Enhanced Catalytic Performance of Titanium Silicalite-1 in Tuning the Crystal Size in the Range 1200-200nm in a Tetrapropylammonium Bromide System

Article abstract of DOI:10.1002/cctc.201500440

A facile method for the size-controllable synthesis of titanium silicalite-1 (TS-1) is presented. A tetrapropylammonium bromide hydrothermal system was used, and the amount of seeds (silicalite-1 suspension) and crystallization time were tuned to control the crystal size. The crystal size could be adjusted from 200 to 1200nm by varying the seed amount from 12 to 0.05wt%. Crystallization time plays a less important role than the seed amount on the crystal size. TS-1 samples with different crystal sizes were characterized and evaluated in propene epoxidation. The catalytic activity and selectivity of propene oxide are enhanced by decreasing the crystal size from 1200 to 200nm because the diffusion limitation is eliminated gradually. The seed is significant for this system, as a lack of seeds leads to poor crystallization and low catalytic activity. The mechanism of the seed function was studied by simulating the transformation process of the seed in the TS-1 synthesis system.

Full text of DOI:10.1002/cctc.201500440

DOI: 10.1002/cctc.201500440
Full Papers
Enhanced Catalytic Performance of Titanium Silicalite-1 in
Tuning the Crystal Size in the Range 1200–200 nm in
a Tetrapropylammonium Bromide System
Yi Zuo,[a] Min Liu,[a] Ting Zhang,[a] Changgong Meng,[a] Xinwen Guo,*[a] and
Chunshan Song*[b]
A facile method for the size-controllable synthesis of titanium
silicalite-1 (TS-1) is presented. A tetrapropylammonium bro-
mide hydrothermal system was used, and the amount of seeds
(silicalite-1 suspension) and crystallization time were tuned to
control the crystal size. The crystal size could be adjusted from
200 to 1200 nm by varying the seed amount from 12 to
0.05 wt%. Crystallization time plays a less important role than
the seed amount on the crystal size. TS-1 samples with differ-
ent crystal sizes were characterized and evaluated in propene
epoxidation. The catalytic activity and selectivity of propene
oxide are enhanced by decreasing the crystal size from 1200
to 200 nm because the diffusion limitation is eliminated gradu-
ally. The seed is significant for this system, as a lack of seeds
leads to poor crystallization and low catalytic activity. The
mechanism of the seed function was studied by simulating the
transformation process of the seed in the TS-1 synthesis
system.
Introduction
Titanium silicalite-1 (TS-1) was first synthesized hydrothermally
by Taramasso et al. in 1983.[1] As a result of its excellent catalyt-
ic activity for selective oxidation reactions, such as alkane oxi-
dation,[2] alkene epoxidation,[3] aromatics hydroxylation,[4]
ketone ammoximation,[5] and desulfurization,[6] TS-1 attracts
much attention. The hydroxylation of phenol with H2O2 over
TS-1 to produce benzenediols was first commercialized by Eni-
chem in 1986.[7] In recent years, the hydrogen peroxide to pro-
pene oxide (HPPO) route has been commercialized by BASF/
Dow Chemical and Evonik/Uhde in Belgium and South Korea,
respectively.[8,9] Some institutes also tried this route in pilot
plants.[10,11] Propene oxide (PO) is an important chemical inter-
mediate mainly in the production of polyether polyol poly-
mers, and the HPPO route catalyzed by TS-1 provides environ-
mental and economic benefits. The traditional commercial
manufacturing routes include the chlorohydrin route and the
Halcon route.[12] The chlorohydrin route produces a lot of pol-
lution, whereas the Halcon route is capital intensive and its
economics depend on those of the coproducts.[13] Despite the
benefits of the HPPO/TS-1 route, many problems still exist
such as the short lifetime and low output of the catalyst. Fur-
thermore, the cost of TS-1 is always an obstacle for its industri-
al application.[14] The expensive template tetrapropylammoni-
um hydroxide (TPAOH) is the biggest influence on the cost of
resources. Therefore, many researchers focus on the reduction
of its dosage or finding an alternative.[15–17]
Müller and Steck used a tetrapropylammonium bromide
(TPABr) template, which has the same alkyl cation as TPAOH at
a much lower cost, to synthesize TS-1.[18] Many subsequent
studies have been performed in TPABr systems.[19,20] Unfortu-
nately, the introduction of BrÀ and decrease of the base
strength leads to a larger TS-1 crystal size, typically, more than
1 mm. Consequently, the catalytic performance is affected neg-
atively by diffusion limitations. The synthesis of hierarchical TS-
1 can eliminate the diffusion limitations, for which typical
methods include post-treatment with alkaline solutions[21–25]
and one-pot synthesis.[26,27] Alternatively, the crystal size of TS-
1 synthesized from the TPABr system has been decreased by
the introduction of different seed types.[28,29] The function of
seed addition is to ensure that nuclei formation does not limit
new crystal formation, which favors the growth of small crys-
tals from many nuclei over the growth of large crystals from
few nuclei. The other important factor that influences the cata-
lytic performance of TS-1 is the coordination state of Ti and
the diffusion properties. Ti coordinated tetrahedrally is the pri-
mary active center for selective oxidation, and we found that
Ti coordinated pentahedrally showed a higher activity than tet-
rahedral Ti.[30] However, the synthesis of TS-1 with a high con-
tent of pentahedral Ti is difficult.
[a] Dr. Y. Zuo, Dr. M. Liu, T. Zhang, Prof. C. Meng, Prof. X. Guo
State Key Laboratory of Fine Chemicals
PSU-DUT Joint Center for Energy Research
School of Chemical Engineering
Dalian University of Technology
2 Linggong Road, Dalian 116024 (P. R. China)
E-mail: guoxw@dlut.edu.cn
[b] Prof. C. Song
EMS Energy Institute
PSU-DUT Joint Center for Energy Research
and Department of Energy & Mineral Engineering
Pennsylvania State University
University Park, PA 16802 (USA)
E-mail: csong@psu.edu
Supporting Information for this article is available on the WWW under
http://dx.doi.org/10.1002/cctc.201500440.
ChemCatChem 2015, 7, 2660 – 2668
2660
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Previously, we obtained TS-1 with a crystal size of approxi-
mately 600 nm by using a crystallized suspension that con-
tained nano-sized TS-1 as the seed in a TPABr hydrothermal
system.[31,32] The smaller crystal size gave rise to an improved
catalytic performance in propene epoxidation and phenol hy-
droxylation. However, the control of the crystal size was diffi-
cult by this method, possibly because of the large seed crystal
size and the difficult dissolution of the Ti-containing seeds.
Herein, a facile method to adjust the crystal size of TS-1 by
changing the seed amount and crystallization time is reported.
Nano-sized TS-1 (200 nm) was obtained from the TPABr
system. The relationship between the catalytic performance
and the crystal size of TS-1 was studied. Furthermore, the
mechanism of the seed function in the preparation of TS-1 is
determined.
crystallinity, which is calculated by comparing the total intensi-
ties of the five characteristic peaks of the samples with that of
the largest one, are quite similar for all samples formed from
seeds. Thus, a crystallization time of 72 h is sufficient if seeds
are added, although the content of seed differed significantly
(0.05–12 wt%). We conclude that nucleation is the rate-limiting
step in the synthesis process, and a small amount of seeds,
which can provide crystallization nuclei, can accelerate crystalli-
zation significantly. The peak at 2q=25.48 is assigned to ana-
tase TiO2, which causes the decomposition of H2O2 easily and
decreases H2O2 utilization. Thus, the generation of anatase TiO2
is disadvantageous for selective oxidation. However, all the
crystallized samples show a similar and weak peak at 2q=
25.48, which indicates that a small amount of anatase TiO2
exists in all the samples.
The FTIR spectra of the samples formed by adding different
amounts of seeds are shown in Figure 2. In the FTIR spectra of
Results and Discussion
Different amounts of seeds
The TS-1 samples synthesized by adding different amounts of
seed were characterized and evaluated for propene epoxida-
tion activity. The preparation processes are described in the Ex-
perimental Section. The powder XRD patterns of the samples
synthesized with different amounts of seeds are shown in
Figure 1. The characteristic peaks of TS-1 with MFI (Mobil Fifth)
topology were sited at 2q=7.8, 8.8, 23.0, 23.9, and 24.48.[33]
Samples are labeled with seed amount (0–12 wt% for samples
in Figure 1) followed by the crystallization time (72 h for sam-
ples in Figure 1). Details of the notation are explained in the
Experimental Section. All the samples show five characteristic
peaks, except TS-1-0-72, which indicates that TS-1 is difficult to
crystallize without seeds in this synthesis system. The relative
Figure 2. FTIR spectra of TS-1 samples synthesized with different amounts of
seeds.
TS-1, the band at n˜ =550 cmÀ1 is characteristic of the double-
frequency vibration in a five-membered ring of MFI topolo-
gy.[34] TS-1-0-72 shows a very weak band at n˜ =550 cmÀ1,
which indicates that the MFI framework has not been formed
completely. This corroborates the XRD characterization. The
bands in other samples are clear and similar because of the
generally completed crystallization. The band at n˜ =960 cmÀ1
,
which is usually assigned to the stretching vibration of SiÀO in
the neighborhood of Ti, is considered to be related to the in-
sertion of Ti into the framework,[35] and the intensity of this
band reflects the amount of Ti inserted into the framework.
However, some researchers attributed this band to defects of
SiÀOH or Ti=O.[36,37] The opinion accepted widely is that the
appearance of this band can be considered as an indirect
proof of the insertion of Ti into the TS-1 framework.[38] The
samples synthesized with seeds all show this band, but that in
TS-1-0-72 is very weak because of the very small amount of Ti
in this TS-1 framework. Framework Ti is considered as the
active center for the selective oxidation reactions. Therefore, it
can be inferred that the catalytic activity of TS-1-0-72 is lower
than that of the other samples.
Figure 1. XRD patterns of TS-1 samples synthesized with different amounts
of seeds.
ChemCatChem 2015, 7, 2660 – 2668
www.chemcatchem.org
2661
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
pattern. UV/vis spectra indicate that the content of framework
Ti decreased with the increasing amount of seed. Chen et al.
also found that the addition of seed caused the increase of
extra-framework Ti.[43] The acceleration of crystallization is dis-
advantageous to the insertion of Ti into the framework. Fan
et al. found that the addition of ammonium salts, which
reduce the crystallization rate and change the crystallization
mechanism, led to a TS-1 nearly free of extra-framework Ti.[44]
The anionic polyelectrolyte was proven to have a similar effect
with ammonium salts.[45] In contrast, a faster crystallization
might decrease the content of framework Ti. The increase of
the seed amount, which leads to the decrease of the relative
content of total Ti, may also be a reason for the reduction of
framework Ti.
Figure 3. UV/vis spectra of TS-1 samples synthesized with different amounts
SEM images of the TS-1 samples formed using different
amounts of seeds are exhibited in Figure 4. All the samples
show an approximately cubic crystal, with the exception again
of TS-1-0-72, which shows mainly amorphous blocks. The size
of the crystallites increased with a reduction of seed content. If
we changed the content of seed from 9 to 0.05 wt%, the TS-
1 crystal size can be tuned from 200 to 1200 nm. A smaller
crystal size is usually expected to show a higher catalytic activi-
ty and PO selectivity because of the elimination of diffusion
limitations. An increase of the seed content beyond 9 wt% did
not provide a clear monotonic trend in crystal size. At a high
seed loading, the limiting function of the seed is alleviated as
is discussed later (Mechanism of seed function). A crystal size
of seeds.
Diffuse reflectance UV/vis spectroscopy is one of the first
spectral techniques used for the detection of the Ti coordina-
tion state in TS-1. The coordination state of Ti is significant for
the catalytic performance of TS-1. The UV/vis spectra of TS-
1 usually have three sets of bands (Figure 3). The band at l
ꢀ210 nm is attributed to the Ti4+O2À to Ti3+OÀ ligand-to-metal
charge transfer,[39] which is characteristic of tetrahedrally coor-
dinated framework Ti. It was reported that the five-membered
ring formed by tetrahedral Ti with H2O2 and alcohol can oxi-
dize propene to propene oxide (PO). Therefore, the content of
tetrahedral Ti is related to the
catalytic activity of TS-1. Anatase
TiO2 shows a band at l=310–
330 nm. The assignment of the
band between l=230 and
280 nm is controversial. Blasco
et al. claimed that it belongs to
a
Ti coordinated octahedrally
that contains a TiÀOÀTi bond,[40]
whereas Geobaldo et al. attribut-
ed it to isolated Ti species that
coordinate with two water mole-
cules.[41] We also studied the as-
signment of this band of Ti in
our previous work,[30,42] and
found that at least two kinds of
Ti species existed in this region.
We conclude that the band at
l=250–280 nm in the samples
in this work can be attributed to
Ti
coordinated
octahedrally
formed by one Ti coordinated
tetrahedrally that bonds with
two water molecules. This Ti spe-
cies is inert for the selective oxi-
dation reaction. The highest
amount of octahedral Ti was ob-
served in TS-1-0-72 because the
TS-1 framework has not been
generated according to its XRD Figure 4. SEM images of TS-1 samples synthesized with different amounts of seeds.
ChemCatChem 2015, 7, 2660 – 2668
www.chemcatchem.org
2662
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
larger than 1200 nm may be obtained if we continue to de-
crease the content of seed. However, a larger crystal is expect-
ed to have a lower catalytic performance because of diffusion
limitations, and its synthesis is more difficult because of the
small amount of seed. Furthermore, the larger crystals ob-
tained in this synthesis system were reported by other re-
searchers.[46] Therefore, we did not continue with a lower con-
tent of seed.
Plots of the crystal size versus the content of seed are
shown in Figure 5. The data fit well to a power function with
a fit equation of [Eq. (1)]:
Crystal size ¼ 468:7 Â ðcontent of seedÞÀ0:3158
ð1Þ
Figure 6. Ar physisorption isotherms of TS-1 samples synthesized with differ-
ent amounts of seeds.
Figure 5. Crystal size of TS-1 plotted against the amount of seed using in
synthesis. (Data points represent experimental data and the continuous
curve represents a power law fit to the data with the equation given on the
plot.)
The degree of fitting measured by the coefficient of determi-
nation (R2) is 0.9968.
Ar physisorption isotherms and pore diameter distribution
curves of TS-1 synthesized with different amounts of seeds are
shown in Figures 6 and 7, respectively. All the samples have
a type I adsorption–desorption curve, except TS-1-0-72, which
indicates that the samples synthesized with seeds contain
many micropores. TS-1-0-72 shows a very low adsorption ca-
pacity at a low relative pressure because of the unformed
framework and micropores. A hysteresis loop appears in some
of the samples at a high relative pressure, which is attributed
to adsorption/desorption from the intercrystal spaces. The
space volume increased with the decrease of crystal size. How-
ever, the hysteresis loop in the isotherm of TS-1-0-72 is much
larger than that in other samples, which is because of the
larger gaps between the amorphous particles in TS-1-0-72. The
crystallized samples show a pore diameter of ꢀ0.5 nm, which
is the inherent channel dimension of TS-1. No regular pore di-
ameter was observed in TS-1-0-72 between 0.3 and 10 nm.
The surface areas and pore volumes calculated according to
Ar physisorption are listed in Table 1. The external surface area
(AEx) of the crystallized samples increased and the microporous
surface area (AMicro) decreased with the decrease of crystal size.
A larger external surface area may be beneficial for the conver-
Figure 7. Pore diameter distribution curves of TS-1 samples synthesized with
different amounts of seeds.
sion of substrates. Overall, the total surface area (AT) changed
only slightly, and non-monotonically, with crystal size. The total
pore volume (VT) increased slightly, probably because of the in-
crease of the intercrystal space volume.
The catalytic performance for propene epoxidation over the
TS-1 samples with different amounts of seeds are shown in
Table 2. Nano-sized TS-1 prepared according to Ref. [47] was
also evaluated for comparison. The conversion of H2O2
(X(H2O2)) over TS-1 first increased and then decreased with the
decrease of crystal size. Two main factors are important in the
oxidation reactions catalyzed by TS-1: diffusion properties and
ChemCatChem 2015, 7, 2660 – 2668
www.chemcatchem.org
2663
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
Table 1. Surface areas, pore volumes, and elemental compositions of the TS-1 samples synthesized with different amounts of seeds.[a]
Sample
AT [m2 gÀ1
]
AMicro [m2 gÀ1
]
AEx [m2 gÀ1
]
VT [cm3 gÀ1
]
VMicro [cm3 gÀ1
]
In the bulk
On the surface
Gel
Si [wt%] Ti [wt%] n(Si/Ti) Si [wt%] Ti [wt%] n(Si/Ti) n(Si/Ti)
TS-1-0-72
TS-1-0.05-72 423
138
0
339
335
336
334
328
305
300
296
138
85
97
102
104
110
124
130
145
0.74
0.25
0.26
0.28
0.29
0.33
0.33
0.35
0.35
0
45.31
45.41
45.45
45.52
45.55
45.59
45.61
45.62
45.69
1.75
1.62
1.56
1.47
1.44
1.38
1.36
1.34
1.26
44.38
48.05
49.95
53.09
54.22
56.64
57.49
58.37
62.16
45.83
45.59
45.57
45.31
45.24
45.42
45.50
45.57
45.66
1.08
1.39
1.41
1.74
1.84
1.60
1.50
1.41
1.29
72.82
56.40
55.43
44.62
42.24
48.71
51.86
55.43
60.62
40.00
40.02
40.04
40.20
40.40
41.20
42.40
43.60
44.80
0.13
0.13
0.13
0.13
0.12
0.12
0.11
0.11
TS-1-0.1-72
TS-1-0.5-72
TS-1-1-72
TS-1-3-72
TS-1-6-72
TS-1-9-72
TS-1-12-72
432
438
442
438
439
430
431
[a] AT, AMicro, and AEx are the total, microporous, and external surface areas, respectively. VT and VMicro are the total and microporous pore volumes, respec-
tively. n(Si/Ti) is the molar ratio of Si/Ti.
In summary, the control of the TS-1 crystal size can be ac-
Table 2. Catalytic performances of propene epoxidation over the
TS-1 samples synthesized with different amounts of seeds.[a]
complished by changing the amount of seed in the TPABr
system. The controllable range is from 200 to 1200 nm and
Catalyst
X(H2O2) [%][b]
S(PO) [%]
U(H2O2) [%]
possibly more. The smallest crystal size obtained at 1708C after
72 h with a seed content of 9 or 12 wt% is in the range of
nano-sized zeolites (not more than 200 nm). The catalytic activ-
ity increased and then decreased with the decrease of crystal
size, which was because of the synergistic effect of the crystal
size and content of framework Ti.
TS-1-0-72
4.4
68.9
70.0
74.4
75.8
74.5
74.5
74.3
67.1
77.0
100
81.6
81.2
82.2
84.2
85.2
85.5
85.3
87.9
86.4
80.3
TS-1-0.05-72
TS-1-0.1-72
TS-1-0.5-72
TS-1-1-72
TS-1-3-72
TS-1-6-72
TS-1-9-72
TS-1-12-72
Nano-sized TS-1
89.0
90.0
92.9
94.6
95.3
96.1
97.8
98.2
95.9
Different crystallization times
The influence of crystallization time on the crystal size and cat-
alytic performance were also studied. The content of seed was
kept constant at 6 wt%. The XRD patterns of the samples crys-
tallized for different times (Figure 8) indicate that the five char-
acteristic peaks began to appear at 15 h. The TS-1 framework
was not formed before 12 h. The relative crystallinity increased
with a prolonged crystallization time until 24 h. SEM images of
the samples (Figure 9) confirmed the XRD results. Mainly amor-
phous substrates were present in TS-1-6-6, and small units
began to form in TS-1-6-12. Then, the small units combined
with each other to generate TS-1 crystals, but the crystal size
was not very uniform in TS-1-6-15. With a further increase of
the crystallization time, the crystals became uniform and the
[a] Reaction conditions: catalyst 0.2 g, 408C, 1 h, propene 0.4 MPa, H2O2
concentration 1.4 molLÀ1. [b] X(H2O2), S(PO), and U(H2O2) are the conver-
sion of H2O2, selectivity of PO, and utilization of H2O2, respectively.
framework Ti content. A smaller crystal size may be expected
to yield a higher X(H2O2) by eliminating diffusion limitations.
However, an increased seed content (the seeds are silicalite-1)
led to a relatively low content of framework Ti (see UV/vis
spectra), which influences the catalytic activity negatively.
Therefore, the net conversion arises from the competing ef-
fects of crystal size and framework Ti content. The molar ratios
of Si/Ti (n(Si/Ti)) in the bulk determined by inductively coupled
plasma atomic emission spectrometry (ICP-AES) are shown in
Table 1. With the increase of seed amount, the n(Si/Ti) in the
bulk decreases, and the extent of the decrease of the actual
n(Si/Ti) is larger than that of the feeding one, which indicates
that less Ti stays in the TS-1 at a faster crystallization rate. This
is mainly because of the decreasing amount of framework Ti.
The selectivity to PO (S(PO)) increased with the decrease of
crystal size because of the gradually eliminated diffusion limita-
tions. All the samples contained a small amount of anatase
TiO2, thus, the utilization of H2O2 (U(H2O2)) values were all
lower than 90%. Most of the Ti in TS-1-0-72 exists as extra-
framework Ti coordinated octahedrally, proven by XRD and
UV/vis spectrophotometry, and was inert for epoxidation and
the decomposition of H2O2. Therefore, the conversion of H2O2
over TS-1-0-72 was much lower than that over other samples,
but the U(H2O2) was similar to that of the other samples.
Figure 8. XRD patterns of TS-1 samples crystallized for different times.
ChemCatChem 2015, 7, 2660 – 2668
www.chemcatchem.org
2664
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
We conclude that the ability
to control the TS-1 crystal size
by tuning the crystallization time
is limited. A reduction of the
time can decrease the crystal
size slightly but also shows
a negative effect on the conver-
sion of H2O2.
Mechanism of seed function
Researchers have expressed dif-
ferent opinions on the function
of seeds in crystallization.[48–50]
The seed may perform as a nu-
cleus directly.[51,52] Others con-
clude that the seed needs to dis-
solve first and functions as
a structure-directing agent.[53,54]
Warzywoda et al.[55] found that
competitive growth occurred on
the seed surface with that in the
liquid phase. Thus, we designed
a series of experiments to study
how the seed changed under
these synthesis conditions.
The experiments are described
in the Experimental Section. SEM
Figure 9. SEM images of TS-1 samples crystallized for different times.
images of the 9 wt% seeds treat-
ed for different times are shown
in Figure 10. The untreated sili-
size increased slightly. The average crystal size increased from
260 to 276 nm if the crystallization time increased from 15 to
72 h. Once the crystal was formed, further crystallization time
hardly affected the TS-1 size.
calite-1 (0 h) shows irregular crystals. The average crystal size is
ꢀ100 nm. After treatment for 1 h, the crystals dissolved and
formed large blocks of substrate. This result is also shown by
XRD characterization (Figure 11), in which the five characteristic
peak intensities decreased clearly. The large blocks may not be
the amorphous substrates, although the crystallinity is quite
low. The very small units with MFI topology also show a low
crystallinity because of the determination limitations of XRD.
We conclude that amorphous substrates cannot perform as
seeds. After treatment for another hour, many small crystals
appeared and the relative crystallinity increased, which indi-
cates that the small units began to reform or recrystallize. As
the treatment time was prolonged, more crystals were gener-
ated, and the large blocks disappeared. This phenomenon
demonstrates that at this amount (9 wt%), the silicalite-1 dis-
solves to form secondary structural units, which perform as
structure-directing agents or nuclei in the synthesis process.
The seed function may differ at a low content. We per-
formed elemental composition analysis, and the elemental
data of the TS-1 samples synthesized with different amounts
of seeds are included in Table 1. The comparison of the ele-
mental compositions in the bulk and that on the surface can
tell us whether the silicalite-1 in the suspension dissolves first
in the synthesis gel or performs as a nucleus directly. If n(Si/Ti)
on the surface increases with the increase of seed amount, it
can be inferred that the silicalite-1 dissolves first and recrystal-
The catalytic performances of propene epoxidation over TS-
1 samples crystallized for different times are presented in
Table 3. TS-1-6-6 and TS-1-6-12 show very poor catalytic activi-
ties because the generation of the framework had just begun
after 6 and 12 h. X(H2O2) increased for samples that had
a longer crystallization time, whereas S(PO) decreased slightly.
Table 3. Catalytic performances of propene epoxidation over the
TS-1 samples crystallized for different times.[a]
Catalyst
X(H2O2) [%][b]
S(PO) [%]
U(H2O2) [%]
TS-1-6-6
0
1.2
null
97.9
97.9
97.6
97.6
96.9
96.2
96.1
null
70.0
77.8
79.7
82.0
83.2
85.2
85.3
TS-1-6-12
TS-1-6-15
TS-1-6-18
TS-1-6-24
TS-1-6-36
TS-1-6-48
TS-1-6-72
66.2
69.4
71.6
72.5
73.1
74.5
[a] Reaction conditions: catalyst 0.2 g, 408C, 1 h, propene 0.4 MPa, H2O2
concentration 1.4 molLÀ1. [b] X(H2O2), S(PO), and U(H2O2) are the conver-
sion of H2O2, selectivity of PO, and utilization of H2O2, respectively.
ChemCatChem 2015, 7, 2660 – 2668
www.chemcatchem.org
2665
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
It is credible that the function
of seed depends on its content/
concentration. There is a balance
between dissolution and recrys-
tallization.[56] If the concentration
is low, the dissolution of the
solid is inhibited, and thus it per-
forms as a nucleus to let the Si
and Ti sources grow on it. If the
concentration increases, the dis-
solution process gains impor-
tance. Therefore, more secon-
dary structural units are generat-
ed and more Si and Ti sources
can grow on them. The crystal
size of synthesized TS-1 is relat-
ed to the average size of the
nuclei. More secondary structural
units mean a smaller average
size of nuclei. Therefore, the
crystal size of TS-1 decreases
with the increase of seed
Figure 10. SEM images of the sample prepared with 9 wt% of seeds treated for different times.
amount. However, there is a minimum size even if more seeds
are added. The size decreased to 200 nm and was constant
with an increased seed amount above 9 wt%. This is because
dissolution becomes the rate-limiting step in crystal growth.
Some researchers suggested that the relative ratio of the
nuclei surface to the Si source controlled the size of TS-1.[57,58]
If dissolution becomes the rate-limiting step, the nuclei surface
will not increase with the increase of seed amount. Therefore,
the size of TS-1 crystals will reach the minimum.
Conclusions
Figure 11. XRD patterns of the sample prepared with 9 wt% of seeds treat-
ed for different times.
A facile size-controllable synthesis of titanium silicalite-1 (TS-1)
by tuning the content of seed and crystallization time was ach-
ieved. The size can be adjusted from 200 to 1200 nm, and
nano-sized TS-1 was obtained in the tetrapropylammonium
bromide hydrothermal system. The content of seed plays
a more important role in size control than the crystallization
time.
lizes with the Si and Ti sources. If n(Si/Ti) on the surface re-
mains constant or even decreases with the increase of seed
amount, it can be inferred that the silicalite-1 performs mainly
as a nucleus directly. The data given in Table 1 indicate that
n(Si/Ti) decreased first and then increased with an increase of
the seed amount. Therefore, we can infer that the solid per-
formed mainly as a nucleus directly if its content was low. SEM
images of the samples that contained 0.5 wt% seed treated for
different times confirm this result (Figure 12). During the 8 h
treatment, the silicalite-1 always retained its morphology. Only
a small amount of aggregation occurred after treatment for
6 h. With the increase of seed amount, more solid tended to
dissolve to secondary structural units. The amount of 1 wt%
seems to be the boundary at which the seed function transi-
tions from the function to serve as a nucleus and through dis-
solution. If the amount is higher than 1 wt%, more solid dis-
solved than performed as a nucleus directly. Therefore, the sili-
calite-1 was dissolved at a content of 9 wt%.
Samples with a small crystal size and large external surface
area show a higher catalytic activity and selectivity of propene
oxide in the propene epoxidation. However, the lower content
of framework Ti, caused by the accelerated crystallization,
shows a negative effect on the conversion of H2O2. The control
of the TS-1 crystal size by tuning the crystallization time is lim-
ited. A reduction of the crystallization time can decrease the
crystal size slightly, but also lead to a lower conversion of H2O2.
The mechanism of the seed function is related to its content.
The seed performs as a nucleus directly at a low content. The
dissolution of the seed to secondary structural units occurred
in each experiment, but a high seed content leads to more dis-
solution. Therefore, the seed needed to be dissolved first at
a high content to facilitate crystal growth.
ChemCatChem 2015, 7, 2660 – 2668
www.chemcatchem.org
2666
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
4000 cmÀ1, and the KBr pellet tech-
nique was used. UV/vis spectra
were obtained by using a Jasco
UV-550 spectrometer from 190–
500 nm, and pure BaSO4 was used
as a reference. Argon physisorp-
tion measurements were per-
formed at liquid argon tempera-
ture by using a Quantachrome Au-
tosorb iQ2 physical sorption appa-
ratus. The total surface area and
pore volume were calculated ac-
cording to the BET and t-plot
methods, respectively. The appear-
ances of the crystals were deter-
mined by using a Hitachi S-4800
scanning electron microscope. X-
ray photoelectron spectra (XPS)
were acquired by using a Thermo
VG ESCALAB250 instrument using
AlKa radiation operated at a con-
stant power of 260 W. The ob-
tained spectra were analyzed by
Figure 12. SEM images of the sample prepared with 0.5 wt% of seeds treated for different times.
using
a
curve-fitting program
XPSPEAK, which was helpful to
find the best fitting results. A PerkinElmer OPTIMA 2000DV optical
emission spectrometer was used to provide the bulk element com-
position.
Experimental Section
Synthesis of TS-1
TS-1 with different crystal sizes was synthesized in a TPABr/ethyla-
mine (EA) hydrothermal system using colloidal silica and titanium
tetrachloride as Si and Ti sources, respectively. TPABr was the tem-
plate, and ethylamine was the base. The molar composition of the
gel was: n(SiO2)/n(TiO2)/n(TPABr)/n(EA)/n(H2O)=1:0.025:0.15:1.0:16.
A suspension that contained nano-sized silicalite-1 was adopted as
seed. The amount of seed was calculated by the weight ratio of
the solid in the suspension to the SiO2 in the colloidal silica, and
was varied from 0 to 0.12 to adjust the crystal size of TS-1. For ex-
ample, the addition of 6 wt% seed means the weight ratio of the
solid in the suspension to the SiO2 in the colloidal silica was 0.06.
The suspension was prepared by mixing tetraethyl orthosilicate,
TPAOH, and water in a glass flask equipped with a condenser and
heated at 808C for 72 h. The molar ratio of the mixture was:
n(SiO2)/n(TPAOH)/n(H2O)=1:0.3:20. The obtained light milky sus-
pension was used as seeds directly without any other treatment.
Epoxidation of propene
The epoxidation of propene was performed in a 400 mL stainless-
steel reactor. First, H2O2/methanol solution (1.4 molLÀ1, 34 mL) and
TS-1 powder (0.2 g) were fed into the reactor. Propene was then
charged to the reactor to reach 0.4 MPa. The reaction was finished
after heating the mixture at 408C for 1 h. The residual H2O2 was
checked by iodometric titration. The products were analyzed by
using a Tianmei 7890F gas chromatograph equipped with a flame
ionization detector (FID) and a PEG-20m capillary column (30 m ”
0.25 m ” 0.4 mm). PO is the target product, and propene glycol
(PG) and its monomethyl ethers (MME) are byproducts. X(H2O2),
S(PO), and U(H2O2) were calculated as follows [Eqs. (2)–(4)]:
XðH2O2Þ ¼ 1ÀnðH2O2Þ=n0ðH2O2Þ
ð2Þ
ð3Þ
ð4Þ
The synthesis procedure was as follows: titanium tetrachloride was
added dropwise into isopropyl alcohol, and the obtained solution
was added to colloidal silica. Finally, TPABr, EA, seed, and water
were added to the colloidal silica gradually. The mixture was stirred
for 30 min, then transferred to a Teflon-lined autoclave and crystal-
lized at 1708C for 6–72 h. The obtained suspension was centri-
fuged, and the solid was washed by centrifugation twice, dried at
1008C overnight, and calcined at 5408C for 6 h to remove the tem-
plate. The synthesized samples were denoted as TS-1-a-t, in which
a and t stand for the content of seed and crystallization time, re-
spectively. For example, TS-1-6-72 was prepared by the addition of
6 wt% seed and crystallized for 72 h.
SðPOÞ ¼ nðPOÞ=ðnðPOÞþnðMMEÞþnðPGÞÞ
UðH2O2Þ ¼ ðnðPOÞþnðMMEÞþnðPGÞÞ=ðn0ðH2O2ÞXðH2O2ÞÞ
in which n0(H2O2) and n(H2O2) are the initial and final molar con-
tents of H2O2, respectively, and n(PO), n(MME), and n(PG) represent
the molar contents of PO, MME, and PG, respectively.
Treatment of the seed
Synthetic sources were added in the same way as with the normal
synthesis procedure, but without the addition of colloidal silica
and titanium tetrachloride (cf. Synthesis of TS-1). Amounts of 9 or
0.5 wt% of seed were used to observe the changes of the seed in
the simulated crystallization conditions. The mixture was heated at
1708C for 1–8 h in a Teflon-lined autoclave. The obtained suspen-
sion was separated, and the solid was dried and calcined at 5408C
for 6 h. The remaining solid was treated silicalite-1, which was char-
acterized by XRD and SEM.
Characterization of TS-1
XRD patterns were recorded by using a Rigaku Corporation Smar-
tLab 9 X-ray diffractometer using CuKa radiation. FTIR spectra were
collected by using a Bruker EQUINOX55 spectrometer from 400–
ChemCatChem 2015, 7, 2660 – 2668
www.chemcatchem.org
2667
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Papers
[29] H. J. Wei, T. Zhao, Y. Q. Liu, G. H. Li, X. Y. Wang, B. J. Li, RSC Adv. 2013, 3,
20811–20815.
Acknowledgements
[30] Y. Zuo, M. Liu, T. Zhang, L. Hong, X. Guo, C. Song, Y. Chen, P. Zhu, C.
Jaye, D. Fischer, RSC Adv. 2015, 5, 17897–17904. See Supporting Infor-
mation.
This work was financially supported by the Fundamental Re-
search
Funds
for
the
Central
Universities
[31] Y. Zuo, X. S. Wang, X. W. Guo, Ind. Eng. Chem. Res. 2011, 50, 8485–8491.
[32] Y. Zuo, X. S. Wang, X. W. Guo, Microporous Mesoporous Mater. 2012, 162,
105–114.
[33] M. M. J. Treacy, J. B. Higgins, Collection of Simulated XRD Powder Patterns
for Zeolites, 5th ed., Elsevier, Amsterdam, 2007, pp. 276.
[34] G. N. Vayssilov, Catal. Rev. Sci. Eng. 1997, 39, 209–251.
[35] G. Tozzola, M. A. Mantegazza, G. Ranghino, G. Petrini, S. Bordiga, G. Ric-
chiardi, C. Lamberti, A. Zulian, J. Catal. 1998, 179, 64–71.
[36] R. M. Boccuti, K. M. Rao, A. Zecchina, Stud. Surf. Sci. Catal. 1989, 48,
133–144.
(2342013DUT13RC(3)704) and China Postdoctoral Science Foun-
dation (2014M551094). The authors also thank Dr. Michael Janik
for his thoughtful discussion.
Keywords: crystal growth
·
heterogeneous catalysis
·
hydrothermal synthesis · silicon · titanium
[1] M. Taramasso, G. Perego, B. Notari, US4410501, 1983.
[37] M. A. Uguina, D. P. Serrano, G. Ovejoer, R. van Grieken, Appl. Catal. A
1995, 124, 391–408.
[2] D. R. C. Huybrechts, L. De Bruycker, P. A. Jacobs, Nature 1990, 345, 240–
242.
[38] S. Bordiga, A. Damin, G. Berlier, F. Bonino, G. Ricchiardi, A. Zecchina, C.
Lamberti, Int. J. Mol. Sci. 2001, 2, 167–182.
[39] E. Jorda, A. Tuel, R. Teisser, J. Kervennal, Zeolites 1997, 19, 238–245.
[40] T. Blasco, M. A. Camblor, A. Corma, J. PØrez-Pariente, J. Am. Chem. Soc.
1993, 115, 11806–11813.
[3] M. Wu, H. L. Song, L. J. Chou, RSC Adv. 2013, 3, 23562–23570.
[4] Z. C. Shan, Z. D. Lu, L. Wang, C. Zhou, L. M. Ren, L. Zhang, X. J. Meng,
S. J. Ma, F. S. Xiao, ChemCatChem 2010, 2, 407–412.
[5] L. B. Yang, F. Xin, J. Z. Lin, Z. Zhuang, R. Sun, RSC Adv. 2014, 4, 27259–
27266.
[41] F. Geobaldo, S. Bordiga, A. Zecchina, E. Giamello, G. Leofanti, G. Petrini,
Catal. Lett. 1992, 16, 109–115.
[6] G. H. Gao, S. F. Cheng, Y. An, X. J. Si, X. L. Xu, Y. M. Liu, H. J. Zhang, P.
Wu, M. Y. He, ChemCatChem 2010, 2, 459–466.
[42] Y. Zuo, M. Liu, L. W. Hong, M. T. Wu, T. Zhang, M. T. Ma, C. S. Song, X. W.
Guo, Ind. Eng. Chem. Res. 2015, 54, 1513–1519.
[7] U. Romano, A. Esposito, F. Maspero, C. Neri, M. G. Clerici, Stud. Surf. Sci.
Catal. 1990, 55, 33–41.
[43] L. Chen, Y. M. Wang, M. Y. He, Mater. Res. Bull. 2011, 46, 698–701.
[44] W. B. Fan, R. G. Duan, T. Yokoi, P. Wu, Y. Kubota, T. Tatsumi, J. Am. Chem.
Soc. 2008, 130, 10150–10164.
[45] J. G. Wang, Y. L. Zhao, T. Yokoi, J. N. Kondo, T. Tatsumi, ChemCatChem
2014, 6, 2719–2726.
[46] Y. Chen, Y. L. Wu, Y. R. Zhang, L. Long, L. Tao, M. D. Yang, N. Tang, J. Mol.
Catal. A 2012, 352, 102–109.
[47] A. Thangaraj, S. Sivasanker, J. Chem. Soc. Chem. Commun. 1992, 123–
124.
[8] A. H. Tullo, P. L. Short, Chem. Eng. News 2006, 84, 22–23.
[9] F. Cavani, J. H. Teles, ChemSusChem 2009, 2, 508–534.
[10] H. Li, C. J. He, M. Lin, W. Wang, X. J. Wu, J. Z. Gao, CN201010511515.X,
2010.
[11] Y. Zuo, M. L. Wang, W. C. Song, X. S. Wang, X. W. Guo, Ind. Eng. Chem.
Res. 2012, 51, 10586–10594.
[12] R. O. Kirk, T. J. Dempsey, Kirk-Othmer Encyclopedia of Chemical Technolo-
gy, Vol. 19, Wiley, New York, 1982, pp. 246.
[13] H. P. Wulff, F. Wattimena, US4021454, 1977.
[48] A. J. H. P. van der Pol, J. H. C. van Hooff, Appl. Catal. A 1992, 92, 93–111.
[49] P. A. Jacobs, J. A. Martens, Synthesis of High-Silica Aluminosilicate Zeo-
lites, Elsevier, Amsterdam, 1987.
[50] R. D. Edelman, D. V. Kudalkar, T. Ong, J. Warzywoda, R. W. Thompson,
Zeolites 1989, 9, 496–502.
[14] T. Iwasaki, M. Isaka, H. Nakamura, M. Yasuda, S. Watano, Microporous
Mesoporous Mater. 2012, 150, 1–6.
[15] M. Padovan, G. Leofanti, P. Roffia, EP311983, 1989.
[16] X. J. Deng, Y. N. Wang, L. Shen, H. L. Wu, Y. M. Liu, M. Y. He, Ind. Eng.
Chem. Res. 2013, 52, 1190–1196.
[51] T. Inui, N. Morinaga, T. Ishihara, T. Kanie, Y. Takegami, J. Catal. 1983, 79,
176–184.
[52] E. Narita, K. Sato, T. Okabe, Chem. Lett. 1984, 1055–1058.
[53] B. M. Zheng, Y. F. Wan, W. Y. Yang, F. X. Ling, H. Xie, X. C. Fang, H. C. Guo,
Chin. J. Catal. 2014, 35, 1800–1810.
[17] D. G. Huang, X. Zhang, T. W. Liu, C. Huang, B. H. Chen, C. W. Luo, E.
Ruckenstein, Z. S. Chao, Ind. Eng. Chem. Res. 2013, 52, 3762–3772.
[18] U. Müller, W. Steck, Stud. Surf. Sci. Catal. 1994, 84, 203–210.
˘
[19] A. Tuel, Y. Ben Taarit, C. Naccache, Zeolites 1993, 13, 454–461.
[20] G. Li, X. W. Guo, X. S. Wang, Q. Zhao, X. H. Bao, X. W. Han, L. W. Lin,
Appl. Catal. A 1999, 185, 11–18.
[54] W. Koetsier, EP81305729.6, 1981.
[55] J. Warzywoda, R. D. Edelman, R. W. Thompson, Zeolites 1991, 11, 318–
324.
[21] A. Silvestre-Albero, A. Grau-Atienza, E. Serrano, J. García-Martínez, J. Sil-
vestre-Albero, Catal. Commun. 2014, 44, 35–39.
[56] Y. Zuo, W. C. Song, C. Y. Dai, Y. P. He, M. L. Wang, X. S. Wang, X. W. Guo,
Appl. Catal. A 2013, 453, 272–279.
[22] Y. Xue, Y. L. Xie, H. J. Wei, Y. Q. Wen, X. Y. Wang, B. J. Li, New J. Chem.
2014, 38, 4229–4234.
[57] G. Majano, L. Delmotte, V. Valtchev, S. Mintova, Chem. Mater. 2009, 21,
4184–4191.
[23] X. X. Wu, Y. Q. Wang, T. Zhang, S. H. Wang, P. X. Yao, W. P. Feng, Y. Lin, J.
Xu, Catal. Commun. 2014, 50, 59–62.
[58] Y. Kamimura, W. Chaikittisilp, K. Itabashi, A. Shimojima, T. Okubo, Chem.
Asian J. 2010, 5, 2182–2191.
[24] Y. Wang, S. J. Zhang, Y. X. Zhao, M. Lin, J. Mol. Catal. A 2014, 385, 1–6.
[25] D. P. Serrano, R. Sanz, P. Pizarro, I. Moreno, S. Shami, Microporous Meso-
porous Mater. 2014, 189, 71–82.
[26] M. J. Wu, X. H. Liu, Y. Q. Wang, Y. L. Guo, Y. Guo, G. Z. Lu, J. Mater. Sci.
2014, 49, 4341–4348.
[27] Q. Lv, G. Li, H. Y. Sun, Fuel 2014, 130, 70–75.
Received: April 22, 2015
[28] Y. H. Zhang, X. S. Wang, X. W. Guo, Chin. J. Fuel Chem. Technol. 2000, 28,
550–554.
Published online on July 27, 2015
ChemCatChem 2015, 7, 2660 – 2668
www.chemcatchem.org
2668
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim