Dehydration of ethanol over zeolites, silica alumina and alumina: Lewis acidity, Br?nsted acidity and confinement effects

Article abstract of DOI:10.1016/j.apcata.2014.12.047

Ethanol dehydration was investigated over commercial H-FER, H-MFI, H-MOR, H-BEA, H-Y and H-USY zeolite samples, and alumina and silica alumina for comparison. The catalysts were characterized using FT-IR spectroscopy of the surface OH groups and of adsorbed CO and pyridine. UV-vis, Raman and TG-DTA were applied to characterize coke, formed more on H-MOR and H-BEA. H-zeolites are definitely more active than silica alumina and alumina on catalyst weight base. The H-MOR sample is the most active but the H-MFI samples with Si/Al₂ ratios 280 and 50 show higher reaction rates per Al ion, H-FER and faujasites show highest ethylene yield (99.9% at 573 K). At lower temperature and higher space velocities, diethyl ether is formed with high yield (>70% at 453-473 K on H-BEA and H-MFI (50)). Overconversion of ethylene mainly to aromatics is observed on H-MFI (50). The different behaviour of protonic zeolites can predominantly be explained by confinement effects on the different zeolite cavities.

Full text of DOI:10.1016/j.apcata.2014.12.047

Accepted Manuscript
Title: Dehydration of Ethanol over Zeolites, Silica Alumina
and Alumina: Lewis Acidity, Brønsted Acidity and
Confinement Effects
´
Author: Thanh Khoa Phung Loriana Proietti Hernandez
Alberto Lagazzo Guido Busca
PII:
S0926-860X(14)00809-6
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2014.12.047
APCATA 15186
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
1-10-2014
23-12-2014
24-12-2014
´
Please cite this article as: T.K. Phung, L.P. Hernandez, A. Lagazzo, G.
Busca, DEHYDRATION OF ETHANOL OVER ZEOLITES, SILICA
ALUMINA AND ALUMINA: LEWIS ACIDITY, BRONSTED ACIDITY
AND CONFINEMENT EFFECTS, Applied Catalysis A, General (2015),
http://dx.doi.org/10.1016/j.apcata.2014.12.047
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.

DEHYDRATION OF ETHANOL OVER ZEOLITES, SILICA ALUMINA AND ALUMINA:
LEWIS ACIDITY, BRØNSTED ACIDITY AND CONFINEMENT EFFECTS
Thanh Khoa Phunga, Loriana Proietti Hernándeza,b, Alberto Lagazzoa, Guido Buscaa,*
aDepartment of Civil, Chemical and Environmental Engineering, University of Genova, p.
J.F. Kennedy 1, 16129 Genova, Italy
b on leave from Departamento de Termodinámica y Fenómenos de Transferencia,
Universidad Simón Bolívar, AP 89000, Caracas 1080, Venezuela
* Tel: (+39) 010 353 6024, Fax: (+39) 010 353 6028, Email: Guido.Busca@unige.it (G.
Busca)
 H-MOR is the most active zeolite in ethanol dehydration at low temperature.
 H-MFI and H-BEA give highest yield to diethyl ether (> 70% at 473 K).
 H-FER and H-FAU give stable 99.9 % yield to ethylene at 573 K.
 Confinement effects best explain the different behaviors of zeolites.
 Zeolites have more active and different Brønsted sites than silica alumina.
Abstract
Ethanol dehydration was investigated over commercial H-FER, H-MFI, H-MOR, H-BEA, H-
Y and H-USY zeolite samples, and alumina and silica alumina for comparison. The
catalysts were characterized using FT-IR spectroscopy of the surface OH groups and of
adsorbed CO and pyridine. UV-Vis, Raman and TG-DTA were applied to characterize
coke, formed more on H-MOR and H-BEA. H-zeolites are definitely more active than silica
alumina and alumina on catalyst weight base. The H-MOR sample is the most active but
the H-MFI samples with Si/Al2 ratios 280 and 50 show higher reaction rate per Al ion, H-
FER and faujasites show highest ethylene yield (99.9% at 573 K). At lower temperature
1
Page 1 of 42

and higher space velocities, diethyl ether is formed with high yield (> 70 % at 453-473 K
on H-BEA and H-MFI (50)). Overconversion of ethylene mainly to aromatics is observed
on H-MFI (50). The different behaviour of protonic zeolites can predominantly be explained
by confinement effects on the different zeolite cavities.
Keywords: Ethanol dehydration; zeolites; silica alumina; ethylene; diethyl ether; coking.
1
Introduction
Ethylene is the main primary intermediate in petrochemistry: world production capacity
exceeds 143,400,000 tpy [1]. Approximately 80 % of the ethylene consumed in US,
Western Europe and Japan is used for production of ethylene oxide, ethylene dichloride,
linear low density and high density polyethylene [2]. Presently, ethylene is produced
commercially primarily by two processes: the steam cracking of hydrocarbon fractions
performed in petrochemical complexes [3] and, to a lower extent, from the separation of
refinery gas mainly obtained as a byproduct of Fluid Catalytic Cracking (FCC) of heavy oils
[4]. Both processes are high energy intensive with large amounts of CO2 green-house gas
emissions.
In the frame of a future possible organic chemistry based on renewables, a number of
alternatives exist for ethylene manufacture. One of them is the MTO (Methanol To Olefins)
process [5], supposing “biomethanol” (e.g. methanol produced using syngases arising
from biomass gasification) is available. Alternatively, bioethanol produced by
lignocellulosics could be the primary intermediate to be converted into ethylene by catalytic
dehydration
C2H5OH  C2H4 + H2O
ΔH298 = + 44.9 kJ/mol
(1)
The reaction (1) is endothermic, but is already largely favoured thermodynamically at
moderate temperatures (e.g. 473-573 K). The dehydration of ethanol to ethylene has
indeed been applied at the industrial level in the sixties using aluminas as the catalysts
[6,7]. Silica alumina is reported to be more active than alumina as an acid catalyst and
also suggested for this reaction [8]. On the other hand, a number of studies reported on
the high catalytic activities of different zeolite catalysts such as H-MFI, H-BEA, H-FAU, H-
FER and H-MOR for ethanol dehydration [9,10,11,12,13,14,15,16,17]. The literature in the
2
Page 2 of 42

field has been recently reviewed by Zhang and Yu [18], that concluded that zeolites might
be unstable for this reaction. In contrast, Fan et. al. judged they are applicable at the
industrial level to produce ethylene from bioethanol [19].
Reaction (1) suffers somehow of the competition with the production of diethyl ether (DEE)
2 C2H5OH  C2H5OC2H5 + H2O
ΔH298 = - 25.1 kJ/mol
(2)
which is exothermic thus also favoured at low or moderate temperature. A number of
mechanistic studies performed with different techniques [20,21,22,23,24,25] were
published concerning ethanol dehydration on zeolites with some disagreement between
the respective conclusions.
The dehydration of ethanol is also largely used as a test reaction for surface acido-basicity
characterization [26,27,28,29]. The aim of this work is to investigate the feasibility of
ethanol dehydration to ethylene over different Brønsted and Lewis acidic solids, such as
zeolites, silica alumina and alumina, and complete [30,31] the picture of ethanol
dehydration mechanisms and the nature of the acidity of catalytically active solids.
2
Experimental.
2.1. Catalysts
Catalysts properties are summarized in Table 1. Commercial zeolites in the ammonium
+
form were calcined at 773 K for 4 hours to thermally decompose NH4 to H+ and NH3(g)
.
2.2. Catalytic experiments
Catalytic experiments were performed at atmospheric pressure in a tubular flow reactor
(i.d. 6 mm) using 0.5 g catalyst (60-70 mesh sieved, to have a ratio between the particle
diameter and the internal reactor diameter near 25) and feeding 7.9% v/v ethanol in
nitrogen with 1.43 h-1 WHSV (total flow rate of 80 cc/min). The carrier gas (nitrogen) was
passed through a bubbler containing ethanol (96%) maintained at constant temperature
(298 K) in order to obtain the desired partial pressures. The temperature in the experiment
was varied stepwise from 373 K to 623 K.
Ethanol conversion is defined as usual:
XEtOH = (nEtOH(in) – nEtOH(out))/nEtOH(in)
While selectivity to product i is defined as follows:
Si = ni/(i(nEtOH(in) – nEtOH(out)))
where ni is the moles number of compound i, and i is the ratio of stoichiometric reaction
coefficients.
3
Page 3 of 42

The outlet gases were analyzed by a gas chromatograph (GC) Agilent 4890 equipped with
a Varian capillary column “Molsieve 5A/Porabond A Tandem” and TCD and FID detectors
in series. In order to identify the compounds of the outlet gases, a gas chromatography
coupled with mass spectroscopy (GC-MS) Thermo Scientific with TG-SQC column (15 m x
0.25 mm x 0.25 m) was used.
2.3. Catalyst characterization
2.3.1. Infrared (IR) spectroscopy experiments.
IR spectra were recorded using Nicolet 380 FT-IR spectrometers. Acidity measurements
were taken using the pure powders pressed into thin wafers and activated in the IR cell
connected with a conventional outgassing/gas-manipulation apparatus by outrgassing at
773 K. The activated samples were contacted with pyridine vapor (pPy ~1 torr) at room
temperature for 15 min, when no further growth of the bands of adsorbed species was
observed; after which the IR spectra of the surface species were collected in continuous
evacuation at room temperature with increasing temperature. CO adsorption was
performed at 130 K (real sample temperature measured by a thermocouple) by the
introduction of a known dose of CO gas inside the low temperature infrared cell containing
the previously activated wafers. The sample was saturated with CO using sufficiently high
CO pressure (up to 20 Torr), until the intensity of the adsorbed species had reached the
maximum. IR spectra of the surface species were collected in continuous evacuation with
increasing temperatures between 130 and 273 K.
2.3.2. TG-DTA experiments
TG-DTA (ThermoGravimetry - Differential Thermal Analysis) experiments were performed
using a TG-DSC Netzsch Gerätebau STA 409 (Germany), with a weight sensitivity of 0.1
mg and equipped with a Netztch410 furnace temperature controller system. For each test,
about 50 mg of the sample was placed in an alumina crucible of 6 mm in diameter and
then introduced inside the furnace. All the samples were analysed in the temperature
interval 323–1073 K, with a nominal heating rate of 10 K/min in static air.
2.3.3. Raman spectroscopy analysis.
Raman spectra were collected over 3-5 mg of catalysts at room temperature on Renishaw
microscope, performing at least 3 analyses at different positions and reducing the
exposure to air for avoiding coke oxidation in case of spent catalysts. A laser of He-Ne
was used at 632.8 nm focused on the sample by a microscope.
2.3.4. UV-Vis spectroscopy analysis.
4
Page 4 of 42

UV-Vis analyses were performed using a Jasco V570 instrument, equipped with a DR
integration sphere for the analysis of spent catalysts powder. All the spectra were recorded
in air at room temperature.
3
Results.
3.1 Characterization of the surface acid sites.
To have information on the nature of the active sites involved in ethanol conversion,
surface acidity characterization studies were performed using IR spectroscopy. We used
this technique to study surface hydroxyl groups and we used molecular probes to test
acidity. In our previous work [30] we remarked that the use of very weak bases (such as
CO) only, as probes, might be not appropriate in this case because of their much lower
reactivity and much higher volatility with respect to the reactant molecule ethanol. The use
of the less volatile and stronger base pyridine may give more reliable results. Thus we
used here both probes to characterize the catalyst surfaces.
3.1.1 IR study of the surface hydroxyl groups.
The IR spectra of the OH hydroxyl groups (OHs) of zeolites are presented in Fig. 1. All
zeolites show a sharp OH stretching band (OH) at ca. 3740-3730 cm-1, which is attributed
to the weakly acidic terminal silanol groups located on the external surface [32]. One or
more bands in the range 3650-3500 cm-1 are also found, attributed to the bridging Si-OH-
Al groups that are exclusively on the inner surface and possess a strong Brønsted acidity
[32]. In the spectrum of our H-BEA sample, the intensity of the OH band of bridging Si-
OH-Al groups is very weak at 3605 cm-1, while in the spectra of H-FER, H-MFI and H-
MOR it is well evident in the range near 3600  10 cm-1. The spectra agree with those
recorded in previous studies [32,33,34]. The two faujasite zeolites were the object of
previous studies [35,36]. The spectrum of the catalyst H-USY clearly shows two OH
stretching bands due to bridging hydroxyl groups at 3627 and 3563 cm-1: the high
frequency (HF) band is due to OH groups located in the supercages (3627 cm-1), the low
frequency (LF) band to OH groups located in the sodalite cages (3563 cm-1). Their low
frequency shoulders, we observe near 3600 cm-1 and 3550 cm-1, were assigned to HF and
LF species, respectively, interacting with residual extraframework (EF) species [37]. The
spectrum of sample H-Y shows an additional split band at 3689 and 3678 cm-1, in the
region usually assigned to OH groups located on EF material, together with the band at
5
Page 5 of 42

3606 cm-1, possibly with several components, in the region of zeolitic hydroxyl groups.
This spectrum, where the distinction of the different zeolitic OHs is difficult, is quite typical
of high Al content H-FAU zeolite [38].
Some absorption in the range of OH stretching of hydroxyl groups on EF species (3700-
3650 cm-1) is found also in the other spectra. In the case of H-BEA a weak band is
observed at 3780 cm-1, also attributed to EF OH groups. Thus the presence of EF species
is evident for HY and H-BEA samples, but cannot be fully excluded in the other cases.
The spectra of alumina and silica alumina in the OH stretching region were discussed
previously [30]. The IR spectrum of the silica alumina sample only shows a strong sharp
band at ca. 3745 cm-1, with a tail extending to lower frequencies. No bands are found in
the region below 3700 cm-1.
3.1.2. IR study of CO adsorption.
In Fig. 2, the spectra of CO adsorbed on H-USY zeolite are presented. The interaction of
CO at 133 K gives rise to four different CO stretching bands (CO) at 2229, 2180, 2156
and 2139 cm-1. The band at 2229 cm-1 is attributed due to CO interacting with strong
Lewis sites, i.e. coordinatively unsaturated Al3+ ions, whose Lewis acid strength is
apparently similar to that of the strongest Lewis acid sites observed on silica alumina and
alumina [30]. The band at 2180 cm-1 is due to CO species interacting with the Brønsted
acidic bridging hydroxyl groups, in this case the OHs located in the supercage. CO
interacting with terminal silanols gives rise to the band at 2156 cm-1, while the band at
2139 cm-1 is due to nearly liquid CO.
Looking at the spectra of OH stretching region in Fig. 2, it is evident that the HF band
(3657 cm-1, OHs located in the supercage) fully disappears, and shifts down to 3286 cm-
1
after interacting with CO, while the LF band (OHs located in the sodalite cages) is
unperturbed, being these sites non accessible to CO. According to the so-called
“hydrogen bonding method” [39] arising from the Bellamy-Hallam-Williams relation, the
extent of shift down suffered by the OH band, ΔOH, can be taken as a measure of the
strength of the H-bonding interaction, thus giving a measure of the Brønsted acid
strength of the surface hydroxyl groups. The shifts down for the HF band of H-USY
zeolite, ΔOH = 347 cm-1, indicates a very strong acidity of these groups. The sites
responsible for the LF band, not accessible to CO are certainly not accessible also to
ethanol, thus are likely not active in the reaction studied here.
6
Page 6 of 42

For the materials under study, the shifting down of the OH groups after interaction with
CO, obtained by similar experiments, are reported in Table 2. In case of H-FER, H-MOR
and H-MFI (50) samples, the measured ΔOH value is also very high, but lower than that
of H-USY. In case of H-Y and H-MFI (280), the position of the perturbed band is at higher
frequency than that of other zeolites, ΔOH being lower than that of other zeolites.
However, it can be noted that the reliability of the hydrogen-bonding method to compare
the acidity of different zeolites from slightly different ΔOH values, can be not fully
reliable, being possible in the zeolite cavities additional interactions such as van der
Waals weak interactions with the zeolite walls.
On the other hand, according to this method, zeolites have much stronger Brønsted
acidity than alumina, in agreement also with the results of pyridine adsorption
experiments [30], Most of silica alumina OHs should also have lower Brønsted acidity
than those of zeolites, except perhaps a small fraction of them that, although being
terminal, gives apparently rise to extensive shift down, thus being possibly almost as
acidic as the bridging sites of most zeolites [32,39]. In addition, the CO stretching
frequency of carbon monoxide adsorbed on OHs presents, on silica alumina, two
components at 2173 and 2156 cm-1, thus at lower frequency than that found on some
zeolites and, in particular, H-USY (2180 cm-1) which appears to be the strongest
Brønsted acid. As for zeolites, these data would suggest that Brønsted acidity is quite
similar among them, H-USY being perhaps the strongest Brønsted acid with H-MFI (280)
and H-Y being the weakest ones, while all other would have similar Brønsted acidity.
Actually, small differences in the OH among zeolites may arise from additional Van
Der Waals interactions as well as from different O-H-(CO) angles caused by the
compression of the CO molecule on the opposite walls of the cavities. These effects are
possible causes of discrepancy between OHs and CO adsorption enthalpies reported
previously [40].Additionally, as proposed by Chakarova and Hadjiivanov [41,42] the
bands of OH groups interacting with CO can also perturbed by Fermi resonances in
some zeolites. In any case, the small differences in the OH among zeolites can be not
entirely due to different Brønsted OH’s acidities.
3.1.3. Surface acidity characterization by IR spectroscopy of adsorption of pyridine.
The spectra of pyridine adsorbed on silica alumina and alumina were discussed previously
[30]. Pyridine adsorbed on alumina gives rise to three 8a components observed at 1589-
7
Page 7 of 42

95, 1610-16 and 1624 cm-1, revealing the existence of at least three different families of
Lewis acid sites, and no significant Brønsted acidity. It seems likely [43,44] that the most
acidic sites are predominantly located on corners and edges in the nanocrystals of
aluminas. In the spectra of the silica alumina sample, the features of pyridine adsorbed on
very strong Lewis sites are also evident (8a mode at 1623 cm-1, 8b mode at 1579 cm-1,
19b mode at 1454-6 cm-1) together with the bands associated to pyridinium ions at 1633-
31 cm-1, 8a, and 1545 cm-1, 19a [30]. This corresponds to the well-known Brønsted acidity
of part of the hydroxyl groups of silica alumina, present together with very strong Lewis
acid sites of the Al3+ type.
The spectra of pyridine adsorbed on zeolites (Fig. 3) all present the features of pyridinium
ions (in particular the 19a band at 1545 cm-1), as the result of their Brønsted acidity. They
also present, after outgassing at r.t., bands of molecular pyridine (in particular the 19b
band at 1445-1440 cm-1, together with 8a bands ion the 1600-1595 cm-1 region), attributed
to H-bonded pyridine. After outgassing at higher temperatures (523 K) the 19b band of
Lewis bonded pyridine is observed, relatively weak, at ca. 1455 cm-1 in the case of H-FER
and H-MOR but definitely intense, relative to the 19a mode of pyridinium ions, not only in
the case of the H-Y sample (where a large amount of EF alumina certainly is present), but
also in H-BEA and in H-USY. In contrast, this band is essentially not found on the two H-
MFI samples, whose Al amount is low. It must be mentioned that, as discussed several
years ago [45], the size of the cavities of FER zeolite does not allow the easy diffusion of
pyridine at room temperature: thus the spectrum recorded on this sample in these
conditions reflects the acidity of the external surface more than that of the internal one.
Additionally, pyridine does not enter the “side pockets” of the MOR porous structure [46]
as well as the sodalite cage and the hexagonal prisms of faujasites [37,35]. Also in the
case of H-BEA, one of the two channels is expected to be diffused by pyridine with much
difficulty, being its diameter too small [33]. The pyridine adsorption experiments confirm
the presence of Brønsted sites in the case of all zeolites, but also shows in most cases
Lewis sites having comparable strengths than those of alumina and silica alumina. Only
the two H-MFI samples do not show significant amounts of pyridine bonded to Lewis acid
sites. Thus, in these cases, that also do not show significant absorption in the region of the
OH of hydroxyl groups bonded to EF material, Al ions can be supposed to be essentially
all in the framework positions. Thus for these materials it can be supposed that the number
of Al ions corresponds to that of active protons. In the other cases the observation of Lewis
8
Page 8 of 42

sites suggests that part of Al ions may be extraframework, thus not corresponding to active
protons.
3.2. Catalytic conversion of ethanol.
In Fig. 4 the effect of temperature on ethanol conversion over the zeolite catalysts is
shown. Data concerning pure alumina and silica alumina [30] are also reported for
comparison. Fig. 5 shows the selectivities (S) to C-containing products of ethanol over
zeolite catalysts as a function of reaction temperature. These experiments were performed
using the same catalysts weight with the same feed conditions.
The catalytic activity on all zeolite samples starts to be significant at 373-393 K, while on
silica alumina and alumina in the same conditions activity is observed at more than 400 K
and 450 K, respectively. In all cases DEE is the main product at low temperature and
conversion. In parallel, conversion at moderate temperature such as 473 K is much higher
on all zeolites (> 45 %) than on silica alumina and alumina (both < 20%). Also the H-MFI
(280) sample, in spite of its very small Al content, that results in a very low density of
protonic sites, is definitely more active, at least in the 373-473 K range, than alumina and
silica alumina. Thus it is evident that zeolites are definitely more active in converting
ethanol than alumina and silica alumina (in terms of activity per weight). The apparent
activation energies measured at low conversion are well above 80 kJ/mol in all cases
except FER where we measured 76 kJ/mol, showing that film diffusion limitation
phenomena (typically corresponding to 5-15 kJ apparent activation energy [47]) do not
influence the activity in the low conversion range. The internal diffusion limitations were
checked using the Weisz-Prater criterion [48] with Knudsen diffusivity calculated with
formulas obtained from the literature [49]. In all cases, we obtained value for Weisz-Prater
number lower than 0.1 at 413 K and lower temperatures, also on FER. Thus, according to
[48], pore diffusion limitation is excluded in the lower temperature range, also on FER.
Certainly, at higher temperature diffusion limitations may become relevant.
Among our zeolite samples an activity (per catalyst weight) trend H-MOR, 16.3 % > H-MFI
(50), 11.5 % > H-BEA, 7.7%  H-FER, 7.0 % > H-MFI (280), 4.6 %  H-USY, 2.5 % > H-Y,
1.2 % is found, e.g. at 413 K. This trend is certainly influenced both by the different
structures and by the different Al and proton contents of the samples. The only conclusion
is that medium pore zeolites are more active in the low conversion range than small and
large pore zeolites. In Table 3 the ethanol conversion rates per Al ions are compared for
9
Page 9 of 42

zeolite samples at 413 K. At this temperature, conversion is in the range 1-17 % on all
catalysts and it is determined by kinetics. The numbers we obtain here are closely similar
to those reported by Chiang and Bhan who used the same FER and MOR samples [25]
and referred the data to the proton content. Indeed, these values are equivalent to the
conversion rates per proton only assuming that all Al ions are in the framework and
generate an active proton each. This assumption, taken e.g. in ref. [25], is actually not
valid in most cases, because part of Al ions are EF or located at the external surface, thus
acting as Lewis acid sites, and also because part of hydroxyl groups may be unaccessible
like in the case of faujasites. In fact, the greater ethanol conversion rates per Al ion are
observed on H-MFI (280), and, to a lesser extent, on H-MFI (50), where the amount of
Lewis acidity is very small or nothing, and EF species are essentially not observed. This is
likely an evidence of the inactivity of part of protons, or of the location of part of Al ions out
of the framework, in the more Al-charged samples. It is interesting that the conversion per
Al site on H-MFI (280) is higher than on H-MFI (50), both catalysts supposed to be EF-
free. It is possible to suppose that isolated protons are most active, possibly because
several protons in the same channel can in some way hinder each other.
Over all catalysts, the selectivity to DEE is high at low temperature when also conversion
is low or moderate (Fig. 5), while the selectivity to ethylene increases by rising temperature
becoming almost total when conversion becomes very high. At low or moderate
temperatures and conversions, the selectivity to ethylene is definitely higher on H-MOR
than on other zeolites, at the same conversion / temperature, at the expense of selectivity
to diethyl ether. In fact, only H-MOR and silica alumina give rise to significant selectivity to
ethylene also at low temperature and conversion. This suggests that indeed H-MOR
structure in some ways favours ethylene production in conditions where DEE is the main
product. Thus, the highest ethanol conversion on H-MOR is not really attributed to its Al
content, relatively high, but at least in part to its structural features.
As a result, maximum yields to DEE are obtained on H-MFI (50) at 453-473 K and on H-
BEA at 473 K (> 70 %) (Fig. 6), while almost total yield (> 99 %) to ethylene is obtained on
H-FER at 523 K and more, and on faujasites, alumina and silica alumina at 573 K and
more.
Additional products, such as hydrocarbons from C3 to C5 and C6+ aromatics are obtained
in significant amount at total ethanol conversion on H-BEA, H-MOR and particularly on H-
MFI (50) (Fig. 5). Interestingly, at high temperature (> 523 K) H-MFI (50) shows a drop in
10
Page 10 of 42

ethylene selectivity and high selectivity to aromatics (toluene, xylene and high-carbon
benzene-derived). Yields to ethylene drop from 92.1 % at 523 K to 10.2 % at 623 K on H-
MFI (50), which nearly corresponds to the start of the reactivity denoted as “ethanol to
hydrocarbons” mostly reported at higher temperatures [50,51,52,53]. Interestingly, both
smaller and bigger pore size zeolites but also the low Al-content H-MFI (280) did not
produce significant amounts of aromatics.
3.3. Effect of space velocity over H-MFI (280) zeolite.
As shown above, H-MFI (280) shows the highest reaction rates per Al ion for ethanol
conversion and DEE synthesis. This is attributed to the fact that only this catalyst is
actually a “single site” one. For this catalyst only, it can be supposed that essentially all Al
ions are in the framework and give rise to an isolated “zeolitic” proton each. Thus over this
catalyst the effect of space velocity (SV) on ethanol conversion rate was investigated. The
ethanol conversion rate definitely depends on contact time, i.e. the lower the space
velocity, the greater the conversion, as expected indeed. On the other hand, the DEE
selectivity clearly has a reverse trend (Fig. 7). At lower space velocities the DEE selectivity
drops on favour of ethylene selectivity. This datum can be interpreted assuming that the
selectivities indeed depend on conversion, DEE being favoured at low conversions, i.e.
when ethanol concentration is still high. This approach agrees with the data reported by
Chiang and Bhan [25], who showed that, at least on H-MOR, DEE formation rate
increases with increasing ethanol pressure while ethylene production do the reverse, at
least in a low pressure range. Alternatively, it may be assumed that at low space velocities
the primary product DEE can convert to ethylene and water as reported elsewhere
[13,54,55]:
(C2H5)2O  2 C2H4 + H2O
(3)
To look deeper at this we considered the selectivities at the same conversion, at different
SV and temperature (Fig. 7, left). As for example, at conversions in the range 81-83 %,
the selectivity to DEE drops from 43.6 % to 8.1 % when temperature is increased from 523
to 573 K in spite of the simultaneous strong increase in space velocity from 0.7 to 7.0 h-1.
This shows that reaction temperature has definitely a stronger effect, at high conversion
levels, to favour ethylene selectivity, than conversion itself, i.e. ethanol partial pressure.
This result also suggests reaction (3) is not the predominant one in forming ethylene.
11
Page 11 of 42

In short, the best interpretation of the data is that the reaction forming ethylene is
associated to higher temperature dependence (i.e. higher activation energies) and lower
ethanol reaction order than the reaction producing DEE. This explains the different
selectivities in different conditions.
3.4. Catalyst stability study.
A study on stability on time on stream was performed on catalysts showing high yield to
ethylene, in conditions allowing conversion approaching 100 %, until 405 min. The data
are summarized in Table 4. H-USY and H-FER catalysts at 573 K show stable 99.9 %
yield to ethylene with 0.1 % yield to propene. Alumina also shows stable activity at 623 K
but the yield to ethylene 98.1-98.9 % is stably affected by coproduction of ethane and
propene.
3.5. Characterization of the exhausted catalysts.
The catalysts discharged from the reactor after the tests performed with increasing
reaction temperature (last step at 623 K) were characterized by UV-Vis, TG-DTA and
Raman spectroscopy after discharge of the reactor in air. While all fresh zeolite catalysts
are white, as expected, some of them darkened upon reaction (Fig. 8). In particular, the
spent H-MOR catalyst is very dark blue, spent H-BEA is dark grey, spent H-MFI (50) is
pale grey, while spent H-FER and H-FAU still remain almost white, just change a little bit
to yellow.
3.5.1. UV-Vis spectroscopy.
UV-Vis spectra of spent catalysts are presented in Fig. 9. According to the previous visual
inspection, the spectra of H-BEA and H-MOR show strong absorptions in the visible
region, while other zeolites show weak absorption at the scale reported in the figure (Fig. 9
left). The spectrum of spent H-MOR shows mainly a broad absorption with the maximum in
the red region, with a pronounced tail in the near IR region. The exact shape of the
absorption is made difficult to be determined by the presence of a discontinuous raise in
absorption observed at 800 nm almost in all spectra due to an instrumental artifact. In any
case some sharper maxima are superimposed to the main broad absorption at 728, 979
and 1091 nm. In contrast, the spectrum of spent H-BEA shows a very strong maximum at
413 nm, in the violet region, with less pronounced components at 494 and 586 nm, and
12
Page 12 of 42

only an absorption tail in the red region. The spectrum observed on spent H-MFI (50) is
very similar to that observed for spent H-BEA but much less intense. Also the spectra
observed on H-FER, H-Y and H-USY are closely similar but far even less intense. A
maximum near 256 nm is more evident in the spectra of H-BEA, H-MFI (50) and H-FER
than on the others.
The spectra in the visible region are due to the * transitions of unsaturated organic
species. In agreement with the visual inspection, the different spectra observed suggest
that more carbonaceous species form on H-MOR and H-BEA, less on H-MFI (50) and
even less on the other zeolites. The different spectra and color of spent H-MOR and H-
BEA point also to a different chemical structure of the carbonaceous species formed on
these catalysts. The spectrum observed on H-BEA (more intense but similar to that found
on H-MFI (50)), with a main maximum a little above 400 nm and components at higher
wavelength, is similar to that observed upon conversion of hydrocarbons and other
organics on other zeolites [56,57]: it can be attributed to linear polyaromatic species such
as e.g. tetracene and pentacene derivatives. The spectrum observed on H-MOR extends
to higher wavelengths, thus being assigned to molecules having larger  electron
delocalization, but is even broader. It can be supposed to be due to a more complex
distribution of molecules with even more aromatic rings. The band centered at 256 nm in
the spectra of H-BEA, H-MFI (50) and H-FER can be due to conjugated dienic species,
precursor of coke species [57].
3.5.2. Raman spectroscopy.
The spent catalysts were also investigated by using Laser Raman spectroscopy. Only in
the case of H-MOR a well-defined Raman spectrum has been obtained (Fig. 10),
confirming that over this catalyst a huge amount of carbonaceous material is present
Several Raman peaks are observed at ca. 1608, 1550, 1390, 1305, 1170 cm-1. The most
intense peaks at 1608 and 1550 cm-1 correspond to C=C stretching vibrations of sp2 C=C
bonds whether in chains or in aromatic rings [58,59]. The higher frequency peak is
typically observed in the Raman spectra of coked zeolites and has been assigned to
“disordered” polyaromatic coke structures, to which the peak at 1390 cm-1 (“breathing”
mode) can be associated [60,61]. The lower frequency component at 1550 cm-1 is
frequently present in graphite-like structures such as carbon nanotubes [62]. The spectrum
confirms a largely aromatic character of the carbonaceous material formed on H-MOR.
13
Page 13 of 42

3.5.3. TG-DTA experiments
The combustion of carbon species formed on the most contaminated spent catalysts has
been studied by TG-DTA. The TG weight loss and DTA curves of spent H-BEA, H-MOR
and H-MFI (50) zeolites as a function of temperature are shown in Fig. 11. It is evident that
the weight loss occurs at different temperatures for each catalyst, suggesting that there
are different species of coke and/or various deposit locations. For H-BEA zeolite, the DTA
curve indicates that a well-defined weight loss (ca 3 % wt/wt) observed starting at ca. 600
K and ending at 880 K, with corresponding an evident exothermic effect. The main heat
evolution is centered at 800 K, with apparently another component around 700 K. In the
case of H-MOR zeolite the main weight loss (> 7 % wt/wt) is observed at higher
temperature (centered at ca. 860 K), confirming a higher amount of coke. For H-MFI (50)
the weight loss is much lower (1 %) and, correspondingly, the DTA peak is also less
evident, centred at ca. 800 K. The DTA peak at ca. 800-900 K can be assigned to the
combustion of “hard” coke, essentially located in the zeolite cavities [63,64]. Maybe, the
apparent splitting of the DTA peak in H-BEA is associated to the two different families of
channels present in the structure.
4
Discussion
Looking at the overall conversion vs increasing temperature curves (Fig. 4), it is evident
that zeolites are definitely more active in converting ethanol than alumina and silica
alumina, based on catalyst weight. Conversion of ethanol mainly to DEE starts at definitely
lower temperature on zeolites, indicating that the different reactivity is not due to a
difference in surface area but to a well different and more active reaction site. This is easily
understood in the case of the comparison of alumina and zeolites. It is very evident, and
well-known indeed, that protonic zeolites display strong Brønsted acidity associated to the
presence of the bridging hydroxyl groups responsible for the well-defined OH stretching
band(s) in the region 3650-3550 cm-1. These species are associated to the protonation of
pyridine (Fig. 3) as well as to the strong interaction with CO giving rise to the shift down of
the OH stretching band (  300 cm-1, Table 2) and a CO stretching at 2174- 2180 cm-1.
Such a kind of Brønsted acidity is not present on alumina, whose OH groups are by far
14
Page 14 of 42

less acidic, but thatpresents Lewis acid sites, whose strength is, in any case, similar to that
of Lewis sites found also in the case of some zeolites.
A more complex discussion is needed when the active sites of silica alumina are
considered. Silica alumina, in fact, displays both strong Lewis acidity (pyridine bands at
1622, 1460 cm-1, adsorbed CO absorbing at 2230 cm-1) and significant Brønsted acidity
(protonation of pyridine, strong adsorption of CO). However, silica aluminas usually do not
present OH stretching bands at 3650-3550 cm-1 associated to bridging OH’s [29,30,
32,34]. This is in particular the case of our sample, which is a commercial cracking catalyst
with typical characteristics, and can be taken as a reference material. This is particularly
evident when considering the activity of the H-MFI (280) sample that, in spite of having a
small Al content, has well evident the OH band of the zeolitic hydroxyl groups at 3606
cm-1 as well as significant activity at 393 K when silica alumina is still fully inactive. Thus
the lower activity of silica alumina cannot be attributed to a low density of sites similar to
those of zeolites. As discussed previously [32,34], silica alumina OH’s are different than
those of zeolites, being essentially terminal although they may bridge upon interaction with
basic molecules and they apparently have strong Brønsted acidity approaching those of
protonic zeolites [65]. In our previous study we concluded that silica alumina has stronger
Brønsted acid sites and similarly strong Lewis acid sites as alumina [30]. On the other
hand, catalytic activities of alumina and silica alumina in ethanol dehydration are very
similar: thus it seems likely that, for ethanol dehydration, silica alumina essentially works
as a Lewis solid acid more than as a Brønsted solid acid. Thus, we attribute the catalytic
activity of protonic zeolites to their Brønsted acid sites, that of silica alumina and alumina
to their Lewis acid sites.
The comparison of the results concerning the catalytic activity and surface properties of
the zeolite catalysts provide evidence of the following points:
1. The H-MOR sample is the most active catalyst in ethanol conversion. This result
agrees with the data reported by Takahara et al. [13]. However, in contrast to the
conclusion of these authors, H-MOR is not the best catalyst for producing ethylene
with highest yields, due to the production of higher hydrocarbons at high
conversion. H-FER, H-Y and H-USY are better catalysts than H-MOR for ethylene
production, with almost total yields.
15
Page 15 of 42

2. The H-MOR sample is also the only zeolite catalyst which shows some selectivity to
ethylene also at low ethanol conversion. This also agrees with previous data
[13,25].
3. Our characterization data show that some of our zeolite catalysts present Lewis
acid sites together with Brønsted sites, thus providing evidence of extraframework
Al ions. This indicates that the assumption that all Al ions produce an active site
(see ref. [25]) is usually not valid, or at least must be ascertained by an accurate
study to demonstrate the complete framework position of Al ions. Also in the case of
our H-MFI samples, where no evidence of extraframework Al species is found,
although the conversion of ethanol is higher for sample having lower SiO2/Al2O3
ratio (again in agreement with previous data [13]), the reaction rate per Al ion is
definitely higher for the sample with higher SiO2/Al2O3 ratio. This shows that density
of sites may have a relevant effect and that Al ions cannot be considered all equally
active even when they are all in the framework.
4. At low temperature and conversion, DEE is produced with high selectivity, while at
high temperature and conversion ethylene becomes the largely predominant
product. This occurs over all catalysts independent of the nature of their acid sites
(Lewis vs Brønsted) and of their porous structure (structural or morphological
microporosity).
5. The selectivity to DEE is (at the same temperature and similar conversion level)
higher for H-MFI (50) and H-BEA than for other zeolites. These catalysts allow > 73
% yields to DEE, slightly better than that obtained in the same conditions over
WO3/TiO2 catalyst [31], but with less simultaneous production of ethylene (9.4 %
versus 14 % yield). On H-MFI (50) at 453 K 71.0 % DEE yield is obtained with only
1.4 % ethylene yield. These may be the best conditions for the highly selective
synthesis of DEE with minimum ethylene coproduction.
6. At high temperature almost total selectivity to ethylene is found on H-FER, H-Y and
H-USY, while other products (higher hydrocarbons) are observed on H-MFI (280),
H-BEA, H-MOR and, in particular, H-MFI (50), in agreement with the results of
Stepanov et al. [66] and Wang et al. [23].
7. After the increasing reaction temperature experiments, coke is produced in
significant amounts on H-MOR, less on H-BEA and much less in the other cases.
16
Page 16 of 42

Experimental data also indicate that carbon residues are larger in molecular size in
H-MOR than in the other cases.
8. The Brønsted acid strength (as measured from IR spectroscopy of CO adsorption)
seems to be similar for all zeolites, in agreement with Xu et al. [67], being possibly
slightly stronger on H-USY and slightly weaker on H-Y and H-MFI (280) than on the
other zeolites, without correlation with catalytic activity and selectivity.
The analysis of these data suggests that differences in Brønsted acid strengths do not play
a main role in determining activities and selectivities in ethanol conversion over acid
zeolite catalysts. The results are better interpreted assuming determinant roles of pore
size and morphology (Fig. 12) through the two different effects, shape selectivity and
confinement, as well as Al content.
As it is well known, shape selectivity is a key phenomenon making forbidden (or strongly
inhibited) reactions involving transition states, intermediates and/or products whose size
exceeds that of the catalyst cavities [68,69], thus somehow favouring competitive
reactions. In contrast, confinement effect directly favours reactions whose transition states
match the cavity size and are stabilized by the cavity [70].
The high ethanol conversion rates we measured for H-MOR can be attributed at least in
part to the higher Al and proton content, ex. with respect to H-MFI samples, we found to be
more active per Al ion. However, the higher ethylene vs DEE selectivity obtained on H-
MOR than on all other solids at low temperature and conversion should be associated to
unique structural features of H-MOR. These peculiar structures are likely the so-called
“side pockets” typical of this structure, as proposed also by Chiang and Bhan [25], where
very likely one only molecule of ethanol can enter, favouring ethanol conversion and
ethylene production, with respect to DEE production that obviously needs the reaction of
two molecules. Thus sites located in the side pockets can be the most active and the most
selective for unimolecular decomposition reaction of ethanol to ethylene. On larger cavities
or open channels and low ethanol conversion this reaction suffers of the competition with
the bimolecular reaction producing DEE.
On the other hand, medium pore zeolites such as H-MFI and H-BEA and, to a lesser
extent, also H-MOR display the highest yields to DEE at moderate temperatures, as a
combination of high ethanol conversion and high selectivity to the ether. It can be
supposed that medium size channels are more favourable for this reaction than both larger
cavities (such as those of faujasites) and smaller cavities such as those of ferrierite. On H-
17
Page 17 of 42

MFI (50) > 70 % yield to DEE is obtained at 453 K, with 72 % conversion and 98 %
selectivity, ethylene being the only (useful) byproduct. This catalyst seems to be very
effective for a heterogeneously catalysed ether synthesis process from (bio)ethanol. The
better performances of medium pore zeolites in forming DEE than both small and large
pore zeolites, suggest that a confinement effect favours DEE formation, additional to the
shape selectivity effect that favours ethylene formation.
The progressive decrease in DEE selectivity and increase in ethylene selectivity found at
increasing temperature and conversion in all cases is explained by the different kinetic
behaviour of the two reactions. Data obtained on H-MFI (280) support the idea that the
change in selectivities is mainly due to the higher activation energy and lower ethanol
reaction order associated to ethylene production reaction with respect to DEE reaction.
The progressive disappearance of the DEE production reaction and the ethylene
production reaction becoming largely dominant is observed on all zeolites and also over
alumina and silica alumina [30] as well as over tungstated titania and zirconia [31]. Indeed,
in all cases, the ethylene yield is really affected in practical conditions by the production of
higher hydrocarbons and by the deactivation of catalyst by coking, more than by the
competition with DEE. Thus, the competition of ethylene with DEE production, investigated
by several authors [25, 28,71], is indeed not relevant in practical conditions for ethylene
synthesis.
Very high selectivity to ethylene and yield approaching 100 % were obtained at high
temperature and conversion on small pore H-FER and on large pore H-Y and H-USY. The
selectivity to ethylene in these high or total conversion conditions is limited by the
production of higher olefins, paraffins and aromatics, occurring mostly in medium pore
zeolites (H-MFI (50), H-MOR and H-BEA). This is particularly the case of H-MFI (50),
which is in fact well-known to be particularly active in producing aromatics from several
hydrocarbon sources such as butenes [72], paraffins [73], as well as vegetable oils
[61,74,75]. Medium pore zeolites are also those producing more coke.
The high selectivity to ethylene at high conversion on H-FER can be explained by a
product shape selectivity effect, being aromatics formation and diffusion in its channels
substantially forbidden for steric reasons [72]. The small size of the ferrierite channels can
be associated also to the limited formation of oligomers of ethylene, possibly associated to
a transition state shape selectivity effect.
18
Page 18 of 42

On the other hand, the increased formation of higher hydrocarbons, aromatics and coke
on medium pore zeolites than on large pore zeolites can be associated to a confinement
effect favouring polymerizations and cyclization/aromatization, due to an optimal size of
the channels for transition states involved in these reactions. The optimal effect for
producing single oligomeric and aromatic molecules is obtained in H-MFI (50) zeolite,
while the larger channels of H-BEA and, even more, of H-MOR could be more effective for
coking. This effect almost does not exist on H-MFI (280), which material also possesses
MFI framework: this suggests that several Brønsted acid sites in the same channels are
also needed for producing this reactivity, “isolated” sites, like those of H-MFI (280), being
not active for this reaction. In addition, this confinement effect should relax in faujasites
whose supercages are far larger.
5
Conclusions.
The main results of this research are the following:
1. Protonic zeolites are definitely more active than silica alumina and alumina in
converting ethanol to DEE and ethylene. This is due to the presence of Brønsted
acidic bridging hydroxyl groups only in the zeolite cavities. Silica alumina and
alumina, that do not present these sites, are supposed to work as Lewis acid
sites in this reaction.
2. H-MOR is the only zeolite catalyst (among those investigated here) which shows
some selectivity to ethylene also at low ethanol conversion. However, according
to our data H-MOR is not the best catalyst for ethylene production at high
conversions.
3. Over all catalysts, at low temperature and conversion diethyl ether is produced with
high selectivity. On H-BEA and H-MFI (50) > 70 % yield to the ether is obtained
at 453-473 K. H-MFI (50) appears to be the best catalyst for DEE synthesis with
poor ethylene coproduction.
4. Over all these catalysts, at high temperature and conversion ethylene becomes the
largely predominant product. This is likely due to higher activation energies and
lower ethanol reaction order of the reaction forming ethylene with respect to the
reaction forming diethyl ether. However, the additional possibility of DEE
cracking to ethylene is also likely at low conversion (successive reaction path).
5. Over both H-FER and H-USY stable 99.9 % yield to ethylene is obtained at 573 K
with no deactivation at least for 405 min. These zeolites outperform alumina and
19
Page 19 of 42

other zeolites at least until this time on stream. In any case, factors affecting
ethylene yields are the coproduction of higher hydrocarbons and catalyst
deactivation by coking. The competition with DEE synthesis is not relevant in
practical conditions.
6. Over H-MFI (50) at full conversion selectivity to ethylene drops while a number of
higher hydrocarbons and high yields to aromatics (~ 29 %) are obtained. This is
not found on the low Al content zeolite H-MFI (280).
7. H-MOR and, to a lesser extent, H-BEA suffer coking much more than all other
catalysts.
8. The catalytic behaviour is interpreted as an evidence of a main role of confinement
effects, which quite evidently favour the formation of diethyl ether at low
conversion and of aromatic hydrocarbons at high conversion over medium pore
cavities. Medium size zeolite porosity seems also to favour coking. Secondarily,
shape selectivity and number of acid sites are also relevant parameters.
Differences in the strength of acid sites of zeolites are not clearly observed and
appear to be not determinant for these reactions within zeolites.
Acknowledgements
The collaboration of M.M. Carnasciali and G. Garbarino for recording Raman spectra is
acknowledged. TKP acknowledges funding by EMMA in the framework of the EU Erasmus
Mundus Action 2. The authors acknowledge University of Genova for partial funding
(PRA2013).
20
Page 20 of 42

Table 1. The properties of investigated catalysts
a
Notation
Commercial name
Manufacturer
Preparation
SiO2/Al2O3
SBET
Na (%)b
mol ratio
-
(m2/g)
~ 190
~ 330
~ 400
~ 400
~ 425
~ 500
~ 680
~ 730
~ 780
AL
Puralox Sba200
Sasol
as received
0.002
-
SA
Silica alumina (13%wt Al2O3)
CP 914C
Strem Chemicals
Zeolyst
as received
11.4
20
H-FER
Calcined at 773 K, 4h
Calcined at 773 K, 4h
Calcined at 773 K, 4h
Calcined at 773 K, 4h
Calcined at 773 K, 4h
as received
0.05
0.05
0.05
0.08
0.05
2.8
H-MFI (280) CBV 28014
Zeolyst
280
50
H-MFI (50)
H-MOR
H-BEA
H-Y
CBV 5524G
CBV 21A
CP 814E
CBV 400
CBV 720
Zeolyst
Zeolyst
20
Zeolyst
25
Zeolyst
5.1
30
H-USY
Zeolyst
as received
0.03
a, b from manufacturer
22
Page 21 of 42

Table 2. Position of OH stretching band of zeolite before and after CO adsorption
before CO
adsorption
3790, 3770, 3725
3747-3740
3635
CO (H-bond)
Samples
AL
SA
H-FER
H-MFI (280)
H-MFI (50)
H-MOR
H-BEA
H-Y
after CO adsorption
Δ
180-110
80, 167 (320)
319
2156
2173,2156
2171
3610
3667, 3580 (3415)
3316
2174
2174
2174
2178
3606
3610
3612
3605
3321
3294
3293
3300
285
316
319
305
2165
3620
3325
295
2180
H-USY
3633
3286
347
24
Page 22 of 42

Table 3. Rate of ethanol conversion per Al ion (105 (mol EtOH reacted) (mol Al)-1 s-1)
at 413 K (conversions are in the range 1-17 %).
H-FER
39.9
H-MFI (280) H-MFI (50)
336.1 155.0
H-MOR
93.0
H-BEA
53.9
H-Y
2.2
H-USY
20.7
25
Page 23 of 42

Table 4. Conversion and selectivities to C-containing products of ethanol over
alumina and zeolites as a function of time on stream.
Time
(min.) conversion
Total
S
Catalyst
S diethyl ether S ethylene S ethane
propene
45
90
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
100.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
98.9
98.8
98.8
98.7
98.5
98.1
98.7
98.5
98.5
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
99.9
100.0
99.9
99.9
99.9
99.9
99.9
99.9
99.9
1.0
1.2
1.2
1.3
1.4
1.8
1.3
1.4
1.4
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.1
0.1
0.1
0.1
0.1
0.1
0.0
0.1
0.0
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.1
0.0
0.1
0.1
0.1
0.1
0.1
0.1
0.1
135
180
225
270
315
360
405
45
AL
623 K
90
135
180
225
270
315
360
405
45
H-FER
573 K
90
135
180
225
270
315
360
405
H-USY
573 K
26
Page 24 of 42

Figure Captions
Fig. 1. FT-IR spectra of pure zeolites after activation at 773 K (OH region)
Fig. 2. FT-IR spectra of surface species (OH region) arising from CO adsorption on H-
USY zeolite. The insert: FT-IR spectra of surface species (CO region) in the same
experiment.
Fig. 3. FT-IR subtraction spectra of surface species arising from pyridine adsorbed on pure
zeolites: (a) at room temperature and (b) outgassing at 523 K.
Fig. 4. Conversion of ethanol (96%) at atmospheric pressure in a fixed-bed tubular quartz
reactor using 0.5 g of catalysts with 1.43 h-1 WHSV in nitrogen from 373 K to 623 K.
Fig. 5. Conversion and selectivities to C-containing products of ethanol over zeolites as a
function of reaction temperature (predominant compounds of others are benzene, toluene,
xylenes, trimethylbenzenes and other mononuclear aromatics).
Fig. 6. Diethyl ether yield as a function of temperature. Reaction condition: at atmospheric
pressure in a fixed-bed tubular quartz reactor using 0.5 g of catalysts with 1.43 h-1 WHSV
in nitrogen from 373 K to 623 K.
Fig. 7. Conversion and selectivities to C-containing products of ethanol over H-MFI (280)
at different space velocities (left); Conversion, diethylether selectivity and ethylene
selectivities on H-MFI (280) at 523 K as a function of space velocity (right).
Fig. 8. The photograph of spent zeolites after reaction with increasing temperature from
373 K to 623 K.
Fig. 9. UV-Vis spectra of spent zeolites after reaction with increasing temperature from 373
K to 623 K.
Fig. 10. Raman spectra of spent H-MOR zeolite after reaction with increasing temperature
from 373 K to 623 K, that was performed in continuous condition: 4000-100 cm-1, 1
accumulation and 1% laser power of He-Ne laser at 632.8 nm.
Fig. 11. TG-DTA curves of spent H-BEA, H-MOR and H-MFI after reaction with increasing
temperature from 373 K to 623 K, under air flow.
Fig. 12. Schematics of the cavities in zeolite structures (IZA database).
27
Page 25 of 42

References
[1]
[2]
[3]
[4]
W.R. True, Oil and Gas Journal, http://www.ogj.com/articles/print/volume-111/issue-
7/special-report-ethylene-report/global-ethylene-capacity-poised-for-major.html
OECD SIDS Initial assessment Profile-Ethylene; INCHIM, 2008
http://www.inchem.org/documents/sids/sids/74851.pdf
R.A. Meyers (ed.), Handbook of Petrochemicals Production Processes, Mc Graw
Hill, 2005, part 6, pp. 6.1-6.63.
The Ethylene Product Stewardship Manual, American Chemistry Council, 2004
http://www.lyondellbasell.com/techlit/techlit/Handbooks%20and%20Manuals/ACC_Ethylen
e_Manual%203096.pdf
[5]
[6]
[7]
[8]
[9]
Q. Wu, W. Xia, A. Takahashi, T. Fujitani, Adv. Mat. Res. 538-541 (2012) 2417-
2420.
Y.C. Hu, in: J.J. McKetta (Ed.), Chemical Processing Handbook, Dekker, New York,
1993, p. 768.
Ethylene from Ethanol, Chematur Engineering Group
http://www.chematur.se/sok/download/Ethylene_rev_0904.pdf (20.05.2013)
M. Caillot, A. Chaumonnot, M. Digne, J.A. van Bokhoven, J. Catal. 316 (2014) 47-
56.
R. Le Van Mao, T.M. Nguyen, G.P. McLaughlin, Appl. Catal. 48 (1989) 265-277.
[10] J.C. Oudejans, P.F. Van den Oosterkamp, H. Van Bekkum, Appl. Catal. 3 (1982)
109-115.
[11] C. de las Pozas, R. Lopez-Cordero, J.A. Gonzalez-Morales, N. Travieso, R. Roque-
Malherbe, J. Mol. Catal. 83 (1993) 145-156.
[12] C.B. Phillips, R. Datta, Ind. Eng. Chem. Res. 36 (1997) 4466-4475.
[13] I. Takahara, M. Saito, M. Inaba, K. Murata, Catal. Lett. 105 (2005) 249-252.
[14] I. Takahara, M. Saito, H. Matsuhashi, M. Inaba, K. Murata, Catal. Lett. 113 (2007)
82-85.
[15] X. Zhang, R. Wang, X. Yang, F. Zhang, Micropor. Mesopor. Mater. 116 (2008) 210-
215.
[16] K. Ramesh, L.M. Hui, Y.-F. Han, A. Borgna, Catal. Commun. 10 (2009) 567-571.
[17] J. Ouyang, F. Kong, G. Su, Y. Hu, Q. Song, Catal. Lett. 132 (2009) 64-74.
[18] M. Zhang and Y. Yu, Ind. Eng. Chem. Res. 52 (2013) 9505-9514.
[19] D. Fan, D.-J. Dai, H.-S. Wu, Mater. 6 (2013) 101-115.
[20] A. Zecchina, S. Bordiga, G. Spoto, D. Scarano, G. Spano, F. Geobaldo, J. Chem.
Soc. Faraday Trans. 92 (1996) 4863-1875.
[21] C.-C. Lee, R.J. Gorte, W.E. Farneth, J. Phys. Chem. B 101 (1997) 3811-3817.
28
Page 26 of 42

[22] J.N. Kondo, K. Ito, E. Yoda, F. Wakabayashi, K. Domen, J. Phys. Chem. B 109
(2005) 10969-10972.
[23] W. Wang, J. Jiao, Y. Jiang, S.S. Ray, M. Hunger, Chem. Phys. Chem. 6 (2005)
1467-1469.
[24] C.M. Nguyen, M.-F. Reyniers, G.B. Marin, Phys. Chem. Chem. Phys. 2 (2010)
9481-9493.
[25] H. Chiang, A. Bhan, J. Catal. 271 (2010) 251-261.
[26] J.H. Kwak, D. Mei, C.H.F. Peden, R. Rousseau, J. Szanyi, Catal. Lett. 141 (2011)
649-655.
[27] S. Roy, G. Mpourmpakis, D.-Y Hong, D.G. Vlachos, A. Bhan, R.J. Gorte, ACS
Catal. 2 (2012) 1846-1853.
[28] J.F. DeWilde, H. Chiang, D.A. Hickman, C.R. Ho, A. Bhan, ACS Catal. 3 (2013)
798-807.
[29] M. Caillot, A. Chaumonnot, M. Digne, J.A. Van Bokhoven, Chem. Cat. Chem. 6
(2014) 832-841.
[30] T.K. Phung, A. Lagazzo, M.Á. Rivero Crespo, V. Sanchez Escribano, G. Busca, J.
Catal. 311 (2014) 102-113.
[31] T. K. Phung, L. Proietti Hernandez, G. Busca,Appl. Catal.A, Gen. 489 (2015) 180-
187
[32] M. Bevilacqua, T. Montanari, E. Finocchio, G. Busca, Catal. Today 116 (2006) 132-
142.
[33] T. Montanari, M. Bevilacqua, G. Busca, Appl. Catal. A 307 (2006) 21-29.
[34] G. Busca, Chem. Rev. 107 (2007) 5366-5410.
[35] T. Montanari, E. Finocchio, G. Busca, J. Phys. Chem. C 115 (2011) 937-943.
[36] T.K. Phung, M.M. Carnasciali, E. Finocchio, G. Busca, Appl. Catal. A 470 (2014)
72-80.
[37] W. Daniell, N.-Y. Topsøe, H. Knözinger, Langmuir 17 (2001) 6233-6239.
[38] H.S. Cerqueira, P. Ayrault, J. Datka, M. Guisnet, Micropor. Mesopor. Mater. 38
(2000) 197-205.
[39] G. Busca, Catal. Today 41 (1998) 191-206.
[40] C. Otero Arean, M. R. Delgado, P.Nachtigall, Ho Viet Thang, M.Rubes, R. Bulanek,
P. Chlubna-Eliasova, Phys. Chem. Chem. Phys., 16 (2014) 10129-10141.
[41] K. Chakarova, K. Hadjiivanov, Chem. Commun. 47 (2011) 1878–1880
[42] K. Chakarova, K. Hadjiivanov, Micropor. Mesopor. Mater. 177 (2013) 59–65.
[43] G. Busca, Catal. Today 226 (2014) 2-13.
[44] T.K. Phung, C. Herrera, M. Ángeles Larrubia, M. García-Diéguez, E. Finocchio, L.J.
Alemany, G. Busca, Appl. Catal. A 483 (2014) 41-51.
[45] M. Trombetta and G. Busca, J. Catal. 187 (1999) 521-523.
29
Page 27 of 42

[46] M. Bevilacqua, A. Gutièrrez Alejandre, C. Resini, M. Casagrande, J. Ramirez and
G. Busca, Phys. Chem. Chem. Phys. 4 (2002) 4575-4583.
[47] O. Levenspiel, Chemical reaction Engineering, Wiley, 1999, pp. 390-391.
[48] M. Albert Vannice, Kinetics of Catalytic Reactions, firth ed., Springer, New York,
2005, p. 63-66.
[49] J. van den Bergh, J. Gascon, F. Kapteijn, in: J. Čejka, A. Corma, and S. Zones
(Eds.), Zeolites and Catalysis, WILEY-VCH Verlag GmbH & Co., KGaA, Weinheim, 2010,
p.363
[50] Y.I. Makarfi, M.S. Yakimova, A.S. Lermontov, V.I. Erofeev, L.M. Koval, V.F.
Tretiyakov, Chem. Eng. J. 154 (2009) 396-400.
[51] M. Inaba, K. Murata, M. Saito, I. Takahara, React. Kinet. Catal. Lett. 88(1) (2006)
135-141.
[52] K. Inoue, M. Inaba, I. Takahara, K. Murata, Catal. Lett. 136 (2010) 14-19.
[53] A.T. Aguayo, A.G. Gayubo, A.M. Tarrío, A. Atutxa, J. Bilbao, J. Chem. Technol.
Biotechnol. 77 (2002) 211-216.
[54] K. Yasunaga, F. Gillespie, J.M. Simmie, H.J. Curran, Y. Kuraguchi, H. Hoshikawa,
M. Yamane, Y. Hidaka, J. Phys. Chem. A 114 (2010) 9098–9109.
[55] M.A. Christiansen, G. Mpourmpakis, D.G. Vlachos, ACS Catal. 3 (2013) 1965-1975.
[56] L.M. Petkovic, D.M. Ginosar, K.C. Burch, J. Catal. 234 (2005) 328-339.
[57] M.J. Wulfers, F.C. Jentoft, J. Catal. 307 (2013) 204-213.
[58] B. Guichard, M. Roy-Auberger, E. Denvers, B. Rebours, A.A. Quoineaud, M. Digne,
Appl. Catal. A 367 (2009) 1-8.
[59] J. Li, G. Xiong, Z. Feng, Z. Liu, Q. Xin, C. Li, Micropor. Mesopor. Mater. 39 (2000)
275-280.
[60] P. Castaño, G. Elordi, M. Olazar, A.T. Aguayo, B. Pawelec, J. Bilbao, Appl. Catal. B
104 (2011) 91-100.
[61] J.A. Botas, D.P. Serrano, A. García, J. De Vicente, R. Ramos, Catal. Today 195
(2012) 59-70.
[62] A.M. Rao, S. Bandow, , E. Richter, P.C. Eklund, Thin Solid Films 331 (1998) 141-
147.
[63] A.R. Pradhan, J.F. Wu, S.J. Jong, T.C. Tsai, S.B. Liu, Appl. Catal. A 165 (1997)
489-497.
[64] H.S. Cerqueira, G. Caeiro, L. Costa, F. Ramoa Ribeiro, J. Mol. Catal. A Chem. 292
(2008) 1-13.
[65] F. Leydier, C. Chizallet, A. Chaumonnot, M. Digne, E. Soyer, A.-A. Quoineaud, D.
Costa, P. Raybaud, J. Catal., 284 (2011) 215-229.
[66] A.G. Stepanov, M.V. Luzgin, V.N. Romannikov, V.N. Sidelnikov, E.A. Paukshtis, J.
Catal. 178 (1998) 466-477.
30
Page 28 of 42

[67] B. Xu, C. Sievers, S.B. Hong, R. Prins, J.A. van Bokhoven, J. Catal. 244 ( 2006)
163-168.
[68] P.B. Weisz V.J. Frilette, J. Phys. Chem. 64 (1960) 382-382.
[69] B. Smit, T.L.M. Maesen, Nature 451 (2008) 671-678.
[70] R. Gounder, E. Iglesia, Acc. Chem. Res. 45 (2012) 229-238.
[71] M.A. Christiansen, G. Mpourmpakis, D.G. Vlachios, ACS Catal., 3 (2013) 1965-
1975.
[72] M. Trombetta, G. Busca, S. Rossini, V. Piccoli, U. Cornaro, A. Guercio, R. Catani,
R. J. Willey, J. Catal. 179 (1998) 581-596.
[73] A. Bhan, W.N. Delgass, Catal. Rev. – Sci. Eng.
50 (2008) 19-151.
[74] T.-A. Ngo, J.S. Kim, S.K. Kim, S.-S. Kim, Energy 35 (2010) 2723-2728.
[75] J.A. Botas, D.P. Serrano, A. García, R. Ramos, Appl. Catal. B 145 (2014) 205-215.
31
Page 29 of 42