Mechanisms of the deactivation of SAPO-34 materials with different crystal sizes applied as MTO catalysts

Article abstract of DOI:10.1021/cs400007v

SAPO-34 materials with comparable Bronsted acid site density but different crystal sizes were applied as methanol-to-olefin (MTO) catalysts to elucidate the effect of the crystal size on their deactivation behaviors. ¹³C HPDEC MAS NMR, FTIR, and UV/vis spectroscopy were employed to monitor the formation and nature of organic deposits, and the densities of accessible Bronsted acid sites and active hydrocarbon-pool species were studied as a function of time-on-stream (TOS) by ¹H MAS NMR spectroscopy. The above-mentioned spectroscopic methods gave a very complex picture of the deactivation mechanism consisting of a number of different steps. The most important of these steps is the formation of alkyl aromatics with large alkyl chains improving at first the olefin selectivity, but hindering the reactant diffusion after longer TOS. The hindered reactant diffusion leads to a surplus of retarded olefinic reaction products in the SAPO-34 pores accompanied by their oligomerization and the formation of polycyclic aromatics. Finally, these polycyclic aromatics are responsible for a total blocking of the SAPO-34 pores, making all catalytically active sites inside the pores nonaccessible for further reactants.

Full text of DOI:10.1021/cs400007v

Research Article
pubs.acs.org/acscatalysis
Mechanisms of the Deactivation of SAPO-34 Materials with Different
Crystal Sizes Applied as MTO Catalysts
Weili Dai, Guangjun Wu, Landong Li,* Naijia Guan, and Michael Hunger*
†
†
,†
†
,‡
†Key Laboratory of Advanced Energy Materials Chemistry (Ministry of Education), College of Chemistry, Nankai University, Tianjin
3
00071, People’s Republic China
‡
Institute of Chemical Technology, University of Stuttgart, 70550 Stuttgart, Germany
S Supporting Information
*
ABSTRACT: SAPO-34 materials with comparable Brønsted
acid site density but different crystal sizes were applied as
methanol-to-olefin (MTO) catalysts to elucidate the effect of
the crystal size on their deactivation behaviors. 13C HPDEC
MAS NMR, FTIR, and UV/vis spectroscopy were employed
to monitor the formation and nature of organic deposits, and
the densities of accessible Brønsted acid sites and active
hydrocarbon-pool species were studied as a function of time-
1
on-stream (TOS) by H MAS NMR spectroscopy. The above-
mentioned spectroscopic methods gave a very complex picture
of the deactivation mechanism consisting of a number of
different steps. The most important of these steps is the formation of alkyl aromatics with large alkyl chains improving at first the
olefin selectivity, but hindering the reactant diffusion after longer TOS. The hindered reactant diffusion leads to a surplus of
retarded olefinic reaction products in the SAPO-34 pores accompanied by their oligomerization and the formation of polycyclic
aromatics. Finally, these polycyclic aromatics are responsible for a total blocking of the SAPO-34 pores, making all catalytically
active sites inside the pores nonaccessible for further reactants.
KEYWORDS: methanol-to-olefin conversion, SAPO-34, crystal size, deactivation mechanism, Brønsted acid sites,
benzene-based carbenium ions, in situ spectroscopy
1
. INTRODUCTION
clarifying the reasons of the deactivation, extensive attention
has been paid to controlling the Brønsted acidity of SAPO-34
As an alternative process for obtaining olefins, methanol-to-
olefin (MTO) conversion on microporous solid acid catalysts
has attracted extensive attention since it was disclosed in the
6,7,22−25
to obtain a highly active catalyst with a long lifetime.
In
addition to the acidity, the crystal size of SAPO-34 is also an
important factor that may influence the catalytic performance,
especially the catalyst lifetime in the MTO, reaction due to
coke formation. SAPO-34 with small crystal size has been
reported to exhibit a prolonged lifetime in MTO conversion
1
−5
1
970s.
During recent decades, a variety of molecular sieves
6
−17
have been explored as possible MTO catalysts,
generally acknowledged that the MTO performances of
molecular sieves are controlled both by their framework
and it is
26−30
2
due to its good resistance to deactivation by coke.
structure and by their acidity. The silicoaluminophosphate
However, to the best of our knowledge, detailed spectroscopic
investigations on the deactivation mechanisms of SAPO-34
catalysts in the MTO process in relation to dependence on the
crystal size and the Brønsted acidity have been missing until
now and are, therefore, the goal of the present work.
In this study, SAPO-34 materials with different crystal sizes
were synthesized and characterized by means of XRD, SEM,
and solid-state NMR spectroscopy. The MTO performances of
these materials were evaluated in a fixed-bed reactor, and the
organic intermediates formed during MTO reaction were
analyzed by in situ FTIR and UV/vis spectroscopy. Specifically,
the dynamic changes in the Brønsted acid sites and benzene-
molecular sieve SAPO-34 with Brønsted acid sites of moderate
acid strength and a well-defined framework with chabazite cages
connected via 8-ring windows is reported to be the most
promising catalyst, giving a maximum light olefin (propene and
ethene) yield of ∼80%. The catalytic and mechanistic aspects
of the MTO reaction on SAPO-34 have been extensively
studied (for an overview, see refs 3−5). According to the well
accepted hydrocarbon pool mechanism, polyalkyl aromatics
confined inside the cages of SAPO-34 are active reaction
intermediates leading to the formation of light olefins upon
activation by Brønsted acid sites at the framework. The small 8-
1,2
́
ring windows of ∼3.8 Å restrict the diffusion of heavy and
branched hydrocarbons and, therefore, lead to a high selectivity
to the desired light olefins.
Received: January 3, 2013
Revised: February 18, 2013
Published: February 20, 2013
A key problem of SAPO-34 for its application as an industrial
18−21
MTO catalyst is the rapid deactivation of this material.
For
©
2013 American Chemical Society
588
dx.doi.org/10.1021/cs400007v | ACS Catal. 2013, 3, 588−596

ACS Catalysis
Research Article
type carbenium ions during the MTO reaction were monitored
by H solid-state NMR spectroscopy using ammonia-loaded
of 8 kHz for 27Al and 31P nuclei, and a rate of 4 kHz was used
for the study of Si nuclei. To avoid hydrolysis of framework
1
29
catalyst samples. On the basis of the catalytic and spectroscopic
results, the deactivation behavior of SAPO-34 with different
crystal sizes in the MTO reaction are revealed, and the
deactivation mechanisms of these catalysts are discussed.
bonds, the hydration of the samples was performed not longer
than 1 day before the NMR characterization by exposing them
to an atmosphere saturated with the vapor of an aqueous
solution of Ca(NO ) at ambient temperature.
3
2
2
.2. MTO Reaction over the SAPO-34 Catalysts. The
2
. EXPERIMENTIAL SECTION
MTO reaction was investigated in a fixed-bed reactor at
atmospheric pressure. Typically, 0.4 g of the SAPO-34 catalysts
(sieve fraction, 0.25−0.5 mm) was placed in a stainless steel
reactor (5 mm i.d.) and activated under flowing nitrogen gas at
723 K for 1 h. After cooling to the reaction temperature of 673
2.1. Preparation and Characterization of the SAPO-34
Catalysts. The SAPO-34 catalysts were synthesized by a
hydrothermal method using the structure-directing agents
TEAOH (tetraethylammonium), TriEA (triethylamine), and
MOR (morpholine) following the procedures described in the
1
2
13
K, C-methanol or C-enriched methanol (see section 2.3.)
was injected at a rate of 0.5 mL/h, corresponding to the weight
2
9,31,32
literature.
The chemical compositions of the synthesis
−1
mixtures and the conditions for the synthesis of different
SAPO-34 samples are summarized in Table 1. Ortho
hourly space velocity of WHSV = 1.0 h . The products were
analyzed by an online gas chromatograph, HP5890/II, with
flame ionization detector and a packed-column Porapak Q to
separate the C −C hydrocarbons. The temperature of the
Table 1. Synthesis Parameters of the SAPO-34 Materials
under Study with Different Crystal Sizes
1
8
column was maintained at 313 K for 15 min and then increased
to 473 K at a heating rate of 10 K/min. The methanol
conversions and the selectivities to the various reaction
products for different times on stream were calculated without
considering the coke contents of the catalysts, which was
exclusively determined at the end of the catalytic experiments.
a
sample
batch composition
temp/K time/h ref
SAPO-34-S
1Al O /1P O /0.6SiO /
473
473
463
24
24
48
29
31
32
2
0
3
2
5
2
.5TEAOH/77H O
2
SAPO-34-M
SAPO-34-L
1Al O /1P O /0.6SiO /2TriEA/
2
3
2
5
2
5
0H O
2
2
.3. Characterization of Entrapped Organic Com-
1A1 O /1P O /0.6SiO /2MOR/
2
3
2
5
2
6
0H O
2
pounds on the Used SAPO-34 Catalysts. The nature of
organic compounds formed on the catalysts during the MTO
reaction was in situ monitored by FTIR and UV/vis
spectroscopy. FTIR spectra were recorded in the diffuse
reflection mode on a Bruker Tensor 27 spectrometer equipped
with a high sensitivity MCT detector cooled by liquid nitrogen
and an in situ reaction chamber. Prior to the FTIR studies, ∼20
mg of the catalyst materials was finely ground and placed in the
chamber. The samples were activated in flowing helium gas at
723 K for 1 h and cooled to 673 K for taking a background
spectrum. Then, methanol was injected into the reaction
a
Molar ratios.
phosphoric acid solution (85 wt %) and aluminumtriisopropy-
late (ATI) were mixed with distilled water and stirred for 2 h,
followed by dropwise addition of the structure-directing agents
and silica sol (neutral, 30 wt %, Alfa Aesar). The mixture was
further stirred for 6 h at room temperature, and the formed gel
was transferred into a Teflon-lined stainless-steel autoclave for
static crystallization. After crystallization at the temperatures
and with the durations given in Table 1 (columns 3 and 4), the
solid products were separated by a centrifuge, washed four
times with demineralized water, and dried at 353 K for 12 h.
The as-synthesized samples were then calcined in flowing
synthetic air at 873 K for 6 h.
−1
chamber at a flow rate of 0.025 mL/h (WHSV = 1.0 h ), and
time-resolved spectra were recorded with a resolution of 4 cm−
and an accumulation of 128 scans. The UV/vis spectra were
recorded in the diffuse reflection mode in the range of 200−600
nm using an AvaSpec-2048 fiber optic spectrometer, an
AvaLight-DH-S deuterium light source by Avantes, and a
glass fiber reflection probe HPSUV1000A by Oxford
Electronics. Before starting the MTO reaction, the glass fiber
reflection probe was placed in the fixed-bed reactor on the top
of the catalyst with a gap of ∼1.0 mm to the catalyst bed.
Reference UV/vis spectra of the catalysts were recorded at
reaction temperature prior to starting the methanol flow.
The organic deposits entrapped inside the pores and cages of
SAPO-34 were characterized by 13C HPDEC MAS NMR
spectroscopy on a Bruker Avance III 400WB spectrometer at a
resonance frequency of 100.6 MHz, with π/2 pulse excitation, a
repetition time of 20 s, a sample spinning rate of 12.0 kHz, and
high-power proton decoupling (HPDEC). 13C-enriched meth-
anol (99%, Cambridge Isotope Laboratories, Inc., USA), mixed
with nonenriched methanol at a ratio of 1:1, was used for
performing the MTO reaction in a standard fixed-bed reactor
before taking samples of the SAPO-34 catalysts after times on-
stream of TOS = 30 min to 1 h. For this purpose, the MTO
reaction was quenched by stopping the methanol flow at
reaction temperature and cooling down within 1 h under
nitrogen flow (60 mL/min). Subsequently, all catalyst samples
studied by NMR spectroscopy were transferred from the fixed-
1
X-ray diffraction (XRD) patterns of the calcined samples
were recorded on a Bruker D8 diffractometer with Cu Kα
radiation (λ = 1.5418 Å) at 5−50° with a scan rate of 2θ =
6
.0°/min.
A JEOL-JSM7500 field emission scanning electron micro-
scope (FESEM) was used to study the crystal morphologies.
The SEM images were recorded after covering the samples with
a thin layer of gold, deposited by sputtering.
The chemical compositions of the calcined samples were
determined by inductively coupled plasma atomic emission
spectrometer (ICP-AES, IRIS Advantage). The surface areas of
the calcined SAPO-34 materials were obtained by means of
nitrogen adsorption at 77 K on a Quantachrome Autosorb 3B
instrument. Before the nitrogen adsorption, the samples were
dehydrated at 473 K for 2 h. The total surface areas were
calculated via the Brunauer−Emmett−Teller (BET) equation.
Solid-state NMR investigations of the catalyst framework
were performed with hydrated samples on a Bruker Avance III
4
00WB spectrometer at resonance frequencies of 104.3, 79.5,
2
7
29
31
and 161.9 MHz for Al, Si, and P nuclei, respectively. The
MAS spectra were recorded upon single pulse excitation of π/6
2
7
29
31
for 2 Al and π/2 for Si and P, with repetition times of 0.5 s
7
29
31
for Al, 20 s for Si, and 30 s for P and sample spinning rates
5
89
dx.doi.org/10.1021/cs400007v | ACS Catal. 2013, 3, 588−596

ACS Catalysis
Research Article
bed reactor into gastight MAS NMR rotors inside a glovebox
purged with dry nitrogen gas to avoid contact with air.
The amounts of entrapped organic compounds after the
MTO reaction were analyzed by thermogravimetric analysis
(
TGA) on a Setram Setsys 16/18 thermogravmetric analyzer.
In a typical measurement, 0.1 g of the used catalyst materials
was heated in an Al O crucible at a constant heating rate of 10
2
3
K/min under flowing synthetic air of 30 mL/min.
1
2
.4. H MAS NMR Characterization of Surface Sites
and Benzene-Based Carbenium Species. The Brønsted
acid sites of the fresh and used SAPO-34 samples were
1
characterized by means of H MAS NMR spectroscopy utilizing
a Bruker Avance III 400WB spectrometer at a resonance
frequency of 400.1 MHz, with π/2 single pulse excitation, a
repetition time of 10 s, and a sample spinning rate of 25.0 kHz
1
using a 2.5 mm MAS NMR probe. Before the H MAS NMR
studies of the fresh SAPO-34 catalysts, these materials were
dehydrated at 673 K in vacuum (pressure below 10− Pa) for
2
1
2 h. Subsequently, the materials were sealed and kept in glass
Figure 1. XRD patterns of the as-synthesized SAPO-34-S (top),
SAPO-34-M (middle), and SAPO-34-L (bottom) materials.
tubes until their transfer into MAS NMR rotors inside a
glovebox purged with dry nitrogen gas. The used SAPO-34
catalysts were obtained after stopping the MTO reaction and
transferring them into MAS NMR rotors without contact to air.
The determination of the number of accessible Brønsted acid
sites was performed upon adsorption of ammonia at room
ppm indicate the presence of tetrahedrally coordinated
framework silicon atoms with four aluminum atoms in the
next-nearest coordination sphere of T atoms (Si(OAl) ) and
4
1
34,35
temperature and by evaluating the H MAS NMR signals
the absence of silicon islands.
The domination signal of the
3
1
caused by ammonium ions (δ1H = 6.5−7.0 ppm). For this
purpose, the catalyst samples were loaded with 100 mbar
ammonia and, subsequently, evacuated at 453 K for 2 h to
P MAS NMR spectra appearing at −28 ppm hints at
tetrahedrally coordinated phosphorus atoms in the alumino-
35
27
phosphate framework. In the Al MAS NMR spectra, strong
signals were observed at 30−44 ppm, which are caused by
tetrahedrally coordinated framework aluminum atoms, whereas
the much weaker signals at −14 to −11 ppm are due to
octahedrally coordinated aluminum atoms formed by coordi-
nation of two additional water molecules to tetrahedrally
coordinated framework aluminum atoms or by extraframework
1
eliminate physisorbed ammonia. Quantitative H MAS NMR
measurements were performed by comparing the signal
intensities of the samples under study with the intensity of an
external intensity standard (dehydrated zeolite H,Na−Y with a
cation exchange degree of 35%). The chemical shift values were
referenced to tetramethylsilane. The decomposition and
simulation of NMR spectra were carried out via the Bruker
software WINFIT.
36
aluminum species. On the basis of the results of the solid-
state NMR characterization, it could be concluded that the local
structures of the framework atoms in all three SAPO-34
materials are very similar and maintained the textual properties
of intact silicoaluminophosphates after their calcination.
3
. RESULTS AND DISCUSSION
3
.1. Physicochemical Properties of the SAPO-34
Catalysts. The XRD patterns of the as-synthesized materials
3.2. Catalytic Performances of the SAPO-34 Catalysts
in the MTO Reaction. In Figure 3, the methanol conversion
and product selectivity curves of the SAPO-34 catalysts in the
MTO reaction are shown. It is obvious that the crystal size has
a great impact on the lifetime of the MTO catalysts under
study. For SAPO-34-S, a methanol conversion of 100% and a
selectivity to light olefins of more than 90% were obtained at
673 K and maintained for a time-on-stream (TOS) of up to 10
h. After TOS = 10 h, the selectivity to C −C olefins gradually
shown in Figure 1 are typical for the chabazite (CHA)
33
framework structure, which indicates that pure SAPO-34
catalysts were obtained. In Figure 2, the FESEM images of the
crystal morphologies of the as-synthesized SAPO-34 materials
are shown. All three SAPO-34 materials consist of crystals with
a cubical shape, whereas their crystal sizes are quite different.
On the basis of their crystal sizes, the samples are assigned
SAPO-34-S (small: 2.5 μm), SAPO-34-M (medium: 6.0 μm),
and SAPO-34-L (large: 20.0 μm). The chemical compositions
and BET surface areas of the calcined SAPO-34 materials under
study are given in Table 2. The n /(n + n + n ) ratios and
2
4
decreased and dropped to ∼5% at TOS = 16 h, while the
methanol conversion decreased from 100% to ∼75%. For
SAPO-34-M, a methanol conversion of 100% and a selectivity
to light olefins of ∼90% was maintained only up to TOS = 4 h.
After TOS = 10 h, the selectivity to light olefins rapidly
decreased to ∼5% while dimethylether (DME) became the
dominant product (∼90%). In comparison with SAPO-34-S
and SAPO-34-M, the deactivation rate of large-crystalline
SAPO-34-L is much more rapid. Typically, 100% methanol
conversion could be only maintained for TOS = 10 min.
Afterward, the methanol conversion decreased dramatically to
40% after TOS = 30 min.
Si
Si
Al
P
surface areas of all three SAPO-34 materials are very similar and
2
occur in the ranges of 0.11−0.12 and 441−471 m /g,
respectively, indicating that the crystal size is exclusively the
deviating parameter.
For a better understanding of the textual properties of
2
9
31
27
SAPO-34 after calcination, Si, P, and Al MAS NMR
spectroscopy (see Figure S1 in the Supporting Information)
was employed to investigate the local structure of the
framework silicon, phosphorus, and aluminum atoms, respec-
tively, in the SAPO-34 samples under study. The 29Si MAS
NMR spectra with signals occurring exclusively at −92 to −89
3.3. TGA Analysis and 13C HPDEC MAS NMR
Investigations of Organic Deposits on the Deactivated
5
90
dx.doi.org/10.1021/cs400007v | ACS Catal. 2013, 3, 588−596

ACS Catalysis
Research Article
Figure 2. SEM images of the as-synthesized SAPO-34-S (left), SAPO-34-M (middle), and SAPO-34-L (right) materials.
Table 2. Chemical Compositions and Surface Areas of the
Calcined SAPO-34 Materials under Study
a
elemental analysis (mmol/g)
Brønsted acid
c
nSi/(nSi
+
BET surface
sites
b
2
sample
SAPO-
nAl
nP
nSi
nAl + nP) area (m /g)
(mmol/g)
6.6
4.9
1.5
0.12
0.12
0.11
441
464
471
1.25
3
4-S
SAPO-
4-M
SAPO-
4-L
6.1
6.0
4.6
4.1
1.4
1.1
1.15
3
1.00
3
a
b
c
Determined by ICP-AES. Determined by N absorption Deter-
2
1
mined by H MAS NMR
SAPO-34 Catalysts. The total weights of organic deposits
formed on SAPO-34-S, SAPO-34-M, and SAPO-34-L after
their deactivation in the MTO reaction were analyzed by TGA
in the temperature range of 293−1073 K. According to Figure
4
, all samples showed a similar weight loss of ∼1.9 wt % at
temperature up to 473 K due to the escape of volatile
compounds. In the high-temperature range of 473−1073 K, the
weight losses for the deactivated SAPO-34-S (TOS = 16 h),
SAPO-34-M (TOS = 16 h), and SAPO-34-L (TOS = 30 min)
were 6.0, 11.5, and 9.0 wt %, respectively. Hence, the amount of
organic deposits formed on SAPO-34-S during the MTO
reaction is obviously lower than that on SAPO-34-M and
SAPO-34-L, which should be a very important reason for the
prolonged lifetime of SAPO-34-S. The amount of organic
deposits formed on SAPO-34-L after TOS = 60 min is only
slightly lower than that on SAPO-34-M after TOS = 16 h,
indicating the much more rapid accumulation of organic
compounds on SAPO-34-L.
The 13C HPDEC MAS NMR spectra of the used SAPO-34
catalysts obtained after the MTO reaction for TOS = 1 h and
Figure 3. Methanol conversion and product selectivities during the
MTO conversion over the SAPO-34 catalysts under study at 673 K.
30 min are shown in Figure 5. For all samples under study,
strong signals of aromatic carbon atoms (125−133 ppm) and
alkyl carbon atoms bound to aromatic rings (12−30 ppm)
accumulated on SAPO-34 samples with larger crystal sizes
(SAPO-34-M and SAPO-34-L) during MTO reaction and vice
versa.
3
7
occur. These signals indicate the formation of polyalkyl
aromatics as dominant organic species over the SAPO-34
catalysts during the MTO reaction. Interestingly, the number of
3.4. In Situ FTIR and UV/Vis Studies of the MTO
Reaction over the SAPO-34 Catalysts. In situ FTIR and
UV/vis spectroscopy were utilized to investigate the nature of
organic compounds formed during the MTO reaction on the
catalysts under study. Figure 6 shows the in situ FTIR spectra
of the species adsorbed and entrapped on the different SAPO-
34 catalysts during the MTO reaction at 673 K and in steps of
30 min. For all three SAPO-34 catalysts under study, strong
methylene carbon atoms (−CH −) in the alkyl groups of
2
polyalkyl aromatics on SAPO-34-S (Figure 5, top), reflected by
the signals at 24−30 ppm, is significantly lower than that of
SAPO-34-M (Figure 5, middle) at TOS = 1 h and SAPO-34-L
(
Figure 5, bottom) at TOS = 30 min. On the basis of this
observation, it can be concluded that polyalkyl aromatics with
long alkyl groups (C , C , etc. instead of C ) are more rapidly
2
3
1
5
91
dx.doi.org/10.1021/cs400007v | ACS Catal. 2013, 3, 588−596

ACS Catalysis
Research Article
Figure 4. TGA curves of the SAPO-34-S (top), SAPO-34-M
(bottom), and SAPO-34-L (middle) materials obtained at 673 K
after TOS = 16 h for SAPO-34-S and SAPO-34-M and after TOS = 60
min for SAPO-34-L.
Figure 6. In situ FTIR spectra recorded during the MTO reaction at
6
73 K over the calcined SAPO-34-S (top), SAPO-34-M (middle), and
SAPO 34-L (bottom) catalysts up to TOS = 30 min.
SAPO-34-L. The wavenumber of 3017 cm− hints at C−H
1
stretching vibrations involving carbon atoms with double bonds
4
2
(
CH or CH ), which are typical for olefins and dienes.
2
Figure 5. 13C HPDEC MAS NMR spectra of the MTO catalysts under
study obtained at 673 K after TOS = 1 h for SAPO-34-S (top) and
SAPO-34-M (middle) and after TOS = 30 min for SAPO-34-L
Considering this assignment and the small width of the bands
at 3017 cm , their observation hints at the presence of light
−1
olefins with a retarded desorption from the SAPO-34 catalysts
with increasing reaction time. The presence of light olefins with
long residence time inside the SAPO-34 pores and cages leads
to an accelerated formation of oligomerization products, such
as aromatics acting as coke deposits, which indicates the end of
the catalyst lifetime. Because of the weak integral intensities of
the FTIR bands caused by entrapped light olefins with retarded
desorption properties, their number is too low for an
(bottom).
−1
bands at 2873−2948 cm due to C−H stretching vibrations
and weak bands at 1608−1621, 1503, 1461−1463, and 1380
−1
cm due to skeletal CC vibrations of aromatics and the
38−41
bending vibrations of C−H bond could be observed.
The
intensities of these bands increased with the progress of the
MTO reaction. In addition, negative bands at 3610 cm− due to
Brønsted acid sites interacting with adsorbate molecules
occurred for all SAPO-34 under study upon starting the
MTO reaction (see insets in Figure 6). This indicates that the
Brønsted acid sites were gradually covered by methanol or
reaction products. Furthermore, with increasing reaction time,
two very narrow bands due to the C−H stretching and methyl
1
13
observation by C HPDEC MAS spectroscopy.
In Figure 7, the in situ UV/vis spectra recorded during the
MTO reaction over the SAPO-34 catalysts under study at 673
K and with a TOS of up to 30 min are shown. For all three
SAPO-34 catalysts, similar organic species were formed during
the MTO reaction, leading to a series of UV/vis bands at ∼240,
270, 330, 390−400, and 450−475 nm. During the first 3 min,
that is, in the induction period of the MTO reaction, a rapid
increase in the UV/vis bands at 240, 270, and 400 nm was
−1
bending vibrations appeared at 3017 and 1304 cm . These
bands reached their maximum intensities for the deactivated
5
92
dx.doi.org/10.1021/cs400007v | ACS Catal. 2013, 3, 588−596

ACS Catalysis
Research Article
Figure 7. In situ UV/vis spectra recorded during the MTO reaction at 673 K over the calcined SAPO-34-S (left), SAPO-34-M (middle), and SAPO-
3
4-L (right) catalysts up to TOS = 30 min.
1
Figure 8. H MAS NMR spectra of the MTO catalysts under study obtained at 673 K for TOS up to 16 h in the case of SAPO-34-S (a), up to 10 h
in the case of SAPO-34-M (b), and up to 30 min in the case of SAPO-34-L (c).
observed which was due to the formation of initial compounds
of the hydrocarbon pool. With further progress of the MTO
reaction (TOS > 3 min), additional bands occurred at 330 and
450−475 nm, and the band intensities at ∼400 nm increased.
5
93
dx.doi.org/10.1021/cs400007v | ACS Catal. 2013, 3, 588−596

ACS Catalysis
Research Article
On the basis of earlier UV/vis studies of organic compounds on
Table 3. Numbers of Accessible Brønsted Acid Sites
(SiOHAl groups), nSiOHAl, and Benzene-Based Carbenium
1
2,43−47
acidic zeolite catalysts,
the bands at 240 and 270 nm
were assigned to UV/vis-sensitive dienes and polyalkyl
aromatics, respectively, and bands at 390 nm were explained
Ions, nbenzenium, on the Calcined and Used SAPO-34 Catalysts
under Study as Determined by H MAS NMR Spectroscopy
for Different Times on Stream (TOS) after Ammonia
Adsorption
1
47
by benzene-type carbenium ions. Bands at ∼400 nm hint at
4
5,48,49
the formation of polycyclic aromatics,
for example, with
two to three condensed aromatic rings at lower wavelengths
and with four condensed aromatic rings at slightly higher
wavelengths than 400 nm. Broad bands occurring at 330 and
a
a
nSiOHAl
nbenzenium
nSiOHAl + nbenzenium
sample
SAPO-
TOS
(mmol/g)
(mmol/g)
(mmol/g)
50
0
1
6
1
0
5
1
1
0
5
1
3
1.25
0.32
0
0
1.25
0.70
0.69
0
4
50−475 nm can be due to dienylic and trienylic carbenium
3
4-S
h
h
0.38
0.69
0
46
ions, respectively. The in situ UV/vis spectra reveal that with
increasing time-on-stream and for larger crystal size, the band at
6 h
0
270 nm due to alkyl aromatics becomes dominating in the high-
SAPO-
34-M
1.15
0.46
0.24
0
0
1.15
0.61
0.60
0
frequency region (see, e.g., Figure 7, right). Simultaneously, the
bands at ∼400 nm due to polycyclic aromatics are increasing in
the low-frequency region.
min
h
0.15
0.36
0
0 h
A direct comparison of the different SAPO-34 catalysts
allows the UV/vis spectra recorded at TOS = 1 min, which are
marked in red (Figure 7). These spectra demonstrate that the
large-crystalline SAPO-34-L is characterized by a very rapid
formation of alkyl aromatics (270 nm). Already at TOS = 10
min, polycyclic aromatics appearing at ∼400 nm dominate the
low-frequency region of the UV/vis spectra of SAPO-34-M and
SAPO-34-L. These large organic species block the cages and
pore entrances of the SAPO-34 catalysts, which leads to a
masking of Brønsted acid sites and a retarded desorption of
reaction products, both of which accelerate the deactivation, as
proved by the strong decrease in the methanol conversion
SAPO-
34-L
1.00
0.48
0.38
0.02
0
1.00
0.57
0.55
0.02
min
0.09
0.17
0
0 min
0 min
a
1
Determined by H MAS NMR after ammonia absorption
0
mmol/g. Interestingly, SAPO-34-S still exhibited a good
MTO activity at this TOS (compare Figure 3). After
deactivation of SAPO-34-S at TOS = 16 h, neither accessible
Brønsted acid sites nor benzene-based carbenium ions could be
detected (Table 3, last column).
According to the spectra in Figure 8b and c and considering
the data summarized in Table 3, the SAPO-34-M and SAPO-
curves in Figure 3.
1
3
.5. H MAS NMR Study of the Fate of Brønsted Acid
Sites and Carbenium Species during MTO Reaction. To
investigate the fate of Brønsted acid sites and active
hydrocarbon pool species on SAPO-34 samples, the calcined
and used catalyst samples taken from the fixed-bed reactor after
3
4-L catalysts showed similar trends in the changes of the
number of accessible Brønsted acid sites and benzene-based
carbenium ions, such as SAPO-34-S; however, in a significantly
accelerated manner. Also for these catalysts, the total
consumption of accessible Brønsted acid sites and benzene-
based carbenium ions led to their strong deactivation.
The above-mentioned findings agree with the observations in
our earlier work focusing on the fate of Brønsted acid sites in
1
different TOS were analyzed by H MAS NMR spectroscopy
without and after loading of ammonia (Figure 8, left and right,
respectively).
1
The H MAS NMR spectrum of calcined SAPO-34-S (TOS
43
=
0) is dominated by a signal at 3.7 ppm, corresponding to
SAPO-34 during the MTO process. Upon regeneration of the
used SAPO-34 catalysts in synthetic air at 873 K, the H MAS
1
bridging OH groups, that is, Brønsted acid sites (Figure 8a, top,
left). Adsorption of ammonia on the accessible Brønsted acid
NMR signals of bridging OH groups occurred again with nearly
the same intensities as observed for the fresh material; hence,
the MTO reaction does not lead to a dehydroxylation of
SAPO-34 catalysts. But during the MTO reaction, these
hydroxyl groups are involved in the formation of adsorbate
complexes of different nature (e.g., alkoxy groups, protonation
of aromatics, hydrogen bondings to deposits etc.), which is
1
site leads to the formation of ammonium ions, causing a H
MAS NMR signal at 6.4 ppm (Figure 8a, top, right). Evaluation
of this ammonium signal gave 1.25 mmol/g accessible Brønsted
acid sites in the calcined SAPO-34-S before its catalytic
application. After TOS = 1 h, strong signals at 1.0−2.3 and
7
.4−8.5 ppm corresponding to hydrogen atoms bound to
43
1
carbon atoms in saturated and unsaturated compounds,
accompanied by a shift and broadening of their H MAS NMR
signals initially appearing at 3.6 ppm.
1
respectively, can be observed in the H MAS NMR spectrum
recorded before and after ammonia loading. Meanwhile, the
intensity of signal at 3.7 ppm corresponding to Brønsted acid
site in SAPO-34 strongly decreased. On the basis of the signal
intensities in the spectrum of the ammonia-loaded sample, the
Brønsted acid site density was calculated to be 0.32 mmol/g. In
this spectrum of the ammonium-loaded sample, a new signal at
3.6. Deactivation Mechanisms of SAPO-34 Materials
Applied as MTO Catalysts. The different spectroscopic
methods utilized in the present work for investigating the
deactivation of SAPO-34 catalysts during the MTO conversion
gave a very complex picture. This picture consists of a number
of elements obtained by in situ UV/vis, FTIR, and solid-state
NMR spectroscopy being complementary methods, sensitive
for very different aspects of the deactivation process.
43
5
.1 ppm assigned to phenylammonium ions occurred, which
is caused by the ammoniation of benzene-type carbenium ions.
The benzene-type carbenium has been recognized as the key
hydrocarbon pool species for MTO reaction using SAPO-34 as
catalyst. According to the evaluation results summarized in
Table 3, the number of benzene-type carbenium ions in SAPO-
Considering the Brønsted acidity as a key property of SAPO-
1
34 materials for their application as MTO catalysts, H MAS
NMR spectroscopy has shown that all three SAPO-34 materials
under studySAPO-34-S, SAPO-34-M, and SAPO-34-L
have a similar number of accessible Brønsted acid sites before
starting the MTO reaction. However, the number of accessible
34-S increased to a value of 0.69 mmol/g after TOS = 6 h,
while the number of accessible Brønsted acid sites decreased to
5
94
dx.doi.org/10.1021/cs400007v | ACS Catal. 2013, 3, 588−596

ACS Catalysis
Research Article
Brønsted acid sites decreased after a short reaction time,
especially for the large-crystalline SAPO-34-L material.
Interestingly, before the strong deactivation of the SAPO-34
catalysts occurred, no accessible Brønsted acid sites could be
4. CONCLUSIONS
In the present work, SAPO-34 materials with comparable
Brønsted acid site densities, but different crystal sizes (2.5 μm
(
SAPO-34-S), 6.0 μm (SAPO-34-M), and 20.0 μm (SAPO-34-
1
detected by H MAS NMR spectroscopy. On the other hand,
L)) were prepared and applied as catalysts in the MTO
conversion. Catalytic investigations demonstrated that the
lifetime of the above-mentioned SAPO-34 catalysts in the
MTO reaction strongly depend on the crystal size: SAPO-34-S
although the number of accessible Brønsted acid sites
decreased, the number of benzene-based carbenium ions
increased during the highly active reaction period of the
SAPO-34 catalysts. The above-mentioned observation indicates
that benzene-based carbenium ions play a key role as
catalytically active hydrocarbon pool compounds in the MTO
conversion over SAPO-34. This statement is supported by the
absence of accessible benzene-type carbenium species in the
strongly deactivated SAPO-34 catalysts.
>
SAPO-34-M > SAPO-34-L. Thermographimetric analysis
indicated that the deactivation of the SAPO-34 catalysts under
study is accompanied by the formation of coke deposits. For
elucidating the mechanism of this catalyst deactivation,
HPDEC MAS NMR, FTIR, and UV/vis spectroscopy were
employed to monitor the formation and nature of organic
deposits during the MTO conversion. By H MAS NMR
spectroscopy, the nature and density of accessible Brønsted acid
sites and active hydrocarbon-pool species were determined as a
function of the time-on-stream (TOS).
The picture of the deactivation mechanism obtained by the
above-mentioned spectroscopic methods was found to be very
complex and consists of the following elements:
1
3
C
As demonstrated by UV/vis spectroscopy, the reaction
period characteristic of the formation of benzene-type
formation is the same as that of the formation of alkyl
aromatics in general (UV/vis band at 270 nm). In the case of
the large-crystalline SAPO-34-L material, 1 C HPDEC MAS
NMR spectroscopy indicated that these alkyl aromatics have
longer alkyl chains consisting of more methylene carbon atoms
1
3
13
(
C HPDEC MAS NMR signals at 24−30 ppm) on the used
SAPO-34-L and SAPO-34-M catalysts in comparison with
those formed on SAPO-34-S. At first, these larger alkyl
aromatics may have the advantage of increasing the selectivity
to ethene. After longer time-on-stream, however, the increasing
organic deposits influence and hinder the reactant diffusion,
which was also demonstrated in our recently published PFG
NMR study on reactant diffusion in SAPO-34 used as MTO
(
i) Already in the active period of the catalyst lifetime, a
rapid and strong decrease of the density of accessible
Brønsted acid sites occurs, while the number of benzene-
1
base carbenium ions increases ( H MAS NMR).
(
ii) The appearance of benzene-based carbenium ions is
accompanied by the formation of alkyl aromatics (UV/
vis).
4
4
catalysts.
Simultaneous with first indications of a catalyst deactivation,
(
iii) More pronounced for SAPO-34 catalysts with large
crystal sizes and longer reaction time, these alkyl
aromatics have longer alkyl chains with methylene
−1
narrow bands of olefinic species (FTIR bands at 3017 cm )
occurred in the FTIR spectra. Their observation by FTIR
spectroscopy hints at an enhanced residence time of these
reaction products in the SAPO-34 cages and pores, which may
be caused by their retarded desorption due to large alkyl
aromatics. Because of the high reactivity of the olefinic reaction
products, a surplus of these species in the SAPO-34 cages and
pores is accompanied by their rapid oligomerization, such as the
formation of polycyclic aromatics, as indicated by UV/vis
spectroscopy (UV/vis bands at ∼400 nm). Finally, these
polycyclic aromatics cause a total blocking of the SAPO-34
pores, which strongly hinders the diffusion of reaction products
and makes all catalytically active sites and species nonaccessible
for further reactants. Since especially the latter deactivation
steps are accompanied by a hindering of the reactant diffusion,
the total decrease in the olefin formation over the large-
crystalline SAPO-34-L occurred after a significantly shorter
MTO reaction time and is much more rapid than that observed
for SAPO-34-S and SAPO-34-M, with significantly smaller
crystal sizes. Furthermore, the accelerated blocking of cage
pores in the large-crystalline SAPO-34-L material is the reason
for the lower total coke formation in SAPO-34-L in comparison
with SAPO-34-S (see TGA data). In the case of SAPO-34-L,
the pores and cages of the outer sphere of the SAPO-34
particles are already blocked, before organic deposits can be
formed in the inner spheres. On the other hand, the weakly
acidic terminal OH groups located at the outer surface of the
SAPO-34 particles keep their accessibility for the reactant
methanol, which is converted, in this situation, with high
selectivity to DME (see Figure 3 for large TOS).
13
carbon atoms ( C HPDEC MAS NMR), which improve
at first the selectivity to light olefins, but also hinder the
diffusivity of reaction products for longer reaction times.
(iv) Simultaneously with the formation of large alkyl
aromatics, olefinic reaction products with enhanced
residence time inside the SAPO-34 cages and pore
were observed (FTIR).
(v) Because of the high reactivity of the olefinic reaction
products with enhanced residence time, the formation of
polycyclic aromatics occurs (UV/vis), which causes a
blocking of the catalyst pores.
(
vi) Finally, because of the pore blocking by the large
aromatic deposits, neither Brønsted acid sites nor
catalytically active carbenium ions are accessible for
1
further reactants ( H MAS NMR). This leads to a strong
decrease in the olefin formation over the SAPO-34
materials in the MTO reaction. In this situation,
methanol is preferentially converted to DME at terminal
OH groups located at the outer surface of the zeolite
particles.
Because of the high significance of the reactant diffusion for
the last steps of the above-mentioned deactivation mechanism,
the total deactivation of the large-crystalline SAPO-34-L
material studied in the present work occurred after a much
shorter MTO reaction time and is much more rapid than
observed for SAPO-34-S and SAPO-34-M with significantly
smaller crystal sizes.
5
95
dx.doi.org/10.1021/cs400007v | ACS Catal. 2013, 3, 588−596

ACS Catalysis
ASSOCIATED CONTENT
Research Article
(
23) Mees, F. D. P.; Van Der Voort, P.; Cool, P.; Martens, L. R. M.;
Janssen, M. J. G.; Verberckmoes, A. A.; Kennedy, G. J.; Hall, R. B.;
Wang, K.; Vansant, E. F. J. Phys. Chem. B 2003, 107, 3161−3167.
24) Bleken, F.; Bjørgen, M.; Palumbo, L.; Bordiga, S.; Svelle, S.;
Lillerud, K.-P.; Olsbye, U. Top. Catal. 2009, 52, 218−228.
25) Dai, W. L.; Wang, X.; Wu, G. J.; Li, L. D.; Guan, N. J.; Hunger,
M. ChemCatChem 2012, 4, 1428−1435.
26) Chen, D.; Moljord, K.; Fuglerud, T.; Holmen, A. Microporous
■
* Supporting Information
S
Additional information as noted in text. This material is
available free of charge via the Internet at http://pubs.acs.org.
(
(
AUTHOR INFORMATION
■
*
(
Corresponding Author
(M.H.) Fax: +49-711-685-64081. E-mail: michael.hunger@itc.
Mesoporous Mater. 1999, 29, 191−203.
(27) Lee, Y. J.; Baek, S. C.; Jun, K. W. Appl. Catal., A 2007, 329,
130−136.
(28) Lee, K. Y.; Chae, H.-J.; Jeong, S.-Y; Seo, G. Appl. Catal., A 2009,
uni-stuttgart.de. (L.L.) Fax: +86-22-2350-0341. E-mail: lild@
nankai.edu.cn.
3
(
69, 60−66.
29) Nishiyama, N.; Kawaguchi, M.; Hirota, Y.; Vu, D. V.; Egashira,
Y.; Ueyama, K. Appl. Catal., A 2009, 362, 193−199.
30) Wang, P. F.; Lv, A. L.; Hu, J.; Xu, J. A.; Lu, G. Z. Microporous
Mesoporous Mater. 2012, 152, 178−184.
31) Tan, J.; Liu, Z. M.; Bao, X. H.; Liu, X. C.; Han, X. W.; He, C. Q.;
Zhai, R. S. Microporous Mesoporous Mater. 2002, 53, 97−108.
32) Iwase, Y.; Motokura, K.; Koyama, T.-R.; Miyaji, A.; Baba, T.
Phys. Chem. Chem. Phys. 2009, 11, 9268−9277.
33) Collection of Simulated XRD Powder Patterns for Zeolites, 5th
revised ed.; Treacy, M. M. J., Higgins, J. B., Eds.; Elsevier: Amsterdam;
007.
34) Hunger, M.; Brunner, E. In Molecular Sieves − Science and
Technology, Karge, H. G., Weitkamp, J., Eds.; Springer-Verlag: Berlin,
004; Vol. 4; pp 201−293.
35) Buchholz, A.; Wang, W.; Xu, M.; Arnold, A.; Hunger, M.
Microporous Mesoporous Mater. 2003, 57, 157−168.
36) Zibrowius, B.; Loeffler, E.; Hunger, M. Zeolites 1992, 12, 167−
74.
37) Spektroskopische Methoden in der organischen Chemie, 8th revised
ed.; Hesse, M., Meier, H., Zeeh, B.; Georg Thieme Verlag: Stuttgart,
Notes
The authors declare no competing financial interest.
(
ACKNOWLEDGMENTS
■
(
This work was supported by the National Basic Research
Program of China (2009CB623502), 111 Project (B12015),
and the Ministry of Education of China (NCET-11-0251).
M.H. is grateful for for financial support by Fonds der
Chemischen Industrie and Deutsche Forschungsgemeinschaft.
(
(
2
(
REFERENCES
■
(
(
1) Sto
̈
cker, M. Microporous Mesoporous Mater. 1999, 29, 3−48.
2
(
2) Chen, J. Q.; Bozzano, A.; Glover, B.; Fuglerud, T.; Kvisle, S.
Catal. Today 2005, 106, 103−107.
(
(
(
3) White, J. L. Catal. Sci. Technol. 2011, 1, 1630−1635.
(
1
(
4) Ilias, S.; Bhan, A. ACS Catal. 2013, 3, 18−31.
5) Olsbye, U.; Svelle, S.; Bjørgen, M.; Beato, P.; Janssens, T. V. W.;
Joensen, F.; Bordiga, S.; Lillerud, K. P. Angew. Chem., Int. Ed. 2012, 51,
810−5831.
6) Djieugoue, M. A.; Prakash, A. M.; Kevan, L. J. Phys. Chem. B
000, 104, 6452−6461.
7) Zhu, Z.; Hartmann, M.; Kevan, L. Chem. Mater. 2000, 12, 2781−
787.
8) Svelle, S.; Olsbye, U.; Joensen, F.; Bjørgen, M. J. Phys. Chem. C
007, 111, 17981−17984.
9) Park, J. W.; Kim, S. J.; Seo, M.; Kim, S. Y.; Sugi, Y.; Seo, G. Appl.
Catal., A 2008, 349, 76−85.
10) Castro, M.; Warrender, S. J.; Wright, P. A.; Apperley, D. C.;
5
(
2
(
2
(
2
(
2
(
011; pp 232−234.
38) Hunger, M. Microporous Mesoporous Mater. 2005, 82, 241−255.
39) Petkovic, L. M.; Ginosar, D. M.; Burch, K. C. J. Catal. 2005, 234,
(
3
(
28−339.
40) Bjørgen, M.; Bonino, F.; Arstad, B.; Kolboe, S.; Lillerud, K.-P.;
Zecchina, A.; Bordiga, S. ChemPhysChem 2005, 6, 232−235.
(
(
1
(
41) Park, J. W.; Seo, G. Appl. Catal., A 2009, 356, 180−188.
42) Binder, H.; Anikin, A.; Kohlstrunk, B. J. Phys. Chem. B 1999,
03, 450−460.
43) Dai, W. L.; Scheibe, M.; Guan, N. J.; Li, L. D.; Hunger, M.
ChemCatChem 2011, 3, 1130−1133.
44) Dai, W. L.; Scheibe, M.; Li, L. D.; Guan, N. J.; Hunger, M. J.
Phys. Chem. C 2012, 116, 2469−2476.
45) Jiang, Y.; Huang, J.; Reddy Marthala, V. R.; Ooi, Y. S.;
Weitkamp, J.; Hunger, M. Microporous Mesoporous Mater. 2007, 105,
32−139.
46) Kirisci, I.; Førster, H.; Tasi, G.; Nagy, J. B. Chem. Rev. 1999, 99,
085−2114.
47) Bjørgen, M.; Bonino, F.; Kolboe, S.; Lillerud, K.-P.; Zecchina,
A.; Bordiga, S. J. Am. Chem. Soc. 2003, 125, 15863−15868.
48) Karge, H. G.; Laniecki, M.; Ziolek, M.; Onyestyak, G.; Kiss, A.;
Kleinschmit, P.; Siray, M. In Zeolites: Facts, Figures, Future;8th
International Zeolite Conference; Amsterdam, Netherlands; 1989;
Studies in Surface Science and Catalysis; Jacobs, P. A., van Santen, R.
A., Eds.; Elsevier: Amsterdam, 1989; Vol. 49; p 1327.
49) Mohan, J. Organic Spectroscopy Principles and Applications; Alpha
Science International Ltd.: Harrow, 2002; pp 128, 137.
50) Park, J. W.; Lee, J. Y.; Kim, K. S.; Hong, S. B.; Seo, G. Appl.
Catal., A 2008, 339, 36−44.
(
Belmabkhout, Y.; Pirngruber, G.; Min, H. K.; Park, M. B.; Hong, S. B.
J. Phys. Chem. C 2009, 113, 15731−15741.
(
(
11) Lee, J. H.; Park, M. B.; Lee, J. K.; Min, H. K.; Song, M. K.;
Hong, S. B. J. Am. Chem. Soc. 2010, 132, 12971−12982.
12) Dai, W. L.; Wang, X.; Wu, G. J.; Guan, N. J.; Hunger, M.; Li, L.
D. ACS Catal. 2011, 1, 292−299.
13) Li, J.; Wei, Y. X.; Liu, G.; Qi, Y.; Tian, P.; Li, B.; He, Y.; Liu, Z.
Catal. Today 2011, 117, 221−228.
14) Wang, Q.; Cui, Z. M.; Cao, C. Y.; Song, W. G. J. Phys. Chem. C
011, 115, 24987−24992.
15) Li, J. Z.; Wei, Y. X.; Chen, J. R.; Tian, P.; Su, X.; Xu, S. T.; Qi,
Y.; Wang, Q. Y.; Zhou, Y.; He, Y. L.; Liu, Z. M. J. Am. Chem. Soc. 2012,
34, 836−839.
16) Teketel, S.; Skistad, W.; Benard, S.; Olsbye, U.; Lillerud, K. P.;
Beato, P.; Svelle, S. ACS Catal. 2012, 2, 26−37.
17) Bhawe, Y.; Moliner-Marin, M.; Lunn, J. D.; Liu, Y.; Malek, A.;
Davis, M. ACS Catal. 2012, 2, 2490−2495.
18) Hereijgers, B. P. C.; Bleken, F.; Nilsen, H. M.; Svelle, S.;
Lillerud, K.-P.; Bjørgen, M.; Weckhuysen, B. M.; Olsbye, U. J. Catal.
009, 264, 77−87.
19) Olsbye, U.; Bjørgen, M.; Svelle, S.; Lillerud, K. P.; Kolboe, S.
(
(
1
(
2
(
(
(
2
(
(
1
(
(
(
(
(
2
(
Catal. Today 2005, 106, 108−111.
(20) Guisnet, M. J. Mol. Catal. A 2002, 182−183, 367−382.
(21) Qi, G. Z.; Xie, Z. K.; Yang, W. M.; Zhong, S. Q.; Liu, H. X.;
Zhang, C. F.; Chen, Q. L. Fuel Process. Technol. 2007, 88, 437−441.
22) Van Niekerk, M. J.; Fletcher, J. C. Q.; O’Connor, C. T. Appl.
Catal., A 1996, 138, 135−145.
(
5
96
dx.doi.org/10.1021/cs400007v | ACS Catal. 2013, 3, 588−596