By-product formation causes leaching of Ti from the redox molecular sieve TS-1

Article abstract of DOI:10.1039/b002055i

Triols, formed as by-products from the titanium silicalite (TS-1) catalysed epoxidation of allylic alcohols, can cause leaching of Ti from the microporous framework.

Full text of DOI:10.1039/b002055i

View Article Online / Journal Homepage / Table of Contents for this issue
By-product formation causes leaching of Ti from the redox molecular sieve
TS-1
Lucinda Davies,a Paul McMorn,a Donald Bethell,b Philip C. Bulman Page,c Frank King,d Frederick E.
Hancockd and Graham J. Hutchings*a
a Department of Chemistry, Cardiff University, PO Box 912, Cardiff, UK CF10 3TB. E-mail: hutch@cf.ac.uk
b Department of Chemistry,University of Liverpool, Liverpool, UK L69 3BX
c
Department of Chemistry, University of Loughborough, Loughborough, UK LE11 3TU
d Synetix, PO Box 1, Billingham, UK TS23 1LB
Received (in Cambridge, UK) 14th May 2000, Accepted 14th August 2000
Triols, formed as by-products from the titanium silicalite
(TS-1) catalysed epoxidation of allylic alcohols, can cause
leaching of Ti from the microporous framework.
The TS-1 was also shown to be an effective catalyst for the
epoxidation of allylic alcohols [Fig. 1(a) and (b)]. In a typical
reaction, crotyl alcohol (0.1 mol) was reacted with aqueous
The titanium(IV) silicalite known as TS-1 has been considered
by many to be a unique redox molecular sieve since its
discovery by Enichem researchers.1,2 TS-1 has been shown to
be an active heterogeneous catalyst for the epoxidation of
alkenes3 and allylic alcohols.4 Two features are responsible for
the uniqueness of TS-1. First, incorporation of Ti into the
microporous silicate framework increases the hydrophobicity5
and enables H2O2 to be used as an oxidant. Other redox
molecular sieves are ineffective with H2O2 and require more
active oxygen sources, e.g. tert-butyl hydroperoxide.6 Sec-
ondly, TS-1 is considered very stable under the reaction
conditions and to date, loss of Ti from the framework of TS-1
has not been studied in detail. These features have been
emphasised recently by Sheldon et al.,7 who also stress that
most papers pay scant attention to the loss of framework
transition metal ions from redox molecular sieves. Recently, it
has been shown that other redox molecular sieves containing Cr
and V leach these metal cations into solution during reaction,
and therefore act as reservoirs for active homogeneous
catalysts.8,9 If leaching is common for transition metals
incorporated into microporous frameworks, then it is surprising
that this should not be the case for Ti in TS-1. We have now
studied the stability of TS-1 with respect to loss of Ti during
reaction and we demonstrate that by-product formation un-
fortunately can lead to the loss of Ti from titanium(IV
)
silicalite.
TS-1 (2.4 wt% Ti) was prepared according to the following
method, which has been used in many previous studies.
Titanium butoxide (3.4 g Aldrich) was added dropwise to
tetraethyl orthosilicate (62.5 g Merck) and the resulting mixture
stirred for 30 min. Tetrapropylammonium hydroxide (40 wt%
in H2O, 76.3 g Alfa) was added dropwise with stirring. The
resulting mixture was then heated at 60 °C for 2 h to evaporate
ethanol produced by the hydrolysis/condensation reactions
between the silicon and titanium reagents. Water was then
added to the synthesis gel to return the volume of the gel to that
prior to the evaporation step. The gel was mixed thoroughly and
autoclaved under autogeneous pressure at 175 °C for 48 h. The
resulting solids were filtered off, washed and dried at 100 °C for
4 h. The materials were calcined at 550 °C in air for 16 h prior
to use. The material was characterised by X-ray diffraction,
scanning electron microscopy, FTIR and reflectance laser
Raman spectroscopy. The data were found to be in agreement
with literature data and the crystallites of TS-1 were found to be
uniform, rhombohedral and 0.2 mm in diameter. The sample of
TS-1 was used as a catalyst for the hydroxylation of phenol with
H2O2 at 70 °C. After 75 min, the products formed were 49%
catechol and 50% hydroquinone; this is in agreement with
previous literature studies10 confirming the quality of the
TS-1.
Fig. 1 Reaction of crotyl alcohol with H2O2 at 50 °C with TS-1 in methanol.
(a) Conversion: (5) crotyl alcohol, (-) H2O2; (b) reaction products: (5)
epoxide, (:) ether diols, (-) triol; (c) Ti leaching.
DOI: 10.1039/b002055i
Chem. Commun., 2000, 1807–1808
This journal is © The Royal Society of Chemistry 2000
1807

View Article Online
H2O2 (30% vol, 0.11 mol) at 50 °C with TS-1 (0.1 g) in
methanol (10 ml) in a stirred flask fitted with a reflux condenser.
Samples were withdrawn at various time intervals and the
products analysed by GC using dimethyl sulfone as an internal
standard. Samples were also filtered hot through a heated sinter
funnel which contained a bed of Celite to remove small particles
of zeolite. Blank experiments confirmed that this method was
effective in removing any particulate Ti-containing material
and, furthermore, that any titanium observed in solution did not
originate from the Celite. The solutions were then analysed by
ICPMS to determine the Ti content [Fig. 1(c)]. Initially, TS-1 is
active and selective for the formation of the oxirane product
and, at this stage, no Ti leaching is observed. However, in a
sequential reaction the oxirane ring opens by nucleophilic attack
by the solvent to form the triol (reaction with water) or ether
diols (reaction with methanol). At this time, Ti is observed to
leach from TS-1 and the pH of the reaction mixture decreases.
These results clearly demonstrate that the structural integrity of
TS-1 is not maintained in the presence of the secondary reaction
products formed under typical reaction conditions.
In a further set of experiments, TS-1 (crystallisation period 2
d) was slurried with a range of reagents at 50 °C for 35 h. No Ti
leaching was observed with hydrogen peroxide in methanol
(30% by vol, < 3 ppm Ti), crotyl alcohol ( < 0.3 ppm Ti),
butane-2,3-diol ( < 2 ppm Ti) or hydrochloric acid (1 M, < 0.7
ppm Ti). Several further experiments were conducted with
glycerol as a model triol. No Ti leaching was observed with
glycerol and hydrochloric acid ( < 2 ppm Ti), or glycerol
together with hydrochloric acid and methanol ( < 1.3 ppm Ti).
However, when TS-1 was slurried with glycerol and hydrogen
peroxide using methanol as solvent, extensive leaching of Ti
was observed (3571 ppm Ti) corresponding to removal of 16%
of the Ti from TS-1. These model experiments confirm that the
necessary condition for Ti-leaching from TS-1 is the combined
presence of a triol together with hydrogen peroxide, and hence
for the successful operation of TS-1 as a truly heterogeneous
catalyst the formation of triols must be avoided.
Fig. 2 Interaction of triols with the active site of TS-1.
so, a further experiment was carried out in which the non-
calcined TS-1 was silanised according to the method of Beck et
al.12 using trimethylsilyl chloride and hexamethyldisiloxane.
As the non-calcined TS-1 retains the template within the intra-
crystalline pores, this procedure effectively silanises just the
exterior surface. On subsequent calcination (550 °C, 4 h), the
template is removed and TS-1 crystallites covered on the
external surface with silica are formed. Using this material as a
catalyst for crotyl alcohol oxidation did not result in any
significant leaching of Ti, yet the catalytic performance was
essentially the same as that observed for the non-silanised
sample [Fig. 1(a), (b)]. These data indicate that it is Ti located
on the external surface of the TS-1 crystallites that is
preferentially leached by the by-product triol.
We would like to thank Synetix, EPSRC, and the DTI/Link
programme on asymmetric catalysis for financial support and
the Department of Earth Sciences, Cardiff University for
ICPMS results.
Notes and references
1 M. Taramasso, G. Perego and B. Notari, US Pat., 4 410 510, 1983.
2 M. Taramasso, G. Manara, V. Fattore and B. Notari, US Pat., 4 666 692,
1986.
3 M. G. Clerici and P. Ingallina, J. Catal., 1993, 140, 71.
4 G. J. Hutchings, D. F. Lee and A. R. Minihan, Catal. Lett., 1995, 33,
369.
It is interesting to comment on the mechanism by which Ti is
removed from the framework of TS-1. Detailed UV–VIS
spectroscopy11 has shown that TS-1 contains four-coordinate
Ti4+ in the absence of water. When TS-1 is suspended in water
or water–H2O2 mixtures, the Ti4+ species becomes six co-
ordinate, retaining three Ti–O–Si bonds, which are considered
to anchor the Ti firmly within the microporous framework. The
solvent molecules, in our case methanol, could be expected to
displace one or more of the water molecules. We suggest that,
when this site is exposed to triols, further Ti–O–Si bonds are
broken and this leads to the irreversible loss of Ti from the
structure (Fig. 2). Since triol by-product formation can be
commonly observed in TS-1 catalysed reactions, we suggest
that Ti leaching from TS-1 is probably more common than has
previously been considered. It is also interesting to determine if
the Ti that is leached during the catalytic reaction originates
from the surface of the TS-1 crystallites. To determine if this is
5 J. Weitkamp, S. Ernest, E. Roland and G. F. Thiele, Stud. Surf. Sci.
Catal., 1997, 105, 763.
6 M. A. Camblor, M. Constantini, A. Corma, L. Gilbert, P. Esteve, A.
Martinez and S. Valencia, Chem. Commun., 1996, 1339.
7 R. A. Sheldon, m. Wallau, I. W. C. E. Ahrens and U. Schuchardt, Acc.
Chem. Res., 1998, 31, 485.
8 H. E. B. Lempers and R. A. Sheldon, Appl. Catal., 1996, 143, 137.
9 W. I. Caravalho, P. B. Varaldo, M. Wallau and U. Schuchardt, Zeolites,
1997, 18, 408.
10 J. A. Martens, P. Buskeni, P. A. Jacobs, A. van der Pol, J. H. C. van
Hoof, C. Ferrini, H. W. Kouwenhoven, P. J. Kooyman and H. van
Bekkum, Appl. Catal. A, 1993, 99, 71.
11 A. Zecchina, S. Bordiga, C. Lamberti, G. Ricchiardi, D. Scarano, G.
Petrini, G. Leofanti and M. Mantegazza, Catal. Today, 1996, 32, 97.
12 J. S. Beck, J. C. Vartulli, W. J. Roth, M. E. Leonwicz, C. T. Kresge,
K. D. Schmitt, C. T. W. Chu, D. H. Olson, E. W. Shepheard, S. B.
McCullen, J. G. Higgins and J. L. Schenter, J. Am. Chem. Soc., 1992,
114, 10 834.
1808
Chem. Commun., 2000, 1807–1808