Catalytic conversion of ethanol over ZSM-11 based catalysts

Article abstract of DOI:10.1016/j.apcata.2016.05.030

The catalytic performance of ZSM-11 zeolites was studied in ethanol conversion to produce light olefins and BTX. Compared to HZSM-5 with same SiO₂/Al₂O₃ ratio, it was found that HZSM-11 possesses significant advantages, resulting in higher stability and lower selectivity to light paraffins. The superior catalytic performance was attributed to the larger size of the intersections and the milder acid strength. Higher SiO₂/Al₂O₃ ratio of HZSM-11 led to lighter and more olefinic products, with less formation of aromatics. Higher (para + meta)/ortho-xylenes ratios were obtained over HZSM-11 with higher SiO₂/Al₂O₃ ratio, while the ratio of toluene/xylenes decreased. The yield of trimethylbenzenes (TMBs) was slightly enhanced over HZSM-11 catalysts with weaker acidity and well correlated with catalyst deactivation. Weak acidity led to faster deactivation and selectivity decay, and the growth of coke precursors inside the intersections was suggested as the main deactivation mechanism. A significant change of the hydrocarbon distribution can be obtained by modifying HZSM-11 with Fe and Ga oxides. Fe doping resulted in an increased selectivity to light olefins and C₄⁺ hydrocarbons; while Ga modification of HZSM-11 showed slightly higher selectivity toward aromatics. The stability of HZSM-11 can be significantly improved by Ga doping.

Full text of DOI:10.1016/j.apcata.2016.05.030

Accepted Manuscript
Title: Catalytic conversion of ethanol over ZSM-11 based
catalysts
Author: Liu Dapeng Liu Yan Eileen Goh Yi Ling Christina
Chu Jia Ying Chuandayani Gunawan Gwie Chang Jie
Armando Borgna
PII:
S0926-860X(16)30310-6
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2016.05.030
APCATA 15894
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
18-2-2016
26-5-2016
31-5-2016
Please cite this article as: Liu Dapeng, Liu Yan, Eileen Goh Yi Ling, Christina
Chu Jia Ying, Chuandayani Gunawan Gwie, Chang Jie, Armando Borgna, Catalytic
conversion of ethanol over ZSM-11 based catalysts, Applied Catalysis A, General
http://dx.doi.org/10.1016/j.apcata.2016.05.030
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Catalytic conversion of ethanol over ZSM-11 based catalysts
Liu Dapeng, Liu Yan, Eileen Goh Yi Ling, Christina Chu Jia Ying, Chuandayani Gunawan Gwie, Chang
Jie* and Armando Borgna*
Institute of Chemical and Engineering Sciences (ICES), A*STAR (Agency for Science, Technology and
Research), 1 Pesek Road, Jurong Island, Singapore 627833, Singapore
* Corresponding author: armando_borgna@ices.a-star.edu.sg and chang_jie@ices.a-star.edu.sg
1

Graphical abstract
110
100
90
110
100
90
Ethylene
Propylene
Butene
ZSM-11
ZSM-5
C2~4 paraffin
C5+
BTX
TMBs
Carbon Balance
40
30
20
10
0
40
30
20
10
0
0
2
4
6
8
10 12 14 16
0
2
4
6
8
10 12 14 16
Time on stream, h
Time on stream, h
With larger size of intersections, ZSM-11 provided superior catalytic performance in ethanol-to-
hydrocarbon (ETH) conversion than ZSM-5. Higher activation energy values for diffusion-controlled
reactions were observed over ZSM-11, facilitating the subsequent conversion of primary ethylene into
more value added products, i.e. propylene and BTX. In addition, the milder acidity of ZSM-11, compare
to ZSM-5 with same SiO2/Al2O3 ratio, made the relevant H-transfer and cracking reactions weaker than
that on ZSM-5, to avoid the formation of low value light paraffins. With promoted with Ga ions, the
selectivity to value-added hydrocarbons and the stability of the ZSM-11 catalyst can be improved.
2

Highlights


MEL structure provided interesting advantages as compared to ZSM-5 during ethanol conversion.
The changes in the catalytic performance were mainly ascribed the larger intersection spaces than
that of ZSM-5 with similar SiO2/Al2O3 ratio.


Fe promotion led to remarkable lower ethane and propane yield, being a promising catalyst for
light olefins. However, it suffers rapid deactivation.
Ga promotion resulted in a significant improvement on stability, along with a slight increase of
aromatics.
Abstract
The catalytic performance of ZSM-11 zeolites was studied in ethanol conversion to produce light olefins
and BTX. Compared to HZSM-5 with same SiO2/Al2O3 ratio, it was found that HZSM-11 possesses
significant advantages, resulting in higher stability and lower selectivity to light paraffins. The superior
catalytic performance was attributed to the larger size of the intersections and the milder acid strength.
Higher SiO2/Al2O3 ratio of HZSM-11 led to lighter and more olefinic products, with less formation of
aromatics. Higher (para+meta)/ortho-xylenes ratios were obtained over HZSM-11 with higher SiO2/Al2O3
ratio, while the ratio of toluene/xylenes decreased. The yield of trimethylbenzenes (TMBs) was slightly
enhanced over HZSM-11 catalysts with weaker acidity and well correlated with catalyst deactivation.
Weak acidity led to faster deactivation and selectivity decay, and the growth of coke precursors inside the
intersections was suggested as the main deactivation mechanism. A significant change of the hydrocarbon
distribution can be obtained by modifying HZSM-11 with Fe and Ga oxides. Fe doping resulted in an
+
increased selectivity to light olefins and C4 hydrocarbons; while Ga modification of HZSM-11 showed
slightly higher selectivity toward aromatics. The stability of HZSM-11 can be significantly improved by
Ga doping.
3

Keywords: Ethanol conversion; Hydrocarbons; ZSM-11; Fe modification; Ga modification
1. Introduction
Ethanol is one of the most important chemicals derived from renewable resources. Currently, bio-
ethanol accounts for almost 90% of the global biofuel production. Driven by innovations [1], world
ethanol production is rapidly increasing and may reach more than 100 billion litres [2].Ethanol can also
be considered as a renewable and versatile platform chemical to produce valuable chemicals through
catalytic processes. Bioethanol-to-hydrocarbons (ETH) is one of the most promising conversion routes in
the ethanol utilization scheme. In the ETH process, ethanol can be converted into ethylene by direct
dehydration. Subsequently, when catalysts with suitable acid and structural properties are employed,
primary product ethylene can be further converted into propylene, butylenes, and BTX (benzene, toluene,
and xylene) via oligomerisation, cracking and aromatisation reactions.
ETH can be efficiently catalysed over various solid acid catalysts such as metal oxides (alumina and
alumina-based catalysts), zeolites (ZSM-5, Beta, Faujasite), SAPO materials and heteropolyacid catalysts
[3-4,5].Indeed, ZSM-5 was widely studied since it is one of the best candidates in ETH, showing
promising selectivity to light olefins and/or higher hydrocarbons [5-6]. ZSM-11 possesses similar
framework and channel size to ZSM-5, but only intersectional straight channels are present in ZSM-11[9].
This leads to significant differences in the relative sizes of the channel intersections of these two zeolites
(50% of which are 30% larger in ZSM-11), and in the relative length of the straight and tortuous channels.
In the methanol to hydrocarbons (MTH) process, it was reported that enhanced propylene selectivity, up
to 50%, was observed by using ZSM-5/11 intergrowth zeolite [10].Considering the effect of topology on
product distribution, significant changes can be expected in the ETO reaction on ZSM-11 catalysts.
However, the catalytic conversion of alcohols over ZSM-11 was much less studied than on ZSM-5. Inoue
4

et al.[11] applied ZSM-11 in the ethanol to propylene process, reporting that 23% of propylene
selectivity can be obtained over W modified ZSM-11(120) at 500 oC. Less attention was paid to catalytic
o
performance at temperature lower than 500 C, mainly because propylene was set as the target product.
Furthermore, the direct comparison between ZSM-5 and ZSM-11 was not performed under the same
conditions.
In this contribution, we report on the catalytic behaviour of ZSM-11 catalysts during ETH. Several
characterization techniques, such as N2-adsorption, XRD, NMR, NH3-TPD, and pyridine adsorption were
used to understand the physicochemical properties of the in-house synthesised ZSM-11 catalysts. In
addition, in-situ DRIFT experiments were carried out to understand the catalytic performance. Using the
well-known ZSM-5 zeolite as a reference, the effect of the structure and acid properties on the catalytic
o
performance in the temperature range of 350-500 C is reported. The effect of the SiO2/Al2O3 ratio and
the promotion effect of transition metal ions (Fe and Ga) on the selectivity and stability of ZSM-11 is also
discussed.
2. Experimental
2.1. Catalyst preparation
A series of ZSM-11 with different SiO2/Al2O3 ratios were hydrothermally synthesised using either
tetrabutylammonium bromide (TBABr) for the SiO2/Al2O3 ratio of 30 or 1,8 diaminooctane (C8DN) for
higher SiO2/Al2O3 ratios (60 and 150) as templating agent, employing slightly modified methods
described in literatures [12,13]. For ZSM-11(30) synthesis, a gel with a molar oxide composition of
2.9Na2O:1Al2O3: 9TBABr:30SiO2:1260H2O was prepared from NaOH, NaAlO2, colloidal silica
(LUDOX HS-40) and TBABr. The gel was transferred into a Teflon lined autoclave, heated at 150 oC and
stirred at 90 rpm for
3
days. Precursor gels with molar compositions of
8.9K2O:1Al2O3:27.3C8DN:60SiO2:3K2SO4:3588H2O and 8.9K2O:1Al2O3:27.3C8DN:150SiO2:3K2SO4:
5

3588H2O were used for preparing samples ZSM-11(60) and ZSM-11(150), respectively. KOH,
Al2SO4·18H2O, C8DN, colloidal silica (LUDOX HS-40) and deionized H2O were mixed and aged at
room temperature for 24 h before hydrothermal crystallisation in autoclave under static conditions at 160
o
oC for 3 days. The resulting crystalline materials were filtered, washed and dried at 70 C overnight and
calcined at 540 oC during 6 h. To obtain the H-form zeolites, all calcined samples were exchanged with 1
o
mol/L NH4NO3 solution under reflux condition for 2 h and then calcined at 500 C for 4h. The derived
HZSM-11 zeolites were labelled as Z30, Z60, and Z150, respectively. Fe and Ga-modified catalysts were
prepared by incipient wetness impregnation (IWI) of Z30 with the corresponding aqueous solution of
nitrate salts, followed by overnight drying at 70 oC and calcination at 500 oC for 4 h. The amount of Fe2O3
and Ga2O3 loading was fixed at 5 wt%. The corresponding catalysts were labelled as Fe/Z30 and Ga/Z30,
respectively.
2.2. Catalyst characterization
Surface area and pore size volume and distribution of the catalysts were measured using N2 adsorption-
o
desorption isotherms at -196 C obtained in a Micromeritics ASAP 2420 system. Prior to the
measurements, all samples were degassed at 573 K until a stable vacuum of ca. 5 mTorr was reached. The
specific surface areas were determined employing the BET method.
X-ray diffraction (XRD) patterns of solid catalysts were obtained using a Bruker D8 X-ray
diffractometer using Cu Kα radiation and a nickel filter at 40 kV and 30 mA. The relative crystallinity
(R.C., %) of ZSM-11 samples was determined from the areas of the diffraction peaks in the 2θ range of
22-25o, taking Z30 as reference.
Solid state NMR spectra were acquired at room temperature on a Bruker Advance 400 MHz (9.4T)
27
spectrometer. The Al MAS NMR spectra were recorded at 104.26 MHz using a spinning rate of 5kHz
29
and 90o pulse length of 1.0µs with a 1s repetition time. The Si NMR spectra with high power proton
6

decoupling were obtained at 79.49 MHz employing a spinning rate of 5 kHz with a 90o pulse length of
4.5µs and repetition time of 10 s. Chemical shifts of Al and Si were measured using aluminum nitrate and
Tetramethylsilane (TMS) as references, respectively.
NH3-TPD experiments were conducted on a Thermo Scientific TPDRO 1100 instrument. 0.2 g of a
calcined sample was placed in a quartz tube and supported using quartz wool. The sample was pretreated
o
o
in Helium flow at 500 C for 2 h. Then, the temperature was decreased to 80 C, and the sample was
treated at this temperature with anhydrous ammonia gas. The sample was then flushed with He (50
mL/min) for 100 min to remove physisorbed ammonia. Temperature programmed desorption of pre-
adsorbed ammonia (NH3-TPD) was carried out from 150 oC to 800 oC.
FT-IR spectra were recorded using a Bio-Rad FT-IR 3000 MX spectrometer, equipped with a mercury
cadmium telluride (MCT) detector and a high temperature DRIFT cell fitted with KBr windows, in the
range of 4000-400 cm-1. The FT-IR spectra were recorded in transmission mode with a resolution of 4 cm-
1. Typically, 256 scans were recorded, averaged and transformed employing Kubelka-Munk method,
taking as a reference the spectra of the same zeolite before adsorption to carry out background subtraction.
o
Before DRIFT measurements, the zeolites were pretreated at 400 C in He flow for 1 h, and then were
cooled down to 35 oC. Saturated ethanol vapor at 35 oC was introduced into the cell by flowing He during
10 min, followed by He purging for 20 min. Subsequently, the cell temperature was increased in steps to
60, 200, and 400 oC under He flow. At each temperature, spectra were recorded after 10 min He purging
and stabilisation.
2.3. Catalytic measurements
Ethanol dehydration reaction was carried out in a continuous-flow fixed-bed reactor at atmospheric
pressure. Typically, 0.5 g of the catalyst was diluted with equal volume of silicon carbide. The catalyst
bed was packed between two layers of quartz wool in the isothermal zone of the reactor. Before
7

o
introducing the reactant, the catalyst was activated under flowing nitrogen at 500 C for 4 h. Catalytic
tests were performed in the temperature range of 350–500 oC. Ethanol was fed by a HPLC infusion pump
(Agilent 1100 series). Liquid ethanol was evaporated and mixed with the inert carrier gas in the
preheating zone, before entering the reactor tube. 5 % Ar mixed with the He carrier gas was used as the
internal standard, allowing for the calculation of the total molar flow of the gas products by monitoring
the Ar peak area in the TCD. Ethylene, always produced during the reaction and detected on both TCD
and FID detectors, was employed as the bridge compound to calculate the real molar flow rates for each
hydrocarbon product. The gas products were analysed by an online gas chromatography (Agilent 6890)
equipped with a flame ionization detector (FID) using a PoraBOND Q capillary column and a thermal
conductivity detector (TCD) using a Hayesep D column. The carbon balance for each GC chromatogram
was calculated by comparing the moles of C in the fed ethanol and in the hydrocarbon products. The
values of the carbon balance range from 90 % to 105 %.
3. Results and discussion
3.1. Physicochemical properties of the catalysts
The textural properties of various catalysts are summarised in Table 1. Z30 shows the highest BET
surface area and mesoporosity, while Z60 and Z150 samples possess a higher fraction of micropore and
lower surface area. This difference might be attributed to the different crystal morphology derived from
the synthesis protocols, which can be observed in the SEM images (Figure S1 in supporting information).
Fe3+ or Ga3+ impregnation causes a slight decrease in both micropore surface and pore volume, indicating
the deposition of metal cations inside the zeolite channels. However, the increased mesoporous volume
may suggest the formation of an amorphous phase, probably due to the dealumination of the framework
[14,15], being consistent with the decrease of relative crystallinity shown by XRD analysis.
8

The 27Al-MAS NMR spectra of various ZSM-11 catalysts, shown in Figure 1, provided valuable
information on the local coordination environment of Al atoms in zeolites. A dominant resonance at 56
ppm is observed for all the samples, which can be assigned to tetrahedrally coordinated aluminium (AlTd)
in the ZSM-11 framework. This is in agreement with earlier investigations using 27Al-MAS NMR of
ZSM-5 samples [16,17].The weak resonance around 0±2 ppm can only be observed for Z30-based
samples, which have the highest Al content, and it can be attributed to extra-framework aluminium in
octahedral coordination (AlOh). It should be noted that the intensity of this resonance becomes stronger
when Z30 was impregnated with Fe and Ga ions. This is on line with the higher mesoporosity revealed by
N2 adsorption, indicating the formation of the extra framework Al2O3 by dealumination.
29
The Si-MAS NMR spectra of ZSM-11 samples show an intense signal at around -112 ppm with a
shoulder at around -115 ppm, both being assigned to Si(4Si, 0Al) sites in the framework of the ZSM-11
structure [16]. The weak resonances in the range of -105 to -109 ppm can be ascribed to the Si(3Si, 1Al)
sites, namely Si in a SiO4 tetrahedral environment with 3 Si atoms and 1Al atoms. It is noteworthy that,
after Ga3+ and Fe3+ modification, a visible weak shoulder emerges at around -100 ppm, which can be
27
attributed to Si(3Si, 1OH) sites [18]. Combined with the increase of (AlOh) species shown by Al-MAS
NMR, it can be concluded that the impregnation with acid aqueous solution may partially dealuminate the
zeolite framework, forming amorphous extraframework Al and Si-OH defects. On the other hand, the
weak intensities of these newly formed species also revealed that the characteristic MEL structure is
dominantly retained after impregnation.
XRD patterns of calcined samples are presented in Figure 2. All ZSM-11samples show the typical
pattern of MEL structure, indicating that a high purity MEL zeolites have been synthesised. Furthermore,
a close inspection of the 2θ region near 45o confirms that there is only one singlet, indicating a pure MEL
structure without MFI intergrowth [19]. The crystallinity of all unmodified HZSM-11 samples is larger
than 90% as reported in Table 1, suggesting that the HZSM-11 framework is well crystallised. In good
agreement with NMR results, XRD also shows that the framework structure of ZSM-11 is retained after
9

impregnation with Fe3+ or Ga3+; however, the relative crystallinity of the samples decreases considerably,
mainly due to the dealumination induced by the aqueous solution used during impregnation. Furthermore,
crystalline phases of Fe3+ or Ga3+ species cannot be detected on the impregnated samples, suggesting that
the impregnated metal oxides are well-dispersed over the zeolite surface.
The acid properties of the ZSM-11 catalysts were investigated by both NH3-TPD and pyridine
adsorption. The NH3-TPD curves (see Figure S2) can be well deconvoluted into two peaks, being
assigned to NH3 desorption from acid sites with different strengths. Generally, it is assumed that the low-
o
temperature (LT) peak, where molecular NH3 desorbed at ca. 200 C, results from weak Lewis acid sites
or strong adsorption on silanol groups, while the high-temperature (HT) peak, where the desorption is
o
observed at ca. 400 C, can be attributed to Brønsted and strong Lewis acid sites [20,21].As shown in
Table 2, compared to HZSM-5 (30), the MEL zeolite with same SiO2/Al2O3 ratio exhibits lower acid
density. These results are in good agreement with previous studies [22]. The total acid density reduces as
the SiO2/Al2O3 ratio increases, along with the shift of the HT peak to lower temperature. Upon
impregnation with transition metal cations, no significant change of total acidity is observed, but the HT
peak significantly shifts to lower temperature.
In order to identify the nature of the acid sites, pyridine adsorption experiments were carried out. The
FTIR spectra of adsorbed pyridine are displayed in Figure 3. It is clear that Z30 shows the highest total
acidity of all HZSM-11 based samples, in well agreement with NH3-TPD results. As expected, the total
acidity of Z60 and Z150 strongly decreases, along with a significant increase of the Brønsted/Lewis (B/L)
ratio. When Z30 was doped with Ga, the total acidity significantly decreases, though the B/L ratio
remains unchanged, indicating that Ga induces similar decreases of both Brønsted and Lewis acidity.
However, when Fe was added, the value for B/L ratio interestingly increases from 1.8 to 2.4, with almost
unchanged total acid density. Correlating the pyridine adsorption results with the NH3-TPD ones, it can
be concluded that mainly weak Brønsted acid sites are generated by Fe impregnation. This change can be
attributed to the electrostatic adsorption of Fe3+ on the silanol groups, eliminating the strong Brønsted
10

sites [23]. At the same time, the framework aluminium can be extracted during Fe3+ impregnation and
new Si-OH defects are formed, generating weak Brønsted acid sites [24].
3.2. Catalytic performance of ZSM-11 catalysts
3.2.1 Effect of MEL structure
In order to explore the effect of the MEL structure, the catalytic performance of Z30 in ETH is
compared with that of HZSM-5(30), as shown in Table 3. 100% ethanol conversion is achieved, except
o
when the reaction is carried out at 350 C over HZSM-5(30), where ethanol conversions slightly lower
than 100 % are obtained; 99.6% and 96.4% ethanol conversion at 1 and 15 h of TOS, respectively. From
the viewpoint of initial selectivity, a different product distribution is observed over Z30 as compared to
HZSM-5(30). First of all, though a similar initial C2 selectivity at 400 oC is observed, the ethylene/ethane
ratio on Z30 is significantly higher. This may be mainly attributed to the lower hydrogen transfer activity
caused by the weaker acidity of Z30. This conclusion is further supported by the lower C4 hydrogen
transfer index (HTI) over Z30, which is defined as the selectivity toward butanes divided by the total
selectivity toward C4. Furthermore, the production of C3 hydrocarbons over Z30, which is exclusively
formed from cracking of oligomers [25], is less favoured than over ZSM-5. In addition, less BTX
(benzene, toluene, and xylenes) is formed over Z30. This also reveals that cracking and aromatisation
activities are also lower than that of HZSM-5(30), mainly due to the relatively lower acidity, as shown in
+
Table 2. As a consequence, higher selectivity toward C4 aliphatic hydrocarbons over Z30 is obtained.
The difference on shape selectivity on these two kinds of zeolites can be analysed comparing the
selectivity toward trimethylbenzene (TMB) and the value of the xylenes/TMB ratio. Table 3 shows that
the initial values of TMB selectivity over Z30 are slightly higher than those over HZSM-5(30), while the
BTX formation is lower. As a result, Z30 exhibits lower ratio of xylenes/TMB, particularly, at high
temperature. This is in good agreement with previous reports on MTH, where the ratio decreases from 2.5
11

on ZSM-5 to 0.5 on ZSM-11 [26]. This can be attributed to the difference in the relative sizes of the
channel intersections of the two zeolites, 50% of which are 30% larger in ZSM-11, as well as to the
difference in relative length of the straight and tortuous channels. The larger intersections and less
tortuous channels of ZSM-11 leads to the formation of TMB from lighter aromatics by facilitating the
further alkylation due to less steric hindrances[27]. However, the effect of the structure on the xylene
distribution is not as significant as in the case of MTH, where the value of (m+p)/o-xylene ratio was 13
on HZSM-5, but only 3 on HZSM-11 [23]. On the contrary, in the case of ETH, the (m+p)/o-xylene ratio
over both zeolites is almost same, being close to 3.5. The distribution of xylene isomers over both zeolites
o
is quite close to the equilibrium, 24/53/24 for p-/m-/o-xylene at 400 C [24], indicating that the xylenes
formation over both zeolites is derived from the same pool of precursors and controlled by
thermodynamics.
From the view point of stability, changes in selectivity with the TOS are observed over all tested
catalysts. While the formation of aliphatic hydrocarbons containing more than 2 carbon atoms and
aromatics decreases with the TOS, the selectivity toward ethylene increases, clearly indicating that the
required active sites for the subsequent ethylene conversion are gradually deactivated as a function of
time. However, different patterns of selectivity decay are observed over Z30 and HZSM-5(30). For
o
instance, the evolution of product distribution at 400 C displayed in Figure 4 shows a maximum for
+
propylene, butene, and C5 liquid hydrocarbons over ZSM-5(30) at 8 h on stream. After that, ethylene
produced by direct dehydration of ethanol becomes dominantly favoured, while the formation of all other
products decreases. Finally, the ethylene selectivity reaches 44.6% after 15 h on stream. In the case of
Z30, although the yield of light paraffins and aromatics shows a declining trend, the selectivity toward
propylene and butene increases during the time on stream. In addition, the oligomerisation sites are
+
deactivated much slower, resulting in an increasing net yield of C5 hydrocarbons. At the end of the
reaction, the selectivity toward ethylene gradually approaches 20.9%, being much lower than the value
observed over HZSM-5(30). The higher stability of Z30 may be ascribed to both lower acidity, which can
12

alleviate coke formation during reaction [28], and the larger space in the intersections, being more
tolerant to blockage by coking.
Due to the complexity of the product distribution, the products are lumped into 4 groups according to
their formation pathways in order to understand the evolution of product distributions as a function of the
reaction temperature. The four groups are: 1) ethylene and propylene, known as light olefins; 2) ethane
+
and propane, which are mainly formed by cracking of oligomers; 3) C4 aliphatic hydrocarbons, which are
produced by oligomerisation reactions; and 4) the pool of aromatics, produced by aromatisation of
aliphatic hydrocarbons. As shown in Figure 5, the continuous increase of light olefins and the monotonic
+
decrease of C4 and aromatics with TOS are observed as a common trend over both zeolites, though the
+
evolutions are faster over HZSM-5(30). However, the C4 yield is kept almost unchanged at temperature
lower than 400 oC over Z30.
From the view point of kinetics, the apparent activation energy for cracking and aromatisation over
MFI and MEL zeolites are calculated using the Arrhenius equation, assuming that the surface coverages
are similar over both zeolites. Significantly different activation energy (Ea) values are observed over these
two zeolites (Figure S3). The values of Ea for cracking and aromatization over HZSM-5(30) are 5.0 and
2.8 kJ/mol, respectively. They are close to the activation energy of ethylene diffusion in HZSM-5 zeolite,
6.31 kJ/mol, calculated from molecular dynamics simulation [29]. However, the corresponding values for
cracking and aromatization are 25.7 and 17.4 kJ/mol over Z30. The higher value of Ea over Z30 indicates
that the temperature dependency on the product distribution is more significant than that over HZSM-
5(30). This different temperature-dependency for Z30 and HZSM-5(30) can be attributed to the different
structure, mainly the larger intersection size in Z30.
The low values of Ea over both zeolites suggest the existence of diffusion limitations on reaction rates.
The diffusion in zeolites is usually an activated process, where the diffusing molecules cannot completely
escape from the potential field of the lattice when its molecular size is comparable with the pore size [30].
This kind of diffusion is known as in configurational regime. According to the gas translation (GT) model,
13

there is an energy barrier imposed by the channels to the diffusing molecules, which retain their gaseous
characteristics when they are moving inside the channels. The diffusive activation energy Ed can be
estimated as the difference between the potentials when the molecules stay in the channels (Φc) and in the
intersections (Φi). A higher Ed can be expected in ZSM-11 because similar Φc is expected on both zeolites
due to the similar channel diameter, while a lower Φi is expected on ZSM-11 due to the larger intersection
size.
Thus, the higher Ed value indicates that the intermediates are more strongly entrapped in the Z30
channels. Therefore, MEL structure can facilitate the formation of oligomers. This is in agreement with
previous reports on MTH, where much less C2-C3 products were observed over ZSM-11 than on ZSM-5,
while the heavy aliphatics were more abundant [26].
o
Lastly, the catalytic performance of Z30 and HZSM-5(30) at higher temperature, 500 C, was also
investigated, as shown in Figure 6. 100 % of ethanol conversion is observed over all catalysts during
time on stream. Around 20% propylene selectivity is obtained over HZSM-5(30) and Z30 which is close
to the propylene selectivity value reported by Inoue et al. [11]. However, higher ethylene and lower BTX
-1
selectivity are observed, due to the much higher space velocity employed in the present work, 9.4 g gcat
h-1. Interestingly, both HZSM-5(30) and Z30 show extremely low values of carbon balance at the initial
o
stage of the reaction. On the contrary, the results obtained at 400 C, shown in Figure 4, indicate that the
carbon balance is close to 100% for all data points, including at the beginning of the reaction. Thus, the
huge carbon loss in the initial stage of the reaction clearly indicates a severe carbon deposition at high
reaction temperatures (≥ 500 °C).
Faster selectivity decay is observed over Z30. This fact, along with the lower BTX yield observed over
Z30, may indicate that the aromatic molecules tend to be trapped in the larger intersections, leading to
coke deposition. Higher SiO2/Al2O3, Z60 gives better carbon balance in the initial stage, but an even
higher selectivity toward ethylene is obtained. Therefore, under our reaction conditions, temperatures
higher than 450 °C lead to fast catalyst deactivation.
14

3.2.2 Effects of SiO2/Al2O3 ratio
o
The spectra of adsorbed ethanol on HZSM-11 zeolites at 200 C are shown in Figure 7. Bands are
observed in the range of 1350~1480 cm-1, which is often assigned to the alkyl groups belonging to light
olefins and their oligomers, and 1500-1650 cm-1 attributed to aromatic rings [31]. A slightly more
pronounced band at 1440~1450 cm-1 is observed over Z30 than over Z60 and Z150, representing the
δ(CH) in H-bonded ethylene and propylene on Z30-based samples. These species were suggested as the
activated intermediates for the oligomerisation process [32]. On the other hand, the band at 960 cm-1
disappears in the spectra of Z60 and Z150, suggesting that ethylene is removed after flowing He for 10
min, while it is still observable over Z30-based catalysts. This could be attributed to the stronger
interaction between ethylene and acid sites and the higher acid site density, which may lead to stronger
adsorption of ethylene inside the channels.
When temperature is further increased to 400 oC (Figure 8), a higher intensity of the band at 1500 cm-1,
representing single ring aromatics, is observed on Z30 than that on Z60 and Z150. Quite clean spectra are
observed over Z60 and Z150 in the range of 1400-1700 cm-1, where only the band of polyaromatics (1598
cm-1) is observed. On the other hand, bands for SiOH groups appear at 3745 cm-1 for all catalysts.
However, the band at 3612 cm-1 due to Si(OH)Al (Brønsted sites) is only observed in the spectra of Z60
and Z150, while it is surprisingly absent on the Z30-based catalysts which possess higher density and
strength of Brønsted acid sites. This observation might indicate a different location of coke precursors.
For Z30, coke precursors may be adsorbed on strong acid sites and then further converted. For instance, it
could be cracked into single ring aromatics. While over Z60 and Z150, these polyaromatics are not
located on the Brønsted acid sites since they may desorb from the relatively weaker acid sites and be
trapped in the intersections [5].
The effect of SiO2/Al2O3 ratio on the ETH performance is studied at 400 oC and summarised in Table 4.
The initial selectivity toward ethylene becomes higher as the SiO2/Al2O3 ratio increases. Indeed, lower
15

acidity leads to a rapid decay of all products except ethylene, mainly because the secondary conversions
of ethylene are significantly inhibited. On the other hand, the decreasing values of C4 HTI, used as an
indicator of hydrogen transfer, suggests that hydrogen transfer is not favoured over HZSM-11 with high
SiO2/Al2O3 ratios. Surprisingly, the selectivity toward TMBs increases slightly on Z60 and Z150,
although the selectivity toward total aromatic becomes lower. It indicates that TMBs may be produced on
other sites than those for aromatisation. The (m+p)/o-xylene ratio increases significantly while the
toluene/xylenes ratio decreases (Figure 9), suggesting that p- and m-xylenes are primarily formed by the
aromatisation of oligomers over HZSM-11 catalysts. When abundant sites for further methyl transfer
reactions are available to produce o-xylene, the xylene distributions closely approaches the equilibrium
value. However, if the active sites exhibit lower acid strength and/or lower acid concentration, the methyl
transfer reactions are controlled by kinetics. As a result, higher fractions of p- and m-xylene are obtained.
Practically, over Z150, the formation of o-xylene cannot be detected. Thus, toluene/xylenes ratio can also
be correlated to the acidity and methyl transfer activity; lower acidity leads to lower toluene/xylenes ratio,
as shown in Figure 8. This might indicate that toluene is also a secondary product formed from xylenes
during ETH conversion.
Ethylene selectivity over Z60 and Z150 approaches 90% after 9 and 7 h of TOS, respectively.
Furthermore, the ethanol conversion over Z150 starts to decrease after few hours of TOS. This indicates
the faster loss of strong Brønsted acid sites as compared to Z30. Acetaldehyde, the dehydrogenation
product of ethanol, is not detected over Z30 and HZSM-5(30); while it is observed over Z150 and Z60
after 2 h and 8 h of TOS, respectively. Since this ethanol dehydrogenation reaction can be catalysed by
silanol sites [33] and weak Brønsted sites, it further indicates that the strong Brønsted acid sites are
rapidly deactivated during the reaction.
The accelerated deactivation during ETH over zeolites with weaker acidity is in contradiction with the
deactivation behaviour found in MTH [34], where the higher selectivity to aromatics led to faster
deactivation over ZSM-type zeolites. However, the deactivation behaviour of ZSM-5 observed in other
16

ETH studies seems to be consistent with our results. Indeed, as previously reported by Madeira et al.[5],
when HZSM-5 zeolites with SiO2/Al2O3 ratios ranging from 16 to 500 were tested in ETH at 350 oC and
30 bar, the catalysts with higher SiO2/Al2O3 ratio suffered more rapid deactivation, although significant
Brønsted acid sites were still retained in the deactivated catalysts. It was suggested that the deactivation
over high SiO2/Al2O3 ratio zeolites was caused by trapping coke precursors in the intersections, blocking
the channels, and therefore losing the accessibility to the active sites located in the channels. This is also
in the line with another deactivation study in MTH over low acidity zeolites [26]. In this study, it was
found that the initial selectivity toward TMBs can be well correlated to the catalysts deactivation. Thus, if
the heaviest fraction can be detected by the on-line gas analysis, the higher TMBs selectivity indicates a
faster deactivation. In the present contribution, the highest TMBs selectivity is observed on Z150 (8.5%),
where the BTX selectivity is only 8.7%. This may suggest the formation of coke precursors from the
aromatics which are trapped inside the intersections. In addition, from our DRIFT results at 400 oC, it can
be concluded that coke precursors, mainly polyaromatics, exist over Z60 and Z150 samples, but they are
not adsorbed on the Brønsted sites. Thus, they could be readily desorbed and trapped in the intersections.
The analysis of coke molecules by continuous wave-electron paramagnetic resonance (CW-EPR) of
deactivated ZSM-5 after ETH [5] showed that coke is composed of radical species, which might self-
catalyse the further growth in weight and size, even when they are desorbed from the acid sites and
trapped inside the intersections. The micropores, which act as the dominant channels to access the acid
sites over zeolites, can be blocked by this coke deposition [35]. Due to this mechanism, the size of the
channel would become narrower because of the trapped coke molecules with TOS. When the channel size
+
is not wide enough to allow ethanol to enter and/or C3 to exit, the reaction will be shifted mainly to the
external surface. When weak acid sites are overriding the strong ones, ethylene is predominantly
produced, and acetaldehyde can be also formed as a major product by dehydrogenation. Thus, the
formation of TMBs can be used as an indicator of coke deactivation in ETH over zeolite catalysts.
17

3.2.3 Effects of metal oxide promotion
DRIFT spectra upon ethanol adsorption at 200 oC on Fe and Ga-modified Z30 samples (Figure 7) show
distinctive bands at 1394 and 1330 cm-1, representing the pair of δ(OH…O) and δ(O-H...O)bonds in H-
assisted adsorbed ethanol [36]. Since the formation of this H-assisted adsorbed intermediate requires the
synergy between Brønsted acid sites and adjacent Lewis acid sites, it seems that the doping cations can
participate in the ethanol activation, acting as adjacent Lewis acid sites. Furthermore, higher residual
ethylene can be observed over Fe/Z30 and Ga/Z30 than on the parent zeolite, as suggested by the band at
960 cm-1. This may be caused by the existence of narrower pores due to the deposition of metal cations
inside the channels, as clearly indicated by the changes in surface area and pore volume after
o
impregnation (Table 1). When temperature is raised to 400 C (Figure 8), the spectra of Fe/Z30 and
Ga/Z30 become similar to the one of Z30. However, the band at 1500 cm-1, representing single ring
aromatics, is more significant over Ga/Z30 than Z30, while over Fe/Z30 the peak at around 1590 cm-1
becomes more intense.
The catalytic ETH performance of Fe/Z30 and Ga/Z30 was studied at 400 oC (Table 4). Over Ga/Z30
catalysts, 100% ethanol conversion is maintained during TOS, while traces of unconverted ethanol are
detected over Fe/Z30 catalyst after 15 h of TOS. The initial ethylene selectivity over Fe/Z30 is
significantly improved as compared to the parent Z30 zeolite. At the same time, the C3 yield over Fe/Z30
is also suppressed, with a much higher propylene/propane value (1.8), as compared to 0.5 over Z30. In
addition, the initial HTI value of C4 and the selectivity toward aromatics are significantly decreased. This
clearly indicates that the subsequent ethylene reactions, such as oligomerisation, cracking, hydrogen
transfer, and aromatisation, are inhibited through weakening the strength of Brønsted acid sites by Fe
doping. On the other hand, Ga doping does not change the initial product distribution significantly, except
for the slight increase of BTX selectivity.
Figure 9 shows that the xylene distribution over both Fe/Z30 and Ga/Z30 are similar to the one on Z30,
being close to the thermodynamic equilibrium. The toluene/xylenes ratio over Fe/Z30 is much lower than
18

that of Z30, while it is close to the one observed over Z150. However, higher toluene/xylene ratio is
observed over Ga/Z30, though the acidity of Ga/Z30 is lower. This might be attributed to the
demethylation activity on the extra-framework Ga cations.
The evolution of product distribution on HZSM-11-based catalysts as function of TOS is shown in
Figure 10 and Figure 11. The product distribution dramatically shifts to light olefins over Fe/Z30; from
26.9% to 72.6% after 15 h of TOS, indicating a rapid loss of active sites for the subsequent reactions.
Accordingly, the selectivity to light paraffins declines from 6.1% to 1.6%, the value of C4 HTI decreases
from 47.9 to 26.0, and the aromatic yield drops from 19.5% to 7.8%, respectively. As in the case of Z150,
acetaldehyde is detected after 9 h of TOS, and approaches 1.5% at the end of TOS. On the other hand,
Ga/Z30 shows the best stability on product distribution among all catalysts tested. During the 15 h of
TOS, light olefin selectivity increases from 8.5% to 17.7%. Cracking and aromatisation are only slightly
inhibited, as light paraffin yield decreases from 15.5% to 9.4% and aromatics from 34.0% to 25.2%
+
during the TOS. As a result, the selectivity toward C4 , i.e. oligomerisation, increases from 41.6% to 47%,
while the HTI value decreases from 83.2 to 64.2.
Although the acid properties after Ga and Fe doping are quite similar, as shown in Table 2, these two
modified zeolites exhibit pronounced differences on the catalytic performance. This should be attributed
to the own catalytic properties of the dopants. It is known that the main purpose of metal cation doping on
Z30 is to moderate the strong Brønsted acid sites by forming new Lewis acid sites, over which cracking,
hydrogen transfer, and aromatisation reactions are catalysed. However, upon Fe promotion, abundant
weak Brønsted acid sites are formed, as indicated by the decrease of the NH3 desorption temperature and
the increase of the B/L ratio. As results, the acid-catalysed H transfer reactions are suppressed [37]. The
maximum yield of aliphatic olefins is obtained at the initial stage of the reaction on Fe/Z30. At the same
time, the weakened acid strength increases the mobility of coke precursors, resulting in the accumulation
of heavy aromatics in the intersections and the blockage of the channels. Therefore, the deactivation
follows a similar mechanism to the one observed on Z150. However, although the acid strength of
19

Ga/Z30 is reduced even more than that of Fe/Z30, a totally opposite behaviour is observed. Ga/Z30
provides a similar initial product distribution to the one observed on Z30. The improved aromatic yield on
Ga/Z30, in spite of its lower acidity as compared to the parent Z30 zeolite, clearly indicates that Ga
cations act as the secondary catalytic site. It is important to note that bifunctional Ga/ZSM-5 catalysts are
commonly used for the aromatisation of light alkanes because Ga species facilitates dehydrogenation
reactions [38]. The primary role of Ga species is to act as a hydrogen remover, namely H-atom spillover
to gallium sites followed by re-combinative desorption based on hydrogen reverse spillover effect [39,40].
Therefore, the inhibition of H transfer activity by lower acid strength is neutralised. In addition, the H
spillover from Ga sites can further alleviate coke formation, leading to a significant stability improvement
on Ga/Z30.
4. Conclusions
In the catalytic conversion of ethanol to hydrocarbons, MEL structure provides interesting advantages
as compared to HZSM-5(30). In fact, HZSM-11 shows lower yield to low-value light paraffins, higher
+
+
initial selectivity toward C4 aliphatics, only mild decrease of C4 selectivity with TOS, and much lower
increase of ethylene selectivity, in spite of the structures are quite similar. These benefits can be mainly
ascribed to the lower acid density and the bigger intersection spaces than HZSM-5 with similar
SiO2/Al2O3 ratio.
As the SiO2/Al2O3 ratio increases, the product distribution significantly shifts to light olefins over
HZSM-11 zeolites, with significant inhibition of ethane and propane formation. Regarding the aromatic
distribution, higher SiO2/Al2O3 ratio leads to higher selectivity to p- and m-xylene, and lower
toluene/xylenes ratio. However, the ethylene selectivity increases rapidly with TOS. Practically, the
ethanol conversion decrease below 100% after 3 h TOS over Z150, along with the formation of
acetaldehyde. The deactivation patterns of the different ZSM-11 zeolites can be well correlated to the
initial selectivity to TMBs. The deactivation seems to be mainly due to coke accumulation in the
20

intersections, according to the TMBs selectivity values and supported by DRIFT results of ethanol
adsorption experiments.
Upon doping with Fe and Ga, the acid density is not significantly modified, but shifts to lower strength.
Fe/Z30 is recognized as a promising catalyst for light olefins, with remarkable low ethane and propane
yield. However, it suffers rapid deactivation, similar to Z150. Ga promotion resulted in a significant
improvement on stability, along with a slight increase of the aromatic yield. The difference between both
promoters is attributed to their own catalytic properties. Fe promotion mainly weakens the strength of
acid sites, inhibiting the acid-catalysed H transfer reactions. However, the mobility of coke precursors is
also enhanced, due to the elimination of strong acid sites. Therefore, the deactivation follows a similar
mechanism to the one on Z150. Since Ga has good dehydrogenation and aromatisation properties, the
inhibition of H transfer activity due to the lower acid strength is neutralised. In addition, the H spillover
from Ga sites can further alleviate the coke formation, leading to a significant stability improvement on
Ga/Z30.
5. Acknowledgements
This work was supported by the Science and Engineering Research Council (Grants 1124004031 &
1124004115) of A*STAR (Agency for Science Technology and Research), Singapore. The authors wish
also to thank Ms Ong Li Li and Mr. Ng Kang Wai Jeffrey for Solid state NMR and NH3-TPD analysis,
respectively.
21

Reference:
[1] Hahn-Hagerdal, B.; Galbe, M.; Gorwa-Grauslund, M. F.; Liden, G.; Zacchi, G. Bio-ethanol – the fuel
of tomorrow from the residues of today. Trends Biotechnol. 2006, 24, 549.
[2] Walker, G. M. Bioethanol: Science and technology of fuel alcohol. Ventus Publishing ApS, 2010
[3] Zhang, M.; Yu Y.; Dehydration of Ethanol to Ethylene.Ind. Eng. Chem. Res. 2013, 52, 9505.
[4] Stocker, M. Methanol-to-hydrocarbons: catalytic materials and their behaviour. Micropor. Mesopor.
Mater. 1999, 29, 3-48.
[5] Maderira,F.; Tayeb, K.; Pinard, L.; Vezin, H.; Maury, S.; Cadran, N. Ethanol transformation into
hydrocarbons on ZSM-5 zeolites: Influence of Si/Al ratio on catalytic performances and deactivation rate.
Study of the radical species role. Appl. Catal. A., 2012, 443-444, 171
[6] Ramesh,K.; Chang, J.; Han, Y.; Borgna, A. Synthesis, Characterization, and Catalytic Activity of
Phosphorus Modified H-ZSM-5 Catalysts in Selective Ethanol Dehydration. Ind. Eng. Chem. Res. 2010,
49, 4080.
[7] Talukdar, A.; Bhattacharyya, K.; Sivasanker, S. HZSM-5 catalysed conversion of aqueous ethanol to
hydrocarbons. Appl. Catal. A: 1997, 148, 357
[8] Kaeding, W.; Butter, S. Production of chemicals from methanol: I. Low molecular weight olefins. J.
Catal. 1980, 61, 155
[9] Xu, R.; Pang, W.; Yu, J.; Huo, Q.; Chen, J. Chemistry and related porous materials: synthesis and
structure, 1st ed., John Wiley and Sons (Asia) Pte Ltd, Singapore, 2007.
[10] Conte, M.; Xu, B.; Davies, T.; Bartley, J.; Carley, A.; Taylor, S.; Khalid, K.; Hutchings, G.
Enhanced selectivity to propene in the methanol to hydrocarbons reaction by use of ZSM-5/11
intergrowth zeolite. Micropor. Mesopor. Mater., 2012,164, 207.
22

[11] Inoue K.; Okabe K.; Inaba M.; Takahara I.; Murata K.; Catalytic conversion of ethanol to propylene
by H-ZSM-11, Reac. Kinet. Mech. Cat., 2010, 101, 227
[12] Gonzalez, G.; Gomes, M.; Vitale, G.; Castro, G. Effect of Al content on phase transitions of zeolite
MEL. Micropor. Mesopor. Mater., 2009, 121, 26.
[13]Rollmann, L.; Valyocsik, E.; U.S. Patent, 4,108,881, 1978.
[14]Chatterjee S.; Greene H.; Effects of catalyst composition on dual site zeolite catalysts used in
chlorinated hydrocarbon oxidation, Appl. Catal. A, 1993, 98, 139
[15] Amin, N.; Anggoro D.; Characterization and activity of Cr, Cu and Ga Modified ZSM-5 for direct
conversion of methane to liquid hydrocarbons, J. Natru. Gas Chem., 2003, 12, 123
[16] Fyfe, C.; Gobbi, G.; Klinowski, J.; Thomas, J.; Ramdas, S. Resolving crystallographically distinct
tetrahedral sites in silicalite and ZSM-5 by solid-state NMR. Nature, 1982, 296, 530.
[17] Kanellopoulos, J.; Unger, A.; Schwieger, W.; Freude, D. Catalytic and multinuclear MAS NMR
studies of a thermally treated zeolite ZSM-5. J. Catal., 2006, 237, 416.
[18] Sazama, P.; Dědeček, J.; Gábová, V.; Wichterlová, B.; Spoto, G.; Bordiga, S. Effect of aluminium
distribution in the framework of ZSM-5 on hydrocarbon transformation. Cracking of 1-butene. J. Catal.,
2008, 254, 180.
[19] Jablonski, G.; Sand, L.; Ward, J. Synthesis and identification of ZSM-5ZSM-11 pentasil intergrowth
structures. Zeolites, 1986, 6, 396.
[20] Sato, H.; Acidity control and catalysis of pentasil zeolites. Catal. Rev. Sci. Eng. 1997, 39, 395.
[21] Lónyi, F.; Valyon, J. On the interpretation of the NH3-TPD patterns of H-ZSM-5 and H-mordenite.
Micropor. Mesopor. Mater., 2001,47, 293.
[22] Jacobs, P.; Martens, J.; Weitkamp, J.; Beyer, H. Shape-selectivity changes in high-silica zeolites.
Faraday Disc. Chem. Soc., 1981, 72, 353.
23

[23] Belokopytov, Y.; Kholyavenko, K.; Gerei, S. An infrared study of the surface properties of metal
oxides: 2. The interaction of ammonia with the surface of Fe2O3, ZnO, MoO3, and V2O5. J. Catal., 1979,
60, 1.
[24] Sanchez, G.; Dlugogorski, B.; Kennedy, E.; Stockenhuber, M. Zeolite-supported iron catalysts for
allyl alcohol synthesis from glycerol. Appl. Catal. A., 2016, 509, 130.
[25] Song, Z.; Takahashi, A.; Nakamura, I., Fujitani, T. Phosphorus-modified ZSM-5 for conversion of
ethanol to propylene. Appl. Catal. A., 2010, 384, 201.
[26] Froment, G.; Dehertogh, W.; Marchi, A. Zeolite Catalyst in the Conversion of Methanol into Olefins.
Catalysis, vol 9, ACS, 1992
[27] Serra, J.; Guillon, E.; Corma, A. Optimizing the conversion of heavy reformate streams into xylenes
with zeolite catalysts by using knowledge base high-throughput experimentation techniques. J. Catal.,
2005, 232, 342.
[28] O’Connor, C.; Kojima, M. Alkene oligomerization. Catal Today, 1990, 6, 329.
[29] Fan J.; Xia Q.; Gong X.; Xiao H.; The influence of temperature on ethane diffusion in H[Al]ZSM-5
studied by molecular dynamics, Chem. Res. Chinese U., 2002, 18, 321
[30] Xiao, J.; Wei., J.; Diffusion mechanism of hydrocarbons in zeolites -1. Theory; Chem. Eng. Sci.,
1992, 47, 1123
[31] Forselv, S. An in-situ FTIR study on the methanol to hydrocarbons reaction over zeolite catalyst,
Ph.D Thesis, Norwegian university of Science and technology, July 2012
[32] Bleken, B.; Mino, L.; Giordanino, F.; Beato, P.; Svelle, S.; Lillerud, K.; Bordiga, P. Probing the
surface of nanosheet H-ZSM-5 with FTIR spectroscopy. Phys. Chem. Chem. Phys., 2013, 15, 13363.
[33] Matsumura, Y.; Hashimoto, K.; Watanabe, S.; Yoshida, S. Dehydrogenation of ethanol over ZSM-5
type zeolites. Chemistry Letters, 1981, 10, 121.
24

[34] Bibby, D.; Milestone, N.; Patterson, J.; Aldridge, L. Coke formation in zeolite ZSM-5. J. Catal. 1986,
97,493
[35] Pinard, L.; Ben Tayeb, K.; Hamieh, S.; Vezin, H.; Canaff, C.; Maury, S.; Delpoux, O.; Pouilloux, Y. ;
On the involvement of radical “coke” in ethanol conversion tohydrocarbons over HZSM-5 zeolite,
Catalysis Today 2013, 218-219, 57-64
[36] Zecchina, A.; Bordiga, S.; Spoto, G.; Scarano, D.; Spano, G.; Geobaldo, F. IR spectroscopy of
neutral and ionic hydrogen-bonded complexes formed upon interaction of CH3OH, C2H5OH, (CH3)2O,
(C2H5)2O and C4H8O with H-Y, H-ZSM-5 and H-mordenite: comparison with analogous adducts
formed on the H-Nafion superacidic membrane. J. Chem. Soc. Faraday, Trans., 1996, 92, 4863.
[37] Vu, D.; Hirota, Y.; Nishiyama, N.; Egashira, Y.; Ueyama, K.; High propylene selectivity in
methanol-to-olefin reaction over H-ZSM-5 catalyst treated with phosphoric acid, J. Japan, Petrol. Inst.
2010, 53, 232-238.
[38] Guisnet, M.; Gnep, N.; Aromatization of propane over GaHMFI catalysts. Reaction scheme, nature
of the dehydrogenating species and mode of coke formation, Catal. Today, 1996, 31, 275-292
[39] Steinberg, K.; Mroczek, H.; Roessner, F., Aromatization of ethane on platinum containing ZSM-5
zeolites, Appl. Catal., 1990, 66, 37-44.
[40] Le Van Mao, R.; Yao, J.; Carli, R.; Gallium containing hybrid catalysts for the aromatization of n-
butane, Appl. Catal. A, 1992, 86, 127-138.
25

Figures
27Al
29Si
Ga/Z30
Fe/Z30
Ga/Z30
Fe/Z30
Z30
Z30
Z60
Z60
Z150
Z150
200
100
0
-100
-200
-60
-80
-100 -120 -140
Chemical shift, ppm
Chemical shift, ppm
Figure 1. 27Al-MAS NMR and 29Si-MAS NMR spectra of HZSM-11 catalysts
26

Ga/Z30
Fe/Z30
Z30
Z60
Z150
ZSM-5
10
20
30
40
50
60
2Theta, o
Figure 2. XRD patterns of the calcined samples
27

B/L ratio
1.8
0.15 mmol/g
0.31 mmol/g
0.08 mmol/g
0.13 mmol/g
Ga/Z30
Fe/Z30
2.4
1.8
0.28 mmol/g
0.16 mmol/g
Z30
Z60
0.16 mmol/g
0.07 mmol/g
0.01 mmol/g
12.5
25.9
0.003 mmol/g
Z150
1700
1650
1600
1550
1500
1450
1400
Wavenumber, cm-1
Figure 3. FTIR spectra of pyridine adsorbed on HZSM-11-based samples at RT and evacuated 200 °C
28

110
100
90
110
100
90
Ethylene
Propylene
Butene
ZSM-11
ZSM-5
C2~4 paraffin
C5+
BTX
TMBs
Carbon Balance
40
30
20
10
0
40
30
20
10
0
0
2
4
6
8
10 12 14 16
0
2
4
6
8
10 12 14 16
Time on stream, h
Time on stream, h
Figure 4. Evolution of product distribution as a function of TOS over HZSM-5(30) and Z30 at 400 oC
29