Cyclodehydration of diethylene glycol over Ag-modified Al2O3 catalyst

Article abstract of DOI:10.1016/j.apcata.2015.03.047

Vapor-phase cyclodehydration of diethylene glycol (DEG) into 1,4-dioxane (DOX) were performed over several solid acid catalysts, such as Al₂O₃, SiO₂-Al₂O₃, Beta and MFI zeolites. All the catalytic activity gradually decreased with time on stream, especially Beta and MFI zeolites with strong acidic property. The catalysts were modified with silver metal for stabilizing the catalytic activity: Al₂O₃ modified with 1 wt.% Ag₂O showed the best catalytic performance. Characterizations such as XRD, TPD and FTIR were performed for analyzing the carbonaceous compound deposited on the catalysts after reaction and investigating the effect of the addition of Ag₂O onto Al₂O₃. Polyethylene glycol was proposed as the carbonaceous compound by FTIR analysis. The modification of Ag was found to decrease the averaged acid strength of Al₂O₃, which is supposed to be effective for inhibiting carbon deposition and stabilizing the catalytic activity. It was also found that H₂, as a carrier gas, was indispensable for stabilizing the conversion of DEG into DOX. Silver metal would work as a remover of the product on the catalyst surface together with H₂ to prevent carbon deposition. Over Al₂O₃ modified with 1 wt.% Ag₂O, a stable DOX selectivity of 90% with a complete conversion was achieved at 250 °C and a contact time of 156.8 g h mol^-1.

Full text of DOI:10.1016/j.apcata.2015.03.047

G Model
APCATA-15381; No. of Pages9
ARTICLE IN PRESS
Applied Catalysis A: General xxx (2015) xxx–xxx
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Cyclodehydration of diethylene glycol over Ag-modified Al O catalyst
2
3
∗
Daolai Sun, Junjie Wang, Yasuhiro Yamada, Satoshi Sato
Graduate School of Engineering, Chiba University, Chiba 263-8522, Japan
a r t i c l e i n f o
a b s t r a c t
Article history:
Vapor-phase cyclodehydration of diethylene glycol (DEG) into 1,4-dioxane (DOX) were performed over
several solid acid catalysts, such as Al2O3, SiO2 Al2O3, Beta and MFI zeolites. All the catalytic activity
gradually decreased with time on stream, especially Beta and MFI zeolites with strong acidic prop-
erty. The catalysts were modified with silver metal for stabilizing the catalytic activity: Al2O3 modified
with 1 wt.% Ag2O showed the best catalytic performance. Characterizations such as XRD, TPD and FTIR
were performed for analyzing the carbonaceous compound deposited on the catalysts after reaction and
investigating the effect of the addition of Ag2O onto Al2O3. Polyethylene glycol was proposed as the
carbonaceous compound by FTIR analysis. The modification of Ag was found to decrease the averaged
acid strength of Al2O3, which is supposed to be effective for inhibiting carbon deposition and stabiliz-
ing the catalytic activity. It was also found that H2, as a carrier gas, was indispensable for stabilizing
the conversion of DEG into DOX. Silver metal would work as a remover of the product on the catalyst
surface together with H2 to prevent carbon deposition. Over Al2O3 modified with 1 wt.% Ag2O, a sta-
Received 8 February 2015
Received in revised form 25 March 2015
Accepted 27 March 2015
Available online xxx
Keywords:
Diethylene glycol
Cyclodehydration
1
,4-Dioxane
Ag-modified Al2O3
Catalyst deactivation
◦
ble DOX selectivity of 90% with a complete conversion was achieved at 250 C and a contact time of
−1
1
56.8 g h mol
.
©
2015 Elsevier B.V. All rights reserved.
1
. Introduction
such as Yb O and Sc O [10,16] prefer the formation of 4-buten-
2 3 2 3
1
-ol in a vapor phase.
In the cyclodehydration of DEG, DOX can be selectively obtained
Cyclodehydration of terminal diols, such as 1,4-butanediol
BDO), 1,5-pentanediol (PDO) and diethylene glycol (DEG), is a
(
by liquid-phase catalytic reaction in the presence of sulfated zir-
conia [1], alumina [14], H-ZSM-5 [17] or Brønsted acidic ionic
liquid (1-butyl-3-methylimidazolium hydrogen sulfate) [18] cat-
alysts (Scheme 1). A maximum DOX yield of 87% is obtained in the
useful method to obtain oxygen heterocycles, such as tetrahydro-
furan (THF), tetrahydropyran (THP), and 1,4-dioxane (DOX), which
are widely used as the solvents in chemical industry [1]. Many
researches dealing with the dehydration of BDO, PDO, and DEG to
useful chemicals have been reported and it is found that acidic cat-
alysts are generally effective for producing the cyclodehydration
◦
reaction using Brønsted acidic ionic liquid as a catalyst at 170 C
[18]. Although vapor-phase reactions are widely used in chemical
industry, there are a few reports dealing with the cyclodehydra-
tion of terminal diols in a vapor phase. Kitayama et al. performed
the cyclodehydration of DEG into DOX over clay mineral sepiolite,
products. In particular, acidic solids such as TiO -supported silico-
2
tungstic acid [2,3], diatomite-supported phosphotungstic acid [4],
and ion exchange resin (Indion-130) [5] are efficient for THF for-
mation from BDO in a liquid phase, while 3-buten-1-ol is mainly
◦
and a maximum DOX yield of 98% was obtained at 260 C in the
initial period of several hours [19]. However, the catalytic activ-
ity decreased with time on stream rapidly. Gradual deterioration
of acidic catalysts is an important issue in the catalytic conversion
under flow conditions.
produced in a vapor phase over base-modified ZrO2 [6,7], In O3
2
[
8,9], and rare earth oxides with base property, such as CeO2 and
Er O [10,11]. Such solid catalysts as sulfated zirconia [1], H-ZSM-
2
3
5
zeolite [12], tin(IV) phosphate [13], and alumina [14] have been
The stability of catalyst is highly required in an industrial pro-
cess. In our group, an effective operation on the stabilization of
catalytic activity of solid acids has been reported [20–22]. For exam-
ple, the vapor-phase dehydration of 1,2-propanediol to propanal
can be catalyzed by acidic SiO2 Al O , whereas both the con-
reported to be effective for the liquid-phase formation of THP from
PDO, while Na-modified monoclinic ZrO [15] and rare earth oxides
2
2
3
∗ Corresponding author. Tel.: +81 43 290 3376; fax: +81 43 290 3401.
E-mail address: satoshi@faculty.chiba-u.jp (S. Sato).
version of 1,2-propanediol and the selectivity to propanal over
SiO2 Al2O3 are deteriorated at 300 C. Ag-modified SiO2 Al2O3
◦
http://dx.doi.org/10.1016/j.apcata.2015.03.047
0
926-860X/© 2015 Elsevier B.V. All rights reserved.
Please cite this article in press as: D. Sun, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.03.047

G Model
APCATA-15381; No. of Pages9
ARTICLE IN PRESS
D. Sun et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
2
O
(CARiACT Q-10 with a mean pore diameter of 10 nm) was sup-
-
H O
2
HO
O
Diethylene glycol (DEG)
OH
plied by Fuji Silycia Chemical Ltd. AgNO , which was used as the
3
O
precursor of the additive Ag, was purchased from Wako Pure Chem-
ical Industries Co., Ltd., Japan.
1,4-Dioxane (DOX)
Ag-modified catalysts were prepared by an incipient wetness
impregnation method using a solution with a prescribed amount
Scheme 1. Cyclodehydration of diethylene glycol to 1,4-dioxane.
of AgNO dissolved in distilled water. The weight percentages of Ag
3
in the support were calculated using the weight of Ag O in the cal-
2
catalyst exhibits stable catalytic activity with high selectivity to
propanal under hydrogen flow conditions. Prior to the reaction,
cined catalysts. An appropriate amount of the AgNO3 solution was
added to a support, and water was evaporated at ambient pres-
sure and 70 C by illuminating samples using a 350-W electric light
◦
Ag O is reduced to metallic Ag, which prevents carbon deposition
2
together with hydrogen [22].
bulb. After the impregnation process, the samples were dried at
◦
In this paper, we investigated the vapor-phase cyclodehydration
of DEG into DOX (Scheme 1) over several commercial solid acids
with modification of Ag. The suitable reaction conditions for the
DOX formation and the behavior of the Ag-modified catalysts were
also studied.
110 C for 12 h, and then calcined for 3 h. The calcination tempera-
◦
ture of Ag-modified catalysts was 450 C. Hereafter, the supported
catalysts are expressed as X-Ag-Y, where X means the support; Y
means the weight percentage of Ag2O. For example, Al2O3-Ag-1
means 1 wt.% Ag2O loaded on Al2O3.
2
. Experimental
2.2. Catalytic reaction
2
.1. Samples
The cyclodehydration of DEG was performed in a fixed-bed
down-flow glass reactor with an inner diameter of 17 mm at an
ambient pressure of either H or N . Prior to the reaction, the cat-
DEG and polyethylene glycol (PEG) with a molar weight
2
2
of 4000 were purchased from Wako Pure Chemical Industries
Co., Ltd., Japan, and were used for the catalytic reaction with-
out further purification. Al O3 (N611N) and SiO2 Al O (N631L)
alyst was placed in the catalyst bed and heated at a prescribed
temperature for 1 h to stabilize the temperature of the catalyst bed.
After the preheating, DEG was fed through the top of the reactor
2
2
3
−
1
were purchased from Nikki Chemical. Beta (Si/Al = 31.9) and MFI
Si/Al = 90.2) zeolites were supplied by Clariant Catalysts Industries
Co., Ltd. The specific surface areas of Al O , SiO Al O , Beta and
at a liquid feed rate of 2.0 g h together with either an H2 or N2
3
−1
−
(
flow of 20 cm min . The reactions were performed at a W/F from
1
26.1 to 261.4 g h mol , where W and F are the catalyst weight and
2
3
2
2
3
2
−1
MFI zeolite are 210, 440, 586 and 431 m g , respectively. Silica
the feed rate of DEG, respectively. The liquid effluents collected
Table 1
Dehydration of DEG over several metal oxide catalysts at 250 C.a
◦
Catalyst
Conversion/%b
Selectivity/%b
DOX
MeOH
EtOH
AA
MDO
EG
TEG
Al2O3
MFI
Beta
SiO2 Al2O3
SiO2
58.4
86.7
91.9
27.1
0.0
67.5
54.0
74.6
30.5
0.0
0.0
0.1
0.1
0.0
0.0
1.0
0.4
1.0
2.2
0.0
1.0
6.5
4.2
5.1
0.0
1.5
6.6
2.5
7.7
0.0
15.1
14.4
8.1
19.5
0.0
1.6
0.4
0.2
4.4
0.0
DOX, 1,4-dioxane; MeOH, methanol; EtOH, ethanol; AA, acetaldehyde; MDO, 2-methyl-1,3-dioxolane; EG, ethylene glycol; TEG, triethylene glycol.
a
Reaction conditions: W/F = 26.1 g h mol−1; H2 flow rate, 20 cm3 min−1
Average activity in the initial 5 h.
.
b
1
00
100
80
60
40
20
Beta
MFI
8
0
Beta
Al O
2
3
6
0
Al O
MFI
2
3
4
0
2
0
SiO -Al O
2 2 3
SiO -Al O
3
2
2
(a)
(b)
0
1
2
3
4
5
0
1
2
3
4
5
Time on stream / h
Time on stream / h
Fig. 1. Changes in conversion and selectivity with time on stream over several solid acid catalyst. Reaction conditions: W/F = 26.1 g h mol 1; H2 flow rate, 20 cm3 min
reaction temperature, 250 C.
−
−1
;
◦
Please cite this article in press as: D. Sun, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.03.047

G Model
APCATA-15381; No. of Pages9
ARTICLE IN PRESS
D. Sun et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
3
O
HO
OH
-H O
2
O
O
Diethylene glycol (DEG)
-H2
1,4-Dioxane (DOX)
Beta-Ag-1
O
O
Decomposition
Beta
SiO -Al O -Ag-1
O
HO
OH
+DEG -H O
2
HO
O
2
2
3
O
OH
+
O
CH OH
3
+
-
H2
O
SiO -Al O
3
2
2
O
OH
Acetal formation
Scheme 2. Probable formation route of the detected products.
MFI-Ag-1
◦
in an ice water trap at 0 C every hour were analyzed by a FID-
GC (GC-8A, Shimadzu) with a 30-m capillary column of TC-WAX
MFI
(
GL-Science, Japan). A GC–MS (QP5050A, Shimadzu) was used for
Al O -Ag-1
2
3
identification of the products in the effluent. 1,2-Propanediol was
used as an internal standard substance.
Al O
2
3
2
.3. Characterization of catalysts
1
00 200 300 400 500 600
Temperature (°C)
The specific surface area of a catalyst sample was calculated with
the Brunauer–Emmett–Teller (BET) method using N adsorption
2
Fig. 2. NH3-TPD profiles of Ag-modified solid acid catalysts.
◦
isotherm at −196 C. The thermo gravimetry-differential ther-
mal analysis (TG-DTA) was performed using Thermoplus 8120E2
(
Rigaku, Japan) under the conditions: sample weight, ca. 10 mg; the
titration using an electric conductivity cell immersed in an aque-
◦
−1
heating rate, 5 C min ; heating range, from the room tempera-
ture to 900 C. In the TG plots, profiles were located to be contacted
each other at around 150–200 C. The temperature-programmed
ous solution of H SO4 to estimate acidity of the catalysts, as has
2
◦
been described in the TPD experiment of adsorbed NH [22,23]. The
3
◦
X-ray diffraction (XRD) patterns of powder samples were recorded
desorption (TPD) of adsorbed NH was measured by neutralization
on a D8 ADVANCE (Bruker, Germany) using Cu K␣ radiation. The
3
5
0
5
0
SiO -Al O (Fresh)
2
2
3
Beta (Fresh)
-5
-5
-10
-15
-20
-
10
15
20
25
-
-
Beta (Used)
(
a)
(b)
SiO -Al O (Used)
2
2
3
-
-25
0
200 400 600 800 1000
0
200 400 600 800 1000
Temperature /°C
Temperature /°C
5
0
5
0
MFI (Fresh)
MFI (Used)
Al O (Fresh)
2
3
-
5
-5
-
10
15
20
25
-10
-15
-20
-25
-
Al O (Used)
2 3
-
(
c)
(d)
-
0
200 400 600 800 1000
0
200 400 600 800 1000
Temperature /°C
Temperature /°C
Fig. 3. TG profiles of solid acid catalysts before and after the reaction. Reaction conditions: H2 flow rate, 20 cm3 min ; reaction temperature, 250 C; W/F = 26.1 g h mol
−1
.
−1
◦
Please cite this article in press as: D. Sun, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.03.047

G Model
APCATA-15381; No. of Pages9
ARTICLE IN PRESS
D. Sun et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
4
diffuse reflectance infrared Fourier-transform (DRIFT) spectra of
catalyst samples were recorded on a spectroscope using FT/IR-4200
0.4
(a)
(
JASCO, Japan).
Al O -Ag-1-Fresh
2
3
0
0
0
.3
.2
.1
0
Al O -Fresh
2
3
3
. Results and discussion
3.1. Cyclodehydration of DEG over solid acid catalysts
Al
2
O
3
-Ag-1-Used
Table 1 summarizes results of the cyclodehydration of DEG over
Al O -Used
several solid acid catalysts, and Fig. 1 shows the changes in con-
version and selectivity with time on stream in each reaction. The
reactions were performed at a W/F of 26.1 g h mol at 250 C. H
was used as a carrier gas and the flow rate of H was 20 cm min
DOX was the main product in all the reactions, while methanol,
ethanol, acetaldehyde, ethylene glycol, 2-methyl-1,3-dioxolane
2
3
−
1
◦
PEG
× 0.1
2
.
3
−1
2
1
000
2000
Wave number / cm
3000
4000
-
1
(
MDO), and triethylene glycol were generated as the by-products.
0.4
The carbon balance was higher than 90% in all the reactions. The
probable formation routes of DOX and the by-products are shown
in Scheme 2. Methanol, ethanol, acetaldehyde, and ethylene glycol
are the decomposition products of DEG. MDO is proposed to
be generated from acetaldehyde and ethylene glycol through
acetal formation, and triethylene glycol must be produced by
the etherification of DEG with by-produced ethylene glycol. Beta
zeolite gave the highest averaged DEG conversion of 91.9% and the
highest DOX selectivity of 74.6% (Table 1), whereas both of them
decreased steeply with time on stream (Fig. 1). The averaged DEG
conversion and selectivity to DOX was 58.4% and 62.5% over Al O ,
respectively. Although the conversion of DEG decreased with time
on stream over Al O , the selectivity to DOX was stable comparing
with those over the other catalysts. Although MFI zeolite gave
a high averaged DEG conversion of 86.7% and 54.0% selectivity
to DOX, the decrease of DEG conversion was rapid. SiO2 Al O
(b)
MFI-Ag-1-Fresh
MFI-Fresh
0
0
0
.3
.2
.1
0
MFI-Ag-1-Used
MFI-Used
PEG
2
3
× 0.1
2
3
1
000
2000
Wave number / cm
3000
4000
-1
2
3
showed poor catalytic activity on DEG conversion into DOX, and
1
0
.5
1
(
c)
also deactivated with time on stream. The reaction did not proceed
◦
Beta-Ag-1-Fresh
Beta-Fresh
over SiO catalyst at 250 C.
2
Deterioration of acidic catalysts in a vapor-phase is a serious
issue in many cases. For example, in the dehydration of 1,2-
propanediol into propanal, the propanal selectivity exceeds 90%
with a completed conversion over strong acidic catalysts such as
silica-supported silicotungstic acid [24] and ZSM-23 [25], whereas
they are rapidly deactivated with time on stream. In the dehy-
dration of 1,2-propanediol into propanal, silica-supported WO3
catalyst, which has weak acidic property, shows a stable propanal
Beta-Ag-1-Used
Beta-Used
.5
0
◦
selectivity of 93.5% at 250 C when 1,2-propanediol is diluted by
PEG
× 0.1
8
0 wt.% water, but it is also deactivated gradually when pure 1,2-
propanediol is used as the reactant [26]. It is generally recognized
that the carbon accumulation proceeds fast over strong acid cat-
alysts, which always results a rapid catalyst deactivation [22].
Fig. 2 shows the acidic property of catalysts evaluated by NH3-
1000
2000
3000
4000
-1
Wave number / cm
Fig. 4. FTIR spectra of (a) Al2O3 series, (b) MFI series, (c) Beta series. The used
catalysts were prepared under the following reaction conditions: H flow rate,
20 cm min ; reaction temperature, 250 C; W/F = 26.1 g h mol .
TPD measurement. Al O showed a weak acidic property, whereas
2
3
2
3
−1
◦
−1
SiO2 Al O , Beta and MFI zeolites had much larger number of
2
3
medium and strong acid sites than Al O . Thus, the weak acidity
2
3
of Al O would cause a stable selectivity to DOX compared to those
2
3
of SiO2 Al O , Beta and MFI zeolites (Fig. 1b).
2
3
Fig. 3 shows the difference in TG profiles between fresh and used
catalysts. The used samples were prepared by being recovered after
the reaction at 250 C for 5 h. In contrast to the fresh catalyst with a
Thus, it is assumed that carbon is relatively difficult to be accu-
mulated on Al O3 because of the weak acidic property of Al O .
2
2
3
◦
The deterioration behaviors are quite similar to those observed in
the dehydration of 1,2-propanediol [22]. The commercial Al O ,
gradual weight decrease, all the used catalysts had a steep decrease
2
3
◦
in weight at temperatures higher than 200 C. The difference in the
SiO2 Al O and Beta zeolite are also applied for the dehydration of
2 3
weight loss between fresh and used catalysts must attribute to the
carbon deposition [20–22]. The carbon content, which was calcu-
lated by the difference in the TG curves between fresh and used
1,2-propanediol: SiO2 Al O and Beta zeolite with a large number
2 3
of medium and strong acid sites give initial propanal selectivities
higher than 60%, which decreases to ca. 40% after 5 h of the reac-
catalysts, was 16.5, 14.6, 9.8, and 6.8 wt.% in the used SiO2 Al O ,
tion; Al O3 with a small number of weak and medium acid sites
2
3
2
Beta, MFI, and Al O , respectively. The amount of accumulated
shows stable catalytic activity, whereas the selectivity to propanal
2
3
carbon on Al O was the smallest among all the tested catalysts.
is low at ca. 40%.
2
3
Please cite this article in press as: D. Sun, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.03.047

G Model
APCATA-15381; No. of Pages9
ARTICLE IN PRESS
D. Sun et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
5
Table 2
Dehydration of DEG over Ag-modified solid acid catalysts at 250 C.a
◦
Catalyst
Conversion/%b
Selectivity/%b
DOX
MeOH
EtOH
AA
MDO
EG
TEG
Al2O3-Ag-1
MFI-Ag-1
Beta-Ag-1
72.6
67.3
91.9
34.9
71.5
52.3
83.7
30.8
0.1
0.1
0.1
0.1
0.8
0.4
1.0
2.2
1.5
3.6
3.9
7.0
1.4
5.5
2.4
7.1
12.7
17.4
8.1
1.3
1.2
0.3
4.5
SiO2 Al2O3-Ag-1
18.9
The abbreviations of the products are the same as those shown in Table 1.
a
Reaction conditions: W/F = 26.1 g h mol−1; H2 flow rate, 20 cm3 min−1
Average activity in the initial 5 h.
.
b
Fig. 4 shows the DRIFT spectra of original and Ag-modified
Al O , Beta, and MFI catalysts before and after the reaction. Because
gave a lower conversion and a lower selectivity to DOX comparing
2
3
with those over MFI without the loading of Ag O. Ag-modified Beta
2
polyethylene glycol (PEG) is considered as the possible component
of the accumulated carbon, the DRIFT spectra of PEG was measured.
zeolite and SiO2 Al O were still deactivated rapidly with time on
2
3
steam (Fig. 5), while the catalytic activity of Al O -Ag-1 was sta-
2
3
−1
PEG showed a clear peak at 1076 cm , which was originated from
ble. A stable DOX selectivity of 71.5% with a DEG conversion of
72.6% was achieved over Al O -Ag-1. Fig. 6a shows the TG profiles
−
1
the vibration of stretching C O [27–29]. A clear peak at 2881 cm
,
2
3
which was assigned to the stretching vibration of CH2 [29,30],
of Al O -Ag-1 before and after the reaction for 5 h, and the carbon
2
3
−
1
was also observed in PEG. The peaks at 1076 and 2881 cm were
observed in the used catalysts such as original and Ag-modified
Al O , Beta, and MFI, whereas they were observed in the fresh cata-
content in the used Al O -Ag-1 was 5.9%, which is lower than that
2 3
of 6.8% in the used Al O without modification of Ag (Fig. 3d).
2
3
2
3
Fig. 2 compares the acidic property of Ag-modified catalysts
with non-modified ones. The NH3 desorption peak of Al O was
lysts. Thus, it is reasonable that the etherification of DEG proceeded
over the catalyst surface, and PEG is a possible component of the
accumulated carbon.
2
3
◦
◦
observed at 251 C, which shifted to 210 C at Ag O loading of
2
1 wt.%. The loading of Ag O into SiO2 Al O , Beta and MFI zeolites
2
2
3
also decreased their averaged acid strength, respectively. A simi-
lar result is observed in Ag-modified SiO2 Al O in our previous
2
3
3
.2. Effect of Ag loading on Al O3
research dealing with 1,2-propanediol dehydration into propanal
22]: the loading of Ag O onto SiO2 Al O decreases the number
2
[
2
2
3
We have reported several papers dealing with the catalyst deac-
of strong and medium acid sites, whereas increases the number
of weak acid sites. As a result, the catalyst deactivation is inhib-
ited and a stable propanal selectivity of 68.5% is achieved. In this
tivation [20–22]. It is found that the modification of acid catalysts
by loading some metal contents is effective for inhibiting the
catalyst deactivation: Cu-modified Al O exhibits stable catalytic
2
3
study, it is proposed that the additive Ag O associates with the
2
activity with high selectivity to DHP in THFA dehydration [20]; Co-
modified Al O stabilizes the conversion of pinacolone to produce
strong acid sites of each catalyst, and it decreases the number of
the strong acid sites and forms new weak acid sites. Although the
2
3
2
,3-dimethyl-1,3-butadiene [21]; Ag-modified SiO2 Al O gives
averaged acid strengths of SiO2 Al O , Beta and MFI zeolites are
2
3
2 3
stable and increased selectivity to propanal in 1,2-propanediol
dehydration [22]. The metal modifiers work effectively for the sta-
decreased by the loading of Ag O, the numbers of medium and
2
strong acid sites are still much larger than that of Al O -Ag-1. Thus,
2
3
bilization in the reactions only in H flow. In this work, Ag-modified
it can be concluded that the weak acid property of Ag-modified
Al O is effective for inhibiting carbon formation and stabilizing
2
catalysts are estimated to inhibit catalyst deactivation under H2
flow conditions.
2
3
the catalytic activity.
Table 2 shows the averaged reaction data of Ag-modified cata-
lysts, and Fig. 5 depicts the changes in conversion and selectivity
to DOX with time on stream. The loading of Ag O onto Beta, Al O
Table 3 summarizes the DEG cyclodehydration over Ag-
modified Al O3 catalysts at different Ag O loadings. Fig. 7 shows
the changes of DEG conversion and DOX selectivity with time on
stream. The selectivity to EG and TEG decreased by Ag modifica-
tion and the selectivity to DOX increased over the Ag-modified
2
2
2
2
3
and SiO2 Al O was found to be efficient for increasing both the
2
3
conversion of DEG and the selectivity to DOX, whereas MFI-Ag-1
1
00
100
Beta-Ag-1
Al O -Ag-1
Beta-Ag-1
80
8
0
0
0
0
2
3
Al O -Ag-1
6
4
2
60
40
20
2
3
MFI-Ag-1
MFI-Ag-1
SiO -Al O -Ag-1
SiO -Al O -Ag-1
2
2
3
2
2
3
(a)
(b)
0
1
2
3
4
5
0
1
2
3
4
5
Time on stream / h
Time on stream / h
Fig. 5. Changes in conversion and selectivity with time on stream over Ag-modified solid acid catalysts. Reaction conditions: W/F = 26.1 g h mol 1; H2 flow rate, 20 cm3 min
reaction temperature, 250 C.
−
−1
,
◦
Please cite this article in press as: D. Sun, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.03.047

G Model
APCATA-15381; No. of Pages9
ARTICLE IN PRESS
6
D. Sun et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
5
0
5
0
Fresh
Fresh
-
5
-5
-10
-15
-
10
15
20
25
-
H (Used)
2
H +N (Used)
2
2
-
-20
(b)
(
a)
-
-25
0
200 400 600 800 1000
0
200 400 600 800 1000
Temperature /°C
Temperature /°C
5
0
Fresh
-5
-
10
15
20
25
-
-
H +CO (Used)
2 2
(
c)
-
0
200 400 600 800 1000
Temperature /°C
3
−1
◦
Fig. 6. TG profiles of Al2O3-Ag-1 before and after the reaction under different carrier gases. Reaction conditions: flow rate, 20 cm min ; reaction temperature, 250 C;
−1
W/F = 26.1 g h mol
.
Al O catalysts. Because Al O -Ag-1 shows the highest averaged
Ag O loadings. The used Al O -Ag-10 shows a diffraction peak of
2 2 3
2
3
2
3
DEG conversion of 97.2% and a relatively stable DOX selectivity of
2.5%, 1 wt.% loading of Ag O is considered to be suitable. Although
Ag, whereas no peak attributed to Ag O is observed. It is suggested
2
8
that the additive Ag O in the metallic form works to inhibit carbon
2
2
the initial conversion and the initial selectivity to DOX were high
over Ag-modified Al O at Ag O loadings of higher than 3 wt.%, they
accumulation during the reaction.
2
3
2
decreased with time on stream.
3.3. Cyclodehydration of DEG over Al O -Ag-1 at different
2
3
Fig. 8 shows XRD patterns of Ag-modified Al O3 at different
reaction conditions
2
Ag O loadings. No peak attributed to Ag is observed at Ag O load-
2
2
◦
ings lower than 3 wt.%. A diffraction peak of Ag O at 2Â = 33.9
JCPDS file 12-793] and a diffraction peak of Ag at 38.4 [JCPDS file
Table 4 shows the cyclodehydration of DEG over Al O -Ag-1 at
2
2
3
◦
[
different contact times, W/F. The conversion of DEG increased with
increasing the W/F, and a complete conversion was obtained at a
W/F of 156.8 g h mol or higher than. The selectivity to methanol
4
-0783] are observed at Ag O loadings of 5 wt.% or higher. This indi-
2
−
1
cates Ag and Ag O are highly dispersed on catalyst surface at small
2
100
100
Al O -Ag-10
2
3
9
8
7
6
5
0
0
0
0
0
90
80
70
60
50
Al O -Ag-3
2
3
Al O -Ag-1
2
3
Al
O -Ag-1
3
2
Al O -Ag-3
2
3
Al O
2
3
Al O -Ag-10
2
3
3
(
a)
Al O
2
(b)
0
1
2
3
4
5
0
1
2
3
4
5
Time on stream / h
Time on stream / h
Fig. 7. Changes in conversion and selectivity with time on stream over Ag-modified Al2O3 at different Ag loadings. Reaction conditions: W/F = 52.3 g h mol−1; H2 flow rate,
0 cm3 min ; reaction temperature, 250 C.
−1
◦
2
Please cite this article in press as: D. Sun, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.03.047

G Model
APCATA-15381; No. of Pages9
ARTICLE IN PRESS
D. Sun et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
7
Table 3
Dehydration of DEG over Ag-modified Al2O3 at different Ag2O loadings at 250 C.a
◦
Catalyst
Conversion/%b
Selectivity/%b
DOX
MeOH
EtOH
AA
MDO
EG
TEG
Al2O3
84.2
96.0
97.2
96.4
93.2
95.2
69.5
81.2
82.5
82.0
82.7
83.3
0.1
0.1
0.2
0.1
0.1
0.3
1.8
1.7
1.8
1.1
0.9
0.9
1.8
1.5
1.7
1.8
1.7
1.9
1.6
1.3
1.1
1.1
0.9
0.7
12.4
7.4
5.6
6.2
5.5
4.1
1.2
0.1
0.1
0.1
0.3
0.3
Al2O3-Ag-0.5
Al2O3-Ag-1
Al2O3-Ag-3
Al2O3-Ag-5
Al2O3-Ag-10
The abbreviations of the products are the same as those shown in Table 1.
a
Reaction conditions: H2 flow rate, 20 cm3 min−1; W/F = 52.3 g h mol
Average activity in the initial 5 h.
−1
.
b
increase the selectivity to the decomposition products such as
methanol, ethanol and acetaldehyde. However, a low reaction
temperature increases the formation of EG and TEG, as the etherifi-
cation of alcohols prefers low reaction temperatures [31,32]. Since
the contact time also affects the selectivity to the products (Table 4),
the appropriate temperatures for DOX formation are different at
different contact times. The highest selectivity to DOX is 79.7% at a
Ag O
Ag
2
Al O -Ag-10-Used
2
3
Al O -Ag-10
2
3
Al O -Ag-5
2
3
−
1
−1
W/F of 26.1 g h mol , and 82.5% at a W/F of 52.3 g h mol .
Al O -Ag-3
Since H2 carrier gas is indispensable for inhibiting the cata-
lyst deactivation in our precious research [20–22], the effect of
the carrier gas was studied at 250 C (Table 6). The reactions
2
3
Al O -Ag-1
◦
2
3
were performed in N , H + N , H + CO , H + NH and H2 flow
2
2
2
2
2
2
3
Al O
2
3
−
1
over Al O -Ag-1 at a W/F of 26.1 g h mol
and a flow rate of
2
3
3
−1
2
0 cm min
individually. H + NH3 and H + CO2 mixed gases,
2
2
1
0
20
30
40
50
60
70
which are helpful to know the effect of acidic and basis sites on the
reaction, respectively, were used to block the sites of catalyst sur-
face. A high averaged DEG conversion of 72.6% and DOX selectivity
2
θ / degree
of 71.5% were obtained in H flow, and they were relatively stable,
Fig. 8. XRD patterns of Ag-modified Al2O3 at different Ag loadings. Used Al2O3-Ag-
0 was tested under the following reaction conditions: H2 flow rate, 20 cm min
reaction temperature, 250 C; W/F = 26.1 g h mol
2
3
−1
;
as shown in Fig. 5. A significant decrease of DEG conversion and
DOX selectivity was observed when the reaction was performed in
H + NH flow (Table 6). This indicates that acid sites on Al O -Ag-
1
◦
−1
.
2
3
2
3
and ethanol increased with increasing the W/F, whereas the selec-
tivity to EG, TEG and MDO decreased. Thus, the changes in the
selectivity indicate that EG and TEG are difficult to form at a
high contact time. However, EG and TEG could be firstly pro-
duced and then decompose to methanol and ethanol at a high
contact time. The highest selectivity to DOX was obtained at a
1 are the active sites for DEG conversion into DOX. There were no
significant differences on the averaged conversion and the selec-
tivity to DOX in N2, H2, H2 + N2, and H2 + CO2 flow, whereas the
catalyst deactivation in N2, H2 + N2 and H2 + CO2 flow proceeded
faster than that in H2 flow. The decreases in DEG conversion from
1 h to 5 h were 16.3, 15.7, 22.8 and 6.3% in N2, H2 + N2, H2 + CO2 and
H2 flow, respectively. The decreases in the selectivity to DOX from
1 h to 5 h were 5.3, 4.7, 10.3 and 1.9% in N2, H2 + N2, H2 + CO2 and H2
flow, respectively. The carbon content was calculated to be 9.8 and
11.9% in Al2O3-Ag-1 used in H2 + N2 and H2 + CO2 flow, respectively
(Fig. 6b and c). Those carbon contents are much higher than 5.9%
observed over Al2O3-Ag-1 in H2 flow (Fig. 6a).
−1
W/F of 156.8 g h mol . Table 4 also shows the averaged reaction
results in the initial 10 h over Al O -Ag-1 and Al O at a W/F of
2
3
2
3
−1
1
56.8 g h mol , Fig. 9 shows the changes of DEG conversion and
DOX selectivity with time on stream. It is obvious that a change
in the selectivity to DOX with time on steam is more stable over
Al O -Ag-1 than that over Al O .
2
3
2
3
Table 5 shows the effect of reaction temperature over Al O -Ag-
at a W/F of 26.1 and 52.3 g h mol . High reaction temperatures
In our previous researches dealing with catalyst deactivation,
the metal-loaded solid acid catalysts show stable catalytic activity
2
3
−1
1
Table 4
Effect of W/F on dehydration of DEG over Al2O3-Ag-1 at 250 C.a
◦
W/F/g h mol 1
−
Conversion/%b
Selectivity/%b
DOX
MeOH
EtOH
AA
MDO
EG
TEG
2
5
6.1
2.3
72.6
97.2
100
100
100
71.5
82.5
90.0
87.7
90.0
87.1
0.1
0.2
0.7
1.5
0.5
0.4
0.8
1.8
3.5
4.2
3.2
3.6
1.5
1.7
2.0
1.7
2.3
2.0
1.4
1.1
0.2
0.3
0.2
0.7
12.7
5.6
0.5
0.2
0.7
2.8
1.3
0.1
0.0
0.0
0.0
0.0
1
2
1
1
56.8
61.4
56.8
c
56.8d
100
The abbreviations of the products are the same as those shown in Table 1.
a
Reaction conditions: H2 flow rate, 20 cm3 min−1
Average activity in the initial 5 h.
Average activity in the initial 10 h.
.
b
c
d
The reaction was performed over pure Al2O3 and the data was averaged in the initial 10 h.
Please cite this article in press as: D. Sun, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.03.047

G Model
APCATA-15381; No. of Pages9
ARTICLE IN PRESS
8
D. Sun et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
Table 5
a
Dehydration of DEG over Al2O3-Ag-1 at different reaction temperatures.
W/F/g h mol 1
−
◦
Temperature/ C
Conversion/%b
Selectivity/%b
DOX
MeOH
EtOH
AA
MDO
EG
TEG
2
275
50
72.6
98.1
97.8
71.5
79.7
71.4
0.1
0.3
1.0
0.8
3.9
5.8
1.5
3.4
8.6
1.4
1.2
1.3
12.7
5.5
4.2
1.3
0.0
0.0
2
6.1
2.3
3
00
50
2
97.2
97.9
100
82.5
76.7
74.0
0.2
0.8
1.7
1.8
4.6
6.7
1.7
4.2
8.0
1.1
0.7
0.4
5.6
3.3
0.7
0.1
0.0
0.0
5
275
00
3
The abbreviations of the products are the same as those shown in Table 1.
a
Reaction conditions: H2 flow rate, 20 cm3 min−1
Average activity in the initial 5 h.
.
b
1
00
100
90
Conversion
Conversion
9
0
Selectivity to 1,4-dioxane
Selectivity to 1,4-dioxane
(
2
a) Al O3
(b) Al O -Ag-1
2 3
8
0
8
0
0
2
4
6
8
10
0
2
4
6
8
10
Time on stream / h
Time on stream / h
Fig. 9. Changes in conversion and selectivity with time on stream over Al2O3 and Al2O3-Ag-1 at a W/F of 156.8 g h mol 1. Reaction conditions: H2 flow rate, 20 cm3 min
reaction temperature, 250 C.
−
−1
;
◦
Table 6
Dehydration of DEG over Al2O3-Ag-1 at different carrier gas flows at 250 C.a
◦
Carrier gas
Conversion/%b
Selectivity/%b
DOX
MeOH
EtOH
AA
MDO
EG
TEG
N2
68.8
70.0
68.0
13.8
72.6
73.8
73.7
74.4
33.0
71.5
0.0
0.0
0.1
0.3
0.1
0.5
0.6
0.4
0.6
0.8
1.7
1.8
2.1
4.0
1.5
1.2
1.3
0.9
0.1
1.4
10.5
10.4
9.3
19.9
12.7
1.5
1.6
1.3
4.4
1.3
H2 + N2
H2 + CO2
H2 + NH3
H2
The abbreviations of the products are the same as those shown in Table 1.
a
Reaction conditions: W/F = 26.1 g h mol−1, total carrier gas flow rate, 20 cm3 min−1; the flow rate of each gas in the mixed gas flow was 10 cm3 min−1
Average activity in the initial 5 h.
.
b
only in H flow [20–22]. In this study, the reaction performed in H2
flow gives the highest DEG conversion, the most stable selectivity
Beta and MFI zeolites. Strong acidic catalysts such as Beta zeolite
showed high catalytic activity, whereas deactivated rapidly with
2
to DOX and the smallest amount of accumulated carbon. This agrees
with the results of our previous researches [20–22], in which we
have concluded that the added metals work as a remover of product
time on stream. Al O3 gave a relatively stable selectivity to DOX,
2
but the conversion of DEG gradually decreased with time on stream.
Metal-modified catalysts were prepared for inhibiting the cata-
lyst deactivation and Al O modified with 1 wt.% Ag O showed the
from the catalyst surface together with H to prevent carbon depo-
2
2
3
2
sition. It is assumed that Ag also plays the same role of the added
best catalytic performance. High and stable DEG conversion and
DOX selectivity can be achieved over the Ag-modified Al O under
metal: Ag metals would dissociate H into hydrogen atoms, which
2
2
3
move to the surface of Al O , which can be explained as spillover
H2 flow. The addition of Ag O onto Al O decreased the averaged
2
3
2 2 3
effect. In other words, Ag together with H absorbed on the support
acid strength of Al O , which is proposed to be effective for inhib-
2 3
2
would assist desorption of the products before the polymerization
of the products and/or carbonization of the polymer proceeds.
iting carbon deposition. The catalytic activity of the Ag-modified
Al O under H flow is more stable than that in the other gas flow,
2
3
2
so that H2 carrier was indispensable for stabilizing the conver-
sion of DEG into DOX. It is proposed that Ag works as a remover
of the product adsorbed on the catalyst surface together with H2
to prevent carbon deposition. A stable DOX selectivity of 90% with
4
. Conclusions
a complete conversion was achieved over the Ag-modified Al O
Vapor-phase cyclodehydration of DEG into DOX was firstly per-
2
3
◦
−1
at 250 C and a W/F of 156.8 g h mol under H flow.
formed over several solid acid catalysts such as Al O , SiO Al O ,
2
2
3
2
2
3
Please cite this article in press as: D. Sun, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.03.047

G Model
APCATA-15381; No. of Pages9
ARTICLE IN PRESS
D. Sun et al. / Applied Catalysis A: General xxx (2015) xxx–xxx
9
References
[15] S. Sato, R. Takahashi, N. Yamamoto, E. Kaneko, H. Inoue, Appl. Catal. A: Gen. 334
2008) 84–91.
(
[
[
16] F. Sato, H. Okazaki, S. Sato, Appl. Catal. A: Gen. 419–420 (2012) 41–48.
17] A. Wali, S.M. Pillai, S. Unnikrishnan, S. Satish, J. Mol. Catal. A: Chem. 109 (1996)
[
[
1] A. Wali, S.M. Pillai, J. Chem. Res. Synop. (1999) 326–327.
2] H. Li, H. Yin, T. Jiang, T. Hu, J. Wu, B. Jiang, Y. Wada, React. Kinet. Catal. Lett. 88
(
149–155.
2006) 233–241.
3] H. Li, H. Yin, T. Jiang, T. Hu, J. Wu, Y. Wada, Catal. Commun. 7 (2006)
78–782.
[
[
18] Y. Wang, W. Li, L. Dai, Chem. Pap. 62 (2008) 313–317.
19] Y. Kitayama, M. Kamimura, K. Wakui, M. Kanamori, T. Kodama, J. Abe, J. Mol.
Catal. A: Chem. 142 (1999) 237–245.
20] S. Sato, J. Igarashi, Y. Yamada, Appl. Catal. A: Gen. 453 (2013) 213–218.
21] S. Sato, N. Sato, Y. Yamada, Chem. Lett. 41 (2012) 831–833.
[
7
[
[
4] X. Cao, J. Ren, M. Liu, C. Zhan, B. Xie, Chin. J. Org. Chem. 33 (2013) 1035–1041.
5] S.H. Vaidya, V.M. Bhandari, R.V. Chaudhari, Appl. Catal. A: Gen. 242 (2003)
[
[
[
[
[
[
[
[
[
3
21–328.
6] N. Yamamoto, S. Sato, R. Takahashi, K. Inui, J. Mol. Catal. A: Chem. 243 (2006)
2–59.
22] D. Sun, R. Narita, F. Sato, Y. Yamada, S. Sato, Chem. Lett. 43 (2014) 450–452.
23] S. Sato, K. Koizumi, F. Nozaki, J. Catal. 178 (1998) 264–274.
[
5
24] K. Mori, Y. Yamada, S. Sato, Appl. Catal. A: Gen. 366 (2009) 304–308.
25] D. Zhang, S.A.I. Barri, D. Chadwick, Appl. Catal. A: Gen. 400 (2011) 148–155.
26] D. Sun, Y. Yamada, S. Sato, Appl. Catal. A: Gen. 487 (2014) 234–241.
27] W. Chen, Z. Zhu, S. Li, C. Chen, L. Yan, Nanoscale 4 (2012) 2124–2129.
28] T. Szabó, O. Berkesi, P. Forgó, K. Josepovits, Y. Sanakis, D. Petridis, Chem. Mater.
[
[
7] Q. Zhang, Y. Zhang, H. Li, C. Gao, Y. Zhao, Appl. Catal. A: Gen. 466 (2013) 233–239.
8] M. Segawa, S. Sato, M. Kobune, T. Sodesawa, T. Kojima, S. Nishiyama, N.
Ishizawa, J. Mol. Catal. A: Chem. 310 (2009) 166–173.
[9] R. Takahashi, I. Yamada, A. Iwata, N. Kurahashi, S. Yoshida, S. Sato, Appl. Catal.
A: Gen. 383 (2010) 134–140.
18 (2006) 2740–2749.
[
[
10] S. Sato, F. Sato, H. Gotoh, Y. Yamada, ACS Catal. 3 (2013) 721–734.
11] S. Sato, R. Takahashi, M. Kobune, H. Inoue, Y. Izawa, H. Ohno, K. Takahashi, Appl.
Catal. A: Gen. 356 (2009) 64–71.
12] C.P. Bezouhanova, F.A. Jabur, React. Kinet. Catal. Lett. 51 (1993) 177–181.
13] S.M. Patel, U.V. Chudasama, P.A. Ganeshpure, Green Chem. 3 (2001) 143–145.
14] Y. Inoue, S. Deguchi, T. Hakushi, Bull. Chem. Soc. Jpn. 53 (1980) 3031–3032.
[
29] Y. Yamada, H. Yasuda, K. Murota, M. Nakamura, T. Sodesawa, S. Sato, J. Mater.
Sci. 48 (2013) 8171–8198.
30] H. Sun, Y. Yang, Q. Huang, Integr. Ferroelectr. 128 (2011) 163–170.
31] F.M. Bautista, B. Delmon, Appl. Catal. A: Gen. 130 (1995) 47–65.
32] R.J.J. Nel, A. de Klerk, Ind. Eng. Chem. Res. 48 (2009) 5230–5238.
[
[
[
[
[
[
Please cite this article in press as: D. Sun, et al., Appl. Catal. A: Gen. (2015), http://dx.doi.org/10.1016/j.apcata.2015.03.047