Efficiency of ethanol conversion induced by controlled modification of pore structure and acidic properties of alumina catalysts

Article abstract of DOI:10.1016/j.apcata.2011.03.014

Bimodal porous alumina catalysts with both continuous macropores and mesopores were prepared using a dual soft template method. The addition of decahydronaphthalene as emulsifier agent successfully contributed to the formation of macropores. The effectiveness of macropore insertion was evaluated in the model reaction of ethanol dehydration at 300 °C. Judging from the results obtained by acidity measurements using temperature programmed desorption of ammonia and the performance of these catalysts in the ethanol dehydration, the introduction of macropores in the catalyst structure enhanced the ethanol conversion.

Full text of DOI:10.1016/j.apcata.2011.03.014

Applied Catalysis A: General 398 (2011) 59–65
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Efficiency of ethanol conversion induced by controlled modification of pore
structure and acidic properties of alumina catalysts
a,∗
b
a
c
Leandro Martins , Dilson Cardoso , Peter Hammer , Teresita Garetto ,
a
a
Sandra H. Pulcinelli , Celso V. Santilli
Instituto de Química, UNESP – Univ Estadual Paulista, Prof. Francisco Degni SN, 14800-900 Araraquara, SP, Brazil
Departamento de Engenharia Química, UFSCar – Univ Federal de São Carlos, Rod. Washington Luis Km 235, 13565-905 São Carlos, SP, Brazil
Instituto de Investigaciones en Catálisis y Petroquímica, (INCAPE) UNL-CONICET RA-3000 Santa Fe, Argentina
a
b
c
a r t i c l e i n f o
a b s t r a c t
Article history:
Bimodal porous alumina catalysts with both continuous macropores and mesopores were prepared using
a dual soft template method. The addition of decahydronaphthalene as emulsifier agent successfully
contributed to the formation of macropores. The effectiveness of macropore insertion was evaluated
Received 19 January 2011
Received in revised form 2 March 2011
Accepted 6 March 2011
◦
in the model reaction of ethanol dehydration at 300 C. Judging from the results obtained by acidity
Available online 11 March 2011
measurements using temperature programmed desorption of ammonia and the performance of these
catalysts in the ethanol dehydration, the introduction of macropores in the catalyst structure enhanced
the ethanol conversion.
Keywords:
Macroporous materials
Mesoporous materials
Alumina
©
2011 Elsevier B.V. All rights reserved.
Emulsion
Ethanol dehydration
1
. Introduction
porous ceramics enclosing macropores and mesopores exhibit
extremely high porosities [4,5], with a significant degree of inter-
connectivity that results in low reactor pressure-drop and high
molecular diffusion rate [6].
Continuous reactors are preferred in most industrial processes
because the reaction proceeds with the highest yield of product
and require the lowest amount of operation costs [1]. In these con-
tinuous catalytic processes granulated heterogeneous catalysts are
needed, since the drawbacks of fine powders, such as high hydraulic
head loss, is avoided.
Alumina is one of the most widely used materials in hetero-
geneous catalysis due to several attractive features [7]. Therefore,
any improvement in reactions yield is extremely profitable, con-
sidering the quantity and variety of chemicals produced through
alumina supported catalysts. Recently, we have reported a simple
and useful procedure for the preparation of hierarchically struc-
tured porous aluminas [4]. This procedure uses a combination of
the sol–gel route and a dual soft template technique involving dis-
persed oil droplets (emulsified “oil in water” system) and block
copolymer micelles. By using this strategy, a series of hierarchi-
cal alumina macro-mesoporous structures presenting controlled
amount of macroporosity were easily obtained.
This paper reports the catalytic results obtained for ethanol
dehydration on several alumina supports to evaluate the role of
mesopores and macropores on the overall reaction activity. Ethanol
dehydration was chosen as a probe reaction to develop a basic
comprehension of the effects of porosity on catalysis of alumi-
nas prepared by a dual soft templating method. Apart from the
characteristic of probe reaction ethanol conversion represents an
important environmental issue, because nowadays the search for
alternative sources of chemicals is an authentic challenge [8].
In the fabrication of these pellets the powder catalysts are
pressed into different forms such as spheres or rods. In this pelleti-
zation process most of the large macropores (typically > 0.5 ␮m)
that are formed in the interparticle region are removed while
both the mesopores and micropores prevail in the pellets [2]. The
diffusion rate of reactants and products in these small pores is low-
ered and this affects the overall kinetics of reaction especially for
those taking place in highly active catalysts [3]. In this context,
the development of porous ceramics with controlled intraparti-
cle macroporosity has attracted considerable attention as potential
catalyst supports. This is in particular the case of bimodal systems
containing macropores which are not formed in the void space
between powder particles. Furthermore, hierarchically fabricated
∗ Corresponding author. Tel.: +55 16 3301 9705; fax: +55 16 3301 9692.
E-mail address: leandro@iq.unesp.br (L. Martins).
0
926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.03.014

6
0
L. Martins et al. / Applied Catalysis A: General 398 (2011) 59–65
◦
2
. Experimental
alysts were dried at 200 C in situ under nitrogen stream before
the experiments. Data were collected every 7 min while the reac-
◦
2.1. Preparation of the porous aluminas
tion temperature was increased from 150 to 300 C. Each data point
corresponds to an average of results obtained for at least four suc-
cessive measurements.
The porous aluminas were prepared according to a pro-
cedure described in [4]. The used reagents were Pluronic
P123 (the tradename for a triblock copolymer, whose nomi-
nal formula is HO(CH CH O) (CH CH(CH )O) (CH CH O)20H,
used as surfactant and mesopore template), decahydronaph-
thalene (DHN, used as emulsion and macropore template),
2
.3. Characterization
2
2
20
2
3
70
2
2
The pore size distribution was determined by mercury intrusion
porosimetry using the AUTOPORE III equipment (Micromeritics).
All samples were degassed before analysis at a vacuum pressure
aluminum isopropoxide (Al(i-OPr) ), n-pentanol (cosurfactant)
3
−
5
below 5 × 10 Pa. The pore diameter was calculated from the
and nitric acid. A homogeneous and clear sol was obtained
with the following molar composition: 0.015Plunonic P123:1Al(i-
Washburn equation [9], using standard surface tension and con-
◦
tact angle values of 0.489 N/m and 135 , respectively. The size of
OPr) :0.1n-pentanol:1.5HNO :35H O.
3
3
2
alumina macropores was also determined from scanning electron
micrographs, recorded using a Philips XL 30 equipment. For this
purpose the samples were fixed on aluminum sample holder and
sputtered with gold.
After the sol preparation, emulsification was performed by
adding DHN under magnetic stirring for 5 min. Gelation was
induced by adding a controlled amount of NH OH solution
4
(
29 wt.%) drop by drop into this medium under mechanical stirring
The powder skeletal density (ꢀs) was determined with an
AccuPyc 1330 equipment (Micromeritics) utilizing helium gas.
Samples were purged 10 times before measurements. Ten repli-
cates were automatically made for each sample, and the mean
up to pH equal to 4.0. The quantity of DHN used was varied accord-
ing to: 0, 50, 60 and 70 wt.%, and the samples were referred as Al-1,
Al-2, Al-3 and Al-4, respectively. A reference sample denoted Al-ref
was prepared following the same procedure however without the
addition of emulsion and surfactant.
value was used. The bulk density (ꢀ ) was measured with a
b
GeoPyc 1360 equipment (Micromeritics), using a free-flowing
dry powder (DryFlo) as the displaced medium. The porosity (P%)
was calculated from the density values by using the relation
In a previous work [4] aluminas containing different DHN quan-
tities (0, 15, 30, 50, 60 and 70 wt.%) were prepared and studied in
detail regarding the preparation method. The results showed that
there was almost no difference in properties between aluminas pre-
pared with DHN content of 15 and 30 wt.%, i.e. only a slight change
in porosity was observed in comparison to the alumina reference.
On the other hand, samples prepared with DHN varying between 50
and 70 wt.% were those with largest differences in porosity. There-
fore, this work was focused only on samples in this concentration
range.
P% = (1 − ꢀ /ꢀs) × 100.
b
Nitrogen adsorption–desorption isotherms were recorded at
liquid nitrogen temperature and relative pressure interval between
0
2
1
.001 and 0.998 on the equipment supplied by Micromeritics (ASAP
◦
010). Samples were evacuated prior to measurements at 200 C for
−
5
2 h under vacuum of 1 × 10 Pa. Surface areas were calculated
following the BET equation up to P/Po = 0.3.
The crystalline phases present in calcined samples were
analyzed by X-ray diffraction (XRD) using a Siemens D5000
diffractometer and Cu K␣ radiation selected by a curved graphite
monochromator. The phase identification was done using the
program X’Pert High Score and the crystallographic pattern file
The calcination was carried out in a conventional muffle oven
◦
by increasing the temperature (5 C/min) from room temperature
◦
◦
◦
to 190 C, then heating (1 C/min) to 600 C and kept at this tem-
perature for 2 h.
Despite the use of significant amounts of decahydronaftalene as
macropore template in preparation of samples (up to 70 wt.%), this
[
04–0875] for ␥-Al O3 and [70–2038] for ␥-Al(OH)3 (gibbsite).
2
◦
Acid sites in calcined aluminas were determined by temperature
compound is completely volatilized at 190 C during calcination of
programmed desorption of ammonia (TPD-NH ) in a Micromerit-
aluminas. The thermal behavior of the as-synthesized samples was
investigated by thermo-gravimetric analysis, performed from room
3
◦
ics ChemiSorb 2705. Samples were treated in helium at 300 C
◦
for 30 min and then exposed to a 1% of ammonia (v/v in helium)
temperature up to 800 C under air flow. The results published in
◦
◦
stream at 100 C until surface saturation. The ammonia excess
Ref. [4] have shown a significant weight loss only up to 500 C. All
samples were calcined at 600 C in air environment and their visual
◦
◦
was removed with a helium flow at 100 C until a constant base-
line signal was obtained. The TPD-NH3 analyses were started by
aspect after that was of a white powder, suggesting the absence of
carbon residue.
◦
◦
heating the sample 10 C/min from 25 to 800 C under helium
flow (60 mL/min). The amount of desorbed ammonia per gram
of sample was estimated by the thermal conductivity detector
response.
2
.2. Catalytic reaction – ethanol dehydration
The XPS measurements were carried out using a commer-
cial spectrometer (UNI-SPECS UHV). The Mg K␣ line was used
The alcohol conversion was monitored in a plug flow reactor
fitted with a thermocouple extending to the center of the cata-
lyst bed. Reaction data were collected under atmospheric pressure
using 50 mg of catalyst. Liquid ethanol (99.8% Merck) was pumped
into the heated reactor (1 mL/h) using a syringe pump where it
was mixed with 25 mL/min of nitrogen gas, delivered by a mass
flow controller, to adjust the reactor feed composition. All the con-
(hꢁ = 1253.6 eV) and the analyzer pass energy was set to 10 eV. The
inelastic background of the C 1s, O 1s and Al 2p core-level spec-
tra was subtracted using Shirley’s method. The binding energies of
the spectra were corrected using the hydrocarbon component of
adventitious carbon fixed at 285.0 eV. The composition of the sur-
face layer was determined from the ratios of the relative peak areas
corrected by sensitivity factors for the corresponding elements. The
deconvoluted spectral components were obtained using multiple
Voigt profiles (70% Gaussian and 30% Lorentzian) without placing
constraints. The width at half maximum (FWHM) varied between
1.5 and 2.2 eV, and the accuracy of the peak positions was ± 0.1 eV.
The small component at high binding energy tail of the O 1s spectra,
attributed to physisorbed carboxyl groups, was subtracted from all
envelope spectra.
◦
nection lines were heated to 150 C to prevent condensation. The
composition of the reactor effluent stream was analyzed using a gas
chromatograph equipped with a flame ionization detector (FID).
The chromatograph was connected online to the reactor outlet.
A DB-1 capillary column was used in the analysis of the product
◦
stream. At a temperature of 50 C in GC analysis, ethene, acetalde-
hyde, ethanol and diethyl ether peaks were observed at 5, 5.9, 6.1
and 7.1 min, respectively. The calibration was carried out using
a mixture of reactant and products of known composition. Cat-

L. Martins et al. / Applied Catalysis A: General 398 (2011) 59–65
61
4
3
2
1
0
Representative scanning electron microscopy (SEM) images of
the alumina samples (Fig. 2) provide additional evidence for the
presence of macropores in the samples prepared with DHN. The
alumina morphology is composed of three-dimensionally inter-
connected pores with circular cross section and diameter varying
from 0.1 to 2 ␮m. Despite the variation of the macropores diame-
ter they were found to show a relatively narrow size distribution
centered on 1 ␮m. It is important to note that the alumina pore
structure presents a good thermal stability, thus after calcination in
macro-
mesopores
Al-4
Al-3
Al-2
Al-1
Al-ref
0.01
◦
air at 600 C the porous structure remained stable and no significant
amount of cracking was observed.
These experimental results demonstrate that aluminas with
pores at two distinct length scales can be produced using adequate
synthesis conditions. Additionally, the findings obtained for the set
of conditions used in this study suggest that the formation of these
two different characteristic pore morphologies can be decoupled
since they can be individually adjusted by controlling the surfactant
or the oil concentration.
1
0
1
0.1
Diameter (µm)
Fig. 1. Effect of DHN amount on the cumulative pore volume determined by Hg-
porosimetry.
Table 1 displays some characteristics of the porous aluminas
such as macro/mesopore volume, porosity and BET surface area,
including the reference sample, Al-ref that was prepared without
addition of surfactant and DHN to evidence the additive contribu-
tion of both templates in generating the observed porosity. It can
be observed that the addition of DHN significantly increases the
3
. Results and discussion
3.1. Porous aluminas characterization
The effects of DHN amount on the cumulative pore size distribu-
3
tion of the calcined aluminas are presented in Fig. 1. A bimodal size
distribution, named as macro- (close to 1 ␮m) and meso- (in the
range of 8–9 nm) pores families can be observed. Sample Al-1 was
prepared without DHN addition, i.e. containing only the surfactant,
for comparison with samples Al-2 to Al-4. The pore size distribution
in the mesopore region is present in all samples from Al-1 to Al-3,
and we attribute this pore family to the micelle templating process.
Sample Al-4 presents only the macropore family, despite the addi-
tion of the surfactant in the gel synthesis. A possible explanation
for the absence of the mesopore family in this sample is the migra-
tion of surfactants from the micelles to the oil/water interface. This
is comprehensible since the amount of DHN is quite high, 70 wt.%.
Furthermore, a considerable increase of the specific macropore vol-
ume can be observed when increasing the DHN quantity from 50
to 70%, in samples Al-2 to Al-4, respectively.
macropore volume up to 2.41 cm /g. This is reflected by the values
of the bulk density and sample porosity, the last increasing from 71
to 94%. Although, no systematic change was observed for the BET
surface area, a considerably higher value was determined for the
Al-3 as compared to the reference.
Fig. 3a shows that as-synthesized samples present a not very
well defined X-ray diffraction pattern. XRD results shown in Fig. 3b
reveal the formation of a typical ␥-Al2O3 structure after calcination
[10], which is independent of the content of DHN and surfactant
templates (Table 1). Additionally, Fig. 3a shows peaks related to
the presence NH4NO3, resulting from the use of NH4OH and HNO3
in the synthesis, the latter is easily washed away. SEM observation
of alumina samples before and after the calcination revealed that
this phase transformation is topotactic, thus both the particles and
the pores morphology remain unchanged.
Fig. 2. Scanning electron micrographies of Al-1 to Al-4 aluminas samples.

6
2
L. Martins et al. / Applied Catalysis A: General 398 (2011) 59–65
Table 1
Pore volume, bulk density, porosity and BET surface area of the alumina samples prepared with different DHN content.
Vol.DHN/Vol.Sola
DHN (wt.%)
Vmeso (cm /g)
3
Vmacro (cm /g)
3
Bulk density (g/cm3)
Porosity (%)
BETArea (m /g)
2
Sample
Al-ref
Al-1
Al-2
Al-3
Al-4
0
0
1.1
1.7
2.6
0
0
50
60
70
0.16
0.50
0.45
0.52
0.08
0.02
0.15
0.54
1.63
2.41
0.85
0.58
0.37
0.32
0.27
71
82
89
92
94
362
352
293
473
417
a
Volume of added DHN in relation to the volume of the alumina sol.
◦
Fig. 3. XRD patterns of alumina samples prepared (a) without the addition of surfactant and DHN and (b) with different DHN quantity and calcined at 600 C.
◦
Brønsted
H
the formation of ␥-Al O3 at 600 C. This dehydroxylation can be
2
represented as ∼OH + ∼OH → ∼O∼ + ꢀ + H O, where ꢀ represents
2
H
H
H H
H
H
O
H
O
H
O
H
O
H
O
H
O
an oxygen vacancy, an electron-deficient metal atom, upon for-
mation of 4- or 5-coordinated AlOx sites from 6-coordinated ones
-H
600 ºC
O
+++
++
2
O
O
O
[
11].
The spectral analysis of the XPS data of calcined samples, which
Al Al Al Al Al Al
AlO AlO AlO AlO AlO AlO
+
Lewis
was focused on O 1s core-level spectra, provided additional infor-
mation about the hydroxyl concentration on the surface of the
alumina samples. The deconvolution of the O 1s spectra into two
components depicted in Fig. 4 and Table 2, allowed to obtain infor-
mation regarding the relative amount of oxygen bonded in the
form of aluminum oxide (530.5 eV) or superficial hydroxyl groups
(531.7 eV). The XPS signal related to the hydroxyl groups decreases
continuously with the addition of DHN. This result highlights a gen-
eral tendency, also observed in NH3-TPD measurements, that the
addition of DHN interferes in the quantity of hydroxyl groups avail-
able on the surface (Fig. 5). Probably the interaction of the surfactant
with aluminum species occurs via oxygen atoms of the surfactant
with the hydrogen atoms of the oxyhydroxides, inducing the for-
mation of hydroxyls at the interface. As already verified, as the
Scheme 1. Representation of the nature of acid sites on alumina surface (acid
strength +++ > ++ > +).
As shown in Scheme 1, not only the structure of aluminas
but also the surface of calcined samples can change completely
upon firing. The transition alumina ␥-Al O3 is formed from
oxyhydroxides and hydroxides under various dehydration and
dehydroxylation conditions, in which these two descriptions rep-
resent the loss of water by desorption of physisorbed water or by
condensation of hydroxyl groups, respectively. Scheme 1 repre-
sents the dehydroxylation condition, in which two adjacent surface
hydroxyls condense to form water molecules on the surface during
2
Fig. 4. O 1s spectra of the Al-ref and Al-4 alumina samples.

L. Martins et al. / Applied Catalysis A: General 398 (2011) 59–65
63
Table 2
XPS signals distribution and amount of acid sites obtained from TPD-NH3 profiles from calcined samples.
Sample
XPS signals (%)
Hydroxyls
TPD acid sites distribution
Total acid sites normalized by
Mass (mmol NH3/g)
Area (␮mol NH3/m2)
Al2O3
Weak (%)
Moderate (%)
Strong (%)
Al-ref
Al-1
Al-2
Al-3
Al-4
73.9
64.6
61.3
59.5
57.5
26.1
35.4
38.7
40.5
42.5
30.0
75.1
78.6
100
34.7
24.9
21.4
0
35.3
7.9
7.2
6.7
8.4
8.3
21.8
20.4
22.8
17.8
20.0
0
0
0
0
100
0
quantity of DHN increases up to 70 wt.%, the surfactants present
in the sol migrate to the DHN/water interface diminishing their
amount in the continuous aqueous phase and interfering strongly
on the presence of surface hydroxyls.
acid sites decrease with the addition of DHN. The differences in the
concentration of these acid sites will strongly affect the catalytic
performance of these aluminas, as we will discuss further on.
Considering that ␥-Al O3 has special importance in catalytic
applications, the samples were evaluated on their acid properties
2
3.2. Catalytic activity on ethanol dehydration reaction
using temperature programmed desorption of NH3 (NH -TPD). An
3
Ethanol dehydration is commonly used to characterize the acid
strength of solid catalysts [13]. Ethanol conversion can be acid or
base catalyzed, therefore the selectivity to a desired product can be
easily related to the character of the surface. Therefore this reac-
tion was selected to examine catalytic performance of the alumina
samples. A simplified ethanol dehydration reaction is depicted in
Scheme 2, showing that there are two pathways for product forma-
ammonia molecule can be retained on the surface of alumina in two
different modes: (1) transfer of a proton from surface hydroxyl to
the adsorbate; this occurs with the surface acting as a Brønsted acid,
corresponding to the strongest mode of interaction and (2) coordi-
nation to an electron-deficient aluminum atom; taking place with
the solid acting as a Lewis acid, created by dehydroxylating of the
surface. If NH adsorption takes place with a Brønsted acid site, the
3
tion based on acid (ka) or basic sites (k ) [13]. The porous aluminas
b
interaction can involve neighboring hydroxyl groups with different
acidic strengths, denoted according to Scheme 1 as ++ or +++. This
studied in this work led mainly to the formation of ethene and
diethyl ether with minor amount of acetaldehyde (<0.5% selec-
tivity), thus demonstrating the prevailing acidic character of the
catalysts. At low temperature diethyl ether is produced, while at
high temperature ethene is predominantly produced, as a conse-
quence of intramolecular dehydration [13]. Besides those products
sketched in Scheme 2, no other was detected in any of the experi-
mental conditions used here.
acid strength distribution results in a broad NH desorption profile.
3
Fig. 5 shows that all aluminas have a large amount of weak
acid sites, probably Lewis sites (temperature of desorption between
◦
1
00 and 230 C). However only samples without or with lower
DHN quantity owns strong acid sites able to retain NH3 from 250
◦
to 300 C or even higher, probably of Brønsted nature. The quan-
tity of desorbed NH3 at high temperatures is quite low, especially
in samples Al-3 and Al-4, indicating that the strong acid sites
present in these alumina surfaces are irrelevant. Evidently, this is
a consequence of an inductive effect of DHN on oxyhydroxides at
alumina surface during the synthesis. The broad desorption peaks
also reveals that the alumina samples possess a heterogeneous dis-
tribution of both type of acid sites. This distribution was already
observed by pyridine adsorption followed by Fourier transform
infrared spectroscopy [12].
◦
Fig. 6 shows the ethanol dehydration results obtained at 300 C
using the porous alumina catalysts. The samples are ordered
according to the increasing porosity and decreasing acidity. The
result is quite surprising, the conversion passes through a mini-
mum (Fig. 6a and b). For the sample Al-ref, which presented the
highest concentration of hydroxyl groups on the surface, the most
efficient conversion was detected. This result is related to the higher
ability of Brønsted acid sites to activate ethanol molecule. Further-
more Fig. 6 shows that for the Al-4, with the highest porosity, the
ethanol conversion is comparable to Al-ref. In spite of Al-4 presents
the lowest acidity, this behavior can be related to easier reactants
and products diffusion in the catalyst pores. Although ethanol is a
small molecule compared to pore dimensions of the studied alu-
mina, its conversion seems to be to some extend a pore diffusion
controlled reaction, that is, the activity appears to be related to
catalyst porosity. Therefore, this behavior highlights both the con-
tribution of catalyst acidity and porosity on the ethanol conversion.
The region of minimum observed for sample Al-2 represents the
worst combination of catalyst acidity and porosity.
Table 2 reports the relative amount of acid sites of type 1, 2 or 3
(
weak, moderate or strong) able to adsorb NH . The total amount
3
of acid sites determined either per mass of catalyst or per surface
area is practically equal for all alumina samples, while the strong
Indeed, ethanol dehydration proceeds initially via the adsorp-
tion of the ethanol molecule and the O–H bond rupture on
acid sites giving rise to a surface ethoxy group. More strongly
CH CH
intramolecular dehydration)
+
H O
2
2
2
(
k
a
CH CH OH
½ CH CH OCH CH
+
½ H O
2
3
2
3
2
2
3
(
intermolecular dehydration)
k
b
+
CH CHO H2
dehydrogenation)
3
(
Fig. 5. NH3-TPD curves (black) and deconvolution profiles (grey) of the aluminas
Scheme 2. Reaction scheme of ethanol conversion on acid (ka) and basic (kb) cata-
prepared with different amounts of DHN.
lysts.

6
4
L. Martins et al. / Applied Catalysis A: General 398 (2011) 59–65
◦
Fig. 6. Catalytic results of ethanol dehydration of alumina samples at 300 C: (a) ethanol conversion, (b) ethanol molar consumption normalized by aluminas surface area,
(
c) product selectivity and (d) products molar formation normalized by aluminas surface area.
adsorbed ethanol favors O–H bond rupture facilitating the nucle-
ophilic attack and consequently giving rise to higher conversion
The change in the relative abundance of Lewis and Brønsted acid
sites could be one of the reasons for the selectivity of ethene or
diethyl ether (Fig. 7b), because intramolecular and intermolecular
dehydration may go through different catalytic sorption processes.
However, from our results, the participation of surface OH groups
or electron-deficient aluminum atoms in the synthesis of diethyl
ether or ethene cannot be ruled out because products formations
do not seem to depend on the relative amount of both acid sites.
The temperature evolution of the dehydration reaction for the
mesoporous sample Al-1 and for the macroporous one Al-4 are
compared in Fig. 7a while 7b shows the product selectivity as a
function of conversion. The results show that sample Al-4 was
the most active in the whole range of the evaluated tempera-
ture, evidencing that the introduction of macropores in the catalyst
structure enhanced the ethanol conversion, probably due to trans-
port effects.
[
13]. For this reason the most acidic sample Al-ref was the most
active. However, in the case of Al-4 sample that showed a rela-
tive weak acidity, facilitated diffusion through its porous structure
might have mainly contributed to the observed ethanol conversion
[
14,15].
The formation of diethyl ether (Fig. 6c and d) is a reaction that
involves two adjacent alcohol molecules adsorbed on two neigh-
boring sites [13,16]. This is the reason why diethyl ether selectivity
is favorable at low conversion levels, thus decreasing with increas-
ing conversion. On the other hand, ethene formation occurs via
intramolecular dehydration and is predominant at high conver-
sions as shown in Fig. 7. At low temperatures (Fig. 7a), not only
the catalyst activity is poor, but also the selectivity of ethylene is
low due to a large number of ethanol converted to diethyl ether.
1
00
a
macroporous alumina
b
1
00
diethyl ether
8
6
4
2
0
0
0
0
0
Al-1
Al-4
8
6
4
2
0
0
0
0 ethene
0
mesoporous alumina
300 350 400
Temperature (ºC)
2
00
250
0
20
40
60
80
100
Conversion (%)
Fig. 7. Catalytic results of ethanol dehydration on mesoporous Al-1 (closed symbols) and macroporous Al-4 (opened symbols) aluminas samples at different temperatures:
a) ethanol conversion and (b) product selectivity as a function of the conversion.
(

L. Martins et al. / Applied Catalysis A: General 398 (2011) 59–65
65
4
. Conclusions
References
[
[
1] K. Tanabe, W.F. Holderich, Appl. Catal. A: Gen. 181 (1999) 399–434.
2] G. Ertl, H. Knözinger, J. Weitkamp, Preparation of Solid Catalysts, Wiley-VCH,
1
Alumina catalysts with varying porosity were prepared by the
combination of the sol–gel method and templating with micelle
and emulsion. For the obtained morphologies the catalytic activity
in ethanol dehydration went through a minimum as a function of
samples porosity. This behavior can be explained by the combina-
tion of surface acidity and porosity. The results clearly show that
macroporous alumina with well designed pore structure allows for
a rapid diffusion and consequently improve the reaction kinetics.
However, especial attention must be given to the synthesis pro-
cedure, because the addition of decahydronaftalene as “oil phase”
negatively interfered on the concentration of superficial hydroxyl
and consequently in the alumina acidity. This work highlights that
both porosity and acidity must be considered in the evaluation of
the overall activity of porous catalysts prepared by similar proce-
dures.
999.
[3] R. Takahashi, S. Sato, T. Sodesawa, K. Arai, M. Yabuki, J. Catal. 229 (2005)
4–29.
4] L. Martins, M.A.A. Rosa, S.H. Pulcinelli, C.V. Santilli, Micropor. Mesopor. Mater.
32 (2010) 268–275.
[5] M.A.A. Rosa, E.P. Santos, C.V. Santilli, S.H. Pulcinelli, J. Non-Cryst. Solids 354
2008) 4786.
2
[
1
(
[
[
6] T. Numaguchi, Catal. Surv. Jpn. 5 (2001) 59–63.
7] C. Marquez-Alvarez, N. Zilkova, J. Perez-Pariente, J. Cejka, Catal. Rev.-Sci. Eng.
50 (2008) 222.
[8] D.P. Fabiano, B. Hamad, D. Cardoso, N. Essayem, J. Catal. 19 (2010) 190–196.
9] E.W. Washburn, Proc. Natl. Acad. Sci. U.S.A. 7 (1921) 115.
10] R. Rinaldi, F.Y. Fujiwara, U. Schuchardt, Appl. Catal. A: Gen. 315 (2006)
4–51.
[11] J.J. Fitzgerald, G. Piedra, S.F. Dec, M. Seger, G.E. Maciel, J. Am. Chem. Soc. 119
1997) 7832–7842.
[
[
4
(
[
[
12] S. Carre, B. Tapin, Appl. Catal. A: Gen. 372 (2010) 26–33.
13] J.I. Di Cosimo, V.K. Diez, M. Xu, E. Iglesia, C.R. Apesteguia, J. Catal. 178 (1998)
499–510.
[
[
[
14] R. Takahashi, S. Sato, S. Tomiyama, T. Ohashi, N. Nakamura, Micropor. Mesopor.
Mater. 98 (2007) 107–114.
15] S.-W. Bian, Y.-L. Zhang, H.-L. Li, Y. Yu, Y.-L. Song, W.-G. Song, Micropor. Mesopor.
Mater. 131 (2010) 289–293.
Acknowledgments
We are deeply grateful to CNPq and FAPESP for the financial
support offered.
16] J.R. Jain, C.N. Pillai, J. Catal. 9 (1967) 322–330.