The role of Bronsted acidity in the selective catalytic reduction of NO with ammonia over Fe-ZSM-5

Article abstract of DOI:10.1016/j.jcat.2009.09.028

The selective catalytic reduction of NO with NH₃ has been studied over Fe-ZSM-5 samples with equal exchange degrees and active sites but different Bronsted acidities. Additionally, the activity of fresh and aged samples in NH₃-SCR with NO (standard SCR) has been compared with that in NH₃-SCR with NO/NO₂ (fast SCR), where oxidation activity is not essential. Using NH₃-TPD-FTIR experiments and DRIFTS measurements, it was shown that the concentration of Bronsted acid sites decreases with an increase in Fe loading, suggesting that, up to Fe/Al = 0.39, about one Bronsted acid site is substituted by one Fe site and that most of the iron is present as a dispersion of cations. A maximum of about 55% of all Bronsted acid sites can be exchanged in aqueous Fe²⁺ solution. These experiments revealed that the acidity of the catalyst is not a crucial factor for high activity and that Bronsted acidity may not be required for adsorbing or activating the ammonia, but it is necessary to bind and disperse the metal ions. On the other hand, the oxidation activity of Fe-ZSM-5 was found to be the main factor that controls the high activity. The altered oxidation activity is largely responsible for the decreased SCR activity after hydrothermal aging.

Full text of DOI:10.1016/j.jcat.2009.09.028

Journal of Catalysis 268 (2009) 297–306
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
The role of Brønsted acidity in the selective catalytic reduction of NO
with ammonia over Fe-ZSM-5
a
b
b
Sandro Brandenberger a, Oliver Kröcher a, , Alexander Wokaun , Arno Tissler , Roderik Althoff
*
a Paul Scherrer Institute, 5232 Villigen PSI, Switzerland
b Süd-Chemie AG Waldheimer Str. 15, 83052 Bruckmühl, Germany
a r t i c l e i n f o
a b s t r a c t
Article history:
The selective catalytic reduction of NO with NH3 has been studied over Fe-ZSM-5 samples with equal
exchange degrees and active sites but different Brønsted acidities. Additionally, the activity of fresh
and aged samples in NH3–SCR with NO (standard SCR) has been compared with that in NH3–SCR with
NO/NO2 (fast SCR), where oxidation activity is not essential.
Received 30 June 2009
Revised 28 August 2009
Accepted 29 September 2009
Available online 29 October 2009
Using NH3–TPD–FTIR experiments and DRIFTS measurements, it was shown that the concentration of
Brønsted acid sites decreases with an increase in Fe loading, suggesting that, up to Fe/Al = 0.39, about one
Brønsted acid site is substituted by one Fe site and that most of the iron is present as a dispersion of cat-
ions. A maximum of about 55% of all Brønsted acid sites can be exchanged in aqueous Fe2+ solution.
These experiments revealed that the acidity of the catalyst is not a crucial factor for high activity and
that Brønsted acidity may not be required for adsorbing or activating the ammonia, but it is necessary to
bind and disperse the metal ions. On the other hand, the oxidation activity of Fe-ZSM-5 was found to be
the main factor that controls the high activity. The altered oxidation activity is largely responsible for the
decreased SCR activity after hydrothermal aging.
Keywords:
NH3–SCR chemistry
Selective catalytic reduction of NO with
ammonia
Metal-exchanged zeolites
Fe-ZSM-5
Active site
Exchange degree
Activity
Ó 2009 Elsevier Inc. All rights reserved.
Fast SCR
Standard SCR
1. Introduction
(with x = 1, 2) [7,11,12], which is the same key compound proposed
by Sun et al. [13] and Yeom at al. (x = 1) [14]. Li et al. [15] showed
Despite several studies on the structure–activity relationship of
Fe-ZSM-5 catalysts for the selective catalytic reduction of nitrogen
oxides (NOx, x = 1, 2) with ammonia (SCR), the debate over the nat-
ure of active sites in this catalyst is ongoing. This applies to the nat-
ure of the iron center, which we have investigated in another study
[1]) as well as to the role of the Brønsted acid sites in the SCR
mechanism. It is well known that for the SCR of NO (standard
SCR) over Fe-ZSM-5, part of the NO first has to be oxidized to ad-
sorbed NO2 species, which is the rate-determining step of the
mechanism. This takes place at iron centers with unknown nucle-
arity [2–7]. However, it is also generally assumed that Brønsted
acidity is required for binding and activating the ammonia, thereby
forming ammonium ions which react further with the surface NOx
species to form N2 and H2O [3,6,8–10].
that the SCR reaction is very efficiently catalyzed by zeolites due to
the protonation of the intermediate [NH4NO2] complex in the SCR
cycle by the Brønsted acid sites, leading to an intermediate product
that decomposes even at room temperature. On the other hand,
very recently Schwidder et al. [16] showed that even with Fe-
ZSM-5 catalysts without Brønsted acidity, a quite large reaction rate
could be achieved for standard SCR between about 250 and 650 °C.
These results support the existence of a promoting effect of
Brønsted acid sites, but also indicate that the Brønsted acidity
may not be indispensable.
Other studies of the SCR mechanism on a broad variety of mate-
rials point in the same direction. Results from Peña et al. [17]
showed that Brønsted acidity is not necessary for the SCR of NO
over different metal oxides on TiO2 at low temperature, but that
Lewis acidity is required. Amores et al. [18] investigated TiO2 cat-
alysts supported with different metal oxides, and concluded from
their results that NHþ4 ions are not involved in standard SCR. This
is in line with a study of this reaction by Liang et al. [19] demon-
strating the importance of Lewis acidity for high activity over
Mn/Al-SBA-15 molecular sieves as well as the minor impact of
Brønsted acidity. Liu et al. [20] compared the reactivity of iron
impregnated Al2O3, TiO2, and BEA zeolite and found the order of
2Z ꢀ NHþ4 þ NO þ NO2 ! 2N2 þ 3H2O þ 2Z ꢀ Hþ
ðwith Z ¼ zeoliteÞ
ð1Þ
It has been proposed that once NH3 has been activated, NO2 reacts
with two adjacent NHþ4 ions to form an active complex, [(NH4)xNO2]
* Corresponding author. Fax: +41 56 310 23 23.
E-mail address: oliver.kroecher@psi.ch (O. Kröcher).
0021-9517/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2009.09.028

298
S. Brandenberger et al. / Journal of Catalysis 268 (2009) 297–306
SCR performance to be Fe/BEA ꢁ Fe/TiO2 > Fe/Al2O3, which is in
the same order as ammonia adsorption capacity. Remarkably, no
evidence that the presence of either Brønsted or Lewis acidity is
a key factor for the activation of ammonia was found. The authors
suggested that the form in which ammonia is held by the support
is not crucial and assumed that the support acts mainly as a reser-
voir for ammonia, which can then migrate to the active site for
reaction with NOx.
In the present study, we tried to obtain a more detailed picture
of the role of Brønsted acidity in the SCR of NO by ammonia in
Fe-ZSM-5 catalysts. For this purpose, catalysts with different
Brønsted acidities, but with identical structure and active site
concentration were studied. In contrast to previous investigations,
a silanization agent was used instead of simple alkali or earth
alkaline metals to specifically reduce the Brønsted acidity without
hampering the iron sites of the catalyst significantly. Additionally,
hydrothermally aged Fe-ZSM-5 samples were used as comparison
materials with both reduced Brønsted acidity and lowered oxida-
tion activity.
the literature in order to obtain low-loaded zeolites, inevitably lead
to a non-homogeneous metal distribution inside the zeolite. This
may be because iron ions will be quantitatively and strongly fixed
on the periphery of the grains [22] in these types of exchange
conditions.
The catalyst with a reduced Brønsted acidity, denoted as Fe-
*
ZSM-5 (0.04) , was prepared by exchanging a hydrothermally
pre-treated NH4–ZSM-5 as a starting material (650 °C, 10% H2O
for 1.5 h).
Fe-ZSM-5 (0.3) was aged under hydrothermal conditions (10%
H2O in air) for 8 h at 650 °C and 800 °C, respectively, in order to
prepare samples with reduced Brønsted acidity as well as lowered
redox activity.
A silanized Fe-ZSM-5 catalyst was prepared using triisopropyl-
chlorosilane (TIPCS) as the silanization reagent. A superstoichio-
metric amount of the organosilane was added to a zeolite/
toluene suspension under nitrogen in a glove box. This mixture
was stirred under reflux overnight under nitrogen, then filtered
and washed with toluene.
We used liquid ion exchange for the preparation of the Fe-ZSM-
5 samples, whereby special attention was paid to controlling the
iron concentration during the exchange process and to keeping
the exchange times constant. Our preparation differs from the
standard liquid ion exchange methods by the use of ammonium
chloride to control the exchange degree, which yields in particu-
larly homogenously exchanged Fe-ZSM-5 samples.
The highly exchanged sample CDV–Fe/ZSM-5 (0.74) was pre-
pared via the chemical vapor ion exchange (CVD) method de-
scribed by Chen and Sachtler [23], since it is not possible to
prepare Fe-ZSM-5 samples with an exchange degree higher than
Fe/Al ꢃ 0.5 by liquid ion exchange. NH4–ZSM-5 was converted
into the acidic form by calcination in air at 450 °C for 2.5 h. A
U-shaped quartz reactor was used for the CVD exchange. The
reactor was built with a porous frit in each of the legs, allowing
the separation of the H-ZSM-5 support and the FeCl3 precursor.
In order to remove adsorbed moisture, 15 g of the H-ZSM-5 was
loaded onto one frit and flushed with nitrogen (100 mL/min) for
1 h at 380 °C. Afterwards, an appropriate amount of anhydrous
FeCl3 (98%, Fluka) was loaded on the second frit and the reactor
was heated to 350 °C for 30 min under N2 (35 mL/min). The
resulting Fe-ZSM-5 material was washed with 200 mL/g deion-
ized water, dried overnight in air at 80 °C, and calcined at
500 °C for 3 h under nitrogen.
For activity measurements, the catalytic materials were wash-
coated on 40.5 mm ꢄ 17.6 mm ꢄ 12.8 mm cordierite monoliths
with a cell density of 400 cpsi. The respective slurries were com-
posed of a liquid phase of 1.5 mL Ludox AS-40 in 40 mL deionized
water and a solid phase of approximately 10 g of Fe-ZSM-5. The
monoliths were coated with the zeolites by immersing the mono-
liths in the slurries, blowing out the excess slurry, and drying them
in an air flow at 110 °C for 1 min. This procedure was repeated un-
til the monoliths were coated with 0.8–1.0 g zeolite, corresponding
to approximately 100 gcat./L. Finally, the coated monoliths were
calcined at 550 °C in air for 3 h.
2. Experimental
2.1. Preparation of the samples
Fe-ZSM-5 samples (Table 1) were prepared by liquid ion ex-
change of NH4–ZSM-5 (Süd-Chemie AG, Si/Al atomic ratio = 13.5).
Within this work, the samples are labeled as Fe-ZSM-5 (n), where
n is the exchange degree Fe/Al. For the ion exchange, 5–15 g/L of
FeCl2ꢂ4H2O and 2–20 g of ammonium chloride, depending on the
desired exchange degree, were added to a suspension of NH4–
ZSM-5 in water. After it was stirred overnight under nitrogen at
80 °C, the resulting Fe-ZSM-5 material was filtered, washed with
200 ml/g of deionized water, dried overnight in air at 80 °C, and
calcined at 500 °C for 3 h under nitrogen. The addition of ammo-
nium chloride to the reaction mixture shifted the equilibrium of
the exchange reaction toward free solvated metal salts. Thus,
micropore diffusion was retained and the dispersion of the metal
improved [21], resulting in a more homogeneous distribution of
the exchange metal. In our opinion, very low concentration ex-
change solutions or short exchange times, as is often described in
Table 1
Sample characteristics of the parent H-ZSM-5 material and the prepared Fe-ZSM-5 catalysts.
Sample
Brønsted acidity (NH3 desorp.
during TPD at ꢃ490 °C) (mol/kg)
Exchange
Brønsted acidity relative
to H-ZSM-5 (DRIFTS)
Total NH3 adsorption capacity
(total NH3 desorp. during TPD) (mol/kg)
Fe (wt.%)a
(exchange degreea)
ratio H+/Feb
H-ZSM-5
0.95
n/a
–
1
3.1
n/a
1.8
3
2.9
2.7
2.8
2.6
2.0
2.2
1.2
0.6
–
Fe-ZSM-5 (0.04)
n/a
n/a
1.0
1.1
0.9
1.1
0.8
0.4
–
0.89
0.22
0.83
0.71
0.6
0.52
0.38
0.43
0.04
0.007
0.006
0.27
0.27
0.57
1.1
1.7
2
2.6
5
2
Fe-ZSM-5 (0.04)d
Fe-ZSM-5 (0.08)
0.49
0.85
0.74
0.67
0.56
0.56
0.48
n/a
Fe-ZSM-5 (0.16)
Fe-ZSM-5 (0.25)
Fe-ZSM-5 (0.3)
Fe-ZSM-5 (0.39)
Fe-ZSM-5 (0.74)
Fe-ZSM-5 (0.3)–TIPCS
Fe-ZSM-5 (0.3), 650 °C, 8 h
Fe-ZSM-5 (0.3), 800 °C, 8 h
0.21
–
–
2
2
0.07c
a
b
c
Iron content measured by ICP–AES.
Calculated by dividing the molar amount of consumed Brønsted acidity (measured by TPD) by the molar amount of exchanged iron (measured by ICP–AES).
Inaccurate value due to very small peak.
Prepared by exchanging hydrothermally pretreated H-ZSM-5 with reduced Brønsted-acidity.
d

S. Brandenberger et al. / Journal of Catalysis 268 (2009) 297–306
299
2.2. Catalyst characterization and activity measurements
350
300
250
200
150
100
50
(a)
Inductively coupled plasma–atomic emission spectroscopy
(ICP–AES) was employed in the quantitative analysis of the ele-
mental composition of all samples. The results of this analysis
are shown in Table 1.
H-MFI-27
0.08
0.16
The Brønsted acidity was measured by FTIR spectroscopy and
temperature-programed desorption (TPD) of ammonia (Table 1).
Infrared spectra of the zeolite surface were recorded with a Nicolet
Magna IR 560 spectrometer equipped with an in situ DRIFT gas cell.
The samples were dried in the cell under 20% oxygen at 550 °C for
10 min, under flowing nitrogen for another 10 min and then the
spectra were recorded at 150 °C. The concentration of Brønsted
acid sites was determined from the area of the band at about
3610 cmꢀ1. By feeding NH3 over the zeolite, all Brønsted acid sites
were saturated, thereby creating a background sample without
acidity but with the same particle size distribution, particle shape,
and packing densities. Very accurate measurements could be per-
formed with this procedure since the background was made of
the sample, which prevented changes in the scattering of light.
For TPD measurements, 100 mg of the sample was heated at
550 °C for 10 min in a flow of 20% O2 and then cooled to RT under
He. NH3 was adsorbed for 1 h at RT from a gas stream containing
1% NH3 in nitrogen, followed by the actual desorption experiment
at a heating rate of 10 °C/min from 30 °C to 800 °C in a He flow. The
concentration of desorbed NH3 in the gas phase was analyzed with
a Nicolet Magna IR 560 spectrometer equipped with a heated gas
measuring cell and a liquid nitrogen cooled MCT detector. The
overlapping peaks were deconvoluted into Gaussian functions
and the absolute number of Brønsted acid sites was quantified by
multiplying the area of the peak at about 490 °C with the feed
gas flow rate.
Exchange
degree
0
0
200
400
600
800
Temperature [°C]
350
300
250
200
150
100
50
(b)
0.25
0.3
0.39
0.74
Exchange
degree
0
The measurements of the SCR activities as well as of the NO and
CO oxidation activities were carried out in a heated quartz reactor.
The concentrations of the gaseous reaction products were deter-
mined by FTIR spectroscopy. Details of the setup of the test appa-
ratus and the gas analysis method have been reported previously
[24]. For SCR measurements, the composition of diesel exhaust
gas was approximated by a model feed gas containing 10% O2, 5%
H2O, 1000 ppm NH3 and 1000 ppm NO (standard SCR), or
500 ppm NO and 500 ppm NO2 (fast-SCR) and N2 balance. NO oxi-
dation was investigated using a feed gas consisting of 10% O2, 5%
H2O, 1000 ppm NO, and N2 as the balance. For CO oxidation activity
measurements, the gas feed consists of 2000 ppm CO and 15% O2
with the remaining balance made up by N2. The gas hourly space
velocity (GHSV) was 52,000 hꢀ1 for feeds containing NO (standard
SCR) up to 450 °C and experiments with CO. At 500 °C, standard
SCR was carried out with a GHSV of 200,000 hꢀ1. The GHSV was
280,000 hꢀ1 for all fast SCR measurements.
0
200
400
600
800
Temperature [°C]
Fig. 1. Measured NH3 concentration during TPD for (a) H-ZSM-5 and Fe-ZSM-5
(0.08, 0.16) and (b) Fe-ZSM-5 (0.25, 0.3, 0.39, and 0.74).
the interaction of Brønsted acid sites with extra-framework alu-
minum species [31].
Fig. 1a and b show that the ion exchange results in a reduction
of Brønsted acidity. At the same time the peaks in the temperature
range below 300 °C tend to become larger with increasing iron
content, which might be explained by the creation of new Lewis
acid sites [32]. This interpretation is hampered by an experimental
error which is significant for weakly adsorbed NH3 species below
300 °C. The height of these peaks was affected by the experimental
conditions, which were slightly different from run to run in the
used TPD instrument. Interestingly, the maximum of the peak at
about 490 °C shifts to lower temperature with increasing iron con-
tent. This effect may be explained by decreasing heats of adsorp-
tion [26] with increasing iron content and decreasing Brønsted
acidity. The change of heat of adsorption, however, is not related
to the NH3 surface coverage, because Hunger et al. [25] measured
slightly increasing heats of desorption for decreasing surface cov-
erages on H-ZSM-5. The peak shifting is rather explained by differ-
ent types of Brønsted sites with slightly different acidities and
affinities to bind iron. An alternative explanation would be the cre-
ation of new Lewis acid sites during ion exchange, desorbing NH3
at about 450 °C. Regardless of the reason for the peak shift, the
overlapping peaks were deconvoluted into Gaussian functions
and the absolute number of Brønsted acid sites was quantified by
multiplying the area of the peak at about 490 °C with the feed
gas flow rate (Table 1). For H-ZSM-5, a Brønsted acid site concen-
tration of 0.95 mol/kg was calculated, which is about 21% lower
3. Results
The NH3 desorption profiles during TPD from H-ZSM-5 and
Fe-ZSM-5 SCR catalysts with different exchange degrees were
compared (Fig. 1a and b). All catalysts were prepared by liquid
ion exchange, except for the sample with Fe/Al = 0.74, which
was prepared by CVD. The measured amounts of desorbed NH3
are given in Table 1. NH3 desorbed in three main peaks at about
140 °C, 230 °C and 490 °C. The peaks observed at 140 °C and
230 °C are caused by weakly adsorbed NH3 on Lewis acid sites,
while the peak at about 490 °C is related to NH3 adsorbed on
Brønsted acid sites [25–28]. The cause of the relatively small
peak at about 650 °C is unclear, but it might be due to strong Le-
wis acid sites [29,30]. Alternatively this peak could account for
especially strong Brønsted acid sites, which are formed from

300
S. Brandenberger et al. / Journal of Catalysis 268 (2009) 297–306
than expected based on the Si/Al ratio of 13.5 in this material. This
result is in line with those obtained from conductometric titration
[33] and another NH3–TPD study [34]. This might be explained by
some dealumination that occurred during the previous calcination
of the zeolite, resulting in a loss of Brønsted acidity [28,33].
The exchange ratio H+/Fe in Table 1, with H+ being the molar
amount of consumed Brønsted acidity, shows that about one
Brønsted acid proton is replaced by one Fe atom up to an exchange
degree of Fe/Al = 0.39, which is in line with the literature values
[35]. In contrast, the highest loaded sample (Fe/Al = 0.74) showed
a smaller H+/Fe ratio, suggesting that above Fe/Al ꢃ 0.4, a part of
the iron formed FexOy particles or clusters instead of replacing H+.
The concentration of Brønsted acid sites was also measured by
FTIR spectroscopy in order to confirm the TPD results (Fig. 2). The
band at 3610 cmꢀ1 was assigned to the OH stretch of Brønsted acid
sites [36], while the band at 3660 cmꢀ1 arose from OH groups asso-
ciated with extra-framework Al [36,37]. The relative concentration
of Brønsted acid sites, expressed as H/Al, was calculated by nor-
malizing the peak area of the band at 3610 cmꢀ1 for the Fe-ZSM-
5 samples to the peak area observed for H-ZSM-5 (Fig. 3). In
Fig. 3, the exchange degree (Fe/Alav) was calculated based on the
concentration of Brønsted acid sites that are available for ion ex-
change (Alav) rather than the concentration of Al determined by
ICP. The measured concentration of the available Brønsted acid
sites in H-ZSM-5 is 0.95 mol/kg, as determined by TPD (data shown
in Table 1). The regression line of the data shown in Fig. 3 has a
slope of ꢀ1.1, proving the conclusion from Fig. 1a and b that
approximately one Brønsted proton is exchanged for one Fe atom.
However, the Fe-ZSM-5 (0.74) sample with Fe/Alav = 0.94 shows a
higher remaining Brønsted acidity (H/Al ꢃ 0.45) than expected on
basis of the exchange degree, revealing that a maximum of about
55% of all Brønsted acid sites can be exchanged in aqueous Fe2+
solution. This finding is consistent with the literature reports
[35,38]. Note that Fe-ZSM-5 (0.74) was exchanged by chemical va-
por deposition of FeCl3, but that it was washed and calcined after-
wards. In accordance with the literature [38], the FTIR signal of the
Brønsted acid sites disappeared completely after H-ZSM-5 had
been loaded with FeCl3 (data not shown) but restored after wash-
ing and calcination.
It is interesting to note that the data shown in Table 1 show
slightly different profiles for the decrease in the amount of
Brønsted acid sites measured by FTIR spectroscopy and NH3–TPD.
NH3–TPD resulted in a slightly lower decrease of the Brønsted acid-
ity with rising iron content compared to FTIR spectroscopy.
Although the difference is only small, it might be explained by
the experimental error or, alternatively, by the creation of new
1.0
H-ZSM-5
Fe/Al
0.08
0.9
0.8
0.7
0.16
0.25
Exchange degree
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.3
0.39
0.74
3750
3700
3650
3600
3550
3500
3450
Wavenumber [cm-1]
Fig. 2. FTIR spectra of the hydroxyl stretching region for Fe-ZSM-5 samples with various exchange degrees.
1.0
0.9
0.8
0.7
Slope: -1.1
0.6
0.5
0.4
0.3
0.2
0.1
0.0
0.0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
.0
1
Fe/Al
av
Fig. 3. Change of H/Al ratio with Fe/Alva ratio for Fe-ZSM-5 samples with various exchange degrees, with Alva corresponding to the concentration of available Brønsted acid
sites determined in Table 1.

S. Brandenberger et al. / Journal of Catalysis 268 (2009) 297–306
301
Lewis acid sites in the metal exchange process, which desorb NH3
in the temperature region used for the quantification of Brønsted
acid sites.
and oxygen and first-order with respect to NO [6,39–42] under the
testing conditions applied in this study, the SCR activity may be ex-
pressed quantitatively by an apparent first-order rate constant.
This can then be used to visualize the dependence of the activity
on exchange degree and remaining Brønsted acidity. When assum-
ing plug-flow reactor conditions and the absence of diffusion lim-
itation, which hold for low and intermediate temperatures, the
apparent first-order rate constant may be calculated from the NO
Fig. 4 shows the NO reduction efficiencies of the Fe-ZSM-5 cat-
alysts with various exchange degrees (Fe/Al). In the high-tempera-
ture range above 500 °C, the conversion level decreases with rising
iron content (with the exception of the zeolite with Fe/Al = 0.74),
whereas in the low temperature range below 350 °C the NO con-
version increases with rising iron content (with the exception of
the zeolite with Fe/Al = 0.74) but decreasing Brønsted acidity. Since
the SCR reaction is independent of the concentrations of ammonia
conversion (X) by:
Å
V
m
k ¼ lnð1 ꢀ XÞ
ð1Þ
Å
where V is the gas flow rate at the reaction temperature and m is
the catalyst mass.
100
Fe/Al
90
Fig. 5 illustrates that at 250 °C, the activity of the catalysts in-
creases with rising iron content about fivefold between Fe/
Al = 0.04 and 0.39, whereas the number of Brønsted acid sites de-
creases significantly.
80
70
60
50
40
30
20
10
0
0.74
0.39
0.3
0.16
0.08
0.04
H-ZSM-5
Fig. 6 compares the SCR activity of two catalysts, both with an
exchange degree of Fe/Al = 0.04, but with different Brønsted acidi-
ties. The Brønsted acidity of the Fe-ZSM-5 samples was previously
determined by DRIFTS to be 89% for sample Fe-ZSM-5 (0.04) and
22% for sample Fe-ZSM-5 (0.04)*, which was related to the acidity
of the parent H-ZSM-5. Both catalysts exhibit equal SCR activity,
indicating that Brønsted acidity might be not decisive for a high
NO conversion. It should be noted that from NH3–TPD (Fig. 7) the
sample Fe-ZSM-5 (0.04)* had an apparent higher remaining
Brønsted acidity of about 50% relative to the parent H-ZSM-5 (Ta-
ble 1), which might be explained by the creation of new Lewis acid
sites due to the hydrothermal aging.
200
300
400
500
600
700
Temperature [°C]
Fig. 4. NO reduction as a function of temperature over Fe-ZSM-5 catalysts with
various exchange degrees (Fe/Al). Reaction conditions: 1000 ppm NO, 1000 ppm,
NH3, 10% O2, 5% H2O, remaining gas feed consists of N2. GHSV = 52,000 hꢀ1
.
In order to further elucidate the influence of Brønsted acidity on
SCR activity, we compared the SCR activity of a fresh Fe-ZSM-5
(0.3) sample with the activity of different Fe-ZSM-5 samples with
equal exchange degree, but reduced Brønsted acidity. We reduced
the Brønsted acidity in two different ways: by aging the Fe-ZSM-5
(0.3) sample at either 650 °C or 800 °C under hydrothermal condi-
tions (10% H2O), and by poisoning the Brønsted acid sites by silan-
ization according to Eq. (2).
120
100
80
60
40
20
0
1
0.9
0.8
0.7
0.6
0.5
0.4
H
SiR3
O
0
0.2
0.4
Fe/Al ratio
0.6
0.8
-
ð2Þ
O
-
Si
Al
+
X-Si(R)3
Si
Al
+
HX
Fig. 5. First-order rate constants for the SCR reaction and Brønsted acidity as a
function of the exchange degree.
100
Fe-ZSM-5 (0.04)*
Fe-ZSM-5 (0.04)
80
60
40
20
0
200
300
400
500
600
700
Temperature [%]
Fig. 6. NO reduction as a function of temperature over two Fe-ZSM-5 samples with equal exchange degree (Fe/Al = 0.04), but with different Brønsted acidities: Fe-ZSM-5
*
(0.04) 89% and Fe-ZSM-5 (0.04) 22% relative to the acidity of the parent H-ZSM-5. Reaction conditions: 1000 ppm NO, 1000 ppm, NH3, 10% O2, 5% H2O, remaining gas feed
consists of N2. GHSV = 52,000 hꢀ1
.

302
S. Brandenberger et al. / Journal of Catalysis 268 (2009) 297–306
350
300
250
200
150
100
50
fresh
aged 650°, 8h
aged 800 °C, 8h
TIPCS
Fe-ZSM-5 (0.04)*
0
0
100
200
300
400
500
600
700
800
Temperature [°C]
Fig. 7. Measured NH3 concentration during TPD of Fe-ZSM-5 (0.04)* and Fe-ZSM-5 (0.3) in the fresh state, after silanization and after aging at 650 °C and 800 °C, respectively.
Both the Brønsted acidity and the number of active iron sites
were reduced by hydrothermal aging. The loss of Brønsted acidity
can thereby ascribed to dealumination and the loss of surface area
[3,43–47]. The NH3 spectrum shown in Fig. 7 shows that hydrother-
mal aging lowered the Lewis acidity as well as Brønsted acidity. The
loss of active iron sites is due to migration of the iron species out of
the ionic exchange sites. As a consequence, inactive metal-oxide
particles with reduced oxidation activity are formed [38,48–51].
In contrast to the hydrothermal aging, the silanization process
should only impact the Brønsted acid sites, whereas it is reasonable
(a) 100
90
80
Fe-ZSM-5
70
Fe-ZSM-5 (TIPCS)
60
50
40
30
20
10
0
200
250
300
Temperature [°C]
(b) 100
90
Fe-ZSM-5 (fresh)
80
70
60
50
40
30
20
10
0
Fe-ZSM-5 (aged, 650°C)
Fe-ZSM-5 (aged, 800 °C)
Fe-ZSM-5 (TIPCS)
200
220
240
260
280
300
Temperature [°C]
Fig. 8. (a) NO oxidation activity of fresh and silanized Fe-ZSM-5 (0.3). (b) NO reduction efficiency as a function of temperature over Fe-ZSM-5 catalysts with equal exchange
degree Fe/Al = 0.3, but with different Brønsted acidities ranging from 52% for the fresh sample, 4% after silanization with TIPCS, 0.7% after aging at 650 °C, and 0.6% after aging
at 800 °C, relative to the parent H-ZSM-5. Reaction conditions: 1000 ppm NO, 1000 ppm, NH3, 10% O2, without H2O, remaining gas feed consists of N2. GHSV = 52,000 hꢀ1
.

S. Brandenberger et al. / Journal of Catalysis 268 (2009) 297–306
303
to assume that the Lewis acid sites and active iron sites remain al-
most unaffected. Please note that the NH3–TPD of the silanized
sample shown in Fig. 7 does not reflect the proportion of the Lewis
acidity and Brønsted acidity correctly. During the TPD experiment,
the O–SiR3 bond is thermally degraded, which releases Brønsted
acid sites. The NH3 desorbing from Lewis acid sites may readsorb
on these freed sites. Niwa et al. have shown that this readsorption
of NH3 on Brønsted acid sites is possible under the applied exper-
imental conditions [52]. Consequently, the Brønsted acidity is in-
creased at the expense of the Lewis acidity.
However, we found the total NH3 adsorption capacities of Fe-
ZSM-5 (0.3) and Fe-ZSM-5 (0.3) TIPCS to be 2.8 and 2.2 mol/kg,
respectively (Table 1). The difference of 0.6 mol/kg between these
two values is equivalent to the 0.56 mol/kg of NH3 desorbed at
490 °C for Fe-ZSM-5 (0.3). This proves that the Brønsted acid sites
are consumed almost completely due to the silanization. Using
in situ DRIFT experiments, we could show that the O–SiR3 group
is stable up to about 300 °C under anhydrous SCR conditions. Using
DRIFT spectroscopy, we measured the Brønsted acidity relative to
the parent H-ZSM-5 to be 52% for the fresh sample, 4% after silan-
ization with TIPCS, 0.7% after aging at 650 °C, and 0.6% after aging
at 800 °C (Table 1). It is interesting to note that the NH3–TPD of the
aged 650 °C sample shows a considerable amount of NH3 desorb-
ing at about 490 °C, whereas the DRIFT measurement reports that
the Brønsted acid sites disappeared almost completely. This may
serve as evidence that some new Lewis acid sites may be formed
during the thermal aging process, which is a known process for
steam-treated alumina [31] that desorbs NH3 at about 490 °C.
To draw reliable conclusions from the experiments, it is impor-
tant that the iron sites remain unaffected by the silanization. In or-
der to prove that there is no migration of iron ions out of the ion
exchange sites due to the silanization, the SCR activity before silan-
ization was compared with the activity after silanization and the
subsequent hydrolysis of the O–SiR3 bond accomplished at
500 °C in an atmosphere with 5% H2O. The SCR activity was com-
pletely restored after hydrolysis (data not shown). Moreover, the
NO oxidation activity of the silanized and fresh Fe-ZSM-5 (0.3)
samples demonstrate that the oxidation activity remains almost
constant (Fig. 8a).
Fig. 8b shows the SCR activity of the fresh catalyst, the two aged
Fe-ZSM-5 samples and the Fe-ZSM-5 sample silanized with TIPCS.
In order to avoid hydrolysis of the Si–O bond, the SCR measure-
ments were performed in an anhydrous atmosphere. Remarkably,
the silanized sample with only 4% acidity, containing about 10
times more iron sites than Brønsted sites, showed only a slightly
lower NO conversion, reduced by about 20% compared to the fresh
Fe-ZSM-5 (0.3). Note that the formation of CO and CO2 is negligible
in all experiments (data not shown), which excludes oxidation of
the isopropyl group. The hydrothermally aged samples, which lost
about 99% of their Brønsted acidity, show a remarkably low activity
compared to the silanized sample, which lost about 96% of its
Brønsted acidity. This may suggest that the loss of redox active iron
100
(a)
90
Fresh
80
Aged, 650°C, 8h
70
Aged, 800°C, 8h
60
50
40
30
20
10
0
200
250
300
350
Temperature [°C]
(b)100
90
80
70
60
50
40
30
20
10
0
Fresh
Aged, 650°C, 8h
Aged, 800°C, 8h
200
250
300
350
Temperature [°C]
Fig. 9. NOx reduction as a function of temperature over Fe-ZSM-5 (0.3) catalysts with various exchange degrees (Fe/Al). (a) Standard SCR. Reaction conditions: 1000 ppm NO,
1000 ppm, NH3, 10% O2, 5% H2O, remaining gas feed consists of N2. GHSV = 52,000 hꢀ1. (b) Fast SCR. Reaction conditions: 500 ppm NO, 500 ppm NO2, 1000 ppm, NH3, 10% O2,
5% H2O, remaining gas feed consists of N2. GHSV = 280,000 hꢀ1
.

304
S. Brandenberger et al. / Journal of Catalysis 268 (2009) 297–306
sites is responsible for the decreased activity rather than the loss of
Brønsted acidity.
b at 250 °C and compared with the Brønsted acidity shown in
Fig. 10. Interestingly, the sample aged at 650 °C exhibited a very
similar Brønsted acidity, but significantly higher standard and fast
SCR activities than the sample aged at 800 °C. In order to under-
stand this observation the oxidation activities of the fresh and aged
catalysts have to be compared, but unlike the experiment de-
scribed in Fig. 8a, the NO oxidation activity cannot be used to as-
sess this difference. It seems that the NO2 produced from NO in
the oxidation process over the iron in the sample desorbs at differ-
ent rates from the various iron sites. For instance, NO2 adsorbs on
Fe sites attached to the zeolite framework but not on Fe-oxide par-
ticles [4,11,55,56]. As the iron sites are altered during aging, and
due to the fact that the rate of the NO oxidation reaction is con-
trolled by the desorption of NO2, [4] no meaningful comparison
of fresh and aged Fe-ZSM-5 samples could be made. Therefore,
we measured the CO oxidation activity because the interaction of
CO with Fe-ZSM-5 is very weak and the oxidation product CO2 is
not adsorbed, [57] which simplifies the interpretation. The results
shown in Fig. 11 illustrate that the evolution of CO oxidation activ-
ity follows the trend found in SCR activity.
With an NO/NO2 = 1/1 gas feed, the SCR of NO with NH3 reaches
the highest rate, and is therefore called fast SCR. At this point, the
oxidation activity of the catalyst is no longer required [5,53]. This
interesting feature of the fast SCR reaction gives us the opportunity
to measure SCR activities independent of the oxidation functional-
ity in the catalyst. In order to differ between the influence of oxida-
tion activity and Brønsted acidity, the standard SCR activity
(Fig. 9a) was compared with NOx conversion under fast SCR condi-
tions (Fig. 9b) before and after aging at 650 °C and 800 °C, respec-
tively. Fig. 9a shows that aging at 650 °C and 800 °C strongly
reduces the NO conversion efficiency, decreasing at 250 °C from
37.4 to 15.9 and 7.7% for the fresh, aged at 650 °C and aged at
800 °C samples, respectively. In contrast, Fig. 9b demonstrates that
under fast SCR conditions the NOx conversion is almost unchanged
after aging at 650 °C and only reduced by 15% after aging at 800 °C.
At 200 °C, the fast SCR conversion for the fresh sample is decreased
to some extent. It is worth mentioning that we observed a high ini-
tial SCR activity when measuring this data point, which, however,
decreased within 15 min to the final steady-state activity reported
in Fig. 9b. This observation is in line with the accumulation of
NH4NO3 deposits in the pores of Fe-ZSM-5 already at 200 °C, there-
by blocking the active sites [2,54]. The reaction rate constants were
calculated from the measured NOx conversion shown in Fig. 9a and
4. Discussion
The NH3–TPD experiments (Fig. 1a and b) and DRIFTS measure-
ments (Fig. 2) show that the concentration of Brønsted acid sites
decreases with an increase in Fe loading, suggesting thereby that
up to Fe/Al = 0.39, about one Brønsted acid site is substituted by
one Fe site (Fig. 3). This let us conclude that most of the iron is dis-
persed as cations and that the Fe complexes, located in the ion ex-
change sites, have in average a single positive charge which is in
line with the results from Marturano et al. [58].
1
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
k, relative (250 °C)
k, relative (250 °C), Fast SCR
Brønsted-acidity
Fig. 4 shows that at T > 500 °C, the NO reduction efficiencies of
Fe-ZSM-5 catalysts decrease with rising iron content (with the
exception of the zeolite with Fe = 0.74). This is related to more pro-
nounced NH3 oxidation activity with rising iron content leading to
a decreased NO reduction efficiency. The oxidation activity goes
thereby trough a maximum for sample Fe-ZSM-5 (0.39) and de-
creases again with rising iron content for sample Fe-ZSM-5 (0.74)
because of the formation of higher clustered iron species, which
are less active in NH3 oxidation as we have proposed elsewhere
[1]. At temperatures below 350 °C, NO conversion increases with
rising iron content (with the exception of the zeolite with
Fe = 0.74) and decreasing Brønsted acidity. This is in line with
the expected trend that the SCR activity is a function of the iron
concentration as well as the assumption that the oxidation of NO
to NOa2ds over the iron sites is the rate-limiting step. The relatively
low activity of Fe-ZSM-5 (0.74) can be explained by the low con-
centration of isolated iron ions, responsible for NO oxidation at
low temperature as proposed from us in another study [1].
The consequently rising activity with decreasing Brønsted acid-
ity shown in Fig. 5 may interpreted in such a way that a high
Brønsted acidity is not decisive for high activity. This assumption
is supported by the results shown in Fig. 6: It is clearly seen that
over the whole temperature range both catalysts show the same
SCR activity, although the Brønsted acidity of sample Fe-ZSM-5
0.016
0.014
Fresh
Aged, 650°C, 8h
Aged, 800°C, 8h
Fig. 10. Relative Brønsted acidity and relative first-order rate constants for the
standard SCR reaction and fast SCR reaction, calculated from Fig. 9 for fresh and
650 °C and 800 °C aged Fe-ZSM-5 (0.3).
1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
k, relative SCR (500 °C)
k, relative CO oxidation (500 °C)
*
(0.04) is only about 1/4 of the acidity of sample Fe-ZSM-5
(0.04). By choosing a very low exchange degree, the formation of
isolated iron ions was strongly favored in both catalysts [59–61],
which ensures that both catalysts contain Fe sites of the same
structure regardless of the Brønsted acidity. One may argue that
for high activity just low Brønsted acidity is required because
NH3 is known to adsorb strongly on the zeolite surface [6]. More-
over, there are still about five times more Brønsted sites than iron
Fresh
Aged, 650°C, 8h
Aged, 800°C, 8h
Fig. 11. CO oxidation activity at 500 °C and relative first-order rate constants for the
standard SCR reaction at 500 °C for fresh and 650 °C and 800 °C aged Fe-ZSM-5
(0.3).
*
sites in sample Fe-ZSM-5 (0.04) . In this light, the 22% of Brønsted
acidity remaining in Fe-ZSM-5 (0.04)* relative to the parent H-

S. Brandenberger et al. / Journal of Catalysis 268 (2009) 297–306
305
ZSM-5 material may be still sufficient. However, Fig. 8b shows that
a silanized sample exhibits only a slightly decreased SCR activity
with respect to the fresh sample even though the Brønsted acidity
is depleted by about 96% (Table 1). This result implies that
Brønsted acidity is indeed of minor importance for SCR activity.
The slightly lower activity of the silanized sample relative to the
fresh one cannot be caused by an alteration of the redox site by sil-
anization because the NO oxidation activity remains almost unaf-
fected (Fig. 8a). Furthermore, the original SCR activity was
restored after hydrolysis of the O–SiR3 bond (data not shown). A
more likely explanation is that either the diffusion of the reactant
is hindered due to the large isopropyl groups, or that there is a
slightly promoting effect of Brønsted acid sites in the SCR reaction
as recently proposed by Schwidder et al. [16].
Fig. 8b also shows that the aged samples show a remarkable
lower activity compared to the silanized sample even though both
lost Brønsted acidity to a similar extent (ꢃ96% for Fe-ZSM-5 TIPCS
and ꢃ99% for aged samples). During hydrothermal aging, the
Brønsted acidity and the number of active iron sites were reduced,
but the silanization only affects the Brønsted acidity, whereas the
oxidation activity can be assumed to be unaffected. Therefore,
the activity pattern shown in Fig. 8b indicates that the primary fea-
ture affected by hydrothermal aging is the redox activity, which is
of particular importance for the SCR reaction. The comparison of
NOx conversion under standard SCR conditions (Fig. 9a) and fast
SCR conditions (Fig. 9b) before and after aging is fully compatible
with this assumption. After aging at 650 °C and 800 °C, the effi-
ciency under standard SCR conditions strongly decreases, whereas
at NO/NO2 = 1/1, where oxidation activity is not required, the effi-
ciency is almost unchanged after aging at 650 °C and only slightly
decreased after aging at 800 °C.
The change in activity can be expressed by a relative compari-
son of the apparent first-order rate constants in order to make
the situation clearer (Fig. 10). The comparison shows that the NO
conversion during standard SCR decreases by about 63% after aging
at 650 °C, but the performance in fast SCR is almost unchanged and
is reduced by only 7% after the hydrothermal treatment even
though the Brønsted acidity disappeared almost completely. More-
over, during aging at 800 °C, the Brønsted acidity remained almost
constant relative to the acidity change during aging at 650 °C, but
the standard SCR as well as the fast SCR activity still decreases.
Both observations support the hypothesis that Brønsted acidity is
of minor importance, but that the oxidation activity may be of ma-
jor importance or even define standard SCR activity. This conclu-
sion agrees with the results shown in Fig. 11; revealing that at
500 °C the evolution of the CO oxidation activity follows the trend
of the SCR activity. Even though this interpretation is somewhat
speculative because there might be different active sites responsi-
ble for CO oxidation and NO oxidation, this result provides evi-
dence that altered oxidation activity is one of the main factors
responsible for decreased SCR activity after hydrothermal aging
of the catalysts.
Of course, the presence of Brønsted acid sites is necessary to
bind and disperse the metal ions. They may also prevent the aggre-
gation of exchanged metal ions. However, both weak and strong
Lewis acid sites in the zeolite framework might act as an adsorp-
tion site for NH3 and therefore as a reservoir of the reductant. Ta-
ble 1 shows that the total NH3 adsorption capacity is still 1.2 mol/
kg after aging at 650 °C and 0.6 mol/kg after aging at 800 °C.
Another interesting conclusion of our study is that the cooper-
ative contribution of two adjacent Brønsted acid sites in the
NH3–SCR of NO, as mentioned in Section 1, is cast into doubt and
may even be impossible. The probability of finding two acid sites
in proximity is very small in ZSM-5 zeolites with reduced Brønsted
SCR reaction, the SCR activity of catalysts with reduced Brønsted
acidity would be expected to be very low. However, our experi-
ments found no such effect.
5. Conclusion
Fe-ZSM-5 samples with equal metal exchange degrees, but with
different Brønsted acidities have been prepared in such a way that
the active iron sites can be assumed to be identical. Moreover, we
reduced the Brønsted acidity in two different ways: by poisoning
the Brønsted acid sites by silanization and by aging the Fe-ZSM-5
(0.3) sample at 650 °C or 800 °C under hydrothermal conditions
(10% H2O). The activity of fresh and aged samples with reduced
Brønsted acidity and a reduced concentration of active iron sites
have been compared under standard SCR conditions. We also per-
formed a similar investigation under fast SCR conditions, where
oxidation activity provided by the iron ions is not required. Using
NH3–TPD–FTIR and DRIFT spectroscopy, we measured the Brønsted
acidity of the samples and found both methods well suited to
quantify the acidity in an accurate manner.
Our experimental results allow us to conclude the following:
ꢅ Brønsted acidity is not required for adsorbing or activating the
ammonia, but it is necessary in order to bind and disperse the
metal ions. Thereby, a maximum of about 55% of all Brønsted
acid sites can be exchanged in aqueous Fe2+ solution.
ꢅ During liquid ion exchange about one Brønsted acid proton is
replaced by one Fe atom up to an exchange degree of Fe/
Al = 0.39, proving the monovalence of the exchanging iron
complexes.
ꢅ At low temperatures, the acidic function of the Fe-ZSM-5 is not a
prerequisite for high activity in the SCR reaction of NO with NH3.
ꢅ The form in which ammonia is held by the support is not crucial
and the support acts mainly as a reservoir for ammonia, which
then migrates to the active site in order to undergo a reaction
with NO.
ꢅ The oxidation function of the Fe-ZSM-5 is one of the main fac-
tors that control the activity in the standard SCR reaction.
ꢅ Altered oxidation activity is responsible for the decreased SCR
activity after hydrothermal aging.
References
[1] S. Brandenberger, O. Kröcher, A. Wokaun, A. Tissler, R. Althoff, Appl. Catal. B, in
preparation.
[2] S. Brandenberger, O. Kröcher, A. Tissler, R. Althoff, Catal. Rev.-Sci. Eng. 50
(2008) 492.
[3] K. Rahkamaa-Tolonen, T. Maunula, M. Lomma, M. Huuhtanen, R.L. Keiski, Catal.
Today 100 (2005) 217.
[4] G. Delahay, D. Valade, A. Guzman-Vargas, B. Coq, Appl. Catal. B 55 (2005) 149.
[5] M. Devadas, O. Kröcher, M. Elsener, A. Wokaun, N. Soger, M. Pfeifer, Y. Demel, L.
Mussmann, Appl. Catal. B 67 (2006) 187.
[6] H.Y. Huang, R.Q. Long, R.T. Yang, Appl. Catal. A 235 (2002) 241.
[7] R.Q. Long, R.T. Yang, J. Catal. 207 (2002) 224.
[8] R.Q. Long, R.T. Yang, J. Am. Chem. Soc. 121 (1999) 5595.
[9] H. Sjovall, L. Olsson, E. Fridell, R.J. Blint, Appl. Catal. B 64 (2006) 180.
[10] A.L. Kustov, T.W. Hansen, M. Kustova, C.H. Christensen, Appl. Catal. B 76 (2007)
311.
[11] R.Q. Long, R.T. Yang, J. Catal. 198 (2001) 20.
[12] J. Eng, C.H. Bartholomew, J. Catal. 171 (1997) 27.
[13] Q. Sun, Z.X. Gao, H.Y. Chen, W.M.H. Sachtler, J. Catal. 201 (2001) 89.
[14] Y.H. Yeom, J. Henao, M.J. Li, W.M.H. Sachtler, E. Weitz, J. Catal. 231 (2005) 181.
[15] M. Li, Y. Yeom, E. Weitz, W.M.H. Sachtler, Catal. Lett. 112 (2006) 129.
[16] M. Schwidder, M.S. Kumar, U. Bentrup, J. Perez-Ramirez, A. Brückner, W.
Grünert, Microporous Mesoporous Mater. 111 (2008) 124.
[17] D.A. Pena, B.S. Uphade, P.G. Smirniotis, J. Catal. 221 (2004) 421.
[18] J.M.G. Amores, V.S. Escribano, G. Ramis, G. Busca, Appl. Catal. B-Environ. 13
(1997) 45.
[19] X. Liang, J.H. Li, Q.C. Lin, K.Q. Sun, Catal. Commun. 8 (2007) 1901.
[20] Z. Liu, P.J. Millington, J.E. Bailie, R.R. Rajaram, J.A. Anderson, Microporous
Mesoporous Mater. 104 (2007) 159.
*
acidity such as the silanized sample or in Fe-ZSM-5 (0.04) . In other
words, if two adjacent Brønsted acid sites were required for the

306
S. Brandenberger et al. / Journal of Catalysis 268 (2009) 297–306
[21] M. Che, O. Clause, C. Marcilly, Preparation of Solid Catalysts, in: G. Ertl, H.
Knozinger, J. Weitkamp (Eds.), Wiley, 1999.
[43] T. Sano, H. Ikeya, T. Kasuno, Z.B. Wang, Y. Kawakami, K. Soga, Zeolites 19
(1997) 80.
[22] J.A.Z. Pieterse, S. Booneveld, R.W. van den Brink, Appl. Catal. B 51 (2004) 215.
[23] H.Y. Chen, W.M.H. Sachtler, Catal. Today 42 (1998) 73.
[24] G. Madia, M. Koebel, M. Elsener, A. Wokaun, Ind. Eng. Chem. Res. 41 (2002)
4008.
[25] B. Hunger, J. Hoffmann, O. Heitzsch, M. Hunger, J. Therm. Anal. 36 (1990) 1379.
[26] N. Katada, H. Igi, J.H. Kim, M. Niwa, J. Phys. Chem. B 101 (1997) 5969.
[27] L. Rodriguez-Gonzalez, E. Rodriguez-Castellon, A. Jimenez-Lopez, U. Simon,
Solid State Ionics 179 (2008) 1968.
[44] W.K. Hall, X.B. Feng, J. Dumesic, R. Watwe, Catal. Lett. 52 (1998) 13.
[45] W. Lutz, B. Fahlke, U. Lohse, R. Seidel, Chem. Tech. (Leipzig) 35 (1983) 250.
[46] K. Ehrhardt, M. Suckow, W. Lutz, Hydrothermal decomposition of
aluminosilicate zeolites and prediction of their long-term stability, in: H.K.
Beyer (Ed.), Catalysis by Microporous Materials, Elsevier, London, 1995, p. 179.
[47] G. Engelhardt, U. Lohse, V. Patzelova, M. Magi, E. Lippmaa, Zeolites 3 (1983)
233.
[48] M. Devadas, O. Kröcher, M. Elsener, A. Wokaun, G. Mitrikas, N. Soger, M. Pfeifer,
Y. Demel, L. Mussmann, Catal. Today 119 (2007) 137.
[28] G. Bagnasco, J. Catal. 159 (1996) 249.
[29] V.R. Choudhary, D.B. Akolekar, J. Catal. 119 (1989) 525.
[30] G.H. Kuehl, H.K.C. Timken, Microporous Mesoporous Mater. 35–36 (2000) 521.
[31] J. Dwyer, A critical evaluation of the concept of Brönsted acidity related to
zeolites, in: P.J. Grobet, W.J. Mortier, E.F. Vansant (Eds.), Innovation in Zeolite
Materials Science, Elsevier, Amsterdam, 1988, p. 333.
[32] M. Mhamdi, S. Khaddar-Zine, A. Ghorbel, Appl. Catal. A-Gen. 337 (2008) 39.
[33] M. Crocker, R.H.M. Herold, M.H.W. Sonnemans, C.A. Emeis, A.E. Wilson, J.N.
Vandermoolen, J. Phys. Chem. 97 (1993) 432.
[49] E.J.M. Hensen, Q. Zhu, M. Hendrix, A.R. Overweg, P.J. Kooyman, M.V. Sychev,
R.A. van Santen, J. Catal. 221 (2004) 560.
[50] O. Kröcher, M. Devadas, M. Elsener, A. Wokaun, N. Soger, M. Pfeifer, Y. Demel,
L. Mussmann, Appl. Catal. B 66 (2006) 208.
[51] M. Devadas, O. Kröcher, A. Wokaun, React. Kinet. Catal. Lett. 86 (2005)
347.
[52] M. Niwa, N. Katada, M. Sawa, Y. Murakami, J. Phys. Chem. 99 (1995)
8812.
[34] Y.J. Li, J.N. Armor, J. Catal. 145 (1994) 1.
[53] G. Tuenter, W.F. Vanleeuwen, L.J.M. Snepvangers, Ind. Eng. Chem. Prod. Res.
[35] L.J. Lobree, I.C. Hwang, J.A. Reimer, A.T. Bell, J. Catal. 186 (1999) 242.
[36] A. Zecchina, S. Bordiga, G. Spoto, D. Scarano, G. Petrini, G. Leofanti, M. Padovan,
C.O. Arean, J. Chem. Soc. Faraday Trans. 88 (1992) 2959.
[37] L.M. Kustov, V.B. Kazansky, S. Beran, L. Kubelkova, P. Jiru, J. Phys. Chem. 91
(1987) 5247.
Dev. 25 (1986) 633.
[54] O. Kröcher, Minimization of NOx-Emissions Through Exhaust Aftertreatment,
in 1st MinNOx Conference, Berlin, Germany, 2007.
[55] A. Sierraalta, R. Anez, M.R. Brussin, J. Catal. 205 (2002) 107.
[56] G.D. Pirngruber, P.K. Roy, R. Prins, Phys. Chem. Chem. Phys. 8 (2006) 3939.
[57] A. Zecchina, M. Rivallan, G. Berlier, C. Lamberti, G. Ricchiardi, Phys. Chem.
Chem. Phys. 9 (2007) 3483.
[38] F. Heinrich, C. Schmidt, E. Löffler, M. Menzel, W. Grünert, J. Catal. 212 (2002)
157.
[39] T. Komatsu, M. Nunokawa, I.S. Moon, T. Takahara, S. Namba, T. Yashima, J.
Catal. 148 (1994) 427.
[40] H. Bosch, F.J.J.G. Janssen, Catal. Today 2 (1988) 369.
[41] M.D. Amiridis, F. Puglisi, J.A. Dumesic, W.S. Millman, N.Y. Topsoe, J. Catal. 142
(1993) 572.
[58] P. Marturano, L. Drozdova, A. Kogelbauer, R. Prins, J. Catal. 192 (2000) 236.
[59] M. Schwidder, M.S. Kumar, K. Klementiev, M.M. Pohl, A. Brückner, W. Grünert,
J. Catal. 231 (2005) 314.
[60] L. Capek, V. Kreibich, J. Dedecek, T. Grygar, B. Wichterlova, Z. Sobalik, J.A.
Martens, R. Brosius, V. Tokarova, Microporous Mesoporous Mater. 80 (2005)
279.
[42] R. Delahay, S. Kieger, N. Tanchoux, P. Trens, B. Coq, Appl. Catal. B 52 (2004)
251.
[61] L.B. Gutierrez, E.E. Miro, M.A. Ulla, Appl. Catal. A 321 (2007) 7.