Interplay between nanoscale reactivity and bulk performance of H-ZSM-5 catalysts during the methanol-to-hydrocarbons reaction

Article abstract of DOI:10.1016/j.jcat.2013.07.009

H-ZSM-5 catalyst powders before and after a steaming post-treatment have been investigated during the Methanol-To-Hydrocarbons (MTH) process at 350 C. Bulk and surface characterization techniques have been combined with in situ Scanning Transmission X-ray Microscopy (STXM) at the aluminum and carbon K-edge to study the changes in acidity, porosity, reactivity, and aluminum distribution upon steaming. It was found that steaming post-treatment has a positive impact on the stability of H-ZSM-5 without inducing important changes in the MTH activity and selectivity. The lower MTH stability of non-steamed H-ZSM-5 catalyst powder is related to the formation of poly-aromatic compounds in the outer regions of the catalyst particles, as probed with in situ STXM. In contrast, a limited amount of poly-aromatics was found in the outer rim of steamed H-ZSM-5 catalyst particles. These differences occur as a result of the generation of mesoporosity as well as the reduction in the number and strength of acid sites after steaming, as evidenced by the nanoscale imaging of adsorbed pyridine with STXM.

Full text of DOI:10.1016/j.jcat.2013.07.009

Journal of Catalysis 307 (2013) 185–193
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Interplay between nanoscale reactivity and bulk performance
of H-ZSM-5 catalysts during the methanol-to-hydrocarbons reaction
a
b
b
c
d
Luis R. Aramburo , Shewangizaw Teketel , Stian Svelle , Simon R. Bare , Bjørnar Arstad ,
e
b
a
a,
⇑
Henny W. Zandbergen , Unni Olsbye , Frank M.F. de Groot , Bert M. Weckhuysen
a
Inorganic Chemistry and Catalysis Group, Debye Institute for NanoMaterials Science, Faculty of Science, Utrecht University, Universiteitsweg 99, 3584 CG Utrecht, The Netherlands
inGAP Center for Research Based Innovation, Department of Chemistry, University of Oslo, N-0315 Oslo, Norway
UOP LLC, a Honeywell Company, Des Plaines, IL 60017, USA
Department of Process Chemistry SINTEF Materials & Chemistry, 0314 Oslo, Norway
Kavli Institute of NanoScience, National Centre for High Resolution Electron Microscopy, Delft University of Technology, PO Box 5046, 2600 GA Delft, The Netherlands
b
c
d
e
a r t i c l e i n f o
a b s t r a c t
Article history:
H-ZSM-5 catalyst powders before and after a steaming post-treatment have been investigated during the
Methanol-To-Hydrocarbons (MTH) process at 350 °C. Bulk and surface characterization techniques have
been combined with in situ Scanning Transmission X-ray Microscopy (STXM) at the aluminum and car-
bon K-edge to study the changes in acidity, porosity, reactivity, and aluminum distribution upon steam-
ing. It was found that steaming post-treatment has a positive impact on the stability of H-ZSM-5 without
inducing important changes in the MTH activity and selectivity. The lower MTH stability of non-steamed
H-ZSM-5 catalyst powder is related to the formation of poly-aromatic compounds in the outer regions of
the catalyst particles, as probed with in situ STXM. In contrast, a limited amount of poly-aromatics was
found in the outer rim of steamed H-ZSM-5 catalyst particles. These differences occur as a result of the
generation of mesoporosity as well as the reduction in the number and strength of acid sites after steam-
ing, as evidenced by the nanoscale imaging of adsorbed pyridine with STXM.
Received 31 October 2012
Revised 7 July 2013
Accepted 12 July 2013
Keywords:
Heterogeneous Catalysis
Zeolites
Methanol-to-hydrocarbons
Coke formation
In situ spectroscopy
Ó 2013 Elsevier Inc. All rights reserved.
1
. Introduction
The Methanol-To-Hydrocarbons (MTH) process has attracted
powders has been investigated in detail. We compare the reactivity
of a calcined (sample name: H-ZSM-5-C) and a mildly steamed
(sample name: H-ZSM-5-500) catalyst powder during the MTH
reaction at 350 °C by catalytic testing and in situ Scanning Trans-
mission X-ray Microscopy (STXM) at the carbon and aluminum
K-edges as well as by a range of bulk and surface characterization
methods. This unique experimental approach allows proposing an
explanation of the large differences noticed in MTH stability be-
tween H-ZSM-5-C and H-ZSM-5-500 in terms of nanoscale differ-
ences in the type and amount of hydrocarbon deposits in
addition to the number and strength of acid sites within individual
zeolite H-ZSM-5 aggregates.
both industrial and academic interest as methanol can be produced
at a large scale from syngas by the direct oxidative conversion of
methane or by the reductive conversion of atmospheric carbon
dioxide with hydrogen [1]. Depending on the reaction conditions
and the type of catalyst used in the MTH process, methanol can
be transformed into olefins (MTO), gasoline (MTG), or other va-
lue-added chemicals [2,3]. Since the first oil crisis, zeolite materials
have been extensively studied as promising catalysts for the MTH
process [4,5]. Among them, H-ZSM-5 is considered as one of the
archetypal catalyst to conduct the MTH reaction [4]. Nonetheless,
the generation of undesired carbon deposits often leads to severe
diffusion restrictions and fast deactivation rates. As a result,
H-ZSM-5 needs to be regenerated, which frequently induces
leaching of framework Al species, resulting in important variations
in activity, selectivity as well as overall stability [6,7].
2. Experimental
2.1. Materials
In the present work, the effect that a mild hydrothermal treat-
ment has on the MTH stability of relevant zeolite H-ZSM-5 catalyst
Two different zeolite H-ZSM-5 aggregate materials, namely a
calcined material (i.e., H-ZSM-5-C) and a mildly steamed material
(
i.e., H-ZSM-5-500), have been investigated. The starting material,
with crystallite dimensions of approximately 200–800 nm, is a
commercial zeolite NH -ZSM-5 catalyst powder provided by Zeo-
lyst (CBV2314, Si/Al = 11.5). To obtain H-ZSM-5-C, the starting
⇑
Corresponding author. Fax: +31 30 251 1027.
E-mail address: b.m.weckhuysen@uu.nl (B.M. Weckhuysen).
4
0
021-9517/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2013.07.009

1
86
L.R. Aramburo et al. / Journal of Catalysis 307 (2013) 185–193
material was calcined in a static oven (N100 Nabertherm) first pre-
heating it at 120 °C (30 min, 2 °C/min) and then increasing the
temperature to 550 °C (360 min, 10 °C/min). The steamed sample
was prepared treating H-ZSM-5-C in a quartz tubular oven (Therm-
oline 79300) during 180 min at 500 °C via saturation of a nitrogen
flow (180 ml/min) with water at 100 °C. Prior to steaming,
the sample was preheated at 120 °C (30 min, 2 °C/min). After the
hydrothermal treatment, the sample was calcined following the
same procedure as that described to obtain H-ZSM-5-C.
similarities in their spectra. During the measurements, the
monochromized X-ray beam was focused to a spot size of
35 Â 35 nm using a Fresnel Zone plate. The pixel size used in the
carbon K-edge STXM image sets was 35 nm.
In a second set of STXM experiments, ex-situ aluminum K-edge
measurements were done at beamline 10.ID.1 of the Canadian
Light Source (CLS, Saskatoon, Canada) using a 35 nm zone plate.
In every case, the samples were placed on silicon nitride windows
(100 nm thickness) and mounted perpendicular to the beam on a
piezoelectric stage. The aluminum stacks were recorded in the
1555–1590 eV range with an energy resolution of 0.15 eV and a
spatial resolution of 70 nm. Aluminum reference compounds were
measured performing a line scan along the samples with a dwell
time of 4 ms and an energy resolution of 0.15 eV. Normalization
and background correction was done subtracting to every spec-
trum a second spectrum obtained from an aluminum-free region.
The aluminum K-edge X-ray absorption spectra (XAS) of these ref-
erence compounds were used to perform a least squares linear
combination fitting of the aluminum stacks obtained from both
H-ZSM-5-C and H-ZSM-5-500. The STXM data files were analyzed
using aXis2000 software.
2.2. X-ray diffraction
X-ray diffraction (XRD) patterns were obtained at room temper-
ature from 5° to 90° 2h with a Bruker-AXS D8 Advance powder
X-ray diffractometer, equipped with an automatic divergence slit,
a Våntec-1 detector and a Cobalt K
.3. Nitrogen physisorption
a 1,2 (k – 1.79026 Å) source.
2
Nitrogen adsorption and desorption isotherms were measured
at À196 °C on a Micromeritics Tristar 3000 instrument.
In a third set of STXM experiments, pyridine adsorption exper-
iments were performed on the interferometrically controlled
microscope at beamline 11.0.2 of the Advanced Light Source
(ALS) at the Lawrence Berkeley National Laboratory (LBNL, Berke-
2.4. Catalytic testing
Catalytic testing experiments were performed on the powders
without pressing and sieving the two catalyst materials under study.
For each experiment, 30 mg of catalyst was placed in a fixed-bed
reactor, activated at 550 °C under oxygen for 1 h, and then cooled
to the reaction temperature (350 °C). The MTH reaction products
were analyzed with an online gas chromatograph (GC) connected
to the outlet of a fixed-bed reactor (i.d. 6 mm) using a heated transfer
line. The inlet of the reactor was connected to helium, which was
ley, USA) [9]. Prior to pyridine adsorption, 20 mg of zeolite sample
À3
was heated (170 °C) in a round-bottom flask under vacuum (10
-
mbar) for 60 min. Subsequently, the samples were prepared adding
0.2 l of pyridine (Acros Organics, 99%). After a waiting time of
l
60 min, pyridine was partially removed by heating the samples
at 400 °C for a period of 60 min. During the STXM experiments,
the samples were placed on silicon nitride windows (100 nm thick-
ness) and mounted perpendicular to the beam on a piezoelectric
stage. The monochromatized X-ray beam was focused to a spot size
of 25 Â 25 nm using a Fresnel zone plate. The absorption at the car-
bon K-edge was measured in the range of 280–310 eV collecting a
series of stacks with an energy resolution of 0.2 eV. After spatially
aligning the stack, PCA was used to obtain the primary components
in the data set. Subsequently, a cluster analysis was performed to
classify pixels according to statistical similarities in their spectra.
The pixel size used in the image sets was 30 nm.
bubbled through methanol kept at 20 °C (PMeOH = 130 mbar) in a sat-
À1 À1
uration evaporator (WHSV = 5.56 g g
h ).
To analyze the coke after reaction, 15 mg of the deactivated cat-
alyst samples was transferred into a Teflon tube where 1 ml of 15%
hydrofluoric acid was added. Subsequent to a waiting time of
3
0 min, 1 ml of dichloromethane, containing hexachloroethane as
the internal standard, was added to the Teflon tube. Then, 1 l of
l
the resulting organic phase was analyzed in an Agilent 6890N GC
equipped with an Agilent 5793 Mass Selective Detector. A HP-
5
0
MS column (60 m, 0.25 mm i.d., stationary phase thickness
.25 m) and an inlet split of 1:5 were used for this purpose.
l
2.6. Temperature-Programmed Desorption
2
.5. Scanning transmission X-ray microscopy
Temperature-Programmed Desorption (TPD) measurements
with ammonia as probe molecule were performed using a Microm-
eritics AutoChemII 2920 apparatus. The sample, 0.15 g in both
cases, was first pretreated in helium (25 ml/min) for 30 min at
600 °C, then cooled down to 100 °C, and saturated with ammonia
to its equilibrium state. Prior to desorption, samples were flushed
in helium for 30 min. Subsequently, ammonia desorption was per-
formed in the range of 100–600 °C at a heating rate of 10 °C/min.
In a first set of in situ STXM experiments, the MTH reaction was
performed on the interferometrically controlled STXM microscope
at the Pollux beamline of the Swiss Light Source (SLS, Villigen, Swit-
zerland) [8]. For this purpose, a Micro-Electro-Mechanical System
(
MEMS) nanoreactor was used to characterize the catalytic behav-
ior of the two zeolite H-ZSM-5 catalyst powders under study. Prior
to reaction, the samples were introduced inside the nanoreactor
and placed in an adaptor that can be translated with nanometer
precision by an interferometrically controlled (x,y,z) piezoelectric
stage. The samples were activated at 450 °C flowing helium (2 ml/
min) during 30 min. Then, the temperature was decreased to
2.7. Sputter depth profiling X-ray Photoelectron Spectroscopy
Sputter depth profiling X-ray Photoelectron Spectroscopy (XPS)
measurements were performed on a Physical Electronics Quantum
2000 Scanning X-ray Photoelectron Spectrometer or a Physical
Electronics Quantera X-ray photoelectron spectrometer. In either
3
50 °C, and methanol (Antonides-Interchema, 99%) was introduced
via saturation of helium flow (2 ml/min) at 0 °C for 400 min. During
reaction, the absorption at the carbon K-edge was analyzed at dif-
ferent times on stream collecting a series of images over small en-
ergy increments, subsequently combining these images to form a
spectral image sequence (stack). The stacks were obtained in the
range of 280–310 eV using an energy resolution of 0.2 eV. In a sec-
ond step, a Principal Component Analysis (PCA) and a cluster anal-
ysis were performed to classify pixels according to statistical
À9
instrument, the base pressure was kept ꢀ1 Â 10 mbar. The XPS
spectra were collected using monochromatic aluminum Ka radia-
tion (1486.6 eV). Charging was neutralized using the combined
low energy electron flood gun and low energy argon ion gun of
the instrument. Survey scans were collected using a pass energy
of 187 eV and region scans with a pass energy of 58 eV. For typical
measurements, H-ZSM-5 zeolites were deposited onto a vitreous

L.R. Aramburo et al. / Journal of Catalysis 307 (2013) 185–193
187
carbon substrate to ensure dispersion. A nominal 5
X-ray beam was used for the analysis. The depth calibration for
the sputter depth profiling was determined from the time required
l
m or 10
l
m
both samples. Most notably, the higher methanol conversion ob-
served for H-ZSM-5-500 for a larger time-on-stream is indicative
for its higher resistance toward deactivation. After 35 h on stream,
H-ZSM-5-C was fully deactivated, while H-ZSM-5-500 presented a
methanol conversion above 55%. A lower intrinsic activity and im-
proved resistance toward deactivation are consistent with a reduc-
tion in the density of acid sites [10]. It may be noted, however, that
the deactivation behavior of zeolite catalysts is a complex issue.
Among material parameters, topology, crystal size, defects as well
as density, and strength of Brønsted acid sites have all been pro-
posed in the literature as relevant parameters for deactivation dur-
ing the MTH reaction [3,11]. Nonetheless, for series of catalysts
prepared by dealumination, the improved stability has been linked
primarily to the lower density of acid sites, as previously demon-
strated for H-beta [12] and H-ZSM-5 [13]. This is caused primarily
due to a reduced tendency toward secondary and coke-forming
reactions when having a lower acid site density [10,14]. In contrast
to the observed activity differences, the selectivity of H-ZSM-5-C
was very similar to that of H-ZSM-5-500. More in particular, as
shown in Fig. 1b and c, for both zeolite H-ZSM-5 powders, the same
compounds, in similar concentrations, were observed in the
effluent.
2
to sputter through a known thickness of SiO on a silicon wafer
using the same ion gun settings. In this manner, the depths of
H-ZSM-5-C and H-ZSM-5-500 were estimated. The Si/Al atomic ra-
tios were calculated using the integrated peak areas of the Si2p and
Al2p making use of an integrated Physical Electronics data analysis
software (Multipak).
2
7
2
.8. Al magic angle spinning nuclear magnetic resonance
27Al Magic Angle Spinning (MAS) Nuclear Magnetic Resonance
(
NMR) experiments were performed at 11.7 T on a Bruker Avance
III spectrometer using a 3.2 mm triple resonance MAS probe head.
The spectra were obtained at room temperature using /12 pulses
p
at an rf-field of 94 kHz, 10000 scans and a recycle delay of 0.5 s.
The MAS rate was 20 kHz. The chemical shifts of 27Al were exter-
3 3
nally referenced to 1 M aluminum nitrate (Al(NO ) (aq)).
3
. Results and discussion
Additional measurements of the hydrocarbon species retained
in the catalyst materials after MTH reaction disclosed important
variations between H-ZSM-5-C and H-ZSM-5-500. As shown in
Fig. 2, H-ZSM-5-C contained a mixture of methylbenzenes (tetraM-
Bs ꢀ pentaMB ꢁ hexaMB) and poly-aromatic species. In contrast,
H-ZSM-5-500 possessed a different amount of methylbenzenes
(tetraMBs > pentaMB ꢁ hexaMB) and the absence of poly-aro-
matic species after MTH reaction. Methylbenzenes are typically
the major group of compounds in the material retained within
the pores of H-ZSM-5 catalysts after the MTH reaction, in particular
for samples with relatively low Al contents [15]. Thus, it appears
likely that the unusually high concentration of heavier species
found in the extract from the H-ZSM-5-C catalyst is related to
the fairly low Si/Al ratio. The fact that comparatively very little
3
3
.1. Performance changes by applying a mild hydrothermal treatment
.1.1. Catalytic testing
With the aim of assessing the influence that a mild steaming
treatment has on the MTH performance of zeolite H-ZSM-5 cata-
lyst powders, H-ZSM-5-C and H-ZSM-5-500 were tested under
realistic MTH conditions at 350 °C. The catalytic performances of
both zeolite H-ZSM-5 samples are compared in Fig. 1.
It can be observed from Fig. 1 that H-ZSM-5-C showed full
methanol conversion in the beginning of the MTH reaction,
whereas H-ZSM-5-500 presented an initial methanol conversion
of ꢀ92%. With increasing time-on-stream, the catalytic activity
progressively dropped, revealing important differences between
Fig. 1. Methanol conversion to hydrocarbons over H-ZSM-5-C (filled symbols) and H-ZSM-5-500 (open symbols) at 350 °C. (b) Selectivity toward C
over H-ZSM-5-C (filled symbols) and H-ZSM-5-500 (open symbols). (c) Percentage composition of aromatics in H-ZSM-5-C (filled symbols) and H-ZSM-5-500 (open symbols).
d) Selectivity toward C –C4, 5+ aliphatic and aromatic species during MTH over H-ZSM-5-C (filled symbols) and H-ZSM-5-500 (open symbols).
1 4
–C species during MTH
(
1
C

1
88
L.R. Aramburo et al. / Journal of Catalysis 307 (2013) 185–193
Fig. 3. 27Al Magic Angle Spinning (MAS) Nuclear Magnetic Resonance (NMR)
spectra of H-ZSM-5-C (a) and H-ZSM-5-500 (b).
resonance in the region of 30–35 ppm, attributed to 5-fold alumi-
num species, was found after steaming [22].
Fig. 2. Retained hydrocarbon species in H-ZSM-5-C (a) and H-ZSM-5-500 (b) after
performing the MTH reaction at 350 °C.
The quantification of the 27Al MAS NMR spectra obtained from
H-ZSM-5-C revealed that 87%, 1%, and 12% of the aluminum show
4
-fold, 5-fold, and 6-fold coordination environments, respectively.
poly-aromatic compounds are found in the extract of the H-ZSM-5-
In contrast, the concentration of 4-fold, 5-fold, and 6-fold alumi-
2
7
5
00 catalyst might be attributed to several reasons. One could
num obtained from the Al MAS NMR spectra of H-ZSM-5-500
was 38%, 48%, and 14%, respectively. Consequently, these findings
reveal a relative increase in the amount of 5- and 6-fold coordi-
nated aluminum after performing a mild hydrothermal treatment.
speculate that the mesopores formed upon steaming contain very
few acid sites, or conversely, that the coke potentially formed in
these mesopores quickly grows into species insoluble in CH
and thus undetectable by the dissolution/extraction protocol
16]. However, as is demonstrated by the ammonia TPD, discussed
2 2
Cl
[
3.5. Temperature-programmed desorption
below, the density of acid sites is lower for the H-ZSM-5-500, and
this is inherently linked to a reduced coking rate and tendency for
aromatics formation.
To investigate the influence that a mild hydrothermal treatment
has on the acidic properties of zeolite H-ZSM-5 powders, tempera-
ture-programmed desorption experiments were performed with
ammonia as probe molecule.
As depicted in Fig. 4, a decrease in the overall acidity (i.e., the
total area below the curves) was found after a mild steaming treat-
ment. More specifically, the total ammonia desorption recorded
from the TPD experiments was 51.9 ml (STP)/g and 27.5 ml
3.2. Physicochemical changes by applying a mild hydrothermal
treatment
A series of bulk, surface, and local characterization techniques
have been applied to gain detailed insight into the physicochemical
properties of H-ZSM-5-C and H-ZSM-5-500.
(
STP)/g for the calcined and steamed catalyst sample, respectively.
Furthermore, the intensity reduction as well as shift of the desorp-
tion peak appearing at higher temperatures is indicative of a loss of
strong acid sites. Accordingly, ammonia TPD informs that steaming
has a significant impact on the acidic properties of H-ZSM-5-500.
3
.3. Nitrogen physisorption and X-ray diffraction
As described in a previous publication [17], H-ZSM-5-C is com-
posed of rectangular parallelepiped shape crystals domains, having
dimensions of 200–800 nm. The agglomeration of these domains
results into the formation of 1–2 lm spheroid particles. Accord-
3.6. X-ray photoelectron spectroscopy
ingly, the nitrogen adsorption and desorption isotherms obtained
for H-ZSM-5-C were characteristic of H-ZSM-5 spheroid aggregates
composed by smaller individual particles [18]. This is illustrated in
Fig. S1 of the Supporting information. In contrast, the isotherms
obtained for H-ZSM-5-500 showed a modification in the hysteresis
cycle appearing at high partial pressures (between 0.5 and 0.9 P/
Po) with respect to the isotherms of H-ZSM-5-C. The more pro-
nounced hysteresis cycle is indicative of the generation of mesopo-
rosity. Importantly, however, the XRD patterns of H-ZSM-5-C and
H-ZSM-5-500, shown in Fig. S2 of the Supporting information, re-
vealed that after applying a mild hydrothermal treatment, the
overall zeolite H-ZSM-5 structure was maintained.
To investigate a possible interplay between the above-men-
tioned physicochemical variations and the amount of aluminum
in the near surface region of the zeolite H-ZSM-5 powders, XPS
sputter depth profiling measurements were performed on H-
ZSM-5-C and H-ZSM-5-500. To this end, the amount of silicon
and aluminum was measured over the first 100 nm from the zeo-
lite surface.
The results of the XPS analysis are summarized in Fig. 5 and
show a substantial decrease in the Si/Al ratio after applying a mild
hydrothermal treatment. Most notably, the average Si/Al ratio over
3.4. Magic angle spinning nuclear magnetic resonance
To obtain information on the coordination changes in alumi-
num taking place by applying a mild hydrothermal treatment,
2
7
Al MAS NMR spectroscopy was performed on H-ZSM-5-C and
H-ZSM-5-500. As depicted in Fig. 3, the NMR spectrum of H-
ZSM-5-C displays peaks that can be directly attributed to 4- and
6
0
-fold aluminum, located in the region of 60–50 [19–21] and
ppm [19–21], respectively. In the case of H-ZSM-5-500, the
NMR spectrum shows an increase in the amount of 6-fold alumi-
num at the expense of 4-fold aluminum species. Additionally, a
Fig. 4. Ammonia-Temperature-Programmed Desorption (TPD) curves of H-ZSM-5-
C (a) and H-ZSM-5-500 (b).

L.R. Aramburo et al. / Journal of Catalysis 307 (2013) 185–193
189
the first ꢀ100 nm decreased from 10.4 for H-ZSM-5-C to 8.2 for H-
ZSM-5-500. This observation indicates a surface re-alumination
after the mild hydrothermal treatment applied. This re-alumina-
tion is most probably due to the mobility of the generated extra-
framework aluminum species.
the generation of tetrahedrally coordinated extra-framework alu-
IV
minum (EFAL ) species [23–25]. These aluminum species have a
2
7
peak in the Al MAS NMR spectra at around 53 ppm and as such
they cannot be easily distinguished from the framework aluminum
(FAL) species in the above described NMR experiments.
4
.2. In situ scanning transmission X-ray microscopy of carbon deposits
4
. Nanoscale imaging of aluminum, carbon deposits and acidity
during MTH
In order to investigate in more detail, the differences in perfor-
To address the question of how the differences observed in the
mance and physicochemical properties between H-ZSM-5-C and
H-ZSM-5-500 STXM have been used to generate chemical maps
of aluminum, carbon deposits, and acidity, the latter indirectly by
using pyridine as probe molecule. The results of these three sets
of advanced STXM experiments are described below.
catalytic performances of H-ZSM-5-C and H-ZSM-5-500 are related
to nanoscopic phenomena taking place at the level of a single zeo-
lite aggregate, a MEMS nanoreactor was applied in combination
with STXM [26]. In this way, it has been possible to measure the
absorption at the carbon K-edge for H-ZSM-5-C and H-ZSM-5-
5
00 during MTH at 350 °C.
Thanks to the narrow energy width and chemical sensitivity of
4
.1. Scanning transmission X-ray microscopy of the state of aluminum
upon steaming
the transitions appearing in the carbon K-edge XAS, the comparison
of the spectral features offers a way to elucidate differences in the
nature of the hydrocarbon species formed during MTH reaction.
More specifically, among the features present in the carbon XAS,
those appearing at 285 eV, 287.6–288.2 eV, and 291–293.5 eV are
of particular interest for the MTH reaction as they correspond,
respectively, to transitions from the carbon 1s to the unoccupied
Complementary to the 27Al MAS NMR data described above,
STXM was applied to gain further insight into the modifications
taking place in the state of aluminum within zeolite H-ZSM-5 cat-
alyst powders. As shown in a recent publication [17], this can be
performed by measuring the absorption at the aluminum K-edge
in combination with a set of 4-, 5-, and 6-fold oxygen coordinated
aluminum-containing minerals. To this end, the state of aluminum
was characterized acquiring aluminum stacks of H-ZSM-5-C and
H-ZSM-5-500. Subsequently, a stack fitting was performed with a
set of aluminum reference compounds, namely albite (4-fold O-
coordinated aluminum mineral), berlinite (4-fold O-coordinated
aluminum mineral), and andalusite (mineral with 5- and 6-fold
O-coordinated aluminum).
The results obtained for H-ZSM-5-C, presented in a recent pub-
lication [17], showed that this sample mainly contained 4-fold
coordinated aluminum next to minor amounts of higher aluminum
coordination environments. More specifically, the spectral fitting
revealed a 92% and 8% contribution of albite and andalusite refer-
ence compounds, respectively, as well as the lack of contribution
from berlinite. In contrast, as shown in Fig. 6, a mild hydrothermal
treatment induces an increase in the relative coordination environ-
ment of aluminum, being in line with the 27Al MAS NMR data. In
other words, the contribution of andalusite increased to 60%, while
that of albite aluminum reference compound decreased to 20%. In
addition to the coordination change taking place upon steaming,
the 20% contribution of berlinite in the spectral fitting provides
experimental evidence for a modification in the local aluminum
environment without varying its coordination number. In this
context, previous studies have reported the presence of different
Ã
Ã
Ã
C@C
p [27–31], CAH r [30–32], and CAC r [33,34] molecular
orbitals. It should be noted, however, that in the present study,
we have limited our analysis to the spectral region going from
2
of some of the transitions appearing at higher energies in the car-
bon K-edge XAS.
83 eV up to 287 eV due to the significant mixing and degeneration
In Fig. 7a and b, the carbon K-edge XAS obtained in the early
stage of the MTH reaction (60–120 min) from the external regions
of the zeolite H-ZSM-5-C and H-ZSM-5-500 aggregates, respec-
tively, are presented. Overall, both XAS displayed different shapes,
being indicative of different catalytic reactions taking place at the
external regions of the zeolite H-ZSM-5 powders. Among the dif-
ferent transitions present in the spectra, it is noteworthy to men-
tion the broadening of the 285 eV peak in the carbon X-ray
absorption spectrum of H-ZSM-5-C, which is related to the pres-
ence of non-equivalent carbon atoms [35]. Importantly, these tran-
sitions can be used as a fingerprint of coke precursor species in the
carbon K-edge XAS since they originate from aromatic species with
high symmetry. To support this hypothesis, the carbon XAS of a se-
lect series of reference compounds containing different amounts of
non-equivalent carbon atoms are presented in Fig. 8. From this fig-
ure, it is suggested that the broadening of the 285 eV peak can be
attributed to the presence of coke precursor species, such as naph-
thalene and/or anthracene.
4
-fold coordinated aluminum species after steaming, arising from
In the later stage of the MTH reaction (280–340 min), the peak
previously present at 285 eV splits in several transitions in the car-
bon X-ray absorption spectrum of H-ZSM-5-C. This observation
indicates a significant increase in the amount of non-equivalent
carbon atoms, which can be attributed to the formation of coke
precursor species. In contrast, the carbon X-ray absorption spec-
trum of H-ZSM-5-500 was similar to that obtained for shorter reac-
tion times. In this case, the band located at 285 eV did not alter its
shape, indicating a limited formation of poly-aromatic species in
the external regions of the zeolite H-ZSM-5-500 aggregate.
4.3. Scanning transmission X-ray microscopy of Brønsted acidity
changes upon steaming
With the aim of elucidating the effect that a mild hydrothermal
treatment has on the acidic properties of a zeolite H-ZSM-5 aggre-
gate at the nanometer scale, pyridine adsorption experiments were
performed in combination with STXM. In this manner, the existence
Fig. 5. X-ray Photoelectron Spectroscopy (XPS) sputter depth profile plot showing
the Si/Al ratio as a function of the sputter depth for H-ZSM-5-C (a) and H-ZSM-5-
5
00 (b).

1
90
L.R. Aramburo et al. / Journal of Catalysis 307 (2013) 185–193
Fig. 6. Coordination and 2-D STXM maps of aluminum within H-ZSM-5-500. The different contributions have been obtained by fitting the nanoscale aluminum K-edge X-ray
absorption spectra of H-ZSM-5-500 with the aluminum K-edge X-ray absorption spectra of albite (a, red), berlinite (b, green) and andalusite (c, blue). (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of this article.)
of different degrees of pyridine interaction could be revealed by
measuring the absorption at the carbon K-edge. Among the features
present in the pyridine carbon K-edge X-ray absorption spectrum,
those appearing at 285 eV, 288 eV, and 291–293.5 eV correspond
Fig. 9a. The data reveal the existence of carbon K-edge XAS origi-
nating from the external and internal regions of the zeolite H-
ZSM-5-C aggregate. Both carbon K-edge XAS displayed a similar
shape although that originating from the external regions of the
aggregate showed a slightly higher intensity in the transitions lo-
cated at around 286.4 eV.
Fig. 9b shows the STXM data for H-ZSM-5-500, revealing the
existence of different carbon K-edge XAS spectra arising from the
internal and external regions of the zeolite aggregate. Both carbon
K-edge XAS displayed a reduction in the broadening of the peak
appearing in the range going from 284 eV up to 287 eV when com-
pared with the spectra of H-ZSM-5-C. This indicates that there are
a smaller number of different pyridine–acid site interactions in
H-ZSM-5-500 after exposing this sample to pyridine. However,
the differences between the carbon K-edge XAS collected from
the internal and external aggregate regions are more pronounced
than those observed for H-ZSM-5-C. In particular, the comparison
of the features appearing in both carbon K-edge XAS revealed a
Ã
Ã
Ã
respectively to carbon 1s ?
2 2 2
p (e u), 1s ? p (b g), and 1s ? r (e g)
transitions [36]. In contrast to the carbon X-ray absorption spec-
trum of benzene, the peak located at 285 eV in the pyridine spec-
trum is divided into a ‘‘b1’’ and ‘‘a1’’ symmetry component due to
the presence of nitrogen in the aromatic ring. The latter transitions
are highly asymmetric as a result of vibrational effects, attributed to
the different degrees of protonation of the pyridine molecule [37].
In other words, a strong protonation can bend the pyridine mole-
Ã
cule further splitting the
p
2
(e u) transitions. On this basis, a more
pronounced broadening of the carbon spectral shape in the range
going from 284 eV up to 287 eV can be explained by a wider variety
of pyridine–acid site interactions.
The results for H-ZSM-5-C, obtained from the analysis per-
formed after pyridine desorption at 400 °C, are summarized in

L.R. Aramburo et al. / Journal of Catalysis 307 (2013) 185–193
191
Fig. 9. Carbon K-edge X-ray absorption spectra of H-ZSM-5-C (a) and H-ZSM-5-500
b) with adsorbed pyridine after desorption at 400 °C. The insets of every image
(
show the cluster analysis index. The colored regions in the cluster index correspond
to the regions of the zeolite H-ZSM-5 aggregate from which the spectra with the
same color coding were obtained. (For interpretation of the references to colour in
this figure legend, the reader is referred to the web version of this article.)
Fig. 7. Carbon K-edge X-ray absorption spectra obtained during the course of the
MTH reaction at 350 °C over H-ZSM-5-C (a) and H-ZSM-5-500 (b). Spectra are
presented in black and gray for the period of 60–120 min and 280–340 min,
respectively. The inserts of every image show the cluster analysis index. In every
case, the X-ray absorption spectra were obtained from the regions depicted in red in
the cluster index. All the spectra have been normalized to the transition located at
mal treatment the core of the zeolite H-ZSM-5 aggregate somehow
preserves its acidic properties to a higher extent as compared to the
outer rim regions. In other words, STXM suggests the presence of an
egg-yolk distribution of acidity for H-ZSM-5-500.
2
85 eV. (For interpretation of the references to colour in this figure legend, the
reader is referred to the web version of this article.)
5
. Interplay between bulk performance, physicochemical
properties, nanoscale reactivity and chemical imaging
When considering the ensemble of data collected in this contri-
bution, it is interesting to note that several material parameters
have been modified as a result of steaming, in particular the mes-
opore volume and acid site density. As expected, the lower density
of acid sites in H-ZSM-5-500 was observed to correlate with a
lower catalyst activity during conventional catalytic testing
(Fig. 1a). However, the effluent selectivity of the H-ZSM-5-C and
H-ZSM-5-500 samples was very similar, and the expected decrease
in aromatics selectivity with a decrease in acid site density was not
observed (Fig. 1b–d). This observation suggests that the change in
3
acid site density observed by NH TPD and XPS analysis of the par-
Fig. 8. Carbon K-edge X-ray absorption reference spectra of benzene (a), naphtha-
lene (b), tetramethylnaphthalene (c), pyridine (d), and anthracene (e). The reference
spectra have been digitalized from Hitchcock’s group gas phase core excitation
database (http://unicorn.mcmaster.ca/).
ent and steamed samples (Figs. 4 and 5) is not sufficient to alter the
relative rates of the various hydrocarbon pool reactions leading to
effluent product formation (i.e., methylation vs. cracking vs. aro-
matization vs. dealkylation). Even so, the parent sample, H-ZSM-
5
-C, deactivated faster than the steamed sample, H-ZSM-5-500
Ã
more pronounced broadening of the carbon 1s ?
p
(e
2
u) transi-
(Fig. 1a). Furthermore, analysis of retained hydrocarbons in the
samples after testing revealed that more poly-aromatic molecules
were deposited in the parent sample compared to the steamed
sample, at the end of the test (Fig. 2). With reference to the open
literature, the difference could be ascribed to a shorter diffusion
path in the steamed sample compared to the parent sample, and
as such, a shorter residence time of aromatic products in the
crystal pores before diffusing out into the gas phase.
tions in the internal regions of the zeolite H-ZSM-5-500 aggregate.
In line with the previous reasoning, the existence of carbon
Ã
1
s ?
p
(e
2
u) transitions spread over a higher energy range in the
pyridine carbon K-edge XAS can be related to the presence of a wider
variety of interactions between pyridine and the acid sites located in
the internal regions of the zeolite H-ZSM-5 aggregate. Accordingly,
the latter observation suggests that after applying a mild hydrother-

1
92
L.R. Aramburo et al. / Journal of Catalysis 307 (2013) 185–193
Scheme 1. Proposed structure–activity relationships for small zeolite H-ZSM-5 aggregates during the MTH reaction and the beneficial effect of a mild hydrothermal
treatment on the overall catalyst stability.
It should, however, be clear that there still exists a discrepancy
between the two characterization approaches described, i.e., nano-
scale reactivity and chemical imaging vs. bulk performance and
physicochemical characterization. Indeed, as the methanol-to-
hydrocarbons reaction is a sequential process, a detailed analysis
of the active catalyst materials becomes far from trivial as the state
of the catalyst material at the top of the bed will be different from
the state of the catalyst material at the bottom. As the catalytic
reactor for the bulk activity measurements substantially differs
in e.g. dimensions and geometry from the employed nano-spectro-
scopic reaction cell [17], it is not possible to trustfully link both
sets of data. Future experiments have therefore to focus on the
proper isolation of individual catalyst particles from different bed
heights and subsequent micro-spectroscopic analysis of the coke
deposits and acidity at the single particle level with STXM. Prefer-
ably, these sets of experiments have then to be conducted at differ-
ent reaction temperatures and partial pressures as well as varying
gas velocities.
loss of a considerable number of acid sites and a re-alumination
of the zeolite H-ZSM-5 aggregate’s surface, as revealed by
ammonia TPD and XPS. STXM at the carbon K-edge on both zeo-
lite H-ZSM-5 powders after pyridine adsorption revealed a more
prominent reduction in the number and strength of the acid
sites present in the outer rim of the zeolite aggregates upon
steaming. In other words, the formation of poly-aromatic species
during MTH goes hand in hand with the core–shell chemical
maps of acidity before steaming.
Scheme 1 summarizes our current ideas related to the im-
proved MTH stability of zeolite H-ZSM-5 powders after applying
a mild hydrothermal treatment. We propose that a lower num-
ber and strength of acid sites present in the outer rim of the
zeolite aggregates in combination with the increased mesoporos-
ity translate into shorter residence time of aromatic products in
the pores before diffusing out into the gas phase. As a result,
there is a decrease in the amount of undesired chemical trans-
formations taking place within the H-ZSM-5-500 sample, which
reduces the formation of poly-aromatic compounds. In other
words, a lower production of poly-aromatic species at the outer
rim of the mildly steamed zeolites together with the formation
of mesoporosity reasonably preserves the diffusion properties
of the reacting and generated molecules, explaining the differ-
ences observed in the catalyst stability between of H-ZSM-5-C
and H-ZSM-5-500.
6
. Conclusions
The effect of a mild hydrothermal treatment on the activity,
selectivity, and stability of zeolite H-ZSM-5 catalyst powders
has been investigated during the MTH reaction at 350 °C. Small
differences in the activity and selectivity have been revealed at
the initial stages of the reaction between a calcined (i.e., H-
ZSM-5-C) and steamed (i.e., H-ZSM-5-500) sample. At longer
times on stream, however, H-ZSM-5-C showed a lower resis-
tance toward deactivation compared to H-ZSM-5-500. The main
difference in the catalytic performance of both samples arises
from the production of poly-aromatic species in the outer re-
gions of the zeolite H-ZSM-5-C aggregate, as observed with
in situ carbon K-edge STXM. The presence of poly-aromatics
in the external catalyst regions obviously hinders the transport
properties of the reacting molecules, in particular methanol,
ultimately leading to severe diffusion limitations in the MTH
process [38]. In contrast, the reduced formation of poly-aro-
matic species within the zeolite H-ZSM-5-500 aggregate pre-
serves its reactant accessibility toward the acid sites, which
has a positive impact on the overall MTH stability of the cata-
lyst material.
Acknowledgments
Dr. B. Watts and Dr. J. Raabe (both from the Swiss Light Source,
Switzerland), Dr. T. Tyliszczak and Dr. A.L.D. Kilcoyne (both from
the Lawrence Berkeley National Laboratory, USA) as well as Dr.
P.S. Miedema (Utrecht University, The Netherlands) are kindly
thanked for their contributions and fruitful discussions. K. Cats
and H. van der Bij (both from Utrecht University, The Netherlands)
are thanked for their help during STXM experiments. We thank
NRSC-C (BMW), NWO-CW Top (BMW) and NWO-CW VICI (FMFdG)
for financial support. S.T.’s Ph.D. grant is financed by inGAP, which
receives funding from the Norwegian Research Council under Con-
tract Number 174873.
Alternatively, it has been shown with aluminum K-edge
Appendix A. Supplementary material
2
7
STXM and Al MAS NMR that steaming has an impact on the
state of aluminum, inducing modifications in its coordination
as well as in its local environment. These changes result in the
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.jcat.2013.07.009.

L.R. Aramburo et al. / Journal of Catalysis 307 (2013) 185–193
193
Sepulveda-Escribano, H.W. Zandbergen, U. Olsbye, F.M.F. de Groot, B.M.
Weckhuysen, ChemCatChem 5 (2013) 1386.
References
[
[
[
[
18] V. Mayagoitia, F. Rojas, I. Kornhauser, J. Chem. Soc. Faraday Trans. 81 (1985)
[
[
[
1] A.Y. Khodakov, W. Chu, P. Fongarland, Chem. Rev. 107 (2007) 1692.
2] T. Mokrani, M. Scurrell, Catal. Rev. Sci. Eng. 51 (2009) 1.
3] M. Bjørgen, F. Joensen, K.P. Lillerud, U. Olsbye, S. Svelle, Catal. Today 142 (2009)
2
931.
19] S.M. Campbell, D.M. Bibby, J.M. Coddington, R.F. Howe, R.H. Meinhold, J. Catal.
61 (1996) 338.
1
9
0;
20] R. Jelinek, B.F. Chmelka, Y. Wu, P.J. Grandinetti, A. Pines, P.J. Barrie, J. Klinowski,
J. Am. Chem. Soc. 113 (1991) 4097.
21] S.M. Cabral de Menezes, Y.L. Lam, K. Damodaran, M. Pruski, Microporous
Mesoporous Mater. 95 (2006) 286.
22] J.X. Chen, T.H. Chen, N.J. Guan, J.Z. Wang, Catal. Today 93 (2004) 627.
23] L. Yingcai, J. Mingyang, S. Yaojun, W. Tailiu, W. Liping, F. Lun, J. Chem. Soc.
Faraday Trans. 92 (1996) 1647.
24] B.H. Wouters, T. Chen, P.J. Grobet, J. Phys. Chem. B 105 (2001) 1135.
25] A. Omegna, M. Haouas, A. Kogelbauer, R. Prins, Microporous Mesoporous
Mater. 46 (2001) 177.
26] J.F. Creemer, S. Helveg, G.H. Hoveling, S. Ullmann, A.M. Molenbroek, P.M. Sarro,
H.W. Zandbergen, Ultramicroscopy 108 (2008) 993.
27] S. Bernard, O. Beyssac, K. Benzerara, N. Findling, G. Tzvetkov, G.E. Brown,
Carbon 48 (2010) 2506.
28] E. Najafi, J.A. Wang, A.P. Hitchcock, J.W. Guan, S. Denommee, B. Simard, J. Am.
Chem. Soc. 132 (2010) 9020.
29] H.A. Katzman, P.M. Adams, T.D. Le, C.S. Hemminger, Carbon 32 (1994) 379.
30] Y. Zubavichus, A. Shaporenko, V. Korolkov, M. Grunze, M. Zharnikov, J. Phys.
Chem. B 112 (2008) 13711.
U. Olsbye, S. Svelle, M. Bjorgen, P. Beato, T.V.W. Janssens, F. Joensen, S. Bordiga,
K.P. Lillerud, Angew. Chem., Int. Ed. 51 (2012) 5810.
4] M. Stocker, Microporous Mesoporous Mater. 29 (1999) 3;
K. Hemelsoet, J. Van der Mynsbrugge, K. De Wispelaere, M. Waroquier, V. Van
Speybroeck, ChemPhysChem 14 (2013) 1526.
5] C.D. Chang, A.J. Silvestry, ChemTech. 10 (1987) 624.
6] P. Magnoux, H.S. Cerqueira, M. Guisnet, Appl. Catal. A-Gen. 235 (2002) 93.
7] H.S. Cerqueira, G. Caeiro, L. Costa, F.R. Ribeiro, J. Mol. Catal. A-Chem. 292 (2008)
[
[
[
[
[
[
[
[
1.
[
[
8] J. Raabe, G. Tzvetkov, U. Flechsig, M. Boege, A. Jaggi, B. Sarafimov, M.G.C.
Vernooij, T. Huthwelker, H. Ade, D. Kilcoyne, T. Tyliszczak, R.H. Fink, C.
Quitmann, Rev. Sci. Instrum. 79 (2008) 113704.
9] A.L.D. Kilcoyne, T. Tyliszczak, W.F. Steele, S. Fakra, P. Hitchcock, K. Franck, E.
Anderson, B. Harteneck, E.G. Rightor, G.E. Mitchell, A.P. Hitchcock, L. Yang, T.
Warwick, H. Ade, J. Synchrotron Radiat. 10 (2003) 125.
[
[
[
[
[
10] M. Guisnet, P. Magnoux, Catal. Today 36 (1997) 477.
11] K. Barbera, F. Bonino, S. Bordiga, T.V.W. Janssens, P. Beato, J. Catal. 280 (2011)
[
[
1
96.
[
[
12] M. Bjørgen, S. Kolboe, Appl. Catal. A-Gen. 225 (2002) 285.
13] S.M. Campbell, D.M. Bibby, J.M. Coddington, R.F. Howe, J. Catal. 161 (1996)
[
[
31] P.R. Haberstroh, J.A. Brandes, Y. Gelinas, A.F. Dickens, S. Wirick, G. Cody,
Geochim. Cosmochim. Acta 70 (2006) 1483.
3
50.
14] A. de Lucas, P. Canizares, A. Durán, A. Carrero, Appl. Catal. A-Gen. 154 (1997)
21.
15] M. Bjørgen, S. Svelle, F. Joensen, J. Nerlov, S. Kolboe, F. Bonino, L. Palumbo, S.
Bordiga, U. Olsbye, J. Catal. 248 (2007) 195.
16] F. Bleken, W. Skistad, K. Barbera, M. Kustova, S. Bordiga, P. Beato, K.P. Lillerud,
S. Svelle, U. Olsbye, Phys. Chem. Chem. Phys. 13 (2011) 2539.
17] L.R. Aramburo, E. de Smit, B. Arstad, M.M. van Schooneveld, L. Sommer, A.
Juhin, T. Yokosawa, H.W. Zandbergen, U. Olsbye, F.M.F. de Groot, B.M.
Weckhuysen, Angew. Chem., Int. Ed. 51 (2012) 3616;
32] D. Solomon, J. Lehmann, J. Kinyangi, B. Liang, K. Heymann, L. Dathe, K. Hanley,
S. Wirick, C. Jacobsen, Soil Sci. Soc. Am. J. 73 (2009) 1817.
33] A.P. Hitchcock, I. Ishii, J. Elec, Spec. Rel. Phenom. 42 (1987) 11.
34] A.P. Hitchcock, D.C. Newbury, I. Ishii, J. Stöhr, J.A. Horsley, R.D. Redwing, A.L.
Johnson, F. Sette, J. Chem. Phys. 85 (1986) 4849.
35] M.L. Gordon, D. Tulumello, G. Cooper, A.P. Hitchcock, P. Glatzel, O.C. Mullins,
S.P. Cramer, U. Bergmann, J. Phys. Chem. A 107 (2003) 8512.
36] W.H.E. Schwarz, T.C. Chang, U. Seeger, K.H. Hwang, Chem. Phys. 117 (1987) 73.
37] S. Aminpirooz, L. Becker, B. Hillert, J. Haase, Surf. Sci. 244 (1991) 152.
38] H. Schulz, Catal. Today 154 (2010) 183.
[
[
[
[
2
[
[
[
[
[
[
L.R. Aramburo, J. Ruiz-Martinez, L. Sommer, B. Arstad, R. Buitrago-Sierra, A.