Microporous and Mesoporous Materials 195 (2014) 284293

Contents lists available at ScienceDirect

Microporous and Mesoporous Materials

journal homepage: www.elsevier.com/locate/micromeso

Differences among the deactivation pathway of HZSM-5 zeolite

and SAPO-34 in the transformation of ethylene or 1-butene to propylene
Eva Epelde, Mara Ibaez, Andrs T. Aguayo, Ana G. Gayubo, Javier Bilbao, Pedro Castao
Chemical Engineering Department, University of the Basque Country, P.O. Box 644, 48080 Bilbao, Spain

a r t i c l e

i n f o

Article history:
Received 21 January 2014
Received in revised form 16 April 2014
Accepted 22 April 2014
Available online 30 April 2014
Keywords:
Ethene
Propene intensication
MFI zeolite
CHA chabazite
Coke deactivation

a b s t r a c t
The deactivation of HZMS-5 and SAPO-34 catalysts has been studied in the transformation of ethylene
and 1-butene under propylene intensication conditions. The deterioration of spent catalysts physical
properties have been quantied and coke has been characterized by TPO and by several spectroscopic
techniques (Raman, 13C NMR, FTIR, FTIR-TPO), in order to determine the effect reaction medium composition and the severity of catalyst shape selectivity have on the nature and location of the coke in the porous structure. The results reveal that the mechanism for coke deactivation consists of two steps: one for
the formation of alkylated aromatics by oligomerization and another for the coke growth-condensation.
The rst step is analogous for both catalysts and it principally depends on the catalyst acid strength and
acid site density. The second step is different for both catalysts: the microporous structure of SAPO-34,
with cavities in the intersections, inhibits the diffusion of alkylated aromatics towards the outside of
the structure, thus blocking active acid sites; whereas, HZSM-5 structure, with a high connectivity and
without cavities, favors the diffusion of the aromatics that evolve for a longer time outside of the micropores. At process conditions, the results demonstrate that the coke formation is faster from ethylene than
from 1-butene, due to the lower reactivity of ethylene for oligomerization-cracking mechanisms as well
as its higher capability for coke formation.
2014 Elsevier Inc. All rights reserved.

1. Introduction
The increasing light olen demand is boosting the intensication of conventional processes for its production from crude oil,
by steam cracking and uid catalytic cracking (FCC) [1,2], and from
other fossil sources (natural gas, coal) by the MTO (Methanol to
Olens) process [3]. New catalytic routes from fossil fuels are currently at different developing stages, such as: (i) the transformation of DME (DTO process) obtained from natural gas or coal [4];
(ii) methane transformation (natural gas) via chloromethane [5];
(iii) coupled methanol-hydrocarbon cracking (CMHC) [6]. In addition, light olen production is a priority goal in biomass valorization routes, by ethanol or bio-oil catalytic transformation [79]
and in the valorization of polyolenic plastic wastes [10,11].
Although the aforementioned processes are selective for light
olen production, they do not satisfy the increasing propylene
demand (with an annual rate of 5.7%), due to the increase in
the production of petrochemicals, such as polypropylene and

Corresponding author. Tel.: +34 94601 8435; fax: +34 94 601 3500.
E-mail address: pedro.castano@ehu.es (P. Castao).
http://dx.doi.org/10.1016/j.micromeso.2014.04.040
1387-1811/ 2014 Elsevier Inc. All rights reserved.

acrylonitrile, among others [12]. Consequently, it is necessary: (i)

to intensify propylene production, by using selective catalysts
and by modifying the reaction conditions in light olen production
processes [13]; (ii) to explore olen interconversion routes, by
selectively transforming ethylene and butenes into propylene
[14,15]. HZSM-5 catalyst is the most widely used for the production of propylene from different raw materials and is conventionally used as an additive in the FCC units for increasing light
olen yield [16].
The suitable kinetic performance of HZSM-5 zeolite is due to its
crystalline three-dimensional structure, characteristic of pentasil
family (MFI topology), congured by ve tetrahedral rings, which
form two types of channels that are perpendicularly crossed. These
ellipsoidal channels, some straight (0.53  0.56 nm) and some in
zig-zag (0.51  0.55 nm), are 10-membered ring openings [17].
This structure provides the HZSM-5 zeolite with a suitable balance
between different properties: shape selectivity, acid strength of the
sites, acid site density, crystalline channel interconnection and
hydrothermal stability. Light olen interconversion takes place
by the well-established oligomerization-cracking mechanism
[18,19], with secondary reactions of isomerization, cyclation,
aromatization, together with the undesirable coke formation. In

E. Epelde et al. / Microporous and Mesoporous Materials 195 (2014) 284293

285

Nomenclature
Roman symbols
B/L
Brnsted/Lewis ratio
Cc
coke content (wt.%)
D, G
bands in the Raman spectrum
dp
pore diameter ()
fc1
fraction of coke type I (%)
Fi
molar owrate of i lump in the product stream ((mol of
C) h1)
F0 and F butene molar owrate in the feed and in the outlet
stream ((mol of C) h1)
Gp
position of G band (cm1)
Gw
width of G band (cm1)

order to intensify propylene production, as well as to attenuate

deactivation by coke, several initiatives for HZSM-5 zeolite alteration have been suggested aimed to: (i) modifying the porous
structure, by generating mesopores and/or hierarchichal structures
by desilication or dealumination [9,20,21]; (ii) decreasing crystal
size [16,22]; (iii) altering acidity by selecting a suitable SiO2/
Al2O3 ratio and by adding different compounds (P, K, La, transition
metals) [2327] and; (iv) incorporating the zeolite in a composite
[28]. These modications result in a decrease in the acid strength
and acid site density, effective for increasing zeolite hydrothermal
stability and for decreasing coke formation, as hydrogen transfer
reactions are disfavored [9,29,30].
In the MTO process, industrially commercialized by UOP/Norsk
Hydro technology, the SAPO-34 catalyst is used [31], of CHA topology, with big cages (0.67  0.11 nm) interconnected through small
windows (0.38  0.38 nm) formed by 8-membered rings [32]. This
porous structure allows forming poly-alkylated benzenes in the
cages, which are active intermediates in the formation of light olens from methanol, and furthermore, it restricts the diffusion of
heavy and branched hydrocarbons in the narrow windows
between the cages [33,34]. In order to increase propylene selectivity and extend the catalytic life (the main problem of this catalyst),
several modications of SAPO-34 have been studied, as are: (i)
decreasing crystal size [35]; (ii) tuning acidity, by oxalic acid treatment or doping with Ce [36]; and, (iii) agglomerating it with ZrO2
as a binder [37]; and, (iv) introducing hierarchichal pore systems
via templating with carbon materials [38].
The modications of HZSM-5 zeolite and SAPO-34 allow obtaining a high olen yield, with high propylene selectivity. However,
both catalysts (largely SAPO-34) show a rapid deactivation by coke,
which justies the growing interest in the knowledge of its formation mechanism. In this work the deactivation by coke of HZSM-5
zeolite and SAPO-34 catalysts has been studied in the transformation of 1-butene and ethylene, under suitable reaction conditions
for intensifying propylene production. The combination of the
results of catalyst properties deterioration and the nature of coke
(studied by TPO, MS/FTIR-TPO, FTIR, Raman and 13C NMR) pursue
improving the knowledge about the effect of catalyst properties
(particularly shape selectivity severity, which is different for both
catalysts) and the composition of the reaction medium (which is
different for the 1-butene and ethylene feeds) on the mechanism
for coke formation. This study is supported on the literature
knowledge about the mechanisms of coke formation on HZSM-5
zeolite and SAPO-34 catalysts used in other catalytic processes
for olen production, such as cracking [3941] and the MTO process [42,43]. Moreover, it is based on previous knowledge about
the spectroscopic analysis methodology of the coke deposited on
acid catalysts [4448].

La
SBET
Sm
Si
Vm
Vp
X

in-plane correlation length (nm)

BET surface area (m2g1)
micropore area (m2g1)
selectivity of i component (%)
micropore volume (cm3g1)
pore volume (cm3g1)
conversion (%)

Greek symbols
eL , eB
molar extinction coefcients (cm lmol1)

2. Experimental
2.1. Catalysts
HZSM-5 zeolite was supplied in ammonium form by Zeolyst
International (SiO2/Al2O3 = 80), and was calcined at 570 C to
obtain the acid form. SAPO-34 (of composition (SiO2)0.08
(Al2O3)2.8 (P2O5)0.98 (HCl)0.02 (C8H19N) 41H2O) was prepared following the method of Lok et al. [49], with tetraethyl ammonium
hydroxide (TEAOH) at 20 wt.% as template. The removal of the
template was carried out by calcination in a mufe furnace at
575 C for 6 h with a heating rate of 5 C min1.
The catalysts were obtained by agglomerating the active phase
(25 wt.%) by wet extrusion with bentonite (Exaloid, 30 wt.%) as a
binder and alumina (Martinswek, 45 wt.%) as inert charge. This
agglomeration is required in order to: (i) confer a suitable mechanical and hydrothermal resistance; (ii) obtain higher accessibility of
the reactants to the active phase, and; (iii) attenuate deactivation
by coke, since the coke deposition on the mesoporous matrix minimizes the blockage at the mouth of micropores [50]. The extrudates were rst dried (110 C, 24 h) and then were sieved to a
particle diameter between 0.15 and 0.3 mm. Finally, the catalysts
were calcined at 570 C for 2 h. This treatment is required in order
to achieve a balance of the acid sites, which allows keeping the catalyst hydrothermally stable and maintaining its kinetic performance throughout reaction-regeneration cycles [46].
The surface area and porous structure were measured by N2
adsorption-desorption (Micromeritics ASAP 2010). The total acidity and acid strength of the catalysts were determined by monitoring the adsorption-desorption of NH3, by combining the techniques
of thermo-gravimetric analysis and differential scanning calorimetry using a Setaram TGDSC calorimeter connected on-line with a
Thermostar mass spectrometer from Balzers Instruments [51]. The
Brnsted/Lewis (B/L) acid site ratio has been determined by analyzing the region of 14001700 cm1 in the FTIR spectrum of
adsorbed pyridine, which has been obtained using a Specac catalytic chamber connected on-line with a Nicolet 6700 FTIR spectrometer. The results have been determined from the ratio
between the intensity of pyridine adsorption bands at 1545 and
1450 cm1 and taking into account the molar extinction coefcients of both adsorption bands (eB = 1.67 cm lmol1 and
eL = 2.22 cm lmol1) [52].
Table 1 summarizes the physical and acid properties of the
fresh active phases and catalysts (after agglomeration). SAPO-34
shows a higher BET surface and micropore volume in comparison
to HZSM-5 zeolite. The catalyst micropore volume corresponds to
the active phase (zeolite or SAPO), whereas the volume of mesoand macropores corresponds to the matrix of the catalyst

286

E. Epelde et al. / Microporous and Mesoporous Materials 195 (2014) 284293

2.3. Coke characterization

Table 1
Properties of the active phases and agglomerated catalysts.
HZSM-5

SAPO-34

Active phase
SBET (m2 g1)
Sm (m2 g1)
Vp (m3 g1)
Vm (m3 g1)
Total acidity (mmolNH3 g1)
Acid strength (kJ mol1
NH3)
B/L ratio at 150 C

556
181
0.11
0.10
0.42
144
0.32

611
288
0.06
0.24
0.64
153

Agglomerated catalyst
SBET (m2 g1)
Sm (m2 g1)
Vp (m3 g1)
Vm (m3 g1)
dp ()

209
68
0.44
0.04
173

215
103
0.20
0.05
37

The coke deposited on the catalysts was characterized by different techniques developed in previous works [4547]. The coke
content was determined by temperature-programmed oxidation
(TPO) in a TGA Q5000 thermobalance (TA Instruments), following
a ramp of 8 C min1 from 250 C to 600 C. FTIR and combined
FTIR-TPO analysis were performed in a Nicolet 6700 (Thermo)
using a transmission cell (60 scans, and resolution of 4 cm1).
These analyses were carried out following a temperature ramp of
5 C min1 from 100 C to 550 C. Solid state CP-MAS (cross-polarization magic-angle spinning) 13C NMR spectra were obtained in a
Broker DXR 300. Raman spectroscopy was performed in a Reinshaw confocal microscope using an excitation source of 514 nm,
and subtracting the uorescence caused by coke.
3. Results

(bentonite and alumina). The total acidity of SAPO-34 (0.64

mmolNH3 g1) is higher than that for HZSM-5 (0.42), however,
the acid strength is similar. It has been proven that the contribution of the catalyst matrix to the acidity is insignicant; thus, the
total acidity per catalyst mass unit is approximately a quarter of
that corresponding to the active phase. The low Brnsted/Lewis
ratio obtained for HZSM-5 zeolite is explained by the signicant
dehydroxylation of HZSM-5 catalyst at high calcination temperatures (570 C) [6]. The Brnsted/Lewis ratio for the SAPO-34 has
not been determined due to the severe steric restriction to the
diffusion of pyridine.

2.2. Reaction equipment and product analysis

Experimental runs were carried out using a reaction equipment
described in a previous work [26]. The xed-bed reactor, made of
316 stainless steel, has an internal diameter of 0.9 cm and a length
of 10 cm. It is located inside a ceramic-covered stainless steel
cylindrical chamber that is heated by an electric resistance and
can operate up to 100 bar and 700 C with a catalyst mass of up
to 5 g. The bed consists of a mixture of catalyst and inert solid (carborundum with an average particle diameter of 0.105 mm) to
ensure bed isothermality and attain sufcient height under low
space time conditions. The temperature is controlled by a digital
TTM-125 series controller and measured by a thermocouple (type
K) situated in the catalyst bed. Two additional temperature controllers are used: one for the furnace chamber and one for the
transfer line between the reactor and the micro-gas chromatograph. The operating variables are controlled by custom software
(Process@ from PID Eng&Tech, Madrid, Spain).
A small fraction of the stream made up of unreacted reactants
and products (diluted in a He stream of 17 cm3 min1) was continuously analyzed online in a micro-gas chromatograph (Agilent
MicroGC 3000A). The remaining stream of reaction products
passed through a Peltier cell at 0 C. A level sensor controlled the
amount of liquid condensate, and the non condensable gas ow
was vented. The micro-GC (with Soprane software) is provided
with 4 analytical modules and the following columns: a molecular
sieve (MS-5A) (10 m) where H2, CO, O2, N2 and CH4 are separated;
Porapak Q (PPQ) (8 m), where oxygenates compounds (MeOH,
DME), C2C3 light olens, CO2 and water are identied; Alumina
(10 m), where C2C5 hydrocarbons are identied (parafns and
olens); and OV-1, where C5C10 fraction and aromatics BTX are
separated. Product stream was analyzed every 4 min. The compounds were identied and quantied using calibration standards
of known concentration. The balance of atoms (C, H) was closed in
all runs above 99.5%.

3.1. Effect of shape selectivity and acidity in the transformation of

ethylene and 1-butene
The catalytic performance was evaluated under the following
operating conditions: 500 C; pressure, 1.35 bar; mass of catalyst,
0.55 g; feed ow rate (1-butene or ethylene) 0.35 cm3 min1, He
ow rate 5 cm3 min1; space time of 6.4 (gcatalyst h) (mol)1; time
on stream, 5 h. Fig. 1 shows the evolution with time on stream
(TOS) with conversion, which has been dened as follows:

F0  F
 100
F0

where F0 and F are the molar ow rates of the reactant (1-butene or

ethylene) in the feed and outlet stream, both terms being expressed
in C atoms contained. 1-butene rapidly reaches (at the reactor
entrance) the thermodynamic equilibrium of the isomerization,
with the rest of butenes (iso-, cis- and trans-butene). Consequently,
F corresponds to the total ow rate of butenes.
The low initial activity of SAPO-34 catalyst is consistent with
the C2C4 olen oligomerization-cracking studies [14,19], which
is explained by the limitation of the oligomerization steps on the
micropores, with a higher inuence on ethylene oligomerization
that is less active than 1-butene [53]. Furthermore, the fast deactivation of SAPO-34 is due to the micropore blockage by the coke
formed within the cavities (cages) between the crystalline channels. This deactivation is faster in the transformation of ethylene

Fig. 1. Evolution with time on stream of ethylene and 1-butene conversion.

287

E. Epelde et al. / Microporous and Mesoporous Materials 195 (2014) 284293

than 1-butene, since the coke formation is favored by a lower

extent of oligomerization-cracking steps. The rapid deactivation
of SAPO-34 in the transformation of ethylene has also been
reported in the literature [14,54]. However, the deactivation
observed in the transformation of 1-butene is slower than that
detailed in [55,56], which is attributed to the agglomeration of
SAPO-34 with bentonite and alumina that contributes to delaying
the micropore blockage.
The lower deactivation rate observed for HZSM-5 catalyst is due
to the absence of cavities in the intersections between crystalline
channels, whose three-dimensional structure favors the diffusion
of the coke precursors towards the outside of the zeolite crystals
avoiding the micropore blockage by coke. This suitable performance of HZSM-5 zeolite is well-known in hydrocarbon cracking
reactions [41,57], particularly in those with light olens as coke
precursors [48].
The selectivities for the different lumps of products calculated
at zero time on stream are summarized in Fig. 2. The products have
been grouped into: (i) methane; (ii) ethylene; (iii) propylene; (iv)
C2C3 light parafns; (v) C4H10, including isobutane and n-butane;
(vi) butenes; (vii) C5+ aliphatics, which include all the olens and
parafns with more than ve carbon atoms, and; (viii) aromatics
BTX (benzene, toluene and xylenes). Lump selectivities have been
calculated as follows:

Si

Fi
 100
F0  F

where Fi is the molar ow rate of the i lump in the product stream,

expressed in C atoms contained.
As a consequence of the severe shape selectivity, it is noteworthy the higher propylene selectivity on the SAPO-34 catalyst,
which is of 79% for ethylene transformation. In addition, the C5+
non aromatic lump is mainly composed of linear C5 olens, whose
kinetic diameter enables the diffusion in the pores. C2-C4 light parafn, and especially BTX selectivity is low as steric limitations inhibit hydrogen transfer and aromatization reactions.

Fig. 2. Selectivity of product lumps at zero time on stream in the transformation of

ethylene (a) and 1-butene (b).

Table 2
Physical properties of the catalysts deactivated after TOS = 5 h.
Catalyst

SBET (m2 g1)

Ethylene
HZSM-5
SAPO-34

91
89

1-Butene
HZSM-5
SAPO-34

97
76

Sm (m2 g1)
8.3
0
10
0

Vp (m3 g1)

Vm (m3 g1)

dp ()

0.35
0.14

0.005
0

197
62

0.37
0.12

0.006
0

185
63

3.2. Deterioration of catalyst properties

Table 2 summarizes the physical properties of the deactivated
catalysts after 5 h time on stream. The comparison between these
values and those for the fresh catalysts (Table 1) shows that the
surface area notably decreases, particularly the micropore area
which is negligible for the SAPO-34 catalyst in both feeds. As the
proportion of micropore volume decreases, the average pore diameter increases for the different catalysts studied. This selective
deterioration of micropore properties reveals that coke deposition
preferably takes place on them. The total micropore blockage in
SAPO-34 catalyst is noteworthy, which indicates that micropores
are blocked by coke. Furthermore, the deterioration of BET surface
on HZSM-5 zeolite is conditioned by micro- and especially mesopore volume loss.
3.3. Coke properties
3.3.1. TG-TPO analysis
Fig. 3 shows the TPO proles corresponding to the combustion
of the coke deposited on the catalysts used in the transformation of
ethylene and 1-butene. Two types of coke are identied: coke I,
which burns at low temperature (maximum of the combustion
peak at 485 C), and coke II, which peaks at 550570 C [58]. Coke
I is related to more hydrogenated species, whereas coke II to more
carbonaceous ones [46]. Under the studied conditions, it is well
established in the literature [41,57] that the origin of the coke is
the condensation of the reaction medium compounds, with steps

Fig. 3. TPO proles for the coke deposited on the catalysts used in the transformation of ethylene and 1-butene.

288

E. Epelde et al. / Microporous and Mesoporous Materials 195 (2014) 284293

that are activated by the acid sites mainly located within the
crystals.
TPO proles in Fig. 3 have been deconvoluted into two gaussian
peaks (as shown in the lower signal in Fig. 3), to quantify the total
coke content (Cc) and the amount of coke I fraction (fI) that are
summarized in Table 3. The coke content is similar for both catalysts in the transformation of ethylene (around 4.6 wt.%), even
though the effect of coke on deactivation is different (Fig. 1).
SAPO-34 catalyst activity is null at 5 h time on stream, whereas
for HZSM-5 catalyst, ethylene conversion is 78%. This higher stability of HZSM-5 zeolite (also observed by comparing the results in
Fig. 1 for 1-butene transformation), for the same coke content, is
due to its porous structure without cages in the crystalline channel
intersections, which avoids the formation of polyaromatic coke
structures within the crystalline structure [57,59]. The presence
of these cages in the SAPO-34 favors the retention of the alkylated
aromatics generated by oligomerization and hinders coke combustion; therefore the maximum coke combustion temperature is
20 C higher for the coke deposited on SAPO-34 (Fig. 3).

3.3.2. FTIR spectroscopy

Fig. 4 shows the two most important regions of molecular
vibration in the infrared spectrum of the coke deposited on the catalysts in the transformation of ethylene and 1-butene. The
observed bands are assigned to [60,61]: 1390 cm1, branched
aliphatics and other bonds univocally unidentied; 1460 cm1, aliphatics and aliphatics linked to aromatics; 1590 cm1, polyaromatic hydrocarbons (PAH) and condensed coke; 1620 cm1,
dienes and conjugated double bonds in carbon chains;
2855 cm1, CH2 groups; 2870 cm1, CH3 groups; 2925 cm1, CH2
Table 3
Total coke content (Cc) and fraction of coke I (f C1) in deactivated catalysts.
Catalyst

Cc (wt.%)

f C1 (%)

Ethylene
HZSM-5
SAPO-34

4.7
4.6

47.1
10.7

1-Butene
HZSM-5
SAPO-34

2.7
4.1

37.8
12.2

(a)

and CH groups; 2955 cm1, CH3 groups; 3030 cm1, simple

and alkylated mono-aromatics.
SAPO-34 catalyst shows a higher intensity of the band at
1620 cm1 (Fig. 4), which is attributed to a higher proportion of
olenic coke on this catalyst. Moreover, the presence of the bands
at 1390 and 1460 cm1 gives evidence of a high concentration of
branched aliphatic species in the coke deposited on both catalysts.
It is also observed that the coke deposited on HZSM-5 catalyst
shows a higher intensity of the band of CH3 groups (2955 cm1)
in proportion to CH2 (2925 cm1), in comparison to those
observed for SAPO-34 catalyst. This result indicates that there exist
a higher proportion of terminal parafnic groups in the coke
formed on HZSM-5 zeolite. It should be mentioned that the nature
and composition of the coke deposited in the transformation of
ethylene is relatively similar to that obtained using the 1-butene
feed. However, in the case of the transformation of ethylene on
SAPO-34 a higher intensity of diene band (1620 cm1) is observed
(Fig. 4) in comparison to the band of coke (1590 cm1), which is
indicative of a higher proportion of olenic coke.
In Fig. 5 the relative intensity of the most representative bands
of the coke deposited on both catalysts in the transformation of
ethylene and 1-butene is compared. The intensity of the bands corresponding to the compounds of aromatic (1590 cm1) and olenic
(1620 cm1) nature is higher for SAPO-34, in both feeds. The high
content of olenic groups in the coke deposited on SAPO-34 catalyst shows the relevant role that olen oligomerization plays in the
formation of the coke that remains retained in the intersection cavities of the crystalline channels. The band at 3030 cm1 indicates
the higher proportion of mono-aromatics in the coke of SAPO-34
in comparison to that deposited on HZSM-5 zeolite. Furthermore,
the ratio between the relative intensity of the bands for aliphatic
groups (CH, CH2 y CH3) is lower for the SAPO-34, which means
that the aliphatic chains are shorter and proportionally in lower
amounts than in the coke deposited on HZSM-5 catalyst. The
higher proportion of aromatics on the coke of SAPO-34 reveals that
the location is closer to the strong acid sites where aromatization
reactions take place.

3.3.3. FTIR-TPO analysis

The continuous monitoring by FTIR analysis during the TPO
experiments of the coke combustion allows analyzing the

(b)

Fig. 4. FTIR spectra of the catalysts spent in the transformation of ethylene and 1-butene in the 1300-1800 (a) and 2800-3100 (b) cm-1 ranges.

E. Epelde et al. / Microporous and Mesoporous Materials 195 (2014) 284293

289

between the maximums observed in TG-TPO (Fig. 3) and FTIRTPO (Fig. 6) analysis is due to the coke sweeping in the former,
which promotes coke aging reactions [62].
The combined results of TG-TPO and FTIR-TPO imply that coke I
burns easily due to its higher H/C ratio and its location in the mesopores of the catalyst. Coke II is either more condensed (as seen in
Fig. 6) or it is presumably located within the micropores of the zeolite. It burns more slowly and requires higher combustion temperatures; thus, the coke II deposited on SAPO-34 needs 20 C more to
burn in comparison to the one deposited on HZSM-5 zeolite,
because of the limited O2 access as well as O2-coke contact. An initial step for the formation of a coke of lower molecular weight
(alkylated aromatics) should be considered that is favored by the
high density of strong acid sites and because of the cages in the
intersection between micropores [63]. Moreover, the aliphatic
compounds of coke I probably evolve towards more condensed
structures, as a high mesopore volume is able to contain coke
molecules.

Fig. 5. Relative intensities of the FTIR bands corresponding to the catalysts spent in
the transformation of ethylene and 1-butene.

evolution of coke composition as well as the characterization of the

two types of coke observed by TPO (Fig. 3) [45,48]. Fig. 6 shows
the evolution of FTIR signals during the combustion process where
the aforementioned most representative bands are identied:
1590 cm1, PAH; 1620 cm1, dienes; 2925 cm1, CH2 groups
and CH; 2955 cm1, CH3 groups.
The combustion results of the coke deposited on HZSM-5 catalyst in the transformation of ethylene (Fig. 6a) and 1-butene
(Fig. 6b) demonstrate the selective combustion of the coke compounds, as a consequence of its heterogeneity; thus, the maximum
of disappearance velocities for CH3 groups peaks at 260 C, for
CH2 groups at 280 C, for dienes at 480 C and for the aromatics
at 505 C. The results of Fig. 6 demonstrate that coke type I and
II (Fig. 3) can be correlated with their composition: coke I is of
more aliphatic nature, whereas coke II is more aromatic and condensed. Furthermore, the composition may also be related to the
location of the coke, since the proximity to the catalytic acid sites
would imply a higher capability for aromatization. Therefore, coke
I is probably externally located whereas coke II is internally
located. It should be mentioned that the temperature difference

3.3.4. Raman spectroscopy

Fig. 7 shows the Raman spectra of the spent catalysts. The main
peaks correspond to [46]: 1350 cm1, denoted D band and assigned
to disordered aromatic structures; 1580 cm1, G band, related to
structured aromatics. The sub-graph in Fig. 7 details the shape
and position of G band whereas in Table 4 the values of its width
(Gw) and position (Gp) are summarized.
For both reactions it is observed that the G band for the coke
deposited on HZSM-5 is narrower and shifts towards higher wavenumber values in comparison to that deposited on SAPO-34. These
differences indicate that the coke deposited on HZSM-5 catalyst is
more structured and less amorphous than that corresponding to
SAPO-34. There exists a potential correlation between Gw
(expressed in cm1) and the in-plane correlation length La (in
nm), which also quanties the coke particle size:

logGw 2:696 logLa 1:954

This equation has been deduced from the experimental data of

other works [61,64]. The calculated values of La are shown in
Table 4. Taking into account the Gp values and the D/G band ratio
obtained and based on previous results [61], it must be considered
that the coke deposited on HZSM-5 zeolite has similar properties to
a nano-crystalline coke, whereas the coke deposited on SAPO-34 is
similar to an amorphous carbon. Based on this last result and

Fig. 6. Evolution of differential intensities of absorbance bands with time during the temperature programmed combustion of the coke deposited on H-ZSM5 catalyst in the
transformation of ethylene (a) and 1-butene (b).

290

E. Epelde et al. / Microporous and Mesoporous Materials 195 (2014) 284293

Fig. 8. 13C NMR spectrum of the catalysts spent in the transformation of ethylene
and 1-butene.
Fig. 7. Raman spectra of the catalysts spent in the transformation of ethylene and
1-butene.

Table 4
Representative parameters of the Raman bands, corresponding to the coke deposited
on the catalysts.

a
b

Catalyst

Gp (cm1)

Gw (cm1)

La (nm)a

D/G

La (nm)b

Ethylene
HZSM-5
SAPO-34

1599.3
1595.9

43.9
69.4

1.12
1.04

2.03
1.18

1.92
1.46

1-Butene
HZSM-5
SAPO-34

1601.3
1594.3

44.9
53.4

1.12
1.09

1.97
1.50

1.89
1.65

Calculated by Eq. (3).

Calculated by Eq. (4).

considering the calculated La values, it should be mentioned that

the growth scheme corresponds to the Ferrari-Robertson expression [65]:

D=G 0:55 La2

principal regions are distinguised [58,67]: 10-40 ppm, aliphatic C

nuclei; and 120150 ppm, aromatic C nuclei. These regions are
desglosed in: 1014 ppm, CH3 groups; 1822 ppm, CH2 groups;
28-32 ppm, CH groups; 129 ppm, aromatic hydrocarbons (AH)
and 141 ppm, polycondensed aromatic hydrocarbons (PAH) [46].
The spectra have been deconvoluted and the intensities of the
bands described above are summarized in Table 5.
The higher width of the aromatic coke band deposited on
HZSM-5 catalyst in comparison to the band for SAPO-34 (ca. 40
vs. 20 ppm respectively) evidences a notable more heterogeneity
of the aromatic species in the coke deposited on HZSM-5 catalyst.
The band at 141 ppm only shown for HZSM-5 catalyst, also raties
the deposition of a more developed coke. This coke size, according
to Raman spectroscopy results (Fig. 7) and pore volume distribution (Table 2), is considerably big enough in order to be located
outside of the micropores. In addition, the coke deposited on

Besides the discrepancy between the La values calculated by

Eqs. (3) and (4) (Table 4), both results reveal a higher condensation
degree as well as higher coke particle size of the structured coke on
HZSM-5 zeolite in comparison to the coke on SAPO-34, this difference being more signicant in the transformation of ethylene. Nevertheless, slight differences are observed in the Raman spectra
between both reactions. Overall, carbonaceous materials with Gw
higher than 50 cm1 (as the case of SAPO-34) have La values lower
than 1 nm [61], whereas materials with a D/G ratio close to 2
(HZSM-5 zeolite) show values around 2 nm [66].
3.3.5. 13C NMR spectroscopy
The bands of 13C NMR for the deactivated catalysts in the transformation of ethylene and 1-butene are shown in Fig. 8. Two

Table 5
Fraction of the bands deconvoluted from the
coke deposited on the catalysts.

13

C NMR proles, corresponding to the

Band

Assignation

HZSM-5

SAPO-34

Ethylene
1014 ppm
1822 ppm
2832 ppm
129 ppm
141 ppm

CH3
CH2
CH
CAH
CPAH

0.180
0.219
0.240
0.351
0.011

0.010
0.090
0.014
0.886
0.000

1-Butene
1014 ppm
1822 ppm
2832 ppm
129 ppm
141 ppm

CH3
CH2
CH
CAH
CPAH

0.079
0.176
0.145
0.545
0.054

0.015
0.146
0.119
0.719
0.000

E. Epelde et al. / Microporous and Mesoporous Materials 195 (2014) 284293

HZSM-5 zeolite has a high proportion of aliphatic chains (10

40 ppm).
The results in Table 5 exhibit a higher aromatic carbon content
in the coke deposited on SAPO-34 (70%) in comparison to the coke
deposited on HZSM-5 zeolite (40%). Furthermore, the insignicant
amount of polyaromatic carbon on the coke deposited on SAPO-34
catalyst is justied by the severe growth limitations due to its location in the micropores of the catalyst (Table 2). The coke deposited
on HZSM-5 is preferably located on the mesopores of the matrix
and therefore, it has lower restrictions for growing by aromatic
condensation or by aliphatic chain extension.

4. Discussion
The results in Section 3 allow analyzing the effect of the reaction medium composition and catalyst shape selectivity on the
magnitude of coke deposition and its structure. The higher coke
formation velocity in the transformation of ethylene is the result
of two circumstances: (i) the higher capability of ethylene for condensation in order to form coke compounds, with respect to other
olens; (ii) its lower reactivity for oligomerization-cracking mechanisms; thus, ethylene concentration in the reaction medium is
higher than the corresponding to 1-butene. This selective effect
of light olen lump concentration on deactivation provides
interesting information to be considered in a kinetic model for
the deactivation of olen interconversion reactions.

291

In addition, differences between the coke composition for both

feeds have been observed, even though these differences are
irrelevant, which is justied by the similar mechanism for coke
formation that implies a rst step for alkylated aromatic formation from the oligomers and a second step for coke growth. The
rst step is favored by a higher density and acid strength of the
sites as well as by enough space available in the micropores in
order to form molecules of high size. The signicant amount of
alkylated aromatics in the coke (Figs. 5 and 8) allows supporting,
by analogy with the deactivation in the MTO process, the role of
these intermediates with a high condensation capability for coke
formation [68,69]. This result is consequence of the presence of
C2-C4 olens in the reaction medium, which evolves together
with the alkylated aromatics by hydrogen transfer, cyclation, aromatization and condensation reactions [41,57].
The most signicant differences come from the different deactivation pathways observed for SAPO-34 and HZSM-5 zeolite, which
are shown in Fig. 9 and detailed below.
The higher shape selectivity severity of SAPO-34 and particularly the existence of cages in the intersections of microporous
channels inhibit the diffusion of alkylated aromatics. These compounds condense to a greater extent by aromatization reactions
until the total blockage of the micropores (Table 2) and the consequent catalytic activity death (Fig. 1). It has even been suggested
that this coke growth causes a unit cell expansion for the SAPO-34
[70]. Furthermore, because of the proximity of the acid sites, the
composition of the coke deposited on SAPO-34 is mainly aromatic

Fig. 9. Scheme of the internal coke precursor retention and of the composition of the external coke deposited on HZSM-5 zeolite and SAPO-34.

292

E. Epelde et al. / Microporous and Mesoporous Materials 195 (2014) 284293

and olenic. Due to the low capability for sweeping alkylated aromatic compounds towards the exterior, the fraction of external
coke is small and the internal coke fraction is formed, based on
the deactivation prole (Fig. 1), at low values of time on stream.
Furthermore, the deactivation of HZSM-5 zeolite catalyst is less
severe (Fig. 1), owing to the expulsion of the alkylated aromatic
compounds formed in the rst step, which are deposited on the
mesopores and grow by aromatic and aliphatic condensation. The
nal composition of the coke is more condensed-structured, with
long parafnic chains (Fig. 8) and a lower olenic content. These
results of composition and growth are based on the external location of the coke outside of the micropores and justify the bimodal
distribution of coke I and II. The good qualities of HZSM-5 zeolite
for facilitating the circulation of the potential coke precursors
and preventing the micropore blockage have been highlighted in
the literature for parafn [41] and polyolen cracking [45]. In addition, using a high gas ow rate and agglomerating HZSM-5 zeolite
in a mesoporous matrix that houses the external coke are suitable
strategies for avoiding the blockage of the pore mouths [43].
Regarding the results obtained, catalyst selection is important
in order to attenuate the deactivation by coke in these reactions.
HZSM-5 zeolite deactivates slower and moderating acid strength
attenuates this deactivation. This strategy is feasible in the transformation of 1-butene at the studied temperature, on a zeolite with
high SiO2/Al2O3 ratio that can be steamed or doped in order to control its acid strength [9,20,21,23,26]. However, the transformation
of ethylene at the studied temperature requires a considerable acid
strength to ensure catalyst activity, thus, the attenuation of acid
strength should be accomplished by a raise in temperature for
maintaining a high conversion level. SAPO-34 deactivates more
rapidly and has more hurdles for modifying its porous structure
and acidity for attenuating the deactivation by coke [71].
The ability of HZSM-5 zeolite for avoiding micropore blockage
under the studied reaction conditions explains the lower deactivation rate of the catalyst than that expected for the coke content
obtained or the deterioration of surface area. HZSM-5 has a highly
connected micropore network that enhances the diffusion, without
cages in their intersections. In addition, the conned molecules
have a signicant mobility across the zeolite and their growth is
limited due to the relatively small diameter of the pores, thus hindering the condensation towards bigger molecules [41]. Furthermore, the mobility of coke precursors enhances their circulation
towards the outside of HZSM-5 zeolite, and therefore a signicant
fraction of the coke corresponds to that deposited on the outside of
the crystals. In the case of SAPO-34, the intersections between the
channels are cages that easily inhibit the circulation of the reactants. Thus, the outward circulation of intermediate compounds
responsible for coke formation is also inhibited and so the coke
deposited on this catalyst is mainly internal coke. Thus, the key
point is that the location of coke in HZSM-5 catalyst would not
be necessarily close to the acid sites; coke is denitively formed
on the acid sites, however it migrates to the exterior of the zeolite
and grows, whereas in the case of SAPO-34 it stays trapped within
the micropores.
In addition to the catalyst properties, the dilution of the feed
with inerts as N2 can contribute to attenuating deposition by coke,
thereby decreasing the concentration of coke precursors and the
velocity for their condensation [41]. Dilution with H2 would presumably be more effective for decreasing coke deposition since it
will also hinder the condensation of coke precursors that takes
place by dehydrogenation [57]. This hydrogenation can be periodically applied for catalyst rejuvenation between reaction steps,
with an expected partial recovery of activity, whose total recovery
will require coke combustion with air.

5. Conclusions
HZSM-5 zeolite and SAPO-34 catalysts are interesting alternatives for the selective light olen production and provide good
prospects for intensifying propylene production from ethylene
and 1-butene. However, their use is conditioned by the fast deactivation by coke. The results reveal that 1-butene shows better prospects than ethylene for obtaining propylene, owing to its higher
reactivity and lower capability for forming coke. Nevertheless, considering coke composition, no signicant differences are observed
that could indicate a different deactivation mechanism in the
transformation of ethylene and 1-butene.
Coke formation on both catalysts takes place following two
steps: the formation of alkylated aromatics and their growth to
form coke. The oligomers are presumably the intermediate compounds of alkylated aromatic formation and oligomerizationcracking reactions, which are crucial for interconverting olens.
The formation of alkylated aromatics as well as the reactivity of
intermediates is conditioned by the density and acid strength of
the sites, and these properties, among others, justify the amount
and impact of the coke.
The shape selectivity of SAPO-34 is a key factor for the high propylene selectivity giving way to high olen (propylene) selectivity
at zero time on stream. However, this catalyst shows a fast deactivation caused by the pore blockage due to the alkylated aromatic
retained in the internal cages of microporous intersections, giving
way to a coke with a high amount of non-condensed rings and of
olenic nature.
HZSM-5 zeolite deactivates more slowly as these alkylated aromatics are swept to the outside of the micropores, and therefore,
they are deposited and grow on the mesopores of the matrix. The
higher micropore diameter and the absence of cages in the intersections enable this external deposition. Consequently, the coke
deposited on HZSM-5 zeolite catalyst has a higher external coke
fraction and is mainly composed of condensed aromatics and long
aliphatic chains. The ability for the internal circulation of coke precursors is favored by agglomerating the HZSM-5 zeolite in a mesoporous matrix, which allows the deposition and evolution of the
coke on the outside of the zeolite thus avoiding the blockage of catalytic acid sites.
Acknowledgements
The nancial support of this work was undertaken by the Ministry of Economy and Competitiveness (MINECO) of the Spanish
Government (CTQ2010-19188 and CTQ2010-19623 projects), by
the Basque Government (Project IT748-13) and by the University
of the Basque Country (UFI 11/39). E. Epelde (BFI08.122) and M.
Ibaez (BFI-2012-203) are grateful for their Ph.D. Grants from the
Department of Education, University and Research of the Basque
Country. The technical and human support provided by SgIker
(UPV/EHU, MICINN, GV/EJ, ESF) is gratefully acknowledged.
References
[1]
[2]
[3]
[4]

T. Ren, M.K. Patel, K. Blok, Energy 33 (2008) 817833.

A. Corma, J. Mengual, P.J. Miguel, Appl. Catal. A: Gen. 417418 (2012) 220235.
T. Mokrani, M. Scurrell, Catal. Rev. 51 (2009) 1145.
N.V. Kolesnichenko, T.I. Goryainova, E.N. Biryukova, O.V. Yashina, S.N.
Khadzhiev, Pet. Chem. 51 (2011) 5560.
[5] U. Olsbye, O.V. Saure, N.B. Muddada, S. Bordiga, C. Lamberti, M.H. Nilsen, K.P.
Lillerud, S. Svelle, Catal. Today 171 (2011) 211220.
[6] D. Mier, A.T. Aguayo, A.G. Gayubo, M. Olazar, J. Bilbao, Appl. Catal. A: Gen. 383
(2010) 202210.
[7] A.G. Gayubo, B. Valle, A.T. Aguayo, M. Olazar, J. Bilbao, Ind. Eng. Chem. Res. 49
(2010) 123131.

E. Epelde et al. / Microporous and Mesoporous Materials 195 (2014) 284293

[8] P.S. Rezaei, H. Shafaghat, W.M.A.W. Daud, Appl. Catal. A: Gen. 469 (2014) 490
511.
[9] A.G. Gayubo, A. Alonso, B. Valle, A.T. Aguayo, J. Bilbao, Appl. Catal. B: Environ.
97 (2010) 299306.
[10] P. Castao, G. Elordi, M. Ibaez, M. Olazar, J. Bilbao, Catal. Sci. Technol. 2 (2012)
504508.
[11] P. Castao, G. Elordi, M. Olazar, J. Bilbao, ChemCatChem 4 (2012) 631635.
[12] J. Plotkin, Catal. Today 106 (2005) 1014.
[13] G. Wang, C. Xu, J. Gao, Fuel Process. Technol. 89 (2008) 864873.
[14] B. Lin, Q. Zhang, Y. Wang, Ind. Eng. Chem. Res. 48 (2009) 1078810795.
[15] T.R. Koyama, Y. Hayashi, H. Horie, S. Kawauchi, A. Matsumoto, Y. Iwase, Y.
Sakamoto, A. Miyaji, K. Motokura, T. Baba, Phys. Chem. Chem. Phys. 12 (2010)
25412554.
[16] X. Gao, Z. Tang, H. Zhang, D. Ji, G. Lu, Z. Wang, Z. Tan, J. Mol. Catal. A: Chem. 325
(2010) 3639.
[17] C. Baerlorcher, W.M. Meier, D.H. Olson, Atlas of Zeolite Framework Types, fth
ed., Elsevier, Amsterdam, 2001.
[18] J.S. Buchanan, J.G. Santiesteban, W.O. Haag, J. Catal. 158 (1996) 279287.
[19] G.L. Zhao, J.W. Teng, Z. Xie, W.M. Yang, Q.L. Chen, Y. Tang, Stud. Surf. Sci. Catal.
170 (2007) 13071312.
[20] D. Tzoulaki, A. Jentys, J. Prez-Ramrez, K. Egeblad, J.A. Lercher, Catal. Today
198 (2012) 311.
[21] L.H. Ong, M. Dmk, R. Olindo, A.C. van Veen, J.A. Lercher, Microporous
Mesoporous Mater. 164 (2012) 920.
[22] H. Konno, T. Okamura, T. Kawahara, Y. Nakasaka, T. Tago, T. Masuda, Chem.
Eng. J. 207208 (2012) 490496.
[23] Z. Wang, G. Jiang, Z. Zhao, X. Feng, A. Duan, J. Liu, C. Xu, J. Gao, Energy Fuels 24
(2009) 758763.
[24] X. Zhu, S. Liu, Y. Song, L. Xu, Catal. Lett. 103 (2005) 201210.
[25] Y. Jin, S. Asaoka, S. Zhang, P. Li, S. Zhao, Fuel Process. Technol. 115 (2013) 34
41.
[26] E. Epelde, A.G. Gayubo, M. Olazar, J. Bilbao, A.T. Aguayo, Chem. Eng. J., in press,
http://dx.doi.org/10.1016/j.cej.2014.04.060.
[27] N. Xue, R. Olindo, J.A. Lercher, J. Phys. Chem. C 114 (2010) 15763
15770.
[28] Y. Jiao, C. Jiang, Z. Yang, J. Zhang, Microporous Mesoporous Mater. 162 (2012)
152158.
[29] S.M.T. Almutairi, B. Mezari, E.A. Pidko, P.C.M.M. Magusin, E.J.M. Hensen, J.
Catal. 307 (2013) 194203.
[30] Q. Yu, X. Meng, J. Liu, C. Li, Q. Cui, Microporous Mesoporous Mater. 181 (2013)
192200.
[31] T. Reddy Keshav, S. Basu, Fuel Process. Technol. 88 (2007) 493500.
[32] D.S. Wragg, D. Akporiaye, H. Fjellvag, J. Catal. 279 (2011) 397402.
[33] W. Dai, M. Scheibe, L. Li, N. Guan, M. Hunger, J. Phys. Chem. C 116 (2011)
24692476.
[34] Q. Qian, J. Ruiz-Martnez, M. Mokhtar, A.M. Asiri, S.A. Al-Thabaiti, S.N. Basahel,
B.M. Weckhuysen, Catal. Today 226 (2014) 1424.
[35] S. Askari, R. Halladj, M. Sohrabi, Microporous Mesoporous Mater. 163 (2012)
334342.
[36] S. Tian, S. Ji, D. L, B. Bai, Q. Sun, J. Energy Chem. 22 (2013) 605609.
[37] S.G. Lee, H.S. Kim, Y.H. Kim, E.J. Kang, D.H. Lee, C.S. Park, J. Ind. Eng. Chem. 20
(2014) 6167.
[38] F. Schmidt, S. Paasch, E. Brunner, S. Kaskel, Microporous Mesoporous Mater.
164 (2012) 214221.

293

[39] K. Qian, D.C. Tomczak, E.F. Rakiewicz, R.H. Harding, G. Yaluris, W.C. Cheng, X.
Zhao, A.W. Peters, Energy Fuels 11 (1997) 596600.
[40] H.S. Cerqueira, C. Sievers, G. Joly, P. Magnoux, J.A. Lercher, Ind. Eng. Chem. Res.
44 (2005) 20692077.
[41] M. Guisnet, L. Costa, F. Rama, J. Mol. Catal. A: Chem. 305 (2009) 6983.
[42] H. Hu, F. Cao, W. Ying, Q. Sun, D. Fang, Chem. Eng. J. 160 (2010) 770778.
[43] H. Schulz, Catal. Today 154 (2010) 183194.
[44] A.T. Aguayo, P. Castao, D. Mier, A.G. Gayubo, M. Olazar, J. Bilbao, Ind. Eng.
Chem. Res. 50 (2011) 99809988.
[45] P. Castao, G. Elordi, M. Olazar, A.T. Aguayo, B. Pawelec, J. Bilbao, Appl. Catal. B:
Environ. 104 (2011) 91100.
[46] M. Ibaez, B. Valle, J. Bilbao, A.G. Gayubo, P. Castao, Catal. Today 195 (2012)
106113.
[47] B. Valle, P. Castao, M. Olazar, J. Bilbao, A.G. Gayubo, J. Catal. 285 (2012) 304
314.
[48] M. Ibez, M. Artetxe, G. Lopez, G. Elordi, J. Bilbao, M. Olazar, P. Castao, Appl.
Catal. B: Environ. 148149 (2014) 436445.
[49] B.M. Lok, C.A. Messina, R.L. Patton, R.T. Gajek, T.R. Cannan, E.M. Flanigen, J. Am.
Chem. Soc. 106 (1984) 60926093.
[50] A.G. Gayubo, A. Alonso, B. Valle, A.T. Aguayo, M. Olazar, J. Bilbao, Fuel 89 (2010)
33653372.
[51] A.T. Aguayo, A.G. Gayubo, R. Vivanco, M. Olazar, J. Bilbao, Appl. Catal. A: Gen.
283 (2005) 197207.
[52] C.A. Emeis, J. Catal. 141 (1993) 347354.
[53] R.J. Quann, L.A. Green, S.A. Tabak, F.J. Krambeck, Ind. Eng. Chem. Res. 27 (1988)
565570.
[54] H. Oikawa, Y. Shibata, K. Inazu, Y. Iwase, K. Murai, S. Hyodo, G. Kobayashi, T.
Baba, Appl. Catal. A: Gen. 312 (2006) 181185.
[55] X. Zhu, S. Liu, Y. Song, L. Xu, Appl. Catal. A: Gen. 288 (2005) 134142.
[56] X. Tang, H. Zhou, W. Qian, D. Wang, Y. Jin, F. Wei, Catal. Lett. 125 (2008) 380
385.
[57] M. Guisnet, P. Magnoux, Appl. Catal. A: Gen. 212 (2001) 8396.
[58] F. Bauer, H.G. Karge, Mol. Sieves 5 (2007) 249364.
[59] B. Valle, A. Alonso, A. Atutxa, A.G. Gayubo, J. Bilbao, Catal. Today 106 (2005)
118122.
[60] H.G. Karge, W. Nieen, H. Bludau, Appl. Catal. A: Gen. 146 (1996) 339349.
[61] J. Robertson, Mater. Sci. Eng. R 37 (2002) 129281.
[62] A. Marcilla, A. Gomez-Siurana, F.J. Valdes, Appl. Catal. A: Gen. 334 (2008) 20
25.
[63] B. Wang, G. Manos, Ind. Eng. Chem. Res. 47 (2008) 29482955.
[64] J. Schwan, S. Ulrich, V. Batori, H. Ehrhardt, S.R.P. Silva, J. Appl. Phys. 80 (1996)
440447.
[65] A.C. Ferrari, J. Robertson, Phys. Rev. B 61 (2000) 1409514107.
[66] A.C. Ferrari, A. Libassi, B.K. Tanner, V. Stolojan, J. Yuan, L.M. Brown, S.E. Rodil, B.
Kleinsorge, J. Robertson, Phys. Rev. B 62 (2000) 11089.
[67] J.L. Bonardet, M.C. Barrage, J. Fraissard, J. Mol. Catal. A: Chem. 96 (1995) 123
143.
[68] Y. Jiang, J. Huang, V.R. Reddy, Microporous Mesoporous Mater. 105 (2007)
132139.
[69] G. Seo, J.H. Kim, H.G. Jang, Catal. Surv. Asia 17 (2013) 103118.
[70] M. Zokaie, D.S. Wragg, A. Gronvold, T. Fuglerud, J.H. Cavka, K.P. Lillerud, O.
Swang, Microporous Mesoporous Mater. 165 (2013) 15.
[71] D. Chen, K. Moljord, A. Holmen, Microporous Mesoporous Mater. 164 (2012)
239250.