Surface organometallic chemistry of titanium on silica-alumina and catalytic hydrogenolysis of waxes at low temperature

Article abstract of DOI:10.1021/om900151a

Ti(CH₂tBu)₄ (1) reacts selectively with the surface silanols of a silica-alumina partially dehydroxylated at 500 °C to provide the monosiloxy species [(=SiO)Ti(CH₂tBu)₃]_SA (2a) and the bisiloxy specie

Full text of DOI:10.1021/om900151a

Organometallics 2009, 28, 5647–5655 5647
DOI: 10.1021/om900151a
Surface Organometallic Chemistry of Titanium on Silica-Alumina and
Catalytic Hydrogenolysis of Waxes at Low Temperature
Cherif Larabi, Nicolas Merle, S ꢀe bastien Norsic, Mostafa Taoufik, Anne Baudouin,
Christine Lucas, Jean Thivolle-Cazat, Aimery de Mallmann,* and Jean-Marie Basset*
Universit eꢀ Lyon 1, Institut de Chimie de Lyon, Lyon, France, Universit eꢀ Lyon 1, CPE Lyon, CNRS,
UMR 5265 C2P2, LCOMS, B a^ timent 308 F - CPE Lyon, 43 Boulevard du 11 Novembre 1918, F-69616
Villeurbanne, France
Received February 25, 2009
Ti(CH tBu) (1) reacts selectively with the surface silanols of a silica-alumina partially dehy-
2
4
droxylated at 500 °C to provide the monosiloxy species [(tSiO)Ti(CH tBu) ] (2a) and the bisiloxy
2
3 SA
species [(tSiO) Ti(CH tBu) ] (2b) in a ca. 40:60 ratio, with concomitant evolution of 1.6 ( 0.2
2
2
2 SA
equiv of neopentane per Ti. These surface complexes were characterized via the combined use of
1
13
1
13
several techniques such as IR spectroscopy, H MAS, C-CP/MAS, 2D H- C HETCOR, and
J-resolved solid-state NMR, as well as mass balance analysis. By treatment under hydrogen at 150 °C
the neopentyl ligands in complexes 2a,b undergo hydrogenolysis and a mixture of supported titanium
1
species is obtained. IR, ESR, H MAS, and DQ solid-state NMR spectroscopies show the pre-
sence of ca. 3% [(tSiO)(M O)TiH ] (3a; M = Si, Al), 5% [(tSiO)(M O)Ti(Me)-H] (3b),
s
2 SA
s
s
SA
III
5-80% [(tSiO)(M O) Ti-H] (3c), and 14% [(tSiO)(M O) Ti ] (3d), along with (SiH ) and
7
s
2
SA
s
2
SA
x
(AlH ) fragments whose formation arise from the opening of adjacent Si-O-M bridges (M=Si, Al).
x
Species 3a-d are efficient catalysts for the hydrogenolysis of waxes with a diesel selectivity higher
than 60%. Comparison with the silica-based system shows a beneficial role of the silica-alumina
support on the activity of the Ti centers, attributed to a direct interaction of the surface with the active
site, which possibly facilitates the β-alkyl transfer, the key C-C bond cleavage step in the proposed
mechanism.
Introduction
example, high diesel selectivities were obtained over Pt/
SiO -Al O (SA) during the hydroprocessing of F-T waxes,
2
2
3
The low-temperature Fischer-Tropsch (F-T) process pro-
duces mainly long-chain n-paraffins with a wide carbon
number range, up to C120, which are virtually free of sulfur
and this was attributed to the mild Brønsted acidity, high
5
surface area, and pore size distribution of the support.
These metal/acid bifunctional catalysts, which are also used
1
6-8
(
<5 ppm) and have low aromatic contents (<1 wt %).
for alkane isomerization,
follow the carbenium ion me-
These long-chain paraffins have few economically interest-
ing applications, except as coatings or lubricants. Conver-
sely, diesel derived from the hydrocracking of F-T waxes
with a cetane number of about 70 can be used as a blending
stock in upgrading lower quality diesel, considering that a
minimum cetane number of 51 is usually required for con-
chanism by (i) dehydrogenation of the alkanes on the metal
sites, (ii) isomerization and cracking of olefins on the acidic
9
sites, and (iii) hydrogenation of new olefins on the metal.
However, the reported hydrocracking catalysts are efficient
only at high temperatures, thus increasing the deactivation
10
by coke formation and poisoning of the acidic sites. There-
fore, it appears interesting to develop new hydrocracking
catalysts running at low temperature and pressure and
2
ventional commercial fuels. Hydrocracking catalysts based
on noble and group 8 metals supported on zeolites or various
other supports have been developed and extensively stu-
leading to an environmentally friendly process (no H S in
2
3
died; amorphous silica-alumina has attracted considerable
interest since, with such a support, higher diesel selectivities
the tail gas).
We and others have reported the preparation and proper-
ties of well-defined silica-supported group 4 metal hydrides
4
were obtained than when using zeolite supports. For
(
5) Calemma, V.; Peratello, S.; Pavoni, S.; Clerici, G.; Perego, C.
*
To whom correspondence should be addressed. Tel: þ33 - 4 72 43 17
4. Fax: þ33 - 4 72 43 17 95. E-mail: demallmann@cpe.fr (A.d.M.);
basset@cpe.fr (J.-M.B.).
Stud. Surf. Sci. Catal. 2001, 136, 307–312(Natural Gas Conversion VI).
(6) Sinfelt, J. H.; Hurwitz, H.; Rohrer, J. C. J. Catal. 1962, 1, 481–483.
(7) Kouwenhoven, H. W.; van Zijll Langhout, W. C. Chem. Eng.
Prog. 1971, 67, 65.
9
(1) Dry, M. E. Appl. Catal., A 1999, 189, 185–190.
(2) De Haan, R.; Joorst, G.; Mokoena, E.; Nicolaides, C. P.
(8) Vaudagna, S. R.; Comelli, R. A.; Canavese, S. A.; Figoli, N. S.
Appl. Catal., A: Gen. 2007, 327, 247–254.
3) Leckel, D.; Liwanga-Ehumbu, M. Energy Fuels 2006, 20, 2330–
336.
4) Leckel, D. Ind. Eng. Chem. Res. 2007, 46, 3505–3512.
J. Catal. 1997, 169, 389–393.
(
(9) Martens, J. A.; Jacobs, P. A.; Weitkamp, J. Appl. Catal., A 1986,
20, 239–281.
(10) St o€ cker, M. Microporous Mesoporous Mater. 2005, 82, 257–292.
2
(
r 2009 American Chemical Society
Published on Web 09/09/2009
pubs.acs.org/Organometallics

5
648 Organometallics, Vol. 28, No. 19, 2009
Larabi et al.
Figure 1. Monitoring the reaction of [Ti(CH
SiO -Al 3-(500) by IR spectroscopy: (a) SiO -Al
b) [(tSiO) Ti(CH tBu)4-x SA supported on SiO -Al
2 4
tBu) ] and
2
2
O
2
2
O
3-(500)
;
(
x
2
]
2
2
O
3-(500)
.
1
Figure 2. C CP/MAS solid-state NMR spectra of 2a,b (40 000
3
Scheme 1. Reaction of 1 with SiO -Al O
2
2
3-(500)
scans, recycle delay 1 s, contact time 1 ms, line broadening
1
applied 50 Hz): (a) 0% C-enriched 2a,b; (b) 30% C-enriched
3
13
2
1
a*,b* (two broad spinning sidebands corresponding to the
13 ppm signal are denoted with stars).
step a β-alkyl transfer, the microscopic reverse of a CdC
2
3
insertion into a metal-carbon bond. Due to their inherent
electrophilicity, titanium complexes are good candidates for
the catalytic hydrocracking of F-T waxes. Herein, we report
the grafting of [Ti(CH tBu) ] (1) and the characterization of
2
4
the supported complexes on silica-alumina (SA), as well as
the preparation and characterization of the hydrides result-
ing from the hydrogenolysis of the grafted species. The
activity of these silica-alumina-supported species in the
hydrogenolysis of F-T waxes under a moderate hydrogen
pressure (1 atm) and mild reaction temperature (typically
180 °C) has been investigated along with the previously
1
1-20
as (tSiO) MH (M = Ti, Zr, Hf)
which are able to
catalyze the hydrogenolysis of alkanes under mild condi-
3
1
8
tions. For example, (tSiO) ZrH, which is very electrophilic,
3
reported silica-supported analogue.
Results and Discussion
Grafting of [Ti(CH tBu) ] (1) on Silica-Alumina. In Situ
can cleave the C-C bonds of propane, butanes, and pentanes
(
but not ethane) under a moderate hydrogen pressure (<1
1,22
2
atm) and at mild temperatures (typically 25-150 °C).
Silica-alumina-supported zirconium hydride has also been
used to cleave the C-C bonds of polyolefins. The mechanism
of such C-C bond cleavage was proposed to involve as a key
2
4
IR Study. Silica-alumina was pressed into a self-supporting
disk, and after calcination under air and treatment under
vacuum at 500 °C (SA or SiO -Al O3-(500)), the major
2
2
surface functional groups were tSiOH and Lewis acid
(
11) Zakharov, V. A.; Dudchenko, V. K.; Paukshtis, E. A.; Karakchiev,
L. G.; Yermakov, Y. I. J. Mol. Catal. 1977, 2, 421–435.
12) Quignard, F.; Choplin, A.; Basset, J. M. J. Chem. Soc., Chem.
Commun. 1991, 1589–1590.
13) Lecuyer, C.; Quignard, F.; Choplin, A.; Olivier, D.; Basset, J. M.
Angew. Chem., Int. Ed. Engl. 1991, 30, 1660–1661.
14) Quignard, F.; Lecuyer, C.; Bougault, C.; Lefebvre, F.; Choplin,
A.; Olivier, D.; Basset, J. M. Inorg. Chem. 1992, 31, 928–930.
15) Quignard, F.; Lecuyer, C.; Choplin, A.; Olivier, D.; Basset, J. M.
J. Mol. Catal. 1992, 74, 353–363.
centers, while no Al -OH surface species were observed
s
2
4
(Figure 1a). The sublimation of [Ti(CH tBu) ] (1) on
(
2 4
SiO -Al O
was then monitored by in situ IR spectros-
copy. The disk turned yellow, and the IR band assigned to
2
2
3-(500)
(
-1
isolated silanol groups (ν(SiO-H) at 3747 cm ) disappeared
(Figure 1b), while a broad band centered at 3700 cm
(
-
1
(
appeared, in agreement with the presence of unreacted
silanols in interaction with the adjacent grafted organome-
(
(
16) Schwartz, J.; Ward, M. D. J. Mol. Catal. 1980, 8, 465–469.
17) Scott, S. L.; Basset, J. M.; Niccolai, G. P.; Santini, C. C.; Candy,
J. P.; Lecuyer, C.; Quignard, F.; Choplin, A. New J. Chem. 1994, 18, 115–
2
5,26
tallic moieties.
Two groups of bands concomitantly
appeared in the 3000-2700 and 1500-1300 cm- regions,
which are respectively assigned to ν(CH) and δ(CH) vibrations
1
1
22.
(
18) Rosier, C.; Niccolai, G. P.; Basset, J. M. J. Am. Chem. Soc. 1997,
119, 12408–12409.
(
19) Dornelas, L.; Reyes, S.; Quignard, F.; Choplin, A.; Basset, J. M.
Chem. Lett. 1993, 1931–1934.
20) Tosin, G.; Santini, C. C.; Baudoin, A.; de Mallmann, A.; Fiddy,
S.; Dablemont, C.; Basset, J. M. Organometallics 2007, 26, 4118–4127.
21) Corker, J.; Lefebvre, F.; Lecuyer, C.; Dufaud, V.; Quignard, F.;
Choplin, A.; Evans, J.; Basset, J. M. Science 1996, 271, 966–969.
22) Lefebvre, F.; Thivolle-Cazat, J.; Dufaud, V.; Niccolai, G. P.;
Basset, J. M. Appl. Catal., A 1999, 182, 1–8.
(23) Dufaud, V. R.; Basset, J. M. Angew. Chem., Int. Ed. 1998, 37,
806–810.
(24) Gora-Marek, K.; Derewinski, M.; Sarv, P.; Datka, J. Catal.
(
Today 2005, 101, 131–138.
(
(25) Nedez, C.; Theolier, A.; Lefebvre, F.; Choplin, A.; Basset, J. M.;
Joly, J. F. J. Am. Chem. Soc. 1993, 115, 722–729.
(26) de Mallmann, A.; Lot, O.; Perrier, N.; Lefebvre, F.; Santini,
C. C.; Basset, J. M. Organometallics 1998, 17, 1031–1043.
(

Article
Organometallics, Vol. 28, No. 19, 2009 5649
1
13
Figure 3. (a) H- C HETCOR spectrum of solid 2a*,b* and traces perpendicular to F2 at (b) 32 ppm and (c) 113 ppm. The spectra
were recorded with 4096 scans, a relaxation delay of 1 s, and a CP contact time of 1 ms. An exponential line broadening of 50 Hz was
applied before Fourier transform.
of neopentyl ligands. The irreversible disappearance of the
free silanol band and the appearance of ν(CH) and δ(CH)
bands as well as the evolution of 2,2-dimethylpropane in the
gas phase is in full agreement with a chemical grafting of 1
and the formation of a tSiOTi bond, as was observed for the
on the carbons directly bound to the titanium center, was
prepared and grafted on SiO -Al O
1
. While the H
MAS spectra of [tSiOTi(*CH tBu) ] (2a*) and [(tSiO)2-
2
2
2
3-(500)
3 SA
Ti(*CH tBu) ] (2b*) are identical with those of 2a,b, the
2 SA
2
1
3
C CP/MAS spectrum shows again the signals at 32,
2
7
grafting of 1 on partly dehydroxylated silica or MCM-41.
Mass Balance Analyses. [Ti(CH tBu) ] (1) in pentane
36, and 113 ppm (Figure 2b), but the intensity of
the 113 ppm signal dramatically increases, which is consis-
tent with its assignment to the carbon directly attached to
2
4
solution was allowed to react at 25 °C for 4 h with
2
as a powder (1.4 OH/nm ; 0.91 mmol/g),
SiO -Al O
the metal: i.e., Ti-*CH tBu). Moreover, the assignments
2
2
3-(500)
2
1
13
followed by filtration, three washing cycles with pentane,
and drying under vacuum at 25 °C to give a yellow solid, 2. A
Ti loading of 2.4-2.9 wt % was obtained according to
elemental analysis, corresponding to 0.50-0.62 mmol Ti/g.
are confirmed by 2D H- C heteronuclear correlation
(HETCOR) spectroscopy, for which short contact times
(1 ms) have been chosen in order to detect only the protons
directly attached to a carbon. Under these conditions, the
2D HETCOR spectrum of 2a*,b* shows two strong correla-
During grafting 1.4-1.8 equiv of tBuCH /Ti evolved, and
3
1
13
the resulting solid contained 11.9-12.4 C/Ti, according to
elemental analyses; these analytical results are consist-
ent with the formation of the two surface species
tions between proton and carbon signals (δ( H)/δ( C)
1.2/32 and 2.4/113) (Figure 3a), but no correlation is ob-
1
3
served for the C signals at 36 ppm, consistent with its
assignment to quaternary carbons. Both correlations at 1.2/
32 and 2.4/113 corroborate their respective attributions to
the methyl (-*CH C(CH ) ; traces in Figure 3b) and methy-
[tSiOTi(CH tBu) ] (2a) and [(tSiO) Ti(CH tBu) ]
2 3 SA 2 2 2 SA
(
2b), in a ratio close to 40:60 (Scheme 1).
1
Solid-State NMR Spectroscopy. The H MAS solid-state
2
3 3
NMR spectrum of solid 2 exhibits only two signals (Figure
S1, Supporting Information), which were assigned by com-
parison with the signals of the monopodal species
lene groups (-*CH tBu; traces in Figure 3c) of the neopentyl
2
ligands. Finally, J-resolved 2D solid-state NMR spectro-
scopy applied to 2a*,b* (Figure 4) shows a quadruplet
1
( JCH=114 Hz) at 32 ppm fully consistent with its assign-
[
Ti(CH tBu) ] , prepared by reaction of [Ti(CH tBu) ] with
(tSiO)Ti(CH tBu) ] and bipodal species [(tSiO)2-
2
3 S
ments to the methyl of the tBu groups (trace in Figure 4b);
the singlet at 36 ppm fits well with the quarternary carbon of
2
2 S
2
4
2
7
SiO2-(700) and MCM-41(200), respectively. The first signal at
.2 ppm corresponds to the methyl resonances of the
CH C(CH ) fragments and the second at 2.4 ppm is
2
8-31
1
the tBu groups (trace in Figure 4c),
1
while the triplet
( J = 107 Hz) at 113 ppm is in agreement with the
-
2
3 3
CH
1
3
attributed to the methylene protons -CH CMe . The
C
CP/MAS solid-state NMR spectrum (Figure 2a) shows an
methylene carbon of the neopentyl fragment (trace in
Figure 4d).
In summary, the reaction of [Ti(CH tBu) ] (1) with
2
3
intense peak at 32 ppm assigned to the methyl groups
-
2
4
CH C(CH ) and two weaker signals at 36 and 113 ppm,
2
SiO -Al O
leads to the formation of the two surface
3 3
2
2
3-(500)
corresponding to the quaternary carbon -CH CMe and the
2
3
methylenic carbons of a neopentyl ligand -CH CMe ,
(28) Terao, T.; Miura, H.; Saika, A. J. Am. Chem. Soc. 1982, 104,
228–5229.
(29) Mayne, C. L.; Pugmire, R. J.; Grant, D. M. J. Magn. Reson.
2
3
1
3
respectively. In order to obtain better quality C NMR
5
1
spectra, the complex [Ti(*CH tBu) ] (1*), 30% C-enriched
3
2
4
1
984, 56, 151–153.
(
30) Miura, H.; Terao, T.; Saika, A. J. Magn. Reson. 1986, 68, 593–
596.
(31) Lesage, A.; Steuernagel, S.; Emsley, L. J. Am. Chem. Soc. 1998,
120, 7095–7100.
(27) Bini, F.; Rosier, C.; Saint-Arroman, R. P.; Neumann, E.;
Dablemont, C.; de Mallmann, A.; Lefebvre, F.; Niccolai, G. P.; Basset,
J. M.; Crocker, M.; Buijink, J. K. Organometallics 2006, 25, 3743–3760.

5
650 Organometallics, Vol. 28, No. 19, 2009
Larabi et al.
1
3
Figure 4. (a) 2D J-resolved solid-state NMR spectrum of the solid 2a*,b*, 30% C-enriched at the R-positions (*) and traces extracted
along the ω1 dimension of the 2D J-resolved spectrum at the carbon chemical shift frequencies (b) 32 ppm, (c) 36 ppm, and (d) 113 ppm.
The spectra were recorded with 1024 scans, a relaxation delay of 2 s, and a CP contact time of 2 ms.
Figure 5. Monitoring the formation of titanium hydrides supported on SiO
2
-Al
2
2
O3-(500) (3a-c) by IR spectroscopy: (a) silica-alumina
dehydroxylated at 500 °C; (b) [(tSiO)Ti(CH tBu) ] and [(tSiO) Ti(CH tBu) ] ; (c) after treatment under H at 150 °C for 4 h;
2
3 SA
2
2 SA
2
-
1
(
d) subtraction (c)-(a), zoom of the 2300-1500 cm region.
species [tSiO-Ti(CH tBu) ] (2a) and [(tSiO) Ti-
resulting from the hydrogenolysis of the neopentyl ligands,
which corresponds to 11.2 ( 1.3 equiv of C/Ti.
ESR and Solid-State NMR Spectroscopy. Samples of solid
2
3 SA
2
(
CH tBu) ] (2b) (Scheme 1) in a ratio close to 40:60, in
2 2 SA
which the titanium atom is grafted to the surface via one or
two covalent (tSiO-Ti) bonds, as evidenced by mass bala-
nce analysis and IR and NMR spectroscopy.
3, prepared by impregnation on silica-alumina powder and
further hydrogenolysis, were analyzed by ESR and H NMR
spectroscopy. While the former technique points out the pre-
1
Preparation of the Titanium Hydride Supported on Sili-
ca-Alumina by Reaction of [(tSiO)Ti(CH tBu) ]
III
sence of ca. 14% of Ti species in solid 3, after a double
and
(tSiO) Ti(CH tBu) ] with H2. In Situ IR Study. When
2
3 SA
[
integration of the characteristic signal at 3439 G (Figure S3,
Supporting Information), the latter gives additional informa-
2
2
2 SA
[
(tSiO)Ti(CH tBu) ] (2a) and [(tSiO) Ti(CH tBu) ] (2b)
2
3 SA
2
2
2 SA
1
tion: the H NMR spectrum of 3 displays several signals at 0.9,
are heated under H for 4 h at 150 °C, the yellow material
2
darkens into a brown solid, 3. The monitoring of the reaction by
infrared spectroscopy (Figure 5) shows the disappearance of ca.
1.3, 1.8 (shoulder), 4.4, and 8.6 ppm (Figure 6a). First, the signals
at 0.9 and 1.3 ppm can be assigned to residual alkyl fragments,
which did not undergo full hydrogenolysis. Second, the signal at
9
0% of the intensity of the ν(C-H) and δ(C-H) bands. In
addition, new bands centered at 1600-1725, 1926, 2195, and
1.8 ppm is consistent with residual silanols, while the signal at 4.4
32,33
-
1
2267 cm appear, which are assigned to ν(TiH), ν(AlH), and
ppm can be readily assigned to (tSi-H) protons.
Finally,
18,23
ν(SiH), respectively.
firmed through the H/D exchange reaction at room temperature
These assignments were further con-
the quite broad downfield signal at 8.6 ppm can be attributed to a
surface titanium hydride, in agreement with a previous assign-
ment of a resonance at 8.5 ppm to the hydride ligand of the
under D (550 Torr): the intensity of vibration bands at
2
1
(
(
600-1725 cm- decreased by ca. 90%, as expected for Ti-H
1
molecular complex {(ArO) Ti-H PMe }. The deconvolu-
34
3
3
3
35
Figure S2, Supporting Information), while the ν(Si-H) bands
2195 and 2267 cm ) and ν(Al-H) band at 1926 cm were
tion of this broad signal between 7 and 10 ppm shows two
-
1
-1
not affected by this treatment, as expected for silicon and
aluminum hydrides. During the hydrogenolysis process,
(32) Campbell-Ferguson, H. J.; Ebsworth, E. A. V.; MacDiarmid,
A. G.; Yoshioka, T. J. Phys. Chem. 1967, 71, 723.
(33) Soignier, S.; Taoufik, M.; Le Roux, E.; Saggio, G.; Dablemont,
C.; Baudouin, A.; Lefebvre, F.; de Mallmann, A.; Thivolle-Cazat, J.;
Basset, J. M.; Sunley, G.; Maunders, B. M. Organometallics 2006, 25,
1569–1577.
-
1
a band of weak intensity also reappeared at 3747 cm corre-
sponding to isolated silanols, which were probably in interaction
-
1
)
with the alkyl ligands (vide supra, ν(OH) band at 3700 cm
and became free when these ligands underwent hydrogenolysis.
(
34) Noth, H.; Schmidt, M. Organometallics 1995, 14, 4601–4610.
(
35) Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calve, S.;
Simultaneously, methane (9 equiv of CH /Ti) and ethane (1.1
4
Alonso, B.; Durand, J. O.; Bujoli, B.; Gan, Z. H.; Hoatson, G. Magn.
Reson. Chem. 2002, 40, 70–76.
equiv of C H /Ti) were observed in the gas phase as the products
2
6

Article
Organometallics, Vol. 28, No. 19, 2009 5651
found previously in the case of zirconium surface organometallic
chemistry, where [(tSiO) ZrH ] species were similarly charac-
2
2
37
terized. The proton resonance at 8.6 ppm does not show such
an autocorrelation and can then be attributed to monohydride
species, which can be described by the general formula
[
alkyl group, including CH ). The two signals revealed upon the
(tSiO)(M O) (R) Ti-H] (x=0, 1, 2; M =Si. Al; R=
2-x
s
x
SA
s
3
deconvolution of the total broad titanium hydride resonance
could then be attributed to a few dihydride species,
[
(tSiO)(M O)TiH ] (3a), around 9 ppm and to monohydride
s 2 SA
species, [(tSiO)(M O) (R) Ti-H]SA, at 8.6 ppm. In addition,
s
2-x
a clear correlation is observed between the resonance at 8.6 ppm
x
and a part of a signal at ca. 1.6 ppm (cross-peaks at ca. 10 ppm in
the ω dimension in Figure 6). This may correspond to a
1
correlation between the proton of the monohydride species
(tSiO)(M O) R Ti-H] (x=0, 1, 2) and either residual
2-x
[
s
x
SA
OH groups located in the vicinity of the tTi-H group but too
far to react with it or residual alkyl groups bound to titanium (x=
1, 2). The precise nature and amount of the different hydride and
Ti species supported on solid 3 could be finally obtained from
the double integration of the ESR signal, from the integration of
III
1
the H MAS NMR signals associated with di- and monohy-
3
8
drides, and from the butanolysis of sample 3.
In conclusion, according to the previously described methods,
the treatment under H2 at 150 °C of the mixture of
1
Figure 6. (a) H MAS solid-state NMR spectrum of [Ti-H]
species obtained as indicated in Scheme 2. (b) DQ rotor-
1
[(tSiO)Ti(CH tBu) ]
2
(2a) and [(tSiO) Ti(CH tBu) ]
3 SA 2 2 2 SA
synchronized 2D H MAS spectrum of [Ti-H] species (see text
and Experimental Section for more details).
(
(
2b) leads to new surface titanium species: [(tSiO)-
M O)TiH ] (3a) [(tSiO)(M O)(Me)Ti-H] (3b), [(tSiO)-
s
2 SA
s
SA
III
different peaks centered at 8.6 and 9.1 ppm in a ca. 92:8 ratio
(Figure S4. Supporting Information). In contrast to IR spec-
troscopy, the H MAS NMR spectrum of the surface species
(M O) Ti-H] (3c),and[(tSiO)(MsO) Ti ] (3d).However,
complexes [tSiOTi(CH tBu) ]
s
2
SA
2
SA
(2a) and [(tSiO) Ti-
2
3 SA
2
1
(CH tBu) ] (2b) were expected to lead to trishydride and
2 2 SA
does not show a resonance indicating the presence of Al-H
dihydride (3a) species, respectively, by hydrogenolysis of the
titanium-carbon bonds. However, since silica-alumina is
amorphous, the various titanium surface species, initially 2a,
b, are surrounded by different local environments: that is,
different types of Si-O-Si and Si-O-Al bridges formed
during the dehydroxylation step. For that reason and as
36
species (expected at ca. 4-5 ppm ), probably due to the broad-
ness of the signal induced by the quadrupolar moment of
trigonal Al centers. To further assign these various signals, 2D
multiple-quantum (MQ) proton spectroscopy under magic angle
spinning (MAS) was applied. To probe structural information
1
15,21,33,39
inherent to proton-proton dipolar couplings, the H DQ MAS
already observed with other metal and oxide supports,
the hydrogenolysis of [tSiOTi(CH tBu) ] (2a) generates
spectrum was recorded for [Ti-H] at a MAS frequency of
ca. 10 kHz (Figure 6b). In this experiment, the evolution of the
DQ coherence created between a pair of dipolar-coupled protons
2
3 SA
first [(tSiO)(M O)TiH ]
s
(3a) and [(tSiO)(M O)(Me)-
s
2 SA
Ti-H]SA (3b; M = Si, Al), together with [Si-H] and
s
is observed during the indirect detection time t . The DQ
1
[Al-H] surface species via the opening of neighboring
Si-O-Si and Si-O-Al bridges and the transfer of an hydride
from the metal to a Si or Al surface atom (Scheme 2). Then, in
most cases, other Si-O-Si or Si-O-Al bridges are close
enough to the titanium centers to allow the surface complex
[(tSiO)(M O)TiH ] (3a) toreact further and yield the major
frequency in the ω dimension corresponds to the sum of the
1
two single quantum frequencies of the two coupled protons and
correlates in the ω dimension with the two corresponding
2
proton resonances. Therefore, the observation of a DQ cross-
peak implies a close proximity between the two protons involved
in this correlation. Autocorrelation peaks are observed on the
diagonal of the 2D spectrum (Figure 6b), in particular for the
s
2 SA
monohydride species [(tSiO)(M O) Ti-H]
(3c) as
s
2
SA
75-80% of the surface titanium species observed by ESR
and H NMR, leaving only ca. 3% of the dihydride species
1
alkyl protons at 0.9 ppm (ca. 2 ppm in the ω dimension, for
1
CH , with z=2, 3), for the remaining OH groups at 1.8 ppm (ca.
[(tSiO)(M O)TiH ] (3a). The other monohydride complex
s
z
2 SA
3
.7 ppm in the ω dimension, probably for geminal >Si(OH)2
[(tSiO)(M O)(Me)Ti-H] (3b) would account for less than
1
s
SA
silanols), and for the signal attributed to (tSi-H) protons
resonances at 4.4 ppm (8.8 ppm in the ω dimension), showing
the presence of silicon dihydride [(M O) SiH ]. A weak auto-
correlation peak is also observed for a proton resonance at ca.
1
(38) The butanolysis consisted of the treatment of sample 3 with tert-
butyl alcohol (40 mbar) for 1 h at 30 °C, followed by evacuation at 80 °C
for 2 h. During this treatment, titanium hydrides disappeared while
s
2
2
t
surface tTiO Bu species appeared, with the concomitant evolution of
8.9 ppm (17.8 ppm in the ω dimension), which unambiguously
dihydrogen (H /Ti=0.85 ( 0.15) and a small amount of methane (CH /
1
2
4
allows the assignment of this resonance to a titanium dihydride,
[
Ti=0.04). This means that less than 5% of the surface titanium was
present as tTiMe species in sample 3 and that methyl was the only alkyl
ligand bound to Ti. The main monohydride Ti species are then
(tSiO)(M O)TiH ] (3a; M =Si, Al; see Scheme 2), as was
s 2 SA s
[
(tSiO)(MsO)(Me)Ti-H] (3b) and [(tSiO)(MsO) Ti-H] (3c), repre-
2
senting respectively less than 5% and 75-80% of the surface Ti species.
After butanolysis, the ESR signal characteristic of Ti species decreased
III
(
(
36) Uhl, W.; Jana, B. Chem. Eur. J. 2008, 14, 3067–3071.
37) Rataboul, F.; Baudouin, A.; Thieuleux, C.; Veyre, L.; Coperet,
from 14% to 3% of the total Ti content.
(39) Vidal, V.; Theolier, A.; Thivolle-Cazat, J.; Basset, J. M.; Corker,
J. J. Am. Chem. Soc. 1996, 118, 4595–4602.
C.; Thivolle-Cazat, J.; Basset, J. M.; Lesage, A.; Emsley, L. J. Am. Chem.
Soc. 2004, 126, 12541–12550.

5
652 Organometallics, Vol. 28, No. 19, 2009
Larabi et al.
Figure 7. Hydrogenolysis of a paraffin wax catalyzed by [Ti-H]/SiO
2
-Al
2
O
3
(1 bar of H
2
, 180 °C): (a) initial hydrocarbon weight
distribution of the wax; (b) hydrocarbon weight distribution obtained after the hydrogenolysis reaction (the decay observed between C8
and C13 is likely due to an analytical artifact: only 92% of the products could be detected by GC with our procedure); (c) weight
distribution of the categorized products obtained after hydrogenolysis.
Scheme 2. Main Structures Proposed for Titanium Hydrides Supported on Silica-Alumina
5
[
% of the surface titanium. In addition to this main process,
[Ti-H]SiO2 and {[Ti-H]SiO2 þ SiO -Al O3-(500)} systems can
2
2
III
(tSiO)(MsO) Ti ] (3d) surface species are also formed,
be explained by the reactivity of the Brønsted acid sites present in
2
SA
accounting for ca. 14% of the observed surface titanium.
Activity in the Hydrogenolysis of FT Waxes. The titanium
hydrides on silica-alumina 3 and titanium hydrides on silica
the pure silica-alumina(500), while they have been eliminated for
[Ti-H]SiO2 and [Ti-H]
by the grafting reaction of
SiO -Al O
3
2
2
18
Ti(CH tBu) on the OH groups of the surface. Indeed, under
2 4
dehydroxylated at 500 °C were tested in the reaction of wax
hydrogenolysis, in order to study the effect of the support on the
activity and selectivity; a low-molecular-weight wax (C20-C50,
Figure 7a) obtained from Aldrich was used in a semicontinuous
flow reactor at 180 °C under a hydrogen stream (1 atm, 20 mL/
min; see the Experimental Section and Figure S5 in the Support-
ing Information). As can be seen in Table 1, after 24 h the
conversion of the wax obtained with titanium hydrides sup-
ported on silica-alumina, 3, is complete, while under similar
conditions the conversion is only 30% with titanium hydrides
supported on silica, [Ti-H]SiO2. Interestingly, solid 3 leads to a
higher diesel selectivity (63%) compared to titanium hydride on
silica (40%), whereas this latter produces more C1-C2 hydro-
carbons (24%) and both catalysts show similar selectivities in
C3-C4 or C5-C9 products (Figure 7b,c). In order to determine
the beneficial role of the silica-alumina support, another experi-
ment was carried out by adding 200 mg of pure SiO -Al O
the same conditions, a blank experiment with silica-alumina(500)
alone gave a very low wax conversion of ca. 3% and produced
only C1-C4 gaseous products. Since the activity observed for
the {[Ti-H]SiO2 þ SiO -Al O3-(500)} system is still much lower
2
2
than that obtained with [Ti-H]
, it thus seems that a
SiO -Al O
3
2
2
direct interaction of the surface with the active site is responsible
for the change in reactivity observed for [Ti-H] com-
SiO -Al O
3
2
2
pared to [Ti-H]SiO2, rather than the intervention of a bifunc-
tional mechanism. Similar support effects have been evidenced in
the field of polymerization, where it is known that alumina-
supported organometallic complexes are usually more reactive
40-45
than their corresponding silica equivalents,
and this has
(40) Marks, T. J. Acc. Chem. Res. 1992, 25, 57–65.
41) Tullock, C. W.; Tebbe, F. N.; Mulhaupt, R.; Ovenall, D. W.;
Setterquist, R. A.; Ittel, S. D. J. Polym. Sci., Part A 1989, 27, 3063–3081.
42) Ballard, D. G. H. J. Polym. Sci., Part A 1975, 13, 2191-2212;
Coord. Polym. 1975, 223-262.
43) Tullock, C. W.; Mulhaupt, R.; Ittel, S. D. Makromol. Chem.
(
(
2
2 3-(500)
(
to [Ti-H]SiO2, under the same experimental conditions. The
results show that the activity is the same as that for [Ti-H]SiO2:
the wax conversion was 34% instead of 30%, but with the
formation of a more important proportion of gaseous C1-C4
products, ca. 56% instead of 33%, and of a lower diesel fraction
Rapid Commun. 1989, 10, 19–23.
(44) Collette, J. W.; Tullock, C. W.; Macdonald, R. N.; Buck, W. H.;
Su, A. C. L.; Harrell, J. R.; Mulhaupt, R.; Anderson, B. C.
Macromolecules 1989, 22, 3851–3858.
(
45) Jezequel, M.;Dufaud,V.;Ruiz-Garcia, M. J.;Carrillo-Hermosilla,
F.; Neugebauer, U.; Niccolai, G. P.; Lefebvre, F.; Bayard, F.; Corker, J.;
Fiddy, S.; Evans, J.; Broyer, J. P.; Malinge, J.; Basset, J. M. J. Am. Chem.
Soc. 2001, 123, 3520–3540.
(
selectivities: C1-C2, 40%; C3-C4, 16%; C5-C9, 29%;
C10-C22, 17%). The difference in selectivity between the

Article
Organometallics, Vol. 28, No. 19, 2009 5653
a
Table 1. Hydrolysis of a Paraffin Wax : Comparison between
Titanium Hydride (75 μmol of Ti) Supported on Silica and
Silica-Alumina
at 500 °C and gives the monosiloxy species [(tSiO)-
Ti(CH tBu) ] (2a) and the bisiloxyl species [(tSiO) Ti-
(CH tBu) ] (2b) in a 40:60 ratio. These surface complexes
2
3 SA
2
2
2 SA
were characterized by the combined use of several techniques
Catalyst
1
such as IR spectroscopy, H MAS, C-CP/MAS, 2D
13
1
13
[Ti-H]silica(500)
[Ti-H]silica-alumina(500)
H- C HETCOR, and J-resolved solid-state NMR as well
as mass balance analysis. By treatment under hydrogen at
150 °C the neopentyl ligands in complexes 2a,b undergo
hydrogenolysis and a mixture of supported titanium species
is obtained. IR spectroscopy, ESR, H MAS, and DQ solid-
state NMR spectroscopy show the presence of ca. 3%
Conversion
wt %)
C1-C2
selectivity (wt %)
C3-C4
selectivity (wt %)
C5-C9 (naphtha)
selectivity (wt %)
C10-C22 (diesel)
selectivity (wt %)
3
0
4
9
8
9
100
(
2
6
7
1
[
[
(tSiO)(M O)TiH ]
s
(3a; M = Si, Al), less than 5%
(tSiO)(M O)(Me)Ti-H] (3b), 75-80% [(tSiO)(M O)
2 SA s
2
3
24
63
s
SA
s
2
III
Ti-H]SA (3c), and 14% [(tSiO)(M O) Ti ] (3d), along
with (SiH ) and (AlH ) fragments via the opening of adja-
s
2
SA
x
x
a
The initial hydrocarbon weight distribution is indicated in Figure 7.
Conditions: 400 mg of wax; 180 °C, PH2=1 bar, flow rate 20 mL/min.
cent Si-O-M bridges (M =Si, Al). Species 3a-d efficiently
s
s
catalyze the hydrogenolysis of a paraffin wax, affording
a diesel selectivity higher than 60%. Comparison with
the silica-supported system has evidenced a beneficial role
of the silica-alumina support on the activity of the
Ti centers, attributed to a direct interaction of the surface
with the active site, which possibly facilitates the β-alkyl
transfer, the key C-C bond cleavage step in the proposed
mechanism.
been often associated with the presence of cationic species
favored on “acidic” alumina and not on “neutral” silica.
The mechanism proposed for such hydrogenolysis reac-
tions involving group 4 metals is based on a key C-C bond
cleavage elementary step, already known in organometallic
chemistry: namely the β-alkyl transfer, which is the micro-
scopic reverse of a CdC double-bond insertion into a
metal-carbon bond and would be the rate-limiting step.
2
3
4
6
Experimental Section
Because of this step, the overall reaction should be thermo-
dynamically unfavorable, except that the hydrogenation of
the olefinic double bonds provides the driving force which
makes the process run. In a first step, a C-H bond in the
paraffinic chain is activated by the metal hydride, the metal
General Procedure. All experiments were carried out by using
standard Schlenk and glovebox techniques for the organome-
tallic synthesis. Solvents were purified and dried according to
standard procedures. C D (SDS) was distilled over Na/benzo-
6 6
0
˚
phenone and stored over 3 A molecular sieves. TiCl
(99%), 1,2-
dibromoethane (99%, Aldrich), 2,2 -biquinoline (98%,
being in a d configuration, through a σ-bond metathesis
4
0
with the formation of a metal-alkyl complex and liberation
of dihydrogen. The second step consists of a β-alkyl transfer,
leading to the formation of an olefin and a smaller alkyl
group σ-coordinated to the metal. Then follows the olefin
insertion into a M-H bond and hydrogenolysis of the
metal-alkyl bonds to give smaller alkanes with the regen-
eration of the metal-hydride bonds. In the case of supported
zirconium hydrides, an influence of the support has been
theoretically evidenced for the β-alkyl transfer: while for a
zirconium center anchored onto silica, the olefin produced
after this step would be directly emitted in the gas phase,
1
3
13
Lancaster), CO
1.7 M in pentane, Aldrich), LiAlH4 (95%, Aldrich, stored
under argon), MgSO (Laurylab), NaHCO (Prolabo), and
2
(99% C, Cambridge Isotopes), tBuMgCl
(
4
3
Vielsmeier reagent (95%, Aldrich, stored under argon) were
used as received. tBuCH Li was prepared from tBuCH Cl
2
2
(
98%, Lancaster, used as received) and Li wire (Aldrich).
[Ti(CH tBu) ] was prepared according to the literature proce-
2
4
4
8
dure. Gas-phase analyses were performed on a Hewlett-Pack-
ard 5890 Series II gas chromatograph equipped with a flame
2 3
ionization detector and an Al O /KCl on fused silica column (50
mꢀ0.32 mm).
4
6
Elemental analyses were performed at the CNRS Central
Analysis Department of Solaize (metal analyses) or in the
Laboratory of Prof. Dormond, LSEO of Dijon (C, H analyses).
IR spectra were recorded on a Nicolet 5700 FT-IR spectrometer,
without strong interaction with the surface, for zirconium
sites supported on alumina, the olefin would be more
strongly π-coordinated to the zirconium center, probably
due to its cationic character on this support, which stabilizes
using an IR cell equipped with CaF windows, allowing in situ
2
4
7
the π-olefin complex. This would also be the case for some
titanium centers supported onto silica-alumina which
would be more electrophilic than those supported onto silica,
thus facilitating the further formation of the Ti-olefin π-
complex in the key β-alkyl transfer step of the proposed
mechanism and may explain the higher activity observed for
studies. Typically, 16 scans were accumulated for each spectrum
-
1
(resolution 2 cm ). Solution NMR spectra were recorded on an
Avance-300 Bruker spectrometer. All chemical shifts were
1
13
measured relative to residual H or C resonances in C D : δ
6 6
1 13
.15 ppm for H and 128 ppm for C.
H MAS and C CP-MAS Solid-State NMR Spectra. These
spectra were recorded on a Bruker Avance-500 spectrometer.
7
1
13
[
Ti-H]
compared to [Ti-H]SiO2.
SiO -Al O
3
1
13
2
2
For specific studies (see below), H MAS and C CP-MAS
solid-state NMR spectra were recorded on a Br u€ ker Avance-500
spectrometer with a double-resonance 4 mm CP-MAS probe.
The samples were introduced under Ar within a glovebox into a
zirconia rotor, which was then tightly closed. In all experiments,
the rotation frequency was set to ca. 10 kHz. Chemical shifts
Conclusion
In conclusion, Ti(CH tBu) (1) reacts selectively with the
2
4
surface silanols of a silica-alumina partially dehydroxylated
1
were given with respect to TMS as external reference for H and
C NMR.
1
3
(
46) Mortensen, J. J.; Parrinello, M. J. Phys. Chem. B 2000, 104,
901–2907.
47) Joubert, J.; Delbecq, F.; Thieuleux, C.; Taoufik, M.; Blanc, F.;
Coperet, C.; Thivolle-Cazat, J.; Basset, J. M.; Sautet, P. Organometallics
007, 26, 3329–3335.
2
(
(48) Cheon, J.; Dubois, L. H.; Girolami, G. S. J. Am. Chem. Soc.
1997, 119, 6814–6820.
2

5
654 Organometallics, Vol. 28, No. 19, 2009
Larabi et al.
Heteronuclear Correlation Spectroscopy. The two-dimen-
sional heteronuclear correlation experiments were performed
(5 mL) and SiO
2
-Al
2
O
3-(500) (1 g) was stirred at 25 °C for
4 h. After filtration, the solid was washed five times with pentane
and all volatile compounds were condensed into another reactor
(of known volume) in order to quantify neopentane evolved
during grafting. The resulting yellow-brown powder was dried
according to the following scheme: 90° proton pulse, t evolu-
1
tion period, cross-polarization (CP) to carbon spins, detection
4
9
of carbon magnetization under TPPM decoupling. The con-
tact time for CP was set to 1 ms. A total of 64 t increments with
024 scans each was collected, and the recycle delay was 1 s (total
acquisition time of 18 h).
J-Resolved Spectroscopy. The two-dimensional J-resolved
-5
1
under vacuum (10 Torr) to yield 1.137 g of 2a,b. Analysis
by gas chromatrography indicated the formation of neopentane
during the grafting (1.6(0.2 NpH/Ti). Anal. Foundfor 2 (wt %):
1
1
Ti, 2.4-2.9; C, 7.5-8.7. H MAS solid-state NMR (300 MHz): δ
2
8,31
13
experiment was performed as previously described:
cross-polarization from protons, carbon magnetization evolves
during t under proton homonuclear decoupling. Simultaneous
to
after
2.4, 1.2 ppm. C CP/MAS solid state NMR: δ 113, 36,
32 ppm.
1
Preparation of 2*, by Impregnation of 1* onto SiO
2
-Al
2
-
1
3
1
80° carbon and proton pulses are applied in the middle of t
1
O3-(500). The C-enriched surface compounds 2a*,b* were
prepared using the procedure described above for the prepara-
tion of 2, using 1* in place of 1 with a reaction time of 4 h at 25
refocus the carbon chemical shift evolution while retaining the
modulation by the heteronuclear JCH scalar couplings. A
Z-filter is finally applied to allow phase-sensitive detection in
1
°C. Anal. Found for 1* (wt %): Ti, 2.9; C, 8.7. H MAS solid-
state NMR (300 MHz): δ 2.4, 1.2 ppm. C CP/MAS solid state
NMR: δ 113, 36, 32 ppm.
1
3
ω
1
. The proton rf field strength was set to 83 kHz during t
1
50,51
49
) and acquisition (TPPM decoupling ).
(
FSLG decoupling
The lengths of carbon and proton 180° pulses were 7 and 6 μs,
respectively, and the Z-filter delay was 200 μs. An experimental
scaling factor, measured as already described, of 0.52 was found;
Monitoring the Synthesis of [Ti(CH tBu) ]/SiO -Al O
2
4
2
2
3-(500)
by IR Spectroscopy. The oxide (25 mg) was pressed into an
17 mm self-supporting disk, adjusted in the sample holder, and
this gave a corrected spectral width of 2452 Hz in the ω
introduced into a glass reactor equipped with CaF windows.
2
The supports were calcined under air and partly dehydroxylated
under vacuum at the desired temperature. Complex 1 was then
sublimed under vacuum at 50 °C onto the oxide disk, which
typically turned yellow. The solid was then heated at 50 °C for
1
52
dimension. The recycle delay was 2 s, and a total of 80 t
increments with 1024 scans each was collected (total acquisition
time 30 h).
1
Two-Dimensional Double-Quantum (DQ) Experiments. These
were performed according to the following scheme: excita-
2
5
h, and the excess of 1 was removed by reverse sublimation at
, which
tion of DQ coherences, t
observable magnetization, Z-filter, and detection. Fourteen
1
evolution, reconversion into
0-60 °C and condensed into a tube cooled by liquid N
2
was then sealed off using a blow torch. An IR spectrum was
recorded at each step.
5
3
pC7 basic elements were used for both excitation and
reconversion steps, with 70 kHz rf field strength (=7ω ); the
R
Preparation of 3 by Treatment under H2 of the Supported
Perhydrocarbyl Complexes 2a,b. General Procedure. In a 375
Z-filter delay was 200 μs. A total of 512 increments of 64 scans each
was collected with 1.5 s recycle delay, which gave a total experiment
time of 14 h.
mL reactor, 1 g of sample 2 ([(tSiO)Ti(CH
[(tSiO) Ti(CH tBu) ] ) was treated with 70 kPa of H . The
2 3
tBu) ]SA and
2
2
2 SA
2
ESR Characterizations. For these measurements, a sample
was loaded in a 4 mm quartz tube in a glovebox and then sealed
under vacuum. ESR spectra were recorded at room temperature
and at 77 K on a Varian E9 spectrometer at ENS-Lyon (Ecole
Normale Sup ꢀe rieure, Lyon) and compared to a vanadyl sulfate
standard recorded under the same conditions. After a double
temperature was increased to 150 °C (1 °C/min) and was
maintained at this temperature for 4 h. The solid turned from
yellow to brown. The gaseous products were then quantified
by GC: methane (9 equiv of CH /Ti) and ethane (1.1 equiv of
4
1
C
2
H
6
/Ti). H MAS solid-state NMR (500 MHz, 25 °C):
δ (ppm) 0.8 (s, CH), 1.8 (s, SiOH), 4.4 (s, SiH and SiH
2
),
III
integration of both signals, the amount of Ti species could be
estimated.
8
signal with Lorentzian functions: see Figure S4 in the Sup-
.6 (s, TiH, 3b,c), 9.1 (s, TiH , 3a) (deconvolution of the
2
Preparation of Tetrakis(2,2-dimethylpropyl)titanium(IV) (1).
The molecular precursor [Ti(CH tBu) ] (1) was prepared from
-1
porting Information); Infrared (cm ): 1600-1730 (broad,
2
4
ν(Ti-H) and ν(TiH )), 1910 (ν(Al-H), 2200, 2260 (m, ν(SiH)
2
[
(
Ti(OEt)
yield 62%).
Preparation of C-Labeled Tetrakis(2,2-dimethylpropyl)tita-
4 2
] and tBuCH Li, following the literature procedure
and ν(SiH )), 2700-2900 (w, ν(C-H)).
48
2
Hydrogenolysis of Waxes. General Procedure. Mechanical
mixtures of titanium hydrides (75 μmol of Ti) supported
on silica-alumina or on silica and wax (400 mg; Aldrich,
ASTM D87; mp 70 °C), were charged using a glovebox into
a stainless steel cylinder reactor which could be isolated
from the atmosphere (see Figure S5 in the Supporting
Information). Two other experiments were carried out simi-
larly, one with a mixture of titanium hydride supported
13
nium(IV) (1*). The preparation of 1* was performed according
to the same procedure as that used for 1, but starting with a 70:30
1
3
2
mixture of nonlabeled tBuCH Cl and 99%- C-monolabeled
1
3
tBu*CH Cl to form 30%- C-labeled tBu*CH Li as alkylating
2
2
agent (yield 60%).
Preparation of SiO
2
-Al
2
O
3-(500). Silica-alumina HA-S-HPV
from Akzo-Nobel, with 25% of aluminum and a specific area of
on silica, 200 mg of pure SiO
wax and the other being a blank experiment, with a mixture
of 200 mg of SiO -Al O and 400 mg of wax. After
2
-Al
2
O3-(500), and 400 mg of
2
4
78 m /g, was calcined at 500 °C under dry air flow for 4 h and
-5
partly dehydroxylated at 500 °C under high vacuum (10 Torr)
for 24 h to give a white solid having a specific surface area of
2
2
3-(500)
connection to the gas lines and purge of the tubes, a flow
of hydrogen (20 mL/min), controlled by a mass flow con-
troller (Brooks) under 1 bar of pressure, was sent upward
into the catalyst bed, which was heated at 180 °C. Hydro-
carbon products were stripped from the liquid medium
by the hydrogen flow. Light hydrocarbons and hydrogen
were analyzed on line by GC (HP 6890 chromatograph
equipped with an Al O /KCl 50 m ꢀ 0.32 mm capillary
2
90 m /g and containing 1.4 OH/nm (0.91 mmol/g).
2
3
Preparation of 2 by impregnation of 1 onto SiO -Al O3-(500).
2
2
2 4
A mixture of 1 [Ti(CH tBu) ] (350 mg, 1.05 mmol) in pentane
(
49) Bennett, A. E.; Rienstra, C. M.; Auger, M.; Lakshmi, K. V.;
Griffin, R. G. J. Chem. Phys. 1995, 103, 6951–6958.
50) Bielecki, A.; Kolbert, A. C.; Levitt, M. H. Chem. Phys. Lett.
989, 155, 341–346.
51) Levitt, M. H.; Kolbert, A. C.; Bielecki, A.; Ruben, D. J. Solid
State Nucl. Magn. Reson. 1993, 2, 151–163.
52) Lesage, A.; Duma, L.; Sakellariou, D.; Emsley, L. J. Am. Chem.
Soc. 2001, 123, 5747–5752.
53) Hohwy, M.; Jakobsen, H. J.; Eden, M.; Levitt, M. H.; Nielsen,
N. C. J. Chem. Phys. 1998, 108, 2686–2694.
(
2
3
1
column and an FID detector for hydrocarbons or with a
(
3
m molecular sieves column and a catharometer for hydro-
gen). Liquid products were condensed at 0 °C and analy-
zed off-line by GC (HP 5890 chromatograph equipped with
da ne t eHc tPo 5r ) .3 0 m ꢀ 0.32 mm capillary column and an FID
(
(

Article
Acknowledgment. We thank Dr Fr ꢀe d ꢀe ric Lef ꢁe bvre and
Sudhakar Chakka for fruitful discussion and the BP
Company for financial support.
Organometallics, Vol. 28, No. 19, 2009 5655
ported titanium hydride, an ESR spectrum of the solid 3 Ti
species supported on silica-alumina500, deconvolution with
Lorentzian functions of the H MAS NMR signal of the
1
titanium hydrides 3, supported on silica-alumina500, and a
scheme of the reactor used for the wax hydrogenolysis. This
material is available free of charge via the Internet at http://
pubs.acs.org.
1
Supporting Information Available: Figures giving an
H
MAS NMR spectrum of 2a,b on silica-alumina500, IR spectra
on the H/D exchange reaction for the SiO -Al
2
2
O3-(500) sup-