The chemistry of the NO/NO2-NH3 fast SCR reaction over Fe-ZSM5 investigated by transient reaction analysis

Article abstract of DOI:10.1016/j.jcat.2008.03.027

We present a systematic study of the chemical steps in the NO/NO₂-NH₃ fast SCR reaction 2NH₃ + NO + NO₂ → 2N₂ + 3H₂O over a commercial Fe-ZSM5 catalyst. The study is based on transient reaction experiments at realistic conditions for removal of NO_x from mobile diesel exhausts. Its goal is to assess and critically evaluate the current ideas on the SCR mechanism, and also to establish to what extent the mechanistic pathways demonstrated for V-based catalysts also apply to Fe-promoted zeolites. Results show that the fast SCR reaction proceeds at low temperature via a global sequence involving NH₄NO₃ or related surface species as intermediates,2NO₂ + 2NH₃ → N₂ + NH₄NO₃ + H₂O,NO + NH₄NO₃ → NO₂ + N₂ + 2H₂O. Such a sequential scheme is the same as that proposed previously for the fast SCR chemistry over V-based catalysts and other zeolite catalysts and thus is considered a general mechanism. It explains all of the available observations for stoichiometry (e.g., optimum NO/NO₂ unit molar ratio), selectivity (e.g., N₂O from NH₄NO₃ decomposition), and kinetics (e.g., rate of fast SCR = rate of nitrate reduction by NO). We further show that the redox reaction between NO and nitrates is the rate-controlling step and is inhibited by ammonia. Remarkably, the same strongly enhanced deNO_x activity observed in the fast SCR reaction also was observed in the absence of gaseous NO₂ but in the presence of surface nitrates. We accordingly propose a general summary of the fast SCR chemistry over V-based and zeolite catalysts that emphasizes the key role of surface nitrates.

Full text of DOI:10.1016/j.jcat.2008.03.027

Journal of Catalysis 256 (2008) 312–322
Contents lists available at ScienceDirect
Journal of Catalysis
www.elsevier.com/locate/jcat
The chemistry of the NO/NO2–NH3 “fast” SCR reaction over Fe-ZSM5 investigated
by transient reaction analysis
Antonio Grossale a, Isabella Nova a, Enrico Tronconi a,∗, Daniel Chatterjee b, Michel Weibel b
a
Dipartimento di Energia, Laboratorio di Catalisi e Processi Catalitici, Politecnico di Milano, Piazza Leonardo da Vinci 32, I-20133 Milano, Italy
Daimler AG, Abteilung GR/VPE, D-70546 Stuttgart, Germany
b
a r t i c l e i n f o
a b s t r a c t
Article history:
We present a systematic study of the chemical steps in the NO/NO2–NH3 fast SCR reaction 2NH3
+
Received 10 March 2008
Revised 22 March 2008
Accepted 27 March 2008
Available online 2 May 2008
NO + NO2 → 2N2 + 3H2O over a commercial Fe-ZSM5 catalyst. The study is based on transient reaction
experiments at realistic conditions for removal of NOx from mobile diesel exhausts. Its goal is to assess
and critically evaluate the current ideas on the SCR mechanism, and also to establish to what extent
the mechanistic pathways demonstrated for V-based catalysts also apply to Fe-promoted zeolites. Results
show that the fast SCR reaction proceeds at low temperature via a global sequence involving NH4NO3 or
related surface species as intermediates,
Keywords:
Urea SCR
Fast SCR
2NO2 + 2NH3 → N2 + NH4NO3 + H2O,
NO + NH4NO3 → NO2 + N2 + 2H2O.
Ammonium nitrate
Zeolite catalysts
Diesel exhaust after treatment
Such a sequential scheme is the same as that proposed previously for the fast SCR chemistry over V-
based catalysts and other zeolite catalysts and thus is considered a general mechanism. It explains all
of the available observations for stoichiometry (e.g., optimum NO/NO2 unit molar ratio), selectivity (e.g.,
N2O from NH4NO3 decomposition), and kinetics (e.g., rate of fast SCR = rate of nitrate reduction by NO).
We further show that the redox reaction between NO and nitrates is the rate-controlling step and is
inhibited by ammonia. Remarkably, the same strongly enhanced deNOx activity observed in the fast SCR
reaction also was observed in the absence of gaseous NO2 but in the presence of surface nitrates. We
accordingly propose a general summary of the fast SCR chemistry over V-based and zeolite catalysts that
emphasizes the key role of surface nitrates.
© 2008 Elsevier Inc. All rights reserved.
1. Introduction
mechanism in close analogy with that of the standard SCR re-
action, but with NO2 serving as a more efficient oxidizing agent
for the vanadium sites than oxygen [3]. Furthermore, for the NO–
NO2/NH3 reacting system, they also reported the occurrence at low
temperatures of two side reactions not observed in the presence of
NO–NH3 only—namely the formation of ammonium nitrate,
In addition to the well-known “standard” SCR reaction,
1
2NH3 + 2NO + O2 → 2N2 + 3H2O,
(1)
(2)
2
the so-called “fast” SCR reaction,
2NH3 + 2NO2 → NH4NO3 + N2 + H2O,
(3)
2NH3 + NO + NO2 → 2N2 + 3H2O,
and the decomposition of ammonium nitrate by NO, described ac-
cording to the following stoichiometry [4]:
◦
plays a critical role at 180–300 C in boosting the denitrification
(deNOx) activity of new generation urea-SCR converters for diesel
vehicles integrated with an upstream preoxidation catalyst that
partially oxidizes NO to NO2 [1].
NH4NO3 ↔ NH3 + HNO3,
2HNO3 + NO → 3NO2 + H2O.
(4)
(5)
Koebel and co-workers [1–4] first extensively investigated the
fast SCR reaction over V2O5–WO3/TiO2 SCR catalysts. To explain
the higher rate of the fast SCR reaction, they proposed a redox
However, reactions (3)–(5) were considered to be side reactions
occurring in parallel to fast SCR (1), not participating in its mech-
anism [4].
In our work aimed at developing a chemically consistent sim-
ulation model of SCR converters for automotive applications [5],
we have addressed mechanistic aspects of the fast SCR chemistry
Corresponding author. Fax: +39 02 2399 3318.
E-mail address: enrico.tronconi@polimi.it (E. Tronconi).
*
0021-9517/$ – see front matter © 2008 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2008.03.027

A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
313
over commercial vanadium-based catalysts by means of transient
reaction analysis [6–10]. Our data have shown that the fast SCR
chemistry proceeds over V2O5–WO3/TiO2 SCR catalysts at low tem-
perature via a sequential scheme, which can be summarized as
comprising two global reactions—ammonium nitrate formation [re-
action (3)] and the following reaction between ammonium nitrate
and NO—formally involving NH4NO3 as an intermediate:
based on spectroscopic evidence and on steady-state reaction data.
In addition, other authors have reported NO2 disproportion [19–21]
and ammonium nitrate formation [16,22] over promoted and un-
promoted zeolites. Based on the analysis of their own data and
of literature data, Kröcher et al. [23] recently proposed a com-
mon SCR reaction scheme for transition–metal zeolites and for
vanadium-based catalysts that is in close agreement both with the
chemistry over a V2O5–WO3/TiO2 catalyst reported in our previous
work [6,7] and with the scheme proposed for the BaNa–Y zeo-
lite [18].
Herein we present a dedicated investigation of the elemen-
tary steps of the fast SCR reaction at low temperature over the
same commercial Fe-ZSM5 catalyst used in a previous SCR reac-
tivity study [26]. Our goal is to assess and critically evaluate the
current ideas on the SCR mechanism, and specifically to estab-
lish in a conclusive manner to what extent the same mechanistic
pathways demonstrated for V-based catalysts also apply to Fe-
promoted zeolite catalysts under fully representative conditions for
automotive applications. For this purpose, we take the same ex-
perimental approach (transient reaction experiments) used in our
previous mechanistic investigation over V2O5–WO3/TiO2, in order
to establish a direct link to the results for V-based catalysts.
NH4NO3 + NO → NO2 + N2 + 2H2O.
(6)
In fact, the sum of (3) and (6) yields the stoichiometry of the fast
SCR reaction (2). Notably, reactions (3) and (6) are similar to those
already reported by Koebel and co-workers, but here they are not
just side reactions, but are intimately related to the fast SCR chem-
istry.
The mechanism of the first step in the fast SCR sequential
scheme—ammonium nitrate formation [reaction (3)]—was clarified
by Koebel’s group [1–4] and implies NO2 dimerization (7), dispro-
portion (8), and successive reactions between nitrous and nitric
acid and NH3 (9), (10), with rapid decomposition of ammonium
nitrite to nitrogen:
2NO2 ↔ N2O4,
(7)
(8)
(9)
N2O4 + H2O ↔ HONO + HNO3,
NH3 + HONO ↔ NH+ + NO2 ↔ [NH4NO2] → N2 + 2H2O,
−
2. Experimental
4
NH3 + HNO3 ↔ NH+4 + NO3 ↔ NH4NO3.
(10)
−
The commercial catalyst used in this work was originally sup-
plied by Daimler in the form of a cordierite honeycomb mono-
lith (400 cpsi—6.5 mils) washcoated with Fe-ZSM5. For testing,
the catalyst was crushed and sieved to 140–200 mesh, to avoid
mass transfer limitations. Samples (160 mg of catalyst powder or
80 mg of catalyst powder diluted with 80 mg of quartz pow-
der) were loaded into a flow-microreactor consisting of a quartz
tube (6 mm i.d.) placed in an electric furnace. The reaction tem-
perature was monitored and controlled by a K-type thermocouple
immersed in the catalyst bed. Mass-flow controllers (Brooks In-
struments) were used to dose He, Ar, NH3, NO, NO2, and O2 in the
gaseous feed stream, while water vapor was added via a saturator
operated at controlled temperature. All of the lines before and af-
Concerning the second step in the fast SCR sequential scheme—
reaction (6) between NO and ammonium nitrate—we have demon-
strated by dedicated transient experiments a mechanism based on:
ammonium nitrate decomposition (10 reverse), successive oxida-
tion of NO to NO2 by nitric acid, which is thus reduced to nitrous
acid (11), and reaction of the latter with NH3 to form N2 via am-
monium nitrite decomposition (9) [6–10]:
NH4NO3 ↔ NH3 + HNO3,
HNO3 + NO ↔ NO2 + HONO,
NH3 + HONO → N2 + 2H2O.
(4) = (10 reverse)
(11)
(9)
◦
ter the reactor were heated to 200 C to prevent H2O condensation
We further observed that the rate-limiting step (6) does not
proceed over V-free WO3/TiO2 and thus is catalyzed by V2O5. The
same results had been previously reported for reaction (5) [4]. Ac-
cording to a redox interpretation of the fast SCR chemistry over
V-based catalysts, the key global reaction (6) actually is associ-
ated with a redox cycle involving the more effective reoxidation
of reduced V-sites by surface nitrates [8,9]; the fast SCR activity of
NO/NO2–NH3 is similar to the activity of NH3 + NO in the absence
of gaseous NO2 but in the presence of either NH4NO3 [6,7,10] or
nitrates prestored onto the vanadium catalyst surface [8,9]. This
rules out the possibility that the fast SCR reaction (2) can proceed
in parallel or consecutively to the nitrate decomposition by NO [re-
action (6)].
There is now a trend in the automobile industry to replace
vanadium-based SCR catalysts with zeolite-based systems to ex-
pand the operating temperature window and address the problems
associated with high-temperature deactivation of the anatase–
rutile TiO2 transition. Zeolites are the new class of automotive SCR
catalysts. Various zeolites have been proposed for this purpose, in-
cluding ZSM-5, mordenite, beta, ferrierite, and Y-zeolite [11]. In the
most active systems, zeolites generally are promoted by transition
metals, such as iron, copper, and silver. These catalysts reportedly
are associated with good deNOx activity in the standard and espe-
cially the fast SCR reactions [11–17].
and NH4NO3 deposition. The species concentrations in the outlet
stream were continuously monitored by a quadrupole mass spec-
trometer (Balzer QMS 200) and a UV analyzer (ABB-LIMAS 11 HV)
in parallel. He was used as carrier gas to enable evaluation of N
balances at steady state. More experimental details are available
elsewhere [5,7,8,26].
Before the experiments, the catalyst was conditioned with a
◦
◦
temperature ramp of 10 C/min up to 600 C in 2% O2 v/v, then
◦
held at 600 C for 1 h. Transient runs consisted of step-response
experiments at 150–170–190 C (transient response method [TRM])
◦
and in temperature-programmed reaction (TPR) runs. In a typical
TRM run, the reactor was kept at constant temperature under a
flow of He + 1% H2O, and step changes (e.g., 0 → 1000 → 0 ppm
or 0 → 500 → 0 ppm) of feed NH3 or NO or NO2 concentra-
tions were imposed. TRM tests were carried out over diluted cat-
alyst beds at 72 or 140 cm3/min (STP), corresponding to GHSV =
−1
8600–23,000 h
if referred to a monolith catalyst. At the end,
◦
◦
a temperature ramp (10 C/min, Tend = 550 C) was run to clean
up the catalyst surface. In the TPR runs, a stream containing NH3
(1000 ppm) and NOx (1000 ppm, with NO/NOx = 1 or 0.5) with
O2 (0 or 2% v/v) and H2O (1% v/v) in He was fed to the reactor
◦
initially at 150 C, and then the reactor temperature was linearly
◦
◦
increased up to 550 C at a heating rate of 20 C/min. Because the
purpose of this work was to address the chemistry of the fast SCR
reaction (2), many runs were performed in the absence of O2 so as
to eliminate contributions of the standard SCR reaction (1).
Concerning the mechanistic features of fast SCR over zeolites,
Weitz et al. [18] proposed a fast SCR pathway over a BaNa–Y ze-
olite similar to that discussed above for V2O5–WO3/TiO2 catalysts,

314
A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
Fig. 1. Formation of ammonium nitrate over the Fe-ZSM5 catalyst. Wcat = 0.080 g, flow rate = 140 cm3/min (STP). Feed = 1% H2O, 2% O2, 1000 ppm NH3 and 1000 ppm
◦
NO2 + He. T = 180 C.
Additional NO2 adsorption runs were performed to study the
formation of surface nitrates and NH4NO3 over undiluted crushed
monolith samples at 120 cm3/min (STP). A few final runs with
step changes in the NH3 feed were devoted to study the fast SCR
Fig. 2A shows storage of NO2 onto the catalyst up to saturation
in the absence of preadsorbed ammonia, accompanied by evolution
of NO. The data, reported previously [26], are consistent with
3NO2 + H2O ↔ NO + 2HNO3.
(12)
◦
dynamics over Fe-ZSM5 at 200 C using a higher feed flow rate
(240 cm3/min STP) in the presence of O2 (2%), to extend the study
Note that the molar ratio NO/ꢀNO2 is approximately equal to 1/3
during the entire adsorption transient. The overall reaction (12)
actually describes the chemisorption of NO2 in the form of ni-
trates and results from the combination of the NO2 dimerization
and disproportion reactions [reactions (7) and (8)] and the oxida-
tion of nitrite species by NO2 [reaction (11 reverse)]. Accordingly,
the results of Fig. 2A agree with the proposed chemistry and with
IR studies documenting the formation of nitrates on zeolite cata-
lysts on exposure to NO2 [18,24,25]. Kröcher and co-workers also
demonstrated reaction (12) in relation to NO2 adsorption onto Fe-
ZSM5 [22] and Cu-ZSM5 [21].
When the same NO2 storage experiment was run in the pres-
ence of ammonia preadsorbed on the catalyst, however, evolution
of N2 instead of NO was observed, as shown in Fig. 2B. Remark-
ably, the N2 concentration detected during the initial stage of the
reaction amounted to about 500 ppm versus a total conversion of
the 1000 ppm of NO2 in the feed. This corresponds to the overall
stoichiometry of ammonium nitrate formation from NO2 and ad-
sorbed ammonia [reaction (3)], according to steps (7)–(10), which
involve disproportion of NO2 [reactions (7) and (8)], as shown in
Fig. 2A, but in this case followed by reaction of nitrites with NH3
rather than with NO2 and leading to N2 via ammonium nitrite de-
composition [reaction (9)]. As a further confirmation, we note that
the total amount of N2 evolved during the experiment in Fig. 2B
(≈0.2 mmol/g) was close to the amount of preadsorbed ammo-
nia (≈0.18 mmol/g), as would be expected from (9). A TPD run
at the end of the NO2 storage phase demonstrated no N2O evolu-
tion, thus confirming that no NH4NO3 was left on the catalyst. On
the other hand, the NO2 evolution from nitrate decomposition was
similar to that observed during the corresponding final TPD part of
the run shown in Fig. 2A.
to conditions closer to automotive applications.
3. Results and discussion
3.1. NH4NO3 formation
◦
Fig. 1 shows a transient run at 180 C addressing the forma-
tion of ammonium nitrate over Fe-ZSM5. After the catalyst was
saturated with NH3 (1000 ppm), at time t = 0 s NO2 (1000 ppm)
was added stepwise to the feed stream. At steady state, the data
indicate equal conversions of NO2 and NH3 (ꢀNO2 = ꢀNH3
=
750 ppm), the formation of N2 (380 ppm), of N2O (20 ppm), and
a lack of about 700 ppm in the overall N balance (total atomic
N = 1300 ppm). These results agree with the stoichiometry of re-
action (3) and correspond to the formation of about 350 ppm of
NH4NO3, which, of course, cannot be detected by our analyzers. In
previous work [6,7,10], we found similar behavior (except for N2O
formation) over a V2O5–WO3/TiO2 catalyst and also documented
that part of the NH4NO3 was decomposed to HNO3 + NH3 and es-
caped undetected from the reactor, later condensing in cold spots
along the downstream lines.
As discussed above, in the case of V-based catalysts and BaNa–Y
zeolites, the formation of ammonium nitrate has been attributed
in the past to NO2 dimerization and disproportion [reactions (7)
and (8)], coupled with the formation and decomposition of ammo-
nium nitrite [reaction (9)] and the reaction between nitric acid and
ammonia [reaction (10)]. We addressed the question as to whether
the same mechanism also applies to Fe-ZSM5 by means of two
◦
NO2 adsorption experiments conducted at 150 C in the absence
Comparing the chemistry involved in NO2 adsorption onto Fe-
ZSM5 in the absence and in the presence of adsorbed ammo-
and presence of preadsorbed NH3 (Figs. 2A and 2B). Both runs con-
sist of a step addition of NO2 only (1000 ppm at t = 0 s) to a feed
stream containing 1% H2O, 2% O2, and He flowing over either a
clean Fe-ZSM5 catalyst or a catalyst sample previously exposed to
◦
nia reveals that at 150 C surface nitrites (or nitrous acid) react
more readily with NH3 according to the ammonium nitrite forma-
tion/decomposition reaction (9) rather than with NO2 according to
the reverse of (11), the reaction observed in the absence of am-
◦
NH3 at 150 C.

A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
315
◦
Fig. 2. Effect of preadsorbed NH3 on NO2 storage on Fe-ZSM5. Wcat = 0.160 g, flow rate = 120 cm3/min (STP). Feed = 1% H2O, 2% O2, NO2 = 1000 ppm + He. T = 150 C.
A = clean catalyst. B = catalyst with preadsorbed NH3.
monia. This reaction is indeed very fast; complete oxidation of
nitrous acid by NO2 occurs already at 150 C, as indicated by the
After the NO2 feed was shut off, the ammonia concentration
level returned to its feed value, and the concentrations of all the
other species dropped to zero. This indicates that no reaction oc-
curred between ammonia and ammonium nitrate at this temper-
ature [6,7,10]. At t = 9600 s, NO (1000 ppm) was added to NH3
in the feed, which resulted in a long transient (prolonged un-
til t ≈ 13,000 s), associated with the conversion of NO and NH3
and the formation of N2 and NO2. The conversion ratio of NO and
NH3 was 1:1 during the entire transient period except the initial
600 s, when a higher NO consumption along with formation of a
few ppm of NO2 were seen. The evolution of nitrogen mirrored
the consumption of NO and NH3 but exceeded it significantly (e.g.,
ꢀNO = 90 ppm, ꢀNH3 = 100 ppm, N2 = 130 ppm).
◦
stoichiometry observed during NO2 adsorption on the clean cata-
lyst (see Fig. 2A and related discussion). Other data (not reported)
◦
show the same behavior already at 50 C. Thus, the preferential re-
action of nitrites with ammonia, even if their oxidation by NO2
is so facile, merits some consideration. Our data also indicate that
ammonia reacts preferentially with surface nitrites, thus forming
N2, rather than with nitrates to form NH4NO3.
3.2. Reactivity of NH4NO3 with NO
To investigate the reactivity of NH4NO3 with nitric oxide over
Fe-ZSM5, ammonium nitrate was generated in situ (i.e., on the cat-
alyst itself). As shown in Fig. 3, 1000 ppm of NO2 along with H2O
(1% v/v) and O2 (0% v/v) were initially fed to the microreactor at
The reactivity observed when NH3 and NO were co-fed to the
Fe-ZSM-5 catalyst cannot be explained by the standard SCR re-
action (1), which exhibits a different stoichiometry and requires
the participation of O2. Rather, it is attributed to the reaction (6)
between NO and the ammonium nitrate deposited onto the cat-
alyst during the previous stage of the experiment. Under similar
operating conditions, we observed the same behavior over a V2O5–
WO3/TiO2 catalyst; we explain this as due to the following reac-
tion:
◦
170 C; subsequently, 1000 ppm of NH3 was added to the feed to
form and build up ammonium nitrate on the catalyst surface ac-
cording to reaction (3). At the end of the first build-up stage, only
NO2 was removed from the feed stream while NH3 was retained,
to limit decomposition of ammonium nitrate to NH3 + HNO3, thus
preserving a significant fraction of the ammonium nitrate formed
on the catalyst. In the final stage of the run, 1000 ppm of NO was
added to the NH3-containing feed stream, still without oxygen. The
same experimental procedure has been used previously to study
the reactivity of NH4NO3 with nitric oxide over V2O5–WO3/TiO2
[6,7].
Fig. 3 shows the temporal evolutions of the outlet concen-
trations of NH3, NO, NO2, N2, and N2O during the NH4NO3
buildup and reactivity run over Fe-ZSM5. During the first stage
(t < 7600 s), when NH3 and NO2 were co-fed to the reactor,
a roughly equimolar consumption of both species was observed
(ꢀNO2 = 830 ppm; ꢀNH3 = 820 ppm), along with formation of
430 ppm of N2, 15 ppm of N2O, and a lack of 760 ppm in the N
balance (total atomic N = 1240 ppm). Such results are clearly
in line with the formation of ammonium nitrate [reaction (1)],
as discussed in the previous section, and point to the forma-
tion/deposition of NH4NO3 salt onto the catalyst surface. A small
fraction of such ammonium nitrate was decomposed to N2O. The
formation of NH4NO3 salt on V2O5–WO3/TiO2 under similar con-
ditions has been documented by IR analysis of a catalyst sample
unloaded from the reactor [10].
1
3
5
NH3 + NO + NH4NO3 → N2 + H2O.
(13)
2
2
2
In view of the 1:1 consumption ratio between NO and NH3, along
with the formation of about 3/2 ppm of N2 for each ppm of con-
verted NO observed over Fe-ZSM5, reaction (13) also explains the
results shown in Fig. 3. Notably, reaction (13) is simply the sum
of reaction (6) between NO and ammonium nitrate, forming NO2,
and reaction (3) between such NO2 and NH3, which rapidly re-
forms part of the converted ammonium nitrate.
Additional runs to explore the temperature dependence of the
◦
reactivity of NH4NO3 with NO were performed at 150 and 190 C
(Figs. 4A and 4B). At both temperatures, the Fe-ZSM5 catalyst was
active in the formation of ammonium nitrate [reaction (3)], con-
firming the expected stoichiometry within experimental error. An
increase in the associated NO2 and NH3 conversions was observed
with decreasing temperature, indicating that NH4NO3 formation
is favored by low temperatures, in agreement with the litera-
ture [1,18]. As for the reaction between ammonium nitrate and
◦
NO +NH3, no conversion was apparent at 150 C (Fig. 4A), whereas
◦
considerable reactivity was detected at 190 C (Fig. 4B). In fact, in

316
A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
Fig. 3. Buildup of ammonium nitrate and its reactivity with NO over Fe-ZSM5. Wcat = 0.080 g, flow rate = 72 cm3/min (STP). Feed = 1% H2O, 0% O2, NH3 = 0–1000 ppm,
◦
NO = 0–1000 ppm, NO2 = 0–1000 ppm + He. T = 170 C.
the latter case, an initial evolution of 450 ppm of N2 was found,
which is in fair agreement with reaction (13) (ꢀNO = 320 ppm)
but far greater than the 130 ppm of N2 observed in the run at
NO2 conversion, it is then reasonable to propose that also reac-
tion (3) (i.e., formation of ammonium nitrate) also was proceeding
along with the fast SCR reaction (2). The formation of NH4NO3 is
in fact corroborated by a lack in the atomic nitrogen balance at
steady state; furthermore, it is compatible with the observed N2
concentrations, as we briefly discuss in what follows.
◦
170 C (Fig. 3).
Based on the aforementioned findings, it seems legitimate to
conclude that a similar reactivity between NO and ammonium
nitrate applies over the present Fe-ZSM5 catalyst, as has been
reported previously for V2O5–WO3/TiO2 catalysts and a BaNa–Y
zeolite [6,7,18,23]. The results shown in Fig. 3 and the related in-
terpretation also agree with the IR data of Busca et al. [25], who
studied the interaction of NO and NO/NO2 with NH3 preadsorbed
onto H-ZSM5 and found that NO2 reacted readily at room temper-
ature with adsorbed ammonium ions, forming nitrates and N2, as
expected from reaction (3), whereas in the presence of NO + NO2,
the catalyst was completely depleted of adsorbed ammonia and of
nitrates at 373 K, consistent with the reduction of nitrates by NO
[reaction (11)] and the resulting overall reaction (2). Furthermore,
the present data demonstrate a strong temperature dependence of
reaction (6) between NO and NH4NO3.
We consider the NO consumption (ꢀNO = 40 ppm) to rep-
resent the extent of the fast SCR reaction (2), whereas the ex-
tent of reaction (3) is obtained by subtracting the amount of NO
converted by the fast SCR reaction from the total NO2 conver-
sion: (ꢀNO2Amm = 420 − 40 = 380 ppm). Accordingly, the total N2
formed by reactions (2) and (3) should amount to 80 + 190 =
270 ppm, which is close enough to the 280 ppm of nitrogen ac-
tually detected at steady state. Furthermore, the observed lack of
320 ppm in the atomic N balance is compatible with the ex-
pected formation of approximately 180 ppm of NH4NO3 due to
reaction (3), after subtracting 10 ppm of N2O from the NH4NO3
decomposition. Based on these findings, we therefore propose that
both reactions (2) and (3) occurred during the first stage of the
transient run in Fig. 5, with the formation and deposition of am-
monium nitrate on Fe-ZSM5. Again, the same behavior had been
reported previously under similar conditions over V2O5–WO3/TiO2
and zeolite catalysts [4,6,7,10,21,26].
We now examine the second stage of the TRM run in Fig. 5.
When NO2 was removed from the feed stream at t ≈ 13,000 s,
a discontinuity in the outlet concentration traces of all species
but NO became apparent. In fact, the ammonia concentration rose
to about 960 ppm, the NO2 trace dropped rapidly to zero, the
N2 concentration decreased to about 60 ppm (mirroring the NH3
evolution), and the NO outlet concentration remained unaltered
at 460 ppm. For pseudo steady-state conditions, such data imply
similar consumption of NO and NH3 (about 40 ppm) and a simul-
taneous formation of 60 ppm of N2, demonstrating a stoichiometry
similar to that reported earlier for the reaction between NO and
NH4NO3 [reaction (6)]. In other words, during this second stage of
the experiment, NO was reacting with the ammonium nitrate de-
posited onto the catalyst for the duration of the first stage. The
resulting transient lasted about 6000 s and in fact ended after de-
pletion of all of the previously accumulated ammonium nitrate.
We now analyze in more detail the chemistry before and after
NO2 was removed from the feed. We have shown that the conver-
sion of NO was due to the fast SCR reaction (2) when NH3, NO, and
3.3. Role of NH4NO3 in the fast SCR reaction
After demonstrating the reactivity of NH4NO3 with NO over Fe-
ZSM5, we proceeded to study the relationship of such a reaction
with the NO/NO2–NH3 (fast SCR) reacting system. In analogy with
our previous studies on a V-based catalyst [6,7], the study relied
on TRM runs involving a first stage in which NH3, NO, and NO2
(1000:500:500 ppm) were co-fed to the Fe-ZSM5 catalyst along
with H2O (1% v/v) but with no O2, followed by a second stage
after stepwise removal of NO2 from the feed stream.
Fig. 5 displays the dynamics of the NH3, NO, NO2, N2, and N2O
◦
outlet concentrations observed during one of such runs at 170 C.
At t > 3000 s, when NO and NO2 were added to NH3 in the feed,
a reaction transient associated with conversion of both NH3 and
NOx and with the formation of N2 and of a few ppm of N2O be-
came apparent. At steady state, the NH3 consumption was around
440 ppm, close to the NOx conversion of 460 ppm. It should be
noted, however, that the NO2 conversion was much greater than
the NO conversion (ꢀNO = 40 ppm, ꢀNO2 = 420 ppm); therefore,
the fast SCR reaction (2) by itself cannot explain the observed be-
havior, because it would result in equimolar conversions of NO and
◦
NO2. In view of the low temperature (170 C) and of the higher

A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
317
Fig. 4. Effect of temperature on the buildup of ammonium nitrate and its reactivity with NO over Fe-ZSM5. Wcat = 0.080 g, flow rate = 72 cm3/min (STP). Feed = 1% H2O,
◦
◦
0% O2, NH3 = 0–1000 ppm, NO = 0–1000 ppm, NO2 = 0–1000 ppm + He. T = 150 C (A), 190 C (B).
NO2 were co-fed to the Fe-ZSM5 catalyst. In contrast, in the sub-
sequent stage, in the absence of NO2, NO consumption was due to
reaction (6) between NO and ammonium nitrate. But despite the
occurrence of two formally different reactions, the experimental
concentration trace of NO shown in Fig. 5 demonstrates no discon-
tinuity between the two stages on the stepwise removal of NO2
from the feed stream. However, recalling that in the presence of
NH3, NO, and NO2 (first stage in Fig. 5), the formation of ammo-
nium nitrate [reaction (3)] was already active along with the fast
SCR reaction (2), this apparent contradiction can be explained by
considering that during the first stage of the run, NO consumption
occurred not directly at the expense of gaseous NO2 [i.e., according
to reaction (2)], but rather via reaction of NO with NH4NO3 [i.e.,
according to reaction (6)], with ammonium nitrate in turn formed
from NO2 and NH3 according to reaction (3). This obviously ex-
plains the absence of any discontinuity of the NO trace between
the two parts of the experiment shown in Fig. 5 and points to a
sequential scheme for the fast SCR chemistry in which ammonium
nitrate (or a related surface species) plays the role of the key in-
termediate and reaction (6) between NO and ammonium nitrate is
the rate-determining step at these low temperatures. In particular,
we emphasize that the data shown in Fig. 5 demonstrate the occur-
rence of the fast SCR in the absence of gaseous NO2 and rule out
significant contributions to deNOx activity from a direct gas-phase
reaction between NO and NO2 [16]. These conclusions regarding
the fast SCR chemistry over Fe-ZSM5 are in close agreement with
our previous findings for a V2O5–WO3/TiO2 catalyst [6,7].
We further investigated the role of reaction (6) between
NH4NO3 and NO in the fast SCR chemistry over the Fe-zeolite at
◦
other temperatures, 150 and 190 C; the results are shown in Fig. 6.
◦
At the end of the first stage of the TRM run at 150 C (Fig. 6A), we
observed no conversion of NO and equimolar conversions of NH3
and NO2 (ꢀNH3 = 410 ppm, ꢀNO2 = 405 ppm), resulting in the
evolution of 200 ppm of N2 and a lack of 400 ppm in the N bal-
ance. These data indicate the occurrence of reaction (3) (formation
of ammonium nitrate) only, whereas the absence of NO conver-

318
A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
Fig. 5. Comparison between activity of fast SCR and activity of NO + NH4NO3 over Fe-ZSM5. Wcat = 0.080 g, flow rate = 72 cm3/min (STP). Feed = 1% H2O, 0% O2,
◦
NH3 = 0–1000 ppm, NO = 0–500 ppm, NO2 = 0–500 ppm + He. T = 170 C.
sion is explained by the adverse effect of low temperature on the
reactivity between NO and NH4NO3, as discussed previously (see
Fig. 4A). In fact, no reactivity of NO was detected during the sub-
sequent transient phase after NO2 was removed from the feed.
catalyst according to reaction (3). Thus, the behavior observed at
the time of ammonia shutoff is consistent with the overall fast
SCR chemistry proposed over Fe-ZSM5, but apparently contradicts
◦
the temperature threshold of 150 C observed for the onset of re-
◦
In contrast, the first stage of the TRM run at 190 C (Fig. 6B) re-
action (6) between NH4NO3 and NO. A possible reason for this
finding is that reaction (6) actually may occur between NO and ni-
tric acid (or surface nitrates) [see reaction (11)] formed by NH4NO3
decomposition according to reaction (10 reverse), and then would
be inhibited by NH3. Reducing the gaseous ammonia concentra-
tion would shift the equilibrium of NH4NO3 decomposition toward
higher HNO3 concentrations, thereby promoting the reactivity with
NO. A similar effect was observed at different temperatures as well.
The formation of surface nitrates and their role in the formation of
ammonium nitrate were already demonstrated in this work by the
NO2 adsorption experiments (see Fig. 2). Ammonia inhibition of
reaction (6) over V2O5–WO3/TiO2 also has been proposed [6,7].
sulted in high equimolar conversions of NH3 and NOx (about 95%)
associated with higher consumption of NO2 compared with NO
(ꢀNO2 = 500 ppm, ꢀNO = 445 ppm), formation of 915 ppm of N2,
and a lack of about 40 ppm in the atomic N balance. Within the
limits of experimental error, such data are again consistent with
the simultaneous occurrence of ammonium nitrate formation (3)
and the fast SCR reaction (2), although with a much greater con-
◦
tribution of the latter than occurred at 170 C. This finding is ex-
plained by the strong promoting effect of temperature on the key
reaction (6) between NO and NH4NO3 documented in Figs. 3 and 4.
◦
The subsequent TRM stage, run at 190 C after NO2 removal,
demonstrated significant conversion of NO and NH3, but only dur-
ing an initial short transient phase. This finding is linked to the
high fast SCR activity and the low NH4NO3 production observed
during the preceding stage in the presence of NH3, NO, and NO2.
3.5. TPR experiments
More information confirming the central role of nitrates in the
fast SCR reactivity is provided in Fig. 7, which compares the tem-
perature dependence of NO conversion measured during four dif-
ferent NO + NH3 TPR experiments with equal space velocities and
heating rates (20 K/min). In the experiment reflected in curve A,
the feed included NO + NH3 (1000 ppm each) + H2O (1% v/v), but
no oxygen. As expected, the observed temporal evolution of the
outlet concentrations, indicating the onset of deNOx activity only
◦
Accordingly, at 190 C, the system was unable to build up a sub-
stantial amount of ammonium nitrate on the catalyst for its sub-
sequent reaction with NO in the absence of NO2, thus confirming
the importance of NH4NO3 in the NO–NO2/NH3 chemistry at low
temperatures.
3.4. NH3 inhibition of NO reactivity with nitrates
◦
above 300 C, is in agreement with the so-called “slow” SCR reac-
Additional interesting information can be extracted from the
tion [4], that is, the poorly active deNOx reaction between NO and
NH3 in the absence of oxygen,
◦
final part of the TRM run at 150 C. As shown in Fig. 6A, imme-
diately after the NH3 feed was shut off (t > 24,000 s), a peak
of deNOx activity involving conversion of NO, formation of N2,
and evolution of NO2 appeared. The transient exhibited two dis-
tinct phases. In the first phase, NO conversion was associated with
equimolar formation of N2 and NO2, consistent with reaction (6)
between NO and the residual ammonium nitrate on the catalyst
surface,
5
3NO + 2NH3 → N2 + 3H2O.
(14)
2
Repeating the same temperature ramp experiment but with 2% O2
v/v in the feed (curve B) revealed greater deNOx activity, due to the
standard SCR reaction (1) instead of reaction (14). Curve C shows
the even much higher deNOx activity resulting from running the
same temperature ramp with a feed containing 1000 ppm of NH3
and 500 ppm each of NO and NO2. Of course, curve C is represen-
tative of the NO conversion in the fast SCR reaction (2). During the
fourth and final TPR experiment, associated with curve D, a feed
containing NO + NH3 (1000 ppm each) + H2O (1% v/v), but no
NH4NO3 + NO → N2 + NO2 + 2H2O.
(6)
In the second phase, however, the outlet concentration of N2 was
greater than that of NO2. This can be explained by a further re-
action of NO2 from reaction (6) with NH3 still adsorbed onto the

A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
319
Fig. 6. Comparison between activity of fast SCR and activity of NO + NH4NO3 over Fe-ZSM5: effect of temperature. Wcat = 0.080 g, flow rate = 72 cm3/min (STP). Feed = 1%
◦
◦
H2O, 0% O2, NH3 = 0–1000 ppm, NO = 0–500 ppm, NO2 = 0–500 ppm + He. T = 150 C (A), 190 C (B).
oxygen or NO2, was passed over a catalyst sample preexposed to
the presence of surface nitrates was virtually identical to that as-
◦
◦
1000 ppm of NO2 + 1% H2O at 60 C. Fig. 7 shows that during the
sociated with the fast SCR reaction up to about 200 C. The drop
◦
◦
initial part of this temperature ramp, up to about 200 C, the evo-
in NO conversion shown by curve D at temperatures above 200 C
lution of the NO conversion matched that seen for the fast SCR TPR
run. As the experiment proceeded, NO conversion dropped sharply,
reaching close to zero, before eventually approaching the behavior
observed for the slow SCR reaction (curve A).
Recalling the data on the buildup of nitrates on the Fe-ZSM5
catalyst on NO2 adsorption (see Fig. 2), a likely explanation of
the similarity between the initial parts of curves C (fast SCR)
and D (NO + NH3 over catalyst pretreated with NO2) in Fig. 7 is
that in both cases, NO and ammonia in the feed were reacting
with surface nitrates, either formed directly via NO2 disproportion
(curve C) or formed previously and stored on the catalyst during
its pretreatment with NO2 (curve D). Notably, once again in this
experiment, the deNOx activity in the absence of gaseous NO2 but
can be attributed to the depletion of surface nitrates. In fact, in
another similar experiment (not reported), in which the catalyst
◦
◦
was pretreated with NO2 at 150 C rather than at 60 C, an earlier
drop of the NO conversion occurred, due to the reduced amount
of nitrates stored at the higher temperature. Eventually, the match
◦
between curves D and A at temperatures above 300 C confirms
the absence of any residual oxidizing agents in the final part of
the temperature ramp experiment over the NO2-pretreated sam-
ple.
Based on the data presented and discussed so far, the reac-
tivity of NO/NO2–NH3 over Fe-ZSM5 at low temperatures appears
to be totally consistent with earlier findings over vanadium-based
catalysts [6,7,23] as well as with the mechanistic proposals for

320
A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
Fig. 7. Temperature-programmed reaction experiments over Fe-ZSM5. Wcat = 0.080 g, flow rate = 72 cm3/min (STP). Curve
A
(slow SCR): Feed = 1% H2O, 0% O2,
NH3 = 1000 ppm, NO = 1000 ppm + He. Curve B (standard SCR): Feed = 1% H2O, 2% O2, NH3 = 1000 ppm, NO = 1000 ppm + He. Curve C (fast SCR): Feed = 1% H2O,
0% O2, NH3 = 1000 ppm, NO = 500 ppm, NO2 = 500 ppm + He. Curve D: Feed = 1% H2O, 0% O2, NH3 = 1000 ppm, NO = 1000 ppm + He over catalyst pretreated with NO2
◦
(1000 ppm) + H2O (1%) at 60 C.
Fig. 8. Fast SCR: TRM run over Fe-ZSM5. Wcat = 0.100 g, flow rate = 240 cm3/min (STP). Feed = 1% H2O, 2% O2, NH3 = 0–1000 ppm, NO = 0–500 ppm,
◦
NO2 = 0–500 ppm + He. T = 200 C.
◦
BaNa–Y [18], all of which depend on the reactivity between NO
and ammonium nitrate or surface nitrates in the fast SCR chem-
istry. Our findings further demonstrate the similarity of the rate
of the fast SCR reaction and the rate of nitrate reduction by NO,
which had not been reported previously on zeolite catalysts.
200 C with 1000 ppm of NH3 added to a feed stream contain-
ing 500 ppm of NO and 500 ppm of NO2 with O2 (2%) and H2O
(1%). The data identify a first transient phase with high deNOx ac-
tivity that was, however, progressively reduced until steady state
was reached, along with a final phase after NH3 shutoff, again as-
sociated with high deNOx activity, which was rapidly exhausted.
Fig. 8 shows a simultaneous conversion of both NO and NO2 at
steady state at the end of the first phase, with NO2 consumption
higher by about 100 ppm, formation of N2 in amounts equal to
3/2 of the NO consumption added to half of the NO2 consump-
tion, and a lack of 100 ppm in the N balance. Such behavior can
be well explained by the chemistry discussed in the previous sec-
tions, invoking ammonium nitrate formation [reaction (3)] and the
consecutive reaction (6) between NO and NH4NO3.
3.6. Fast SCR dynamics under NH3 pulsed feed
We analyzed the dynamic behavior of the NO/NO2–NH3 re-
acting system over Fe-ZSM5 under a stepwise feed of ammonia
at low temperature at conditions closer to those prevailing in
real automotive SCR converters—namely in the presence of oxygen
and at higher space velocities. Fig. 8 shows the temporal evo-
lution of the outlet species concentrations during a TRM run at

A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
321
Somewhat different behavior can be observed in the second
4. Conclusion
phase of the run shown in Fig. 8. After stepwise removal of NH3
from the feed, a peak in NO conversion appeared, accompanied
by evolution of N2 and of NO2, which temporarily exceeded their
feed concentrations of 500 ppm. In analogy to the discussion of
the final stage in Fig. 6A, such dynamics can be attributed to the
reaction of NO with ammonium nitrate and to the inhibiting effect
of ammonia. In fact, removing NH3 from the gas phase promoted
reaction (6) between NO and NH4NO3 still present on the cata-
lyst, thereby releasing NO2 and N2 until the surface species was
depleted.
Our findings demonstrate conclusively that the chemistry pro-
posed for the fast SCR reaction at low temperatures over V2O5–
WO3/TiO2 [6,7,10,23] and over zeolite systems [13,18,22] is able to
interpret the steady-state and dynamic features of the NO/NO2–
NH3 system over commercial Fe-ZSM5 catalysts, as well under
conditions representative of diesel exhaust after treatment, as pro-
posed previously [23]. Accordingly, the role of NO2 in the fast SCR
reactivity is to form surface nitrites and nitrates via dimerization
and disproportionation/heterolytic chemisorption, the role of NO is
to reduce nitrates to nitrites, and the role of NH3 is to enable rapid
and selective decomposition of nitrites to nitrogen via formation
of unstable ammonium nitrite. It was pointed out previously that
this chemistry can explain the optimal 1:1 molar ratio of NO and
NO2 in the fast SCR reaction [6,13,15]. It also can explain the full
range of selectivity resulting from varying the NO2/NOx feed ra-
tio [13,24,26]; incomplete reduction of nitrates by NO (the critical
step R7 in Table 1) is responsible for the undesired formation of
NH4NO3 at very low temperatures and of N2O at low to interme-
diate temperatures.
◦
More TRM runs at higher temperatures (225–300 C) (not re-
ported) demonstrated higher activity, approaching total conversion
of NOx. The reaction dynamics observed during the first stage of
the TRM runs, after the addition of NH3 to the feed stream, were
qualitatively similar to the run shown in Fig. 8, but transients were
faster due to the higher deNOx activity. In contrast, the final tran-
sient phase after the removal of NH3 from the feed became shorter
with increasing temperature, and eventually disappeared. This is
consistent with the reduced storage of surface species (adsorbed
NH3 and nitrates) at higher temperatures.
Although the general SCR chemistry emerging from the present
work essentially confirms existing concepts, here for the first time
it has been evaluated over zeolite systems on a step-by-step basis
by means of transient reaction analysis. Thanks to this powerful
technique, additional novel results have been obtained over Fe-
ZSM5:
3.7. Rephrasing the reactions in terms of surface species
Throughout this paper, all of the reactions involved in the NO–
NO2/NH3 chemistry have been expressed conventionally in terms
of molecular species. But, as is apparent from the data presented
so far, the actual catalytic mechanism likely involves related sur-
face adspecies, as discussed, for example, in the analysis of the
data given in Figs. 6 and 7. Accordingly, HONO and HNO3 in the
reaction steps should be replaced with nitrite and nitrate species,
respectively.
Fig. 9 shows a schematic summary of the fast SCR chemistry in
terms of surface species. The related basic reaction steps are given
in Table 1, along with data on the decomposition of the overall
reactions observed in this work. Thus, for example, the reaction
sequence yielding the fast SCR reaction (2) can be rephrased in
terms of the elementary steps in Table 1 as
1. The reduction of nitrates by NO (or, conversely, the oxida-
tion of NO by nitrates) is the rate-limiting step in the fast
SCR chemistry at low temperature; the reverse reaction (i.e.,
oxidation of nitrites by NO2) is very fast even at very low tem-
perature.
+
2NH3 + H2O ↔ 2NH4 + O2−
,
R4
R1 + R2
R7
2NO2 + O2− ↔ NO2− + NO3
,
−
NO−3 + NO ↔ NO2 + NO
,
−
2
2NH+4 + 2NO2 → 2N2 + 4H2O.
2*R5
−
Fig. 9. Proposed reaction scheme for NO/NO2–NH3 SCR at low temperature.
Table 1
The fast SCR chemistry
Basic reaction steps in NO/NO2–NH3 SCR chemistry over V-based and zeolite catalysts
Involving NO2 only
2NO2 ↔ N2O4
R1 see (7)
R2 see (8)
R3 see (11 reverse)
−
N2O4 + O2− ↔ NO2− + NO3
−
−
NO2 + NO2 ↔ NO + NO3
In the presence of NH3
+
2NH3 + H2O ↔ 2NH4 + O2−
R4 NH3 adsorption
R5 see (9)
−
NH+4 + NO2 ↔ [NH4NO2] → N2 + 2H2O
−
NH+4 + NO3 ↔ NH4NO3
R6 see (10)
In the presence of NO
NO + NO3 ↔ NO2 + NO2
−
−
R7 see (11)
Global reactions observed in this work, and their relationship with the basic reactions above
−
3NO2 + O2− ↔ NO + 2NO3
(12) = R1 + R2 + R3
2NO2 + 2NH3 → NH4NO3 + N2 + H2O
NO + NH4NO3 → NO2 + N2 + 2H2O
(3) = R1 + R2 + R4 + R5
(6) = R6 reverse + R7 + R5
∗
1
2
NO + NH3
+
1 NH4NO3
→
3 N2
+
5 H2O
(13) = (6) + (3)
2
2
2
2NH3 + NO + NO2 → 2N2 + 3H2O
(2) = R4 + R1 + R2 + R7 + 2*R5

322
A. Grossale et al. / Journal of Catalysis 256 (2008) 312–322
2. Ammonia inhibits such a redox step, likely by blocking the
[6] C. Ciardelli, I. Nova, E. Tronconi, D. Chatterjee, B. Bandl-Konrad, Chem. Com-
mun. (2004) 2718.
surface nitrates; high NH3 concentrations have a detrimental
effect on fast SCR kinetics at low temperature [26].
3. The same enhanced fast SCR deNOx activity also is seen in the
absence of gaseous NO2, provided that the catalyst has been
saturated with nitrates. This important result is not obvious
from the previous SCR literature.
[7] I. Nova, C. Ciardelli, E. Tronconi, D. Chatterjee, B. Blandl-Konrad, Catal. To-
day 114 (2006) 3.
[8] E. Tronconi, I. Nova, C. Ciardelli, D. Chatterjee, M. Weibel, J. Catal. 245 (2007) 1.
[9] I. Nova, C. Ciardelli, E. Tronconi, D. Chatterjee, M. Weibel, Top. Catal. 42–43
(2007) 43.
[10] C. Ciardelli, I. Nova, E. Tronconi, B. Bandl-Konrad, D. Chatterjee, M. Weibel, B.
Krutzsch, Appl. Catal. B: Environ. 70 (2007) 80.
[11] K. Rahkamaa-Tolonen, T. Maunula, M. Lomma, M. Huuhtanen, R. Keiski, Catal.
Today 100 (2005) 217.
[12] M. Schwidder, M.S. Kumar, K. Klementiev, M.M. Pohl, A. Brückner, W. Grünert,
J. Catal. 231 (2005) 314.
[13] O. Kröcher, M. Devadas, M. Elsener, A. Wokaun, N. Söger, M. Pfeifer, Y. Demel,
L. Mussmann, App. Catal. B: Environ. 66 (2006) 208.
[14] M. Wallin, C.J. Karlsson, M. Skoglundh, A. Palmqvist, J. Catal. 218 (2003) 354.
[15] M. Li, Y. Yeom, E. Weitz, W.M.H. Sachtler, J. Catal. 235 (2005) 201.
[16] H. Sjövall, L. Olsson, E. Fridell, R.J. Blint, Appl. Catal. B: Environ. 64 (2006)
180.
In summary, our work emphasizes the key role of surface ni-
trates in the fast SCR chemistry. Their reactivity with NO is critical
to explaining the strongly enhanced deNOx activity and governs
the fast SCR kinetics at low temperature, whereas their buildup
and depletion controls the SCR dynamics in parallel to ammonia
adsorption–desorption. Identifying the elementary steps of fast SCR
reactions over the Fe-ZSM5 catalyst is an important step in deriv-
ing chemically consistent kinetic schemes, as well as a key stage
in the development of new catalytic systems and integrated af-
tertreatment technologies.
[17] G. Delahay, D. Valade, A. Guzma-Vargas, B. Coq, Appl. Catal. B: Environ. 55
(2005) 149.
[18] Y.H. Yeom, J. Henao, M.J. Li, W.M.H. Sachtler, E. Weitz, J. Catal. 231 (2005)
181.
[19] G. Li, C.A. Jones, V.H. Grassian, S.C. Larsen, J. Catal. 234 (2004) 401.
[20] M.S. Kumar, M. Schwidder, W. Grünert, U. Bentrup, A. Brückner, J. Catal. 239
(2006) 173.
[21] J. Despres, M. Koebel, O. Kröcher, M. Elsener, A. Wokaun, Microporous Meso-
porous Mater. 58 (2003) 175.
Acknowledgments
The authors thank one of the reviewers for critical suggestions
that helped improve the quality of the paper.
[22] M. Devadas, O. Kröcher, M. Elsener, A. Wokaun, N. Söger, M. Pfeifer, Y. Demel,
L. Mussmann, Appl. Catal. B: Environ. 67 (2006) 187.
[23] O. Kröcher, in: Proceedings of Congress MinNOx—Minimization of NOx Emis-
sions through Exhaust Aftertreatment, Berlin, February 1–2, 2007, Paper H030-
02-286-7.
[24] M. Devadas, O. Kröcher, M. Elsener, A. Wokaun, G. Mitrikas, N. Söger, M. Pfeifer,
Y. Demel, L. Mussmann, Catal. Today 119 (2007) 137.
[25] V. Sanchez-Escribano, T. Montanari, G. Busca, Appl. Catal. B: Environ. 58 (2005)
19.
References
[1] G. Madia, M. Koebel, M. Elsener, A. Wokaun, Ind. Eng. Chem. Res. 41 (2002)
3512.
[2] M. Koebel, M. Elsener, G. Madia, Ind. Eng. Chem. Res. 40 (2001) 52.
[3] M. Koebel, G. Madia, F. Raimondi, A. Wokaun, J. Catal. 209 (2002) 159.
[4] M. Koebel, G. Madia, M. Elsener, Catal. Today 73 (2002) 239.
[5] C. Ciardelli, I. Nova, E. Tronconi, D. Chatterjee, T. Burkhardt, M. Weibel, Chem.
Eng. Sci. 62 (2007) 5001, and references therein.
[26] A. Grossale, I. Nova, E. Tronconi, Catal. Today (2007), doi:10.1016/j.cattod.
2007.10.117.