Catalytic and mechanistic study of lean NO2 reduction by isobutane and propane over HZSM-5

Article abstract of DOI:10.1016/j.molcata.2005.12.033

In this study, the selective catalytic reduction (SCR) of NO₂ by two saturated hydrocarbons, propane and isobutane, has been investigated over acidic zeolite HZSM-5. The activity for NO₂ reduction and the selectivity for N₂ formation correlate well with the intracrystalline acidity (Br?nsted acid sites) of the zeolite. Isobutane is a more efficient reducing agent than propane due to its tertiary hydrogen being more readily activated than the secondary hydrogen in propane. In situ DRIFT step-response experiments with NO₂ and isobutane show formation of surface bound NO⁺, isocyanate, unsaturated hydrocarbons and amine species. The NO⁺ species seems to play an important role in the reduction of NO₂ probably by reacting with alkenes formed from carbenium ion adsorbates. These species likely form isocyanates, amine species and finally nitrogen.

Full text of DOI:10.1016/j.molcata.2005.12.033

Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
Catalytic and mechanistic study of lean NO2 reduction
by isobutane and propane over HZSM-5
Dongmei Zhaoa,c,∗, Hanna Harelind Ingelstena,b, Magnus Skoglundha,b, Anders Palmqvista,b,∗∗
¨
a Competence Centre for Catalysis, Chalmers University of Technology, SE-412 96 Go¨teborg, Sweden
b Applied Surface Chemistry, Chalmers University of Technology, SE-412 96 Go¨teborg, Sweden
c Department of Chemistry, Go¨teborg University, Kemivaegen 10, SE-412 96 Go¨teborg, Sweden
Received 28 July 2005; received in revised form 21 December 2005; accepted 22 December 2005
Available online 3 February 2006
Abstract
In this study, the selective catalytic reduction (SCR) of NO2 by two saturated hydrocarbons, propane and isobutane, has been investigated over
acidic zeolite HZSM-5. The activity for NO2 reduction and the selectivity for N2 formation correlate well with the intracrystalline acidity (Brønsted
acid sites) of the zeolite. Isobutane is a more efficient reducing agent than propane due to its tertiary hydrogen being more readily activated than
the secondary hydrogen in propane. In situ DRIFT step–response experiments with NO2 and isobutane show formation of surface bound NO+,
isocyanate, unsaturated hydrocarbons and amine species. The NO+ species seems to play an important role in the reduction of NO2 probably by
reacting with alkenes formed from carbenium ion adsorbates. These species likely form isocyanates, amine species and finally nitrogen.
© 2006 Elsevier B.V. All rights reserved.
Keywords: HZSM-5; Lean NOx reduction; Propane; Isobutane; FTIR
1. Introduction
lean NOx reduction although at higher temperatures. However,
in general the NOx conversion is low for HC-SCR catalysts and
the active temperature range is narrow. By identifying the key
factors that control the HC-SCR process, both the conversion
and active temperature interval for NOx reduction could how-
ever be improved. The mechanistic proposal that HC-SCR over
zeolite-based catalysts proceeds via intermediates containing
carbon–nitrogen bonds is now widely accepted. This has been
shown by different categories of experiments [7–17], among
which N2 formation has been demonstrated from compounds of
the intermediate types envisaged, under HC-SCR reaction con-
ditions. Such intermediate compounds can be nitrosoalkanes or
nitrocompounds(nitroalkanesandnitroalkenes). Nitrosoalkanes
can be converted to N2 via oximes, nitriles and NH containing
species, over transition metal exchanged zeolites [18–25], and
acidic zeolites [20,23]. Nitroalkanes and nitroalkenes can be
converted to isocyanate, which subsequently can be hydrolyzed
to NH species and further to N2. This has been reported over,
e.g. Co-ZSM-5 [17,26], Cu-ZSM-5 [27] and H-ferrierite [10];
however, the details of the reaction still remain unclear. Forma-
tion of ammonia has been demonstrated for some systems, e.g.
H-ferrierite [10], H-Cu-ZSM-5 [28], H-MOR [29,30] and Pd-
H-MOR [31], in which the NH3-SCR reaction is proposed to be
Today diesel and lean burn gasoline engines are attractive
alternatives to the conventional Otto engine due to lower fuel
consumption and hence lower CO2 emissions. However, an
important disadvantage with these alternatives is that they oper-
ate at high air-to-fuel ratios that give oxygen excess in the
exhaustgas, andthusmakeitdifficulttoreduceNOx (NO + NO2)
formed during the combustion. This calls for efficient techniques
for lean NOx reduction.
Iwamoto et al. [1,2] and Held et al. [3] found early that
Cu-ZSM-5 is an active catalyst for reduction of NO by hydrocar-
bons in oxygen excess. Furthermore, Iwamoto et al. showed the
synergetic effect of oxygen for the selective catalytic reduction
of nitrogen oxides by hydrocarbons (HC-SCR). Hamada et al.
[4–6] observed that a redox system is not essential for this reac-
tion, and that acidic zeolites, as well as alumina are active for
∗
∗∗
Corresponding author. Tel.: +46 31 772 2886; fax: +46 31 772 2853.
Corresponding author. Tel.: +46 31 772 2961; fax: +46 31 160 062.
E-mail addresses: dongmei@chem.gu.se (D. Zhao),
adde@chem.chalmers.se (A. Palmqvist).
1381-1169/$ – see front matter © 2006 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2005.12.033

14
D. Zhao et al. / Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
Table 1
the final step in the reaction pathway. Furthermore, a reaction
between gas phase NOx and nitrogen containing deposits has
been suggested [15,16]. A large number of experiments show C,
N and O, or C and N containing intermediates in the gas phase
during the SCR reaction. Formation of HCN and/or HNCO was
observed during the HC-SCR reaction with C2H4 and C3H6 over
Cu-ZSM-5 [32–35]. Cyanogens were observed over Cu-ZSM-
5 during HC-SCR using isobutane as the reducing agent [28],
and various nitriles were detected using C3 and C4 alkanes over
Fe-ZSM-5 [25]. Furthermore, acrylonitrile was observed over
Cu/H-ZSM-5 [23] using propane, and over H-MOR [30] using
propene. Nitroethene has also been reported using ethene and
NO2 over H-ferrierite [11].
Already in an early stage of the research NO2 was recognized
to have an important role in the first step of HC-SCR reactions
[5]. The HC-SCR activity of acidic zeolites has been found to
increase when NO2 is used as NOx source instead of NO [17].
This characteristic has been used to improve the SCR perfor-
mance by combining two functionally different catalysts, the
first being an NO oxidation catalyst and the second an active
catalyst for the SCR of NO2 [36,37]. The dual pore system cata-
lyst concept is also based on this strategy, but with two different
types of zeolites mixed in one catalyst [38,39].
A possible reaction pathway for lean NO2 reduction by
propane over HZSM-5 has been presented previously [40]. The
key initiating steps are likely the formation of NO+ species
on the catalyst surface and the activation of the hydrocarbon
over Brønsted acid sites, forming carbenium ion adsorbates and
subsequently alkenes. A reaction between the alkenes and the
NO+ species may form oximes (R-CNO), which can easily be
rearranged to isocyanate species (R-NCO) [41]. Hydrolysis of
the isocyanate species with water can thus form amine species,
which may in turn react with the NO+ species, or with NO2 in
the gas phase, forming N2. In the present study we focus on the
selective catalytic reduction of NO2 by propane and isobutane
overaseriesofHZSM-5sampleswithvaryingSiO2/Al2O3 ratio.
The objective is to study the activity and reaction mechanism for
lean reduction of NO2 correlated with the type of hydrocarbon
and the surface acidity.
Textural properties of the HZSM-5 zeolites used in the experiments
HZSM-5(40)
HZSM-5(234)
HZSM-5(569)
Crystallinity (%)
SiO2/Al2O3
Na2O (wt%)
MiPVa (ml/g)
SAb (m2/g)
95
98
234
0.01
0.183
407
100
569
0.03
0.183
377
39.7
<0.05
0.177
405
a
Micropore volume.
BET surface area.
b
filter disc fused into the tube 28 cm from the upper end). The
reactorwasheatedbyatubularfurnace(controlledbyEurotherm
818S) and the temperature was measured using a thermocouple
(K-type) located inside the reactor, 2 mm above the catalyst bed.
This reactor system is also described in detail previously [40].
The catalytic samples were pre-conditioned for 30 min at 500 ◦C
in a gas composed of 10% O2 (99.9996%) in helium (99.9996%)
before exposure to the reactants. The gas flow was controlled
by separate mass flow controllers, and the total flow rate was
maintained constant at 1500 ml (NTP)/min corresponding to a
space velocity of about 50,000 h−1. The inlet feed composition
was 10% O2 (99.9996%), 1040 ppm NO2 (0.983% in He) and
300 ppm propane (2% in He) or 200 ppm isobutane (2% in He),
balanced with helium. The reaction temperature was decreased
step-wise (100 ◦C/step) from 650 to 350 ◦C for C3H8 and from
450 to 250 ◦C for i-C4H10. The products were monitored on-line,
and NOx (i.e. the sum of NO and NO2) and NO were determined
by an Eco Physics, CLD 700 EL ht chemiluminescence analyzer.
Nitrogen dioxide was calculated from the measured NOx and
NO concentrations. An Agilent 6850 gas chromatograph fitted
with a HP-PLOT Molesieve column and a thermal conductivity
detector was used to quantify N2. Water, CO2, NOx and unre-
acted hydrocarbons were removed by a liquid nitrogen cooled
cold trap, located upstream of the gas chromatograph. Carbon
monoxide, CO2 andN2OweremeasuredbyphotoacousticFTIR
spectroscopy, employing a Bruel & Kjaer, Type 1301 analyzer.
Other possible nitrogen containing reaction products, if present,
could be detected by the same technique.
The catalyst performance was studied under steady state con-
ditions at each temperature, which was achieved within around
10 min after changing conditions. At each steady state condi-
tion, the catalysts were exposed to the reactants for about 1 h to
complete the measurement for all the components. The catalysts
were stable during the entire experiment.
2. Experimental
2.1. Catalyst preparation
The catalysts used in this study were HZSM-5 zeo-
lites, received from Albemarle Catalysts BV, with different
SiO2/Al2O3 molar ratios (40, 234 and 569), denoted HZSM-
5(40), HZSM-5(234) and HZSM-5(569), respectively. The tex-
tural properties of the samples are shown in Table 1. The samples
were pressed into discs, and then ground and sieved to obtain a
size fraction between 300 and 500 ␮m.
2.3. In situ DRIFT studies
The in situ Fourier Transform Infrared (FTIR) spectroscopy
measurements were carried out using a BioRad FTS 6000 spec-
trometer in diffuse reflection (DRIFT) mode, equipped with a
continuous flow reactor chamber with KBr windows (Harrick
Scientific Praying Mantis DRIFT cell). The set-up is described
in detail elsewhere [42,43]. About 30 mg of the calcined zeo-
lite sample was placed on a tungsten grid (6.2 mm × 3.2 mm),
which is situated over the outlet of the reaction chamber. The
temperature was measured with a K-type thermocouple. The
2.2. Activity studies
The catalytic activity for NOx reduction and the selectivity
for N2 formation were investigated in a vertical flow reactor
(21 mm i.d., 58 cm long quartz tube with a thin sintered quartz

D. Zhao et al. / Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
15
gases, Ar(99.995%), O2 (99.95%), NO2 (5000 ppminAr), C3H8
(1% in Ar) and isobutane (2% in Ar), were introduced via mass
flow controllers (Bronkhorst Hi-Tech) to the DRIFT cell. The
samples were initially pre-treated in nitrogen dioxide and oxy-
gen (1000 ppm NO2 and 10% O2, balanced with Ar, 500 ◦C,
40 min) followed by oxygen (10% O2 in Ar, 500 ◦C, 1 h) and
then flushed in pure Ar (500 ◦C, 15 min) at a total flow rate of
300 ml (NTP)/min (the flow rate was kept constant throughout
the experiments). Background spectra (40 scans at a resolution
of 1 cm−1) were collected for each temperature, under Ar expo-
sure. Fresh samples were used for all experiments.
In situ reaction experiments were performed for the sam-
ples HZSM-5(40) and HZSM-5(234) with isobutane as reduc-
ing agent. The samples were pre-treated and the background
spectra were recorded as mentioned before. The reaction exper-
iments were performed in three sequences. Sequence 1, feeding
1000 ppm NO2 and 10% O2 in Ar to the DRIFT cell, the co-
adsorption of O2 and NO2 on the catalyst surface was recorded
in the spectra as a function of time (typically 0, 20 s, 1, 2, 3, 5,
7, 10, 15 and 20 min or until saturation). Sequence 2, the reduc-
ing agent, isobutane, was added to the feed; thus, the reaction
NO2 + O2 + isobutane started over the catalyst and IR spectra
were recorded as a function of time. Sequence 3, when the reac-
tion in sequence 2 reached steady state, the supply of reducing
agent was switched off, leaving the sample only exposed to NO2
and O2, and new spectra were recorded as a function of time.
Fig. 1. FTIR spectra of the three different HZSM-5 samples collected at steady
state during Ar flow at 350 ◦C. The measured signals of the baselines are off-set
by −0.04, −0.01 and 0.03 V, respectively, for clarity.
cate that the Brønsted acidic sites most likely are the active
sites for this reaction and that the HC-SCR activity is related
to the amount of acidic sites. This is in accordance with pre-
vious reports both on non-zeolite catalysts [6,43,51] and acidic
zeolites [4,5,40]. Moreover, as shown in Figs. 2a and 3a, the
maximum conversion of NOx is observed over the HZSM-5(40)
sample, both for isobutane and propane, however, at different
temperatures, 350 ◦C for isobutane and 450 ◦C for propane. Fur-
thermore, isobutane exhibits higher maximum NOx conversion
(54%) than propane (39%).
When relating the conversion of NOx over the three samples
with their aluminum content, which may be assumed to corre-
spond to the number of Brønsted acid sites, it seems that the
activity per site is somewhat higher for the samples with the
higher SAR, indicating higher turn-over frequencies over these
samples. This is possibly related to differences in the strength
of the acid sites between the samples. However, the purpose of
the sample comparison was to find the sample giving the highest
conversion of NOx and for this reason it appears that the total
acidity is of greater importance making the HZSM-5(40) sample
the one showing highest overall conversion.
3. Results
3.1. Evaluation of the catalytic activity
3.1.1. Characterization of surface acidity
The surface acidity was investigated using DRIFT spec-
troscopy under Ar exposure (Fig. 1). The absorption peaks at
3737 and 3600 cm−1 are assigned to O H stretching vibrations
originating from silanol groups (Si OH) and Si(OH)Al bridg-
inghydroxylgroups, respectively[23,44–50]. Theresultsclearly
demonstrate the difference in acidity between the different sam-
ples. The sample with higher alumina content, HZSM-5(40),
contains significant amounts of Brønsted acid sites, whereas the
other two samples, HZSM-5(234) and HZSM-5(569), contain
considerably less or negligible amounts of these sites. Detailed
results from ammonia adsorption experiments, over these sam-
ples, have been reported in a previous paper [40].
3.1.2. Influence of surface acidity on the NOx conversion
Figs. 2a and 3a show the NOx conversion over the catalysts,
HZSM-5(40), HZSM-5(234) and HZSM-5(569), for SCR of
NO2 with isobutane and propane, respectively, as a function
of temperature, allowing about 10 min to reach steady state at
each temperature.
For isobutane (Fig. 2a), the highest activity for NOx reduction
is obtained with the HZSM-5(40) sample at 350 ◦C, followed,
in decreasing order of activity, by HZSM-5(234) and HZSM-
5(569). When using propane as the reducing agent, the same
order of activity is observed for the maximum NOx conversion
(Fig. 2b) at 450 ◦C for the three samples. These results indi-
3.1.3. Dependence of product selectivity on the type of
hydrocarbon
Since the HZSM-5(40) sample gave the highest NOx con-
version among the catalysts studied, the following studies were
focused on this catalyst. Figs. 2b and 3b show the conversion of
NO2 and hydrocarbon (isobutane or propane), and the selectivity
for N2 and NO formation as a function of temperature.
The NO2 conversion in percent is defined as:
100 × nNO conv.
2
CNO conv.
=
2
n
NO2 add.

16
D. Zhao et al. / Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
Fig. 3. (a) Temperature dependence of the NOx conversion in the SCR of
NO2 with propane over a series of acidic HZSM-5 samples. HZSM-5(40)
(squares), HZSM-5(234) (triangles) and HZSM-5(569) (diamonds). (b) Temper-
ature dependence of the conversion of NO2 (diamonds) and propane (crosses),
and the selectivity of N2 (triangles), NO (squares) and N2O (circles), in the SCR
reaction of NO2 with propane over the HZSM-5(40) sample. Gas composition:
300 ppm propane, 1040 ppm NO2, 10% O2 in He, total flow 1500 ml (NTP)/min.
Fig. 2. (a) Temperature dependence of the NOx conversion in the SCR of
NO2 with isobutane over a series of acidic HZSM-5 samples. HZSM-5(40)
(squares), HZSM-5(234) (triangles) and HZSM-5(569) (diamonds). (b) Temper-
ature dependence of the conversion of NO2 (diamonds) and isobutane (crosses),
and the selectivity of N2 (triangles), NO (squares) and N2O (circles), in the
SCR reaction of NO2 with isobutane over the HZSM-5(40) sample. Gas com-
position: 200 ppm isobutane, 1040 ppm NO2, 10% O2 in He, total flow 1500 ml
(NTP)/min.
propane as the reductant, however in different temperature
regions(250–450 ◦Cforisobutaneand450–650 ◦Cforpropane).
The selectivity to N2 increased slightly (from 46 to 57%) in the
active temperature range for i-C4H10, while the corresponding
value for propane decreased (from 51 to 29%) with increasing
temperature. The selectivity to NO was approximately constant
(increased from 47 to 48%) for isobutane and increased (from 61
to 81%) for propane in their respective active temperature range.
The change in NO selectivity corresponds well with the change
in selectivity to N2. This indicates that the main N containing
products in the reaction are NO and N2. Some N2O formation
was also observed during the reaction, but the concentration
was low or negligible (below 20 ppm) at temperatures of 350 ◦C
and above, whereas at 200 and 250 ◦C somewhat higher N2O
concentrations were obtained in the isobutane case. The con-
version of hydrocarbons appeared to be different for isobutane
and propane. Isobutane was converted to almost 100% within its
active temperature range (250–450 ◦C), while the conversion of
propane was only 62% at 450 ◦C, which subsequently increased
to 100% when the temperature reached 650 ◦C. The conversion
where nNO conv. is the number of NO2 molecules converted and
2
nNO add. is the number of NO2 molecules added.
2
The hydrocarbon conversion in percent is defined as:
100 × nHC conv.
CHC conv.
=
nHC add.
where nHC conv. is the number of hydrocarbon molecules con-
verted and nHC add. is the number of hydrocarbon molecules
added.
The N2, NO and N2O selectivity in percent is defined as:
100 × nform. × n
S =
,
n
NO2 conv.
where nform. is the number of N2, NO or N2O molecules formed,
n the number of nitrogen atoms in the molecule and nNO conv.
2
is the number of NO2 molecules converted.
As shown in Figs. 2b and 3b, the conversion of NO2 reached
about 95% or higher with isobutane and 90% or higher with

D. Zhao et al. / Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
17
of i-C4H10, which correlates with the corresponding slightly
decreased NOx conversion for i-C4H10 (Fig. 2a) over the same
temperature range, indicates that the reaction mainly occurs
between NOx and i-C4H10. This is verified by the selective
formation of N2, which increased in the same active temper-
ature range. However, within the active temperature range for
propane, the NOx reduction over the HZSM-5(40) sample actu-
ally decreased; hence, the higher conversion of propane at higher
temperatures was most likely caused by reaction with oxygen.
This correlates well with the N2 formation, which decreased in
the corresponding temperature range. From this it follows that
isobutane is effectively activated at lower temperatures (250 ◦C)
compared to propane and that the reaction with NO2 is favored
rather than the oxidation with oxygen. This results in a wider
active temperature window (250–450 ◦C) in NOx reduction for
isobutane compared to propane for which high activity for NOx
reduction appeared only around 450 ◦C.
CO peaks is stronger when the temperature is higher, indicating
more CO formation at higher temperatures (Fig. 4). Meanwhile,
theN2Oformationishigheratlowtemperaturesandnegligibleat
higher temperatures. Furthermore, peaks in the N H stretching
region, 3200–3400 cm−1, which are more pronounced at 350 ◦C,
may indicate formation of organic amine species. The peak at
3180 cm−1, whichlikelyoriginatesfromthealkeneC Hstretch-
ing, indicates formation of alkenes. Species like HCN, HNCO
and/or NH3 were not observed during these experiments.
The corresponding gas phase FTIR spectra for the HZSM-
5(234) sample are shown in Fig. 5, at 250, 350 and 450 ◦C. The
NO2, CO2 and NO are observed by the presence of their char-
acteristic peaks, as described previously for the HZSM-5(40)
sample. The peaks for CO and N2O change in a similar way
as observed for the HZSM-5(40) sample, i.e. at lower temper-
ature the N2O formation is higher. The intensity of the N H
stretching peaks (3200–3400 cm−1) is similar to that observed
for the HZSM-5(40) sample (gas phase). Formation of HCN
and NH3 could not be observed. However, there is a shoulder
at 2280 cm−1 (Fig. 5) appearing at 250 ◦C in the spectrum, typ-
ical for the –NCO group in gas phase, which may indicate the
presence of HNCO or –NCO species.
3.1.4. Reaction products and possible reaction
intermediates
There are several reports in the literature, where C, N and
O containing species have been observed as by-products in the
gas phase (such as HCN, HNCO, NH3, etc.) during the SCR
reaction. The gas phase FTIR spectra recorded in this study dur-
ing the activity experiments over the HZSM-5(40) sample are
shown in Fig. 4, at 250, 350 and 450 ◦C. Even though the exact
structures of some of the molecules involved in these spectra
are uncertain, the characteristic absorption peaks of different
functional groups can be recognized. The peaks at 1630 and
1595 cm−1 typically originate from asymmetric stretching of
the NO2 group and the peaks at 2915 and 2885 cm−1, from the
C H stretching of isobutane. The main products CO2 (peaks at
3.2. In situ DRIFT studies; formation of surface bound
species
The mechanism of lean NO2 reduction with isobutane was
studied by in situ DRIFT step–response experiments. Below
2000 cm−1 the background noise level is high, which makes
identification of the peaks difficult. This noise is probably related
to perturbed vibration frequencies of the zeolite framework
[49,50,53].
2355 and 2360 cm−1) and NO (peaks at 1890 and 1830 cm−1
)
are observed at all temperatures investigated. The two doublets
at 2100 and 2165 cm−1 and at 2200 and 2210 cm−1 are typi-
cally CO and N2O absorption peaks [52]. The intensity of the
3.2.1. Co-adsorption of NO2 and O2
Fig. 6a shows the DRIFT spectrum of reaction sequence
1, after introduction of 1000 ppm NO2 and 10% O2 over the
Fig. 4. Gas phase IR spectra recorded at 250 ◦C (light grey), 350 ◦C (grey) and 450 ◦C (black) during activity experiments over the HZSM-5(40) sample. Gas
composition: 1040 ppm NO2, 200 ppm isobutane and 10% O2 in He. The insets represent enlargement of the 3400–3000 and 2250–2050 cm−1 regions, characteristic
for N–H and unsaturated C–H stretching vibrations, and for isocyanates, N2O and CO, respectively.

18
D. Zhao et al. / Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
Fig. 5. The gas phase IR spectra recorded at 250 ◦C (light grey), 350 ◦C (grey) and 450 ◦C (black) during the activity experiments over the HZSM-5(234) sample. Gas
composition: 1040 ppm NO2, 200 ppm isobutane and 10% O2 in He. The insets represent enlargement of the 3400–3000 and 2250–2050 cm−1 regions, characteristic
for N–H and unsaturated C–H stretching vibrations, and for isocyanates, N2O and CO, respectively.
HZSM-5(40) sample at 250 ◦C. A positive peak at 2140 cm−1
appears and its intensity grows with time. This peak has been
assigned, by Hadjiivanov et al. [44], to adsorbed NO+ species.
Negative peaks at 3610 and 3745 cm−1 are also observed, which
are assigned to O H stretching vibrations originating from
Brønsted acid sites and silanol groups, respectively [44–50].
The negative peak at 3610 cm−1 reflects the NO+ groups, which
are mainly adsorbed on Brønsted acid sites. Two broad peaks,
centred around 2880 and 2490 cm−1, appear which probably
originate from the perturbed O H stretching vibrations caused
by hydrogen bonds with H in the OH group [49]. The NO2/O2
experiment was repeated at 350 ◦C, and the same peaks at 2140,
3610 and 3745 cm−1 are observed in the spectrum (Fig. 6b),
though with lower intensity, except for the two broad water
peaks, which almost disappeared due to higher temperature.
The corresponding spectra (Fig. 6c) for the HZSM-5(234)
sample showed the same positive NO+ peak at 2140 cm−1; how-
ever, its steady state intensity was much lower compared to the
HZSM-5(40) sample at 350 ◦C. The weaker intensity is likely
due to the lower number of Brønsted acid sites present in this
sample. No broad zeolite hydroxyl peaks are observed at this
temperature.
3.2.2. Reduction of NO2 by isobutane
Fig. 7a shows the step–response DRIFT experiments when
200 ppm isobutane is introduced to the mixture of 1000 ppm
NO2 and 10% O2 over the HZSM-5(40) sample at 250 ◦C
(sequence 2). A broad feature evolves in the 3500–2000 cm−1
region in which also more narrow peaks develop. The growth of
the two broad peaks, centred around 2880 and 2490 cm−1, indi-
cates the formation of water from the reaction. The presence of
the peaks in the range 2000–2300 cm−1 suggests formation of
new species containing double bonds (e.g. –N C O) or triple-
Fig. 7. Step–response DRIFT experiments over the HZSM-5(40) sample at
250 ◦C, using isobutane as the reducing agent. Light grey corresponds to time
direct after the gas switch and time is then increased until steady state is
reached, which is illustrated with black. (a) Isobutane is turned on; 1000 ppm
NO2, 200 ppm C4H10, 10% O2 in Ar (sequence 2). (b) Isobutane is turned off;
1000 ppm NO2, 10% O2 in Ar (sequence 3). For clearance the log Io/I values
have been off-set.
Fig. 6. DRIFT experiments with 1000 ppm NO2 and 10% O2 in Ar (sequence
1). Light grey corresponds to time direct after the gas switch and time is then
increased until steady state is reached, which is illustrated with black. (a) HZSM-
5(40) at 250 ◦C, (b) HZSM-5(40) at 350 ◦C and (c) HZSM-5(234) at 350 ◦C.
For clearance the log Io/I values have been off-set.

D. Zhao et al. / Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
19
Fig. 8. Step–response DRIFT experiments over the HZSM-5(40) sample at
350 ◦C, using isobutane as the reducing agent. Light grey corresponds to time
direct after the gas switch and time is then increased until steady state is
reached, which is illustrated with black. (a) Isobutane is turned on; 1000 ppm
NO2, 200 ppm C4H10, 10% O2 in Ar (sequence 2). (b) Isobutane is turned off;
1000 ppm NO2, 10% O2 in Ar (sequence 3). For clearance the log Io/I values
have been off-set.
bonds (e.g. –C N and –C C) [46]. Peaks at 2257 cm−1 are often
assigned to isocyanate (–N C O) and/or nitrile (–C N) species
[16,23,50,54–56]. The assignment of peaks in this region to
adsorbed species with well defined structure is difficult because
cyanides, isocyanides, cyanates, isocyanates and carbonyl all
exhibit strong IR peaks between 2300 and 2000 cm−1 [56]. The
2140 cm−1 (NO+) peak is observed as a shoulder growing on
the large feature. There are indications of peaks growing in
the 2500–2400 cm−1 region, which may be due to t-butylnitrile
N-oxide [41], but could also be assigned to bidentate nitrate
[23]. In the region 3000–2800 cm−1, C H stretching vibrations
occur [16,23,57]. In particular, a peak at 2950 cm−1 can prob-
ably be attributed to isobutane [24] (in gas phase or loosely
bound to the surface), and a peak at 2995 cm−1 can be due
to presence of t-butyl groups [58]. A clear peak appears and
develops with time at 3133 cm−1, which can be assigned to
C H stretching vibrations of unsaturated hydrocarbons [16,24]
and/or N H stretching vibrations of amine species or ammo-
nia [24,59,60]. There are also indications of a shoulder in the
region 3400–3200 cm−1, which probably originates from N H
stretching vibrations [16,23,24,28,61]. Moreover, the concomi-
tant increase in intensity of the negative peaks at 3745 and
3610 cm−1 implies adsorbates on the Brønsted OH groups,
likely isobutane. The experiment was repeated over the same
catalyst at 350 ◦C and the results are shown in Fig. 8a. In the large
feature region, 3500–2000 cm−1, the two broad zeolite hydroxyl
peaks appear more flat than at 250 ◦C indicating less adsorption
of water at this temperature. All the other peaks described for
250 ◦C were observed also at 350 ◦C. The shoulder that appears
in the NH stretching region (3400–3200 cm−1) is much more
pronounced at 350 ◦C than at 250 ◦C, indicating that more –NH
surface species are formed at this temperature. At both temper-
atures the isocyanate species (2257 cm−1) and the unsaturated
hydrocarbons and/or amine species (3135 cm−1) increase imme-
diately after isobutane has been introduced. At the same time the
Brønsted acid sites are blocked, likely by these species (Fig. 9).
Fig. 9. Intensity variation with time for isocyanate species (2257 cm−1), amine
species and/or unsaturated hydrocarbons (3135 cm−1) and Brønsted acid sites
(3610 cm−1), when 200 ppm isobutane is added to the reaction gas mixture
(1000 ppm NO2 and 10% O2) at: (a) 350 ◦C and (b) 250 ◦C. The values are
normalized between 0 and 1, and between −1 and 0, respectively.
At 350 ◦C the 3135 cm−1 peak reaches a maximum where after
it decreases. This is possibly related to the water formation in
the reaction. Isocyanate species are proposed to hydrolyze with
water forming amine species; however, the water is likely con-
sumed initially and the formation may not be sufficient for a
total hydrolysis of isocyanate species.
In Fig. 10 the results from the NO2 reduction with isobutane
over the HZSM-5(234) sample at 350 ◦C are shown. The results
are similar to the results for the HZSM-5(40) sample with the
–NCOand/or–CN(peakat2257 cm−1)speciesgrowing, andthe
Brønsted acid sites (peaks at 3610 cm−1) becoming increasingly
affected with time. However, there is a large difference in the
intensities of the peaks, which are much weaker for the HZSM-
5(234) sample compared to the HZSM-5(40) sample, as a result
of less adsorbed species on the catalyst surface.
3.2.3. Reactivity of the adsorbed complexes formed during
NO2 reduction
The chemical stability of the adsorbed reaction complexes
over HZSM-5(40) towards NO2 has been investigated (sequence
3). When the NO2 reduction has reached steady state in sequence
2, the supply of isobutane was turned off, still allowing NO2 + O2
through the sample (sequence 3). This sequence was per-
formed both at 250 and 350 ◦C, and the results are shown in
Figs. 7b and 8b, respectively. The light grey line represents the

20
D. Zhao et al. / Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
complexes are adsorbed on the catalyst surface, the complexes
seem to be more reactive than over the HZSM-5(40) sample.
This is indicated by less blockage of hydrocarbons over this
sample and further supported by the C H peaks at 3133 cm−1
,
which decrease with time upon exposure to NO2 + O2. The neg-
ative peak at 3610 cm−1, originating from the Brønsted acid
sites, decreased considerably more compared to the change over
the HZSM-5(40) sample, indicating a higher recovery of active
Brønsted acid sites.
4. Discussion
In a previous study, we investigated the reaction mechanism
for NO2 reduction by propane, as an example for saturated
hydrocarbons, over acidic HZSM-5 zeolites [40]. The present
study focuses on the differences in NO2 SCR activity between
two saturated hydrocarbons: propane and isobutane and on the
reaction mechanism for NO2 SCR with isobutane.
The HZSM-5(40) sample shows the highest NOx conver-
sion and the highest N2 formation for all temperatures studied
and for both reducing agents (isobutane and propane) due to
its higher content of Brønsted acid sites. Moreover, the present
study shows that the NO2 reduction in O2 excess is dependent on
the type of hydrocarbon used. Over the most active HZSM-5(40)
sample, isobutane was effectively activated at 250 ◦C, and the
effectiveness, by which isobutane reduces NOx, was sustained
over a wider temperature range compared to propane.
The influence of mass transfer limitations on the reduction
of NOx is considered by estimating the Weiszs-modulus [62],
in accordance with calculations presented previously [40], and
the results are shown in Table 2. This estimation compares the
reaction rate and the diffusion rate of the reactants at the active
surface. When the reaction rate is high, the reaction can essen-
tially take place in the first small part of the catalyst bed, and
hence there may be transport limitations in this part of the cata-
lyst bed. In most cases of the present study, the catalytic reaction
Fig. 10. Step–response DRIFT experiments over the HZSM-5(234) sample
at 350 ◦C, using isobutane as the reducing agent. Light grey corresponds to
time direct after the gas switch and time is then increased until steady state is
reached, which is illustrated with black. (a) Isobutane is turned on; 1000 ppm
NO2, 200 ppm C4H10, 10% O2 in Ar (sequence 2). (b) Isobutane is turned off;
1000 ppm NO2, 10% O2 in Ar (sequence 3). For clearance the log Io/I values
have been off-set.
reaction at steady state (sequence 2), which is the starting point
of the exposure to NO2 + O2, and the black line represents the
end of the exposure, at which no more change can be observed in
the spectra. As can clearly be seen, the reaction complex formed
at 250 ◦C is chemically very stable towards gas phase NO2,
since no changes are observed in the DRIFT spectra (Fig. 7b)
after 5 min exposure to NO2 in 10% O2. However, at 350 ◦C,
as shown in Fig. 8b, some peaks decrease in intensity. The two
broadhydroxylpeakscentredat2490and2880 cm−1 decreaseto
a certain extent and the isocyanate (–NCO) and/or nitrile (–CN)
peak at 2257 cm−1, and the NO+ peak at 2140 cm−1 decrease
correspondingly. This may indicate that the adsorbed isocyanate
and/or nitrile species on the surface are hydrolyzed with trace
amounts of adsorbed water (formed on the surface during the
reaction) and the hydrolysis products likely react with NO+
species [25], leading to a decrease in the NO+ peaks. When
the adsorbed water is mostly consumed, the hydrolysis reaction
ceases and no more decrease in intensity can be observed. The
remaining peaks likely originate from hydrocarbons or deriva-
tives of hydrocarbons hydrogen-bonded to the zeolite hydroxyl
groups. Thepeakat3135 cm−1, originatingfromC Hstretching
vibrations of unsaturated hydrocarbons, only slightly decreases
in intensity, and the negative peak at 3610 cm−1 was also slightly
decreased in intensity implying a slight recovery of Brønsted
acid sites.
The corresponding results obtained for the HZSM-5(234)
sample at 350 ◦C are shown in Fig. 10b. The light grey line repre-
sents the start of the exposure to NO2 + O2 (in Ar) and the black
linerepresentstheendoftheexposure, atwhichnomorechanges
are observed in the spectra. The changes are more pronounced
for this sample compared to HZSM-5(40). The intensity of
the two broad bands (centred around 2490 and 2880 cm−1),
the nitrile and/or isocyanate species (peak at 2257 cm−1) and
NO+ species (2140 cm−1) decreases with time. The decrease
may be caused by the hydrolysis of –CN/–CNO species with
trace amounts of adsorbed water, and the reaction of hydroly-
sis products with NO+ species. Even though less amounts of
Table 2
The values used for estimation of the Weiszs-modulus at 450 ◦C for propane and
at 350 ◦C for isobutane
Propane
Isobutane
robs (mol/(s kgcat))
Φ
robs (mol/(s kgcat))
Φ
HZSM-5(40)
NO2 cat.part.
NO2 zeolite.part.
HCcat.part.
4.0E−04
6.1E−04
2.1E−04
3.2E−04
0.43
0.52
0.77
0.95
5.75E−04
8.84E−04
2.07E−04
3.18E−04
0.53
0.65
0.99
1.22
HCzeolite.part.
HZSM-5(234)
NO2 cat.part.
NO2 zeolite.part.
HCcat.part.
1.7E−04
2.6E−04
1.2E−04
1.8E−04
0.18
0.23
0.44
0.54
3.39E−04
5.21E−04
1.67E−04
2.57E−04
0.31
0.38
0.80
0.99
HCzeolite.part.
HZSM-5(569)
NO2 cat.part.
NO2 zeolite.part.
HCcat.part.
1.4E−04
2.1E−04
3.5E−05
5.4E−05
0.15
0.18
0.13
0.16
2.20E−04
3.38E−04
8.47E−05
1.30E−04
0.20
0.25
0.41
0.50
HCzeolite.part.

D. Zhao et al. / Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
21
is likely not influenced by mass transfer limitations. This means
that the diffusion of the reactants, i.e. NO2 and propane or isobu-
tane, to the active sites within the zeolite channels [63] does not
limit the SCR reaction. However, in four cases the values of
the Weiszs-modulus are close to or higher than 1. The critical
point is reached for isobutane over HZSM-5(40) and HZSM-
5(234), when the isobutane conversion is high (close to 100 and
72%, respectively) and for propane over the most acidic sample,
HZSM-5(40), when the propane conversion is 65%.
The initiation of the reaction between NO2 and hydrocar-
bons has been proposed to proceed via either a radical type
or electrophilic mechanism [64]. The radical type reaction is
mostly proposed over transition metal exchanged zeolite sys-
tems with either unsaturated or saturated hydrocarbons, and
the electrophilic mechanism is demonstrated over acidic zeo-
lite catalysts only for unsaturated hydrocarbons, in particular
ethene [30] and propene [29]. The initiation of the SCR reac-
tion with saturated hydrocarbons over acidic zeolite catalysts
is not clearly demonstrated in the literature. Burch and Hayes
[57] suggested that the activation of the C H bond over acidic
zeolite catalysts proceeds via donation of Brønsted acid protons
to the hydrocarbon, forming an adsorbed carbonium ion, which
may, in turn, be dehydrogenated to an adsorbed carbenium ion.
The subsequent reaction with NOx can be relatively facile. We
have previously suggested formation of the carbenium ions as
an important substep in lean NOx reduction with propane over
HZSM-5 [40]. Here this mechanism may also explain the cause
of the differences in activity between propane and isobutane.
Tertiary carbons (in isobutane) have higher affinity to accept
protons leading to more stable tertiary carbenium ions than the
secondary carbons in propane.
It is suggested in the literature that when both C, O and
N containing adsorbates are formed on the surface, NO2 can
react with these adsorbates and eventually form N2 [15,16,24].
However, NO2 will most probably not react directly with
adsorbates (following an Eley–Rideal mechanism), instead the
adsorbed species containing only N and O (from co-adsorption
of NO2 + O2) likely react with the C, O and N containing
adsorbed complexes probably via a Langmuir–Hinshelwood
mechanism [15]. Based on in situ spectroscopy we have pro-
posedareactionpathforSCRofNO2 withpropaneoverthesame
catalytic system, and we observe the formation of isocyanate
and amine species on the catalyst surface [40]. In the present
study these species were clearly observed in the in situ DRIFT
spectroscopy measurements when isobutane was used as the
reducing agent for lean NOx reduction, over HZSM-5(40) and
HZSM-5(234). The growth and decline of these species in the
step–response experiments strongly indicate their involvement
inthereaction. Furthermore, thecorresponding changesin inten-
sity of the Brønsted OH stretching vibrations indicate that these
species most likely are formed on the Brønsted acid sites. These
observations further support our proposed pathway for the SCR
reaction, which most likely starts by the formation of adsorbed
carbenium ions and NO+ species. These species will then further
react to form adsorbed isocyanate species, which may hydrolyze
with water forming adsorbed amine species. Adsorbed NO+
species are then likely to react with these amine species to
form N2. In addition, in the present study, species containing
–NCO were observed also in the gas phase at 250 ◦C over the
HZSM-5(234) sample, which further support formation of iso-
cyanate or HNCO during the reaction. Acke and Skoglundh [65]
reported that HNCO reacted to almost 100% over a Pt/HZSM-5
catalyst above 210 ◦C for the NO reduction in O2 excess. Liu
et al. [27] observed the emergence of ethyl isocyanate below
330 ◦C over Cu-MFI under similar conditions when the catalyst
became deactivated by surface deposits. This may explain why
the –NCO/HNCO species peak at 2280 cm−1 was not observed
at 350 and 450 ◦C (in the gas phase) in the present study due
to its instability at higher temperature. Furthermore, this peak is
not seen in the gas phase over the HZSM-5(40) sample, proba-
bly because it is consumed in further reactions as soon as it is
formed over this sample.
The obvious difference observed in the DRIFT study between
propane and isobutane is the formation of surface adsorbates on
the catalysts, which is considerably more pronounced for isobu-
tane than for propane, shown by the formation of broad peaks
in the region of 2000–3500 cm−1. This is most likely due to the
fact that tertiary carbons in isobutane are considerably easier
to activate than secondary carbons in propane, which results in
higher coverage of surface adsorbates (saturated and/or unsatu-
rated hydrocarbons) for isobutane.
5. Concluding remarks
The lean reduction of NO2 by saturated hydrocarbons over
acidic HZSM-5 zeolites is strongly dependent on the type of
hydrocarbon where isobutane is a more effective reducing agent
than propane showing a wider active temperature window, start-
ing at a lower temperature. Key steps in the reaction of NO2 are
likely the formation of NO+ on the surface and the activation
of the hydrocarbon. This is supported by in situ DRIFT studies
showing formation of NO+, isocyanate and amine species on the
zeolite surface as well as –CNO/HCNO species in the gas phase
during the reaction of NO2 and isobutane. The differences in
reactivity between the two hydrocarbons stem from the fact that
the Brønsted acid sites activate isobutane at a lower temperature
than propane. This is because the tertiary hydrogen in isobu-
tane is more reactive than the secondary hydrogen in propane.
This supports a reaction mechanism for lean NO2 reduction over
acidic zeolites where the hydrocarbon is activated over Brønsted
sitesformingcarbeniumionadsorbates, whichmayformalkenes
that react with NO+ species forming isocyanate species, amine
species and finally nitrogen.
Acknowledgements
The authors would like to thank Dr. Sander van Donk, Albe-
marle Catalysts BV, for fruitful discussions and for providing the
zeolites. This work has been performed within the Department
¨
of Chemistry, Goteborg University, and the Competence Centre
for Catalysis, which is hosted by Chalmers University of Tech-
nology and financially supported by the Swedish Energy Agency
and the member companies: AB Volvo, Volvo Car Corporation,
Scania AB, GM Powertrain Sweden AB, Haldor Topsøe A/S,

22
D. Zhao et al. / Journal of Molecular Catalysis A: Chemical 249 (2006) 13–22
Perstorp AB, Albemarle Catalysts BV and The Swedish Space
Corporation.
[30] T. Gerlach, M. Baerns, Chem. Eng. Sci. 54 (1999) 4379.
[31] K. Shimizu, F. Okada, Y. Nakamura, A. Satsuma, T. Hattori, J. Catal.
195 (2000) 151.
[32] F. Radtke, R.A. Koeppel, A. Baiker, Appl. Catal. A 107 (1994) L125.
[33] F. Radtke, R.A. Koeppel, A. Baiker, Environ. Sci. Technol. 29 (1995)
2703.
[34] F. Radtke, R.A. Koeppel, A. Baiker, Chem. Commun. (1995) 427.
[35] F. Radtke, R.A. Koppel, A. Baiker, Catal. Today 26 (1995) 159.
[36] S. Akaratiwa, T. Nanba, A. Obuchi, J. Okayasu, J.O. Uchisawa, S.
Kushiyama, Top. Catal. 16 (2001) 209–216.
[37] H. Hamada, Catal. Surv. Jpn. 1 (1997) 53.
[38] T. Holma, A. Palmqvist, M. Skoglundh, E. Jobson, Appl. Catal. B 48
(2004) 95.
[39] J.A. Martens, A. Cauvel, F. Jayat, S. Vergne, E. Jobson, Appl. Catal. B
29 (2001) 299.
References
[1] M. Iwamoto, H. Hamada, Catal. Today 10 (1991) 57.
[2] M. Iwamoto, H. Yahiro, S. Shundo, Y. Yu-u, N. Mizuno, Shokubai
(Catalyst) 32 (1990) 430.
[3] W. Held, A. Koenig, T. Richter, L. Puppe, SEA Paper 900496 (1990).
[4] H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito, M. Tabata, Appl. Catal. 64
(1990) L1.
[5] H. Hamada, Y. Kintaichi, M. Sasaki, T. Ito, M. Tabata, Appl. Catal. 70
(1991) L15.
[6] Y. Kintaichi, H. Hamada, M. Tabata, M. Sasaki, T. Ito, Catal. Lett. 6
(1990) 239.
[40] H.H. Ingelsten, D. Zhao, A. Palmqvist, M. Skoglundh, J. Catal. 232
(2005) 68.
[41] A. Obuchi, C. Wogerbauer, R. Koppel, A. Baiker, Appl. Catal. B 19
(1998) 9.
[7] I.O.Y. Liu, N.W. Cant, Catal. Surv. Asia 7 (2003) 191.
[8] X. Wang, H.Y. Chen, W.M.H. Sachtler, J. Catal. 197 (2001) 281.
[9] T. Nanba, S. Masukawa, J. Uchisawa, A. Obuchi, Chem. Lett. 33 (2004)
924.
[42] A. Hinz, M. Skoglundh, E. Fridell, A. Andersson, J. Catal. 201 (2001)
[10] T. Nanba, A. Obuchi, Y. Sugiura, C. Kouno, J. Uchisawa, S. Kushiyama,
J. Catal. 211 (2002) 53.
247.
˚
[43] H.H. Ingelsten, A. Hildesson, E. Fridell, M. Skoglundh, J. Mol. Catal.
[11] T. Nanba, A. Obuchi, H. Izumi, Y. Sugiura, J.Y. Xu, J. Uchisawa, S.
Kushiyama, Chem. Commun. (2001) 173.
[12] N.W. Cant, D.C. Chambers, A.D. Cowan, I.O.Y. Liu, A. Satsuma, Top.
Catal. 10 (2000) 13.
[13] H.Y. Chen, X. Wang, W.M.H. Sachtler, Phys. Chem. Chem. Phys. 2
(2000) 3083.
[14] H.Y. Chen, X. Wang, W.M.H. Sachtler, Appl. Catal. A 194 (2000) 159.
[15] H.Y. Chen, T. Voskoboinikov, W.M.H. Sachtler, J. Catal. 186 (1999) 91.
[16] H.Y. Chen, T. Voskoboinikov, W.M.H. Sachtler, Catal. Today 54 (1999)
483.
A 209 (2004) 199.
[44] K. Hadjiivanov, J. Saussey, J.L. Freysz, J.C. Lavalley, Catal. Lett. 52
(1998) 103.
[45] J. Szanyi, M.T. Paffett, J. Catal. 164 (1996) 232.
[46] M. Trombetta, G. Busca, S. Rossini, V. Piccoli, U. Cornaro, A. Guercio,
R. Catani, R.J. Willey, J. Catal. 179 (1998) 581.
[47] G. Boskovic, T. Vulic, E. Kis, P. Putanov, Chem. Eng. Technol. 24
(2001) 269.
[48] B.I. Mosqueda-Jimenez, A. Jentys, K. Seshan, J.A. Lercher, Appl. Catal.
B 43 (2003) 105.
[17] A.D. Cowan, N.W. Cant, B.S. Haynes, P.F. Nelson, J. Catal. 176 (1998)
329.
[49] M. Wallin, C.-J. Karlsson, A. Palmqvist, M. Skoglundh, Top. Catal.
30/31 (2004) 107.
[18] T. Beutel, B. Adelman, W.M.H. Sachtler, Catal. Lett. 37 (1996) 125.
[19] B.J. Adelman, T. Beutel, G.D. Lei, W.M.H. Sachtler, Appl. Catal. B 11
(1996) L1.
[50] P.E. Fanning, M.A. Vannice, J. Catal. 207 (2002) 166.
[51] H.H. Ingelsten, M. Skoglundh, E. Fridell, Appl. Catal. B 41 (2003) 287.
[52] NIST
Standard
Reference
Database
69,
[20] J.J. Wu, S.C. Larsen, J. Catal. 182 (1999) 244.
[21] E.V. Rebrov, A.V. Simakov, N.N. Sazonova, E.S. Stoyanov, Catal. Lett.
58 (1999) 107.
[22] E.V. Rebrov, A.V. Simakov, N.N. Sazonova, E.S. Stoyanov, Catal. Lett.
64 (2000) 129.
[23] F. Poignant, J.L. Freysz, M. Daturi, J. Saussey, Catal. Today 70 (2001)
197.
[24] H.Y. Chen, T. Voskoboinikov, W.M.H. Sachtler, J. Catal. 180 (1998)
171.
[25] C.A. Jones, S.C. Larsen, Catal. Lett. 78 (2002) 243.
[26] N.W. Cant, A.D. Cowan, A. Doughty, B.S. Haynes, P.F. Nelson, Catal.
Lett. 46 (1997) 207.
[27] I.O.Y. Liu, N.W. Cant, B.S. Haynes, P.F. Nelson, J. Catal. 203 (2001)
487.
[28] F. Poignant, J. Saussey, J.C. Lavalley, G. Mabilon, Catal. Today 29
(1996) 93.
[29] T. Gerlach, F.W. Schutze, M. Baerns, J. Catal. 185 (1999) 131.
http://webbook.nist.gov/chemistry, 27 January 2005.
[53] R. Szostak, Molecular Sieves, Van Nostrand Reinhold, New York, 1989.
[54] N.W. Hayes, W. Grunert, G.J. Hutchings, R.W. Joyner, E.S. Shpiro,
Chem. Commun. (1994) 531.
[55] S.K. Park, H. Choo, L. Kevan, Phys. Chem. Chem. Phys. 3 (2001) 3247.
[56] F. Solymosi, T. Bansagi, J. Catal. 156 (1995) 75.
[57] R. Burch, M.J. Hayes, J. Mol. Catal. A 100 (1995) 13.
[58] M. Xin, I.C. Hwang, S.I. Woo, J. Phys. Chem. B 101 (1997) 9005.
[59] E. Jobson, A. Baiker, A. Wokaun, J. Chem. Soc. Faraday Trans. 86
(1990) 1131.
[60] N.Y. Topsoe, H. Topsoe, J.A. Dumesic, J. Catal. 151 (1995) 226.
[61] S.K. Park, Y.K. Park, S.E. Park, L. Kevan, Phys. Chem. Chem. Phys. 2
(2000) 5500.
[62] P.B. Weiszs, C.D. Prater, Adv. Catal. 6 (1954) 143.
[63] F. Witzel, G.A. Sill, W.K. Hall, J. Catal. 149 (1994) 229.
[64] R. Brosius, J.A. Martens, Top. Catal. 28 (2004) 119.
[65] F. Acke, M. Skoglundh, Appl. Catal. B 20 (1999) 235.