Aluminium chloride: A new aluminium source to prepare SAPO-34 catalysts with enhanced stability in the MTO process

Article abstract of DOI:10.1016/j.apcata.2013.12.016

The MTO process produces lower olefins from natural gas or coal via methanol, providing an interesting route to obtain valuable petrochemicals from carbon sources alternative to petroleum. Small-pore silicoaluminophosphate SAPO-34 has been proven an efficient catalyst for the MTO process, showing exceptionally high selectivity to lower olefins. However, these catalysts undergo rapid deactivation due to deposition of high molecular weight hydrocarbons on the pore entrances, which completely blocks the internal channels of the SAPO-34 crystals. In the present work, we explore the use of very dilute zeotype precursor solutions to reduce the size of the crystals of these materials, improving greatly their catalytic behaviour in the MTO process by increasing the catalyst life time.

Full text of DOI:10.1016/j.apcata.2013.12.016

Applied Catalysis A: General 472 (2014) 72–79
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Aluminium chloride: A new aluminium source to prepare SAPO-34
catalysts with enhanced stability in the MTO process
∗
Teresa Álvaro-Mu n˜ oz , Carlos Márquez-Álvarez, Enrique Sastre
Instituto de Catálisis y Petroleoquímica, ICP-CSIC, C/Marie Curie, 2, 28049 Madrid, Spain
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 26 July 2013
Received in revised form
The MTO process produces lower olefins from natural gas or coal via methanol, providing an interest-
ing route to obtain valuable petrochemicals from carbon sources alternative to petroleum. Small-pore
silicoaluminophosphate SAPO-34 has been proven an efficient catalyst for the MTO process, showing
exceptionally high selectivity to lower olefins. However, these catalysts undergo rapid deactivation due
to deposition of high molecular weight hydrocarbons on the pore entrances, which completely blocks the
internal channels of the SAPO-34 crystals. In the present work, we explore the use of very dilute zeotype
precursor solutions to reduce the size of the crystals of these materials, improving greatly their catalytic
behaviour in the MTO process by increasing the catalyst life time.
2
8 November 2013
Accepted 11 December 2013
Available online 17 December 2013
Keywords:
MTO
SAPO-34 nanocrystals
Aluminium chloride
Dilute precursor solutions
©
2013 Elsevier B.V. All rights reserved.
1
. Introduction
is an attractive synthesis route because methanol can be obtained
from a wide variety of materials.
The global energy crisis has renewed the interest in developing
Methanol can be efficiently produced from syngas [1–3]
obtained by natural gas reforming or carbon gasification, and it
might even provide an environmentally carbon neutral alternative
to fossil carbon sources [4], if produced by chemical recycling of
carbon dioxide via hydrogenation [5] or from syngas obtained by
biomass gasification [6].
technologies to produce energy, fuels and chemicals from resources
other than traditional fossil fuels (oil, coal, gas). A large number of
base chemicals are carbon-containing compounds (in many cases
obtained from olefins), so alternative sources for their production
are very limited. It is not necessary to emphasize the importance of
petroleum in global economy at present. It is mainly used as source
of fuels in energy production plants, as an energy source in domestic
and industrial uses, in transportation, but also, although in a much
lesser extent, as raw material in chemicals synthesis. The petroleum
transformation represents a large percentage of the total chemi-
cal processes, despite the huge problem associated to the limited
known global reserves, which are not expected to increase in the
short-term. Therefore, due to the high oil prices and limitation of
resources, it is necessary to develop new processes for the pro-
duction of compounds that are traditionally derived from certain
petroleum fractions.
Small-pore silicoaluminophosphate SAPO-34 (chabazite frame-
work type) has been proven an efficient catalyst for the MTO
process. This catalyst is exceptionally selective to lower olefins
(selectivities over 80% to C2 –C4 olefins have been obtained)
[7,8] because it has small pore entrances which allow only the
diffusion of linear hydrocarbons thus inhibiting production of aro-
matic hydrocarbons. However, the main drawback of this catalyst
is its rapid deactivation. Deactivation starts when aromatics and
heavy branched compounds are formed inside the large cages.
These molecules cannot diffuse though the porous structure of
the catalyst because their kinetic diameter is larger than the pore-
opening size. Therefore, they remain inside the cages where they
can form carbonaceous deposits that block the pore openings and
prevent the access of molecules to the active sites [9,10]. Coke
compounds may adsorb on the acid sites, poisoning the catalyst
active centres. Olsbye and col. studied the deactivation process of
several zeolitic materials in the methanol-to-hydrocarbons (MTH)
and MTO processes [11–14]. They concluded that the formation of
coke during the MTO reaction on ZSM-5 crystals significantly differs
from that on SAPO-34 crystals. Pore blocking in ZSM-5 zeolite has
Light olefins such as ethylene, propylene and butylenes are
important intermediates for the petrochemical industry. They can
be produced via several processes using mainly fossil, and there-
fore non-renewable, raw materials. In this sense, the production of
light olefins from methanol (Methanol-to-Olefins or MTO process)
∗ Corresponding author. Tel.: +34 915854796.
E-mail address: teresa.alvaro@icp.csic.es (T. Álvaro-Mu n˜ oz).
0
926-860X/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcata.2013.12.016

T. Álvaro-Mu ˜n oz et al. / Applied Catalysis A: General 472 (2014) 72–79
73
Table 1
Synthesis conditions used to prepare SAPO materials from gels with molar composition 1Al2O3:1P2O5:0.6SiO2:xTEAOH:yH2O. Crystallization was maintained at 423 K for 5
days.
Catalyst
x
y
Silicon sourceb
pH gel
pH cryst
Phase
Al(OH)3-La
AlCl3-L-3
AlCl3-L-6
AlCl3-T-6
1
3
6
6
40
65
110
110
L
L
L
T
6.8
1.4
8.5
8.5
7.3
0.2
9.3
8.4
Chabazite
Cristobalite
Chabazite
Chabazite
a
Reference sample synthesized with aluminium hydroxide [21].
L: Ludox; T: TEOS.
b
been related to the formation of polyaromatic compounds at the
channel intersections or on the outer surface, while in SAPO-34
it arises from polyaromatic molecules formed in the large cavi-
ties. Both processes lead to catalyst deactivation [15,16] but the
differences in pore architecture of these materials affect the coke
formation and therefore the activity of the catalyst. In this respect,
ZSM-5, although less sensitive to deactivation as compared to
SAPO-34, has a low selectivity towards light olefins.
[34]. This article mainly deals with zeolite synthesis, but also alu-
minophosphate systems, namely, AlPO-5 (AFI), AlPO-11 (AEL), and
AlPO-18 (AEI), are discussed. In the present work, we explore the
use of very dilute zeotype precursor solutions to obtain SAPO-34
with decreased crystal size, which is shown to improve its catalytic
performance.
2. Experimental
Numerous studies indicate that the morphology of SAPO-34
crystals can affect their properties and applications and various
methods have been suggested to improve catalyst lifetime in the
MTO reaction [9,17–21]. Catalyst deactivation may be strongly cor-
related with particle size due to the diffusion limitations of the
guest molecules in the micropores [22]. For large catalyst crystals,
residence time for hydrocarbons is high because of the long diffu-
sion path. Aromatic compounds cannot escape from the pores of
SAPO-34 and successive polymerizations readily occur because of
the long reaction time. Chen et al. [9] examined the effect of crystal
size of SAPO-34 on its selectivity and deactivation in the MTO reac-
tion. In large crystals, the reactions were controlled by limitations
in the diffusion of both methanol and the intermediate product
dimethyl ether (DME). Furthermore, the coking rate increased with
increasing crystal size. It was found that SAPO-34 catalysts with
crystal size of less than 500 nm showed remarkable lifetime in the
MTO reaction. Lee et al. used various templates to obtain SAPO-
2.1. Synthesis of SAPO-34 molecular sieves
Hydrothermal synthesis of SAPO-34 samples was carried
out using tetraethylammonium hydroxide as template and alu-
minium chloride as aluminium source. The gel composition was:
1Al O :1P O :0.6SiO :xTEAOH:yH O. The molar composition of
2
3
2
5
2
2
the reaction mixtures and the synthesis conditions for the different
SAPO-34 materials obtained are given in Table 1. For comparative
purposes, a reference material was synthesized by the conventional
procedure, using Al(OH)3 as aluminium source, according to the
method published previously [21]. Experimental conditions (tem-
perature and crystallization times) have been adjusted in order to
obtain pure phases of SAPO-34. The same silicon to aluminium ratio
(0.3) was used for all the synthesis gels prepared.
In a typical synthesis to get pure SAPO-34, the aluminium source
(aluminium chloride, 99% or aluminium hydroxide, reagent grade,
both from Sigma Aldrich) was added slowly to a dilute phos-
phoric acid solution (prepared using orthophosphoric acid, Riedel
de Haën, 85%), and the mixture was vigorously stirred for 2 h to
obtain a uniform gel. Silica solution (Sigma–Aldrich, Ludox SM-
30, 30 wt% suspension in water) or tetraethylorthosilicate (TEOS,
Merck, 98%) was added dropwise to this mixture, followed by
addition of the template, tetraethylammonium hydroxide (TEAOH,
Sigma–Aldrich, 35 wt% in water). Finally, the mixture was stirred
for about 4 h. The gel was then transferred into Teflon-lined stain-
3
4 catalysts with similar physico-chemical properties but different
crystal size and discussed the effect of crystal size on the catalytic
performance, especially on deactivation behaviour [20]. They con-
cluded that deactivation is related to diffusion limitation and that
catalysts with smaller crystal size show slower deactivation due to
the higher proportion of accessible cages near the external surface.
Yao et al. [23] reported a polymer-assisted dry gel synthesis method
aiming to control crystal growth. Using this approach, SAPO-34
nanocrystals were successfully formed from a precursor gel con-
taining polyacrylamide, but SAPO-34 crystals larger than several
micrometres were also formed simultaneously.
3
less steel autoclaves with a capacity of 40 cm , which were heated
statically at 423 K under autogeneous pressure for 5 days.
The resulting solids were collected by centrifugation, washed
with water and ethanol and dried at room temperature overnight.
The organic template and the water trapped within the micropores
of the as-synthesized solids were removed by calcination at 823 K
prior to catalyst testing. Complete removal of the organic molecule
was assessed by thermogravimetric analysis.
Other authors have tried to improve the diffusivity of the prod-
ucts creating hierarchical pores into the SAPO-34 crystallites by
different synthesis methods [24,25] or using carbon materials
(
nanoparticles, nanotubes) as templates [26]. In all these cases
an important improvement in the lifetime of the catalysts was
observed.
The above mentioned reports suggest that control of the size of
SAPO-34 crystals is a very important factor for improving the cat-
alytic activity and lifetime of the catalyst. However, small SAPO-34
crystals with a uniform particle size have been obtained mainly
through fractionation of the products. In this sense, nowadays
several attempts to synthesize SAPO-34 and other zeolitic and
mesoporous materials with small crystal size have been carried
out using different approaches: synthesis from colloidal solutions
2
.2. Characterization
Powder X-ray diffraction (XRD) patterns of as-synthesized and
calcined samples were recorded on a Philips X’PERT diffractome-
ter using CuK␣ radiation with a nickel filter. The textural data (pore
volume and BET surface area) were determined by nitrogen adsorp-
tion measurement using a Micrometrics ASAP 2010 volumetric
apparatus. Previous to measure the nitrogen adsorption/desorption
isotherms samples were degassed at 623 K under vacuum for at
least 20 h. The crystal size and morphology were examined by scan-
ning electron microscopy (SEM) using a JEOL JSM 6400 or a Philips
XL30 microscope, operating at 20 kV. Elemental mapping by EDX
[
27,28] or applying microwave heating [28–33].
One promising route for preparation of microporous and partic-
ularly zeolite nanocrystals is the use of clear solutions or colloidal
suspensions as precursor media. A review summarizing the dif-
ferent systems studied was published from Tosheva and Valtchev

7
4
T. Álvaro-Mu ˜n oz et al. / Applied Catalysis A: General 472 (2014) 72–79
has carried out using a ultrahigh resolution FEI-NOVA NanoSEM
1,2
2
30 FESEM instrument.
AlCl -T-6
3
The organic content of the samples was determined by ele-
mental analysis with a Perkin-Elmer 2400 CHN analyser and
by thermogravimetric analysis (TGA) using a Perkin-Elmer TGA7
instrument. TG analyses were carried out at a heating rate of
1,0
0,8
AlCl -L-6
3
2
0 K/min under air flow. Chemical analysis of Al, P and Si in calcined
samples was performed by inductively coupled plasma optical
emission spectrometry (ICP-OES, Perkin-Elmer 3300DV instru-
ment) after sample dissolution by alkaline fusion.
0
0
0
,6
,4
,2
AlCl -L-3
3
29
Si CP/MAS NMR spectra were recorded at room temperature
using a Bruker AV-400-WB spectrometer operating at 79.5 MHz,
with a 4 mm probe spinning at 10 kHz. A ␲/2 pulse of 3 ␮s, contact
time of 6 ms and recycle delay of 5 s were used. The chemical shifts
were referenced to tetramethylsilane (TMS), taken as 0 ppm.
Al(OH) -L
3
0,0
Ammonia temperature programmed desorption (NH -TPD) was
3
performed using a Quantachrome ChemBET-3000 TPR/TPD equip-
ment. Typically, 100 mg of sample pellets (20–40 mesh) were
pre-treated at 823 K for 1 h in helium flow (25 mL/min) and sub-
sequently cooled to the adsorption temperature (400 K). A gas
mixture of 5.0 vol% NH3 in He was then allowed to flow over the
sample for 4 h at a rate of 15 mL/min. Afterwards, a 25 mL/min
helium flow was passed over the sample, while maintaining the
temperature at 400 K for 30 min to remove weakly adsorbed NH3,
and, finally, the temperature was increased to 823 K at a rate of
10
20
30
40
50
2
θ (
º)
Fig. 1. XRD patterns of as-synthesized samples.
crystallize and only a dense phase was obtained, as indicated by the
XRD pattern of sample AlCl -L-3 (Fig. 1). Therefore, the ratio was
increased to TEAOH/Al O = 6 in order to reach a pH value suitable
for the crystallization of the desired material. With this new com-
position two gels were prepared using two different silicon sources,
Ludox and TEOS, to study its effect on the physico-chemical proper-
ties of the resulting material. The X-ray powder diffraction patterns
3
2
3
1
0 K/min.
2
.3. Catalyst testing
(
Fig. 1) confirmed the structure type SAPO-34 (CHA framework
code) for both as-synthesized samples, AlCl -L-6 and AlCl -T-6,
according to patterns reported for this structure [35]. Neverthe-
less, slight differences respect to the pattern of the reference sample
Methanol conversion to olefins was tested at 623, 673 and 723 K
3
3
using a laboratory plant with a continuous down flow packed bed
reactor, fully automated and controlled from a personal computer
(
PiD Eng&Tech Microactivity Reference), operating at atmospheric
Al(OH) -L can be observed (Fig. 1), which can be attributed to differ-
3
pressure. Catalyst weight (1.0 g; 20–30 mesh pellets size) and
methanol flow rate (0.025 ml/min) were adjusted in order to obtain
ent crystallite shape and size for both types of samples. Practically
no loss of crystallinity was observed when the as-synthesized sam-
ples were heated at 823 K in order to remove the organic molecules
(Supplementary Information, Fig. I), confirming their thermal sta-
bility under calcination conditions. The XRD patterns of the two
samples prepared with aluminium chloride using different silicon
sources exhibit comparable peak width. This fact suggests that both
samples present similar crystal size.
Some selected scanning electron microscopy (SEM) pho-
tographs of the SAPO-34 materials synthesized are presented in
Fig. 2. It is observed that samples prepared with aluminium chlo-
ride have very similar morphology, although the sample prepared
−1
a weight hourly space velocity (WHSV) of 1.2 h . Previous to the
reaction, samples were pre-treated under nitrogen flow at 723 K
for 1 h. During the reaction, nitrogen was used as an inert dilu-
ent gas and co-fed with methanol into the reactor with a constant
methanol/nitrogen ratio of 1/1 mol. Methanol was fed as liquid
using a Gilson 307 HPLC pump, vaporized and mixed with the nitro-
gen stream in a pre-heater at 473 K. The reaction products were
analyzed on-line by gas chromatography using a Varian CP3800 gas
chromatograph equipped with flame ionization (FID) and thermal
conductivity (TCD) detectors, with a Petrocol DH50.2 capillary col-
umn and a Porapack Q 80–100 mesh packed column for separation
of hydrocarbons and oxygenates, respectively.
with TEOS (AlCl -T-6) possesses slightly smaller crystal size. In the
3
micrographs, the typical cubic-like rhombohedra morphology can
be clearly observed, with crystals approximately 200–300 nm in
size. However, the sample prepared with aluminium hydroxide
3
. Results and discussion
(
Al(OH) -L) presents a quite different crystal shape and size. As
3
it can be observed in the micrographs, sample Al(OH) -L presents
3
3
.1. Crystal structure and morphological analysis
plate-like crystals with size around 500 nm × 300 nm and several
tenths of nm thickness. Therefore, it was found that using highly
The synthesis conditions were varied in order to obtain SAPO-
4 crystals using a dilute solution. The molar composition of the
diluted precursor solution (samples AlCl -L-6 and AlCl -T-6, pre-
3
3
3
pared with aluminium chloride) the crystal size of this material can
be decreased.
reaction mixtures and the synthesis conditions for the different
materials obtained are given in Table 1. A sample prepared with
aluminium hydroxide [21] is also included as a reference mate-
rial. Relatively large amounts of template (tetraethylammonium
hydroxide) were used in order to neutralize the hydrochloric acid
released into the medium due to hydrolysis of aluminium chloride,
for SAPO-34 crystallization is favoured around neutral pH. Initially,
the amount of template was increased from the TEAOH/Al O ratio
3.2. Thermogravimetric and elemental analysis
Thermogravimetric analyses (TGA) were performed aiming to
verify the incorporation of the SDA molecules in the structure
of the as-made samples and their subsequent complete elimina-
tion after calcination prior to the use of SAPO-34 materials in the
catalytic reactions. The TGA profiles of the samples studied are
plotted in Fig. 3. These results show three different weight loss
2
3
of 1 used in the conventional synthesis method to a ratio of 3.
Although the amount of TEAOH used was high, the pH of the syn-
thesis gel (1.4) was too acidic to allow the SAPO-34 material to

T. Álvaro-Mu ˜n oz et al. / Applied Catalysis A: General 472 (2014) 72–79
75
Fig. 2. SEM images of calcined samples.
Table 2
Elemental (CNH) and thermogravimetric analyses of the as-synthesized samples.
Sample
Weight loss (%)a
Organic contentb
wt (%)
I (T < 473 K)
II (473 < T < 823 K)
III (>823 K)
C/N Exper.
Template molec. per cage
Al(OH)3-L
AlCl3-L-6
AlCl3-T-6
3.1
6.4
6.7
14.2
11.5
14.4
1.8
2.8
1.5
13.9
14.3
14.6
7.34
7.26
7.31
1.00
1.07
1.06
a
From TGA analyses.
From chemical analyses.
b
steps. The first weight loss (I), at temperatures below 473 K, can be
attributed to removal of adsorbed water. The second weight loss
at least, three different steps in the temperature range from ca.
623 K to 773 K. These results suggest that there is a different inter-
action between the organic cations and the SAPO framework in
both types of samples, and that part of the organic occluded in
(
II), between 473 K and 823 K, is due to the decomposition of the
template. Finally, the third weight loss (III), at temperatures higher
than 823 K, is associated with the combustion of organic residues
occluded in the channels and cages of the SAPO-34. Quantitative
data of the different weight losses are presented in Table 2.
The derivative plot presented in Fig. 3 shows that the kinet-
the pores of sample Al(OH) -L is more weakly bonded to the inor-
3
ganic network. On the other hand, it is observed that the weight
loss at temperatures below 473 K is higher for samples prepared
with aluminium chloride. Therefore, these samples show a more
ics of template removal in sample Al(OH) -L is markedly different
hydrophilic behaviour than sample Al(OH) -L.
3
3
from that of samples obtained with AlCl . In samples prepared with
aluminium chloride, the profile exhibits a single peak centred at
The total weight losses of steps II and III are in good agreement
with the organic content determined by CHN elemental analysis
(Table 2), confirming the previous assignment of weight losses II
and III to removal of template. Moreover, the C/N ratios determined
by elemental analysis are very close to 8, the value correspond-
ing to TEAOH, which indicates that the organic SDA molecules
occluded or adsorbed in the different samples were not decom-
posed during the hydrothermal treatment. Based on the topological
structure of SAPO-34 molecular sieve, the average number of tem-
plate molecules per cage in the different samples was calculated
3
7
23 K that can be associated to decomposition of tetraethylammon-
ium species. In contrast, the profile of sample Al(OH) -L reveals,
3
100
0
9
9
8
8
7
5
0
5
0
5
-
2
(Table 2). In all the cases the organic content corresponds to one
molecule of SDA (TEAOH) incorporated per unit cell.
-4
The chemical composition of samples obtained by ICP-OES is
presented in Table 3. In samples prepared with aluminium chloride,
the Si/Al+P ratio is very close to that of the synthesis gels, and this
-
-
-
6
value is slightly higher for sample Al(OH) -L. Although the amount
3
8
of silicon incorporated in the network is similar for AlCl3-L-6 and
AlCl -L-6
3
AlCl -T-6, the different proportion of phosphorous and aluminium
3
AlCl -T-6
3
in the solids suggests that the silicon distribution in the frame-
work is rather different in both samples. This will be discussed later
on the basis of 29Si-NMR analysis. EDX elemental mapping of the
three samples was also collected (Figs. III–V and Table I of Supple-
mentary Information). Oxygen, aluminium, silicon and phosphorus
were detected. The distribution of the elements was very
10
Al(OH) -L
3
2
00
400
600
800
Temperature (ºC)
Fig. 3. Thermogravimetric analyses of SAPO-34 samples.

7
6
T. Álvaro-Mu ˜n oz et al. / Applied Catalysis A: General 472 (2014) 72–79
Table 3
Elemental composition.
Sample
Molar composition
Si/(Al+P)gel
Si/(Al+P)solid
Si incorporationa
NH3 TPD (mmol/g)
Al(OH)3-L
AlCl3-L-6
AlCl3-T-6
Si0.17Al0.43P0.40O2
Si0.13Al0.48P0.39O2
Si0.13Al0.40P0.47O2
0.15
0.15
0.15
0.20
0.15
0.15
1.31
1.00
1.00
0.43
0.64
0.67
a
The level of silicon incorporation is defined as the molar ratio: [Si/(Si+Al+P)]solid/[Si/(Si+Al+P)]gel.
homogenous on the entire particles of the samples with lower Si
content. The contents of Si, Al and P determined by both techniques,
EDX and ICP-OES, are very similar in the case of samples AlCl3-L-6
creates a negative charge per every Si atom in the framework, which
is usually balanced by the positive charge of the organic molecules
occluded within the microporous structure. However, in the case
and AlCl3-T-6. However, for sample Al(OH) -L, the amount of sil-
of sample AlCl -T-6, in addition to this band, centred at −87 ppm,
3
3
icon obtained by EDX, is slightly lower than that determined by
ICP-OES. This result might indicate that the later sample contains
a small amount of amorphous silica, which would contribute to
the total silicon content determined by ICP-OES but would not be
detected in a local analysis of crystals by EDX.
other signals can be observed at −94 ppm and −99 ppm, and also,
with lower intensity, at −105 ppm and −110 ppm, which have been
previously attributed to Si(3Al), Si(2Al), Si(1Al) and Si(0Al), respec-
tively. The presence of these bands can be explained if silicon is
incorporated into the framework via substitution mechanism 3
(SM3), in which the incorporation of Si occurs via a simultaneous
substitution of a pair of adjacent Al and P atoms by two Si atoms.
The higher intensity of the resonances appearing at lower field in
3.3. Textural properties
sample AlCl -T-6 compared to sample AlCl -L-6 indicates a higher
Calcined materials
were
analyzed
by
nitrogen
3
3
contribution of substitution mechanism SM3 in the former sample
which should result in a lower Al/P ratio, in agreement with the
adsorption–desorption at 77 K in order to determine their textural
properties. All the samples possess an important microporous
contribution, typical of these kind of materials, and a smaller, but
significant, external one (Fig. II in Supplementary Information).
Pore volume and surface area values calculated from the isotherms
are collected in Table 4. It can be observed that all the samples
chemical analysis results (Table 3). Sample Al(OH) -L presents a
3
spectrum clearly different from the other samples, with a more
intense signal at −110 ppm, indicating aluminosilicate domains
(commonly referred to as Si islands) in the SAPO network [37–39].
2
Depending on the relative contribution of both SM3 and SM2 sub-
stitution mechanisms, the size and concentration of Si islands will
be different [40–42]. Some authors have proposed that the strength
of the acid sites generated at the border of the Si islands is higher
than that of the acid sites created by the isolated Si atoms, and
have surface area in the range 620–650 m /g. However, it should
be noticed that the non microporous component, both in surface
area and volume, is significantly higher for sample AlCl -T-6. These
3
differences can be attributed to a smaller crystal size and higher
intercrystalline porosity of the sample synthesized with TEOS as
silicon source.
that the strength increases as the value of n in the Si(OAl)n(OSi)
4−n
environments decreases [40]. Thus, a higher number of acid sites
are generated through the SM2 mechanism, while substitution via
SM2+SM3 yields less but stronger acid sites. Therefore, according to
.4. 29Si MAS NMR
3
spectra shown in Fig. 4, it can be expected that sample AlCl -T-6 has
3
The incorporation of Si to the SAPO-34 framework has been
2
9
acid sites with higher strength on average than sample AlCl3-L-6,
due its lower proportion of isolated Si atoms. Furthermore, sample
studied by Si MAS NMR (Fig. 4). It has been reported that, when
the amount of silicon incorporated to the SAPO framework is low,
Si atoms are found in a unique Si(4Al) environment resulting from
the substitution of phosphorous by silicon in the aluminophosphate
framework [36]. However, if the silicon content is higher, multiple
silicon environments can occur as a result of the combination of the
single P by Si substitution mechanism with the simultaneous sub-
stitution of a pair of neighbouring Al and P atoms by two Si atoms.
This can be observed in Fig. 4, where various distinct resonances can
be distinguished in the range −87 to −110 ppm, attributed to dif-
29
Al(OH) -L would possess the strongest acid sites as its Si NMR
3
spectrum indicates that it contains the largest proportion of acid
sites at the border of Si islands.
1,4
1,2
1,0
0,8
0,6
0,4
-94: -Si (3Al)-
-99: -Si (2Al)-
-105: -Si (1Al)-
110: -Si (0Al)-
ferent Si(nAl) environments (n = 0–4). Sample AlCl -L-6 presents a
3
-87: -Si (4Al)-
broad band centred at −87 ppm attributed to Si(4Al) environments,
corresponding to Si atoms in phosphorous positions, although this
band has a certain asymmetry, showing a small contribution to
lower field. This suggests the presence of a small fraction of Si in
Si(nAl) environments with n lower than 4. Therefore the silicon
substitution is occurring mainly via mechanism 2 (SM2: a phospho-
rous atom substituted by a silicon atom). This mechanism leads to
Si surrounded by 4 Al atoms in the second coordination shell, and
-
Al(OH) -L
3
AlCl -L-6
AlCl -T-6
3
3
Table 4
0,2
0,0
Textural properties of calcined materials.
2
3
Sample
Surface area (m /g)
Pore volume (cm /g)
Smicro
Sext
Stotal
Vmicro
Vext
Vtotal
-
40
-50
-60
-70
-80
-90
-100
-110
-120
-130
S-Al(OH)3-L
S-AlCl3-L-6
S-AlCl3-T-6
608
588
558
44
31
76
652
619
633
0.26
0.25
0.23
0.29
0.17
0.36
0.55
0.42
0.59
Chemical shift (ppm)
Fig. 4. 29Si CP/MAS NMR spectra of calcined samples.

T. Álvaro-Mu ˜n oz et al. / Applied Catalysis A: General 472 (2014) 72–79
77
Al(OH) -L. This improved performance of the catalysts prepared
with aluminium chloride can be attributed to their smaller crystal
size compared to sample Al(OH)3-L.
The catalysts performance was strongly dependent on the reac-
tion temperature used. At 623 K, all catalysts showed a very short
lifetime reaching high conversion levels during two hours or less
and showing subsequently a fast decay of conversion. At higher
3
Al(OH) -L
3
AlCl -L-6
3
AlCl -T-6
3
temperatures the results were quite different. Samples AlCl -L-6
3
and AlCl -T-6 rendered complete conversion of oxygenates (both
3
methanol and dimethylether) and high selectivity towards short
chain olefins (C2 –C4 ) during 6 h at 673 and 723 K. However, the
catalyst lifetime decreased when the temperature was increased
from 673 to 723 K. At 673 K, the conversion was maintained above
8
0% during 10 h while at 723 K, conversion levels were lower
than 70% after 7 h of reaction. These results can be explained
assuming that deactivation of acid sites due to formation of heavy
hydrocarbon species takes place first for the strongest sites. After
deactivation of the strong acid sites in the initial period of reaction,
the remaining weaker sites would not be active enough to catalyze
the transformation of DME to olefins at 623 K, which would explain
the decrease of conversion observed at this temperature after a
short time of reaction. At higher reaction temperature (673 K), the
less acidic sites would be able to transform DME into light olefins,
thus increasing the catalyst lifetime. A further increase of temper-
ature to 723 K decreases the lifetime due to a faster deactivation
of the catalyst, as the formation of pore-blocking products in the
course of the reaction would be enhanced, making the active sites
of the catalyst less accessible for reactant molecules.
The main problem of this kind of catalysts in the MTO process is
the rapid deactivation attributed to the deposition of high molecu-
lar weight hydrocarbons on the pore entrances [7,43,44]. However,
it is possible to optimize the synthesis of these materials to improve
the catalyst lifetime [45–48]. At all the temperatures tested, sam-
ples prepared with aluminium chloride retained high conversion
during more time than the conventional sample. There are several
factors that could explain the poorer behaviour of the conventional
sample compared with these new materials. First, this sample has
higher Si content in the solid and higher proportion of Si(nAl) envi-
ronments with n equal to 2 or 1. These Si species are associated
to acid centres with higher acid strength [49] and, therefore, are
responsible for the transformation of short chain olefins to higher
molecular weight compounds, which causes the catalyst deactiva-
tion. The second factor affecting the stability of the catalyst is the
larger crystal size of this sample, which has been also correlated
with a lower resistance of the catalysts to deactivation [9,19,21].
All the catalysts gave selectivity to light olefins up to ca. 90%
at conversion levels close to 100% (Fig. 7). The decrease of con-
version was accompanied by a decrease of the selectivity to light
olefins that was more pronounced at 723 K. These results can be
attributed to the deactivation of acid sites, which leads to the for-
mation of important amounts of methane and aromatics at long
time of reaction, especially at 723 K.
4
50
500
550
600
650
700
750
800
Temperature (K)
Fig. 5. NH3-TPD plots of calcined SAPO-34 catalysts.
3
.5. Acidity
The acidity of the catalysts has been evaluated by TPD of ammo-
nia (Fig. 5 and Table 3). The area under the TPD profile indicates
the total amount of ammonia desorbed, which is equivalent to the
number of acid sites, whereas the desorption temperature indi-
cates the acid strength of the sites, the desorption temperature
increasing with the acid site strength. The three SAPO-34 sam-
ples show different TPD profiles. Sample AlCl -L-6 presents a major
desorption peak at low temperature (with maximum at around
25 K), which is an indication of site uniformity and of a relatively
weak acid strength. The TPD profile shows some tailing towards
higher temperatures, indicating the presence of a small propor-
3
5
tion of acid sites with higher strength. However, samples AlCl -T-6
3
and Al(OH)3 exhibit broader desorption profiles, clearly showing
an important contribution of a desorption band with maximum at
around 610 K, corresponding to strong acid sites, overlapping with
the low temperature ammonia desorption peak attributed to weak
acid sites. These results are in agreement with the 29Si NMR results
described previously. The spectrum of sample AlCl -L-6 presents a
3
broad band centred at −87 ppm attributed to Si(4Al) environments,
which give rise to bridging Si-OH-Al hydroxyl groups of weak acid
strength, in accord with the presence of mainly a low tempera-
ture desorption peak at in the ammonia TPD. In contrast, the 29Si
NMR spectra of samples AlCl -T-6 and Al(OH) -L show an impor-
3
3
tant contribution of bands assigned to Si(3Al), Si(2Al) and Si(1Al)
environments. The bridging Si-OH-Al hydroxyl groups associated to
silicon atoms located in those environments are expected to exhibit
stronger acidity, and would be responsible for the presence of the
high temperature band in the TPD of ammonia. On the other hand,
silicon atoms in Si(0Al) environments, that is, surrounded by 4 sil-
icon atoms, do not contribute to the formation of acidic hydroxyl
groups. Accordingly, the higher proportion of these Si centres found
For all the catalysts, the ethylene/propylene ratio increased with
the reaction temperature (Fig. 8). During the period of high conver-
sion of oxygenates, this ratio was around 0.7 at 623 K, 1.0 at 673 K
and 1.6 at 723 K. The raise in ethylene/propylene ratio with temper-
ature has been previously attributed to the secondary reactions of
oligomerization and cracking, which are favoured when the reac-
tion temperature increases [50,51]. Some differences among the
catalysts could be observed in the distribution of the reaction prod-
ucts. The ethylene/propylene ratio was slightly higher for sample
AlCl -T-6 than for sample AlCl -L-6. This result can be attributed
in sample Al(OH) -L might explain its lower overall content of acid
sites determined by ammonia TPD (Table 3).
3
3
.6. Catalytic activity
The catalytic activity of these materials in the MTO reaction was
−
1
studied at 623, 673 and 723 K and a WHSV of 1.2 h as it has been
described previously in Section 2 (Figs. 6 and 7). It can be observed
that samples prepared with aluminium chloride had a very similar
stability in the MTO reaction at the different temperatures studied
and this stability was always higher than that shown by sample
3
3
to the fact that the sample prepared with TEOS possesses stronger
acid sites (as 29Si NMR and TPD of ammonia results indicate), which
favours the cracking of heavy compounds to light olefins, leading
to higher amounts of ethylene than propylene.

7
8
T. Álvaro-Mu ˜n oz et al. / Applied Catalysis A: General 472 (2014) 72–79
100
623 K
673 K
723 K
AlCl -L-6
3
80
60
40
20
0
AlCl -T-6
AlCl -T-6
3
3
AlCl -L-6
3
Al(OH) -L
Al(OH) -L
3
3
Al(OH) -L
3
AlCl -L-6
3
AlCl -T-6
3
0
1
2
3
4
5
6
0
2
4
6
8
10
12
0
1
2
3
4
5
6
7
8
9
Reaction time (h)
Fig. 6. Conversion of oxygenates (MeOH and DME) vs. reaction time. Test conditions: T = 623, 673 and 723 K, WHSV = 1.2 h−1, 1 g of catalyst.
100
80
60
40
20
0
Al(OH) -L
Al(OH) -L
Al(OH) -L
3
3
3
AlCl -L-6
AlCl -L-6
AlCl -L-6
3
3
3
623 K
AlCl -T-6
673 K
AlCl -T-6
723 K
AlCl -T-6
3
3
3
0
1
2
3
4
5
6
0
2
4
6
8
10
12
0
1
2
3
4
5
6
7
8
9
Reaction time (h)
Fig. 7. Selectivity to short chain olefins (C2–C4) in the MTO reaction as a function of time-on-stream. Test conditions: T = 623, 673 and 723 K, WHSV = 1.2 h−1, 1 g of catalyst.
From the results presented here, it can be concluded that the
crystal size of SAPO-34 catalysts would be the main parameter
to control the catalytic behaviour in order to obtain more stable
materials with longer lifetime. Moreover, it is necessary to obtain
materials with the appropriate proportion of Si(nAl) environments,
associated to acid sites of moderate acid strength, to maximize the
transformation of methanol to light olefins while limiting the trans-
formation of short olefins to higher molecular weight compounds.
4
3
3
2
2
1
1
0
0
,0
,5
,0
,5
,0
,5
,0
,5
,0
Al(OH) -L
Al(OH) -L
Al(OH) -L
673 K
3
3
623 K
3
AlCl -L-6
AlCl -L-6
AlCl -L-6
3
3
3
AlCl -T-6
AlCl -T-6
AlCl -T-6
3
3
3
7
23 K
0
1
2
3
4
5
6
0
2
4
6
8
10
12
0
2
4
6
8
Reaction time (h)
Fig. 8. Ethylene/propylene ratio obtained in the MTO reaction at different temperatures as a function of time-on-stream. Test conditions: T = 623, 673 and 723 K,
−
1
WHSV = 1.2 h , 1 g of catalyst.

T. Álvaro-Mu ˜n oz et al. / Applied Catalysis A: General 472 (2014) 72–79
79
The results demonstrate that samples synthesized with aluminium
chloride, which present smaller crystal size, rendered the best cat-
alytic performance.
[15] A.T. Aguayo, A.E.S. del Campo, A.G. Gayubo, A. Tarrío, J. Bilbao, J. Chem. Technol.
Biotechnol. 74 (1999) 315–321.
[
16] D. Chen, H.P. Rebo, K. Moljord, A. Holmen, Ind. Eng. Chem. Res. 38 (1999)
241–4249.
4
[
[
[
17] A.J. Marchi, G.F. Froment, Appl. Catal. 71 (1991) 139–152.
18] M.J. van Niekerk, J.C.Q. Fletcher, C.T. O’Connor, Appl. Catal. 138 (1996) 135–145.
19] N. Nishiyama, M. Kawaguchi, Y. Hirota, D. Van Vu, Y. Egashira, K. Ueyama, Appl.
Catal. A: Gen. 362 (2009) 193–199.
4
. Conclusions
SAPO-34 catalysts have been synthesized using aluminium chlo-
[20] K.Y. Lee, H.-J. Chae, S.-Y. Jeong, G. Seo, Appl. Catal. A: Gen. 369 (2009) 60–66.
[
[
21] T. Álvaro-Mu n˜ oz, C. Márquez-Álvarez, E. Sastre, Catal. Today 179 (2012) 27–34.
22] Y. Hirota, K. Murata, M. Miyamoto, Y. Egashira, N. Nishiyama, Catal. Lett. 140
ride. Using highly diluted precursor solutions the crystal size of
this material can be decreased. This method provided rhombohe-
dral crystals with dimensions of 200–300 nm, while an optimized
sample prepared by the conventional procedure was formed by
plate-like crystals with size around 300 nm × 500 nm and several
tenths of nm thickness. It was shown that this decrease in the
crystal size was the main factor allowing to obtain materials with
improved catalytic performance in the MTO process, which retain
high conversion of oxygenates for longer reaction time.
(
2010) 22–26.
[23] J. Yao, H. Wang, S.P. Ringer, K.-Y. Chan, L. Zhang, N. Xu, Microporous Mesoporous
Mater. 85 (2005) 267–272.
[
[
24] H.Q. Yang, Z.C. Liu, H.X. Gao, Z.K. Xie, J. Mater. Chem. 20 (2010) 3227–3231.
25] S.T. Yang, J.Y. Kim, H.J. Chae, M. Kim, S.Y. Jeong, W.S. Ahn, Mater. Res. Bull. 47
(2012) 3888–3892.
[26] F. Schmidt, S. Paasch, E. Brunner, S. Kaskel, Microporous Mesoporous Mater.
64 (2012) 214–221.
1
[
[
[
27] H. van Heyden, S. Mintova, T. Bein, Chem. Mater. 20 (2008) 2956–2963.
28] C.S. Cundy, J.O. Forrest, Microporous Mesoporous Mater. 72 (2004) 67–80.
29] J.-W. Jun, J.-S. Lee, H.-Y. Seok, J.-S. Chang, J.-S. Hwang, S.-H. Jhung, Bull. Korean
Chem. Soc. 32 (2011) 1957–1964.
Acknowledgements
[
30] C. Gabriel, S. Gabriel, E.H. Grant, E.H. Grant, B.S.J. Halstead, D. Michael, P. Mingos,
Chem. Soc. Rev. 27 (1998) 213–223.
We are thankful for the financial support of the Spanish Ministry
of Economy and Competitiveness, project MAT2012-31127. TAM
acknowledges CSIC for a PhD grant.
[31] S.-E. Park, D.S. Kim, J.-S. Chang, W.Y. Kim, Catal. Today 44 (1998) 301–308.
[
[
32] X. Xu, W. Yang, J. Liu, L. Lin, Adv. Mater. 12 (2000) 195–198.
33] D.P. Serrano, M.A. Uguina, R. Sanz, E. Castillo, A. Rodrı ı´ guez, P. Sánchez, Micro-
porous Mesoporous Mater. 69 (2004) 197–208.
[
[
34] L. Tosheva, V.P. Valtchev, Chem. Mater. 17 (2005) 2494–2513.
35] B.M. Lok, C.A. Messina, R.L. Patton, R.T. Gajek, T.R. Cannan, E.M. Flanigen, US
Patent 4,440,871 (1984).
Appendix A. Supplementary data
[
36] C.S. Blackwell, R.L. Patton, J. Phys. Chem. 88 (1984) 6135–6139.
Supplementary material related to this article can be
found, in the online version, at http://dx.doi.org/10.1016/
j.apcata.2013.12.016.
[37] L. Xu, A. Du, Y. Wei, Y. Wang, Z. Yu, Y. He, X. Zhang, Z. Liu, Microporous Meso-
porous Mater. 115 (2008) 332–337.
[
[
[
[
38] A. Izadbakhsh, F. Farhadi, F. Khorasheh, S. Sahebdelfar, M. Asadi, Z.F. Yan, Micro-
porous Mesoporous Mater. 126 (2009) 1–7.
39] J. Tan, Z. Liu, X. Bao, X. Liu, X. Han, C. He, R. Zhai, Microporous Mesoporous
Mater. 53 (2002) 97–108.
References
40] S. del Val, T. Blasco, E. Sastre, J. Pérez-Pariente, J. Chem. Soc., Chem. Commun.
(
1995) 731–732.
41] J.A. Martens, C. Janssens, P.J. Grobet, H.K. Beyer, P.A. Jacobs, Stud. Surf. Sci. Catal.
9 (1989) 215–225.
[
[
[
1] A. Gotti, R. Prins, J. Catal. 178 (1998) 511–519.
2] I. Wender, Fuel Process. Technol. 48 (1996) 189–297.
3] C.V. Ovesen, B.S. Clausen, J. Schiotz, P. Stoltze, H. Topsoe, J.K. Norskov, J. Catal.
4
[
[
42] A.M. Prakash, S. Unnikrishnan, K.V. Rao, Appl. Catal. A: Gen. 110 (1994) 1–10.
43] A. Gronvold, K. Moljord, T. Dypvik, A. Holmen, Stud. Surf. Sci. Catal. 81 (1994)
1
68 (1997) 133–142.
[
[
[
[
[
[
4] G.A. Olah, A. Goeppert, G.K.S. Prakash, J. Org. Chem. 74 (2009) 487–498.
5] C. Song, Catal. Today 115 (2006) 2–32.
6] C.N. Hamelinck, A.P.C. Faaji, J. Power Sources 111 (2002) 1–22.
7] J. Liang, H. Li, S. Zhao, W. Guo, R. Wang, M. Ying, Appl. Catal. 64 (1990) 31–40.
8] M. Stöcker, Microporous Mesoporous Mater. 29 (1999) 3–48.
9] D. Chen, K. Moljord, T. Fuglerud, A. Holmen, Microporous Mesoporous Mater.
399–404.
[
[
[
[
[
44] D. Chen, H.P. Rebo, A. Gronvold, K. Moljord, A. Holmen, Microporous Meso-
porous Mater. 35–36 (2000) 121–135.
45] P.F. Wang, A.L. Lv, J. Hu, J.A. Xu, G.Z. Lu, Microporous Mesoporous Mater. 152
(
2012) 178–184.
46] T. Álvaro-Mu n˜ oz, C. Márquez-Álvarez, E. Sastre, Catal. Today 213 (2013)
19–225.
47] H.-G. Jang, H.-K. Min, J.K. Lee, S.B. Hong, G. Seo, Appl. Catal. A: Gen. 437–438
2012) 120–130.
48] T. Álvaro-Mu n˜ oz, C. Márquez-Álvarez, E. Sastre, Catal. Today 215 (2013)
08–215.
2
9 (1999) 191–203.
2
[
[
[
10] S. Wilson, P. Barger, Microporous Mesoporous Mater. 29 (1999) 117–126.
11] T.V.W. Janssens, S. Svelle, U. Olsbye, J. Catal. 308 (2013) 122–130.
12] D. Mores, J. Kornatowski, U. Olsbye, B.M. Weckhuysen, Chem. Eur. J. 17 (2011)
(
2
874–2884.
2
[
[
13] D. Mores, E. Stavitski, M.H.F. Kox, J. Kornatowski, U. Olsbye, B.M. Weckhuysen,
Chem. Eur. J. 14 (2008) 11320–11327.
[49] G. Sastre, D.W. Lewis, J. Chem. Soc., Faraday Trans. 94 (1998) 3049–3058.
[50] M. Popova, C. Minchev, V. Kanazirev, Appl. Catal. A: Gen. 169 (1998) 227–235.
[51] X. Wu, M.G. Abraha, R.G. Anthony, Appl. Catal. A: Gen. 260 (2004) 63–69.
14] D.S. Wragg, M.G. O’Brien, F.L. Bleken, M. Di|Michiel, U. Olsbye, H. Fjellvåg,
Angew. Chem. Int. Ed. 51 (2012) 7956–7959.