Applied Catalysis B: Environmental 142–143 (2013) 861–876

Contents lists available at SciVerse ScienceDirect

Applied Catalysis B: Environmental

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

Experimental and modeling study of a dual-layer (SCR + PGM) NH3 slip monolith
catalyst (ASC) for automotive SCR aftertreatment systems. Part 1. Kinetics for the
PGM component and analysis of SCR/PGM interactions
Massimo Colombo a , Isabella Nova a , Enrico Tronconi a,∗ , Volker Schmeißer b ,
Brigitte Bandl-Konrad b , Lisa Zimmermann c
a
Dipartimento di Energia, Laboratorio di Catalisi e Processi Catalitici, Politecnico di Milano, Piazza Leonardo da Vinci 32, I-20133 Milano, Italy
b
Daimler AG, 019-G206 RD/RPE, 70546 Stuttgart, Germany
c
Daimler AG, 019-D121 TP/PME, 70546 Stuttgart, Germany

a r t i c l e i n f o a b s t r a c t

Article history: We present herein the first of two parts in the development and validation of a chemically and physically
Received 28 July 2012 consistent mathematical model of a commercial dual-layer (SCR + PGM) monolithic NH3 slip converter
Received in revised form 17 October 2012 (ASC). The overall project followed a systematic approach of growing complexity, and its results empha-
Accepted 29 October 2012
size the beneficial features of a dual-layer configuration with the SCR catalyst on top of the PGM
Available online 12 November 2012
component. Specifically, we report in this paper NH3 /O2 /NO NO2 steady-state and transient kinetic
runs performed over the PGM component of the dual-layer NH3 slip catalyst. The PGM component was
Keywords:
tested in a representative temperature range (150–550 ◦ C) in the form of precursor washcoat powders at
Urea SCR
Ammonia slip
high space velocities in order to gain kinetic information. From these data an original global PGM kinetic
ASC model was developed, which fully accounts for the effects of temperature and of NO2 /NOx feed ratio (0–1)
Zeolite catalysts on NH3 oxidation. The model considers NO2 inhibition on NO oxidation, as well as a novel NO2 inhibition
Diesel exhaust aftertreatment effect on the NH3 oxidation reactions. Comparative NH3 /O2 /NO NO2 steady-state runs were performed
also over two combinations of SCR + PGM powders (sequential double-bed and mechanical mixture). The
N2 selectivity was greater over the mechanical mixture, as in this configuration the unselective NH3 oxi-
dation products (NOx ) formed over the PGM catalyst had a chance to further react selectively with NH3
over the SCR catalyst. Such a positive interaction between the PGM and the SCR catalytic chemistries was
satisfactorily predicted by a model involving the simple superposition of the PGM and SCR kinetics. In the
following part of the project the herein developed PGM kinetics, together with consistent SCR kinetics,
will be incorporated in a novel dual-layer monolith catalyst model and validated against both lab-scale
and engine test bench data collected over dual-layer ASC systems.
© 2012 Elsevier B.V. All rights reserved.

1. Introduction depending on the operating conditions in order to sustain high lev-

els of NH3 adsorption and of De-NOx conversion. At the same time,
To meet current and future vehicle emission regulations like NH3 emissions from the aftertreatment system into the environ-
EURO VI and Tier 2, the application of exhaust aftertreatment sys- ment must be avoided. For this purpose a so called NH3 slip catalyst
tems is needed in addition to engine internal measures [1]. Selective (ASC) is more and more frequently introduced in modern exhaust
catalytic reduction by NH3 (NH3 -SCR) has become an established gas aftertreatment (EGA) systems, often as a small monolith slice
process to reduce NOx emissions from lean exhaust gas (e.g. diesel after the SCR unit, with the purpose of oxidizing unreacted NH3
exhausts) from passenger cars as well as from light- and heavy- leaving the SCR brick [2].
duty vehicles. The reducing agent NH3 is generated on board from Platinum and other precious metals (PGM) are typically adopted
an aqueous urea solution (AdBlue® ), carried in a separate tank and as very active elements for NH3 oxidation [10]. However, they also
injected into the exhaust gas upstream of the SCR catalyst. NH3 is exhibit a poor selectivity to N2 , i.e. the desired final product. As
adsorbed onto the SCR catalyst and then converted by NOx reduc- a solution, it has been shown that the combination of a PGM-
tion reactions. The injected amount of AdBlue® must be controlled based washcoat with a SCR washcoat according to a dual-layer
architecture is beneficial regarding both conversion and selectiv-
ity performances [11]. Specifically, if the SCR layer is placed on top
∗ Corresponding author. Tel.: +39 02 2399 3264; fax: +39 02 2399 3318. of the oxidation layer, we can expect in principle an increase of the
E-mail address: [email protected] (E. Tronconi). N2 selectivity. In fact, NOx formed by unselective NH3 oxidation

0926-3373/$ – see front matter © 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.apcatb.2012.10.031
862 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876

representative T-range and in the presence of NOx , including NO2 )

Nomenclature in order to develop a dedicated global kinetic model. In addition,
the potential synergies of the catalytic chemistries associated with
Ci gas-phase concentration of species i [mol/m3 gas] the PGM and the SCR active phases were also investigated by
Ej rate parameter for T-dependence of reaction j [K] comparing two different reactor configurations, involving either
k◦ j logarithm of rate constant of reaction j at Tref [–] the sequential arrangement of the two catalysts in a double-bed
kads-PGM rate constant for NH3 adsorption [1/s] configuration, or their intimate mixing in a mechanical mixture.
KNN rate parameter in (Eq. (16)) [m3 /mol] Finally, the last stage of the project, to be reported in a forth-
KNO rate parameter in (Eq. (17)) [m3 /mol] coming companion publication, will involve the integration of the
KNHNOOX rate parameter in (Eq. (18)) [m3 /mol] SCR/PGM kinetic models in a specifically developed mathematical
KNO2 -PGM rate parameter in (Eq. (19)) [m3 /mol] model of dual-layer monolithic catalysts [12], followed by the
eq
KNO equilibrium constant of reaction (R.4) [1/bar0.5 ] lab-scale and full-scale validation of the developed mathematical
2
PO2 partial pressure of O2 [bar] model against data collected over the single layered SCR and PGM
Q flow rate [m3 /s] monoliths, as well as over the full dual-layer ASC monolith catalyst.
R ideal gas constant [J/mol K]
Ri rate of formation of ith species [mol/m3cat s] 2. Methods
rk rate of reaction k [mol/m3cat s]
T temperature [K] 2.1. Experimental
Tref reference temperature [K]
v gas linear velocity [m/s] The commercial dual-layer ASC system herein studied was com-
VCAT volume of catalyst bed [m3 ] posed of state-of-the-art PGM and SCR catalysts: the PGM catalyst
z reactor axial coordinate [m] was a Pt/Al2 O3 -based system, while a Fe-zeolite was used as NH3 -
SCR catalyst component. The original ASC system was a dual-layer
Greek symbols monolith with an SCR layer coated on top of a PGM one; for the
˛ molar feed ratio between NH3 and NOx [–] present study however the two catalysts were supplied in the form
ε void fraction of catalyst bed [–] of precursor powders of the monolith washcoats.
jPGM surface coverage of adsorbed species j on PGM cat- Kinetic runs were carried out in a flow-microreactor consisting
alyst [–] of a quartz tube (6 mm i.d.) placed in an electric oven. Three differ-
jSCR surface coverage of adsorbed species j on SCR cata- ent reactors were prepared for the present study: one was loaded
lyst [–] with the PGM catalyst only, while two others were loaded with
i,k stoichiometric coefficient of species i in reaction k both SCR and PGM catalysts. In the latter case two different con-
 volumetric fraction of PGM catalyst in mechanical figurations were studied, named in the following as “Double Bed”
mixture [–] and “Mechanical Mixture” configuration. In the case of the Double
˝jPGM PGM catalyst: adsorption capacity of species j Bed configuration two consecutive catalyst beds were loaded in the
[mol/m3cat ] microreactor, where the SCR catalytic bed was followed by a PGM
˝jSCR SCR catalyst: adsorption capacity of species j layer, as shown in Fig. 1A. On the opposite, the Mechanical Mixture
configuration involved the intimate mixing of the two powders into
[mol/m3cat ]
a single catalytic bed, as illustrated in Fig. 1B. In both the “Dou-
ble Bed” and “Mechanical Mixture” configurations the amount of
PGM catalyst loaded in the reactor was the same and equal to that
in the PGM layer must diffuse back through the SCR layer above, loaded in the case of runs over the PGM catalyst only, while the rel-
where they can be selectively converted by the residual NH3 to ative proportion of the two catalysts was representative of typical
nitrogen. Thus, such a particular SCR-ASC system enables minimal SCR/PGM washcoat load ratios in dual-layer ASC monoliths.
NH3 breakthrough at increased NH3 and NOx conversions. In all cases before being loaded in the reactor the catalyst pow-
With such a high degree of complexity of modern EGA systems, ders were dried and sieved (140–200 mesh), and diluted with pure
modeling plays an important role as part of the global automotive cordierite powders up to a total load of 160 mg, which provided
development process. The physico-chemical processes on the SCR an acceptable depth of the catalytic bed and, at the same time,
catalyst are already well understood and modeled [3–9], whereas prevented bypass. The reactor outlet was directly connected to a
this is not the case for the ASC, for which only a few literature quadrupole mass spectrometer (Balzers QMS 200, for analysis of
reports exist [2,10] wherein its functionality has been investigated N2 O and N2 ) and, in parallel, to a UV-analyzer (ABB LIMAS 11HW,
and mathematically described in order to enable the simulation for analysis of NO, NO2 and NH3 ). NH3 , NO, NO2 , N2 O, O2 and He
of vehicle tailpipe emissions. Particularly, modeling of the dual- were dosed from bottled calibrated gas mixtures by mass flow con-
layer ASC configuration requires a dedicated study of the processes trollers, while water vapor was added by means of a saturator.
prevailing in each single washcoat layer (SCR + PGM), as well as The catalyst temperature was measured by a K-type thermocou-
a suitable structure of the converter model in order to describe ple immersed in the catalyst bed. In the case of the Double Bed
related synergetic effects. Whereas 1D models of single-layer wash- configuration the thermocouple was immersed in the SCR catalyst
coated catalysts are common in the SCR literature, modeling of bed: based on independent diagnostic measurements of the tem-
dual-layer washcoats has been rarely published so far. perature profile along the axial direction of the reactor, we assumed
The present paper reports part of a systematic work aimed at the same temperature also for the following PGM layer, being both
the development of a chemically consistent mathematic model of layers placed within the isothermal zone of the oven.
dual-layer ASC converters. A preliminary step involved the catalytic The catalysts were pre-conditioned in a T-ramp at 5 ◦ C/min up
activity study and the kinetic modeling of the SCR component in the to 600 ◦ C in 8% O2 (v/v) and 8% H2 O (v/v) followed by hold at
ASC system [3], namely a state-of-the-art Fe-zeolite SCR catalyst. 600 ◦ C for 5 h. Kinetic runs included mainly isothermal steady-
This paper deals with the second step of the model development: state experiments performed in the 150–550 ◦ C temperature range.
the NH3 oxidation over the PGM component of the ASC system At steady state conditions the N-balance always closed within
was herein studied under typical automotive conditions (e.g. in a ±10%, consistently with the accuracy of the analysis apparatus.
 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876 863

Fig. 1. Tested configurations of SCR/PGM combined systems.

The NH3 adsorption/desorption on the PGM catalytic system was z = reactor axial coordinate (m); Ri = rate of formation of ith species,
instead studied by isothermal Transient Response (TRM) runs (NH3 computed according to:
step feed) followed by Temperature Programmed Desorption (TPD)
experiments. Typical feed concentrations of NOx and NH3 ranged 
NR
between 0 and 1000 ppm, always in the presence of O2 (8%, v/v), Ri = rk · vi,k (7)
H2 O (8%, v/v) and balance He. The system was operated at atmo-
k=1
spheric pressure with a Gas Hourly Space Velocity (GHSV) in the
range 325,000–1,093,000 cm3 /h gcat (STP). A detailed description
where rk is the rate of reaction k and i,k is the stoichiometric coef-
of the experimental equipment and procedures can be found else-
ficient of species i in reaction k;
where [13].
˝j = catalyst adsorption capacity of adsorbed species j
Experimental results are frequently presented in the paper in
(mol/m3cat );
terms of NH3 conversion and products yields, calculated according
 j = surface coverage of adsorbed species j (−).
to the following equations:
The rate parameters included in the adopted reaction rates were
    estimated by global multiresponse non-linear regressions based on
NHin out
3 − NH3
NH3 conversion =   · 100 (1) the least squares method. For this purpose the BURENL routine,
NHin
3 developed by Prof. Guido Buzzi-Ferraris, has been used [15,16].
As described in the “Methods” section, the interaction between
 
2 · Nout
2
SCR and PGM components of the ASC system was studied in two dif-
N2 yield =   
in
 · 100 (2) ferent reactor configurations named “Double Bed” and “Mechanical
NHin
3 + NOx Mixture”. PGM/SCR “Double Bed” reactor simulations were based
  on the sequential implementation of SCR and PGM kinetics within
NOout
x the just described reactor model, where the SCR kinetics were those
NOx yield =     · 100 (3)
published in [3]. In view of the arrangement of the catalytic beds
NH3 in
+ NOin
x
(SCR catalyst followed by the PGM catalyst), the output of the SCR
  layer simulation was used as the input for the simulation with the
2 · N2 Oout
N2 O yield =   
in
 · 100 (4) PGM model.
NHin
3 + NOx On the opposite, the simulation of the kinetic data collected
over the SCR/PGM “Mechanical Mixture” configuration called for
a dedicated model accounting for the simultaneous presence of
2.2. Reactor model two different active phases in different amounts. Again, a hetero-
geneous one-dimensional plug-flow dynamic reactor model that
In analogy with the approach used for the SCR component assumes the catalytic bed to be isothermal and isobaric has been
of the ASC [3], experimental data collected over the powdered employed. Perfect mixing and homogeneous distribution of the two
PGM catalyst have been analyzed according to a heterogeneous active phases were assumed, resulting in the following unsteady
one-dimensional dynamic plug-flow reactor model, assuming the differential material balance equations for adsorbed and gaseous
catalyst bed to be isothermal and isobaric. species:
The model is based upon the following unsteady material bal-
ance differential equations for gaseous and adsorbed species [14]:
∂jSCR
SCR adsorbed phase : ˝jSCR = −(1 − ) · RjSCR (8)
∂Ci ∂C ∂t
gaseous phase : ε = −v i + (1 − ε)Ri (5)
∂t ∂z
∂jPGM
∂ PGM adsorbed phase : ˝jPGM = − · RjPGM (9)
adsorbed phase : ˝j i = Ri (6) ∂t
∂t

where ε = void fraction of catalyst bed (−); Ci = gas-phase con- ∂Ci ∂Ci
Gas phase : ε = − · + (1 − ε) RiTOT (10)
centration of ith species (mol/m3 ); v = gas linear velocity (m/s); ∂t ∂z
864 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876

100 100

90 A 90 B
80 80

70 70
NH3 conversion, %

60 60

N2 yield, %
50 50

40 40

30 30

20 20

10 10

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

100 100

90 C 90 D
80 80

70 70
NOx yield, %

60 60
N2O yield, %

50 50

40 40

30 30

20 20

10 10

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 2. NH3 oxidation on PGM catalyst: NH3 = 500 ppm; H2 O = 8% (v/v); O2 = 8% (v/v); carrier gas = He, SV = 1,093,000 cm3 /h/ga.p. (STP).

where: or 1000 ppm of NH3 in the presence of water (8%, v/v) and balance
helium, was followed by a temperature ramp in order to complete

nr-SCR 
nr-PGM
RiTOT = (1 − ) · i,k rkSCR +  · i,k rkPGM (11) ammonia desorption (TPD).
The experiment (see Fig. A in Supplementary material) showed
k=1 k=1
no evidence of ammonia adsorption onto the catalyst surface dur-
VPGM−CAT ing the transient phase. Additionally, the ammonia signal stayed at
= (12)
VTOT−CAT zero ppm throughout the TPD run, indicating no detectable release
VTOT−CAT = VSCR−CAT + VPGM−CAT (13) of ammonia stored during the adsorption phase. However, the lack
of a detectable storage capacity for ammonia (at least under the
Applying the general equations (Eqs. (5)–(7)) (or (Eqs. (8)–(13)) studied conditions) does not rule out the adsorption of the same
to the relevant chemical species present in the reacting system, a species on the catalyst surface:
system of partial differential equations (PDE) is obtained. A dedi-
cated FORTRAN code has been used for numerical integration of the NH3 + ∗ ↔ NH3 ∗ (R.1)
PDE system according to the method of lines. The discretization of
the variables along the axial coordinate was based on finite differ- Indeed, literature reviews on the ammonia oxidation mech-
ences (FD), using typically a grid with 31 equispaced points. The anism over PGM catalysts report the unanimous acceptance
resulting system of ordinary differential equations was then inte- of reaction steps which consider adsorption of both reactant
grated over time using the LSODI library routine, based on Gear’s molecules (NH3 and O2 ) [17]. Moreover, there is general agreement
method for stiff ODEs. that ammonia adsorbs in on-top position on Pt [18] and that such
an adsorption is a non-activated process.
3. Results and discussion

3.1. Identification of the reaction network over the PGM catalyst 3.1.2. NH3 O2
Upon feeding both ammonia and oxygen to the reactor, NH3
3.1.1. NH3 adsorption–desorption oxidation is expected to proceed, as already widely reported in
The ammonia adsorption and desorption properties of the PGM literature [10,17–19]. N2 , N2 O and NOx were identified as main oxi-
catalyst have been investigated by transient two-stage runs: a TRM dation products, with selectivity being a function of the noble metal
run at constant temperature (namely 150 ◦ C), stepwise feeding 500 content and of the catalyst temperature.
 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876 865

550 550
Experimental data 550
500 500
Thermodyn. Eq. 500
450 450 NO 450
T
400 NO 400 400

Concentration, ppm
350
NO
350

Temperature, °C
350
Concentration, ppm

300 300 300

250 250
250
200
200 200 T
150
150 150
100
NO2 100
100 NO2 50
50
NO2
50 0
0 -50
0
5000 10000 15000 20000 25000
-50 Time, s
50 100 150 200 250 300 350 400 450 500 550 600 650
Temperature, °C
Fig. 4. NO oxidation on PGM catalyst, transient runs: NO = 500 ppm; H2 O = 8%
(v/v); O2 = 8% (v/v); carrier gas = He, SV = 1,093,000 cm3 /h/ga.p. (STP).
Fig. 3. NO oxidation on PGM catalyst, steady state data: NO = 500 ppm; H2 O = 8%
(v/v); O2 = 8% (v/v); carrier gas = He, SV = 1,093,000 cm3 /h/ga.p. (STP).
giving NO2 as the only reaction product and reaching a maximum
consumption of about 300 ppm around 350 ◦ C.
Fig. 2 shows the results collected during steady-state NH3 oxida-
It is well known in literature [20] that noble metals are active in
tion runs performed over the PGM catalyst sample. The feed stream
the oxidation of NO according to the reversible reaction:
consisted of 500 ppm ammonia in the presence of water (8%, v/v),
oxygen (8%, v/v) and balance helium. In Fig. 2A the NH3 conversion NO + 1/2O2 ↔ NO2 (R.5)
is plotted versus the catalyst temperature, while Fig. 2B–D show
and the concentration traces in Fig. 3 are consistent with the stoi-
respectively the N2 , NOx (NO + NO2 ) and N2 O yields for the same
chiometry reported above. Since reaction (R.5) is reversible, the
tests.
theoretical concentration levels of NO and NO2 calculated accord-
Focusing first on NH3 conversion (Fig. 2A), a steep reaction light-
ing to the thermodynamic equilibrium constraints are also shown
off was observed between 200 ◦ C and 225 ◦ C, with NH3 reaching
in Fig. 3 (dashed lines). It is evident that below 400 ◦ C the NO
100% conversion already at 225 ◦ C. The evolution of reaction prod-
conversion is kinetically limited, being far from the equilibrium
ucts started in the same T-window, showing a significant change
values. On the opposite, outlet NO and NO2 concentrations consis-
of products selectivity with temperature. N2 was the first species
tent with thermodynamic equilibrium constraints are approached
to be produced, reaching a maximum yield of about 60% at 225 ◦ C.
above 400 ◦ C, resulting in a decrease of NO conversion as the tem-
A further increase of the catalyst temperature resulted then in the
perature is increased.
progressive decrease of N2 production. N2 O also showed a maxi-
An interesting feature of NO oxidation over the investigated
mum behavior, centered around 250 ◦ C. Above 250 ◦ C the NOx yield
PGM catalytic system was highlighted during catalyst heat-up
started to monotonically increase, reaching a maximum of about
transients. Fig. 4 shows the temporal evolution of species outlet
80% at 550 ◦ C. Correspondingly a progressive reduction of both N2
concentrations and temperature for the T-steps run whose steady-
and N2 O yields was observed. The same test was also repeated with
state values were plotted in Fig. 3. Between 175 and 250 ◦ C, as soon
a lower oxygen feed content (2%, v/v, results not shown), showing
as the catalyst temperature was stepwise increased, the NO trace
negligible variations in terms of both NH3 conversion and N2 , N2 O
and NOx yields. Notably, in spite of the same amount of NOx pro-
duced, the relative proportions of NO and NO2 were significantly 60
different in the two tests, being controlled by the NO oxidation
reaction (R.5). 50
Species outlet concentrations in the studied temperature range
in
are consistent with the stoichiometry of the following global NO2 = 0 ppm
reactions, which proceed in parallel and prevail at different tem- 40
NO Conversion, %

peratures, in line with previous literature indications [10,17,18]:

30
2NH3 + 3/2O2 → N2 + 3H2 O (R.2)

2NH3 + 2O2 → N2 O + 3H2 O (R.3) 20

in
NO2 = 250 ppm
2NH3 + 5/2O2 → 2NO + 3H2 O (R.4)
10
NO + 1/2O2 ↔ NO2 (R.5)

0 in
3.1.3. NO O2 NO2 = 500 ppm
A set of experimental runs was dedicated to study the catalyst 100 150 200 250 300 350
behavior in the presence of NO and O2 : steady state activity tests
Temperature, °C
feeding to the reactor NO (500 ppm), O2 (8%, v/v), H2 O (8%, v/v)
and balance helium were performed in the 150–550 ◦ C temperature Fig. 5. NO oxidation on PGM catalyst, NO2 concentration effect: NO = 500 ppm;
range. Fig. 3 shows steady-state species outlet concentrations mea- NO2 = 0-250-500 ppm; H2 O = 8% (v/v); O2 = 8% (v/v); carrier gas = He,
sured during one of such runs: NO started to be converted at 150 ◦ C SV = 1,093,000 cm3 /h/ga.p. (STP).
866 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876

A
100 100

B
80 80
NH3 conversion, %

N2 yield, %
60 60

40 40

0 ppm NO
20 100 ppm NO 20
250 ppm NO
500 ppm NO
0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

C
100 100

D
80 80
NOx yield, %

N2O yield, %

60 60

40 40

20 20

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 6. NH3 /NO reacting system on PGM catalyst, NO concentration effect: NH3 = 500 ppm; NO = 0-100-250-500 ppm, NO2 /NOx = 0; H2 O = 8% (v/v); O2 = 8% (v/v); carrier
gas = He, SV = 1,093,000 cm3 /h/ga.p. (STP).

suddenly dropped with a simultaneous evolution of NO2 . Imme- 3.1.4. NH3 NO O2

diately thereafter, however, although the temperature was kept The effect of NO on NH3 oxidation was studied in steady state
constant and a steady-state behavior was expected, the NO signal runs where NO (100–250–500 ppm) was fed to the reactor together
started slowly but steadily to rise again, while the release of NO2 with 500 ppm of NH3 , 8% (v/v) O2 , 8% (v/v) H2 O and balance
correspondingly decreased. helium.
According to the literature on transient experiments performed Results of such tests are compared in Fig. 6 with those
on noble metal systems similar to the present PGM catalyst [20], the obtained in the absence of NO in the feed stream (green square
slow decrease in NO oxidation activity is ascribed to the strongly dots). The figure shows the comparison in terms of NH3 con-
oxidizing nature of NO2 , which is responsible for the formation of version (Fig. 6A) and N2 , NOx and N2 O yields (Fig. 6B–D,
noble metal oxides/chemisorbed oxygen on the catalyst surface, respectively).
reducing the catalyst activity. The decreasing NO conversion is thus The addition of increasing amounts of NO to the feed stream
a result of the NO2 formed during the reaction, which oxidizes the Pt had a great impact on both NH3 conversion and products yields. In
catalyst. The dynamic behavior observed during the NO oxidation the low temperature range (T < 250 ◦ C) increasing the NO feed con-
run in Fig. 4 is hence in accordance with literature evidence. tent drastically promoted the NH3 conversion, which was limited
In order to collect further evidence on the role of NO2 in slowing to less than 15% at 200 ◦ C in the absence of NO, but approached
down the NO oxidation rate, steady-state NO oxidation exper- 100% when co-feeding 500 ppm of NO (Fig. 6A). This means a drop
iments with 500 ppm NO, 8% (v/v) O2 and 8% (v/v) H2 O were of about 50 ◦ C of the reaction light-off temperature in comparison
performed at different NO2 feed contents, namely 250 ppm and with the NH3 O2 reacting system. However, the N2 yield (Fig. 6B)
500 ppm. For each run, the steady-state NO conversions are plot- progressively decreased below 275 ◦ C with the addition of increas-
ted versus temperature in Fig. 5. Since NO2 decomposition to NO ing amounts of NO, indicating a greater formation of undesired
occurs above 350 ◦ C, only data up to 350 ◦ C are shown. It is observed oxidation products: in particular, the N2 O yield showed indeed
that the steady-state NO conversion decreases with increasing NO2 , an opposite trend, growing with the NO feed content (Fig. 6D).
especially in the high temperature range: significant NO2 inhibition This experimental evidence is well supported by the literature
of the NO oxidation activity is thus clearly established. [10,17,18]. The enhanced N2 O production in the presence of NO was
 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876 867

100 100
A B
80 80
NH3 conversion, %

60 60

N2 yield, %
40 40
0 ppm NO2
100 ppm NO2
20 20
250 ppm NO2
500 ppm NO2
0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

100 100
C D
80 80
NOx yield, %

60 60
N2O yield, %

40 40

20 20

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 7. NH3 /NO2 reacting system on PGM catalyst, NO2 concentration effect: NH3 = 500 ppm; NO2 = 0-100-250-500 ppm, NO2 /NOx = 1; H2 O = 8% (v/v); O2 = 8% (v/v); carrier
gas = He, SV = 1,093,000 cm3 /h/ga.p. (STP).

attributed to the occurrence of the following unselective Standard- Fig. 7 shows the experimental results obtained feeding 500 ppm
SCR reaction: of NH3 and alternatively 100–250–500 ppm of NO2 to the reactor
while flowing oxygen (8%, v/v), water (8%, v/v) and balance helium.
2NH3 + 2NO + 3/2O2 → 2N2 O + 3H2 O (R.6) Temperature steps were performed while continuously feeding the
mentioned reactant mixture. Experimental results are also com-
A different behavior was observed at high temperatures, namely pared with those obtained in the absence of NO2 in the feed stream
above 400 ◦ C, with the N2 and N2 O yields not significantly affected (green square dots).
by NO addition. Finally, the trend of NOx yield as a function of The addition of NO2 to the feed resulted in a lower consumption
temperature (Fig. 6C) highlights that NO was consumed at low tem- of NH3 (Fig. 7A) in the 200–275 ◦ C temperature range. Complete
peratures (150–225 ◦ C T-range) and produced at higher ones, up to conversion of NH3 was indeed approached at 225 ◦ C in the absence
550 ◦ C, where the NOx yield exceeded 90% independently of the of NO2 , while total conversion was achieved only at 275 ◦ C when
NO feed content. Summarizing, the main finding associated with co-feeding also 500 ppm of NO2 . Above this temperature complete
the analysis of N2 and NOx yields as a function of temperature is ammonia conversion was approached for all reacting mixtures. The
that the N2 O production was highly enhanced below 275 ◦ C in the N2 yield was also influenced by the presence of NO2 (Fig. 7B):
presence of NO. the addition of 100 ppm of NO2 apparently suppressed N2 forma-
tion, with a decrease of the maximum yield from about 60% down
3.1.5. NH3 NO2 O2 to about 40%. However, nitrogen evolution resulted enhanced by
NO2 is commonly present in the feed stream to mobile SCR con- greater concentrations of NO2 , with a shift of the maximum yield
verters due to the inclusion of an upstream preoxidation catalyst to higher temperatures with growing NO2 feed contents. Above
in the configuration of aftertreatment systems. In addition, under 300 ◦ C almost the same yield was obtained both in the presence
unfavorable operative conditions NO2 might undergo an incom- and in the absence of NO2 . Fig. 7C shows how NOx , actually NO2 ,
plete selective reduction by NH3 : therefore, it could also represent was partially consumed at temperatures as high as 250 ◦ C; on
a possible side product at the outlet of the SCR converter. Hence, further increasing the catalyst temperature the NOx yield mono-
being NO2 a likely component of the ASC feed mixture, the inves- tonically increased independently of the feed mixture composition,
tigation of the NH3 /NO2 reacting system was also included in the approaching almost 100% at 550 ◦ C. The formation of N2 O (Fig. 7D)
present work. showed instead a progressive decrease with increasing NO2 feed
868 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876

A
100 100

B
80 80
NH3 conversion, %

N2 yield, %
60 60

40 40

0 ppm NOx
20 20
150 ppm NOx
250 ppm NOx
500 ppm NOx
0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

100 100
C D
80 80
NOx yield, %

N2O yield, %
60 60

40 40

20 20

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 8. NH3 /NOx reacting system on PGM catalyst, NOx concentration effect: NH3 = 500 ppm; NOx = 0-150-250-500 ppm, NO2 /NOx = 0.5; H2 O = 8% (v/v); O2 = 8% (v/v); carrier
gas = He, SV = 1,093,000 cm3 /h/ga.p. (STP).

content: a maximum yield of about 30% at 300 ◦ C was recorded in 3.1.6. NH3 NO NO2 O2
the presence of 500 ppm of NO2 , against a maximum of about 60% Finally, the complete NH3 /NO/NO2 O2 reacting system was also
in the absence of NO2 . At high temperatures (T > 400 ◦ C), in analogy investigated in steady state runs: 500 ppm of NH3 were fed to the
with what observed for the N2 yield, the amount of produced N2 O reactor together with 150–250-500 ppm of NOx (NO2 /NOx = 0.5),
seemed to be independent of NO2 . 8% (v/v) O2 , 8% (v/v) H2 O and balance helium. Temperature steps
From the analysis of NH3 conversion and products yields we can were performed while continuously feeding the mentioned reac-
speculate that the NH3 oxidation reactions were inhibited by NO2 , tants mixture.
specifically in the low-T region (T < 275 ◦ C), resulting in a decrease of Fig. 8 shows collected experimental results in terms of NH3
the NH3 conversion with increasing NO2 feed contents. The experi- conversion (Fig. 8A), N2 yield (Fig. 8B), NOx yield (Fig. 8C) and
mental evidence of such an inhibition effect was clear, but a detailed N2 O yield (Fig. 8D) in comparison with results collected in the
mechanistic study of this phenomenon was not in the scope of this absence of NOx in the feed stream (green square dots). Fig. 8A shows
study and will be performed in the next future. that the addition of increasing amounts of NOx had only a slight
In addition to the inhibited NH3 oxidation reactions, the NOx impact on the NH3 conversion, with a slight decrease of catalyst
consumption observed in the low-T range justified the introduction activity in the 200–250 ◦ C temperature region. NOx had negligi-
of two additional reactions describing the reactivity between NH3 ble effects on the N2 yield over the whole T-range (Fig. 8B). The
and NO2 : NOx yield (Fig. 8C) shows a trend similar to those observed for the
NH3 /NO and NH3 /NO2 reacting systems: NOx were partially con-
2NH3 + 2NO2 → N2 O + N2 + 3H2 O (R.7) sumed at temperatures below 250 ◦ C, then on further increasing
the catalyst temperature the NOx yield monotonically increased
4NH3 + 3NO2 → 7/2N2 + 6H2 O (R.8) independently of the feed mixture composition, approaching 100%
at 550 ◦ C. Regarding N2 O formation (Fig. 8D), the highest yields
The occurrence of these reactions was independent of the oxy- were recorded either in the absence of NOx or in the presence of
gen presence in the feed stream, as verified performing a diagnostic the highest NOx feed content (500 ppm), with a maximum centered
test in the absence of oxygen (results not shown). around 250 ◦ C.
 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876 869

In summary, it seems that in the low-T region (T < 275 ◦ C) Notably, a re-parameterized form of the desorption rate con-
NO2 inhibits both the NH3 oxidation reactions and the reactivity stant was used, with the purpose of reducing correlation of the
between NH3 and NO (reaction R.6). Indeed, the activity enhance- rate parameter estimates during parameter fitting. The same re-
ment provoked by NO in the feed mixture, as pointed out by Fig. 6, parameterized form of the rate constant was also implemented in
was not observed in the presence of NO2 . Furthermore, there was all the rate equations defined in the following paragraphs.
no direct evidence of additional reactions with respect to those
identified during the investigation of the NH3 O2 , NH3 /NO/O2 and 3.2.2. NH3 oxidation and non-selective standard SCR
NH3 /NO2 /O2 reacting systems. Three reactions, namely (R.2), (R.4) and (R.6), were included in
the kinetic model to describe the reactivity between NH3 and O2 ,
3.2. Kinetic model for the PGM catalyst both in the presence and in the absence of NO.
For the three reactions we assumed first order kinetics with
Based on the experimental data analyzed in Section 3.1, respect to the NH3 surface coverage, and a negative dependence
the following global reaction network was identified for the on the NO2 concentration. Experimental runs performed over the
NH3 /O2 /NO NO2 reactivity over the PGM component of the stud- PGM catalyst showed indeed that NH3 oxidation as well as the reac-
ied ASC system: tivity between NH3 and NO are both inhibited by NO2 , which could
be present either as the product of NO oxidation or as a component
NH3 adsorption/desorption : NH3 + ∗ ↔ NH3 ∗ (R.1)
of the feed mixture. Indeed, a detailed mechanistic investigation of
∗
NH3 oxidation to N2 : 2NH3 + 3/2O2 → N2 + 3H2 O (R.2) the NO2 inhibition phenomenon was beyond the scope of the PGM
∗ kinetic model development. In view of our current lack of under-
NH3 oxidation to NO : 2NH3 + 5/2O2 → 2NO + 3H2 O (R.4)
standing, we have therefore chosen a purely empirical approach
NO oxidation to NO2 : NO + 1/2O2 ↔ NO2 (R.5) in order to describe NO2 inhibition of the kinetics. Accordingly,
we introduced LHHW-type terms in the denominators of rate Eqs.
(16)–(18), thus accounting for their negative order dependence on
Non-Selective Standard SCR : 2NH3 ∗ + 2NO + 3/2O2 the NO2 concentration, without however necessarily invoking the
competitive adsorption of NH3 and NO2 on the same catalytic site.
→ 2N2 O + 3H2 O (R.6) Such inhibition terms thus are not associated with a single NO2
adsorption constant, but rather with multiple fitting parameters.
Additionally, the assumed kinetics were zero order with respect
NH3 /NO2 to N2 and N2 O : 2NH3 ∗ + 2NO2 to oxygen, based on the observed negligible variations of catalyst
→ N2 + N2 O + 3H2 O (R.7) activity and products selectivity over the 2%–8% (v/v) O2 concen-
tration range. Finally, for reaction R.6 first order kinetics were
considered with respect to the NO gas phase concentration,
0 −E PGM

exp kNN NN (1000/T − 1000/473) · NH
NH3 oxidation to N2 : rNN =   3
(16)
1 + KNN · CNO2
0 −E PGM

exp kNO NO (1000/T − 1000/473) · NH
NH3 oxidation to NO : rNO =   3
(17)
1 + KNO · CNO2

0 PGM C

exp kNHNOOX − ENHNOOX (1000/T − 1000/473) · NH NO
NH3 /NO/O2 to N2 O : rNHNOOX =   3
(18)
1 + KNHNOOX · CNO2

It is worth emphasizing that Fig. 2, as well as literature data on

PGM catalysts, show formation of N2 O from NH3 oxidation also in
NO2 -SCR to N2 : 4NH3 ∗ + 3NO2 → 7/2N2 + 6H2 O (R.8) the absence of NO in the feed stream. Notwithstanding such exper-
imental and literature evidence, no direct pathway from NH3 + O2
Based on this set of reactions, a consistent global kinetic model to N2 O was considered in the herein developed kinetic scheme. The
was developed, as presented and discussed in the following. sum of (R.4) and (R.6) results however in the stoichiometry of NH3
oxidation to N2 O:
3.2.1. NH3 adsorption/desorption
2 NH3* 5 2 O2 2 NO 3H 2O
In a first attempt of model development, as the experimental (R.4)
data presented in Section 3.1 showed negligible ammonia adsorp- *
NH NO 3 4 O2 N2O 3 2 H 2O
tion capacity onto the PGM catalyst, the description of adsorption 3 (R.6)
phenomena was avoided. However this resulted in the incorrect
2 NH3* 2O2 N 2O 3H 2O
description of catalyst activity in the high temperature regime, (R.3)
where NH3 adsorption likely becomes rate controlling for the oxi- The present reaction scheme can thus in principle describe also
dation reactions [10]. For this reason, and in line with literature the N2 O formation observed in the absence of NO in the feed stream.
evidence [17,18], ammonia adsorption was assumed to proceed
on Pt sites as a non-activated and molecular process, while an 3.2.3. NO oxidation
Arrhenius rate law was used to describe the desorption process. The oxidation of NO to NO2 is a reversible reaction, well stud-
Accordingly, the following rate equations were adopted: ied in the literature over PGM catalytic systems [21–23]. Different
 
NH3 adsorption : PGM
rads = kads−PGM CNH3 1 − NH (14) rate expressions as well as mechanistic studies can be traced in the
3
literature on this topic. Two main aspects characterize the kinetics
NH3 desorption : of this reaction: the first one is that thermodynamic equilibrium
  1000 1000
0 PGM (15) constraints must be respected, the second one is that a strong
rdes = exp kdes-PGM − Edes-PGM − · NH
T 473 3 inhibiting effect of NO2 on the NO oxidation over PGM systems
870 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876

has been extensively reported [21,22]. In line both with literature 100
evidence and with the collected data, a consistent reaction rate was NH3

NH3 conversion, N2, N2O & NOx yield, %

adopted (Eq. (19)), which takes into account both the equilibrium 90
NOx
constraints and the aforementioned inhibiting effect of NO2 : 80

NO oxidation to NO2 :
  1000 70
0 1000 N2 O
rNOox = exp kNOox−PGM − ENOox−PGM − 60
T 473
 eq
 (19)
50
CNO PO2 − CNO2 /KNO
×  2 40
1 + KNO2 −PGM · CNO2
30

20
3.2.4. Reactions involving NH3 and NO2 N2
Two reactions were included in the kinetic model to account for 10
the reactivity of NH3 with NO2 :
0
NH3 /NO2 to N2 and N2 O : 2NH∗3 + 2NO2 → N2 + N2 O + 3H2 O 100 150 200 250 300 350 400 450 500 550 600

(R.7) Temperature, °C

Fig. 9. NH3 oxidation on PGM catalyst, kinetic fit: NH3 = 500 ppm; H2 O = 8% (v/v);
NO2 − SCR to N2 : 4NH∗3 + 3NO2 → 7/2N2 + 6H2 O (R.8) O2 = 8% (v/v); carrier gas = He, SV = 1,093,000 cm3 /h/ga.p. (STP). Symbols: experimen-
tal results; lines: model fit.
Second order overall kinetics in ammonia surface coverage and
NO2 gas phase concentration were adopted for both reactions (R.7)

100 100
A B
80 80
NH3 conversion, %

N2 yield, %

60 60

40 40

NO2/NOx = 0
20 20
NO2/NOx = 0.5
NO2/NOx = 1
0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

100
100
C D
80
80
N2O yield, %
NOx yield, %

60 60

40 40

20
20

0
0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 10. NH3 /NO NO2 reacting system on PGM catalyst, effect of NO2 /NOx ratio and kinetic fit for alpha = 1: NH3 = 500 ppm; NOx = 500 ppm; NO2 /NOx = 0, 0.5, 1; H2 O = 8%
(v/v); O2 = 8% (v/v); carrier gas = He, SV = 1,093,000 cm3 /h/ga.p. (STP). Symbols: experimental results; lines: model fit.
 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876 871

A
100 100

B
80 PGM 80
Mechmix
NH3 conversion, %

Double bed
SCR

N2 yield, %
60 60

40 40

20 20

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

100 100

C D
80 80

N2O yield, %
NOx yield, %

60 60

40 40

20 20

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 11. NH3 oxidation, comparison of different reactor configurations: NH3 = 500 ppm; H2 O = 8% (v/v); O2 = 8% (v/v); carrier gas = He. SVPGM = 1,093,000 cm3 /h/ga.p. (STP),
SVSCR = 690,000 cm3 /h/ga.p. (STP).

and (R.8). This resulted in similar rate equations, namely (Eqs. (20) calculated (solid line) ammonia conversion and N2 , N2 O and NOx
and (21)). yields are plotted versus the catalyst temperature. The model
  1000 1000
NH3 /NO2 to N2 and N2 O : 0
rNHNOO = exp kNHNOO − ENHNOO − PGM C
· NH (20)
NO2
T 473 3

  1000 1000
NO2 − SCR to N2 : 0
rNOO = exp kNOO − ENOO − PGM C
· NH (21)
NO2
T 473 3

Reactions (R.1), (R.2) and (R.4)–(R.8), with the related rate equa-
tions (Eqs. (14)–(21)), thus provide a global kinetic scheme for the
PGM component of the investigated ASC system.
provides a fairly good description of the data over the whole
150–550 ◦ C T-range. The onset of NH3 conversion is indeed well
3.3. Kinetic fit captured as well as the evolution of products yields as a function of
catalyst temperature. Minor deviations can be observed between
Experimental data collected over the powdered PGM cat- 250 ◦ C and 300 ◦ C for the N2 yield and above 400 ◦ C for the NOx
alysts have been analyzed according to the heterogeneous yield.
one-dimensional plug-flow dynamic reactor model described in For the NO O2 reacting system (see Fig. B in Supplemen-
Section 2. Rate parameters of (Eq. (19)) were independently esti- tary material) the model was able to describe the experimentally
mated from NO oxidation tests while a global multiresponse observed trend of steady-state NO conversions over the whole
non-linear regression based on the least squares method applied investigated temperature range, both in the kinetically (T < 400 ◦ C)
to the whole set of runs involving NH3 provided estimates of the and in the thermodynamically (T > 400 ◦ C) controlled regimes.
remaining rate parameters in Eqs. (14)–(21). Fig. 10 illustrates the effects of the NO2 /NOx feed ratio on NH3
Kinetic fit results are presented in Figs. 9 and 10. Fig. 9 refers conversion and N2 , NOx and N2 O yields when feeding to the PGM
to the NH3 oxidation runs, where experimental (symbols) and catalysts equimolar amounts of NH3 and NOx (˛ = 1). Experimental
872 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876

100 100

A B
80 80
NH3 conversion, %

60 60

N2 yield, %
40 40

20 20

Mechmix
Double bed
0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

100 100

C D
80 80
NOx yield, %

N2O yield, %
60 60

40 40

20 20

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 12. NH3 /NO reacting system on SCR + PGM catalysts, Double Bed versus Mechanical Mixture: NH3 = 750 ppm; NO = 250 ppm; H2 O = 8% (v/v); O2 = 8% (v/v); carrier
gas = He, SVPGM = 1,093,000 cm3 /h/ga.p. (STP). SVSCR = 690,000 cm3 /h/ga.p. (STP).

data (symbols) are plotted versus temperature together with fit catalytic chemistries in relation to limiting (segregated versus
results (solid lines). mixed) spatial arrangements of the two catalysts. Two catalytic
The effect of the NO2 /NOx ratio in terms of NH3 conversion is bed layouts have been in fact designed to achieve these goals: (i)
well captured by the model, with a decrease of catalyst activity a Double Bed configuration, with the first bed and the second bed
on increasing the NO2 content. The qualitative trends of N2 , NOx consisting of the SCR catalyst and of the PGM catalyst, respectively,
and N2 O yields are also well captured as a function of both tem- provides information on the sequential combination of the respec-
perature and NO2 /NOx ratio, though the quantitative agreement is tive catalytic activities; (ii) a single bed configuration, wherein the
not always perfect. Some deviations can indeed be observed below reactor is loaded with a mechanical mixture of the two catalyst
275 ◦ C for NO2 /NOx = 0.5 and above 300 ◦ C for NO2 /NOx = 1. powders, can suitably provide experimental evidence of local par-
The same qualitative results are observed also in case of excess allel interactions between the catalytic chemistries.
ammonia in the feed stream (see Fig. C in Supplementary material Like for the PGM catalyst, both the “Double Bed” and “Mechani-
for fit results with ˛ = 5). cal Mixture” configurations were studied by means of steady state
runs addressing the main SCR reacting systems, namely NH3 /O2 ,
3.4. Analysis of SCR/PGM interaction NH3 /NO/O2 , NH3 /NO2 /O2 and NH3 /NO NO2 /O2 .

As discussed in Section 2, the dual-layer ASC configuration 3.4.1. NH3 O2

involves the presence of a PGM catalytic layer beneath the SCR In order to figure out the role that the combination of different
coating. A micro-flow reactor loaded with powder is inherently catalysts plays on reactivity and selectivity of ammonia oxidation,
unsuitable for reproducing the real ASC system configuration: Fig. 11 compares data collected over the SCR catalyst only, the
above all, no arrangement of catalytic beds exists which can PGM catalyst only and over the two SCR/PGM combined systems.
effectively simulate the NOx back-diffusion and reaction with The analysis is based on NH3 conversion, N2 , NOx and N2 O yields
the ammonia pre-adsorbed on the SCR catalyst. Accordingly, the obtained when feeding to the catalysts 500 ppm of NH3 , 8% (v/v)
fundamental study performed over SCR-PGM combined systems O2 , 8% (v/v) H2 O and balance Helium.
definitely does not claim to faithfully simulate the real ASC sys- Observing the data in Fig. 11, it is clearly apparent that the cata-
tem behavior. The aim of this investigation was rather to elucidate lyst composition dramatically affects both the ammonia oxidation
experimentally the interactions between the SCR and the PGM activity and the products selectivity. The SCR catalyst shows the
 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876 873

100 100
A B
80 80
NH3 conversion, %

60 60

N2 yield, %
40 40

20 20
Mechmix
Double bed
0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

100 100

C D
80 80

N2O yield, %
NOx yield, %

60 60

40 40

20 20

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 13. NH3 /NO2 reacting system on SCR + PGM catalysts, Double Bed versus Mechanical Mixture: NH3 = 750 ppm; NO2 = 250 ppm; H2 O = 8% (v/v); O2 = 8% (v/v); carrier
gas = He. SVPGM = 1,093,000 cm3 /h/ga.p. (STP). SVSCR = 690,000 cm3 /h/ga.p. (STP).

weakest oxidation performance, resulting in incomplete ammonia conversion was indeed obtained already at 275 ◦ C, with N2 yields
oxidation with almost total selectivity towards N2 at all tem- above 40% over the whole 150–550 ◦ C T-range. Correspondingly,
peratures, as well-established in the literature for Fe-zeolite SCR lower NOx and N2 O yields compared with PGM catalyst only and
catalysts [3,7,8,24]. Double Bed configuration were measured between 300 ◦ C and
On the opposite, the PGM catalyst exhibits the highest ammonia 500 ◦ C.
oxidation activity, with NH3 conversion above 90% already at 200 ◦ C The direct comparison between the two ASC combined sys-
(Fig. 11A). However the same catalyst is also the one associated with tems (Double Bed versus Mechanical Mixture) points out how the
the highest yield of NOx in the high T-range (see Fig. 11C). mechanical mixture of SCR and PGM powders entails the occur-
When the SCR catalyst is followed by the PGM one (“Double rence of interactions between the chemistries prevailing over the
Bed” configuration), the overall activity is strengthened due to the catalysts, which instead do not take place in the Double Bed con-
presence of the PGM catalyst: NH3 conversion and products yields figuration where the two catalysts are segregated. The different
are indeed almost overlapped with those of the PGM only cata- selectivity observed in the simultaneous presence of both cata-
lyst below 300 ◦ C, with some deviations only in terms of N2 yield. lysts are due in fact to the competition between the SCR and PGM
However, on moving to higher temperatures the contribution of reactivities, as the SCR reaction products can react over the PGM
the upstream SCR layer becomes significant, as pointed out by the catalyst and vice versa. A similar picture was observed by Long et al.
increased N2 yield above 500 ◦ C (Fig. 11B) and the corresponding [25] who studied a noble metal promoted Fe-ZSM-5 for Selective
decrease of the NOx yield (Fig. 11C). This trend is likely related to Catalytic Oxidation of NH3 . Long et al. [25] attributed the reduced
the partial selective conversion of NH3 to N2 in the SCR catalyst N2 O and NOx production to the SCR features of the combined cata-
bed. As a result a lower amount of unreacted ammonia is available lyst: N2 O and NO generated by NH3 oxidation over the noble metal
for the oxidation to NO occurring in the following PGM bed. can be indeed further reduced to N2 by unreacted ammonia, thus
The best compromise between activity and selectivity was improving simultaneously both ammonia conversion and N2 selec-
achieved with the PGM/SCR mechanical mixture: complete NH3 tivity.
874 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876

100 3.4.2. NH3 NOx O2

The combined SCR/PGM systems were comparatively analyzed
90 NH3
in terms of steady state activity and products yields in the case
NH3 conversion, N2 & NOx yield, %

80 of a feed mixture composed of 750 ppm of NH3 , 250 ppm of NOx ,

NOx 8% (v/v) O2 , 8% (v/v) H2 O and balance Helium. The comparisons of
70
NH3 conversion, N2 , NOx and N2 O yields are shown for NO2 /NOx
N2 O
60 feed ratios of 0 (Fig. 12) and 1 (Fig. 13). For NO2 /NOx of 0.5 refer to
50
Fig. D in Supplementary material.
For all NO2 /NOx ratios no significant differences between “Dou-
40 ble bed” and “mechanical mixture” configurations were apparent
30 N2 in terms of NH3 conversion. On the opposite, the different arrange-
ment of the two catalysts in the reactor played a role in determining
20
products yields. In terms of N2 yields the picture was complicated
10 and varied depending on the NO2 /NOx feed ratio. In the case of
NO2 /NOx = 0 (Fig. 12B), a higher N2 yield was recorded over the
0
mechanical mixture below 350 ◦ C, while at higher temperatures the
100 150 200 250 300 350 400 450 500 550 600 Double Bed configuration prevailed. A constantly higher N2 yield
Temperature, °C over the Double Bed (Fig. 13B) was also recorded in the case of
NO2 /NOx = 1.
Fig. 14. NH3 oxidation on Double Bed (SCR + PGM) configuration, model Focusing on the NOx yield, a common qualitative trend is shown
simulations: NH3 = 500 ppm; H2 O = 8% (v/v); O2 = 8% (v/v); carrier gas = He.
SVPGM = 1,093,000 cm3 /h/ga.p. (STP). SVSCR = 690,000 cm3 /h/ga.p. (STP). Symbols:
independently of the NO2 /NOx ratio (Figs. 12C and 13C): at low
experimental results; lines: model simulations. temperatures the Double Bed and Mechanical mixture configura-
tions exhibited the same NOx yields, while at high temperatures
(above 350 ◦ C or 450 ◦ C depending on the NO2 /NOx ratio) the lower

100 100

90
A 90 B
80 80

70 70
NH3 conversion, %

N2 yield, %

60 60

50 50

40 40

30 30 NO2/NOx = 0
NO2/NOx = 0 NO2/NOx = 0.5
20 20
NO2/NOx = 0.5 NO2/NOx = 1
10 10
NO2/NOx = 1
0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

100 100

90 C NO2/NOx = 0 90 D
80 NO2/NOx = 0.5 80 NO2/NOx = 0
NO2/NOx = 1 NO2/NOx = 0.5
NOx yield, %

70 70
N2O yield, %

NO2/NOx = 1
60 60

50 50

40 40

30 30

20 20

10 10

0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 15. NH3 /NO NO2 reacting system on Double Bed (SCR + PGM) configuration, effect of NO2 /NOx ratio, model simulations: NH3 = 750 ppm; NOx = 250 ppm; H2 O = 8%
(v/v); O2 = 8% (v/v); carrier gas = He. SVPGM = 1,093,000 cm3 /h/ga.p. (STP). SVSCR = 690,000 cm3 /h/ga.p. (STP). Symbols: experimental results; lines: model simulations.
 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876 875

NOx yields were always obtained in the case of the Double Bed 100
configuration.
NH3
In the high temperature region (T > 350 ◦ C), the general decrease 90

NH3 conversion, N2 & NOx yield, %

of the NOx yield in favor of the N2 yield over the Double Bed configu- 80
ration was likely related to the occurrence of the NH3 -SCR reactions NOx
70
in the upstream SCR bed. The temperature increase boosted indeed
the ammonia and NOx consumption over the SCR catalyst in the 60
Double Bed configuration, providing a reduced ammonia and NOx 50
supply to the PGM catalyst bed downstream, thus resulting in a
reduced NOx evolution. 40
Finally, the effect of the reactor configuration on the N2 O yield 30 N2
resulted to be a function of the NO2 /NOx feed ratio: when NO was
20
the only source of NOx (Fig. 12D) the mechanical mixture exhibited
a lower N2 O yield in the whole 150–550 ◦ C T-range in comparison 10 N2 O
with the Double Bed configuration. The latter system showed how-
0
ever lower N2 O yields when NO2 was present in the feed stream
(Fig. 13D). 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C

Fig. 16. NH3 oxidation on Mechanical Mixture (SCR + PGM) configuration,

3.5. Predictive simulations of SCR/PGM combined systems model simulations: NH3 = 500 ppm; H2 O = 8% (v/v); O2 = 8% (v/v); carrier gas = He.
SVPGM = 1,093,000 cm3 /h/ga.p. (STP). SVSCR = 690,000 cm3 /h/ga.p. (STP). Symbols:
experimental results; lines: model simulations.
Experimental data collected over both Double Bed and Mechan-
ical mixture configurations were systematically compared with
predictive model simulations, as shown in Figs. 14–17. The simula-
tions were generated simply assuming the superposition of the SCR

100 100

A B
80 80
NH3 conversion, %

N2 yield, %

60 60

40 40

20 NO2/NOx = 0 20
NO2/NOx = 0.5
NO2/NOx = 1
0 0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C
100
100
C D
80
80
NOx yield, %

60
N2O yield, %

60

40 40

20
20

0
0

100 150 200 250 300 350 400 450 500 550 600 100 150 200 250 300 350 400 450 500 550 600
Temperature, °C Temperature, °C

Fig. 17. NH3 /NO NO2 reacting system on Mechanical Mixture (SCR + PGM) configuration, effect of NO2 /NOx ratio, model simulations: NH3 = 750 ppm; NOx = 250 ppm;
H2 O = 8% (v/v); O2 = 8% (v/v); carrier gas = He. SVPGM = 1,093,000 cm3 /h/ga.p. (STP). SVSCR = 690,000 cm3 /h/ga.p. (STP). Symbols: experimental results; lines: model simulations.
876 M. Colombo et al. / Applied Catalysis B: Environmental 142–143 (2013) 861–876

and PGM kinetics independently determined over the two pow- Such a positive interaction between the PGM and the SCR catalytic
dered catalyst components of the ASC system. chemistries was satisfactorily predicted by a model involving the
Fig. 14 shows simulation results for the Double Bed configu- simple superposition of the PGM and SCR kinetics.
ration in the case of the NH3 /O2 reacting system. A good match In summary, the present preliminary fundamental study has
between experimental data (symbols) and simulations (solid lines) confirmed a positive interaction between SCR and PGM catalytic
is evident for both ammonia conversion and NOx yield, while some chemistries, which beneficially affects the N2 selectivity. This
deviations are present between 275 ◦ C and 500 ◦ C for the N2 and promising synergy needs to be further verified in the real dual-
N2 O yields. layer monolith configuration of the ASC system. Accordingly, in
For the same configuration, the effect of NO2 /NOx is shown in the following part of the project the herein developed PGM kinet-
Fig. 15. Simulations (solid lines) resulted in good agreement with ics, together with consistent SCR kinetics, are incorporated in a
experiments (symbols) in the whole T-range for NO2 /NOx feed novel dual-layer monolith catalyst model, and validated against
ratios of either 0 or 1. In the case of NO2 /NOx = 0.5 a good match both lab-scale and engine test bench data collected over dual-layer
between experiments and simulations was found above 275 ◦ C, ASC catalysts. The related results will be reported in a forthcoming
while in the low-T range the overestimation of catalyst activity paper.
(Fig. 15A, pink line vs. diamonds) resulted in the somewhat incor-
rect prediction of N2 and NOx yields (Fig. 15B and C, pink line vs. Appendix A. Supplementary data
diamonds).
Predictive simulations of NH3 conversion and products yields Supplementary data associated with this article can be
in the case of the mechanical mixture configuration are shown in found, in the online version, at http://dx.doi.org/10.1016/j.
Fig. 16 for the NH3 /O2 reacting system. In this case model predic- apcatb.2012.10.031.
tions resulted in very good agreement with experimental data over
the whole 150–550 ◦ C T-range, but for minor deviations of the NOx References
yield above 450 ◦ C.
A good simulation quality is also apparent above 350 ◦ C [1] T.V. Johnson, SAE Technical Paper 2011-01-0304 (2011).
[2] A. Scheuer, W. Hauptmann, A. Drochner, J. Gieshoff, H. Vogel, M. Votsmeier,
in Fig. 17, showing the effect of the NO2 /NOx ratio on NH3 Applied Catalysis B: Environmental 111–112 (2012) 445–455.
conversion and products yields for the mechanical mixture. [3] M. Colombo, I. Nova, E. Tronconi, V. Schmeißer, B. Bandl-Konrad, L. Zimmer-
Significant deviations in terms of N2 and N2 O yields were mann, Applied Catalysis B: Environmental 111–112 (2011) 106–118.
[4] M. Iwasaki, K. Yamazaki, H. Shinjoh, Applied Catalysis A: General 366 (2009)
instead observed comparing model predictions and experimen-
84–92.
tal data below 350 ◦ C, specifically in the case of NO2 /NOx = 0 and [5] P. Markatou, J. Dai, A. Johansson, W. Klink, M. Castagnola, T.C. Watling, M.
1. Tutuianu, SAE Technical Paper 2011-01-1304 (2011).
On the whole, the results in Figs. 14–17 emphasize that the [6] L. Olsson, H. Sjövall, R.J. Blint, Applied Catalysis B: Environmental 87 (2009)
200–210.
overall performance of combined SCR + PGM systems results from a [7] A. Schuler, M. Votsmeier, P. Kiwic, J. Gieshoff, W. Hautpmann, A. Drochner, H.
simple superposition of the respective catalytic chemistries, which Vogel, Chemical Engineering Journal 154 (2009) 333–340.
however provides different results depending on the relative spa- [8] H. Sjövall, R.J. Blint, A. Gopinath, L. Olsson, Industrial & Engineering Chemistry
Research 49 (2010) 39–52.
tial distribution of the two catalyst components. [9] D. Chatterjee, T. Burkhardt, M. Weibel, I. Nova, A. Grossale, E. Tronconi, SAE
Technical Paper 2007-01-1136 (2007).
4. Conclusions [10] A. Scheuer, M. Votsmeier, A. Schuler, J. Gieshoff, A. Drochner, H. Vogel, Topics
in Catalysis 52 (2009) 1847–1851.
[11] A. Scheuer, O. Hirsch, R. Hayes, H. Vogel, M. Votsmeier, Catalysis Today 175
As a part in the development and validation of a chemically and (2011) 141–146.
physically consistent mathematical model of a commercial dual- [12] M. Colombo, I. Nova, E. Tronconi, Chemical Engineering Science 75 (2012)
75–83.
layer (SCR + PGM) monolithic NH3 slip converter (ASC), we have [13] C. Ciardelli, I. Nova, E. Tronconi, M. Ascherfeld, W. Fabinski, Topics in Catalysis
reported in this paper on NH3 /O2 /NO NO2 steady-state and tran- 42–43 (2007) 161–164.
sient kinetic runs performed over the PGM component of the NH3 [14] C. Ciardelli, I. Nova, E. Tronconi, B. Konrad, D. Chatterjee, K. Ecke, M. Weibel,
Chemical Engineering Science 59 (2004) 5301–5309.
slip catalyst, herein tested in the form of precursor washcoat pow- [15] P.L. Villa, P. Forzatti, G. Buzzi-Ferraris, G. Garone, I. Pasquon, Industrial & Engi-
ders. From these data an original global PGM kinetic model was neering Chemistry Process Design and Development 24 (1985) 12–19.
developed, which fully accounts for the effects of temperature and [16] G. Buzzi-Ferraris, G. Donati, Chemical Engineering Science 29 (1974)
1504–1509.
of NOx co-feed (with variable NO2 /NOx feed ratio in the 0–1 range)
[17] R. Kraehnert, M. Baerns, Chemical Engineering Journal 137 (2008)
on NH3 oxidation. The model considers both the well known NO2 361–375.
inhibition of NO oxidation, as well as a novel NO2 inhibition effect [18] E.V. Rebrov, M.H.J.M. de Croon, J.C. Schouten, Chemical Engineering Journal 90
on the NH3 oxidation reactions. The fit reveals that NH3 adsorption (2002) 61–76.
[19] Y. Li, J.N. Armor, Applied Catalysis B: Environmental 13 (1997)
becomes the rate limiting step at intermediate-high temperatures. 131–139.
Comparative NH3 /O2 /NO NO2 steady-state runs were per- [20] D. Bhatia, R.W. McCabe, M.P. Harold, V. Balakotaiah, Journal of Catalysis 266
formed also over two different combinations of SCR + PGM (2009) 106–119.
[21] W. Hauptmann, M. Votsmeier, J. Gieshoff, A. Drochner, H. Vogel, Applied Catal-
powders, loaded in the test micro-reactor either as a sequential ysis B: Environmental 93 (2009) 22–29.
double-bed or as a mechanical mixture, respectively. While the NH3 [22] L. Olsson, E. Fridell, Journal of Catalysis 210 (2002) 340–353.
oxidation activity was similar, being controlled by the PGM compo- [23] L. Olsson, H. Persson, E. Fridell, M. Skoglundh, B. Andersson, The Journal of
Physical Chemistry B 105 (2001) 6895–6906.
nent, the N2 selectivity was greater over the mechanical mixture, [24] K. Kamasamudram, N.W. Currier, X. Chen, A. Yezerets, Catalysis Today 151
due to the fact that in this configuration the unselective NH3 oxi- (2010) 212–222.
dation products (NOx ) formed over the PGM catalyst had a chance [25] R.Q. Long, R.T. Yang, Catalysis Letters 78 (2002) 353–357.
to further react selectively with more NH3 over the SCR catalyst.