a major peak centred at δ Ϫ78
(linewidth ≈ 350 Hz), which corresponds to fully condensed T3
᎐
units C H Si*(OSi᎐) . A minor peak centred at δ Ϫ72 is also
᎐
2
3
3
observed and corresponds to T2 units C H Si*X(OSi᎐) with
᎐
᎐
2
2
3
X = OH or OEt. This demonstrates that the network of silicon
entities is highly condensed. Spectra of compound 2 obtained
by 13C CP MAS are presented in Fig. 9 (ti = 0 µs) and NMR data
corresponding to the 13C vinyl resonances are presented in Table
3. The CP spectrum has two peaks located at δ 135 and 129 and
two minor resonances assigned to residual ethoxy groups (δ 58
Conclusion
Octameric vinylsilasesquioxane 1 has been studied by X-ray dif-
fraction and 29Si/13C CP MAS NMR spectroscopy. The X-ray
data showed an uncertainty about the vinyl CH2 positions
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626
1623

View Online
INEPT. The high resolution of the spectra of 1 allows one to
use this compound as a safe secondary reference for the set-up
of 29Si and 13C CP NMR experiments.
The amorphous gel 2 was also studied by 29Si/13C CP NMR
spectroscopy. It is a good example for applying NMR analysis
techniques, which werecarefullyset with welldefined compounds
(such as crystals) to amorphous derived materials. The IRCP
experiment also showed a strong reduction in dipolar coupling
through motional averaging. Therefore, extreme caution should
be taken when considering quantitative CP experiments. This
couldbeespeciallyimportantinthestudyof amorphousmaterials
(bridged organofunctional silasesquioxanes, for instance), for
which 13C CP NMR spectroscopy is used in the quantification of
bridging reactions between the different cubane cores.
Experimental
Syntheses
Compound 1 and derived gel 2. Triethoxyvinylsilane (Fluka)
was used as received. The purity of this monomeric compound
was checked by standard 13C and 29Si liquid-state NMR spec-
troscopy and INEPT 29 [relative to SiMe4: δSi Ϫ58.7; δ (᎐CH)
᎐
C
130.5; δ (᎐CH ) 135.8; δC(CH2) 58.3; δC(CH3) 18.3;
᎐
C
2
|1J(13C᎐29Si)| = 119 Hz as shown by 13C INEPT]. Acidic water
(HCl, 0.1 mol lϪ1) (0.35 g, 19.7 mmol) was added to Si(C2-
H3)(OEt)3 (2.50 g, 13.1 mmol) with vigorous stirring, without
any solvent (H2O :Si = 1.5:1). After 2–3 min the mixture
became homogeneous. At room temperature and after several
days, single crystals of compound 1 were obtained as transpar-
ent parallelepipeds. These were extracted and washed with
dried ethanol. They were insensitive to air moisture and found
suitable for structure determination. Upon ageing the remain-
ing liquid mixture led to a transparent gel which could be
heated at 80 ЊC, leading to a dry amorphous powder called
xerogel (compound 2). Such powders were studied without
further treatment. The yield of the crystallized phase was esti-
mated to be a few %. Several single crystals were analysed by
X-ray diffraction, leading to one unique crystallographic struc-
ture, corresponding to compound 1. The IR spectrum was
recorded and acted as ‘fingerprint’ of 1. Many experiments
were repeated to obtain crystals, the molecular structure of
which was systematically checked by IR spectroscopy.
Fig. 10 Low-rotation-speed 13C CP MAS spectrum (lower) of com-
pound 2. Only the vinyl CH and CH2 patterns are simulated (upper):
νrot = 1000 Hz, Ξ (13C) = 75.46 MHz, φ = 7 mm, Ns = 1688, tCP = 5 ms,
r.d. = 6 s, l.b. = 10 Hz. Isotropic resonances are labelled by arrows
Cadmium acetylacetonate. Acetylacetone (pentane-2,4-dione,
Fluka) (1 cm3, 9.68 mmol) was slowly added at room tem-
perature and with stirring to an alkaline aqueous solution of
cadmium chloride (6 cm3 of KOH solution at pH 14 ϩ 20 cm3
of CdCl2 solution, [Cd2ϩ] = 0.25 mol lϪ1). The white precipitate
immediately obtained was filtered off and washed with iced
water. Crystalline cadmium acetylacetonate was identified by
X-ray diffraction: according to the crystallographic data given
in ref. 53, a powder diffractogram was simulated using FULL-
PROF program.57 The experimental diffractogram and the
simulation were strictly identical. Solid-state 13C CP MAS
NMR (contact time = 2 ms, recycle delay = 6 s, number of
Fig. 11 Evolution of the magnetizations (compound 2) versus inver-
sion time ti for vinylic CH and CH2 signals: ᭹, δ 129; ᭿, δ 135
owing to static disorder or rapid reorientation by jumping. The
latter is in accord with a careful analysis of CP dynamics, which
revealed the rapid anisotropic reorientation of the vinyl groups
around the Si᎐C bonds. The frequency of the motion was cer-
tainly higher than 105 Hz, leading to a strong reduction in
dipolar coupling. Adjacent CH and CH2 groups showed very
different residual dipolar coupling, as shown by IRCP. These
observations were explained by the simple geometrical charac-
teristics of the vinyl groups: the angle between the C᎐H and the
Si᎐C bond axis was nearly the magic angle, leading to a very
strong reduction in the dipolar coupling. Nevertheless, the CH
polarization inversion showed two regimes as expected, allow-
ing a complete spectral editing and definite assignments. The
IRCP sequence can be safely used in solid-state NMR spectro-
scopy for the determination of proton multiplicities and can
be compared to the liquid-state NMR sequences DEPT and
scans = 856): δ (C᎐O) 199.2, 193.8, 189.8 and 187.1; δ (᎐CH)
᎐
᎐
C
C
102.1 and 100.6; δC(CH3) 30.2, 29.7 and 29.3. These isotropic
chemical shifts are in very good agreement with those observed
by Takegoshi et al.58
Spectroscopy
Infrared spectra were recorded at room temperature as pressed
KBr pellets with a Nicolet Fourier-transform spectrometer
(range 400–4000 cmϪ1); liquid 13C and 29Si NMR spectra at
room temperature on Bruker AC-300 (13C, 75.46 MHz) and
MSL-400 (29Si, 79.50 MHz) spectrometers, respectively, using a
Bruker VSP probe. Chemical shifts are referenced to SiMe4.
1
The 13C INEPT sequence29 (including decoupling from H and
refocusing) was also used to ascertain assignments (see Results
1624
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626

View Online
and Discussion). Solid-state 13C and 29Si CP MAS experiments
were carried out using Bruker MSL-300 and MSL 400 spectro-
meters (13C, 75.46; 29Si, 59.62 and 79.50 MHz) equipped with
Bruker probes including external lock (4 and 7 mm probes).
Zirconia rotors were used. Solid samples were spun at 0.5–4
kHz. Fluctuations in MAS speed were smaller than ±2 Hz over
several hours. The magic angle was carefully set using the 79Br
resonance of KBr. The matching condition for the Hartmann–
Hahn cross-polarization (1H 90Њ pulse length: 5 µs for the 4 mm
probe and 6.5 µs for the 7 mm probe) was set on adamantane
(13C) and compound 1 (29Si). It should be noted that both radio-
frequency channel levels have to be very carefully set. Indeed, it
has been demonstrated that under mismatched conditions the
results of IRCP sequences may be false.48 In order to check
the matching of the Hartmann–Hahn condition, the IRCP
seqence was tested on a glycine sample which acts as a typical
‘rigid’ methylene CH2 group (see Results and Discussion). The
test was done periodically during all 13C IRCP MAS experi-
ments. Contact times (tCP) were optimized by standard variable-
contact-time experiments. Relaxation delays were 6–15 s. The
13C and 29Si shielding-tensor analyses,39 as well as deconvolu-
tion of lines, were obtained by using the WINFIT program
developed by Massiot et al.40
References
1 M. G. Voronkov and V. I. LavrentЈyev, Top. Curr. Chem., 1982, 102,
199.
2 F. J. Feher and T. A. Budzichowski, Polyhedron, 1995, 22,
3239.
3 R. H. Baney, M. Itoh, A. Sakakibara and T. Suzuki, Chem. Rev.,
1995, 95, 1409.
4 D. A. Loy and K. J. Shea, Chem. Rev., 1995, 95, 1431.
5 C. L. Frye and W. T. Collins, J. Am. Chem. Soc., 1970, 92, 5586.
6 P. A. Agaskar and W. G. Klemperer, Inorg. Chim. Acta, 1995, 229,
355 and refs. therein.
7 F. J. Feher, T. A. Budzichowski and K. J. Weller, J. Am. Chem. Soc.,
1989, 111, 7288.
8 F. J. Feher and K. J. Weller, Organometallics, 1990, 9, 2638.
9 K. Larsson, Ark. Kemi, 1960, 16, 203, 209, 215.
10 I. A. Baidina, N. V. Podberezskaya, V. I. Alekseev, T. N. Martynova,
S. V. Borisov and A. N. Kanev, Zh. Strukt. Khim., 1979, 20, 648;
N. V. Podberezskaya, S. A. Magarill, I. A. Baidina, S. V. Borisov,
L. E. Gorsh, A. N. Kanev and T. N. Martynova, J. Struct. Chem.
(Engl. Transl.), 1982, 23, 422.
11 G. Koellner and U. Müller, Acta Crystallogr., Sect. C, 1989, 45,
1106.
12 F. J. Feher and T. A. Budzichowski, J. Organomet. Chem., 1989, 373,
153.
13 H. B. Bürgi, K. W. Törnroos, G. Calzaferri and H. Bürgy, Inorg.
Chem., 1993, 32, 4914.
14 G. Calzaferri, R. Imhof and K. W. Törnroos, J. Chem. Soc., Dalton
Trans., 1994, 3123.
15 K. W. Törnroos, H. B. Bürgi, G. Calzaferri and H. Bürgy, Acta
Crystallogr., Sect. B, 1995, 51, 155.
Crystallography
Crystals of compound 1 were prepared as described above,
16 K. W. Törnroos, G. Calzaferri and R. Imhof, Acta Crystallogr., Sect.
C, 1995, 51, 1732.
sealed in Lindemann capillaries and studied at 293 K.
17 K. W. Törnroos, Acta Crystallogr., Sect. C, 1994, 50, 1646.
18 M. Bärtsch, P. Bornhauser, G. Calzaferri and R. Imhof, J. Phys.
Chem., 1994, 98, 2817.
19 B. J. Hendan and H. C. Marsmann, J. Organomet. Chem., 1994, 483,
33.
20 J. Kowalewski, T. Nilsson and K. W. Törnroos, J. Chem. Soc.,
Dalton Trans., 1996, 1597.
21 E. Lippmaa, M. A. Alla, T. J. Pehk and G. Engelhardt, J. Am. Chem.
Soc., 1978, 100, 1929.
22 G. Engelhardt, D. Zeigan, D. Hoebbel, A. Samoson and E. Lippmaa,
Z. Chem., 1982, 22, 314.
23 J. D. Lichtenhan, Y. A. Otonari and M. J. Carr, Macromolecules,
1995, 28, 8435.
24 S. E. Yuchs and K. A. Carrado, Inorg. Chem., 1996, 35, 261.
25 A. Sellinger and R. M. Laine, Macromolecules, 1996, 29, 2327.
26 T. N. Martynova and V. P. Korchkov, J. Organomet. Chem., 1983,
248, 241.
Crystal data. C16H24O12Si8, M = 633.04, trigonal, space group
¯
R3 (no. 148), a = 13.533(2), c = 14.222(2) Å, γ = 120Њ,
U = 1012(5) Å3 (by least-squares refinement for 25 reflections in
the range 13.9 < θ < 14.5Њ, λ = 0.710 69 Å), Z = 3, Dc = 1.398 g
cmϪ3, F(000) = 984. Colourless parallelepipeds. Crystal dimen-
sions 0.4 × 0.3 × 0.3 mm, µ(Mo-Kα) = 0.4 mmϪ1
.
Data collection. Enraf-Nonius CAD-4 diffractometer, ω–2θ
mode with ω scan width = 0.80 ϩ 0.34 tan θ, ω-scan speed 2–
20.1 minϪ1, graphite-monochromated Mo-Kα radiation, 999
reflections measured (1 р θ р 25Њ; h 0–13, k 0–13; lϪ16 to 16),
882 independent, giving 332 with I у 3σ(I) (merging R =
0.0166). Two standard reflections were used (intervals: 100
reflections and 60 min). Variation of standards: 1%.
27 T. N. Martynova and T. I. Chupakhina, J. Organomet. Chem., 1988,
345, 11.
28 A. Pines, M. G. Gibby and J. S. Waugh, J. Chem. Phys., 1973, 59,
569.
29 A. E. Derome, Modern NMR Techniques for Chemistry Research,
Pergamon, Oxford, 1991, p. 129.
30 C. J. Brinker and G. W. Scherer, Sol-Gel Science: The Physics and
Chemistry of Sol-Gel Processing, Academic Press, San Diego,
1990.
31 W. R. Schmidt, L. V. Interrante, R. H. Doremus, T. K. Trout,
P. S. Marchetti and G. E. Maciel, Chem. Mater., 1991, 3, 257.
32 C. A. Fyfe and J. Niu, Macromolecules, 1995, 28, 3894.
33 C. K. Johnson, ORTEP, Report ORNL-3794, Oak Ridge National
Laboratory, Oak Ridge, TN, 1965.
34 G. Socrates, Infrared Characteristic Group Frequencies, Wiley,
Chichester, 1994, p. 42.
35 D. Hoebbel, I. Pitsch, D. Heidemann, H. Jancke and W. Hiller,
Z. Anorg. Allg. Chem., 1990, 583, 133.
36 I. Pitsch, D. Hoebbel, H. Jancke and W. Hiller, Z. Anorg. Allg.
Chem., 1991, 596, 63.
37 A. R. Grimmer, R. Peter, E. Fechner and G. Molgedey, Chem. Phys.
Lett., 1981, 77, 331.
38 R. K. Harris, T. N. Pritchard and E. G. Smith, J. Chem. Soc.,
Faraday Trans. 1, 1989, 85, 1853.
39 J. Herzfeld and A. E. Berger, J. Chem. Phys., 1980, 73, 6021.
40 D. Massiot, H. Thiele and A. Germanus, Bruker Rep., 1994, 140, 43.
41 N. J. Clayden, C. M. Dobson, L. Y. Lian and D. J. Smith, J. Magn.
Reson., 1986, 69, 476.
Structure solution and refinement. Direct methods (Si, O, C),
full-matrix least-square refinements with anisotropic thermal
parameters in the last cycles for all non-H atoms and hydrogens
in calculated positions (assuming sp2 C atoms). Occupancies of
C(11), C(21)–C(24) positions were also refined. As µ(Mo-Kα)
was rather low, no absorption correction was applied. Sixty-
eight refined parameters. Extinction correction method:59 sec-
ondary extinction 36(9). The weighting scheme was w = 1. Final
R and RЈ values were 0.0503 and 0.0480; (∆/σ)max = 0.15. A final
Fourier-difference calculation showed residual electron density
in the range Ϫ0.2 to ϩ0.15 e ÅϪ3. The SHELXS 8660 system of
computer programs was used for the direct methods. Refine-
ment was performed with the CRYSTALS program.61 Atomic
scattering factors, corrected for anomalous dispersion, were
obtained from ref. 62.
Atomic coordinates, thermal parameters, and bond lengths
and angles have been deposited at the Cambridge Crystallo-
graphic Data Centre (CCDC). See Instructions for Authors,
J. Chem. Soc., Dalton Trans., 1997, Issue 1. Any request to the
CCDC for this material should quote the full literature citation
and the reference number 186/438.
Acknowledgements
42 E. M. Menger, D. P. Raleigh and R. G. Griffin, J. Magn. Reson.,
1985, 63, 579.
43 T. M. Duncan, A Compilation of Chemical Shift Anisotropies, The
Farragut Press, Chicago, 1990, p. C-13.
We thank Professor P. Boch for his support and Professor
R. M. Laine for providing different octameric silasesquioxanes.
S. Mace is also gratefully acknowledged for experimental help.
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626
1625

View Online
55 F. Lauprêtre, L. Monnerie and J. Virlet, Macromolecules, 1984, 17,
44 A. M. Orendt, J. C. Facelli, A. J. Beeler, K. Reuter, W. J. Horton,
P. W. Cutts, D. M. Grant and J. Michl, J. Am. Chem. Soc., 1988, 110,
3386.
45 M. Mehring, Principles of High Resolution NMR in Solids, Springer,
Berlin, 1983, p. 50.
1397.
56 J. Hirschinger and M. Hervé, Solid State NMR, 1994, 3, 121.
57 J. Rodrigues-Carvajal, FULLPROF 93, An Advanced Rietveld
Code, Institut Laue-Langevin, Grenoble, 1993.
46 M. Alla and E. Lippmaa, Chem. Phys. Lett., 1976, 37, 260.
47 S. J. Opella and M. H. Frey, J. Am. Chem. Soc., 1979, 101, 5854.
48 X. Wu and K. W. Zilm, J. Magn. Reson., 1993, A102, 205.
49 T. Terao, H. Miura and A. Saika, J. Am. Chem. Soc., 1982, 104,
5228.
58 K. Takegoshi, K. J. Schenk and C. A. McDowell, Inorg. Chem.,
1987, 26, 2552.
59 A. C. Larson, Crystallographic Computing, eds. F. R. Ahmed,
J. R. Hall and C. P. Huber, Munksgaard, Copenhagen, 1970,
p. 291.
50 X. Wu, S. T. Burns and K. W. Zilm, J. Magn. Reson., 1994, A111,
29.
60 G. M. Sheldrick, SHELXS 86, Program for the Solution of Crystal
Structures, University of Göttingen, 1986.
51 X. Wu, S. Zhang and X. Wu, Phys. Rev. B, 1988, 37, 9827.
52 L. Müller, A. Kumar, T. Baumann and R. R. Ernst, Phys. Rev. Lett.,
1974, 32, 1402.
53 E. N. Maslen, T. M. Greaney, C. L. Raston and A. H. White,
J. Chem. Soc., Dalton Trans., 1975, 400.
61 D. J. Watkin, J. R. Carruthers and P. W. Betteridge, CRYSTALS,
An Advanced Crystallographic Program System, Chemical
Crystallography Laboratory, University of Oxford, 1988.
62 International Tables for X-Ray Crystallography, Kynoch Press,
Birmingham, 1974, vol. 4.
54 C. Bonhomme, J. Maquet, J. Livage and G. Mariotto, Inorg. Chim.
Acta, 1995, 230, 85.
Received 29th January 1997; Paper 7/00700K
1626
J. Chem. Soc., Dalton Trans., 1997, Pages 1617–1626