Well-defined silica-supported calcium reagents: Control of schlenk equilibrium by grafting

Article abstract of DOI:10.1002/chem.200802512

Calcium reagents Ca(α-Me₃Si-2-Me₂N-benzyl) ₂·2thf (1) and Ca[N(SiMe₃)₂] ₂·2thf (2) reacted with silica partially dehydroxylated at 700 °C to afford materials that bear (≡ SiO)Ca(α-Me ₃

Full text of DOI:10.1002/chem.200802512

DOI: 10.1002/chem.200802512
Well-Defined Silica-Supported Calcium Reagents: Control of Schlenk
Equilibrium by Grafting
[
a]
[b]
[a]
[b]
Rꢀgis M. Gauvin,* Frank Buch, Laurent Delevoye, and Sjoerd Harder*
Abstract: Calcium reagents Ca(a-
Me Si-2-Me N-benzyl) ·2thf (1) and
py, and elemental analysis. Reaction of
2 with one equivalent of the bulky sila-
zation, with evidence showing that cat-
alysis occurs through supported species.
In styrene polymerization, a marked in-
fluence of the surface acting as a ligand
on the stereoselectivity of the reaction
was observed, as syndiotactic-rich poly-
styrene (88% of r diads) was obtained.
These results illustrate that grafting of
calcium benzyl or amide compounds
on a silica surface is a new concept to
prevent ligand exchange through the
Schlenk equilibrium. Heteroleptic cal-
cium complexes that cannot be synthe-
sized as stable molecular species in so-
lution can be obtained as silica-sup-
ported species which have been shown
to be catalytically active.
3
2
2
Ca[N
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(SiMe ) ] ·2thf (2) reacted with
nol (tBuO) SiOH,
a
silica-surface
mimic, afforded the homoleptic bis-sil-
AHCTUNGTRENGNyUN loxide calcium derivative through
3
2
2
3
silica partially dehydroxylated at 7008C
to afford materials that bear (ꢀ
SiO)Ca(a-Me Si-2-Me N-benzyl)·1.6thf
ligand exchange (Schlenk equilibrium),
and a derivative was isolated and struc-
turally characterized. Preliminary stud-
ies have shown that both grafted
benzyl and amide derivatives are active
in olefin hydrosilylation, intramolecular
hydroamination, and styrene polymeri-
3
2
(
SiO -1)
and
(ꢀSiO)Ca
[N-
2
ACHTUNGTRENNUNG( SiMe ) ]·1.3thf (SiO -2) fragments, re-
3 2 2
spectively. Due to the bulk of the sup-
ported complexes, the silanol groups
are only partially metalated: 50% in
SiO -1 and 70% in SiO -2. In the case
2
2
of SiO -2, a parallel SiMe -capping side
2
3
reaction affords in fine a silanol-free
surface. The materials were character-
ized by IR spectroscopy, 1D and 2D
solid-state high-field NMR spectrosco-
Keywords: calcium · heterogeneous
catalysis · silica · supported cata-
lysts · surface chemistry
Introduction
intermediate reactive centers into inactive species. Even if,
from a simplified molecular point of view, the silica surface
may be considered as a bulky alkoxide, it still differs from
soluble models or bulky ligands in two ways: 1) the latter
often bury the metal center within a pocket that makes it
unreactive or less reactive and 2) molecular model ligands
may sterically and electronically mimic the silica surface but
do not take into account the enforced immobility of silica-
supported species. Distinct chemistry may thus be expected
from grafted systems.
Heterogenization of a catalyst is generally applied in the de-
velopment of cleaner processes based on recyclable materi-
als. Although this is one of the most interesting aspects of
catalyst immobilization, it is equally important to note that
the chemistry of a catalyst can be significantly altered by its
[2]
[3]
[1]
anchoring onto a support. Indeed, grafting may block mul-
timolecular deactivation processes or prevent aggregation of
Here we present the first studies on grafting of calcium
reagents on a silica surface. Our interest in grafted calcium
species stems from the rapidly growing application of calci-
[
a] Dr. R. M. Gauvin, Dr. L. Delevoye
Unitꢀ de Catalyse et de Chimie du Solide UMR CNRS 8181
Ecole Nationale Supꢀrieure de Chimie
BP 90108, 59652 Villeneuve d’Ascq Cedex (France)
Fax : (+33)320-43-67-54
[4]
um reagents as green catalysts. As calcium is one of the
cheapest and most accessible metals worldwide, we are not
interested in recycling a potential calcium catalyst, but our
motivations are based on two practical aspects. First, instead
of application in a batch process grafted catalysts can be ad-
vantageously used in continuous processes. Second, organo-
calcium chemistry is often plagued by the Schlenk equilibri-
E-mail: regis.gauvin@ensc-lille.fr
[
b] F. Buch, Prof. Dr. S. Harder
Anorganische Chemie, Universitꢁt Duisburg-Essen
Universitꢁtsstrasse 5, 45117 Essen (Germany)
Fax : (+49)201-183-2621
E-mail: sjoerd.harder@uni-due.de
[5]
um that is well-known for Grignard reagents. Calcium cat-
alysts in general are composed of an inactive spectator
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200802512.
4382
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 4382 – 4393

FULL PAPER
S
R
ligand (L ) and a reactive ligand (L ), responsible for cata-
lyst activity. A Schlenk equilibrium between heteroleptic
and homoleptic complexes would result in an undefined
mixture of potential catalytically active species (Scheme 1).
We report here experimental procedures for grafting of
calcium reagents and extensive analyses of the immobilized
species (elemental analyses, IR spectroscopy, and various
state-of-the-art solid-state NMR techniques). In addition,
we describe preliminary studies on the use of grafted calci-
um catalysts in alkene hydrosilylation, intramolecular
alkene hydroamination, and polymerization of styrene.
R
Scheme 1. Schlenk equilibrium via an associative transition state; L and
L represent reactive and spectator ligands, respectively.
Results and Discussion
S
Reaction of Ca
ACHTUNGTNERUNNG( DMAT) ·2thf (1) with SiO -700: synthesis
2 2
and characterization of SiO -1: Overnight stirring of a
2
As ligand exchange likely proceeds through a bimolecular
associative process, immobilization of the calcium reagent
could prevent such Schlenk equilibria. In addition, a solid
support can also be used to prevent aggregation of calcium
reagents. Thus, distinct highly reactive mononuclear calcium
sites may be obtained.
yellow toluene solution of Ca ACHTNUGTNERNUG( DMAT) ·2thf (1) with SiO -
2 2
700 affords, after washing with toluene, a yellow material
(SiO -1), the color of which fades on exposure to air
2
(Scheme 2). Elemental analysis indicates a Ca loading of
1.26 wt%, which corresponds to 0.55 Ca per square nanome-
À2
ter. As the initial silanol concentration is 1.1 nm , half of
The complexes Ca
A
H
U
G
R
N
U
G
the silanol groups reacted with the molecular precursor to
afford the grafted species. The origin of incomplete con-
sumption of silanol groups could be the bulk of the support-
2
3
[6]
[7]
Me N-benzyl) and Ca[N ACHTUNGTRNENUG( SiMe ) ] ·2thf (2) are versatile
3 2 2
2
reactive precursors in calcium chemistry and show catalyt-
ic activity in various reactions. Therefore, we targeted
ed Ca
ACHTUNGTRENNU(GN DMAT)·xthf fragment. This hinders approach of the
these reagents for immobilization on a silica surface. As
calcium is divalent we must selectively prepare sites com-
prising a single calcium-to-surface bond. This should leave a
catalytically active bond or at least a reactive functionality
amenable to activation. In this regard, the support of choice
appears to be Aerosil 380 that has been dehydroxylated at
008C under a secondary vacuum. This type of nonporous
silica, which we designate here as SiO -700, bears isolated si-
lanol groups as the sole type of protic site, a prerequisite for
synthesis of the targeted single-site catalysts. In addition,
molecular precursor and thus prevents reaction of residual
[11]
silanol groups. The contents of nitrogen (0.50 wt%) and
carbon (8.01 wt%) indicate N/Ca and C/N ratios of 1.15 and
18.51, respectively. This is consistent with the formation of
(ꢀSiO)Ca
ACHTUNGTRENNU(GN DMAT)·1.6thf surface species. Such a stoichiome-
try indicates that the silica surface, which can be regarded as
a ligand, is slightly bulkier than the bidentate DMAT ligand
(in comparison: 1 bears two THF ligands per Ca atom).
We also monitored DMAT-H formation during the graft-
7
1
ing process by H NMR spectroscopy on a mixture of Ca-
the high reactivity of complexes 1 and 2 towards the acidic
ACHTUNGTRNEN(GU DMAT) ·2thf (1) and SiO -700 in C D . As the concentra-
2 2 6 6
silanol groups in SiO -700 warrants efficient grafting. Con-
tion of DMAT-H does not increase after measurement of
the first spectrum (10 min) the reaction of the benzyl calci-
um complex with the silica support is fast. Quantification of
catalyst loading by using ferrocene as an internal standard
gives a Ca content of 1.29 wt%, a value close to that ob-
tained from elemental analysis (1.26 wt%). This confirms
that protonolysis of the calcium–carbon bond is indeed the
main reaction and that the proposed stoichiometry is cor-
rect.
2
sidering that the density of silanol groups is about
À2 [10]
1
.1 nm , which translates to an average spacing of 11 ꢃ,
treatment of SiO -700 with a solution of 1 or 2 would
2
merely give isolated heteroleptic calcium benzyl or amide
groups (Scheme 2). Furthermore, the lack of porosity of the
Aerosil support will avoid dependence of the catalytic per-
formance on diffusion of reagents and products to and from
surface active sites.
This protonolysis reaction was also confirmed by infrared
studies on SiO -1 (Figure 1). The IR spectrum of SiO -700
2
2
À1
shows a typical sharp isolated silanol signal at 3747 cm ,
which is not observed in IR spectra of SiO -1. Instead, a
2
À1
broad, low-intensity signal centered at 3620 cm indicates
the presence of residual silanol groups, as already deduced
from the measured calcium loading. The broadness of this
peak is most probably due to interaction with neighboring
[
11]
2
3
grafted species. The n
A
H
U
T
E
N
G
A
H
U
G
R
N
U
G
À1
À1
the 3097–3020 cm and 2981–2785 cm regions, respective-
ly, of the spectrum of SiO -1 are similar to those of the mo-
2
lecular precursor, as are the peaks related to aromatic rings
À1
À1
[12]
in the 1600 cm and 1490–1450 cm regions.
2 2
Scheme 2. Preparation of SiO -1 and SiO -2.
Chem. Eur. J. 2009, 15, 4382 – 4393
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4383

R. M. Gauvin, S. Harder et al.
features characteristic of the expected grafted species: a
sharp peak at d=0.0 ppm for the Me Si group, signals at d=
3
6
.8–6.0 ppm for the aromatic protons, a signal at d=2.6 ppm
originating from the Me N groups, and a broad distribution
2
between d=4 and 0.5 ppm that could be assigned to the
THF ligand(s) and to the methyne proton. The grafted spe-
cies contains both rigid (DMAT) and mobile (THF) moiet-
ies for which the relaxation behavior is expected to vary
considerably. Therefore, it may be possible to filter the sig-
1
1
nals in order to selectively observe a fragment. H– H
double-quantum MAS spectroscopy (DQ-MAS) provides
information about the spatial proximity of protons by means
of dipolar recoupling pulse sequences. In the BABA (back
[13]
to back) sequence, for example, rotor-synchronized pulses
are used to create double-quantum coherence and, after a t1
evolution period for 2D acquisition, convert them to single-
quantum coherences. The BABA sequence can also be used
to filter out the signals of mobile protons from those in a
rigid configuration. Indeed, the correlation information ob-
tained by increasing the number of rotor cycles is in direct
competition with the simultaneous relaxation phenomenon,
which occurs during pulse excitation and reconversion. Ap-
plying four rotor cycles for excitation and reconversion peri-
2 2 2
Figure 1. DRIFT spectra of SiO -700, SiO -2, and SiO -1.
In addition, SiO -1 was investigated by high-field state-of-
2
the-art solid-state NMR spectroscopy. This is a powerful
technique to obtain information on a material at the molec-
ular level. It provides information on the environment of
observed nuclei through their chemical shifts and also from
correlation between the different sites.
ods to SiO -1 affords a spectrum in which the THF proton
2
Solid-state NMR spectra were recorded on an 18.8 T
signals are of considerably lower intensity, and thus the
spectrum of the DMAT ligand alone is revealed. The fil-
tered spectrum allows for detection of the rather weak ben-
zylic methyne proton signal at d=0.9 ppm. (Figure 2d). The
1
spectrometer (800.13 MHz H frequency). Figure 2a pres-
1
ents the H MAS NMR spectrum of SiO -1. It has several
2
assignment of the ArCHSiMe proton signal is consistent
3
1
with the corresponding H NMR data of the molecular pre-
cursor 1 in solution (d=0.9 and 1.24 ppm). An intermediate
situation is observed for spectra recorded with the sequence
synchronized with one and two rotor cycles (Figure 2b and
c, respectively).
The spatial proximities between the protons of SiO -1 can
2
be assessed by using double-quantum correlation spectrosco-
py. Use of the back-to-back train pulse synchronized with
four rotor cycles affords a correlation map comprising sever-
al off-diagonal signals (Figure 3).
Two pairs associate the aromatic signals around d=
6
.8 ppm with the d=0.0 ppm and d=2.6 ppm peaks
(
Figure 3, top left: interactions A and B, respectively). The
third off-diagonal correlation involves the signals at d=2.6
and 0.0 ppm (interaction C). The observed proximity be-
tween aromatic signals and the Me Si or Me N protons is in
3
2
agreement with the structure of the proposed surface spe-
cies.
The third correlation between Me N and Me Si groups
2
3
may not be expected on account of their relatively large spa-
tial separation (crystal structure data suggest a distance of
4
–5 ꢃ). However, the fact that a considerable number of
protons are involved in this specific correlation (9H from
Me Si and 6H from Me N) might explain the observed
cross-peak. The lack of correlations between DMAT and
THF protons could be due to several factors. The broadness
of their signal could dilute the information in the spectral
noise, or the THF protons could be too “mobile” to effi-
1
Figure 2. H MAS NMR spectra of SiO
2
-1 acquired using a) a single-
3
2
pulse sequence, b) an excitation/reconversion pulse train of one rotor
period, c) two rotor periods, and d) four rotor periods (800.13 MHz,
0 kHz spinning speed, 64 transients except for a), for which 16 transients
were recorded). The excitation/reconversion trains were composed of
back-to-back pulses of 2.1 ms).
2
4384
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 4382 – 4393

Silica-Supported Calcium Reagents
FULL PAPER
1
1
Figure 3. H– H 2D double quantum MAS spectrum of SiO
2
-1 acquired with an excitation/reconversion pulse train of four rotor periods, expanded views,
and correlation assignments (800 MHz, 20 kHz spinning speed, pulse length of 2.1 ms, 16 transients were recorded for each of the 400 t1 increments, the
relaxation delay was set to 5 s).
ciently interact with others on the grafted species. This is in
line with our earlier conclusions based on one-dimensional
H NMR spectra using the BABA sequence (vide supra).
A close-up of the high-field region of the correlation spec-
trum reveals the presence of at least two other interactions,
designated here as D and E (Figure 3, top right). These con-
cern correlations between Me N and Me Si protons with the
2 3
ArCHSiMe proton resonating at d=0.9 ppm, and thus con-
3
1
firm its assignment. This is in agreement with the fact that
no significant on-diagonal signal is detected at this chemical
shift, that is, it originates from an isolated CH proton.
Chem. Eur. J. 2009, 15, 4382 – 4393
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4385

R. M. Gauvin, S. Harder et al.
An enlargement of the aromatic region shows cross-peaks
that allow assignment of the various signals (Figure 3,
bottom left). Three main off-diagonal interactions can be
observed: F, between d=6.8 and 6.4 ppm; G, between d=
6.6 and 6.0 ppm; and H, between d=6.4 and 6.0 ppm. We
also know that an intense correlation involves the proton
resonating at d=6.8 ppm with the Me Si group (interaction
3
A). A broader correlation (B, lower intensity) is observed
between the low-field aromatic protons at d=6.8–6.6 ppm
(
undistinguished) and the Me N substituent. This would
2
allow one to assign a chemical shift of d=6.8 ppm to the
3
CH group ortho to the CHSiMe group (C H), and of d=
3
13
2
Figure 4. C CP-MAS NMR spectrum of a) 1 (64 transients) and b) SiO -
6
6
.6 ppm to the CH ortho to the Me N moiety (C H). The re-
maining assignments can then be proposed on the basis of
the three interaromatic correlations: d=6.4 ppm for C H
and d=6.0 ppm for C H. This assignment would be in good
2
1 (2560 transients) at 100.62 MHz with a relaxation delay of 5 s. The con-
tact time was set to 1 ms at a radio-frequency field of 50 kHz. H decou-
pling at 90 kHz was applied during C acquisition.
1
4
13
5
agreement with solution data. Precursor 1 in solution (in
the THF ligands, which indicates their different environ-
ments. In contrast, the solution spectrum of 1 is composed
of a double set of signals for the DMAT ligands (indicating
slow exchange between the two diastereomers) but only one
set for the THF ligands (indicating fast exchange between
C D ) consists of two slowly interconverting diastereomers.
6
6
3
4
5
6
For the aromatic C H-C H-C H-C H part the following sets
of H NMR signals were observed: d=7.12, 6.89, 6.38, 6.72
1
and 7.18, 6.85, 6.55, 6.27 ppm. Both compare well to the 6.8,
[6]
6.4, 6.0, 6.6 ppm sequence proposed for the aromatic signals
the two different THF ligands).
1
3
in SiO -1.
The C CP-MAS NMR spectrum of SiO -1 comprises five
2
1
2
3
The C CP-MAS NMR spectra of 1 and SiO -1 are dis-
distinct patterns (Figure 4b). The aromatic carbon atoms
resonate in the range of d=130–110 ppm. Compared to the
spectrum of the molecular precursor, the two low-field aro-
matic signals from the DMAT quaternary carbon atoms are
not detected in the spectrum of the grafted species under
these experimental conditions. The coordinated THF mole-
2
1
3
played in Figure 4. The solid-state C spectrum of crystal-
line 1 features two sets of sharp well-resolved signals, ac-
counting for the aromatic carbon atoms (d=149.0–
1
10.9 ppm), CH and Me N groups (d=49.6–39.2 ppm), THF
2
a- (d=70.0 and 68.8 ppm) and b-carbon atoms (d=25.8 and
4.6 ppm), and Me Si moieties (d=4.5 and 2.5 ppm). As
2
cules give rise to two signals, at d=69.0 (C ) and 24.7 ppm
3
a
only one of the two diastereomers of precursor 1 is found in
the crystal structure (the meso form with R and S chirality
at the benzylic carbon atoms), the two sets of signals must
arise from the nonsymmetry of this particular diastereomer
(C ). The Me N and methyne carbon signals cannot be dis-
tinguished, as a broad peak is observed at d=43 ppm. The
b
2
[6]
Me Si carbon atoms resonate at d=À0.5 ppm, which is shift-
3
ed somewhat to high field compared to 1.
1
13
(
nonsymmetry is also observed in the crystal structure). This
The H– C CP HETCOR NMR spectrum of SiO -1 was
2
is confirmed by observation of a double set of signals for
recorded to confirm and refine the assignments (Figure 5).
1
13
Figure 5. a) H– C CP HETCOR NMR spectrum of SiO
2
-1. The CP conditions are identical to those used for the spectra in Figure 4. A total of 1024
transients was added for each of the 64 t1 increments. The relaxation delay was set to 5 s. The 2D processing included an exponential line broadening of
0 Hz in both dimensions. b) Close up of the Me N/methyne area.
2
2
4386
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 4382 – 4393

Silica-Supported Calcium Reagents
FULL PAPER
À1
This pulse sequence includes a cross-polarization contact
time to correlate carbon nuclei through space with neighbor-
ing protons. Accordingly, the spectrum comprises the ex-
pected correlations between protons and carbon atoms of
the aromatic ring, of THF a- and b-methylene groups (re-
spective HÀC couples: d=4.1/69.0, and 1.5/24.7 ppm), and
The absence of an IR band at 3747 cm , characteristic for
isolated silanol groups, indicates their complete consump-
tion (Figure 1). Infrared signals similar to those of the mo-
lecular precursor 2 can be observed for modified silica SiO -
2
À1
À1 [12]
2 (n
A
H
U
G
R
N
U
G
A
H
U
G
R
N
U
G
Regarding the fate of the ammonia released in the silylation
side reaction, no NH-related bands (3380 and 3320 cm )
are detected, that is, coordinated NH is not present in the
À1
of the Me Si fragment (HÀC couple: d=0/À0.5 ppm). As
3
1
3
shown in Figure 5, the C signal of the Me N and CH
2
3
groups at about d=43 ppm gives rise to several cross-peaks.
Three of them are indicative of MeN moieties in different
environments, at d /d of 3.3/41.8, 2.7/43.0, and 2.7/43.8 ppm
material. This also shows that the possible side reaction be-
tween CaN
ACHTUNGTRNENU(G SiMe ) and NH , which could give grafted
3 2 3
[16]
CaNH , did not occur under these reaction conditions.
H
C
2
(
a maximum of four MeN signals is expected). Only in the
Solid-state NMR spectra of SiO -2 display the expected
2
1
last-named couple is residual interaction with the Me Si
features for the proposed chemical composition. The
H
3
group detected, which may be an indication of the actual
configuration of the silica-grafted species. However, at this
stage, further refinement of the assignment appears difficult.
Most interestingly, a cross-peak associates the methynic
proton at d=0.9 ppm with a signal centered at d=42.7 ppm
MAS NMR spectrum comprises a sharp high-field signal at
d=À0.73 ppm with a shoulder at d=À0.57 ppm (Figure 6).
13
in the C dimension, fully in line with the observed values
for the CaCH moiety in solution: the two diastereomers
1
13
show H and C signals at d=0.89/1.24 ppm and 42.8/
[6]
47.1 ppm, respectively.
Reaction of Ca[N
ACHTUNGTRENNUNG( SiMe ) ] ·2thf (2) with SiO -700: synthe-
3 2 2 2
sis and characterization of SiO -2: Reaction of 2 with SiO -
2
2
7
00 suspended in pentane gave, after subsequent washing
À6
with pentane and drying under high vacuum (10 mmHg),
an off-white material (SiO -2). Grafting of 2 onto SiO -700
2
2
is expected to follow a similar pattern to grafting of the re-
[14]
lated lanthanide compounds Ln[N
ACHTUNGTRNNGE(U SiMe ) ] . In the latter
3 2 3
case, the monopodal species (ꢀSiO)Ln[N
AHCTUNGTRNNEUG( SiMe ) ] is
3 2 2
formed through protonolysis of a Ln–amido bond by a sila-
nol group, and a concomitant side reaction was observed in
which some of the silanol groups were silylated by the re-
leased (Me Si) NH to give ꢀSiOSiMe and NH (via transi-
3
2
3
3
ent Me SiNH ). Thus, a silanol-free surface was obtained in
3
2
[15]
fine.
Elemental analysis indicates a Ca loading of 1.72 wt%,
which corresponds to a surface density of 0.75 Ca per
1
Figure 6. Top: H MAS NMR spectrum of SiO
2
-2 (400.13 MHz, 12.5 kHz
square nanometer. In comparison, grafting of SiO -700 with
13
2
spinning speed, relaxation delay of 5 s, 16 transients); bottom: C CP
MAS NMR spectrum of SiO -2 (100.62 MHz, 12.5 kHz spinning speed,
À2
Ln[N
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(SiMe ) ] gave only a Ln content of 0.48nm (Ln=Y,
3
2
3
2
La, Nd, Sm). This difference most probably stems from the
considerably smaller size of the grafted CaN(SiMe ) group
compared to the Ln[N(SiMe ) ] group. This decrease in
1024 transients), with a relaxation delay of 5 s. The contact time was set
1
to 1 ms at a radio-frequency field of 50 kHz. H decoupling at 80 kHz
AHCTUNGTRENNUNG
3
2
1
3
was applied during the C acquisition.
AHCTUNGTRENNUNG
3
2 2
steric hindrance for the remaining silanol groups favors fur-
ther metalation versus trimethylsilylation. The modest steric
bulk of the CaN
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(SiMe ) group also explains a higher Ca
These signals are assigned to the Me Si groups of the N-
3
2
3
loading than observed for the grafted calcium benzyl species
A
T
N
T
E
N
N
(SiMe ) and ꢀSiOSiMe units, respectively. Broad signals
3
2
3
À2
(
0.55nm ). The Ca/N ratio of 0.98 is in agreement with the
due to the THF ligands are observed at d=3.17 and
formation of monografted surface species CaN
A
H
U
G
R
N
U
G
1.3 ppm, whereby the latter is partially masked by the Me Si
3
2
3
SiO -700 bears 1.1 OH per square nanometer, of which 0.75
peaks.
2
1
3
are bound to calcium, 0.35 silylated silanol groups (ꢀSiO-
Accordingly, the C CP-MAS NMR spectrum comprises
three main peaks at d=3.8, 24.8, and 69.5 ppm which corre-
SiMe ) per square nanometer should have been formed.
3
Thus, the C/Ca ratio of 12.1 corresponds to 1.3 THF per cal-
cium center. This is consistent with a surface species formu-
spond to Me Si and the b- and a-carbons of calcium-bound
THF moieties, respectively (Figure 6). Noteworthily, a high-
3
lated as (ꢀSiO)CaN
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(SiMe ) ·1.3thf.
field shoulder on the Me Si signal may be indicative of
3
2
3
ꢀ
SiOSiMe . The concomitant presence of CaNSiMe3 and
3
Chem. Eur. J. 2009, 15, 4382 – 4393
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4387

R. M. Gauvin, S. Harder et al.
ꢀ
SiOSiMe groups was confirmed by observation of the ex-
3
29
pected
Si CP-MAS NMR signals at d=À13.4 and
[
12,15,17]
12.4 ppm, respectively.
Molecular model of grafted calcium catalysts: The heteroge-
neous character of the grafted calcium amide and benzyl
species prevents a detailed study on their bonding situation.
As insights into the possible bonding modes involving the
silyl ether-rich silica surface and the metal would afford a
better understanding of the supported species, we investigat-
ed the synthesis and structural characterization of a model
system. A molecular ligand that mimics the isolated silanol
functionality on the surface of SiO -700 is the bulky ether-
2
[18]
rich silanol (tBuO) SiOH. Reaction of (tBuO) SiOH with
3
3
Ca ACHTUNGTRENNUNG( DMAT) ·2thf (1) or Ca[N CAHTUNGTENRNUN(G SiMe ) ] ·2thf (2) could give
2 3 2 2
a surface-like heteroleptic complex (Scheme 3).
2
+
À
À
2
Figure 7. Crystal structure of [Ca
] ACHTUNGRTENNUN[G (tBuO) SiO ] AHCTUNGTRENUNGN[ OH ] ·thf; hydrogen
4 3 6
atoms and methyl groups (tBu) have been omitted for clarity.
Table 1. Selected interatomic distances [ꢃ] in the crystal structure of
2
+
À
À
[
Ca ] ACHTGRUNETNUN[G (tBuO) SiO ] ACTHNUTRGENUNG[ OH ] ·thf.
4 3 6
2
Ca1ÀO1
Ca1ÀO5
Ca1ÀO9
Ca1ÀO10
Ca1ÀO25
Ca2ÀO9
Ca2ÀO13
2.128(4)
2.264(3)
2.390(5)
2.429(4)
2.366(3)
2.283(3)
2.427(3)
Ca2ÀO14
Ca2ÀO25
Ca2ÀO26
Ca2ÀO27
Ca3ÀO5
Ca3ÀO6
Ca3ÀO17
2.435(3)
2.345(5)
2.340(3)
2.386(7)
2.411(3)
2.410(3)
2.244(3)
Ca3ÀO25
Ca3ÀO26
Ca4ÀO13
Ca4ÀO17
Ca4ÀO18
Ca4ÀO21
Ca4ÀO26
2.288(3)
2.269(5)
2.227(3)
2.452(5)
2.416(4)
2.165(4)
2.380(3)
Scheme 3.
The equimolar reaction of either 1 or 2 with (tBuO) SiOH
was monitored by H NMR. The sharp silanol peaks
3
1
changed to broad signals, indicative of slow exchange. More
importantly, even though the correct 1/1 stoichiometry was
used, in both cases half of the calcium reagent did not react.
Apparently, the Schlenk equilibrium is fully to the side of
the homoleptic complexes (Scheme 3). This underscores the
importance of grafting as a tool in the syntheses of hetero-
leptic calcium reagents.
The approximate C symmetry is broken by coordination of
i
a THF ligand to Ca2. This structure type is also found in a
recently published aryloxide hydroxide calcium cluster of
2
+
À
À
composition [Ca 6A
] ACHTUNGTRENUN[GN ArO ] CHTUGNERTNNNU[G OH ] , which was obtained by
4
2
[
20]
controlled hydrolysis of Ca ACHTUNGTREN(NUNG OAr) . This particular frame-
2
work, which also has been observed in several other crystal
structures of aryloxide calcium complexes, must be a highly
The homoleptic complex Ca ACHTNUGTERNNGNU[ OSi ACTHUNGTRNENUNG( tBuO) ] could be pre-
3 2
[
21]
2+
À
pared in quantitative yield by reaction of 2 with two equiva-
lents of the bulky silanol. The product that remains after re-
moving all volatile materials is extremely soluble in pentane,
and even at À808C no crystals could be obtained. After
three weeks at room temperature, however, large colorless
crystalline blocks started to separate. The crystal structure
favored structural type. As suggested, the [M ]4
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
[RO ] -
6
À
A
H
U
G
R
N
U
G
2
[
20]
À
3
2
+
gands show additional coordination of a tBuO side arm to
2
+
À
2+
showed a complex of composition [Ca
]
A
H
U
G
R
N
U
G
one of the Ca ions. The CaÀO bonds for terminal ligands
4
À
3
6
À
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
[OH ] ·thf. The presence of an isolated OH ligand was con-
(2.128(4)–2.165(4) ꢃ) are shorter than those for bridging li-
2
firmed by the observation of a relatively sharp IR signal at
gands (2.244(3)–2.452(5) ꢃ). The terminal and bridging CaÀ
À1
3675 cm , typical for calcium hydroxide complexes (a re-
O bond lengths compare well to those in a recently pub-
[22]
cently reported dimeric b-diketiminate calcium hydroxide
lished silsesquioxane calcium complex. The CaÀO bonds
to the ether side arms are generally the longest (2.410(3)–
2.435(3) ꢃ), but still significantly shorter than those in the
À1 [19]
complex showed an IR signal at 3697 cm ). Apparently
the long crystallization time caused partial hydrolysis to give
a less soluble siloxide–hydroxide cluster.
[22]
silsesquioxane calcium complex (2.635(3)–2.844(4) ꢃ).
The crystal structure of this cluster is shown in Figure 7,
and selected interatomic distances are summarized in
Table 1. The framework of the structure can be best de-
scribed as two fused Ca/O cubes with two missing corners.
The crystal structure nicely shows the variety of bonding
modes that could play a role in silica-grafted organocalcium
species.
4388
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 4382 – 4393

Silica-Supported Calcium Reagents
FULL PAPER
Scheme 4. Catalytic applications of grafted calcium catalysts.
[
a]
Preliminary catalytic studies on SiO -1 and SiO -2: To ex-
3
Table 2. Catalytic hydrosilylation of activated alkenes by PhSiH .
2
2
plore the scope of silica-grafted calcium amide and benzyl
groups in catalysis, we investigated their catalytic activity in
the hydrosilylation of activated alkenes, intramolecular
alkene hydroamination, and in styrene polymerization (pro-
Entry Substrate Product
Catalyst
T
t
Conversion
[%]
[
b]
[8C] [h]
SiO
1
2
2
-1
-2
20
20
<0.1
<0.1
>98
>98
1
2
R
posed catalytic cycles are shown in Scheme 4; L represents
SiO
2
50
50
1
1
>98
>98
the reactive ligand DMAT or N
Ca(DMAT) ·2thf has been shown to be an effective cata-
lyst in the homogeneous hydrosilylation of conjugated al-
ACHTUNGTNERUNNG( SiMe ) ).
3 2
ACHTUNGTRENNUNG
2
SiO
1
2
-1
50
20
1.5
>98
>98
3
4
<0.1
[4e]
kenes.
Apart from considerable activity, it also showed
very good regioselectivity. In some cases the regioselectivity
could be controlled conveniently by choice of solvent or by
SiO
2
2
2
-2
-1
50
50
2 (16)
1
93 (98)
>98
[4e]
switching to more polar catalysts (based on Sr or K). Also
Ca[N(SiMe ) ] ·2thf shows good activity in this highly atom
efficient key transformation.
Supported catalysts SiO -1 and SiO -2 were tested in the
SiO
1
50
50
0.5 (36)
16
35 (47)
>98
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
5
6
3
2 2
[4j]
SiO -2
2
50
1 (24)
30 (63)
2
2
hydrosilylation of styrene, cyclohexadiene, and 1,1-diphenyl-
ethylene (DPE) with PhSiH . The results are compared with
[c]
SiO
1
2
-1
-2
50
50
2
3
>98
>98
3
7
those of the ungrafted catalysts under homogeneous condi-
tions (Table 2). The reactions were carried out with 5 mol%
catalyst loading (based on active groups) in benzene or THF
at temperatures ranging from 20 to 508C. The conversion of
[
c]
8
SiO
2
50
3
>98
1
[a] General conditions: 0.5 mmol alkene, 0.5 mmol PhSiH , 300 mg sol-
vent, catalyst concentration calculated per active Ca amide or Ca benzyl
substrates to products was monitored by H NMR spectros-
3
copy. The following conclusions can be drawn: 1) In all
cases initiation of the catalyst takes place by reaction of
fragment:
0.036 mmol). [b] Measured by NMR spectroscopy. [c] Solvent=
[D ]THF.
2 2
1 and 2 (0.025 mmol), SiO -1 (0.026 mmol), and SiO -2
(
R
R
(
SiO )CaL with PhSiH . The product PhH SiL could be
2 3 2
8
detected, and presumably the actual catalyst is a grafted cal-
cium hydride: (SiO )CaH (Scheme 4). 2) The observed con-
2
versions after a given time are in most cases similar to or
slightly lower than those found under homogeneous condi-
tions. An exception is found for DPE, the most sterically
hindered alkene, which under homogeneous conditions also
reacts more slowly than styrene or cyclohexadiene (at least
under apolar conditions). 3) Similar to runs under homoge-
neous conditions, the regioselectivity for the hydrosilylation
of DPE depends on the polarity of the solvent: in an apolar
medium exclusively branched products are formed (Table 2,
entries 5 and 6), whereas in THF the linear silanes were ob-
tained (Table 2, entries 7 and 8). Formation of the latter
The catalytic activity of SiO -1 and SiO -2 in the hydrosi-
2
2
lylation of alkenes shows that the silica surface bears reac-
tive calcium benzyl and amide functionalities. However, it is
hitherto unclear whether catalysis proceeds in a homogene-
ous or heterogeneous fashion. The fact that bulky substrates
like DPE react considerably more slowly with silica-grafted
catalysts is an indication for a heterogeneous process.
During the catalytic hydrosilylation of DPE, the solid cata-
lyst turns bright red, whereas the solvent (benzene or THF)
is colorless. This is a strong indication for a (ꢀSiO)CaC-
[4e]
should be explained by an alternative mechanism.
(Me)Ph resting state.
2
Chem. Eur. J. 2009, 15, 4382 – 4393
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4389

R. M. Gauvin, S. Harder et al.
As even prolonged reaction times never lead to full con-
version in the hydrosilylation of DPE, for slower reactions
catalyst decomposition can be an issue. A possible route for
catalyst decomposition could be reaction of the proposed
surface hydride (ꢀSiO)CaH·nthf with a ꢀSiÀOÀSiꢀ bond of
product released we can conclude that 44% of all initial
active sites bind a cyclic amide. As we could not detect a
coupling pattern for CH D, it is likely that the catalytic in-
2
termediate in Scheme 4 isomerizes by a 1,3-H shift to a
more stable calcium amide complex, which is released after
the silica to give (ꢀSiO) Ca·nthf and ꢀSiH (Scheme 5a).
CD OD addition. We did not find any released substrate
2
3
H C=CHCH C(Ph) CH NH on addition of CD OD. There-
2
2
2
2
2
3
fore, it is likely that the product was chemically bound to
the surface and not physisorbed, as in this case desorbed
products would also comprise a fraction of unconverted sub-
strate bearing in mind the conversion of 90%.
The grafted benzyl calcium catalyst was also investigated
in the polymerization of styrene. Earlier work showed that
benzyl calcium complexes are efficient initiators for living
[4a]
polymerization of styrene. Whereas the homoleptic calci-
um catalyst Ca(DMAT) ·2thf (1) gives largely atactic poly-
styrene, polymer obtained with the heteroleptic initiator Ca-
(DMAT)(9-Me Si-fluorenyl)·thf shows a remarkable syndio-
AHCTUNGTRENNUNG
2
AHCTUNGTRENNUNG
3
tacticity which strongly depends on the styrene concentra-
[23]
tion (in pure styrene 92% r diads are found). Recently,
we found that extending the fluorenyl ligand by substitution
at the 2- and 7-positions gave a large increase in tacticity
control: even under dilute conditions syndiotacticities up to
Scheme 5. Proposed reaction paths accounting for surface hydride deacti-
vation.
[24]
9
2% (r diads) were found. As the silica surface can be
R
Cleavage of the ꢀSiOÀCa(L ) bond by PhSiH to give
seen as a very extensive ligand, efficient tacticity control in
styrene polymerization is expected.
3
R
ꢀ
SiOÀSiH Ph and HCa(L ) may also be envisioned
2
(
Scheme 5b). However, in the latter case a soluble calcium
Polymerization of styrene with the initiator SiO -1 was
2
hydride would be released from the surface and higher cata-
lytic activity should be observed.
In the case of entry 5 (Table 2), reloading an isolated and
either performed in solution (10% styrene in cyclohexane,
508C) or in the bulk (100% styrene, 208C). Whereas solu-
tion polymerizations with initiator 1 give full conversion of
monomer within 30 min, polymerization with grafted initia-
washed SiO -1 catalyst with substrates gave further conver-
2
sion, although at a much lower rate (20% after 30 h). This
could be either due to catalyst deactivation (Scheme 5a) or
cleavage (Scheme 5b). Analysis of the used catalyst by IR
and MAS NMR spectroscopy show indications for forma-
tor SiO -1 is considerably slower. After one hour only 4%
2
polymer could be isolated. Polymerization over 20 h gave
18% conversion. The much slower polymerization rates are
likely due to the fact that monomer transport to the surface-
bound active sites is increasingly hindered by the growing
polymer chains. The obtained polymer also shows a some-
[12]
tion of ꢀSiH surface species, but not ꢀSiOÀSiH Ph.
2
Therefore, catalyst deactivation is the more likely decompo-
sition route.
what broadened molecular weight distribution (M =1.68ꢄ
n
4
[12]
The proposed catalyst in the hydrosilylation of alkenes is
a monomeric surface hydride, which is expected to be highly
reactive. Calcium-mediated intramolecular hydroamination
10 , PDI=1.45) indicative of different active sites during
1
3
polymerization. Analysis by C NMR spectroscopy, howev-
er, revealed a considerable syndiotacticity of 88% (r diads).
These observations all point to a heterogeneous polymeri-
zation reaction, as leaching of the grafted catalyst would
give essentially atactic polystyrene.
[4e,g,i]
of aminoalkenes
proceeds through a calcium amido in-
termediate (Scheme 4), which should be much less reactive
and less prone to decomposition. We investigated the intra-
molecular hydroamination of H C=CHCH C(Ph) CH NH
2
Polymerization in neat styrene gave a polymer with a bi-
modal molecular weight distribution, the main fraction of
2
2
2
2
and found that catalyst SiO -2 catalyzes ring closure to give
2
5
À1
the expected product (90% after 16 h, 10 mol% cat., C D ,
which was a polymer with an M of 1ꢄ10 gmol , whereas
6
6
n
2
08C). This is considerably slower than under homogeneous
the smaller fraction shows an extremely high M value of 2ꢄ
n
6
À1 [12]
conditions with catalyst 2 (>98% after 1 h, 10 mol% cat.,
C D , 208C). Apparently the ring-closure process on the
10 gmol . We assign the high molecular weight fraction
to leaching of a small part of the catalyst into the solution
to give a homogeneous propagating species which can grow
6
6
silica surface is sterically hindered. Neither the isolated cata-
lyst nor the mother liquor is catalytically active towards a
second loading of substrate. There is, however, an indication
that the catalytic reaction is heterogeneous. A used catalyst
was isolated, washed, and suspended in C D . Addition of
1
3
rapidly. The C NMR analysis of the polymer confirms in-
volvement of at least two propagating species: the signal in-
tensities do not follow Bernoullian statistics. Instead signals
for a polymer of considerable syndiotacticity (from hetero-
geneous polymerization) overlap with signals from a fully
atactic polymer (from homogeneous polymerization). Be-
6
6
1
CD OD led to immediate appearance of H NMR signals
for the ring-closure product. From quantification of the
3
4390
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 4382 – 4393

Silica-Supported Calcium Reagents
FULL PAPER
cause styrene is less easy to dry than cyclohexane, leaching
of the propagating species might proceed by reaction with
water. This could generate mobile Ca(OH)2 which, in
Schlenk equilibrium with a grafted propagating species,
could form mobile growing chains.
spectrometer. Solid-state MAS NMR spectra were recorded on a Bruker
1
13
Avance 400 spectrometer ( H: 400.13 MHz, C: 100.62 MHz) and
1
Avance II 800 ( H: 800.13 MHz). Chemical shifts are given with respect
1
13
to TMS for H and C. Proton decoupling was performed by using the
[25]
SPINAL64 composite pulse sequence.
Diffuse-reflectance infrared
spectra were collected by using a Harrick cell on a Nicolet Avatar spec-
trometer fitted with a MCT detector. Elemental analyses were carried
out at the Service Central d’Analyse du CNRS (Ca), and in the Service
dꢅAnalyse Elꢀmentaire, LSEO, Universitꢀ de Bourgogne (C, H, N). Aer-
Conclusion
2
À1
osil 380 silica (Degussa, specific area 380 m g prior to heat treatment)
À6
was heated under secondary vacuum (10 mmHg) at 5008C over 15 h
Homoleptic calcium benzyl or silylamide reagents react
readily with isolated surface silanol groups of partially dehy-
drated silica. Appropriate support pretreatment allows se-
lective formation of monopodal, heteroleptic species (ꢀ
followed by 4 h of heating at 7008C (designated here as SiO
and stored in a glove box. The starting materials Ca(DMAT) ·2thf,
and
2
-700),
[
6]
A
H
U
G
R
N
N
2
[
7]
[26]
Ca[N
tBuO)
Synthesis of SiO
A
H
U
G
R
N
U
G
3
)
2
[
]
2
18b]
·2thf,
1-amino-2,2-diphenyl-pent-4-ene,
(
3
were prepared according to reported procedures.
R
2
-1: In a glove box, one leg of a double-Schlenk appara-
tus was loaded with molecular precursor 1 (555 mg, 9.3ꢄ10 mol) dis-
SiO)Ca(L )·nthf with a reactive calcium–ligand bond and a
À4
spectator SiOÀCa linkage. Due to the bulk of these calcium-
based fragments, full coverage of the surface is not reached.
Whereas unconverted residual silanol groups are still pres-
ent on the surface after grafting of the benzyl derivative,
complete capping with trimethylsilyl groups is observed in
the amido case. The grafted calcium benzyl and amide frag-
ments have been characterized with a variety of analytical
methods (IR, 1D, and 2D solid-state NMR, elemental analy-
sis).
Preliminary results show that the grafted calcium benzyl
and amide groups are active in catalytic hydrosilylation of
alkenes and intramolecular hydroamination of alkenes. Al-
though recycling of the catalysts was not very efficient, key
experiments demonstrate that reactions were catalyzed by
supported calcium species. Moreover, grafted calcium
benzyl species can be used to initiate styrene polymeri-
zation. Simple grafting of the homoleptic initiator Ca-
solved in 20 mL of toluene. The other leg was charged with SiO
2
-700
(
1.0 g) suspended in 15 mL of toluene. The yellow solution of 1 was
added to the silica support by filtering through the sintered glass separat-
ing the two Schlenk tubes. The reaction mixture was stirred for 15 h at
room temperature. The supernatant liquid was then separated by filtra-
tion into the other leg, from which toluene was transferred by trap-to-
trap distillation back into the leg containing the modified support in
order to wash away residual molecular precursor. This operation was re-
peated four times until colorless washing fractions were obtained. The re-
sulting yellow powder SiO
2
-1 was then dried under secondary vacuum
À6
(
3.10 mmHg) at 408C for 6 h. Elemental analysis (%, average of two
measurements): C 8.19, H 1.17, N 0.51, Ca 1.26. The amount of DMAT-
H released during grafting was quantified by the following procedure: in
a glove box, an NMR tube was loaded with SiO -700 (38 mg), ferrocene
2
(6.5 mg, 34.9 mmol), and 1 (15 mg, 25.1 mmol). The sample was then
sealed and vigorously shaken. Due to the orange coloration from ferro-
cene, no color change could be observed. After deposition of the sus-
1
pended silica, the H NMR spectrum was recorded (no further evolution
was observed). From the relative integration of the protons of ferrocene
and of the SiMe
3
groups of DMAT-H (100:32), it can be calculated that
A
C
H
T
U
N
G
T
R
E
N
N
U
N
G
(DMAT) ·2thf (1), which produces mostly atactic polystyr-
12.4 mmol of 1 reacted with silica, which corresponds to a mass of calcium
2
ene, gave a heterogeneous initiator that produced polystyr-
of 0.50 mg. Considering the initial mass of SiO
of 1.29 wt%.
2
, this gives a Ca loading
ene of considerable syndiotacticity, albeit at the expense of
lower activity. This demonstrates the potential of grafting as
a simple tool to enforce higher selectivity in catalytic trans-
formations.
Synthesis of SiO
2
-2: White material SiO
2
-2 was prepared by the proce-
dure described above for SiO
2
-1 from molecular precursor 2 (500 mg, 9.9
À4
1
0
2
mol) and SiO -700 (1.15 g). Elemental analysis (%, average of two
measurements): C 6.23, H 1.32, N 0.59, Ca 1.72.
The intended synthesis of a molecular model system for
the grafted species did not give the heteroleptic product, but
instead ligand exchange in solution resulted in formation of
homoleptic products. This emphasizes the fact that grafting
of organocalcium reagents on the silica surface successfully
prevented formation of homoleptic species by the Schlenk
equilibrium. Thus, grafting could be a powerful new method
in the organometallic chemistry of the heavier alkaline-
earth metals Ca, Sr, and Ba, which is often plagued by
Schlenk equilibria. It could also prevent aggregation and
thus enable the syntheses of more reactive monomeric alka-
line earth metal reagents.
2+
À
À
Synthesis of [Ca
.15 mmol) was added to a solution of Ca[N
0.075 mmol) in 0.50 mL of C D . After 30 min full conversion of Ca[N-
]
4
A
H
U
G
R
N
U
[(tBuO)
3
SiO ]
6
A
H
U
G
R
N
N
[OH ]
2
·thf: (tBuO)
3
SiOH (40 mg,
0
A
H
U
G
R
N
N
3 2 2
) ] ·2thf (38 mg,
6
6
3 2 2
AHCTUNGERTUNNNG( SiMe ) ] ·2thf was observed in the NMR spectra. The solvents were re-
moved by vacuum evaporation and the remaining precipitate was dried
under high vacuum and dissolved in small amounts of pentane. After
0 days the product was isolated in the form of large, slightly yellow, crys-
talline blocks (yield: 27 mg, 59%). M.p. 1538C (decomp). H NMR
(300 MHz, C D , 208C): d=3.68 (m, 4H; thf), 1.53 (brs, 172H;
2
1
6
6
(tBuO)
3
SiO), 1.73 ppm (m, 4H; thf); the OH signals could not be ob-
served, likely on account of their low intensity (2H) and broad character;
1
3
C NMR (75 MHz,
(CH ); IR (Nujol): n˜ =3675 (OH), 1193, 976, 699 cm ; elemental analy-
sis calcd (%) for C76 (1847.0): C 49.42, H 9.39; found: C
8.72, H 9.12.
General procedure for catalytic hydrosilylation of alkenes: All alkene
6 6 3 3
C D , 208C): d=32.4 (C ACHUNTGRENNU(G CH ) ), 73.4 ppm (C-
À1
A
H
U
G
R
N
U
G
3 3
)
H
4 6
172Ca O27Si
4
substrates were dried by stirring over freshly ground CaH
2
overnight. In
À1
Experimental Section
a dry Schlenk tube 83 mg of the catalyst (SiO -1 (0.315 mmolg ):
2
À1
0
.026 mmol; SiO
2
-2 (0.430 mmolg ): 0.036 mmol) was suspended in
General considerations: Manipulations were carried out under an argon
atmosphere in a glove box or by using Schlenk techniques. Solvents were
dried with conventional reagents and stored in a glove box over 3 ꢃ mo-
lecular sieves. Solution NMR analyses were run on a Bruker Avance 300
300 mg of C (or [D
6
D
6
8
]THF). Subsequently, PhSiH
3
(0.5 mmol) and the
alkene substrate (0.5 mmol) were added and the resulting suspension was
stirred and heated to 508C. Generally, a color change to red was ob-
served at the beginning of the reaction. The conversion was followed by
Chem. Eur. J. 2009, 15, 4382 – 4393
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4391

R. M. Gauvin, S. Harder et al.
1
taking samples at regular time intervals and analyzing them by H NMR
spectroscopy and GC/MS (see Supporting Information).
[2] a) F. J. Feher, T. A. Budzichowski, Polyhedron 1995, 14, 3239–3253;
b) R. Murugavel, A. Voigt, M. Ganapati Walawalkar, H. W. Roesky,
Chem. Rev. 1996, 96, 2205–2236; c) R. Duchateau, Chem. Rev. 2002,
Procedure for catalytic intramolecular hydroamination: 1-Amino-2,2-di-
phenylpent-4-ene was dried overnight by stirring a solution in hexane at
1
02, 3525–3542.
[
3] a) L. Bourget-Merle, M. F. Lappert, J. R. Severn, Chem. Rev. 2002,
102, 3031–3066; b) C. Floriani, Chem. Eur. J. 1999, 5, 19–23; c) L.
Turculet, T. D. Tilley, Organometallics 2004, 23, 1542–1553; d) C.
Janiak, H. Schumann, Adv. Organomet. Chem. 1991, 33, 291–393.
4] a) S. Harder, F. Feil, K. Knoll, Angew. Chem. 2001, 113, 4391;
Angew. Chem. Int. Ed. 2001, 40, 4261; b) M. H. Chisholm, J. Galluc-
ci, K. Phomphrai, Chem. Commun. 2003, 48; c) Z. Zhong, S. Schnei-
derbauer, P. J. Dijkstra, M. Westerhausen, J. Feijen, Polym. Bull.
2003, 51, 175; d) M. R. Crimmin, I. J. Casely, M. S. Hill, J. Am.
Chem. Soc. 2005, 127, 2042; e) F. Buch, J. Brettar, S. Harder, Angew.
Chem. 2006, 118, 2807; Angew. Chem. Int. Ed. 2006, 45, 2741;
f) M. R. Crimmin, A. G. Barrett, M. S. Hill, P. A. Procopiou, Org.
Lett. 2007, 9, 331; g) S. Datta, P. W. Roesky, S. Blechert, Organome-
tallics 2007, 26, 4392; h) F. Buch, S. Harder, Organometallics 2007,
6
08C on freshly ground CaH
2
. In a dry Schlenk tube the catalyst (SiO
2
-2
À1
(
166 mg, 0.430 mmolg , 0.072 mmol) was suspended in C
6
D
6
(600 mg)
and 1-amino-2,2-diphenylpent-4-ene (0.5 mmol) was added. The conver-
sion was followed by taking samples at regular time intervals and analyz-
ing them by H NMR spectroscopy (see the Supporting Information). In
one experiment after full conversion (>98%) the solid catalyst was sepa-
rated, washed with hexane once, and dried under high vacuum. The dried
together with cyclohexane as an internal
standard. After recording a H NMR spectrum, the sample was quenched
3
with CD OD. The H NMR spectrum of the quenched sample showed
1
[
catalyst was suspended in C
6
D
6
1
1
the appearance of the product signals.
Procedure for the polymerization of styrene: Polymerization of styrene
was performed in a thermostated 100 mL stainless steel Bꢆchi reactor at
normal pressure in cyclohexane (1m styrene solution, 1 mm catalyst SiO -
2
based on reactive DMAT groups, 508C) or in pure styrene (1 mm cata-
-1 based on reactive DMAT groups, 208C). The reactor was
and distilled from
nBuLi) and dry styrene (11.5 mL, ca. 100 mmol, freshly distilled from
CaH and stored over alox pearls). A suspension of the initiator SiO -1
317 mg, 0.1 mmol in 1 mL cyclohexane) was added via a port. The usual
appearance of a red color indicated that the polymerization started im-
mediately. After a given polymerization time (between 1 and 20 h), the
mixture was quenched with oxygen-free methanol. Evaporation of all sol-
vents and drying under high vacuum (0.01 Torr, 1208C, 2 h) yielded the
2
6, 5132; i) A. G. Barrett, M. R. Crimmin, M. S. Hill, P. B. Hitch-
1
cock, G. Kociok-Kçhn, P. A. Procopiou, Inorg. Chem. 2008, 47,
lyst SiO
2
7
366–7376; j) F. Buch, S. Harder, Z. Naturforsch. B 2008, 62, 169–
loaded with dry cyclohexane (90 mL, dried with CaH
2
1
77.
[
5] W. Schlenk, W. Schlenk, Jr., Ber. Dtsch. Chem. Ges. 1929, 62, 920–
24.
2
2
9
(
[
[
6] S. Harder, F. Feil, A. Weeber, Organometallics 2001, 20, 1044–1046.
7] M. Westerhausen, W. Schwarz, Z. Anorg. Allg. Chem. 1991, 604,
1
27.
[
8] For an rare example of immobilization of an alkaline earth metal
amide on mesoporous silica, see: a) C. Zapilko, R. Anwander, Stud.
Surf. Sci. Catal. 2005, 158, 461–468. For selected examples of graft-
ing of transition-metal benzyl complexes, see: b) D. G. H. Ballard,
E. Jones, R. J. Wyatt, R. T. Murray, P. A. Robinson, Polymer 1974,
1
3
polymer. The polymers were analyzed by GPC and C NMR spectrosco-
py ([D ]tetrachloroethane). The tacticity of the polymer was checked by
analyzing the C NMR signal for Cipso in the phenyl ring.
2
1
3
[12,23]
2
+
À
À
Crystal structure of [Ca ] AHCTUNGTERUNGNN[ (tBuO) SiO ] ACTHNUTGNRUEGN[ OH ] · ACHTNUTGERNNNU(G thf): X-ray diffraction
data were measured on a Siemens SMART CCD diffractometer at
4 3 6
2
1
5, 169–174; c) N. G. Maksimov, G. A. Nesterov, V. A. Zakharov,
[
27]
P. V. Stchastnev, V. F. Anufrienko, Y. I. Yermakov, J. Mol. Catal.
1978, 4, 167–179; d) S. Nemana, B. C. Gates, Langmuir 2006, 22,
À708C, and the structure solved with SHELXS-97
SHELXL-97. PLATON was used for geometry calculations and
graphics. Crystal data: C76 , M =1847.04, triclinic, space
group P1, a=14.0780(13), b=14.8270(13), c=15.6881(15) ꢃ, a=
and refined with
[
28]
[29]
8
214–8220; for selected examples of grafting of transition-metal
H
172Ca
4
O
27Si
6
r
amido complexes, see: e) Q. Zhuang, K. Tanaka, M. Ichikawa, J.
Chem. Soc. Chem. Commun. 1990, 1477–1478; f) R. Anwander, R.
Roesky, J. Chem. Soc. Dalton Trans. 1997, 137–138; g) A. O. Bouh,
G. L. Rice, S. L. Scott, J. Am. Chem. Soc. 1999, 121, 7201–7210.
9] a) J. M. Thomas, R. Raja, D. W. Lewis, Angew. Chem. 2005, 117,
3
9
1
2.338(5), b=107.983(5), g=117.059(5)8, V=2709.5(5) ꢃ , Z=1, 1calcd =
.131 Mgm , F ACHTUNGTRENNUNG( 000)=1006, m ACTHUNGTRENNUNG( MoKa)=0.328 mm . Of the 53379 mea-
sured reflections, 22610 were independent (Rint =0.036, qmax =27.78) and
8252 observed [I>2s(I)]. The final refinement converged to R1=
.0696 for [I>2s(I)], wR2=0.2064, and GOF=0.99 for all data. The
À3
À1
[
1
0
6
614–6641; Angew. Chem. Int. Ed. 2005, 44, 6456–6482; b) C.
Copꢀrt, M. Chabanas, R. Petroff Saint-Arroman, J.-M. Basset,
Angew. Chem. 2003, 115, 164–191; Angew. Chem. Int. Ed. 2003, 42,
final difference Fourier synthesis gave min./max. residual electron density
À3
of À0.57/+0.71 eꢃ . All hydrogen atoms were placed on calculated po-
1
56–181.
sitions and were refined in a riding mode. The structure was refined as
an inversion twin and the Flack parameter converged to 0.49(3). On ac-
count of the poor crystal quality, the hydrogen atoms of the two hydrox-
ide ligands could not be located. CCDC-703000 contains the supplemen-
tary crystallographic data for this paper. These data can be obtained free
of charge from The Cambridge Crystallographic Data Centre via
www.ccdc.cam.ac.uk/data_request/cif.
[
[
10] L. T. Zhuravlev, Langmuir 1987, 3, 316–318.
11] a) E. Le Roux, M. Chabanas, A. Baudouin, A. de Mallmann, C. Co-
pꢀret, E. A. Quadrelli, J. Thivolle-Cazat, J.-M. Basset, W. Lukens,
A. Lesage, L. Emsley, G. J. Sunley, J. Am. Chem. Soc. 2004, 126,
1
2
3391–13399; b) R. M. Gauvin, A. Mortreux, Chem. Commun.
005, 1146–1148.
[
[
12] See the Supporting Information.
13] I. Schnell, S. P. Brown H. Y. Low, H. Ishida, H. W. Spiess, J. Am.
Chem. Soc. 1998, 120, 11784–11795.
[
[
14] R. M. Gauvin, L. Delevoye, R. Ali Hassan, J. Keldenich, A. Mor-
treux, Inorg. Chem. 2007, 46, 1062–1070.
15] a) N. R. E. N. Impens, P. van der Voort, E. F. Vansant, Microporous
Mesoporous Mater. 1999, 28, 217–232; b) W. Hertl, M. L. Hair, J.
Phys. Chem. 1971, 75, 2181–2185; c) D. W. Sindorf, G. E. Maciel, J.
Phys. Chem. 1982, 86, 5208–5219.
Acknowledgements
We thank the CNRS and MNESR for financial support and Bertrand
Revel (Centre Commun de Mesures RMN, USTL) for his assistance with
NMR spectroscopy. Nord/Pas de Calais Region, Europe (FEDER),
CNRS, French Ministry of Science, USTL, and ENSCL are thanked for
funding the Bruker 18.8 T spectrometer. The Deutsche Forschungsge-
meinschaft is kindly acknowledged for partial funding of this project in
[16] The CaN
3 2
ACHTUNGTNERNUN(G SiMe ) functionality has been shown to react with a large
excess of NH to give a CaNH functionality: C. Ruspic, S. Harder,
3
2
Inorg. Chem. 2007, 46, 10426–10433. The absence of ammonialysis
2
the framework of the Advanced Materials priority program (AM -net).
during the synthesis of SiO -2 may be due to the much lower con-
2
Prof. Dr. R. Boese and D. Blꢁser are thanked for measurement of the X-
ray diffraction data.
centration of NH3 formed in situ.
[17] M. Westerhausen, Inorg. Chem. 1991, 30, 96–101.
[
3
18] a) For use of (tBuO) SiOH as model for silica, see: A. Fischbach,
M. G. Klimpel, M. Widenmeyer, E. Herdtweck, W. Scherer, R. An-
[
1] J. Guzman, B. C. Gates, Dalton Trans. 2003, 3303–3318.
wander, Angew. Chem. 2004, 116, 2284–2289; Angew. Chem. Int.
4392
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 4382 – 4393

Silica-Supported Calcium Reagents
FULL PAPER
Ed. 2004, 43, 2234–2239; b) For its synthesis, see: J. Beckmann,
Appl. Organomet. Chem. 2003, 17, 52–62.
[24] D. F.-J. Piesik, K. Hꢁbe, S. Harder, Eur. J. Inorg. Chem. 2007, 5652–
5661.
[
[
[
19] C. Ruspic, S. Nembenna, A. Hofmeister, J. Magull, S. Harder, H. W.
Roesky, J. Am. Chem. Soc. 2006, 128, 15000–15004.
20] W. Teng, M. Guino-o, J. Hitzbleck, U. Englich, K. Ruhlandt-Senge,
Inorg. Chem. 2006, 45, 9531–9539.
21] a) V.-C. Arunasalam, I. Baxter, S. R. Drake, M. B. Hursthouse,
K. M. Abdul Malik, D. J. Otway, Inorg. Chem. 1995, 34, 5295–5306;
b) J. Utko, S. Przybylak, L. B. Jerzykiewicz, S. Szafert, P. Sobota,
Chem. Eur. J. 2003, 9, 181–190.
[25] M. Ernst, J. Magn. Reson. 2003, 162, 1–34.
[26] S. Hong, S. Tian, M. V. Metz, T. J. Marks, J. Am. Chem. Soc. 2003,
125, 14768.
[27] G. M. Sheldrick, SHELXS-97, Program for Crystal Structure Solu-
tion, 1997, Universitꢁt Gçttingen, Germany.
[
28] G. M. Sheldrick, SHELXL-97, Program for Crystal Structure Re-
finement, 1997, Universitꢁt Gçttingen, Germany.
[
29] A. L. Spek, PLATON, A Multipurpose Crystallographic Tool 2000,
Utrecht University, Utrecht, The Netherlands.
[
[
22] V. Lorenz, S. Blaurock, F. T. Edelmann, Z. Anorg. Allg. Chem. 2008,
6
34, 441.
Received: December 1, 2008
Published online: March 13, 2009
23] F. Feil, S. Harder, Macromolecules 2003, 36, 3446–3448.
Chem. Eur. J. 2009, 15, 4382 – 4393
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
4393