Epoxidation of propylene over titanosilicate-1 in fixed-bed reactor: Experiments and kinetics

Article abstract of DOI:10.14233/ajchem.2014.15519

The epoxidation of propylene with hydrogen peroxide over titanosilicate-1 is studied in a fixed-bed reactor with methanol/water as the solvent. The effects of methanol concentration (40-70 wt %), hydrogen peroxide concentration (5-20 wt %), pressure (1.8-3 MPa) and temperature (35-50 °C) on the reaction are investigated. The results show that the reaction rate decreases with increasing reaction pressure, which is opposite to that reported in the literature. The activation energy of the reaction is 57.4 kJ mol^-1 and the orders with respects to hydrogen peroxide and propylene are 0.71 and 0.29, respectively. The intrinsic kinetic modeling results show that an Eley-Rideal (H₂O₂ adsorbed) kinetic model incorporating the effects of hydrogen peroxide, propylene, methanol and propylene oxide satisfactorily agrees with the experimental results.

Full text of DOI:10.14233/ajchem.2014.15519

Asian Journal of Chemistry; Vol. 26, No. 4 (2014), 943-950
ASIAN JOURNAL OF CHEMISTRY
http://dx.doi.org/10.14233/ajchem.2014.15519
Epoxidation of Propylene Over Titanosilicate-1 in Fixed-bed Reactor: Experiments and Kinetics
1,2
1,*
1
1
1
LINA WANG , YAQUAN WANG , GUOQIANG WU , WENPING FENG , TENG ZHANG ,
1
1
1
1
RUMIN YANG , XING JIN , HAINAN SHI and SHUHAI WANG
1Key Laboratory for Green Chemical Technology of the Ministry of Education, School of Chemical Engineering & Technology, Tianjin
University, Tianjin 300072, P.R. China
2Xi'an University of Science and Technology, Xi'an 710054, P.R. China
*Corresponding author: E-mail: wlnkiddy@163.com
Received: 21 March 2013;
Accepted: 15 January 2013;
Published online: 15 February 2014;
AJC-14675
The epoxidation of propylene with hydrogen peroxide over titanosilicate-1 is studied in a fixed-bed reactor with methanol/water as the
solvent. The effects of methanol concentration (40-70 wt %), hydrogen peroxide concentration (5-20 wt %), pressure (1.8-3 MPa) and
temperature (35-50 °C) on the reaction are investigated. The results show that the reaction rate decreases with increasing reaction pressure,
which is opposite to that reported in the literature. The activation energy of the reaction is 57.4 kJ mol-1 and the orders with respects to
hydrogen peroxide and propylene are 0.71 and 0.29, respectively. The intrinsic kinetic modeling results show that an Eley-Rideal (H2O2
adsorbed) kinetic model incorporating the effects of hydrogen peroxide, propylene, methanol and propylene oxide satisfactorily agrees
with the experimental results.
Keywords: Titanosilicate-1, Epoxidation, Fixed-bed reactor, Kinetic model.
(TS-1), Ti-MCM-41 and Ti-MCM-48 supported gold catalysts.
Recently, Nguyen et al.5 studied the photocatalytic epoxidation
of propylene by molecular oxygen. However, the gas-phase
epoxidation of propylene catalyzed by the currently available
catalysts exhibited extremely low propylene oxide yield and
selectivity, a considerable gap still existed between industrial
applications and the fundamental academic research.
INTRODUCTION
Propylene oxide, which is an essential chemical material
used in the synthesis of polyether polyols, propylene glycols
and propylene glycol ethers1,2, has become increasingly
important in the chemical industry since 1950s3,4. The global
consumption of propylene oxide increased from approximately
3.9 million tons in 1991 to 6 million tons in 2009, which is
predicted to grow over 7 % annually from 2009 to 2014 owing
to the increasing demand5.
Compared to the propylene oxide production processes
mentioned above, the process of epoxidizing propylene with
hydrogen peroxide to propylene oxide (HPPO) catalyzed by
TS-1 is advantageous in the aspects of co-products and environ-
mental friendliness27,28. Figs. 1 and 2 depict the epoxidation
of propylene with hydrogen peroxide over TS-1.
Propylene oxide is currently produced by two different
commercial processes, i.e., the chlorohydrin process and the
hydroperoxidation process1. However, both of these processes
suffer from inevitable disadvantages. The former one is not
environmentally friendly due to the utilization of environ-
mentally hazardous chlorine and the production of chlorinated
organic byproducts as well as calcium chloride. The latter
process produces equimolar amounts of co-products, styrene
or tert-butanol and the viability of this process depends on the
existence of a market for such co-products. In recent years,
seeking novel approaches instead of the conventional ones to
produce propylene oxide has attracted worldwide attention
and many studies concerning the gas-phase propylene
epoxidation using molecular oxygen or the mixture of oxygen
and hydrogen have been reported6-20. For example, Haruta
et al.21-26 studied the gas-phase propylene epoxidation by the
mixture of oxygen and hydrogen over TiO2, titanosilicate-1
CH3CH
OH
CH2
OH
H+/OH-
CH3CH
CH2
H2O
+
O
Fig. 1. Main reaction
CH3CH
H+/OH-
CH2
OCH3
CH3CH
CH2
CH3CH
CH2
+
CH3OH
O
OH
OCH3 OH
Fig. 2. Side reaction
Until now, the industrialized HPPO processes of Dow/
BASF and Evonik/Uhde are the most mature, which have
been commercialized in Belgium, Korea and Thailand28-31. A
higher H2O2 conversion and propylene oxide selectivity can

944 Wang et al.
Asian J. Chem.
be obtained over TS-1 or modified TS-1 using CH3OH/H2O
Silica sol was purchased from Sigma-Aldrich. Tetrapropyl-
ammonium hydroxide was prepared through ion-exchange.
Synthesis of titanosilicate-1:Titanosilicate-1 was prepared
according to the method described in the literature45,46. A
solution of 2.2 g butyl titanate, 23.7 g dried isopropyl alcohol
and 94.5 g tetraethyl orthosilicate were added into a solution
of 15 wt % tetrapropyl-ammonium hydroxide (TPAOH) in
water with vigorous magnetic stirring under N2 atmosphere.
The resulting mixture was continuously stirred for 2 h to evapo-
rate ethanol formed. During this time, distilled water was added
to compensate for the weight loss. The molar ratio of SiO2/
TiO2 was 70 in the final solution. Then the mixture was
transferred to a 300 mL PTFE lined stainless steel autoclave
and heated in an oven at 175 °C, under autogenous pressure
without stirring, for 72 h. After cooling to the room tempera-
ture, the crystalline product was separated from the liquid by
filtration, washed with water to pH = 7, dried for 6 h at 120 °C
and finally calcined for 6 h at 550 °C in air.
as the solvent32-35
.
Previous studies36-38 have reported the effects of solvents
such as methanol/water and isopropanol and other operating
conditions on the epoxidation of propylene with hydrogen
peroxide. Especially, studies on the kinetics of propylene
epoxidation have also been reported with slurry reactors39,40
.
Liang et al.39 developed a kinetic model derived from the Eley-
Rideal mechanism with H2O2 adsorbed on the active sites and
free C3H6. Shin et al.40 proposed a kinetic model following the
Langmuir-Hinshelwood mechanism with C3H6 and H2O2 both
adsorbed on the active sites. However, slurry reactors, either
continuous or batch, require the separation of catalysts, sugges-
ting extra cost and operation time with the abrasion of the
catalyst. Moreover, the severe back mixing in the continuous
operation and the high ratios of liquid to catalysts will lead to
possible side reactions, resulting in the decrease of propylene
oxide selectivity. Due to these disadvantages described above,
slurry reactors are not suitable for the reaction from both the
industrial and the economic points of view.
Synthesis of TS-1/SiO2: TS-1/SiO2 was prepared accor-
ding to the procedure described in the literature45. The mixture
of 1.98 g silica sol, 4 g TS-1 and 0.16 g tianqing powder was
grinded for 40 min and then extruded into φ1 mm strips with
an extrudating machine. Titanosilicate-1/SiO2 was dried for
6 h at 120 °C and finally calcined for 6 h at 550 °C in air. The
strip TS-1/SiO2 was cut into cylindrical pieces of φ1 mm ×
1 mm for the use in the fixed-bed reactor. In the testing to
remove the internal diffusion effects, the strips were grounded
into 20-40, 40-60 and 60-80 mesh.
Based on the requirements and characteristics of the HPPO
process and the engineering considerations, fixed-bed reactors
are currently regarded more suitable. However, in contrast to
the massive research on slurry reactors reported, investigations
on the application of fixed-bed reactors have seldom been
accessed. Pan et al.41 studied the effects of operating conditions
on the propylene epoxidation over TS-1 in a fixed-bed reactor
and high H2O2 conversion (> 94 %) and propylene oxide
selectivity (> 90 %) were obtained under optional conditions.
Li et al.42,43 reported that the supported TS-1 used in a fixed-
bed reactor exhibited a long catalytic life (200 h) and the studies
concerning the effects of solvents and sodium ions on the
epoxidation of propylene in the fixed-bed reactor were succe-
ssively reported by them. Recently, the epoxidation of propylene
over modified TS-1 zeolite catalysts in a fixed-bed reactor was
studied by Dong and Guo44 and the catalytic life of the modified
TS-1 was more than 1000 h with H2O2 conversion and propylene
oxide selectivity being above 95 and 95 %, respectively.
Nevertheless, modeling of hydrogen peroxide to propylene
oxide over TS-1 in the fixed-bed reactor has not been explored
in detail hitherto. Therefore, accurate understanding of the
reaction kinetics is imperative in order to establish a realistic
model of propylene oxide reactor for industry applications.
Thereby motivated, the present work aims to systema-
tically examine the propylene epoxidation catalyzed by TS-1
and develop a kinetic model for a fixed-bed reactor. Continuous
reactions were carried out under high-pressure and the effects
of temperature, pressure, methanol concentration and hydrogen
peroxide concentration on the propylene epoxidation were
studied. Four kinetic models were proposed to describe the
formation of propylene oxide to correlate with the experimental
results to elucidate the reaction mechanism.
Experimental apparatus and procedure: The schematic
diagram of the experimental apparatus for the epoxidation of
propylene is illustrated in Fig. 3.
8.82 g Titanosilicate-1 (15 mL) diluted with 8.55 mL
ceramic particles was added into a stainless steel tube reactor
(L = 1200 mm, d = 6 mm) and a little bit of stainless steel wire
was put into the entrance and exit of the reactor respectively
to avoid the dropping of catalysts. A stainless steel tube of
2 mm diameter with a thermocouple was embedded in the
reactor to determine the temperature. The reaction temperature
was controlled by a thermostatic circulating water bath with
the constant flow rate of 6 L min-1 and the temperature
difference along the reactor was less than 0.5 °C. The high
pressures were realized with N2 and the propylene was always
in liquid and excess in the experiments (the molar ratios of
C3H6/H2O2 were from 1.65 to 4). After the pressure reached
the needed one, the flow rate of N2 was kept at 40 mL min-1.
When the pressure and temperature were stabilized at the
required ones, C3H6 (liquid) was introduced into the reactor at
0.244 mL min-1 by a high-pressure micro-pump, the mixed
solutions of H2O2/CH3OH/H2O with different flow rates were
introduced into the reactor with another pump to react with
C3H6 (30 wt % H2O2 in water was used to make up the mixture
of H2O2/CH3OH/H2O in all the experiments except the one
with 70 wt % CH3OH, which was made up by 50 wt % H2O2
in water. The concentration of H2O2 was expressed in wt %
H2O2 in CH3OH + H2O + H2O2. The concentration of H2O2
was expressed in wt % H2O2 in CH3OH + H2O + H2O2. After
reaction, the products entered the condenser with alcohols
which was cooled to 2 °C. At the bottom of the apparatus, a
liquid-receiver was used to collect the samples for the analysis.
EXPERIMENTAL
Methanol (AR), isopropyl alcohol (AR), 30 wt and 50 wt
% hydrogen peroxide (AR), tetraethyl orthosilicate (AR) and
butyl titanate (AR) were purchased from Tianjin Guang fu
Fine Chemical Research Institute Co., Ltd., China. Propylene
(99.99 %) was purchased from Tianjin Division of SINOPEC.

Vol. 26, No. 4 (2014)
Epoxidation of Propylene Over Titanosilicate-1 in Fixed-bed Reactor: Experiments and Kinetics 945
In this work, the reaction rates (r) were calculated based
on the rtotal per gram catalyst, i.e.,
T
N2
r = rtotal/mcat
(6)
In repeating the experiments, it was found that both the
reaction rates and the propylene oxide selectivity exhibited
the errors within 5 %.
Water
Thermocouple
Reactor
C3H6(L)
RESULTS AND DISCUSSION
Catalyst bed
Effects of operating conditions: To avoid the formation
of hot spots, the reaction has been conducted in a down-flow
mode for the convenience of removing the reaction heat during
reactions. In addition, the reactor was heated with circulating
water, which at the same time also removes the reaction heat47.
The temperature difference between the entry and the exit of
the coolant was maintained within 0.5 °C to keep the reaction
temperature fluctuation within a smaller range. Furthermore,
small TS-1 particles (φ1 mm × 1 mm) was used to remove the
mass and heat transfer gradients43. Hanika and Hajkova48
proved that reactions with the particles of this dimension
exhibited no temperature peak. And also, the catalysts were
diluted with ceramic particles, which could improve the mass
and heat transfer to eliminate hot spots. Therefore, it is believed
that there should be no hot spots in the reaction.
Water
N2
H2O2/CH3OH/H2O
Coolant out
Pressure gauge
Condenser
Coolant in
P
Receiver
Sampling
Fig. 3. Schematic diagram of experimental apparatus
In the preliminary experiments, we found that at the liquid
flow rate of the mixture of H2O2/CH3OH/H2O ≥ 0.29 mL/min,
the external diffusion effect disappeared and at the catalyst
particle sizes ≤ φ1 mm × 1 mm, the internal diffusion effect
could be neglected49. So in the kinetic experiments, the catalyst
particle sizes of φ1 mm × 1 mm were used and the liquid flow
rates were all above 0.29 mL/min. In addition, in the testing
to the decomposition of hydrogen peroxide over TS-1 at 35 to
55 °C, it was found that the decomposition of hydrogen per-
oxide at 35 to 50 °C was less than 6 % and could be neglected.
Effect of methanol concentration: It is known that
methanol is the best among many kinds of protic or aprotic
solvents for the epoxidation of propylene27. Therefore, metha-
nol/water mixtures were used as the solvent and the effect of
the methanol concentration on the reaction was studied.
As shown in Fig. 4, the reaction rate and propylene oxide
selectivity increased with the rising methanol concentration,
but did not change appreciably when the methanol concen-
tration increased from 65 to 70 %.
The reaction was carried out at 5-10 wt % H2O2, 40-70 wt %
CH3OH, T = 35-50 °C, P = 1.8-3.0 MPa (N2). During the
reaction, the liquid samples were collected at given intervals
(0.5 h).
Product analysis:The products of the reaction were analyzed
with a HP 4890 gas chromatography (GC) equipped with a
flame ionization detector (FID) and an OV-1701 (30 m × 0.25
mm × 0.33 µm) capillary column with nitrogen as the carrier
gas. The unreacted H2O2 was measured by iodometric titration.
The conversion of H2O2 and propylene oxide selectivity were
calculated based on the starting amount of H2O2 according to
the following two equations, respectively.
0
nH2
– nH2
0
O
O
2
2
XH2
=
×100 %
×100 %
O
(1)
(2)
2
nH2
nPO
O
2
SPO
=
nPO + nMME + nPG
XH2O and SPO denote the conversion of H2O2 and the selec-
2
On one hand, the increase of the reaction rate with increa-
sing the methanol concentration could be attributed to the
increase in the solubility of propylene in H2O2/CH3OH/H2O.
On the other hand, the results could be also explained by the
5-membered ring mechanism of epoxidation proposed in the
literature, in which the increase of the methanol concentration
improved the formation rate of the 5-membered ring active
adducts27,50-52. When the methanol concentration exceeded 65
wt %, the reaction rate did not change appreciably. The water
concentration decreased with the rising methanol concen-
tration, as a result, the side reactions were reduced and propylene
oxide selectivity increased. Therefore, 65 wt % was the suitable
concentration for the reaction.
tivity of propylene oxide, respectively. n0 and n represent the
initial and the final amount in moles, respectively. MME and
PG denote propylene glycol monomethyl ethers and propylene
glycol, respectively.
The total reaction rate can be calculated as follows:
– rtotaldV = Fn0,H dXH
(3)
O
O
2
2
2
2
which can be converted to
– rtotaldV = FVC0H dXH
(4)
O
O
2
2
2
2
From which the total reaction rate, rtotal, can be obtained
as follows:
Effect of hydrogen peroxide concentration: The effects
of hydrogen peroxide concentration on the reaction rate and
dXH2
O
– rtotal = C0H2
2
O
(5)
2
d(V / FV )

946 Wang et al.
Asian J. Chem.
0.008
H2O2 changed significantly; indicating that side reactions were
promoted at more than about 25 % H2O2. The order of the
H2O2 concentration would be obtained in the following part.
Effect of pressure: The effect of pressure on the reaction
is shown in Fig. 6. The reaction rate decreased with the rising
pressure, but propylene oxide selectivity increased slightly. In
addition, the reaction rate did not change appreciably after
the pressure increased from 2.5 MPa to 3 MPa.
98
96
94
92
90
88
0.29 mL min–1
0.40 mL min–1
0.58 mL min–1
0.70 mL min–1
0.006
0.004
0.002
0
0.0100
96
0.0075
92
0.29 mL min–1
0.40 mL min–1
40
50
60
70
0.58 mL min–1
0.0050
CH3OH concentration (%)
88
0.70 mL min–1
Fig. 4. Effect of methanol concentration on reaction rate and propylene
oxide selectivity. Reaction conditions: T= 40 °C, P= 2.5 MPa, 10.64
wt % H2O2 with H2O in balance
0.0025
84
propylene oxide selectivity are shown in Fig. 4. The results
indicated that the reaction rate increased with the rising
hydrogen peroxide concentration, but propylene oxide selec-
tivity decreased conversely.
0.0000
80
1.8
2.1
2.4
2.7
3.0
It has been reported that the reaction order with respects
to hydrogen peroxide ranged between zero and unity39,40
Pressure (MPa)
.
Fig. 6. Effect of pressure on reaction rate and propylene oxide selectivity.
Reaction conditions: T = 40 °C, 10.64 wt % H2O2, 65 wt % CH3OH,
24.36 wt % H2O
According to this conclusion, the increase of the reaction rate
with the increasing hydrogen peroxide concentration shown
in Fig. 5 was reasonable.
Pan et al.41 reported that the reaction rate increased with
the increasing pressure in the fixed-bed reactor. However, the
results obtained in this work differed from those reported
markedly. The effect of pressure on the solution in liquid phase
100
90
80
70
60
0.016
0.012
0.008
0.004
0.000
≠
≠
can be expressed as (∂lnk/∂p)T = (-DVm )/RT53. ∆Vm is the
0.29 mL min–1
0.40 mL min–1
0.58 mL min–1
0.70 mL min–1
activation volume of the reaction, which is equal to the partial
≠
molar volume of the activated complex (Vm ) minus the sum
of the partial molar volume of the reactants (ΣVRi), that is,
∆Vm≠ =Vm≠ -ΣVRi. Therefore, if ∆Vm≠ > 0, the reaction rate will
decrease with the rising pressure. In the HPPO reaction, the
activated complex in the reaction is believed to a 5-membered
ring active adduct with a large molar volume51, which is larger
than the sum of the partial molar volume of H2O2 and C3H6.
Therefore, the results that the reaction rate decreased with the
rising pressure were reasonable.
Effect of temperature: The effect of the reaction tempe-
rature was studied by examining the performance of TS-1 in
the temperature range of 35-50 °C. Fig. 7 showed that the
reaction rate increased with the increasing temperature.
However, the propylene oxide selectivity decreased, which led
to the reduction of the propylene oxide production rate.
Especially, when the reaction temperature increased from 45
and 50 °C, the propylene oxide selectivity changed obviously,
indicating that more alkyl ether of propylene glycol was
produced in this temperature range.
5
10
15
20
H2O2 concentration (%)
Fig. 5. Effect of hydrogen peroxide concentration on reaction rate and
propylene oxide selectivity. Reaction conditions: T= 40 °C, P = 2.5
MPa, the H2O concentration is 7/3 times the corresponding H2O2
concentration with CH3OH in balance
The water concentration increased with the increasing
hydrogen peroxide concentration, which accelerated the side
reactions, as a result, the propylene oxide selectivity decreased.
It was noteworthy that the selectivity between 10 and 20 %
Reaction kinetics: Eventhough the commercial-scale
propylene oxide plant based on hydrogen peroxide has been
employed hitherto, the reaction mechanism is still controversial

Vol. 26, No. 4 (2014)
Epoxidation of Propylene Over Titanosilicate-1 in Fixed-bed Reactor: Experiments and Kinetics 947
0.015
96
dation of propylene. Therefore, the adsorptions of methanol
and hydrogen peroxide on TS-1 active sites should be taken
into consideration in deducing the kinetic models. Due to the
large concentration of methanol used in this experiment and
its strong adsorption on TS-1 active sites, it was considered
that the epoxidation occurred on TS-1 with adsorbed methanol
molecules. In addition, the adsorption of H2O molecules and
the diffusion were neglected for the hydrophobicity of TS-1
and the adsorption of propylene was studied for the competitive
adsorption with other reactants on Ti active sites39. The product
propylene oxide could inhibit the epoxidation for the
competitive adsorption with the reactants on Ti active sites.
Judging from it, the adsorption of propylene oxide might also
have its contribution to the kinetics.
94
92
90
88
86
0.29 mL min–1
0.40 mL min–1
0.58 mL min–1
0.70 mL min–1
0.010
0.005
0.000
Therefore, three steps including the adsorption of H2O2
and C3H6, surface reaction and desorption of propylene oxide
occurred in the reaction; however, the results in the section of
effect of methanol concentration and effect of hydrogen
peroxide concentration showed that the rate of epoxidation
was correlative to the concentration of reactants and increased
with the concentration of C3H6 or H2O2, so the controlling step
of the whole reaction was the surface reaction.
35
40
45
50
Temperature (°C)
Fig. 7. Effect of temperature on reaction rate and propylene oxide
selectivity. Reaction conditions: P = 2.5 MPa, 10.64 wt % H2O2, 65
wt % CH3OH, 24.36 wt % H2O
and the kinetics for fixed bed reactors has not yet been reported.
Thus, the kinetics was studied below to explore the reaction
in a fixed-bed reactor.
K
1
(1)
(2)
(3)
(4)
(5)
(6)
ꢀꢀꢀ⇀
↽ꢀꢀꢀ
S+ ROH
S·ROH
K
Reaction rate equation: The propylene epoxidation is a
consecutive reaction, yielding propylene oxide by the reaction
of propylene with H2O2. Therefore, the reaction rate equation
could be expressed as follows:
2
ꢀꢀꢀ⇀
S·ROH + H2O
S·ROH + C3H
S·ROH·H O
2 ↽ꢀꢀꢀ
2
2
K
3
ꢀꢀꢀ⇀
S·ROH·C H
6 ↽ꢀꢀꢀ
3
6
k
1
S·ROH·H2O2 + C3H6 →S·ROH·PO + H2O
Ea



α
β
r = A × Exp. –

× C
× CC3
H
6
K
4
H
O
2
2
ꢀꢀꢀ⇀
S·ROH + PO
↽ꢀꢀꢀ
RT
S·ROH·PO


k
2
Based on the experimental data of the reactions at 35 to
50 °C shown in Figs. 4-7, the rate equation was obtained by
the statistical software 1STOPT and the fitting results were
listed in Table-1.
S·ROH + C3H6 + H2O2 →S·ROH·PO+ H2O2
k
3
S·ROH·C3H6 + S·ROH·H2O2 →
(7)
S·ROH·PO + S·ROH + H2O
Proposed reaction steps
TABLE-1
ESTIMATED DATA OF α, β, A, Ea
A
Ea (kJ mol-1)
R
F
α
β
Four kinetic models of the reactions were established
based on the proposed mechanism above, which were in accor-
dance with the Eley-Rideal or Langmuir Hinshelwood
mechanisms54.
0.71
0.29 1.87 × 107
57.4
0.962
517.6
F, Statistics and probability; R, correlation coefficient
(1) Eley-Rideal mechanism, H2O2 adsorbed, C3H6 free,
surface reaction controlling.
Based on the estimation data listed in Table-1, the activa-
tion energy of the reaction was obtained as 57.4 kJ mol-1 and
the kinetic equation (35-50 °C) was expressed as follows:
k1K2CH CC
O
H
6
2
2
3
r2 =
5.74×104
1 + K1CCH
+ K2CH2 + K3CC3 + K4CPO
O2 H
6




OH
0.29
r =1.87×107 × Exp. –
×C 0.71 ×CH O
3 6
3

H O
2
2
RT
(2) Eley-Rideal mechanism, C3H6 adsorbed, H2O2 free,
surface reaction controlling.

It could be derived from the kinetic equation that the order
of the reaction was 0.71 with respect to hydrogen peroxide,
which was consistent with the result in Fig. 5.
k1K3CH CC
O
H
6
2
2
3
r1 =
1 + K1CCH
+ K2CH2 + K3CC3 + K4CPO
O2 H
6
OH
3
Kinetic Model
(3) L-H mechanism, C3H6 and H2O2 adsorbed on the single
active sites, surface reaction controlling.
Reaction mechanism: Owing to the formation of 5-mem-
bered rings mechanism described by Clerici et al.27, the 5-
membered ring intermediate among Ti active sites, H2O2 and
alcohol was formed with the decrease of the system energy,
which could account for the excellent activity of the epoxi-
k3K2K3CH CC
O
H
6
2
2
3
r3 =
2
(1 + K1CCH
+ K2CH2 + K3CC3 + K4CPO
)
OH
O
H
6
3
2

948 Wang et al.
Asian J. Chem.
(4) L-H mechanism, C3H6 and H2O2 adsorbed on different
active sites, surface reaction controlling.
concentrations. The existence of propylene oxide concentration
in the denominator of the model equation showed the
decreasing effect on the reaction. Therefore, propylene oxide
should be removed from the reaction system as soon as possible
in the practical process.
k4K2K3CH CC
O
H
6
2
2
3
r4 =
(
1 + K3CC )(1+ K1CCH
+ K2CH2 + K4CPO
)
H
OH
O
2
3
6
3
The reaction rate constant and equilibrium adsorption
constants of the four models at 40 °C were calculated and the
obtained values were summarized in Table-4.
The propylene concentration (CC3H ) in the models was a
6
function of the propylene solubility, which was related to the
reaction temperatures, pressures and mass ratios of methanol/
water. Therefore, according to the coefficient equation of Peng-
Robinson equation, the solubility of propylene at different
reaction temperatures, pressures and mass ratios of methanol/
water was obtained by the software ASPEN.
TABLE-4
PARAMETERS OF THE FOUR MODELS AT 40 °C
Model 1 Model 2 Model 3 Model 4
k(mol L-1g-1 cat min-1)
K2 (mol L-1)
K3 (mol L-1)
0.83
2.19
0.02
5.90
0.0001
2.22
0.012
0.64
Estimation of kinetic parameters: One reaction rate
constant and three adsorption equilibrium constants were
present in every kinetic model. The effect of temperature on
the rate constant and adsorption equilibrium constants was
assumed to obey with the Arrhenius equation and van't Hoff
equation, respectively. Based on the experimental data of the
propylene epoxidation in the fixed-bed reactor, the parameters
of the models were estimated (Table-2) by the statistical
software 1STOPT and the statistical results of the models were
listed in Table-3.
-
0.28
-
17.04
2.01
27.26
K4 (mol L-1)
0.15
0.00005
According to Model 1, the absorption equilibrium cons-
tants of H2O2, C3H6 and propylene oxide at 40 °C were K2 =
2.19, K3 = 17.04 and K4 = 2.01 respectively as showed in
Table-4. The values of the adsorption constants were consistent
with the polarities of these species.
As displayed in Table-3, the Eley-Rideal mechanism fitted
the experimental data well, especially with the assumption that
the epoxidation reaction took place between H2O2 adsorbed
on Ti active sites and propylene in free state (Model 1). Model
1 exhibiting the optimal fitting result was regarded as the
expected model which was consistent with the model obtained
by Liang et al.39 based on the results obtained in a slurry reactor.
Simulation of the reaction using Eley-Rideal and
Langmuir-Hinshelwood models: The reaction rates of epoxi-
dation of propylene were calculated according to the estimated
parameters of the four models and the calculated reaction rates
were plotted against the experimental ones as shown in Figs.
8 and 9. The results showed that the experimental data
satisfactorily fitted the Eley-Rideal mechanism, whereas they
exhibited poor agreements with the Langmuir-Hinshelwood
mechanisms.
TABLE-3
STATISTICAL RESULTS OF THE MODELS
Conclusion
Model 1
0.9642
558.6
Model 2
0.9373
304.1
Model 3
0.9450
350.2
Model 4
0.8970
172.8
R
F
Operating conditions such as methanol concentration,
pressure, hydrogen peroxide concentration and reaction tempe-
rature remarkably affected the reaction rate and propylene
oxide selectivity. The rates of propylene oxide production were
observed to increase with the rising methanol concentration,
reaction temperature and hydrogen peroxide concentration,
however, decrease with the rising reaction pressure. Propylene
oxide selectivity increased with the rising methanol concen-
tration and reaction pressure, but decreased with the increasing
hydrogen peroxide concentration and reaction temperature.
F, Statistics and probability; R, correlation coefficient
As shown in Model 1, although CH2O and CC3H6 existed in
2
both the numerator and denominator of the kinetic equation,
the concentrations in the denominator had less influence on
the reaction rate. Therefore, the reaction rate increased with
the rising reactant concentration. That was why the reaction
rate would increase with the increasing CH3OH and H2O2
TABLE-2
KINETIC PARAMETERS OF VARIOUS MODELS
Model 1
Model 2 Model 3
Model 4
K
4
4
4
4
4
4.41×10
3.29 ×10
4.17×10
3.14×10
5
6
5
5
4.17×1 0 ×Ex p ( -
)
1.80×10 ×Exp ( -
)
1.09×10 ×Exp（-
)
1.46×10 ×Exp (-
)
RT
RT
RT
RT
K2
4
4
3
6.53×10
2.18×10
3.04×10
-9.76×10
3
5
6
8.02×10 ×Exp (-
)
2.81×10 ×Exp (-
)
2.58×10 ×Exp (-
)
−5.75×Exp (
)
RT
RT
RT
RT
K3
K4
4
4
4
4
3.25×10
5.60×10
3.65×10
4×10
−6
2
5
-6
8.30×10 ×E x p (
)
2.34×10 ×Ex p (-
)
3.42×10 ×Exp (-
)
3.59×10 ×Ex p (
)
RT
RT
RT
RT
3
4
4
4
3.45 ×10
2.39×10
1.89×10
3.80×10
3
3
2
0.04 × E x p (
)
−1.35×10 ×Exp (-
)
2.82×10 ×Exp (-
)
1.16×10 ×Exp (-
)
RT
RT
RT
RT

Vol. 26, No. 4 (2014)
Epoxidation of Propylene Over Titanosilicate-1 in Fixed-bed Reactor: Experiments and Kinetics 949
0.008
Fn0,H = initial flow rate of moles of H2O2 (mol min-1)
O
2
2
FV = flow rate of volume (L min-1)
V = reactor volume (L)
ki = reaction rate constant of species i (mol L-1g-1 cat min-1)
Ki = adsorption equilibrium constant of species i (mol L-1)
α, β = orders of CH2O2 and CC3H6
A = preexponential factor (mol L-1g-1 cat min-1)
Ci = concentration of species i (mol kg-1 or mol L-1)
Ea = apparent activation energy of the reaction (kJ mol-1)
mcat = catalyst mass (g)
∆Vm≠ = the activation volume of the reaction
Vm≠ = the molar volume of the activated complex
ΣVRi = the sum of the molar volume of the reactant i
0.006
0.004
0.002
0.000
Eley-Rideal C3H6 adsorbed
Eley-Rideal H2O2 adsorbed
ACKNOWLEDGEMENTS
This work has been supported by Natural Science Foun-
dation of China with Grant No. 21276183.
0.000
0.002
0.004
0.006
0.008
rexp. (mol L–1 min–1 g–1 cat)
REFERENCES
Fig. 9. Experimental rates and calculated rates according with Eley-Rideal
models (rexp. vs. rcal.
)
1. T.A. Nijhuis, M. Makkee, J.A. Moulijn and B.M. Weckhuysen, Ind.
Eng. Chem. Res., 45, 3447 (2006).
0.008
2. C. Neri, B. Anfossi, A. Esposito and F. Buonomo, Process for the
Epoxidation of Olefinic Compounds. US Patent 4833260 (1989).
3. K. Weissermel, H.J. Arpe and C.R. Lindley, Industrial Organic Chemistry;
Wiley-VCH: Weinheim (2003).
4. D. Kahlich, U. Wiechern and J. Lindner, Propylene Oxide in Ullmann's
Encyclopedia of Industrial Chemistry; JohnWiley & Sons Inc: Hoboken
(2009).
0.006
0.004
0.002
0.000
5. V.H. Nguyen, H.Y. Chan, J.C.S. Wu and H. Bai, Chem. Eng. J., 179,
285 (2012).
6. T.A. Nijhuis, B.J. Huizinga, M. Makkee and J.A. Moulijn, Ind. Eng.
Chem. Res., 38, 884 (1999).
7. R. Meiers and W.F. Holderich, Catal. Lett., 59, 161 (1999).
8. G. Jenzer, T. Mallat, M. Maciejewski, F. Eigenmann and A. Baiker,
Appl. Catal. A, 208, 125 (2001).
9. A.K. Sinha, S. Seelan, S. Tsubota and M. Haruta, Top. Catal., 29, 95
(2004).
10. N. Yap, R.P. Andres and W.N. Delgass, J. Catal., 226, 156 (2004).
11. A.K. Sinha, S. Seelan, M. Okumura, T.Akita, S. Tsubota and M. Haruta,
J. Phys. Chem. B, 109, 3956 (2005).
L-H sole-site
L-H dual-site
12. J.L. Zhao, J.C. Zhou, J. Su, H.C. Guo, X.S. Wang and W.M. Gong,
AIChE J., 53, 3204 (2007).
0.000
0.002
0.004
0.006
0.008
13. Z.H. Suo, M.S. Jin, J.Q. Lu, Z.B. Wei and C. Li, J. Nat. Gas Chem., 17,
184 (2008).
rexp. (mol L–1 min–1 g–1 cat)
Fig. 10. Experimental rates and calculated rates according with L-H models
(rexp. vs. rcal.
14. T. Liu, P. Hacarlioglu, S.T. Oyama, M.F. Luo, X.R. Pan and J.Q. Lu, J.
Catal., 267, 202 (2009).
)
15. H.W. Yang, D.L. Tang, X.N. Lu and Y.Z. Yuan, J. Phys. Chem. C, 113,
8186 (2009).
16. G.W. Zhan, M.M. Du, D.H. Sun, J.L. Huang, X. Yang, Y. Ma, A.R.
Ibrahim and Q.B. Li, Ind. Eng. Chem. Res., 50, 9019 (2011).
17. J.H. Huang, E. Lima, T. Akita, A. Guzman, C.X. Qi, T. Takei and M.
Haruta, J. Catal., 278, 8 (2011).
The reaction rate equation was obtained with Ea = 57.4 kJ
mol-1 and the orders with respects to hydrogen peroxide and
propylene were 0.71 and 0.29, respectively. The Eley-Rideal
(H2O2 adsorbed) kinetic model satisfactorily predicted the
experimental results and could be utilized to describe the
reaction mechanism.
18. G.W. Zhan, M.M. Du, J.L. Huang and Q.B. Li, Catal. Commun., 12,
830 (2011).
19. J. Su, J.C. Zhou, C.Y. Liu, X.S. Wang and H.C. Guo, Chin. J. Catal.,
31, 1195 (2010).
20. M.M. Du, G.W. Zhan, X. Yang, H.X. Wang, W.S. Lin, Y. Zhou, J. Zhu,
L. Lin, J.L. Huang, D.H. Sun, L.S. Jia and Q.B. Li, J. Catal., 283, 192
(2011).
Nomenclature
Xi = conversion of species i (%)
ni0 = initial number of moles of species i (mol)
ni = number of moles of species i (mol)
Si = selectivity to product I (%)
21. B.S. Uphade, M. Okumura, S. Tsubota and M. Haruta, Appl. Catal. A,
190, 43 (2000).
22. A.K. Sinha, S. Seelan, T. Akita, S. Tsubota and M. Haruta, Appl. Catal.
A, 240, 243 (2003).
rtotal = total reaction rate (mol L-1 min-1),
r = reaction rate (mol L-1 g-1 cat min-1)
23. C.X. Qi, T. Akita, M. Okumura, K. Kuraoka and M. Haruta, Appl.
Catal. A, 253, 75 (2003).

950 Wang et al.
Asian J. Chem.
24. J. Huang, T. Takei, T. Akita, H. Ohashi and M. Haruta, Appl. Catal. B,
95, 430 (2010).
39. X.H. Liang, Z.T. Mi, Y.L. Wu, L. Wang and E.H. Xing, React. Kinet.
Catal. Lett., 80, 207 (2003).
25. C.X. Qi, J.H. Huang, S.Q. Bao, H.J. Su, T. Akita and M. Haruta, J.
Catal., 281, 12 (2011).
40. S.B. Shin and D. Chadwick, Ind. Eng. Chem. Res., 49, 8125 (2010).
41. S.C. Pan, Z. Tang, T.S. Tao and Y.Y. Feng, Adv. Fine Petrochem., 2, 38
(2001).
26. J. Huang, E. Lima, T. Akita, A. Guzmán, C. Qi, T. Takei and M. Haruta,
J. Catal., 278, 8 (2011).
42. G. Li, X.S. Wang, H.S. Yan, Y.Y. Chen and Q.S. Su, Appl. Catal. A,
218, 31 (2001).
27. M.G. Clerici, G. Bellussi and U. Romano, J. Catal., 129, 159 (1991).
28. F. Cavani and J.H. Teles, Chem. Sus. Chem., 2, 508 (2009).
29. G.F. Thiele and E. Roland, J. Mol. Catal. Chem., 117, 351 (1997).
30. J.H. Teles, A. Rehfinger, P. Bassler, A. Wenzel, N. Rieber and P. Rudolf,
Method for the production of propylene oxide. US Patent 6756503,
(2004).
43. G. Li, X. Wang, H.Yan,Y. Liu and X. Liu, Appl. Catal. A, 236, 1 (2002).
44. Y. Dong and X. Guo, Acta Petrol. Sin., 26, 677 (2010).
45. M. Taramasso, G. Perego and B. Notari, Preparation of porous crystal-
line synthetic material comprised of silicon and titanium oxides, US
Patent 4410501 (1983).
31. M.G. Clerici, Oil Gas Eur. Mag., 32, 77 (2006).
32. S. Park, K.M. Cho, M.H. Youn, J.G. Seo, S.H. Baeck, T.J. Kim, Y.M.
Chung, S.H. Oh and I.K. Song, Catal. Lett., 122, 349 (2008).
33. G. Paparatto, A. Forlin and P. Tegon, Process for the Preparation of
Olefin Oxides, US Patent 7442817 (2008).
46. D.P. Serrano, R. Sanz, P. Pizarro, I. Moreno, P. de Frutos and
S. Blázquez, Catal. Today, 143, 151 (2009).
47. H. Thomas, H. Willi, T. Georg and S. Joerg, Process for the epoxidation
of olefins, Eur. Patent 1247805 (2002).
48. J. Hanika and A. Hajkova, Chem. Prum., 37, 398 (1987).
49. H. Liu, G.Z. Lu, Y.L. Guo, Y. Guo and J.S. Wang, Chem. Eng. J., 116,
179 (2006).
34. V.Arca,A. Boscolo Boscoletto, N. Fracasso, L. Meda and G. Ranghino,
J. Mol. Catal. Chem., 243, 264 (2006).
35. W.G. Cheng, X.S. Wang, G. Li, X.W. Guo and S.J. Zhang, J. Catal.,
255, 343 (2008).
50. G.N. Vayssilov, Catal. Rev., Sci. Eng., 39, 209 (1997).
51. G. Bellussi, A. Carati, M.G. Clerici, G. Maddinelli and R. Millini, J.
Catal., 133, 220 (1992).
36. X.W. Liu, X.S. Wang, X.W. Guo and G. Li, Catal. Today, 93-95, 505
(2004).
52. P.E. Sinclair and C.R.A. Catlow, J. Phys. Chem. B, 103, 1084 (1999).
53. M.G. Evans and M. Polanyi, Trans. Faraday Soc., 31, 875 (1935).
54. R.H. Perry and D.W. Green, Perry's Chemical Engineers' Handbook;
McGraw-Hill: New York (2007).
37. G. Li, J. Meng, X. Wang and G. Xinwen, React. Kinet. Catal. Lett., 82,
73 (2004).
38. Z.G. Zhang, J.N. Kang and Y. Wang, React. Kinet. Catal. Lett., 92, 49
(2007).