Effect of propylene glycol monomethyl ether and rust impurities on TS-1 deactivation in propylene epoxidation

Article abstract of DOI:10.1016/j.cattod.2012.09.039

The properties of TS-1 catalyst in continuous reaction of propylene epoxidation, as well as the effects of by-product propylene glycol monomethyl ether (PGME) on catalytic performance, was investigated in a low-pressure fixed bed reactor with chemical pure methanol as solvent. As the rust could be doped from the pipings and solvent containers, its impact on catalyst deactivation was then studied by using industrial-grade methanol. Finally the catalyst before and after deactivation was contrastively characterized by means of micro-structure analysis, TPD and TPR. The results showed that self-developed TS-1 catalyst performed well in the reaction, and the by-product propylene glycol monomethyl ether would not depress the catalytic activity in 80 h. The presence of rust could not only block active sites leading to partial loss of activity, but also increase the surface acidity, which would induce side reactions and decrease the selectivity of propylene oxide.

Full text of DOI:10.1016/j.cattod.2012.09.039

Catalysis Today 212 (2013) 169–174
Contents lists available at SciVerse ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Effect of propylene glycol monomethyl ether and rust impurities on
TS-1 deactivation in propylene epoxidation
Li Yi-chuan, Shen Ben-xian, Zhao Ji-gang∗
State Key Laboratory of Chemical Engineering, East China University of Science and Technology, Shanghai 200237, China
a r t i c l e i n f o
a b s t r a c t
Article history:
The properties of TS-1 catalyst in continuous reaction of propylene epoxidation, as well as the effects of
by-product propylene glycol monomethyl ether (PGME) on catalytic performance, was investigated in
a low-pressure fixed bed reactor with chemical pure methanol as solvent. As the rust could be doped
from the pipings and solvent containers, its impact on catalyst deactivation was then studied by using
industrial-grade methanol. Finally the catalyst before and after deactivation was contrastively charac-
terized by means of micro-structure analysis, TPD and TPR. The results showed that self-developed TS-1
catalyst performed well in the reaction, and the by-product propylene glycol monomethyl ether would
not depress the catalytic activity in 80 h. The presence of rust could not only block active sites leading
to partial loss of activity, but also increase the surface acidity, which would induce side reactions and
decrease the selectivity of propylene oxide.
Received 14 March 2012
Received in revised form
14 September 2012
Accepted 18 September 2012
Available online 16 April 2013
Keywords:
Propylene oxide
TS-1
Catalytic performance
Deactivation
Characterization
© 2013 Elsevier B.V. All rights reserved.
1. Introduction
chain length, the diffusivity of these oligomers decreases rapidly,
until they literally get stuck in the zeolite pores and block the
As one of the important organic intermediate chemicals, more
than four million tons of propylene oxide (PO) have been glob-
ally produced every year. However, in two traditional methods
of manufacturing PO, chlorhydrin process has serious pollution
to the environment whereas the co-oxidation process produces a
large amount of joint products. Consequently, developing a short,
clean process for producing PO is needed to replace the currently
used industrial processes. Since the MFI-type titanosilicate, TS-1
was synthesized [1], direct epoxidation of propylene to propyl-
ene oxide (PO) over TS-1 using hydrogen peroxide (H2O2) as the
oxidant, is regarded as the important development trend of PO
green production [2–7], for its mild reaction conditions, high cat-
alytic activity and high atom economy without pollution problems
[2,3].
Although TS-1 gives excellent catalytic activity of the propyl-
ene epoxidation, the deactivation of the catalyst occurs on stream
[8]. The deactivation mechanism of TS-1 is one of very impor-
tant subjects in the industrialization of propylene epoxidation, but
only a few reports on the TS-1 deactivation have been reported.
Thiele et al. [3] thought that PO and the by-product propylene gly-
col monomethyl ether (PGME) could form oligomers within the
zeolite micropores under acid catalysis reaction. With increasing
access of reactants to the catalytic sites. On the other hand, the
activity of TS-1 for propylene epoxidation by hydrogen peroxide
in methanol medium decreases due to titanium leaching from the
catalyst framework [9]. Chen et al. [10] reported that pore plug-
ging was resulted from by-products with large dimensions such
as dimers or oligomers and leaching framework titanium would
simultaneously deactivated the catalyst. On the basis of simula-
tion experiment, Yan et al. [11] reported that the deactivation
was mainly caused by the increasing amount of dimers and trim-
mers formed by PO. Liu et al. [12] studied deactivated catalyst
by desorption in the presence of N2, the deactivation mechanism
they postulated was that adsorbed PO and PGME at active sites
of TS-1 could suppress the catalytic activity, and the polyethers
formed by PO and PGME, as well as the dimers and trimmers
formed by PO could block the micropores or occupy the active
sites.
In the present study, the performance of TS-1 catalyst in a
continuous reaction system for propylene epoxidation, and the
impact of the by-product PGME on catalytic activity are studied in
a low-pressure fixed bed reactor with chemically pure methanol as
solvent. Considering the corrosion of pipings and solvent containers
in the pilot-plant, the influence of rust on TS-1 deactivation is exer-
cised by using chemical-grade methanol as solvent under typically
reaction conditions. Subsequently the mechanism of deactivation
caused by rust is investigated by characterizing the sample of fresh
and deactivated TS-1 catalyst.
∗
Corresponding author. Tel.: +86 2164252916.
E-mail addresses: zjg@ecust.edu.cn, zjg13122752231@hotmail.com (Z. Ji-gang).
0920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2012.09.039

170
L. Yi-chuan et al. / Catalysis Today 212 (2013) 169–174
Fig. 1. Schematic diagram of experiment.
2. Experimental
a continuous-flow fixed bed reactor and its typical conditions [15]
were listed in Table 1. The amount of catalyst loaded in the reactor
was 4 g.
2.1. Materials
The residual H2O2 was measured by iodometric titration [16].
Reaction products were quantifiably calculated by a relative
correction factor method [17] and analyzed on a HP-1 gas chro-
matography using a flame ionization detector and a capillary
column (50 m × 0.32 m × 1.05 m). PO was the main product in this
reaction, propylene glycol monomethyl ether (PGME) and propyl-
ene glycol (PG) were the byproducts. The reaction results were
described by the following evaluating indicators:
Propylene (purity >95%) was provided by Gao Qiao SINOPEC
(Shanghai, China). Hydrogen peroxide (30 wt%) was obtained from
Jin Lu Co. Ltd (Shanghai, China). Methanol and propylene oxide
were provided by Ling Feng Chemicals Co. Ltd (Shanghai, China).
They were analytical reagent-grade.
2.2. Synthesis of TS-1
nH0
− niH
O
O
2
2
2
2
In a typical preparation [13], 1.92 g tetrapropylammonium bro-
mide (TPABr, templating agent) was dissolved in 80 ml deionized
water, then 20 ml colloidal silica NS-30(30 wt%, silica sources)
was added under stirring for 1 h. Amounts of acetylacetone and
tetrabutyl titanate (TBOT) were added in 10 ml isopropyl-alcohol
orderly; this solution was called titanium alkoxide complex and
was added and stirred for an hour, then 3.5 g n-butylamine was
added. The mixture was then heated at 50 ◦C and 60 ◦C for an hour,
respectively, under stirring.
The transparent gel, whose composition is as follows:
SiO2:xTiO2:0:06TPA:45H2Owhere 0 < x < 0:04 was then trans-
ferred into a Teflon-lined autoclave and was heated at 175 ◦C under
autogenous pressure for three days.
H2O2 conversion : XH
=
× 100%
O
2
2
nH0
O
2
2
nPO
PO selectivity : SPO
=
× 100%
nPO + nPGME + nPG
nPO
PO yield coefficient : YPO
=
× 100%
id
PO
n
XH , SPO, and YPO denote the conversion of H2O2, the selectiv-
ity of PO and the yields of PO, respectively. The nPO, nPGME, and nPG
stand for the number of moles of PO, PGME, and PG, respectively.
O
2
2
The n0H
and niH
were the initial mole content and the final mole
O
2
O
2
2
2
After crystallization, the autoclave was cooled, and the solid was
recovered by filtration. The zeolite was washed several times with
deionized water and dried overnight at 110 ◦C in an oven. Occluded
organics were removed by heating the sample in an air flow at
550 ◦C for about 6 h.
TS-1 molecular sieve catalyst was granulated according to the
patent [14],the particle size of the catalyst used was from 20 mesh
to 40 mesh.
Table 1
Conditions of catalytic experiments.
Conditions
Details
Solvent
Temperature/K
Pressure/MPa
Initial H2O2 concentration/wt%
Mole ratio of H2O2 to C3H6
WHSV of propylene/h−1
LHSV of H2O2 solution/h−1
Particle size of catalyst/meth
Reactor
Methanol
323
0.7
2.1
1:2.5
2.1
32.4
20–40
PFR
2.3. Catalytic experiments
The experimental procedures of the direct epoxidation reaction
are shown in Fig. 1. The epoxidation of propylene was carried out in
Reactor dimensions
L/D = 50

L. Yi-chuan et al. / Catalysis Today 212 (2013) 169–174
171
Table 2
Instrumental and operating conditions for ICP-AES measurements.
100
Parameter
Type or amount
R.F. power (kw)
1.15
0.6
1.0
14
30
100
15
20
Carrier gas (Ar) flow rate (L min−1
)
90
80
70
60
Auxiliary gas (Ar) flow rate (L min−1
)
Coolant gas (Ar) flow rate (L min−1
Nebulizer flow (psi)
)
Pump rate (r min−1
)
X
H2O2
Observation height (mm)
Integration time (s) on-axis
off-axis
S
PO
5
Wavelength (nm)
Fe 259.940
content of H2O2, respectively. The nid represent the mole content
of PO that all the H2O2 fully react with C3H6 would produce.
PO
0
20
40
60
80
100
120
140
2.4. Characterization of TS-1
Reaction time (h)
The crystal structure was identified by X-ray diffraction (XRD)
(Rigaku D/max- 2550VB/PC) analyses using Cu-K␣ radiation. Spec-
tra were collected from 2ꢀ = 5◦ to 50◦ in 0.02◦ steps. The crystal
morphology was examined by a scanning electron microscopy
(SEM) (JEOL Model JSM-6360 LV, Japan). Framework IR spectra
were recorded on a Nicolet Magna-IR550, using KBr wafers.
Inductively coupled plasma atomic emission spectrometry(ICP-
AES) was used for Fe determination by an IRIS Advantage ER/S
inductively coupled plasma spectrometer (TJA, USA). The opera-
tion conditions and the wavelengths were summarized in Table 2.
The sample of TS-1 was pre-dissolved by aqua regia.
Fig. 2. Long-run test of the catalyst for propylene epoxidation.
catalyst effectively catalyzes the epoxidation of propylene as both
the H2O2 conversion and PO selectivity are maintained at about 90%
during 140 h on steam, and therefore the stability is impressively
good.
3.2. PGME effect on catalytic performances
PGME is the overriding by-product of propylene epoxidation
[16], formed by consecutive reactions of target reaction. Propylene
epoxidation was operated in the reactor and PGME was added to
The micro-structure of TS-1 was analyzed on
a Micro-
meritics ASAP 2020 including single- and multipoint BET
(Brunauer–Emmett–Teller) surface area, Langmuir surface area,
pore volume and pore-length distributions in the mesopore ranges
by the BJH (Barrett–Joyner– Halenda) algorithm, and the distribu-
tions of micropore volume and pore size by the methods of H–K
(Horvath-Kawazoe) with Saito & Foley model for cylindrical pores
as well.
H2-temperature-programmed reduction (H2-TPR) was per-
formed on a Micromeritics AutoChem II 2920 chemical adsorption
instrument. Using mixture gas H2–Ar (9.92% H2) as reduction gas.
Typically, the sample (0.2 g) was first pretreated in a quartz reactor
with a high-purity He gas flow at 120 ◦C for 1 h. After the sample was
cooled to room temperature, the H2–Ar mixture was introduced
into the reactor at a rate of 5 cm3 min, and the temperature was
raised to 700 ◦C at a rate of 10 ◦C/min. The consumption of H2 was
monitored by a thermal conductivity detector (TCD).
NH3-temperature-programmed desorption (NH3-TPD) was also
carried out with the Micromeritics AutoChem II 2920 chemical
adsorption instrument. Typically, the sample (0.2 g) loaded in the
quartz tube was first pretreated with a high-purity He gas flow at
250 ◦C for 1 h. The adsorption of NH3 was performed in an NH3–He
(9.93% NH3) mixture for 0.5 h, and then the remaining or weakly
adsorbed NH3 was purged by high-purity He. TPD was performed
in the He flow by raising the temperature to 800 ◦C at a rate of
10 ◦C/min. The desorbed NH3 was detected by the mass spectrom-
eter.
the raw material (0.5 wt%) continuously. From the trend of XH
,
O
SPO, and YPO shown in Fig. 3, it can be seen that the PO selectivit2y,
H2O2 conversion and yield of PO are monotonous at about 85%, 68%
and 57% in several, and they are distinctly stable along the reaction
time with PGME injection.
2
Since the by-product formation is due largely to strong acid
sites [3], the pH value of feed solution is adjusted to 5.5 for forc-
ing the rate of PO selectivity to be reduced at about 80%, in order
to further investigate the effect on catalyst of PGME. The reaction
status is exhibited in Fig. 4, it is observed that during the reaction
of approximately 80 h, the H2O2 conversion persists at 95% stably
as before and the PO selectivity fluctuates slightly near 80% all the
90
80
X
H2O2
S
PO
YPO
70
60
50
40
3. 2 Results and discussion
3.1. Catalytic performances
In order to prevent the formation of by-products, pH value of the
feed solution in the tests using chemically pure methanol as solvent
was adjusted to about 7 by ammonia solution (0.1 mol/L) [16]. The
level of catalytic reaction is shown in Fig. 2, it is clearly that the TS-1
0
20
40
60
80
Reaction time (h)
Fig. 3. Effect of PGME on catalytic performance for propylene epoxidation.

172
L. Yi-chuan et al. / Catalysis Today 212 (2013) 169–174
100
90
80
70
60
50
40
XH2O2
SPO
Fig. 6. XRD spectra of TS-1.
0
10
20
30
40
50
60
70
80
that titanium was incorporated into the framework of TS-1 catalyst,
and there was no distinct transformation occurred in crystal struc-
ture during the reaction with rust. This result can also be confirmed
by their SEM images (Fig. 7).
Reaction time /h
Fig. 4. Effect of by-product on catalytic performance for propylene epoxidation.
The image of TS-1-USED in Fig. 8a was absolutely different
from the image of TS-1-FRESH in Fig. 8b. After reaction, the
surface of catalyst was covered by a layer of tawny material.
The measured concentration of Fe in TS-1-USED by ICP-AES was
9.058 0.010 ␮g/g. That means the tawny material was fine par-
ticles of rust. The infrared spectra in Fig. 9 exhibited a strong
absorption band at 960 cm−1 that revealed the existence of frame-
work metallic ions and particularly framework Ti in TS-1. But the
intensity of band at 960 cm−1 of TS-1-USED was weakened as com-
pared with TS-1-FRESH, this might be the result of combined effects
of framework Ti and the rust adhered on catalyst.
The micro-structure analysis results are given in Table 3. As can
be seen from the table, surface area and pore volume of catalyst all
diminish for a certain extent after reaction, indicating that part of
pores are covered by the rust so that the reactants cannot reach the
active site resulting in the reduction of H2O2 conversion.
time, revealing that by-product PGME would not cause perceptible
deactivation in this reaction system in at least 80 h.
3.3. Catalytic performances with rust impurities
At the time using industrial-grade methanol as solvent in the
plant, the pipings of device and containers holding solvent could
introduce rust impurities into reactor by the feed stream scouring.
Feed solvent was pumped into reactor through a rusty pipe then
epoxidation reaction was carried out under the typical conditions,
and the catalytic performances were presented in Fig. 5. Only 60 h,
H2O2 conversion, PO selectivity and the yield of PO started to fall
and separately dropped by 12.5%, 20.4% and 25.8% until the reaction
of 96 h, which illustrated that the rust do give rise to a certain loss
in catalyst activity.
The micropore distribution by HK method and mesopore dis-
tribution by BJH method are depicted in Figs. 10 and 11. It can be
observed that a loss of accessible micropore volume in the sorp-
tion isotherms occurred (Fig. 10), and the uniform distribution of
mesopores disappeared (Fig. 11). After reaction the surface area,
pore volume and distribution of TS-1 are probably altered by the
blockage of channels resulting from the rust adsorbed.
Redox property of a catalyst is an important factor affecting
its catalytic performances, particularly in selective oxidation reac-
tions. The H2 absorption peak appeared in the H2-TPR process
should be H2 consumption peak during the sample was deoxidated
by H2, referred to as the H2-TPR peak. The H2-TPR profiles for the
sample of TS-1-FRESH and TS-1-USED are shown in Fig. 12. A H2-
TPR peak is observed at 201.5 ◦C for the TS-1-USED sample while
no obvious peak for the TS-1-FRESH sample, revealed that the TS-1
can be reduced after the catalyst is polluted by rust in feed solu-
tion. This result clearly indicates that the components from rust
3.4. Characterization of TS-1 catalysts
The fresh TS-1 catalyst was named TS-1-FRESH. After the epox-
idation of propylene with rust, the deactivated catalyst taken out
of reactor was named TS-1-USED. Their X-ray diffraction patterns
were shown in Fig. 6. The characteristic peaks attributed to the MFI
structure obviously appeared in the patterns, while no peak was
observed at 2ꢀ = 25.3◦ and a twin peak at 2ꢀ = 45◦. It was indicated
100
90
80
70
60
50
Table 3
Micropore parameters of TS-1 catalyst before and after reaction.
Catalyst
Surface area/m2 g−1
Pore volume/cm3 g−1
X
H2O2
SBET
SLANGMUIR
Ssp
SBJH
Vsp
VBJH
S
PO
40
30
TS-1-FRESH
TS-1-USED
382
296
513
399
382
294
284
174
0.2197
0.1816
0.1523
0.1160
Y
PO
SBET, multipoint BET surface area; SLANGMUIR, Langmuir surface area; Ssp, single-
point BET surface area (P/P0 = 0.20); SBJH, surface area of mesopores between 1 nm
and 110 nm by the BJH method; Vsp, pore volume by single-point BET method
(P/P0 = 0.983, pore width less than 117 nm); VBJH, cumulative desorption volume
of mesopores between 1 nm and 110 nm by the BJH method.
0
10
20
30
40
50
60
70
80
90 100
Reaction time (h)
Fig. 5. Effect of rust in liquid feed on catalytic epoxidation of propylene.

L. Yi-chuan et al. / Catalysis Today 212 (2013) 169–174
173
Fig. 7. SEM image of TS-1.
Fig. 8. Image of TS-1.
will be adsorbed on the TS-1 surface and the H2-TPR peak is just
the reduction peak for iron species. This further interpreted that the
rust will cause deactivation of TS-1 zeolite catalyst in the propylene
epoxidation.
As the acid sites are detrimental to propylene epoxidation over
the TS-1 catalyst, the increase of surface acidity is likely to cause
side reactions, reducing the selectivity of PO. NH3-TPD curves of
TS-1-FRESH and TS-1-USED are shown in Fig. 13. It is observed that
the sample of TS-1-FRESH has one desorption peak at about 90 ◦C
whereas the sample of TS-1-USED has another peak at higher tem-
perature corresponding to relatively strong acidity. That means the
amount of acidity on the catalyst surface is highly enhanced due
to the appearance of the large number of stronger acid sites. It has
been clarified that the joined rust can depress the catalytic perform-
ances by changing the number of acid sites on the TS-1 surface. This
also confirms the situation that reducing the pH value of feed can
promote the generation of PGME during the study in Section 2.2.
0.14
0.12
0.10
0.08
0.06
TS-1-FRESH
TS-1-USED
0.04
0.02
0.00
4
6
8
10
12
14
16
18
20
Pore Width (Å)
Fig. 9. IR spectra of TS-1.
Fig. 10. H–K pore volume plot.

174
L. Yi-chuan et al. / Catalysis Today 212 (2013) 169–174
4. Conclusions
The PO selectivity and H2O2 conversion of 90% were obtained
during the reaction of approximately 140 h, while the by-product
propylene glycol monomethyl ether would not disturb the cat-
alytic activity over 80 h. Nevertheless, in the pilot-plant of PO, the
components of rust from pipings and containers lead to a notice-
able decline in the catalytic performance of propylene epoxidation,
connected to not only decrease of BET surface area and surface mor-
phology alterations but also significant increase of surface acidity.
In order to raise the property and stability of TS-1 catalyst in epox-
idation of propylene, the rust must be removed from the feedstock
and the acidity of catalyst must be controlled in the pilot-scale test
reaction systems.
TS-1-FRESH
TS-1-USED
Acknowledgements
500
10
100
50
20
Pore Width (Å)
We are grateful to the National High Technology Research and
Development Program of China (863 Program: 2012AA040306)
and Tianjin Dagu Chemical Co., Ltd. (China) for financial
support.
Fig. 11. BJH pore volume plot.
201.5
References
[1] M. Taramasso, G. Perego, B. Notari, US Patent 4410 501 (1983).
[2] M.G. Clerici, G. Bellussi, U. Romano, Journal of Catalysis 129 (1991)
159–167.
[3] G.F. Thiele, E. Roland, Journal of Molecular Catalysis 117 (1997)
351–356.
TS-1-USED
[4] A. Tuel, Zeolites 16 (1996) 108–117.
[5] X.S. Wang, X.W. Guo, G. Li, Catalysis Today 74 (2002) 65–75.
[6] W. Liz, Urethane Technology 21 (2004) 4.
[7] C. Perego, A. Carati, P. Ingallina, M.A. Mantegazza, G. Bellussi, Applied Catalysis
A: General 221 (2001) 63–72.
[8] Q.F. Wang, W. Li, J.X. Chen, Y.L. Wu, Z.T. Mi, Journal of Molecular Catalysis A:
Chemical 273 (2007) 73–80.
TS-1-FRESH
[9] L. Davies, P. McMorn, D. Bethell, P. Page, F. King, F.E. Hancock, G.J. Hutchings,
Chemical Communications 18 (2000) 1807–1808.
[10] J.X. Chen, Z.T. Mi, Y.L. WU, W. Li, Z.H. Li, Journal of Fuel Chemistry and Technol-
ogy 31 (2003) 80–85.
[11] H.S. Yan, X.S. Wang, J. Liu, X.W. Guo, Acta Petrolei Sinica (Petroleum Processing
Section) 18 (2002) 46–51.
[12] X.W. Liu, X.S. Wang, R.P. Wang, X.W. Guo, G. Li, Journal of Molecular Catalysis
17 (2003) 247–254(China).
0
100
200
300
400
500
600
700
Temperature (oC)
[13] W.G. Xiao, B.X. Shen, J.G. Zhao, Y.T. Wei, C. Zhang, L. Wang, H. Ling, X. Chen, CN
Patent 1751996 (2005).
Fig. 12. H2-TPR profiles of catalyst.
[14] B.X. Shen, W.G. Xiao, J.G. Zhao, Y.T. Wei, Z.M. Zhang, H. Chen, L. Wang, C. Zhang,
CN Patent 1911516 (2007).
[15] J.G. Zhao, B.X. Shen, W.G. Xiao, C. Zhang, L. Wang, Energy sources.
Part A: Recovery, Utilization, and Environmental Effects 31 (2009)
108–117.
[16] J.G. Zhao, B.X. Shen, W.G. Xiao, C. Zhang, L. Wang, Acta Petrolei Sinica
(Petroleum Processing Section) 21 (2005) 44–49.
TS-1-FRESH
TS-1-USED
[17] J.R. Jiao, B.X. Shen, J.G. Zhao, Energy Sources part A: Recovery, Utilization, and
Environmental Effects 33 (2011) 1147–1154.
90.8
88.6
313.0
0
100
200
300
400
500
600
700
800
Temperature (oC)
Fig. 13. NH3-TPD profiles of catalyst.