Stability enhancement of H-mordenite in dimethyl ether carbonylation to methyl acetate by pre-adsorption of pyridine

Article abstract of DOI:10.1016/S1872-2067(09)60081-4

The carbonylation of dimethyl ether to methyl acetate over H-mordenite (HMOR) and pyridine-modified HMOR was compared. The catalytic stability of HMOR was improved significantly by pyridine pre-adsorption, and a yield of methyl acetate ～30 was still obtained after 48 h on stream at 473 K. In situ infrared spectroscopy and ammonia temperature-programmed desorption revealed that pyridine preferentially occupied the acidic sites in 12-membered ring pores but not the acidic sites in 8-membered ring pores. ¹²⁹Xe NMR studies suggested that the channels of HMOR were blocked by coke in the reaction but those in the pyridine-modified HMOR were not. The acidic sites in the 12-membered ring pores were responsible for the deactivation of HMOR, and the reaction can be directed to occur mainly on the acidic sties in the 8-membered ring pores by the selective adsorption of pyridine in the 12-membered ring pores.

Full text of DOI:10.1016/S1872-2067(09)60081-4

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
CHINESE JOURNAL OF CATALYSIS
Volume 31, Issue 7, 2010
Online English edition of the Chinese language journal
RESEARCH PAPER
Cite this article as: Chin J Catal, 2010, 31: 729–738.
Stability Enhancement of H-Mordenite in Dimethyl Ether
Carbonylation to Methyl Acetate by Pre-adsorption of
Pyridine
LIU Junlong1, XUE Huifu1, HUANG Xiumin1, WU Pei-Hao2, HUANG Shing-Jong3, LIU Shang-Bin2,*,
SHEN Wenjie1,#
1State Key Laboratory of Catalysis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China
2Institute of Atomic and Molecular Sciences, “Academia Sinica”, Taipei 10617, Taiwan, China
3Department of Chemistry, Taiwan University, Taipei 10617, Taiwa, China
Abstract: The carbonylation of dimethyl ether to methyl acetate over H-mordenite (HMOR) and pyridine-modified HMOR was compared.
The catalytic stability of HMOR was improved significantly by pyridine pre-adsorption, and a yield of methyl acetate ~30% was still ob-
tained after 48 h on stream at 473 K. In situ infrared spectroscopy and ammonia temperature-programmed desorption revealed that pyridine
preferentially occupied the acidic sites in 12-membered ring pores but not the acidic sites in 8-membered ring pores. 129Xe NMR studies
suggested that the channels of HMOR were blocked by coke in the reaction but those in the pyridine-modified HMOR were not. The acidic
sites in the 12-membered ring pores were responsible for the deactivation of HMOR, and the reaction can be directed to occur mainly on
the acidic sties in the 8-membered ring pores by the selective adsorption of pyridine in the 12-membered ring pores.
Key words: dimethyl ether; carbonylation; methyl acetate; H-mordenite; pyridine; stability
The present industrial production of acetic acid is mainly by
methanol carbonylation in the Monsanto or BP CativaTM ho-
mogeneous processes, which use rhodium or iridium catalysts
and methyl iodide as promoter [1]. Due to concerns over en-
vironmental pollution and production cost, the trend in recent
methanol carbonylation is to avoid the use of corrosive hal-
ides and expensive precious metals and to rely on a heteroge-
neous process using solid acid catalysts [2–4]. For example,
Y-type zeolites and sulfated zirconia catalysts have been in-
vestigated for the vapor phase carbonylation of methanol [2],
but the main product is dimethyl ether (DME) with only trace
amounts of acetic acid and hydrocarbons at 473 K. Upon in-
creasing the temperature to 573 K, the yield of hydrocarbons
was increased, but the yield of acetic acid was increased only
marginally. On the other hand, over H-ZSM-5 and
H-mordenite (HMOR), acetic acid was formed with a yield of
0.4% at 473 K [2]. The catalytic performance of these zeolites
could be improved by incorporating copper, which gave about
2% methyl acetyl yield at 523 K [2]. However, these solid
acids suffered rapid deactivation by coking, especially at ele-
vated temperatures [3]. Keggin polyoxometallate clusters
containing Rh or Ir were also identified as active catalysts for
methanol carbonylation with a methyl acetate yield of ca. 30%
at 493 K, but this decreased to 16% when the methanol feed
was replaced by DME [3–6]. Again, these catalysts were also
found to form hydrocarbons, leading to severe deactivation by
the deposition of carbonaceous residues.
Iglesia and co-workers [7–10] recently reported that HMOR
and H-ferrierite (HFER) are good catalysts for the carbonyla-
tion of DME to methyl acetate (MA), and that HMOR is the
more active and selective catalyst. A common feature of these
two types of zeolites is the existence of 8-member ring (8-MR)
channels, in which the reaction mainly occurs [9]. Kinetic and
spectroscopic studies confirmed that the carbonylation of DME
involves an interaction between the reactant and surface hy-
droxyl (–OH) groups on the zeolite catalyst to form methoxy
Received date: 24 December 2009.
*Corresponding author. Tel: +886-2-2366-8230; Fax: +886-2-23620200; E-mail: sbliu@sinica.edu.tw
#Corresponding author. Tel: +86-411-84379085; Fax: +86-411-84694447; E-mail: shen98@dicp.ac.cn
Foundation item: Supported by the National Natural Science Foundation of China (20973166).
Copyright © 2010, Dalian Institute of Chemical Physics, Chinese Academy of Sciences. Published by Elsevier BV. All rights reserved.
DOI: 10.1016/S1872-2067(09)60081-4

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
LIU Junlong et al. / Chinese Journal of Catalysis, 2010, 31: 729–738
species, which was followed by the insertion of CO molecules
to form acetyls in the rate-determining step. These acetyls
further react with DME to produce MA and regenerate the
surface methoxy species [7–9,11]. When DME was replaced
by methanol, however, the reaction was slow due to the strong
inhibition by water generated by methanol dehydration to
DME. Theoretical calculations showed that the unique cata-
lytic properties observed for these zeolites with 8-MR chan-
nels are due to the size of the pore where the active sites were
located and also the unique orientation of the methoxy groups
in the pore channels [12]. More importantly, the transition state
during the insertion of CO into the methoxide fitted perfectly
in the 8-MR channels. The negative effect of water on the
carbonylation reaction rate was attributed to their strong ad-
sorption on the acidic sites, which reduced available methoxy
groups for CO.
dure yielded a HMOR-10 catalyst with a Si/Al ratio of 9.6.
To prepare the pyridine-modified catalysts, 600 mg HMOR
(40–60 mesh) was loaded into a stainless steel fixed-bed re-
actor (8 mm i.d.), heated to 773 K at a rate of 10 K/min under
N2 flow (30 ml/min), and maintained at this temperature for 1
h. After cooling to 573 K, the sample was purged with a 1.3%
pyridine-98.7% N2 mixture (30 ml/min) for 1 h and then
flushed with N2 (30 ml/min) for 1 h, which yielded the
Py-HMOR catalysts.
1.2 Catalyst characterization
Field-emission scanning electron microscopy (FESEM)
images were recorded using a Philips Fei Quanta 200F in-
strument operated at 20 kV. The samples were placed on a
conductive carbon tape on an aluminum sample holder.
These pioneering and fundamental studies give the reaction
mechanism and role of the unique 8-MR channels in DME
carbonylation. However, the stability of the HMOR catalyst,
which is important to the industrial application of this process,
was not much discussed. It is well known that the addition of a
small amount of amine to the feedstream can significantly
promote the selectivity in various processes by poisoning the
strong framework Brönsted acid sites of the zeolites, e.g., in the
epoxidation of cyclohexanol [13]. It should be noted that the
addition of amine usually improved the selectivity of the reac-
tion but decreased the catalytic activity considerably [14,15].
In this work, we report a procedure to improve the stability of
HMOR while keeping a high MA yield during DME carbon-
ylation. This procedure was the use of the pre-adsorption of
pyridine on HMOR, which led to preferential blocking of the
acidic sites in the 12-MR pores while leaving the acidic sites in
8-MR pores undisturbed. Consequently, the pyridine-modified
HMOR catalyst showed significantly enhanced stability for
DME carbonylation with MA yield exceeding 30% for at least
48 h on stream at 473 K.
Fourier transform infrared (FT-IR) spectra were recorded on
a Bruker Vector-22 instrument with a resolution of 2 cm–1. The
samples were pressed into a self-supporting wafer and evacu-
ated (2×10–2 Pa) in the IR cell at 723 K for 5 h. Adsorption of
pyridine was conducted at 298–573 K for 5 min to ensure a
saturated loading prior to the acquisition of the IR spectrum at
room temperature. All spectra were normalized using the
overtone and combination vibrations of HMOR zeolite at 1 876
and 1 985 cm–1. In situ IR spectra were obtained using a high
temperature and pressure cell. The sample was pressed into a
self-supporting disk and mounted into the cell. It was then
purged with He at 723 K for 4 h. The adsorption of MA was
conducted at room temperature, and the catalyst sample was
then heated in a He flow. DME carbonylation was performed
by introducing a 5% DME-50% CO-2.5% N2-42.5% He mix-
ture at 473 K with the cell pressurized to 1.0 MPa. Spectra were
then recorded as a function of reaction time.
Temperature-programmed
desorption
of
ammonia
(NH3-TPD) experiments were carried out using a U-shape
quartz tube reactor on a chemisorption apparatus (AutoChem II
2920). For each run, 100 mg sample was heated to 473 K at a
rate of 10 K/min, maintained at the this temperature for 1 h
under N2 flow (30 ml/min), and then purged with a 10%
NH3-90% He gas mixture (30 ml/min) at 473 K for 30 min.
Subsequently, physisorbed ammonia was removed by purging
the sample with a 0.6% H2O-99.4% N2 mixture (30 ml/min) at
473 K for 1 h. After cooling to room temperature, the sample
was heated to 873 K at a rate of 10 K/min under He flow (20
ml/min). The outlet gas was collected after passing through a
solid KOH trap to remove water and a dry ice trap to trap
pyridine. The amount of desorbed NH3 was then measured by a
thermal conductivity detector (TCD).
1 Experimental
1.1 Catalyst preparation
A commercially available Na-mordenite (NaMOR) with a
Si/Al ratio of 6.4 and surface area of 434 m2/g was used as the
precursor. The HMOR sample was obtained from NaMOR by
ion exchange. NaMOR (100 g) was dispersed in 1 L of
NH4NO3 aqueous solution (1 mol/L) at 353 K for 3 h followed
by filtration and washing with distilled water. After repeating
the ion exchange three times, the resulting solid was dried at
383 K for 12 h and then calcined at 773 K for 6 h in air, which
yielded the HMOR-6 catalyst. In order to adjust the Si/Al ratio,
the HMOR-6 sample was then treated with a 2 mol/L HNO3
solution (40 ml/g) at 373 K for 10 h followed by drying at 383
K for 12 h and calcination at 773 K for 6 h in air. This proce-
129Xe NMR spectra of xenon adsorption on the samples
were recorded on a Bruker Avance-300 instrument operated at
a Larmor frequency of 83.012 MHz using a single pulse se-
quence with a π/2 pulse of 15 μs and a recycle delay of 2 s.
Typically, 128 000 free induction delay signals were accumu-

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
LIU Junlong et al. / Chinese Journal of Catalysis, 2010, 31: 729–738
lated for each spectrum. Diluted Xe gas was used as the
chemical shift reference. The details of the experimental pro-
cedure can be found elsewhere [16].
flow for 1 h. After cooled to 573 K, the sample was purged
with a 1.3% pyridine-98.7% N2 mixture (30 ml/min) for 1 h,
and then flushed with N2 for another 1 h before carrying out
the DME carbonylation test.
Temperature-programmed oxidation (TPO) measurements
of the samples were conducted with a U-shape micro-reactor
connected to a mass spectrometer (QMS-200, Omnistar). Be-
fore the experiment, the sample (100 mg) was pre-treated at
473 K for 1 h under He flow (40 ml/min). After cooling to
room temperature, the sample was heated to 1 073 K at a rate of
10 K/min in a flowing 20% O2-80% He mixture (60 ml/min).
The effluent was monitored by a mass spectrometer.
2 Results and discussion
2.1 DME carbonylation
Figure 1(a) compares the conversions of DME and product
selectivities during DME carbonylation over the HMOR-6 and
Py-HMOR-6 catalysts as a function of time on stream (TOS).
For the HMOR-6 catalyst, an induction period of 2.5 h was
observed during which the conversion of DME reached 37%.
The conversion of DME was maintained at this level for an
additional 2.5 h, followed by a rapid deactivation to a conver-
sion of DME of only 6% at 18 h on stream. The selectivity for
MA was more than 98% for about 5 h, but it rapidly decreased
to only 50% at 18 h on stream. On the other hand, the selec-
tivities for methanol and hydrocarbons (HCs) were practically
zero during the initial stage, but they increased significantly to
35% and 14%, respectively, at 18 h on stream. This reaction
pattern indicated that the typical HMOR lost activity due to
coking during the course of reaction and that the deposition of
carbonaceous residues (cokes) in the pore channels inhibited
the formation of MA but promoted the formation of significant
amounts of methanol and hydrocarbons.
1.3 Catalytic test
DME carbonylation was conducted with a continuous flow
fixed-bed stainless steel reactor. In the case of HMOR, 600 mg
of catalyst (40–60 mesh) was loaded into the reactor, heated to
773 K at a rate of 10 K/min under N2 flow (30 ml/min) and
maintained at this temperature for 1 h. After the catalyst sample
was cooled to 473 K, a 5% DME-50% CO-2.5% N2-42.5% He
mixture was introduced at a gas hourly space velocity (GHSV)
of 1 250 ml/(g·h) and the reactor was pressurized to 1.0 MPa.
The effluent was analyzed by an online gas chromatograph
(Agilent 6890N) equipped with a TCD and a flame ionization
detector (FID). N2 was used as an internal standard.
Similar reaction conditions were applied to the Py-HMOR
catalysts. The HMOR was initially heated at 773 K under N2
40
20
HMOR-10
Py-HMOR-10
(b)
30
15
10
5
Py-HMOR-6
20
(a)
HMOR-6
10
0
0
100
100
Py-HMOR-6
Py-HMOR-10
80
HMOR-6
90
80
HMOR-10
HMOR-10
60
40
30
20
10
20
HMOR-6
10
Py-HMOR-10
Py-HMOR-6
0
4
0
15
HMOR-10
10
HMOR-6
2
0
5
0
0
10
20
30
40
50
0
10
20
30
40
50
Time on stream (h)
Time on stream (h)
Fig. 1. Conversion of DME and selectivities for MA, methanol, and HCs during DME carbonylation over the HMOR-6 and Py-HMOR-6 catalysts (a)
and HMOR-10 and Py-HMOR-10 catalysts (b). Reaction conditions: 473 K, 5% DME-50% CO-2.5% N2-42.5% He, 1250 ml/(g·h), 1.0 MPa.

Evaluation Only. Created with Aspose.PDF. Copyright 2002-2021 Aspose Pty Ltd.
LIU Junlong et al. / Chinese Journal of Catalysis, 2010, 31: 729–738
For the Py-HMOR-6 catalyst, an increased induction period
to 11 h was observed. The conversion of DME gradually in-
creased to 33% and was maintained at this level for 48 h on
stream without deactivation. The selectivity for MA increased
from 90% in the initial stage to a plateau value of 99% while
the selectivity for methanol decreased progressively from 10%
in the initial stage to less than 1% at extended TOS. Interest-
ingly, unlike over the HMOR-6 catalyst, there was no detect-
able formation of hydrocarbons over the Py-HMOR-6 catalyst.
Apparently, the pre-treatment by the adsorption of pyridine
significantly enhanced the stability of HMOR during DME
carbonylation. Long-term stability test further revealed that a
rather stable performance was obtained over the Py-HMOR-6
catalyst with MA selectivity of more than 99% while the con-
version of DME only slightly decreased from 33% to 20% after
250 h on stream.
8-MR pores as the active sites for DME carbonylation. How-
ever, the induction period was prolonged over Py-HMOR-6
because of the diffusion limitation of pyridine adsorption in the
large channels.
Similar reaction patterns were observed on the HMOR-10
and Py-HMOR-10 catalysts except for their lower DME con-
versions, as illustrated in Fig. 1(b). Over the HMOR-10 cata-
lyst, the conversion of DME approached 16% during the in-
duction period of about 3 h. It was maintained at this level for 4
h followed by a rapid deactivation with the formation of sig-
nificant amounts of methanol and HCs. However, this deacti-
vation rate was much lower than that over the HMOR-6 cata-
lyst, and the conversion of DME only decreased to 9% after 18
h on stream. Again, the Py-HMOR-10 catalyst exhibited con-
siderably enhanced stability. The conversion of DME gradually
increased to about 13% during the induction period of 13 h and
it was maintained at this level for 50 h on stream. MA was
produced as the main product with a selectivity of 99% while
the selectivity for methanol was less than 1%. Apparently,
acidic leaching (dealumination) of the HMOR-6 sample de-
creased the Brönsted acid sites in the 8-MR channels and
caused a remarkable loss in activity. Simultaneously, it also
eliminated the acidic sites in the 12-MR channels, which gave a
slow deactivation rate. For the two HMOR samples with dif-
ferent Si/Al ratios, pyridine pre-adsorption in the 12-MR pores
significantly improved the stability during DME carbonylation.
However, although it caused diffusion limitation of the reac-
tants and products as indicated by the extended induction pe-
riod.
It was previously proposed that DME reacts with the Brön-
sted acid sites of HMOR during the induction period to form
surface methoxy species and methanol, followed by the inser-
tion of CO into the C–O bond of the tertiary surface-bound
methyl to produce acetyl [7–9]. This reaction intermediate then
reacted with DME to produce methyl acetate while regenerat-
ing a surface methyl group to complete the catalytic cycle. As a
result, MA was formed as the sole product at steady state,
typically with a superior selectivity of more than 99%. How-
ever, over the HMOR-6 catalyst, both the conversion of DME
and selectivity for MA declined significantly after 5 h on
stream, whereas the selectivity for methanol increased to 35%
due to the rapid deactivation by coking. Presumably, the sig-
nificant amounts of water produced together with hydrocar-
bons led to the hydration of DME to methanol. In this context,
the Py-HMOR-6 catalyst exhibited a better catalytic perform-
ance and stability than the HMOR-6, which indicated that the
pre-adsorption of pyridine effectively inactivated the acidic
sites in the 12-MR pore channels and left the acidic sites in the
2.2 FESEM and NH3-TPD
Figure 2 shows the FESEM images of the HMOR and
Py-HMOR catalysts. The HMOR-6 sample exhibited a tradi-
tional block-like morphology with the size of ~5 μm in length
Fig. 2. FESEM images of the HMOR-6 (a, b), Py-HMOR-6 (c, d), HMOR-10 (e, f), and Py-HMOR-10 (g, h) samples.