Effect of structure of the redox molecular sieve TS-1 on the oxidation of phenol, crotyl alcohol and norbornylene

Article abstract of DOI:10.1039/b503241e

A range of crystalline TS-1 samples with different morphologies as well as the corresponding TS-1 precursor structures have been synthesised using hydrothermal crystallisation. The materials have been characterised using powder X-ray diffraction, IR and Raman spectroscopy and electron microscopy. The materials were used as catalysts for the oxidation of crotyl alcohol, phenol and norbornylene and, in particular, the reactivity of the precursor structures was contrasted with crystalline TS-1. The oxidation of crotyl alcohol, selected as a relatively non-reactive substituted alkene, did not require the TS-1 structure for reactivity and TS-1 precursor structures are active, although crystalline TS-1 was found to be more reactive than the precursor structures. In contrast, phenol hydroxylation is only catalysed by crystalline TS-1. The reaction of phenol is observed to occur only on the exterior surface of large TS-1 crystallites. With smaller crystallites of TS-1, i.e. the size range of interest for catalysis, the rapid subsequent reaction of hydroquinone makes it difficult to determine whether reaction occurs solely on the exterior of the crystallites or at sites within the porous structure. Hence it is suggested that this reaction has limited scope as a probe reaction for the reactivity of sites within the crystallites. It is, however, feasible that phenol hydroxylation is a viable probe reaction for TS-1 type structural units. Norbornylene was studied as an example of a reactant too large to enter the internal pore structure of TS-1 and hence only reaction at pore mouths and external surface sites was possible. Larger TS-1 crystallites were more active for this substrate than suggested by surface area considerations. The results are discussed in terms of the selection of model reactions for the study of TS-1 catalysts. The Owner Societies 2005.

Full text of DOI:10.1039/b503241e

View Article Online / Journal Homepage / Table of Contents for this issue
R E S E A R C H P A P E R
Effect of structure of the redox molecular sieve TS-1 on the
oxidation of phenol, crotyl alcohol and norbornylenew
a
a
b
c
d
Owain J. Kerton, Paul McMorn, Donald Bethell, Frank King, Frederick Hancock,
e
e
f
a
Andrew Burrows, Christopher J. Kiely, Simon Ellwood and Graham Hutchings*
a
School of Chemistry, Cardiff University, Cardiff, UK CF10 3TE. E-mail: hutch@cf.ac.uk
Department of Chemistry, University of Liverpool, Liverpool, UK L69 3BX
Johnson Matthey Catalysts, P.O. Box 1, Billingham, Teeside, UK TS23 1LB
Johnson Matthey PCT, 28 Cambridge Science Park, Milton Road, Cambridge, UK CB4 0FP
Department of Materials Science and Engineering, Lehigh University, 5 East Packer Avenue,
Bethlehem, PA 18015-3195, USA
b
c
d
e
f
Quest International, Ashford, Kent, UK TN24 0LT
Received 10th March 2005, Accepted 24th May 2005
First published as an Advance Article on the web 9th June 2005
A range of crystalline TS-1 samples with different morphologies as well as the corresponding TS-1 precursor
structures have been synthesised using hydrothermal crystallisation. The materials have been characterised using
powder X-ray diffraction, IR and Raman spectroscopy and electron microscopy. The materials were used as
catalysts for the oxidation of crotyl alcohol, phenol and norbornylene and, in particular, the reactivity of the
precursor structures was contrasted with crystalline TS-1. The oxidation of crotyl alcohol, selected as a relatively
non-reactive substituted alkene, did not require the TS-1 structure for reactivity and TS-1 precursor structures are
active, although crystalline TS-1 was found to be more reactive than the precursor structures. In contrast, phenol
hydroxylation is only catalysed by crystalline TS-1. The reaction of phenol is observed to occur only on the
exterior surface of large TS-1 crystallites. With smaller crystallites of TS-1, i.e. the size range of interest for
catalysis, the rapid subsequent reaction of hydroquinone makes it difficult to determine whether reaction occurs
solely on the exterior of the crystallites or at sites within the porous structure. Hence it is suggested that this
reaction has limited scope as a probe reaction for the reactivity of sites within the crystallites. It is, however,
feasible that phenol hydroxylation is a viable probe reaction for TS-1 type structural units. Norbornylene was
studied as an example of a reactant too large to enter the internal pore structure of TS-1 and hence only reaction
at pore mouths and external surface sites was possible. Larger TS-1 crystallites were more active for this substrate
than suggested by surface area considerations. The results are discussed in terms of the selection of model
reactions for the study of TS-1 catalysts.
researchers in 1983, was an important milestone in the field
of oxidation research. Since then, TS-1 has been shown to be
an effective catalyst for alkene and alcohol oxidation
Following the discovery of TS-1 a number of other metal-
Introduction
Selective oxidation remains an essential pathway for the func-
tionalisation of petroleum based feedstocks and considerable
attention continues to be given to the design of heterogeneous
catalysts capable of operating under mild reaction conditions.
Until recently, these oxidation reactions were carried out with
gas-phase reactants with two-phase (gas/solid) heterogeneous
catalysts, and elevated temperatures were required to activate
8–10
11,12
containing molecular sieves and mesoporous materials have
13,14
been used as oxidation catalysts
the identification of new Ti-containing porous oxidation cata-
and there is still interest in
15–18
lysts.
An essential feature of TS-1, identified in this earlier litera-
8
ture, is that it is effective for the epoxidation of alkenes with
1
the reactants. More recently, attention has been focussed on
identifying catalysts that can operate with liquid-phase reac-
tants at ambient or low temperatures. To achieve reaction
under these milder conditions it is necessary that activated
forms of dioxygen are used, e.g. hydrogen peroxide or organic
electron depleted carbon–carbon double bonds, e.g. allylic
alcohols, but in general lower reactivities are observed than
8
with non-substituted alkenes. However, the epoxides are
19
20
susceptible to hydrolysis and rearrangement. The oxidation
of these substrates is of immense interest to the fine chemicals
industry. TS-1 exhibits a second important feature, namely
that it is significantly more robust to loss of Ti through
leaching than other Ti-containing structures developed to date.
Previously, we have studied the stability of TS-1 in detail and
have shown that at short reaction times (r24 h) there is
2
–4
hydroperoxides with transition-metal molecular sieves, and
5
6,7
sodium hypochlorite or iodosylbenzene with immobilised
salen-modified transition-metal cations. In this respect the
discovery of the redox molecular sieve TS-1, by Enichem
w Electronic supplementary information (ESI) available: Fig. S1: IR
spectra of TS-1 materials (template : Si = 1.0 mol ratio). Fig. S2:
Scanning electron micrographs of silicalite (template : Si = 1.0 mol
ratio). Fig. S3: Powder X-ray diffraction patterns of silicate (template :
Si = 1.0 mol ratio). Fig. S4: IR spectra of silicate (template : Si = 1.0
mol ratio). Fig. S5: Scanning electron micrographs of TS-1
with template : Si mol ratio = 0.35–0.63. See http://www.rsc.org/supp
data/cp/b5/b503241e/
minimal loss of Ti when oxidising allylic alcohols with hydro-
21–23
gen peroxide.
used as a model reaction for TS-1 is the hydroxylation of
Another oxidation that has been extensively
2
4–32
phenol, and in particular this reaction has been used to
evaluate the role of Ti in framework positions inside the pores
or Ti active sites on the surface of the crystallites.
3
1
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 2 6 7 1 – 2 6 7 8
2671

View Article Online
Given the importance of TS-1, it is not surprising that
considerable interest has been given to its synthesis as well as
Ti-free silicalite-1. TS-1 is prepared by a slow hydrothermal
crystallisation from an aqueous gel containing Ti and Si
sources and tetrapropylammonium (TPA) cation as the struc-
Oxidation reactions
Crotyl alcohol and norbornylene epoxidation
Catalysts were studied for the epoxidation of crotyl alcohol
using the following standard procedure. Crotyl alcohol (1.0 g)
and aqueous hydrogen peroxide (1.87 g, 30 vol%) were stirred
together in methanol (10 ml) in a round-bottomed flask placed
in an oil-bath at 60 1C. The catalyst (0.1 g) was added and
small aliquots were removed, cooled and analysed using gas
chromatography using pentan-1-ol as an internal standard.
Similar experiments were carried out using norbornylene.
3
3
ture directing template. Crystallisation is known to occur via
3
4–36
globular structures
posites,
formed from inorganic–organic com-
and these globular structures can be formed at
shorter crystallisation times with TPAOH than with
3
3,37
3
4,38
TPABr. A detailed study of the factors controlling crystal-
3
lisation was carried out by van Hooff and van der Pol and it
was shown that the concentration of OH plays a crucial role
9
ꢀ
in the crystallisation process and the TPAOH concentration
controlled the crystallite size of TS-1 and consequently its
reactivity.
Phenol hydroxylation
Catalysts were studied for the hydroxylation of phenol using
the following standard procedure. Phenol (1.0 g) and aqueous
hydrogen peroxide (1.2 g, 30 wt%) were stirred together in
methanol (10 ml) in a round-bottomed flask placed in an oil
bath at 80 1C. The catalyst (0.1 g) was added and as the
reaction proceeded small aliquots were removed, cooled and
analysed using gas chromatography using pentan-1-ol as an
internal standard.
The details of the crystallisation process involved in the
synthesis of TS-1 are now well understood, and there is
agreement that crystalline TS-1 with relatively small crystallite
3
9
sizes are the most effective as oxidation catalysts. However,
there has been less research interest given to the study of the
reactivity of the precursor structures to TS-1 as compared with
the immense number of studies concerning crystalline TS-1.
We have now addressed this aspect and in this study we
investigate the reactivity of TS-1 precursors and contrast them
with crystalline TS-1 prepared with representative morpholo-
gies. The use of these materials for the oxidation of crotyl
alcohol, phenol and norbornylene is reported and we comment
on the appropriateness of these reactions as models for com-
parative studies.
Material characterisation
Materials were characterised using a range of techniques.
Powder X-ray diffraction was carried out using a Siemens
D500 diffractometer which was equipped with a copper target
and a primary monochromator. Data was collected in a flat
plate mode. IR spectra were obtained using a Perkin Elmer
System 2000 FTIR spectrometer fitted with a Mir TGS detec-
tor. Samples were diluted with KBr (1% sample) and pressed
into discs. Surface areas were determined using nitrogen
adsorption according to the BET method using a Micromeri-
tics ASAP 2000 instrument. Scanning electron microscopy was
carried out using a Cambridge S-360 electron microscope.
Samples for transmission electron microscopy (TEM) analysis
were prepared by dispersing the catalyst powder onto a lacey
carbon film supported on a copper mesh grid. TEM observa-
tions were made in a JEOL 2000 EX high-resolution electron
microscope operating at 200 kV. Raman spectra were obtained
using a Renishaw Ramascope spectrometer.
Experimental
Synthesis of TS-1
TS-1 (2.5 mol% Ti) was prepared in a similar manner to that
2
proposed by Taramasso et al. Titanium(IV) butoxide (0.425 g,
Aldrich) was added slowly to tetraethyl orthosilicate (10.3 g,
Aldrich) and the mixture was stirred for 2 h to give a clear
solution. To the solution, the template tetrapropylammonium
hydroxide (25.4 g, Alfa, 40% in water, K o0.9 ppm, Na o0.05
ppm, template: tetraethyl orthosilicate mol ratio ¼ 1.0) and
deionised water (15.2 g) were added slowly with stirring at
2
5 1C. The clear gel was stirred at 25 1C for 1 h and then stirred
at 60 1C for 5 h to aid hydrolysis and, during this procedure,
the volume was maintained by addition of deionised water. The
resulting solution was heated in a Teflon-lined stainless-steel
autoclave at 175 1C for a range of times. The material was then
recovered by centrifugation, washed with deionised water,
dried at 100 1C and then calcined in air at 550 1C for 24 h.
Additional preparations were carried out using lower tetra-
propylammonium hydroxide concentrations (tetrapropylam-
monium hydroxide : tetraethyl orthosilicate mol ratio ¼
Results
Synthesis of TS-1 precursors and crystallites
This part of the study involved the synthesis of a set of TS-1
precursors and crystalline TS-1 with a range of crystallite sizes
and morphologies that are representative of TS-1 materials
used in previous catalysis studies. The materials will used in the
subsequent part of this study for the oxidation of crotyl
alcohol, phenol and norbornylene. Two series of TS-1 samples
were synthesised, one series focussed on the effect of synthesis
time and the other focussed on the effect of template : Si mol
ratio in the crystallisation gel.
0
.35–1.0) using appropriate amounts of tetraproylammonium
hydroxide.
Synthesis of silicalite-1
Titanium free silicalite-1 was synthesised as a blank catalyst
containing no Ti, using the same method as TS-1 but without
the addition of the titanium(IV) butoxide.
Synthesis of TS-1 (mol ratio template : Si = 1.0): effect of
crystallisation time. A series of materials were prepared at
1
75 1C with crystallisation times ranging from 2 h to 12 days
and these were characterised using powder X-ray diffraction
Fig. 1), SEM (Fig. 2) and IR spectroscopy (see ESIw). Unfortu-
(
Reaction of TS-1 with chlorotrimethylsilane
nately, insufficient material was synthesised for crystallisation
times of less than 14 h to enable powder X-ray diffraction to be
carried out. The preparation method we selected did not
involve stirring, and in this static preparation method initially
(2–16 h) no crystalline TS-1 is formed, rather honeycombed,
amorphous particles are prepared. Crystalline TS-1 is initially
detected at ca. 18 h (cf. the material synthesised at 16 h which is
amorphous) as shown by powder X-ray diffraction (Fig. 1).
This procedure was carried out according to the method of
4
Beck et al. Non-calcined TS-1 (0.5 g) was refluxed with
0
chlorotrimethylsilane (10 g, Aldrich) and hexamethyldisiloxane
(
15 g, Aldrich) for 16 h. The treated TS-1 was recovered
following evaporation of the volatile reagents, washed with
acetone and dried (25 1C). Prior to use as a catalyst, the
material was calcined (550 1C, 24 h) to remove the template.
2
672
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 2 6 7 1 – 2 6 7 8
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5

View Article Online
determined by XRD, is only observed at 22 h crystallisation
time. The globular phase (22 h sample) has an XRD pattern that
is similar to that for the final TS-1 material, with the exception
that a much broader reflection at 2y ¼ 261 is observed. Subse-
quently, at longer crystallisation times the globular phase is
transformed to the orthorhombic crystallites that are character-
2
istic of TS-1. Once this second phase transition is complete,
there is no further noticeable change in the morphology or
crystallinity. All these observations are fully consistent with the
observations made in previous definitive studies that have
3
3–39
focussed primarily on the synthesis of TS-1.
The IR spectra (see ESIw) are also consistent with the
powder X-ray diffraction and SEM characterisations. Two
ꢀ
1
absorption bands at 450 and 560 cm are characteristic of
the MFI pentasil zeolite framework and this is only observed
Fig. 1 Powder X-ray diffraction patterns of TS-1 materials (templa-
te : Si ¼ 1.0 mol ratio): (a) TS-1 18-h; (b) TS-1 20-h; (c) TS-1 22-h; (d)
TS-1 1-day; (e) TS-1 2-day; (f) TS-1 6-day treated with chlorotri-
methylsilane; (g) TS-1 12 d treated with chlorotrimethylsilane.
ꢀ1
following crystallisation for 16–18 h. The IR band at 960 cm
could be characteristic of Ti–O stretching in the TS-1 structure
and, again, this is observed following crystallisation for 16–18 h.
ꢀ
1
However, the band at 960 cm is also characteristic of silanol
sites (i.e. Si–OH) and may indicate a defect site caused by the
absence of a tetrahedral framework atom, or formed by the
SEM (Fig. 2) shows that distinct, facetted crystallites are
formed at 18 h crystallisation, although some irregular, possi-
bly non-crystalline material, is also present. The crystallite size
range is quite broad (1–20 mm, average size 10 mm) but this
does not change markedly with crystallisation time, although
the irregular non-crystalline material is absent following crys-
tallisation for 2 days.
Taking the XRD and SEM analyses together, it is possible to
observe the sequence of processes occurring during the crystal-
lisation process as a function of time in a non-stirred reaction
mixture. The large honeycomb-type particles observed up to 16 h
crystallisation are transformed to a much smaller, globular
phase via a break-down of the honeycomb structure. The total
time for this transformation is relatively short, ca. 4 h, and the
globular phase, which has the TS-1 MFI pentasil structure, as
4
1–44
addition of a different tetrahedral framework atom.
ꢀ
1
Hence, the band at 960 cm may be a consequence of Ti
incorporation. This was confirmed from a study of the crystal-
lisation of Ti-free silicalite using the same reaction conditions
as the TS-1 series, i.e. Si ¼ template ratio and crystallisation
conditions, but without the addition of the Ti-source. Char-
acterisation using SEM, XRD and IR spectroscopy (see ESIw)
showed that in the absence of Ti, crystalline material only
starts to form after 32 h crystallisation and all the materials
appear similar over the crystallisation range from 30 h to 12
days with an average crystallite size of ca. 10 mm. In addition
the IR spectra confirm the presence of the MFI pentasil
ꢀ1
framework, but the band at 960 cm is absent.
Fig. 2 Scanning electron micrographs of TS-1 materials (template : Si ¼ 1.0 mol ratio) for a range of crystallisation times: (A) TS-1 14-h; (B) TS-1
8-h; (C) TS-1 20-h; (D) TS-1 2-day; (E) TS-1 12-day; (F) TS-1 2-day treated with chlorotrimethylsilane; (G) TS-1 12-day treated with
chlorotrimethylsilane (all space bars ¼ 20 mm).
1
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 2 6 7 1 – 2 6 7 8
2673

View Article Online
TEOS mol ratios with the 2-day crystallisation time. FT-IR
spectra of TS-1 samples (TPAOH : TEOS mol ratio 0.35–0.63)
are also consistent with those in the literature as a band at 960
ꢀ1
cm is observed that is indicative of the incorporation of Ti
ꢀ
1
two bands are present at 500–600 cm that arise as a result of
the formation of the MFI structure of TS-1. No significant
differences are observed between the spectra of the various
samples.
It is clear from the microscopic studies (see ESIw) that a
definite change in crystallite size and morphology is observed
with an increase in TPAOH content within the gel. At a
template : Si mol ratio ¼ 0.35 scanning electron microscopy
reveals particles with an average particle size of approximately
Fig. 3 Raman spectra of TS-1 (template : Si ¼ 1.0 mol ratio).
1
50 nm. No marked change in the size of the particles is
observed on increasing the template : Si mol ratio to 0.40,
however, it is observed that the particles are cubic. At a
template : Si mol ratio 0.48 TS-1 crystallites increase in size
to an average diameter of approximately 250 nm. However, it
is apparent that some ‘coffin’ like morphologies are present
with an approximate dimensions of 600 ꢁ 250 nm. On increas-
ing the template : Si mol ratio to 0.5 a major transformation
occurs within the sample. It is apparent that there is a ‘switch’
in morphology from small cubic structures (B250 nm in
diameter) to the larger ‘coffin’-like structures. When the tem-
plate : Si mol ratio is further increased to 0.54 crystallites have
an average size of between 0.5–3 mm, irregular crystallite
morphologies are present although some are cubic and ‘coffin’
like structures. This is also the case with template : Si mol
ratios of 0.56 and 0.58, although particle sizes increase to
approximately 4 mm. At a template : Si mol ratio ¼ 0.63 all
crystallites exhibit a ‘coffin’ like structure, as is the case of
template : Si mol ratio ¼ 1.0, with an average diameter of
approximately 10 mm.
The laser Raman spectra of 2- and 12-day TS-1 are shown in
Fig. 3. It is apparent that the two samples exhibit significant
differences in the spectra. The 12-day TS-1 sample gives a
Raman spectrum that is almost identical with that for silicalite.
This indicates the 12-day sample may be deficient in silanol
defects. The 2-day TS-1 sample shows bands with Raman shifts
4
4
ꢀ1
which are characteristic of Ti–O (900 cm TiO stretching in
ꢀ1
TiO octahedra in titanate, 446 and 611 cm for rutile and
6
ꢀ1
3
95, 517, 638 cm for anatase).
The TS-1 sample prepared following crystallisation for 2 and
2 days were reacted with chlorotrimethylsilane and hexa-
1
methyldisiloxane and the materials were subsequently calcined.
No effect on the crystallite size was observed using this
procedure (Fig. 1). The TS-1 materials prepared by 2- and
1
were also characterised using N adsorption and the results are
2-day crystallisation and treated using a range of conditions
2
given in Table 1. It is apparent that, as expected, the high pore
volume is only developed following calcination to remove the
template. In addition, silanisation of the TS-1 does not affect
the surface area or pore volume when the template is still
present. This suggests that the reaction with chlorotrimethylsi-
lane and hexamethyldisiloxane and the subsequent calcination
only results in the formation of a silica layer being deposited on
the exterior surface of the TS-1 crystallites.
Transmission electron microscopy studies. In a complimen-
tary study to the SEM investigations on the series of TS-1
structures prepared in this study, TS-1 samples prepared
following crystallisation for 2 and 12 days with template : Si
mol ratios of 0.35 and 1.0 were selected for detailed character-
isation using transmission electron microscopy. The micro-
graphs of typical crystals prepared using the template : Si ratio
of 1.0 are shown in Fig. 4. The dominant morphology in the
Synthesis of TS-1 (2-day crystallisation): effect of template : Si
mol ratio. A second set of TS-1 samples were prepared using a
2
-day crystallisation time but with varying template : Si mol
ratio (0.35–1.0). The materials were characterised by X-ray
diffraction, IR spectroscopy and electron microscopy (see
ESIw). XRD patterns of TS-1 samples (TPAOH : TEOS mol
2
ratio 0.35–0.63) are consistent with those of the literature with
double reflections observed at 2y ¼ 24.1 and 29.31. No
significant differences are observed between the diffraction
patterns of the samples prepared using different TPAOH : -
Table 1
N
2
adsorption data for TS-1 (template : Si mol ratio ¼ 1.0)
Sample
Conditions
Surface
2
area/m g
Pore volume/
ꢀ1
ꢀ1
ml g
TS-1 2 day
Uncalcined
Uncalcined
2
0.5
0.0073
b
nd
a
Calcined at 550 1C
Calcined at 550 1C
528
532
0.217
0.215
a
TS-1 12 day
Uncalcined
29
15
0.0763
0.0393
0.153
a
Uncalcined
Calcined at 550 1C
356
c
b
nd
Silicalite
Calcined at 550 1C
382
a
Reacted with chlorotrimethylsilane and calcined according to the
c
nd: not determined. Crystallised for 12
40 b
method of Beck et al.
days.
Fig. 4 TEM micrographs of TS-1 (template : Si ¼ 1.0 mol ratio): (a)
TS-1 2-day; (b) TS-1 12-day; (c) TS-1 2-day, treated with chlorotri-
methylsilane.
2
674
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 2 6 7 1 – 2 6 7 8
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5

View Article Online
case of the 2-day TS-1 sample were micron scale angular
lozenge-shaped crystals, as shown in Fig. 4(a). These were
often decorated with much smaller anatase crystallites ranging
from 30–50 nm in diameter. The 12-day TS-1 sample, on the
other hand, contained three distinct morphologies. The domi-
nant morphology, shown in Fig. 4(b), comprised of ‘clean’
micron scale angular TS-1 crystals which were similar in size
distribution to those seen in the 2-day TS-1 sample. In addi-
tion, there were micron sized agglomerates of 5–10 nm anatase
nanoparticles (Fig. 4(c)) and distinct ‘star-shaped’ agglomer-
ates of rutile needles (Fig. 3(d)). On reaction with chlorotri-
methylsilane followed by calcination, the morphology of the
TS-1 crystallites remained unchanged but the decoration with
nanoparticles of TiO was absent.
2
Transmission electron micrographs for TS-1 prepared using
template : Si mol ratio of 0.35 with crystallisation for 2 and 12
days are shown in Fig. 5. The 12-day TS-1 sample shows well
crystalline orthorhombic TS-1 particles with an average dia-
meter of approximately 100–150 nm (Fig. 5(a)), however, the
2-day TS-1 material shows more irregular shaped particles are
formed (Fig. 5(b)).
Fig. 5 TEM micrographs of TS-1 (template : Si ¼ 0.35 mol ratio): (a)
TS-1 2-day; (b) TS-1 12-day.
Oxidation reactions
Three substrates were selected for detailed study using the
range of TS-1 structures prepared. Crotyl alcohol was selected
as an example of a substrate with an oxidisable carbon–carbon
double bond. Phenol was selected as it is often used as a probe
molecule to determine the relative reactivity of active sites on
the exterior surface of TS-1 and sites within the internal porous
structure. Norbornylene was selected as an example of a
substrate that is too large to enter the pore structure of TS-1
and consequently can only react on the exterior surface.
Crotyl alcohol epoxidation was investigated and the results
are given in Table 2 for reaction for 24 h, together with a
summary of the TS-1 crystallite size and morphology. The
epoxidation of crotyl alcohol can produce four products, three
1
9
of which are formed from competing ring-opening reactions
Scheme 1
a
Table 2 Effect of TS-1 synthesis conditions on the oxidation of crotyl alcohol
Product selectivity (%)
b
Template : Si mol ratio
Crystallisation time/h
Crystal shape and size /mm
Conv. (%)
Epoxide
Triol
Ether diols
No catalyst
Silicalite-1
—
288
48
288
48
48
48
48
48
48
48
48
48
14
18
20
22
24
48
48
144
288
—
0
0
—
—
68
74
61
64
63
58
51
56
58
47
52
48
47
62
79
48
49
51
46
50
—
—
7
—
—
25
22
20
25
24
33
37
33
34
39
38
42
44
30
17
41
40
38
37
40
m, 10 (mt)
s, 0.15
c, 0.15
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
a
.35
.35
.40
.44
.48
.50
.52
.54
.56
.58
.63
.0
94
95
92
94
88
89
62
64
56
48
52
45
44
48
60
57
49
44
52
60
4
c, 0.15
c, 0.15
9
9
c, 0.25; m, 0.6 ꢁ 0.25
c, 0.25; m, 3–5 ꢁ 0.25
c, 0.3; m 2
m, 2
13
9
12
11
8
m, 4
m, 4
14
10
10
9
m, 10
a
a; m, 5
.0
.0
g; 10
g; 10
8
.0
4
.0
m, 10
11
11
9
.0
m, 10 (mt)
m, 10 (mt)
m, 10 (mt)
m, 10 (mt)
c
.0
.0
.0
17
10
b
Reaction conditions: Crotyl alcohol (1.0 g), H
2 2
O (1 eq.), TS-1 (0.1 g), methanol (10 ml), 60 1C, 24 h. Average particle size determined from
electron microscopy, s spherical/irregular, c cubic, m coffin shaped monoclinic, a amorphous, g globular, mt multiply twinned, minor component in
40
c
italics. Reacted with chlorotrimethylsilane and calcined according to the method of Beck et al.
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 2 6 7 1 – 2 6 7 8
2675

View Article Online
a
Table 3 Effect of TS-1 synthesis conditions on the hydroxylation of phenol
b
Template : Si mol ratio
Crystallisation time/h
Crystal shape and size /mm
Conv. (%)
Catechol : hydroquinone (mol ratio)
No catalyst
Silicalite-1
—
—
m, 10 (mt)
a
0
0
—
—
288
14
1
1
1
1
1
1
1
1
1
0
0
0
a
.0
.0
.0
.0
.0
.0
.0
.0
.0
0
—
—
18
20
a; m, 5
g, 10
g, 10
0
0
—
—
22
24
0
m, 10
38
46
4.1
0
1 : 3.2
1 : 3.2
1 : 3.0
—
48
m, 10 (mt)
m, 10 (mt)
m, 10 (mt)
m, 10 (mt)
s, 0.15
c, 0.15
s, 0.15
144
288
48
c
0
—
.35
.35
.35
48
288
21.0
18.0
6.0
1 : 2.0
1 : 2.1
1 : 0
c
48
b
Reaction condition: Phenol (1.0 g), H
2 2
O (1 eq.), catalyst (0.1 g), methanol (10 ml) at 80 1C for 24 h. Average particle size determined from
c
electron microscopy, s spherical/irregular, c cubic, m coffin shaped monoclinic, a amorphous, g globular, mt mutiply twinned. Samples reacted
with chlorotrimethylsilane and calcined according to the method of Beck et al.
40
Scheme 3
The oxidation of larger substrates was investigated to study
the reactivity of molecules that are too large to diffuse into the
micropores of TS-1. Initial experiments were carried out with
cyclooctene, however, it was found that a non-catalysed homo-
geneous oxidation masked the heterogeneously catalysed reac-
tion. This is an important consideration in selecting substrates
for catalytic reactions with porous materials. In view of this,
the oxidation of norbornylene was investigated (Scheme 3) and
the results are given in Table 4. Both small crystallite (0.15 mm)
and large crystallite (10 mm) TS-1 were active for this oxida-
tion, and the large crystallite material was only ca. 3–4 times
less active than the smaller crystallite material, which is con-
siderably less than that expected for the difference in crystallite
size.
Scheme 2
(
Scheme 1). It is apparent that all the TS-1 samples tested show
activity for this epoxidation reaction and the epoxide is the
major reaction product.
Phenol hydroxylation was investigated and the results are
given in Table 3, together with a summary of the structure and
morphology of the TS-1 samples. The hydroxylation of phenol
produces two products, hydroquinone and catechol, together
with polymerisation products that remain on the catalysts as
coke (Scheme 2).
Discussion
Comments on the reactivity of crotyl alcohol and phenol
2
Silicalite, TiO (anatase), 2- and 12-day non-calcined TS-1
were all found to be inactive for phenol hydroxylation under
these reaction conditions. No reaction was observed when a
homogeneous catalyst, titanium acetylacetonate, was used.
This indicates that, should Ti-leaching occur during the oxida-
tion reactions, there would be no homogeneously catalysed
component to the overall reaction. However, no significant Ti-
leaching was observed from TS-1 samples during the phenol
hydroxylation studies for the short reaction times utilised in
this study.
With respect to the reactivity of crotyl alcohol there are a
number of features that can be commented on. First, as
expected, the catalytic activity is related to the crystallite size,
with the smallest crystallites giving the highest activities and
selectivity to the epoxide. The well crystalline small crystallites
prepared after 12-day crystallisation with a template : Si mol
ratio of 0.35 give the highest activity and selectivity. The
activity is similar to that of the 2-day sample but the selectivity
to the epoxide is markedly higher. Secondly, the activity
a
Table 4 Oxidation of norbornylene
b
a
Norbornylene conversion (%)
Material
Template : Si mol ratio
Crystal shape and size /mm
Norbornylene oxide selectivity (%)
None
Silicalite-1
—
—
—
3.2
0.8
14
—
—
14
17
b
m, 10 (mt)
m, 10 (mt)
s, 0.15
c
TS-1
1.0
0.35
c
TS-1
42
a
2 2
Reaction conditions: Norbornylene (1.0 g), H O (1 eq.), catalyst (0.1 g), methanol (10 ml) at 60 1C for 24 h. Average particle size determined
c
from electron microscopy, s spherical/irregular, m coffin shaped monoclinic, mt mutiply twinned. Crystallisation time 288 h.
2
676
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 2 6 7 1 – 2 6 7 8
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5

View Article Online
declines with increasing crystallite size as does the selectivity to
the epoxide. It could be expected that the ring-opening reaction
of the epoxide could be preferentially catalysed on the surface
of the TS-1 crystallites or at the pore openings, since the larger
transition state for this reaction would be favoured at these
sites, and this may be a factor in the decreased epoxide
selectivity observed as the crystallite size increases. However,
even the large crystallite TS-1 that has been reacted with
chlorotrimethylsilane and calcined is active and the catalytic
results are very similar for the non-treated material. It should
be noted that crotyl alcohol has a relatively small van der
Waals radius and so can readily diffuse into the channels within
both the large (10 mm) and small (0.2–0.3 mm) crystallites.
Hence, we consider that diffusion of the products within the
porous structure plays a role, since the diffusion pathway will be
significantly longer in the larger crystallites and this provides
reaction pathways for the ring-opening reactions that are cata-
lysed by the weak acid sites present within the pores of TS-1.
Thirdly, and most importantly, the amorphous and globular
precursor structures to TS-1 are also active for this reaction
with similar selectivity and activity to the corresponding
crystalline TS-1 formed at longer crystallisation time. This
indicates that the formation of crystalline TS-1 is not a
prerequisite for this reaction.
makes it difficult to make a full range of definitive comments
concerning the relationship between structure and activity,
since selectivity is dependent on conversion, a number of
important novel observations are still possible as for many
related samples the conversions observed for crotyl alcohol
oxidation are very similar. With respect to crotyl alcohol
oxidation, selected as a relatively non-reactive substituted
alkene, did not require the TS-1 structure for reactivity as
demonstrated by the high reactivity observed with the amor-
phous and globular precursor materials, although crystalline
TS-1 is more reactive, indicating that model reactions of this
type are not dependent on the TS-1 structure. In contrast,
phenol hydroxylation is only catalysed by crystalline TS-1.
This shows that the oxidation reactions are very sensitive to the
degree of structural order within the material. In particular,
phenol is not readily oxidised and the TS-1 12-day sample
comprising large well ordered crystallites of TS-1 is inactive
under the standard reaction conditions. This may be due to
two factors, namely the van der Waals cross-section and the
electronic nature of the molecule. However, norbornylene is
much more readily oxidised than phenol over the same 12-day
TS-1 sample and norbornylene is too large to enter the micro-
porous structure of TS-1. The observed reactivity with nor-
bornylene is due solely to the oxidation on the exterior
crystallite surface of TS-1, or in the vicinity of the pore mouths
on the exterior of the crystallites. Since phenol shows no
activity with the large crystallite TS-1 12-day sample, this
could indicate that the electronic structure of phenol may be
the dominant factor. Indeed, phenol has a smaller van der
Waals radius than the pores of TS-1 and, consequently, should
readily diffuse into the mesopores of the crystallites. In view of
this, on the basis of molecular size alone, the reactivity of
phenol should be higher than norbornylene, but this is not
observed. However, another factor masks the reactivity of
phenol with TS-1 as rapid deactivation can mask activity. This
has been previously discussed in detail by Clerici and Ingalli-
For the oxidation of phenol using large crystallite TS-1 (ca.
1
0 mm, prepared using template : Si mol ratio ¼ 1.0) very few
samples showed any catalytic activity, although all were con-
firmed by the characterisation methods to be well crystalline
TS-1. The hydroquinone : catechol ratio was observed to be
independent of the reaction time and was ca. 3.0 for these large
crystallites. The highest activity was observed for the material
formed with a 2-day crystallisation period. Interestingly, reac-
tion of this sample with chlorotrimethylsilane followed by
calcination resulted in complete loss of the catalyst activity
for phenol hydroxylation. Extended crystallisation times led to
a decrease in activity and the sample prepared with 12-day
crystallisation was found to be inactive.
4
5
46
na and Langhendries et al. with reference to the size of TS-1
crystallites.
The activities of the crystalline TS-1 with smaller crystallites
ca. 0.15 mm, prepared using template : Si mol ratio 0.35),
(
The low relative reactivity of phenol as compared with the
other substrates used in this study, makes it a poor probe
molecule to determine the activity of TS-1 samples. This
indicates that the phenol hydroxylation reaction should not
be used in isolation as a test reaction for TS-1 samples. In
addition to low reactivity, one of the products, hydroquinone,
can readily be oxidised to quinone which can subsequently
polymerise in the microporous structure. This effect, which has
crystallised for 2 and 12 days, were very similar. However, it
is interesting to note that the activity of the 10 mm TS-1 was
significantly higher than the 0.15 mm TS-1 for the 2-day
crystallisation materials. In contrast, the activity of the 12-
day samples was reversed as the 10 mm TS-1 was inactive.
Reaction of the 2-day TS-1 small crystallite (0.2–0.3 mm)
material with chlorotrimethylsilane with subsequent calcina-
tion decreased the conversion. This result contrasts with the
3
2
been discussed extensively previously, will change the relative
molar ratios of the observed products. The ratio of hydro-
quinone : catechol from phenol hydroxylation has been used by
many researchers to quantify the reactivity of external/internal
active sites. Since the subsequent oxidation of the two pro-
ducts, hydroquinone and catechol, can vary, this implies that
this reaction cannot be used as a probe for the reactivity of the
internal and external crystallite active sites. This is further
emphasised in the present study for the samples treated with
chlorotrimethylsilane. The 2-day TS-1 (10 mm) sample treated
with chlorotrimethylsilane was observed to be inactive for
phenol hydroxylation, whereas the non-treated material was
the most active in the series of samples. Even for the very small
crystallite sample (0.2–0.3 mm), treatment with chlorotri-
methylsilane dramatically decreased the reactivity for phenol
hydroxylation. In contrast, the 2-day TS-1 (10 mm) sample
treated with chlorotrimethylsilane is almost as active for crotyl
alcohol epoxidation as the non-treated sample, and there is no
effect on the product selectivity. These observations tend to
confirm that the oxidation of phenol occurs at active sites on
the exterior crystallite surface or within the pores close to the
surface of the crystallites.
2
-day large crystallite TS-1 sample treated with chlorotrimethyl-
silane in the same way, when complete loss of activity is
observed. However, only catechol is observed as a product
for the 2-day treated TS-1 (0.15 mm) sample after 24 h reaction
time (Table 3). Catechol has a larger van der Waals radius than
hydroquinone and, consequently, could be expected to be
formed to a lesser extent on the basis of shape selectivity. Even
at shorter reaction times, e.g. up to 6 h, hydroquinone was not
observed, but quinone is observed in an equal amount to
catechol. Quinone is formed readily from the oxidation of
hydroquinone, but would also be expected to polymerise
rapidly within the micropores of the TS-1 catalyst, as has been
3
2
described previously.
Comments on structure–activity relationship and the use of
phenol hydroxylation as a probe reaction
From the catalysis studies presented in this work, there are
clear differences between the reactivity for the oxidation of
crotyl alcohol, phenol and norbornylene with the TS-1 sam-
ples. It should be noted that the catalysis studies have com-
pared data at set reaction times and consequently a range of
conversions and selectivities have been observed. While this
There is one important feature of the phenol hydroxylation
reaction that gives it a considerable advantage as a probe
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 2 6 7 1 – 2 6 7 8
2677

View Article Online
1
7
8
I. Halsz, M. Agarwal, E. Senderov and B. Marcus, Stud. Surf. Sci.
Catal., 2003, 145, 435.
O. A. Kholdeeva, M/S. Mel’gunov, A. N. Shmakov, N. N.
Trukhan, V. V. Kriventsov, V. I. Zaikovskii, M. E. Malyshev
and V. N. Roamannikov, Catal. Today, 2004, 91–92, 205.
reaction, namely no activity is observed for homogeneously
catalysed processes. This indicates that leached-Ti from the
catalyst structure will not influence the observed reactivity for
phenol hydroxylation. The same is not true for other sub-
strates, and we have previously shown that leached-Ti can play
1
19 G. J. Hutchings and D. F. Lee, J. Chem. Soc., Chem. Commun.,
1994, 1095.
2
3
a role in the observed reactivity for crotyl alcohol oxidation.
2
2
0
1
A. Bhaumik and T. Tatsumi, J. Catal., 1999, 182, 349.
L. J. Davies, P. McMorn, D. Bethell, P. C. B. Page, F. King, F. E.
Hancock and G. J. Hutchings, Chem. Commun., 2000, 1807.
L. J. Davies, P. McMorn, D. Bethell, P. C. B. Page, F. King, F. E.
Hancock and G. J. Hutchings, J. Catal., 2001, 198, 319.
Indeed, homogeneous Ti species catalyse the formation of the
triol and the Ti-leaching is related to the concentration of triol
present. Hence, once leaching commences during the oxidation
of a substrate that can form a triol, the effect becomes auto-
2
2
2
3
catalytic. In the present study, it is apparent that larger
crystallite TS-1 (10 mm) forms more triol than the smaller
crystallite TS-1 (0.15 mm). This is considered to be due to the
longer diffusion pathway in the larger crystals which will
enhance the probability of hydrolysis occurring.
The results of this study show that care needs to be exercised
in selecting model or test reactions for the evaluation of the
catalytic activity of the redox molecular sieve TS-1, since for
reactions of allylic alcohols well ordered structures, although
more active, are not a prerequisite, whereas the reactivity of
larger molecules is significantly influenced by the electronic
nature of the molecule.
23 L. J. Davies, P. McMorn, D. Bethell, P. C. B. Page, F. King, F. E.
Hancock and G. J. Hutchings, Phys. Chem. Chem. Phys., 2001, 3,
6
32.
2
2
4
5
B. Notari, Stud. Surf. Sci. Catal., 1987, 37, 413.
G. Perego, G. Bellussi, C. Corno, M. Taramasso, F. Buonomo
and A. Esposito, Stud. Surf. Sci. Catal., 1986, 28, 129.
A. Thangaraj, R. Kumar, S. P. Mirajkar and P. Ratnasamy,
J. Catal., 1990, 130, 1.
2
6
27 J. S. Reddy, R. Kumar and P. Ratnasmy, Appl. Catal., 1990, 58,
L1.
28
29
30
31
32
A. Thangaraj, R. Kumar and P. Ratnasamy, J. Catal., 1991, 131,
94.
A. Tuel, S. Moussa-Khouzami, Y. Ben Tarrit and C. Naccache,
J. Mol. Catal., 1991, 68, 45.
D. R. C. Huybrechts, I. Vaesen, H. X. Li and P. A. Jacobs, Catal.
Lett., 1991, 8, 237.
A. J. H. P. van der Pol, A. J. Verduyn and J. H. C. van Hooff,
Appl. Catal. A, 1992, 92, 113.
2
Acknowledgements
We thank Quest International and Synetix for financial
support.
U. Wilkenho
W. Gammon, P. A. Jacobs and E. van Steen, J. Catal., 2001, 203,
01.
¨
ner, G. Langhendries, F. van Laar, G. V. Baron, D.
2
3
3
R. Gougeon, L. Delmotte, D. Le Nouen and Z. Gabelica, Micro-
porous Mesoporous Mater., 1998, 26, 143.
References
1
R. M. Contractor, H. S. Horowitz, G. M. Sisteer and E. Bordes,
Catal. Today, 1997, 37, 51.
34 A. Cizmek, B. Subotica, K. Damir, V. Babic-Ivancic and A.
Tonjec, Microporous Mater., 1997, 12, 267.
2
3
M. Tarramasso, G. Perego and B. Notari, US Pat., 4410501, 1983.
A. J. H. P. Van der Pol, A. J. Verduyn and J. H. C. van Hooff,
Appl. Catal., 1992, 92, 133.
35 S. L. Burkett and M. E. Davis, Chem. Mater., 1995, 7, 920.
36 S. L. Burkett and M. E. Davis, Chem. Mater., 1995, 7, 1453.
37 C. E. A. Krirchhock, R. Ravishankar, F. Verspeurt, P. J. Grobet,
P. A. Jacobs and A. Martens, J. Phys. Chem. B, 1999, 103, 4965.
38 Q. Li, B. Mihailova, D. Creaser and J. Sterte, Microporous
Mesoporous Mater., 2001, 43, 51.
39 J. C. Van Hooff and A. J. Van der Pol, Appl. Catal. A: Gen., 1992,
92(2), 93.
40 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. B. Higgins and J. L. Schlenter,
J. Am. Chem. Soc., 1992, 114, 10834.
4
5
6
R. A. Sheldon, J. Mol. Catal. A, 1996, 107, 75.
S. B. Ogunwumi and T. Bein, Chem. Commun., 1997, 901.
P. Piaggo, P. McMorn, C. Langham, D. Bethell, P. C. B. Page,
F. E. Hancock and G. J. Hutchings, New J. Chem., 1998, 1167.
P. Piaggio, C. Langham, P. McMorn, D. Bethell, P. C. B. Page, F.
E. Hancock, C. Sly and G. J. Hutchings, J. Chem. Soc., Perkin
Trans. 2, 2000, 143.
M. G. Clerici and P. Ingallina, J. Catal., 1993, 140, 71.
M. G. Clerici, G. Bellusi and V. Romano, J. Catal., 1991, 129, 159.
M. G. Clerici and G. Bellusi, Eur. Patent, 1989, 315247.
T. Tatsumi, M. Yako, M. Nakamura, Y. Huyara and H. Tomi-
naga, J. Mol. Catal., 1993, 78, L41.
7
8
9
0
1
1
1
41 E. Duprey, P. Beaunier, M. A. Springuel-Huet, F. Bozon-Ver-
´
duraz, J. Fraissard, J.-M. Manoti and J.-M. Bregeault, J. Catal.,
1997, 165, 22.
12
13
14
15
16
B. Kraushaar-Czarnetski and J. H. C. van Hooff, Catal. Lett.,
1989, 2, 43.
M. A. Camblor, M. Constantini, A. Corma, L. Gilbert, P. Esteve,
A. Martinez and S. Valencia, Chem. Commun., 1996, 1330.
F. Rey, G. Sankar, T. Maschmeyer, J. M. Thomas, R. G. Bell and
G. N. Greaves, Top. Catal., 1996, 3, 121.
42 G. L. Marra, G. Artioli, A. N. Fifch, M. Milanesio and C.
Lamberti, Microporous Mesoporous Mater., 2000, 40, 85.
43 Q. Zhao, X. H. Bao, X. W. Han, X. M. Liu, D. L. Tan, L. W. Lin,
X. W. Guo, G. Li and X. S. Wang, Mater. Chem. Phys., 2000, 66,
41.
44 G. Ricchiardi, A. Damin, S. Bordiga, C. Lamberti, G. Spano, F.
Rivetti and A. Zecchina, J. Am. Chem. Soc., 2001, 123, 11409.
45 M. G. Clerici and P. l. Ingallina, J. Catal., 1993, 140, 71.
46 G. Langhendries, D. E. De Vos, G. V. Baron and P. A. Jacobs,
J. Catal., 1999, 187, 453.
M. Kang, W.-J. Hong and M.-S. Park, Appl. Catal., B, 2004, 53,
195.
A. Tharanj, S. M. Kuznicki and G. S. Koermer, Int. Pat.,
WO 2000023186, 2000 (assigned to Engelgard Corp.).
2
678
P h y s . C h e m . C h e m . P h y s . , 2 0 0 5 , 7 , 2 6 7 1 – 2 6 7 8
T h i s j o u r n a l i s & T h e O w n e r S o c i e t i e s 2 0 0 5