Deoxygenation of methylesters over CsNaX

Article abstract of DOI:10.1016/j.jcat.2008.06.012

The deoxygenation of methyl octanoate over a CsNaX zeolite catalyst has been investigated as a model reaction for the production of de-oxygenated liquid hydrocarbons from biodiesel. Several operating parameters were investigated, such as the type of basic catalyst used, the co-reactant incorporated in the reactor as a solvent of the liquid feed, and the reaction temperature. The CsNaX zeolite used in the study was prepared by ion exchange of NaX with CsNO₃/CsOH solution. A significant role of the solvent (co-reactant) was found on the activity, selectivity, and stability of the catalyst. That is, when methanol was co-fed enhanced stability and decarbonylation activity were observed. By contrast, when nonane was used, the catalyst deactivated rapidly and the selectivity to coupling products was enhanced. Temperature programmed desorption (TPD) of methyl octanoate and methanol, as well as flow catalytic studies suggest that methyl octanoate first decomposes to an octanoate-like species. The decomposition of such species leads to the formation of heptenes and hexenes as major products. Octenes and other hydrogenated products are formed in lower amounts via hydrogenation by hydrogen produced on the surface by methanol decomposition, but not from gas phase H₂, followed by dehydration. When the polarizable Cs cation is not present in the catalyst, reduced activity and formation of undesired products, such as aromatics and pentadecanone, occur. Similarly, non-zeolitic basic catalysts, such as MgO, exhibit low activity and low selectivity to de-oxygenated liquid hydrocarbons.

Full text of DOI:10.1016/j.jcat.2008.06.012

Journal of Catalysis 258 (2008) 199–209
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
Deoxygenation of methylesters over CsNaX
∗
Tawan Sooknoi 1, Tanate Danuthai 2, Lance L. Lobban, Richard G. Mallinson, Daniel E. Resasco
School of Chemical, Biological and Materials Engineering, University of Oklahoma, Norman, OK 73019, USA
a r t i c l e i n f o
a b s t r a c t
Article history:
The deoxygenation of methyl octanoate over a CsNaX zeolite catalyst has been investigated as a model
reaction for the production of de-oxygenated liquid hydrocarbons from biodiesel. Several operating
parameters were investigated, such as the type of basic catalyst used, the co-reactant incorporated in
the reactor as a solvent of the liquid feed, and the reaction temperature. The CsNaX zeolite used in
the study was prepared by ion exchange of NaX with CsNO3/CsOH solution. A significant role of the
solvent (co-reactant) was found on the activity, selectivity, and stability of the catalyst. That is, when
methanol was co-fed enhanced stability and decarbonylation activity were observed. By contrast, when
nonane was used, the catalyst deactivated rapidly and the selectivity to coupling products was enhanced.
Temperature programmed desorption (TPD) of methyl octanoate and methanol, as well as flow catalytic
studies suggest that methyl octanoate first decomposes to an octanoate-like species. The decomposition
of such species leads to the formation of heptenes and hexenes as major products. Octenes and other
hydrogenated products are formed in lower amounts via hydrogenation by hydrogen produced on the
surface by methanol decomposition, but not from gas phase H2, followed by dehydration. When the
polarizable Cs cation is not present in the catalyst, reduced activity and formation of undesired products,
such as aromatics and pentadecanone, occur. Similarly, non-zeolitic basic catalysts, such as MgO, exhibit
low activity and low selectivity to de-oxygenated liquid hydrocarbons.
Received 15 April 2008
Revised 4 June 2008
Accepted 8 June 2008
Available online 18 July 2008
Keywords:
Basic zeolites
CsNaX
Biofuels
Methylesters
Deoxygenation
Decarbonylation
Deacetalation
Hydrogenation
TPD
TPRx
© 2008 Elsevier Inc. All rights reserved.
1. Introduction
may contain less carbon than its fatty acid counterpart, the for-
mer reactions require much less hydrogen than the latter [8].
Moreover, the properties of the fuel obtained by decarbonyla-
The interest in renewable fuels as replacements for fossil fuels
has rapidly increased over the past few years. Among the vari-
ous fuels-from-biomass investigated, fatty acid methyl-ester bio-
fuels (FAME) obtained by trans-esterification of triglycerides from
natural oils and fats with methanol have received considerable at-
tention [1–4]. They exhibit high cetane number and are considered
to burn cleanly; however, there is growing concern about the fun-
gibility of these fuels with conventional petroleum-derived diesel
due to the oxidative and thermal instability [5–7]. Reduction of the
oxygen content in the fuel would readily improve the stability of
the fuel and therefore its utilization potential. Various processes
including hydrogenolysis [8,9], decarbonylation [10,11], and decar-
boxylation [12,13] of FAME have been proposed to transform the
biodiesel into the hydrocarbon base fuel.
tion/decarboxylation are not significantly different from those ob-
tained from hydrogenolysis [12]. Typically, decarbonylation takes
place over supported noble metal catalysts at relatively high tem-
◦
perature (>350 C) [14]. For example, Pd/C has been found to be an
effective catalyst for decarbonylation/decarboxylation [12]. How-
ever, CO produced from the reaction may competitively adsorb
on the metal surface, leading to a reduction or even loss in the
catalytic activity as conversion increases. Hence, the required to-
tal and hydrogen partial pressures are generally high, in order to
keep the surface clean and facilitate the oxygenate–metal surface
interaction, while reducing catalyst deactivation [15].
An alternative family of catalysts that may overcome some of
these limitations is that of solid bases. This is because the FAME is
fairly electrophilic while the decarbonylated by-product, CO, is nu-
cleophilic. Thus, while the adsorption of FAME over basic catalyst
may be strong and the reaction can be readily promoted at atmo-
spheric pressure, the by-product CO should promptly leave the ba-
sic surface. As a result, considerably lower temperature and hydro-
gen partial pressure may be required. Among possible solid base
catalysts, low-silica zeolites containing highly polarizable cations,
such as cesium, may be good candidates. These catalysts have been
found to exhibit relatively high activity towards reactions involving
oxygenates [16,17], as well as hydrogenation and hydrogenolysis of
The decarbonylation and decarboxylation have advantages over
hydrogenolysis because, while the hydrocarbons thus produced
Corresponding author. Fax: +1 (405) 325 5813.
E-mail address: resasco@ou.edu (D.E. Resasco).
Permanent address: Department of Chemistry, Faculty of Science, King
*
1
Mongkut’s Institute of Technology, Ladkrabang, Bangkok 10520, Thailand.
2
Permanent address: Petroleum and Petrochemical College, Chulalongkorn Uni-
versity, Patumwan, Bangkok 10330, Thailand.
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.06.012

200
T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
◦
acrylonitrile and propionitrile [18]. It has been suggested that, in
addition to the need for strong basicity, an active catalyst requires
the co-existence of acid and basic sites for reactions such as hy-
drogenation and hydrogenolysis [19]. In fact, an alkali-exchanged
zeolite such as the CsNaX contains conjugate acid–base pairs, in
which the exchangeable alkali cation has Lewis acidity and the
oxygen that bridges Si and Al near the exchangeable cation has
basicity [20].
ple was initially pretreated in a flow of 2% O2/He for 2 h at 450 C.
◦
Then, the sample was cooled in He flow to 150 C. A 100 μl pulse
of methyl octanoate was injected to the sample, and a He flow
was introduced for 3 h at 150 C in order to remove the excess
methyl octanoate. The sample was then heated to 900 C at a rate
of 10 C/min. Masses (m/z) of 2, 28, 56, and 74 were monitored to
determine the evolution of hydrogen, carbon monoxide, hydrocar-
bons, and methyl octanoate, respectively.
◦
◦
◦
The purpose of this work is to investigate deoxygenation reac-
tions on Cs and NaX zeolite catalysts using methyl octanoate as a
model feed that may mimic some of the most relevant reactions
involved in the refining of biodiesels. The effects of varying re-
action conditions such as temperature, hydrogen partial pressure,
feed concentration, presence of competing molecules, as well as
catalyst formulations were investigated in a flow reactor. In addi-
tion, temperature programmed techniques were employed to elu-
cidate possible reaction pathways thought to be relevant in the
deoxygenation of FAME.
2.4. Temperature-programmed desorption (TPD) and temperature
programmed reaction (TPRx) of methanol
Methanol was chosen as a probe for the zeolite surface and as a
co-reactant to help reduce the rate of catalyst deactivation during
the conversion of methyl octanoate. Therefore, the desorption and
decomposition of methanol over CsNaX and NaX catalysts were
investigated by TPD and TPRx techniques in the same apparatus
described in the previous section. In each run, 50 mg of the sam-
2. Experimental
◦
ple was initially pretreated in a flow of 2% O2/He for 2 h at 450 C.
Then, the sample was cooled in He flow to room temperature.
In the desorption experiment, 10 μl of methanol was repeatedly
injected over the sample at room temperature until reaching sat-
uration, as indicated by a steady MS signal. After removal of the
excess methanol by flowing He for 3 h at room temperature, the
2.1. Catalyst preparation and characterization
Commercial molecular sieve (UOP type 13X, NaX) and MgO
(LOT No. 130337) were obtained from Fluka and Schweizerhall,
respectively. The as-received materials were calcined at 450 and
◦
◦
◦
sample was heated to 900 C at a rate of 10 C/min. For the TPRx
experiment, methanol was continuously fed through the sample
bed at a flow rate of 1 ml/h. The sample was linearly heated from
700 C, respectively. The Cs-containing zeolite (CsNaX) was pre-
pared by ion exchange of molecular sieve 13X with 0.1 M CsNO3/
◦
CsOH (4/1 v/v) at 80 C for 24 h. The solid material was filtered
◦
◦
◦
room temperature to 700 C at a heating rate of 10 C/min. Masses
(m/z) of 2, 28, 31, and 45 were monitored to determine the evolu-
tion of hydrogen, carbon monoxide, methanol, and dimethylether,
respectively.
and left to dry at 80 C overnight. The sample was then calcined at
◦
450 C for 2 h in a flow of dry air. It has been demonstrated that
crystallinity in CsNaX samples is largely retained after this treat-
ment [21]. To investigate the effect of excess Cs in the zeolite, the
CsNaX sample was washed repeatedly with water, a treatment that
eliminated a significant fraction of Cs.
Surface areas of the CsNaX, before and after washing, and NaX
catalysts were determined by nitrogen adsorption at 77 K using a
Micromeritics ASAP 2000 apparatus. X-ray diffraction (XRD, Bruker
AXS D8Discover) was employed to investigate the catalyst struc-
ture. The hydrogen uptake of the catalysts was measured by tem-
perature programmed reduction (TPR) in an apparatus constructed
for this purpose.
2.5. Catalytic activity measurements for the deoxygenation of methyl
octanoate
The deoxygenation of methyloctanoate was carried out at at-
ꢀꢀ
mospheric pressure, in a fixed bed flow reactor made with 1/4
quartz tube. Prior to the reaction, the catalyst samples were pre-
◦
treated in situ under flow of 2% O2/He at 450 C for 2 h. Then, the
◦
catalyst bed was cooled to the reaction temperature (425 C) un-
2.2. Temperature-programmed desorption (TPD) of isopropylamine
der flow of He (25 ml/min). The reactant feed of 10 wt% methyl
octanoate in methanol solvent was introduced into the reactor us-
ing a syringe pump via a heated vaporization port. Alternatively, in
some runs, nonane was used as a solvent instead of methanol. Un-
reacted feed and products were quantified using an online GC-FID
equipped with a capillary HP-5 column, following a temperature
program to optimize product separation.
The catalyst acidity was tested by the amine TPD technique.
First, 50 mg of sample was pretreated at 450 C in a flow of 2%
◦
O2/He for 1 h. After the pretreatment, the sample was cooled in
He to room temperature and then consecutive 10 μl pulses of iso-
propylamine (IPA) were injected to the sample until saturation was
achieved, as indicated by a constant m/z = 44 signal in the MS.
After removing the excess IPA by flowing He for 3 h, the sam-
◦
◦
ple was linearly heated to 700 C at a heating rate of 10 C/min.
Masses (m/z) of 44, 41, and 17 were monitored to determine the
evolution of desorbing species and MS fragments. The signal 44 is
only due to IPA; by contrast, 41 and 17 are due to propylene and
ammonia, respectively, but also to MS fragments of IPA. To deter-
mine whether there is desorption or decomposition, one follows
the evolution of the three signals simultaneously. In this case, no
decomposition was observed and the three signals kept the same
ratio throughout the desorption. The amount of desorbed IPA was
calibrated with 50 μl pulses of 2% propylene in He (m/z = 41).
2.6. Temperature-programmed oxidation (TPO) of coke deposits after
reaction
Temperature-programmed oxidation (TPO) was employed to in-
vestigate and quantify the nature and amount of coke deposited on
ꢀꢀ
the spent catalysts. 30 mg of the sample was packed in the 1/4
◦
quartz tube reactor. The temperature was ramped to 900 C with a
heating rate of 10 C/min and the TPO profiles were recorded un-
◦
der flow of 2% O2/He. The CO2 produced by the oxidation of the
coke deposits was further converted to methane over a methana-
◦
tion catalyst (15% Ni/Al2O3) in the presence of hydrogen at 400 C.
2.3. Temperature-programmed desorption (TPD) of methyl octanoate
The corresponding methane was then analyzed online by an FID
detector. The amount of oxidized coke was calibrated using 100 μl
pulses of pure CO2 injected to the same system.
The evolution of pre-adsorbed methyl octanoate over CsNaX
catalysts was followed by TPD. In each run, 50 mg of the sam-

T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
201
Table 1
Elemental analysis data of NaX and CsNaX catalysts
Catalysts
Surface area
(m2/g)
Total composition
Si/Al
Cs/Al
NaX
Washed CsNaX
CsNaX
745
506
468
1.2
1.2
1.2
–
0.47
0.62
Fig. 2. TPR of CsNaX (before and after washing) and NaX catalyst; all of them treated
◦
in air at 450 C before starting the TPR.
Fig. 1. Evolution of m/z = 44 during TPD of isopropylamine.
3. Results
3.1. Characterization of the fresh catalysts
The NaX and CsNaX catalysts were characterized by XRD and
BET surface area analysis. The XRD pattern of CsNaX catalyst
showed almost the same pattern as that of NaX, which confirms
that the CsNaX catalyst contains practically the same crystalline
phase after the cesium exchange. The atomic bulk (ICP) compo-
sitions of NaX and CsNaX catalysts are presented in Table 1. The
analysis indicates that a significant fraction of the initial Na ions
have been exchanged by Cs. This significant amount of ion ex-
change plus some excess Cs occluded in the zeolite resulted in a
noticeable drop in BET surface area and pore volume, as compared
to those of the original NaX. This is because the Cs cation is much
larger than the Na cation and the exchange leads to a reduction of
the available pore volume and surface area. This observation is in
good agreement with previous reports [22,23]. However, it must be
noted that, as the CsNaX sample was dried without washing dur-
ing the preparation, excess Cs cations were left on the sample as
bulk particles of cesium oxide, cesium carbonate, cesium peroxide,
or cesium superoxide. Therefore, a fraction of the observed drop
in BET may be due to pore blockage by these species. The wash-
ing treatment eliminated part of this Cs and resulted in a partial
recovery of the BET area.
Fig. 3. TPD profiles of adsorbed methanol over CsNaX and NaX catalysts.
is consistent with the presence of some residual acidity on this
catalyst.
The hydrogen consumption of the two fresh zeolites catalysts
was quantified by the temperature programmed reduction (TPR)
technique. As shown in Fig. 2, a significant amount of hydrogen
◦
consumption can be observed at around 550 C over the CsNaX
zeolites whilst no hydrogen was consumed by the NaX sample.
Relatively less hydrogen consumption can be observed as succes-
sive washing steps were applied to the CsNaX sample. Therefore, it
is clear that the high hydrogen consumption observed on the un-
washed CsNaX is due to the reduction of the excess cesium species
[24,25]. The Cs bulk species (oxides or carbonates) may be formed
from excess Cs during the activation in air. The hydrogen consump-
tion increases with the amount of excess cesium species in the
sample.
The TPD of isopropylamine suggested that alkali-exchanged ze-
olite possesses some acidity. As shown in Fig. 1, the evolution of
◦
adsorbed IPA at lower temperature (∼180 C) corresponds to the
weak adsorption of IPA on the zeolite framework, while that at
higher temperature may represent acid sites, but not the strong
Brønsted sites typically found in HY or HZSM-5. In those cases, the
IPA does not desorb at high temperatures as a whole molecule, but
rather it decomposes into propylene and ammonia. In the present
zeolites, there is no decomposition, but rather desorption. Quan-
tification of the desorbed amount resulted in about 2.9 μmol/gcat
for the weakly adsorbed IPA on CsNaX catalyst, but no strongly ad-
sorbed species. By contrast, on the NaX zeolite, about 4.6 μmol/gcat
were weakly adsorbed and 3.6 μmol/gcat strongly adsorbed, which
3.2. Temperature-programmed desorption (TPD) and reaction (TPRx) of
methanol
The evolution profiles obtained from TPD of adsorbed methanol
are shown in Fig. 3. A significantly higher amount of undissoci-
ated methanol is desorbed from the NaX catalyst than from CsNaX.
Moreover, while the NaX zeolite exhibited two desorption peaks at
◦
temperatures around 150 and 300 C, the CsNaX zeolite only dis-
played the lower-temperature peak. Similar differences in the TPD

202
T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
Fig. 4. (a) Temperature-programmed reaction of continuous flow methanol over CsNaX. (b) H2/CO signal of continuous flow methanol over CsNaX.
profiles of adsorbed methanol on NaX and CsNaX have been pre-
viously observed [26]. These differences were suggested to arise
from a stronger interaction of the adsorbate with the Na cations
in NaX, as compared to those in CsNaX. However, other adsorp-
tion studies [27,28] would indicate that one could also assign the
lower temperature TPD peak to methanol adsorption on sites lo-
cated in the supercages of the NaX and the higher temperature
peak to sites located in the smaller sodalite cages. The difference in
binding strength in the two sites can be explained in terms of the
shorter-range interactions taking place in the small cavities of the
sodalite cages. Accordingly, the disappearance of the high tempera-
ture peak from the TPD of the CsNaX indicates that the presence of
the much larger Cs cations in the small sodalite cage dramatically
reduces its accessible pore volume, thus preventing the adsorption
of methanol on these sites. Based on this description, adsorption
on the CsNaX zeolite would occur only in the larger supercages,
but even this peak is smaller than the corresponding one on NaX,
due to the larger extent of blocking by Cs.
To investigate the temperature range that is more relevant to
the reaction studies, the decomposition of methanol over CsNaX
was examined by temperature programmed reaction (TPRx) under
continuous flow of methanol. As shown in Fig. 4a, the decomposi-
Fig. 5. TPD profile of adsorbed methyl octanoate on CsNaX zeolite catalyst.
◦
tion of methanol starts at 300 C, and in agreement with previous
studies [29], hydrogen (H2), and carbon monoxide (CO) are the
main decomposition products. However, it is interesting to note
that they do not appear at the same time. To make this trend more
clearly visible, the H2/CO ratio is plotted in Fig. 4b. A sharp rise in
◦
this ratio is observed above 300 C, but it goes back to a constant
low value at higher temperatures. This observation suggests that
methanol primarily decomposes into H2 that desorbs and a frag-
ment that remains on the surface. The species retained on the sur-
face has been previously identified by 13C NMR [30] and suggested
to be a formate-like species. As the temperature is increased, this
formate-like species can further decompose evolving CO. Beyond
this temperature the continuous decomposition produces both CO
and H2.
The desorption characteristics of methyl octanoate over CsNaX
catalyst are shown in Fig. 5. Unlike the case of methanol, the TPD
of adsorbed methyl octanoate reveals no evolution of the parent
methyl octanoate molecule, suggesting that this is strongly ad-
sorbed. Instead, decomposition products are observed. CO and H2
Fig. 6. Conversion of 10 wt% methyl octanoate over CsNaX using nonane and
methanol as solvent as a function of time on stream. Reaction conditions: 425 C,
1 atm, W /F = 198 g h/mol, 25 ml/min of He.
◦
◦
were initially evolved above 300 and 400 C, respectively. From
the areas, we have quantified the amount of CO evolved below
450 C to be about 1.6 μmol/gcat. This amount of CO is produced
◦
3.3. Reaction under continuous flow of methyl octanoate
by the decomposition of methyl octanoate at the methoxyl group.
3.3.1. CsNaX catalyst
On the other hand, a larger amount, 6.8 μmol/gcat of CO, is pro-
duced at temperatures higher than 450 C from the decarbonyla-
◦
The effect of co-feeding methanol was investigated during the
decarbonylation of methyl octanoate over the basic CsNaX zeo-
lite catalyst. The conversion obtained by using either nonane or
methanol as a solvent is compared in Fig. 6. It can be seen that
when nonane was used, a high activity was observed at the be-
tion of methyl octanoate. It is important to note that, contrary to
◦
methanol, essentially no hydrogen is evolved up to about 400 C.
A significant amount of hydrogen (4 μmol/gcat) was only produced
◦
at ∼450 C.

T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
203
Fig. 7. Product distribution from the reaction of 10 wt% methyl octanoate using (a) nonane and (b) methanol as solvent as a function of time on stream. Reaction conditions:
◦
425 C, 1 atm, W /F = 198 g h/mol, 25 ml/min of He. Heptenes (a), hexenes ("), octenes (2), pentadecanone (E), coupling ester (Q), octane (1), tetradecene (F),
heptane (e), hexane (!).
Table 2
bonylation of methyl octanoate over CsNaX. In addition, significant
amounts of hexene are obtained. As discussed below, we believe
that hexenes are products of deacetalation. In significantly lower
concentrations, octenes and coupling products, together with small
amounts of octanes and oxygenates, were also observed. As shown
in Figs. 8a and 8b, the yields of heptenes and hexenes increase
parallel to the overall conversion as a function of W /F , suggest-
ing that both are primary products. That is, hexene is not a sec-
ondary product arising from the cracking of heptene. By contrast,
as shown in Fig. 8b, the evolution of octene follows a different pat-
tern; it seems to reach a plateau at an overall conversion near 40%
and the start of this plateau coincides with the increase in the for-
mation of secondary products such as lower oxygenates, octanoic
acid, and tetradecene (see Fig. 8c). On the other hand, the coupling
products exhibit a drop in the concentration as a function of W /F ,
indicating that these heavier compounds tend to react further into
lighter compounds as the catalyst bed increases.
Product distribution from reaction of 10 wt% methyl octanoate in methanol over
CsNaX at various space time
W /F (g h/mol)
99
198
396
Conversion
24.0
37.2
62.3
Yield (%)
Hexenes
n-Hexane
Heptenes
n-Heptane
Octenes
n-Octane
4.5
0.2
8.9
0.1
3.0
0.4
0.6
1.9
0.8
0.3
2.0
1.2
8.8
0.1
14.6
0.2
6.1
0.8
0.5
2.0
1.1
0.3
1.5
1.1
19.3
0.3
26.1
0.8
6.7
1.0
0.4
3.3
2.2
0.6
0.9
0.8
Octanal
Oxygenates
Octanonic acid
Tetradecene
Pentadecanone
Coupling ester
A major consideration in the production of fuels from biomass
is the consumption of hydrogen [31], which is typically needed
in reactions involving deoxygenation steps. We have investigated
the impact of the presence of hydrogen on the overall activity and
product distribution observed in the conversion of 10 wt% methyl
octanoate in methanol over the CsNaX catalyst. Table 3 summa-
rizes the conversion and product distribution obtained on CsNaX
at various reaction conditions.
◦
Reaction conditions: 425 C, 1 atm, 25 ml/min of He.
ginning of the reaction; however, this activity was rapidly lost by
catalyst deactivation. On the other hand, while a relatively lower
activity was observed when methanol was cofed into the reac-
tion system, a much better stability with time on stream was
obtained.
The product distributions from the decarbonylation of 10 wt%
methyl octanoate over CsNaX using nonane and methanol as sol-
vents are compared in Figs. 7a and 7b, respectively. In the reaction
co-feeding nonane, the dominant product was 8-pentadecanone,
a coupling product of methyl octanoate. Deoxygenated products,
namely octene, heptene and hexane were also observed, but with
lower selectivities. When methanol was used as a solvent, very
small amounts of coupling products (i.e. 8-pentadecanone) were
obtained. In this case, the reaction products were mostly hep-
tenes, hexenes, and octenes; that is, the dominant reactions were
decarbonylation (elimination of CO), deacetalation (elimination of
CH3CHO), and hydrogenation/dehydration (elimination of O). It is
clear that while the conversion of methanol itself is rather low
(∼5%), its presence greatly affects the product distribution and sta-
bility of the catalyst.
Interestingly, very similar activity, stability (not shown) and
◦
product selectivity were observed at 425 C when comparing the
reaction under H2 and under He. The main decarbonylation prod-
ucts, C7 and C6 olefins, were dominant under both gases, indicat-
ing little effect of gas-phase hydrogen in this reaction. It is also
interesting to note that the octene yield obtained when using He
was higher than when using H2 as a carrier gas. This result demon-
strates that the decarbonylation over CsNaX does not require gas-
phase hydrogen when methanol is used as a solvent. The presence
of hydrogen also had no effect on the catalyst stability, indicating
that hydrogenation of carbonaceous residues does not take place.
To investigate the effect of reaction temperature, a run was con-
◦
◦
ducted at 400 C, rather than at 425 C. As summarized in Table 3,
◦
the conversion decreased compared to the same reaction at 425 C.
The changes in selectivity are consistent with higher activation
energies for the decarbonylation and deacetalation reactions com-
pared to the rest, particularly the coupling reactions. That is, while
Table 2 summarizes the product distribution from the reaction
of 10 wt% methyl octanoate in methanol at various space times
(W /F ). As discussed above, the major product is C7 hydrocarbons
(i.e. heptenes), which is the expected product from direct decar-
◦
the selectivity to hexene and heptene slightly decreases at 400 C,
the selectivity to the coupling products increased.

204
T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
Table 3
Product distribution from reaction of 10 wt% methyl octanoate over CsNaX using
methanol as solvent at various conditions
Conditions
Carrier gas
Reaction temperature ( C)
He
425
H2
425
He
400
◦
Conversion
Decabonylation/coupling ratio
Olefin/paraffin ratio
37.2
11.4
25.7
34.8
8.3
24.5
10.8
5.0
14.2
Selectivity (%)
Hexenes
n-Hexane
Heptenes
n-Heptane
Octenes
23.7
0.3
39.4
0.5
16.4
2.2
1.3
5.4
3.0
0.8
4.0
3.0
22.3
0.9
40.7
0.9
10.9
1.4
2.0
6.6
3.4
1.1
6.0
3.7
20.2
1.8
35.8
1.8
9.2
0.9
1.8
8.3
3.7
0.9
9.2
6.4
n-Octane
Octanal
Oxygenates
Octanonic acid
Tetradecene
Pentadecanone
Coupling ester
Reaction conditions: W /F = 198 g h/mol.
Fig. 9. Product distribution from the reaction of 10 wt% methyl octanoate in
methanol over NaX zeolite catalyst as a function of time on stream. Reaction con-
◦
ditions: 425 C, 1 atm, W /F = 198 g h/mol, 25 ml/min of He. Hexenes ("), hep-
tenes (a), multisubstituted aromatics (E), heptane (e), octenes (2), hexane (!),
octane (1), and coupling hydrocarbons (F).
Fig. 8. Conversion (a), yields of major products (b), yields of minor products (c) from
the reaction of 10 wt% methyl octanoate in methanol over CsNaX as a function of
◦
W /F . Reaction conditions: 425 C, 1 atm, 25 ml/min of He. Heptenes ("), hex-
enes (!), octenes (a), octane (e), heptane (2), hexane (1), oxygenates (Q), pen-
tadecanone (P), coupling ester (F), octanoic acid (E), octanal ( ), tetradecene (9).
3.3.2. NaX and MgO catalysts
The deoxygenation activity of methyl octanoate over NaX and
MgO catalysts were also studied and compared to that on CsNaX
catalyst. Fig. 9 illustrates the product distribution obtained over
NaX as a function of time on stream, keeping the same condi-
tions as those used over CsNaX (compare with Fig. 7b). In signifi-
cant contrast with the product distribution obtained on CsNaX, it
Fig. 10. Conversion of 10 wt% methyl octanoate in methanal over NaX as a function
of time on stream, compared to that on CsNaX. Reaction conditions: 425 C, 1 atm,
W /F = 198 g h/mol, 25 ml/min of He.
◦

T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
205
◦
Fig. 11. Conversion (a) and product distributions (b) of 10 wt% methyl octanoate in methanol from the reaction over CsNaX and MgO catalysts. Reaction conditions: 425 C,
1 atm, W /F = 198 g h/mol, 25 ml/min of He. (A) Hexenes, (B) hexane, (C) heptenes, (D) heptane, (E) octenes, (F) octane, (G) octanal, (H) oxygenates, (I) octanoic acid,
(J) tetradecene, (K) pentadecanone, (L) coupling ester.
was observed that on this catalyst hexenes, heptenes, and multi-
substituted aromatics were the dominant products at the begin-
ning of the reaction. Then, as the catalyst deactivated, the aromatic
products tended to disappear and only hexenes and heptenes were
obtained as main products at longer time on stream. Moreover, on
this catalyst, methanol conversion to dimethyl ether was observed
to a significant extent, while absent on the Cs-containing catalyst.
Both cyclization and etherification are reactions typically catalyzed
by acids, which indicates that the NaX zeolite used in this study
contains some acidity, in agreement with previous reports [32].
Also, typical of acid catalysts, rapid deactivation is observed with
NaX (see Fig. 10). As shown below, this deactivation is probably
due to coke formation over the acid sites, leading to a more rapid
decrease in the production all the acid catalyzed products, namely
multi-substituted aromatics, small alkanes, and dimethyl ether. It
is interesting that as the acid sites become deactivated, decarbony-
lation can still proceed and deactivates at a lower rate, leading to a
gradual decrease in hexenes and heptenes with time on stream. In
line with this concept, the activity of washed CsNaX falls between
that of CsNaX and that of NaX. Again, the additional (unselective)
activity and more rapid deactivation, as compared to the unwashed
CsNaX, can be related to the small amount of acid sites generated
after the successive washing treatments.
Fig. 12. TPO profiles of spent CsNaX catalysts from the reaction of 10 wt% methyl
using nonane and methanol as solvent.
also be seen that the oxidation of the deposits from these reactions
◦
appeared at a relatively lower temperature (∼280 C), as compared
◦
To compare to a nonzeolitic basic catalyst, the reaction of 10%
methyl octanoate in methanol was investigated under the same
conditions as those used for the zeolite catalysts. The conversion
and product distribution obtained over this catalyst are shown
in Figs. 11a and 11b, respectively. It can be seen that a signifi-
cantly lower conversion was obtained over MgO compared to that
over CsNaX. In addition to differences in basicity, the lower cat-
alytic activity of MgO can be attributed to its lower surface area
(<40 m2/g), as compared with that of the zeolites (>400 m2/g).
Moreover, MgO not only shows differences in the activity level, but
also in selectivity, which is much lower towards the target hydro-
carbon products and higher to undesired condensation products
(such as 8-pentadecanone) compared to the high selectivity to C7–
C8 hydrocarbons displayed by CsNaX.
to that of conventional hydrocarbon coke deposits (>550 C). This
suggests that the deposits on CsNaX are mainly high-molecular-
weight oxygenates rather than the typical polynuclear aromatic
coke.
Fig. 13 shows the TPO profiles of the coke deposited during the
reaction of 10% methyl octanoate in methanol over NaX, compared
to CsNaX. It is clear that on NaX not only a significantly higher
amount of deposits was formed, but also there is an important
fraction of the deposits that decomposes at significantly higher
◦
temperature (350 C), as compared with that in CsNaX. In line with
the concepts discussed above, this additional deposit may be de-
rived from the residues resulting from acid-catalyzed reactions.
While the oxidation temperature of these species is higher than
that observed with the basic CsNaX catalyst, it is much lower than
that typically observed over strong acid catalysts. Hence, these de-
posits should not be referred to as polynuclear aromatic coke, but
rather as oxygenated aromatic compounds formed over the weak
acid sites of NaX.
3.4. Characterization of spent catalysts
The coke residues left on the catalysts after the conversion of
methyl octanoate were quantified by TPO. Fig. 12 shows that when
nonane was used as a solvent, more than twice as much coke was
deposited on CsNaX as when methanol was used. This difference in
coke deposition explains the different deactivation rates observed
when nonane was used instead of methanol as a solvent. It can
4. Discussion
The rapid deactivation of the CsNaX zeolite catalyst observed
during the reaction of methyl octanoate in nonane is presumably

206
T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
formate-like species (13 C NMR signal ∼166 ppm)
Under our reaction conditions, we expect that both surface for-
mate and surface carbonate species will be present. These species
can readily compete with the adsorption of methyl octanoate, thus
lowering its surface coverage and inhibiting the self-condensation
reaction. As a result, this adsorption inhibition suppresses the
formation of high-molecular weight oxygenates, resulting in bet-
ter catalytic stability. Also, consistent with the above discussion,
a lower initial activity can be expected for the reaction in the pres-
ence of methanol, as experimentally observed. Since there is no
significant difference in activity and product selectivity when hy-
drogen is used as carrier gas, we may suggest that the hydrogen,
produced in situ from the methanol decomposition, can further
facilitate the hydrogenation of methyl octanoate and its primary
products. Hence, various olefins and alkanes, such as octene, hex-
ane and heptane were obtained from the reaction using methanol
as a solvent.
Fig. 13. TPO profiles of spent CsNaX and NaX catalysts from the reaction of 10 wt%
methyl octanoate in methanol.
due to the adsorption of methyl octanoate on the highly basic
sites of the zeolite that may not be greatly affected by the pres-
ence of the hydrocarbon solvent. Such strong adsorption may lead
to a rapid deactivation by site (and probably pore) blockage. Un-
der these conditions, the adsorbed methyl octanoate can react
with each other (self-condensation) forming the observed high-
molecular-weight oxygenates (e.g. 8-pentadecanone). Such com-
pounds cannot readily diffuse out from the pores of the zeolite and
lead to further blockage. The formation of high-molecular weight
products is usually derived from (i) condensation (mostly aldol
type) of the carbanion intermediates with available electrophilic
species [33], or (ii) self-condensation [34]. It must be noted that
these coupling products do not necessarily form a hard coke, since,
as the TPO results show, they can be removed at relatively low ox-
As suggested by TPD results (Fig. 5), the methyl octanoate
is strongly adsorbed on basic sites of CsNaX. Due to the highly
polarizable nature of methyl octanoate, the interaction between
carbonyl ester and the basic framework oxygen would readily
weaken the carboxylate C–O bond, leading to the decomposi-
tion of the ester into two surface-aldehydes, as generally ob-
served over typical basic catalysts [37]. In this case, the de-
composed methoxyl group would form the formate-like species
(surface-formaldehyde) while the other fragment would form an
“octanoate-like species” (surface-octanaldehyde), in parallel with
the formation of the formate-like species discussed above. This
formate-like species would then decompose into carbon monox-
ide and hydrogen, in a manner similar to that observed with
methanol. This primary decomposition of the methoxyl group is
evidenced in the present study by the observed CO evolution prior
◦
idation temperatures (i.e. below 300 C).
By contrast, the use of methanol as a solvent generates a com-
pletely different situation. First, the catalytic stability of CsNaX is
greatly enhanced. Methanol being a highly polar molecule, it can
competitively interact with the zeolite basic sites [18]. Several re-
ports [35,36] have suggested that adsorption of methanol derives
mainly from interactions between the methoxyl oxygen and the
exchangeable cation and interactions between the hydroxyl hydro-
gen and the framework oxygen:
◦
to the formation of hydrocarbon (∼300–400 C). The fact that the
amount of CO evolved at this temperature range (∼1.6 μmol/g) is
relatively small compared to that evolved at higher temperatures
(∼6.8 μmol/g), suggests that the decomposition of the formate-like
species was not the dominant path, but the species was retained
on the surface up to higher temperatures.
+
+
When M is Cs , a highly polarizable cation, the interaction
+
between M –O is relatively weaker than that with higher Lewis
acid Na . However, over CsNaX, the framework O–H interaction
+
becomes stronger and hence this weakens the O–H of methanol.
As shown in the TPD and TPRx studies, this interaction leads to
the decomposition of methanol to yield primarily a surface CH2O
and hydrogen. The surface CH2O would then form a formate-
like species, which can be further converted into a carbonate-like
species as the temperature increases, as evidenced by NMR signals
at 166 and 171 ppm, respectively [30].

T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
207
“octanoate-like species”
Chart 1
It is proposed that decomposition of such “octanoate-like
species” leads to the formation of the two major hydrocarbon
products observed on the CsNaX catalyst, heptene and hexene (see
Chart 1). In the first case, the well-known β-hydrogen elimina-
tion [38] and subsequent decomposition would result in heptene
as major product of the reaction, with parallel evolution of CO and
H2. This reaction may well be referred to as a reversible reaction of
the typical hydroformylation [39]. In the second case, the decom-
position of a cyclic-like intermediate can lead to the formation of
more electrophilic acetaldehyde (enol-form) and hexene, as shown.
The acetaldehyde may be evolved as a by-product or may undergo
further aldol condensation/alkylation to form C3–C5 oxygenates, as
observed in small amounts by GC-MS.
It must be noted that while hydrogen is not required for the
production of heptenes and hexenes it is necessary for the forma-
tion of octenes and alkanes via hydrogenation/dehydration. These
reactions can indeed occur on cesium-exchanged zeolites [40]. As
discussed above, due to the highly polarizable cesium cations,
the basic sites associated with it can readily decompose the
methanol/methoxyl group, creating a hydrogen surface fugacity
well above that in equilibrium with hydrogen in the gas phase
[41,42]. With such high virtual pressure of hydrogen, the adsorbed
“octanoate-like species” may well be hydrogenated forming pri-
marily an octanol that then rapidly undergoes dehydration to form
octenes:
The small amounts of octane, heptane and hexane in the prod-
uct give evidence for the hydrogenation activity of CsNaX (see
Table 2). However, the extent of this reaction is small in compar-
ison to that of the octanoate-like species decomposition. This is
because olefins are less adsorptive on the basic sites, as compared
to the more electrophilic octanoate-like species. These hydrogenol-
ysed/hydrogenated products can be particularly promoted in the
reaction with methanol, but not readily with nonane (Fig. 7).
The observed small amounts of 8-pentadecanone and long-
chain olefins suggest another reaction path on CsNaX. We believe
that 8-pentadecanone is not derived from the “octanoate-like” in-
termediate responsible for the production of heptenes and hexene.
Instead, this large product is likely to arise from the decomposition
of an acid anhydride. The adsorbed methyl octanoate may undergo
a condensation reaction, forming dioctanoic acid anhydride. The
acid anhydride may undergo a decarboxylation to ketone [43] on
the basic sites since the highly negative framework charge of oxy-
gen would act as an acceptor for the acidic CO2. This path would
explain the high yield of 8-pentadecanone observed in the reaction
when nonane is used as a solvent.
Due to its low mobility, 8-pentadecanone is likely to stay on
surface long enough to be hydrogenated by the adsorbed H to
form alcohol, but such a heavy alcohol would undergo dehydra-
tion, forming a long-chain inner olefin. In fact, as W /F increases,
8-pentadecanone is seen to be readily converted and this is consis-
tent with the observed increase in the yields of long-chain olefins
at longer space times.
The comparison of the catalytic behavior of CsNaX with those of
NaX and MgO indicates that the presence of polarizable Cs cations
result in a catalyst with unique performance. First, on the latter
catalysts, the decarbonylation/deacetalation activity is not as pro-
nounced as over CsNaX. Also, since the Na cation is much harder
than the Cs cation [44] it holds the negative framework charge
more tightly and thus it provides less Lewis basicity to the frame-
work oxygen [45]. That is, the presence of Na as exchangeable
cation, leads to an increase in the (weak) acidity of the Lewis sites
in the zeolite, as evidenced by IPA-TPD. This weak acidity would
explain the formation of aromatic products and the rapid deacti-
vation over NaX. Consistent with this view, a higher amount of
coke deposits was found on NaX, with an important fraction of the
carbonaceous species decomposed at relatively high temperature
(Fig. 13).

208
T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
It should be noted that the yield of hexene is higher than that
5. Conclusions
of heptene over NaX, but the opposite is true for CsNaX. This dif-
ference suggests that in the case of NaX additional hexenes are
produced by cracking of higher molecular weight compounds, not
only from the deacetalation that occurs in CsNaX. The additional
yields due to cracking would also be applied to heptene and other
hydrocarbons formed, and hence these reactions give a higher ini-
tial conversion over this catalyst. It is also interesting to note that,
after the weak acid sites are deactivated in NaX, the production of
heptene and hexene slowly decreases with time on stream (Fig. 9).
As proposed above, these products arise from both cracking over
acid sites and by decarbonylation/deacetalation over basic sites. As
the acid sites are deactivated while the latter are still active, a
gradual decrease only in these two products with time on stream
is observed. Accordingly, since after about 250 min on stream the
acid function has been completely eliminated, the residual activ-
ity can be ascribed to the decarbonylation/deacetalation activity of
NaX.
The suggestion that the alkali-exchanged zeolites contain both
acid and basic functions, is in agreement with earlier studies
[46,47] which indicate that NaX can be regarded as an am-
photeric catalyst, interacting with polar oxygenates in a manner
totally different from that observed over CsNaX. For example,
over NaX, methanol competes with methyl octanoate for adsorp-
tion and consequently DME is largely produced by dehydration
of methanol. Also, on this catalyst, in addition to decarbonyla-
tion, methyl octanoate undergoes acid-catalyzed reactions, such as
cracking, isomerization, alkylation, dimerization, and even aroma-
tization. Hence, multi-substituted aromatics are obtained over NaX
since methanol can readily act as a methylating agent over weakly
acidic sites. When the acid sites are deactivated, the cracking prod-
ucts and aromatics virtually disappear. Only decarbonylation and
deacetalation activity due to negative framework charge of NaX re-
mains. However, this is to a much lower extent than on CsNaX. In
addition, it can be seen that on NaX, without Cs, no hydrogenation
activity can be observed and the use of methanol as a solvent does
not facilitate the maintenance of the desirable reactions.
Finally, by analyzing the behavior observed on the MgO cata-
lyst, one can conclude that not only the basicity is required for
decarbonylation, but also the highly polar environment of the ze-
olite micropore seems to play an essential role in adsorption and
decomposition of the adsorbed species. Although MgO is well re-
garded as a highly basic catalyst, its two-dimensional surface can-
not readily facilitate the adsorption and decomposition of methyl
octanoate. In addition, a much lower surface area, as compared to
that of the zeolites, would show a significant effect on activity.
Hence, on this catalyst only small amounts of deoxygenate prod-
ucts were obtained, as compared to that on CsNaX. By contrast, the
MgO catalyst seems to display enough basicity to activate the cou-
pling reaction of methyl octanoate, thus resulting in the formation
of 8-pentadecanone. As mentioned above, the 8-pentadecanone is
formed via the decomposition of an acid anhydride intermediate.
As the acid anhydride is a “hard” species, it would possess a better
interaction and decomposition to 8-pentadecanone over the “hard”
MgO base.
It can be concluded that the decarbonylation/deacetalation ac-
tivity of methyl octanoate can occur at high rate and for a long
time on stream over CsNaX catalyst when co-feeding methanol.
Methyl octanoate strongly adsorbs on CsNaX basic sites and can-
not be desorbed unless decomposed. When a weak adsorbent as
nonane is co-fed CsNaX rapidly deactivates. By contrast, when
methanol is co-fed with methyl octanoate, the catalyst stability is
greatly enhanced due to the presence of decomposed fragments of
methanol on the surface. These fragments are formate-like species
that prevent self-condensation and formation of higher molecular
weight oxygenates. The TPD results suggest that the decarbony-
lation of methyl octanoate proceeds via primary decomposition
at the methoxyl group, presumably producing an octanoate-like
species as intermediate. The direct decomposition of this species
gives heptenes and hexenes as main products. Octenes and other
hydrogenated products are formed by hydrogenation/dehydration,
in which the surface hydrogen produced from methanol decompo-
sition plays an important role.
When Cs is not present (NaX zeolite) the catalyst basicity is
much lower and the weak acid sites dominate the behavior of
the catalyst. The net results are a decrease in the decarbonyla-
tion/deacetalation activity, together with an increase in the selec-
tivity toward undesired products. The poor performance displayed
by the MgO catalyst indicates that not only the basicity is required
for decarbonylation, but also the highly polar environment charac-
teristic of the zeolite micropore seems to play an essential role.
In summary, the use of the basic characteristics of Cs cation
exchanged in zeolite X and methanol as a co-reactant appears as
an effective combination for the conversion of methyl-esters to hy-
drocarbons, a reaction that may have practical application in the
refining of biodiesel with low hydrogen consumption.
Acknowledgments
This work was financially supported by the Oklahoma Secretary
of Energy and the Oklahoma Bioenergy Center. The authors would
like to thank Mr. Surapas Sitthisa for TPR and TPO analysis. One
of the authors (T.S.) thanks the Fulbright Thai visiting program for
support, and (T.D.) thanks the Thailand Research Fund for a schol-
arship.
References
[1] D.E. Lopez, K. Suwannakarn, D.A. Bruce, J.G. Goodwin Jr., J. Catal. 247 (2007)
43.
[2] I.N. Martyanov, A. Sayari, Appl. Catal. A 339 (2008) 45.
[3] D.A.G. Aranda, R.T.P. Santos, N.C.O. Tanpanes, A.L.D. Ramos, O.A.C. Antunes,
Catal. Lett. 122 (2008) 20.
[4] M.D. Serio, R. Tesser, L. Pengmei, E. Santacesaria, Energy Fuels 22 (2008) 207.
[5] G. Knothe, Fuel Process. Technol. 86 (2005) 1059.
[6] G. Knothe, A.C. Matheaus, T.W. Ryan III, Fuel 82 (2003) 971.
[7] G. Knothe, Fuel Process. Technol. 88 (2007) 669.
[8] W.F. Maier, P. Grubmiller, I. Thies, P.M. Stein, M.A. McKervey, P.R. Schleyer,
Angew. Chem. Int. Ed. 18 (2003) 939.
[9] D.S. Brands, G. U-A-Sai, E.K. Poels, A. Bliek, J. Catal. 186 (1999) 169.
[10] T.A. Foglia, P.A. Barr, J. Am. Oil Chem. Soc. 53 (1976) 737.
[11] I. Kubickova, M. Snare, K. Eranen, P. Maki-Arvela, D.Y. Murzin, Catal. Today 106
(2005) 197.
While MgO can readily decompose methanol to CO and H2 [48],
only small amounts of octene and hydrogenated products can be
found over this catalyst. It is believed that the absence of highly
polarisable cesium cations and the absence of the highly polar en-
vironment in the restricted pores of the zeolites limit the ability
of MgO to adsorb and dissociate hydrogen which results in lower
surface fugacity of hydrogen than in the case of the zeolites. There-
fore, hydrogenation cannot be readily promoted on MgO and yields
of octene, heptane and hexane are relatively small.
[12] P. Maki-Arvela, I. Kubickova, M. Snare, K. Eranen, D.Y. Murzin, Energy Fuels 21
(2007) 30.
[13] M. Snare, I. Kubickova, P. Maki-Arvela, K. Eranen, J. Warna, D.Y. Murzin, Chem.
Eng. J. 134 (2007) 29.
[14] G.N. Rochafilho, D. Brodzki, G. Djega-Mariadassou, Fuel 72 (1993) 543.
[15] M. Snare, I. Kubickova, P. Maki-Arvela, D. Chichova, K. Eranen, D.Y. Murzin,
Fuel 87 (2008) 933.
[16] I. Rodriguez, H. Cambon, D. Brunel, M. Lasperas, J. Mol. Catal. A Chem. 130
(1998) 195.
[17] X.F. Zhang, E.S.M. Lai, R. Martin-Aranda, K.L. Yeung, Appl. Catal. A Gen. 261
(2004) 109.

T. Sooknoi et al. / Journal of Catalysis 258 (2008) 199–209
209
[18] T. Sooknoi, J. Dwyer, J. Mol. Catal. A Chem. 211 (2004) 155.
[33] J. Li, J. Tai, R.J. Davis, Catal. Today 116 (2006) 226.
[19] P. Belirame, P. Fumagallip, G. Zuretti, Ind. Eng. Chem. Res. 32 (1993) 26.
[20] D. Barthomeuf, Catal. Rev. 38 (1996) 521.
[21] M. Lasperas, H. Cambon, D. Brunel, I. Rodriguez, P. Geneste, Microporous
[34] N.D. Plint, N.J. Coville, D. Lack, G.L. Nattrass, T. Vallay, J. Mol. Catal. A Chem. 165
(2001) 275.
[35] D.F. Plant, A. Simperler, R.G. Bell, J. Phys. Chem. B 110 (2006) 6170.
[36] P. Mignon, P. Geerling, R. Schoonheydt, J. Phys. Chem. B 110 (2006) 24947.
[37] H.N. Wright Jr., H.J. Hagemeyer Jr., US Patent 3,714,236 (1973).
[38] N. Takahashi, H. Matsuo, M. Kobayashi, J. Chem. Soc., Faraday Trans. 1 80 (1984)
629.
[39] G. Cuma, P. Famulari, M. Marchetti, B. Sechi, J. Mol. Catal. A Chem. 218 (2004)
211.
[40] T. Sooknoi, J. Dwyer, Stud. Surf. Sci. Catal. 97 (1995) 423.
[41] E. Iglesia, J.E. Baumgartner, G.L. Price, J. Catal. 134 (1992) 549.
[42] M. Boudart, Catal. Lett. 3 (1989) 111.
[43] M. Glinski, W. Szymanski, D. Lomot, Appl. Catal. A 281 (2005) 107.
[44] R.G. Pearson, J. Am. Chem. Soc. 85 (1963) 3533.
[45] Y. Okamato, M. Ogawa, A. Maezawa, T. Imanaka, J. Catal. 112 (1988) 427.
[46] F. Horst, F. Hartmut, G. Ekkehard, H. Bernd, J. Herve, K. Christine, K. Olaf,
K. Knut, Phys. Chem. Chem. Phys. 1 (1999) 593.
Mater. 7 (1996) 61.
[22] M. Lasperas, H. Cambon, D. Brunel, I. Rodriguez, P. Geneste, Microporous
Mater. 1 (1993) 343.
[23] P.E. Hathaway, M.E. Davis, J. Catal. 116 (1989) 263.
[24] Y. Sun, Z. Liu, P. Pianetta, D. Lee, J. Appl. Phys. 102 (2007) 074908.
[25] E.A. Podgornov, I.P. Prosvirin, V.I. Bukhtiyarov, J. Mol. Catal. A Chem. 158 (2008)
337.
[26] R. Schenkel, A. Jentys, S.F. Parker, J.A. Lercher, J. Phys. Chem. B 108 (2004) 7902.
[27] N. Takahashi, A. Mijin, T. Ishikawa, K. Nebuka, H. Suematsu, J. Chem. Soc. Fara-
day Trans. I 83 (1987) 2605.
[28] F. Yagi, H. Tsuji, H. Hattori, Microporous Mater. 9 (1997) 237.
[29] S. Imamura, T. Higashihara, Y. Saito, H. Aritani, H. Kanai, Y. Matsumura, N.
Tsuda, Catal. Today 50 (1999) 369.
[30] M. Hunger, U. Schenk, M. Seiler, J. Weitkamp, J. Mol. Catal. A Chem. 156 (2000)
153.
[47] J.A. Lercher, A. Jentys, A. Brait, Mol. Sieves Sci. Technol. 6 (2008) 153.
[48] D.C. Foyt, J.M. White, J. Catal. 47 (1977) 260.
[31] G.W. Huber, A. Corma, Angew. Chem. Int. Ed. 46 (2007) 7184.
[32] H.B. Schwarz, S. Ernst, J. Karger, B. Knorr, G. Seiffert, R.Q. Snurr, B. Staudte, J.
Weitkamp, J. Catal. 167 (1997) 248.