Reaction pathway of the reduction by CO under dry conditions of NO x species stored onto PtBa/Al2O3 Lean NO x Trap catalysts

Article abstract of DOI:10.1016/j.jcat.2010.06.014

The reduction by CO under dry condition of NO_x species stored at 350 °C onto a PtBa/Al₂O₃ Lean NO_x Trap catalyst is investigated by means of transient response methods (CO-TPSR and CO-ISC experiments) and complementary FTIR spectroscopy. It is shown that the reduction by CO under nearly isothermal conditions of nitrates stored onto PtBa/Al₂O₃ occurs through a Pt-catalyzed pathway which leads mainly to nitrogen and does not proceeds via the thermal decomposition of stored nitrates with release of NO_x in the gas phase. The CO-TPSR and CO-ISC experiments and the combined FTIR study show that (i) the reduction of nitrates by CO occurs via a consecutive reaction scheme with formation of nitrites and then of surface isocyanate/cyanate species followed by the reaction of these species with residual nitrites to give nitrogen; (ii) the reaction of NCO species with nitrites to give nitrogen is slower than the reduction of nitrates to give nitrites and then NCO species; (iii) upon contact with oxygen at increasing temperature, NCO species can be re-oxidized to surface nitrites at first and then to surface nitrates; (iv) the formation of nitrogen is observed during the reduction of stored NO_x by CO, and during the oxidation of surface NCO species upon oxygen addition and upon admission of NO in the presence of excess O₂. In all cases, the formation of nitrogen can be explained by the reaction between surface NCO species and nitrites. The mechanism for the reduction of stored NO_x by CO under dry and near isothermal conditions is discussed.

Full text of DOI:10.1016/j.jcat.2010.06.014

Journal of Catalysis 274 (2010) 163–175
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Reaction pathway of the reduction by CO under dry conditions of NOx species
stored onto PtABa/Al2O3 Lean NOx Trap catalysts
a
b
b
b
Pio Forzatti a, Luca Lietti a, , Isabella Nova , Sara Morandi , Federica Prinetto , Giovanna Ghiotti
*
a Laboratory of Catalysis and Catalytic Processes and NEMAS Centre of Excellence, Dipartimento di Energia, Politecnico di Milano – p.za Leonardo da Vinci, 32 – 20133 Milano, Italy
b Dipartimento di Chimica IFM and NIS Centre of Excellence, Università di Torino, via P. Giuria 7, 10125 Torino, Italy
a r t i c l e i n f o
a b s t r a c t
Article history:
The reduction by CO under dry condition of NOx species stored at 350 °C onto a PtABa/Al2O3 Lean NOx
Trap catalyst is investigated by means of transient response methods (CO-TPSR and CO-ISC experiments)
and complementary FTIR spectroscopy. It is shown that the reduction by CO under nearly isothermal con-
ditions of nitrates stored onto PtABa/Al2O3 occurs through a Pt-catalyzed pathway which leads mainly to
nitrogen and does not proceeds via the thermal decomposition of stored nitrates with release of NOx in
the gas phase. The CO-TPSR and CO-ISC experiments and the combined FTIR study show that (i) the
reduction of nitrates by CO occurs via a consecutive reaction scheme with formation of nitrites and then
of surface isocyanate/cyanate species followed by the reaction of these species with residual nitrites to
give nitrogen; (ii) the reaction of NCO species with nitrites to give nitrogen is slower than the reduction
of nitrates to give nitrites and then NCO species; (iii) upon contact with oxygen at increasing tempera-
ture, NCO species can be re-oxidized to surface nitrites at first and then to surface nitrates; (iv) the for-
mation of nitrogen is observed during the reduction of stored NOx by CO, and during the oxidation of
surface NCO species upon oxygen addition and upon admission of NO in the presence of excess O2. In
all cases, the formation of nitrogen can be explained by the reaction between surface NCO species and
nitrites. The mechanism for the reduction of stored NOx by CO under dry and near isothermal conditions
is discussed.
Received 25 March 2010
Revised 21 June 2010
Accepted 25 June 2010
Available online 11 August 2010
Keywords:
Lean NOx Trap
PtABa/Al2O3
Reduction by CO
Transient response methods
FTIR
Nitrates
Nitrites
Isocyanates
Reaction pathway
Reaction mechanism
Ó 2010 Elsevier Inc. All rights reserved.
1. Introduction
not initiated by the thermal decomposition of nitrates/nitrites
ad-species with release of NOx in the gas phase but involves a Pt-
Lean NOx Traps (LNTs), also referred to as NOx storage-reduction
(NSR) catalysts, represent a viable technology for the reduction of
NOx emissions from light duty diesel engines and lean burn gaso-
line engines [1–3]. Their operation comprises a lean period of
30–90 s during which the NOx emitted in the exhaust gases are ad-
sorbed on the catalyst surface in the form of nitrites and nitrates,
followed by a short 3–5 s rich period during which the stored
NOx are reduced by H2, CO and unburned hydrocarbons to give
N2. The LNT catalysts typically consist of a NOx storage component,
such as Ba or K, and of a noble metal such as Pt. This latter cata-
lyzes the oxidation of NO and the combustion of CO and unburned
hydrocarbons during the lean period, and the reduction of stored
NOx during the rich period; the catalysts components are dispersed
on high surface area support like alumina.
catalyzed pathway which is active at low temperature and leads
to nitrogen [4,5]. It has also been proposed that under near isother-
mal conditions N2 is formed exclusively via a consecutive reaction
scheme which involves the fast reaction of H2 with stored NOx to
give ammonia followed by the slower reaction of ammonia with
residual stored NOx to give N2 [6–8]. On the opposite, Pihl et al.
[9] and Cumaranatunge et al. [10] suggested that upon switch from
lean to rich conditions the stored NOx are at first released to the gas
phase. In the operation of a real system, the NOx release is caused
primarily by the exotherm generated by the reaction between fuel
and excess oxygen in the gas phase or oxygen stored in the catalyst
upon the rich spike. However, the decrease in O2 and NOx gas-
phase concentrations, which destabilizes the stored nitrates, may
also contribute to the NOx release. Then, the regeneration of the
NOx Trap occurs via dissociative adsorption of gaseous NO and of
hydrogen at metal Pt to form PtAN, PtAO and PtAH species fol-
lowed by recombination of surface PtAN species to give nitrogen
and by recombination of surface PtAN and PtAH species to give
ammonia according to a parallel pathway. However, the selective
catalytic reduction of NOx with ammonia and the oxidation of
ammonia to give nitrogen are also considered in the schemes.
Several studies were published on the reduction of stored NOx
using hydrogen as reducing agent [3–17]. In previous works, some
of us have shown that the reduction by H2 under near isothermal
conditions of nitrates stored onto PtABa/Al2O3 and PtAK/Al2O3 is
* Corresponding author. Fax: +39 02 706 38173.
E-mail address: lica.lietti@polimi.it (L. Lietti).
0021-9517/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2010.06.014

164
P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
Several authors have also postulated a spillover mechanism
137 m2/g and 0.81 cm3/g for the PtABa/Al2O3 sample. The Pt dis-
persion, as determined by H2 chemisorption at 0 °C, was ꢀ70% [24].
involving oxygen from Pt to Ba during the NOx storage and a re-
verse spillover involving nitrites/nitrates from Ba to Pt during the
reduction [11–13]. This explains why the NOx storage and reduc-
tion is improved when the Pt is in close proximity to the Ba
[11,14–16] and also accounts for the participation to NSR of a large
percentage of Ba [17].
The reduction by CO of NOx stored on LNT catalysts has at-
tracted lower interest. It has been reported that stable isocya-
nates/cyanates are formed during the reduction of stored NOx by
CO; these species are hydrolyzed into ammonia which then react
with NO or stored NOx in the presence of oxygen to give nitrogen
[18–20]. The reactivity of isocyanates/cyanates with O2 to give
N2 and CO2 was observed over PtARh/Ba/Al2O3 and Ag/Al2O3 cata-
lysts by Daturi and coworkers using FTIR spectroscopy [20,21] and
2.2. Reactivity tests
The reactivity tests were performed over the PtABa/Al2O3 and
the Ba/Al2O3 samples previously conditioned. In the case of the
PtABa/Al2O3 sample, conditioning involves a few adsorption/
reduction cycles with NO/O2 (1000 ppm NO and 3% v/v O2 in He)
and H2 (2000 ppm in He) at 350 °C, respectively, with an inert
purge (He) between the two phases. Conditioning was done until
a reproducible behavior was obtained and this typically required
2 or 3 cycles [4,7]. For the Ba/Al2O3 binary catalyst, which does
not adsorb significant amounts of NO [25] and do not react with
H2, the adsorption was carried out with NO2 (1000 ppm in He),
and the adsorbed NOx were removed by heating at 500 °C. Like
PtABa/Al2O3, also in the case of Ba/Al2O3 large amounts of NOx
are stored on the catalyst surface in the form of nitrate species [25].
Temperature-programmed surface reaction experiments (TPSR)
were carried out in the presence of CO as reducing agent to analyze
reactivity of adsorbed nitrates. Before performing TPSR runs, the
catalysts were saturated with NOx at 350 °C by imposing a rectan-
gular step feed of NO (0 ? 1000 ppm ? 0) in flowing He + 3% v/v
O2. For the Ba/Al2O3 catalyst, the adsorption was carried out with
NO2 (1000 ppm in He) instead of NO + O2. The NO + O2 (or NO2)
flow was maintained up to catalyst saturation and then the inlet
NO + O2 (or NO2) concentration was decreased to zero. The samples
were then kept under inert He flow for about 1 h at 350 °C to pro-
voke the desorption of weakly adsorbed species (He purge). The
samples were then cooled down to RT under flowing He [4,7]
and then heated at 15 °C/min up to 400 °C in He + 2000 ppm CO
(CO-TPSR).
The storage/reduction of NOx over Ba/Al2O3 and PtABa/Al2O3
was also performed at constant temperature by imposing stepwise
changes in the inlet concentration of the reagents (Isothermal Step
Concentration experiments, ISC), according to the following se-
quence. After catalyst conditioning, the storage of NOx (lean phase)
was accomplished at 350 °C for about 150 s using the same condi-
tions reported for CO-TPRS experiments. After storage, the NO + O2
(or NO2) concentration was decreased to zero, while maintaining
the same temperature (He purge). The reduction of the stored
NOx was then carried out by imposing a stepwise change in the
CO inlet concentration (0 ? 2000 ppm ? 0) in flowing He (Rich
phase). Three or four cycles have been carried out until a steady
cyclic catalytic behavior was obtained.
by Scholz et al. over a PtABa/c-Al2O3 catalyst performing lean-rich
cycling experiments [22]. The hydrolysis of surface NCO species to
give ammonia and carbon dioxide has been also reported by Daturi
and coworkers [20]. Furthermore, FTIR analysis performed at low
temperatures (i.e. 150 °C) over Pd/Al2O3 catalysts showed that iso-
cyanate/cyanate species react upon adding NO/O2 mixtures to the
feed flow. The mechanism of the reduction of stored NOx by H2 and
by CO over LNT catalysts under dry conditions has been discussed
by Szailer et al. [19]. The proposed mechanism assumes as a first
step the thermal decomposition of stored NOx with release of
NOx in the gas phase followed by dissociative re-adsorption onto
the Pt sites with rearrangement, formation of adsorbed NCO spe-
cies which eventually react with NOx to give nitrogen.
In this paper, a systematic and quantitative study of the reduc-
tion by CO under dry conditions of NOx stored onto a PtABa/Al2O3
LNT catalyst is carried out to clarify the reaction pathway and the
mechanism of the reaction. Transient reactivity methods and com-
plementary FTIR spectroscopy are employed to obtain information
on the nature, reactivity and evolution of reagents, intermediates
and products both in the gaseous phase and at the catalyst surface.
In order to prevent the thermal decomposition of nitrates with re-
lease of NOx in the gas phase during the lean/rich switches that
complicates the description of the reacting system, experiments
have been performed under nearly isothermal conditions, i.e. in
the absence of significant temperature effects during cycling. This
has been accomplished using low CO concentration and by sepa-
rating the lean and the rich phases by an inert purge in between.
Accordingly, a more precise rationalization of the catalytic behav-
ior and a better understanding of the chemical processes occurring
during the lean/rich cycling could be obtained.
The experiments were carried out in a flow micro-reactor sys-
tem consisting of a quartz tube reactor (7 mm internal diameter)
directly connected to a mass spectrometer (Balzers QMS 200). In
order to minimize intra- and inter-particle diffusion limitations,
2. Experimental
2.1. Materials
60 mg of catalyst with small particle diameter (100–120 lm) and
a total flow rate of 100 cc/min (at 1 atm and 0 °C) were used in
each run. The flow rates of the gases were measured and controlled
by mass-flow controllers (Brooks 5850 TR), and the gases were
mixed before entering the reactor. The reactor was inserted into
an electric furnace driven by a PID temperature controller/pro-
grammer (Eurotherm 2408). A K-type thermocouple (outer diame-
ter 0.5 mm) directly immersed in the catalyst bed was used to
measure and control the catalyst temperature. The following
mass-to-charge (m/e) ratios were used to monitor the concentra-
tion of products and reactants: 2 (H2), 14 (N2), 17 (NH3), 18
(H2O), 28 (CO), 30 (NO), 32 (O2), 44 (N2O or CO2), and 46 (NO2).
The mass-spectrometer data were quantitatively analyzed using
the fragmentation patterns and the response factors determined
experimentally from calibration gases. Relevant interferences in
the mass-to-charge signals were taken into account in determining
Ba/Al2O3. and PtABa/Al2O3 catalysts were prepared by the incip-
ient wetness impregnation method, using aqueous solutions of
Ba(CH3COO)2 (Strem Chemical, 99%) and Pt(NH3)2(NO2)2 (Strem
Chemicals, 5% Pt in ammonium hydroxide) to impregnate the c-
alumina support calcined at 700 °C (Versal 250 from La Roche
Chemicals, surface area of 200 m2/g and pore volume of 1.2 cm3/g).
In the case of the PtABa/Al2O3 catalyst, the impregnation was car-
ried out in sequential manner: the alumina support was first
impregnated with the Pt-containing solution and then with the
Ba acetate solution. After each impregnation step, the catalysts
were dried at 80 °C overnight and then calcined at 500 °C for 5 h.
The final loading was 20/100 w/w for the Ba/Al2O3 catalyst and
1/20/100 w/w for the PtABa/Al2O3 catalyst. The following surface
areas and pore volumes were measured by N2 adsorption–desorp-
tion at 77 K: 133 m2/g and 0.82 cm3/g for the Ba/Al2O3 sample;
the products composition.
A gas chromatograph (HP 6890)

P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
165
equipped with a Poraplot Q and a 5 Å molecular sieve capillary col-
umn was also used for the analysis of CO2, N2O, and H2O, and of O2,
N2, and CO, respectively.
and accordingly the presence of CO in the feed does not affect sig-
nificantly the temperature threshold of nitrate thermal
decomposition.
Further details about the experimental apparatus and proce-
dures can be found elsewhere [4,6,7].
3.1.2. CO-ISC experiment and FTIR study
The reduction by CO of nitrates stored onto Ba/Al2O3 was also
studied by transient experiment under isothermal conditions
(CO-ISC experiment) at 350 °C. After adsorption of NO2 at the same
temperature, 2000 ppm of CO was admitted to the catalyst. After
CO admission, the outlet concentration of CO approaches quickly
the feed level, indicating a very limited reaction between adsorbed
nitrates and CO and thus indicating that stored NOx are poorly
reactive in the presence of CO at the adsorption temperature of
350 °C. Also very small amounts of CO2 are formed that confirmed
the poor activity of CO toward nitrates, along with very few ppm of
NOx.
2.3. FTIR study
Absorption/transmission IR spectra were run on a Perkin-Elmer
FTIR System 2000 spectrophotometer equipped with a HgACdATe
cryo-detector, working in the range of wavenumbers 7200–
580 cmꢁ1 at a resolution of 1 cmꢁ1 (number of scans ꢀ10). For IR
analysis, powder samples were compressed in self-supporting
discs (10 mg cmꢁ2) and placed in a commercial heated stainless
steel cell (Aabspec), allowing thermal treatments in situ under vac-
uum or controlled atmosphere and the simultaneous registration
of spectra at temperatures up to 600 °C.
Before the NOx storage, the samples were fully conditioned, i.e.
submitted to: (i) outgassing at 550 °C for 30 min, (ii) four or five
conditioning cycles consisting of NO2 adsorption at 350 °C, fol-
lowed by outgassing at 550 °C, necessary to transform the bulky
BaCO3 present on calcined powder into a BaO dispersed phase
[26], (iii) oxidation at 550 °C for 30 min, cooling in oxygen down
to the temperature requested.
NOx storage was carried out at 350 °C by admitting freshly pre-
pared NO/O2 mixtures (pNO = 5 mbar, pO2 = 20 mbar) on PtABa/
Al2O3 discs, or NO2 (pNO2 = 5 mbar) on BaAAl2O3 discs, up to cata-
lysts saturation (ca. 20 min), followed by outgassing at the same
temperature for few seconds. The ratio O2/NO used during storage
is rather small compared to real operation. The spectra of the
stored NOx were collected at 350 °C or after cooling down the discs
at the chosen temperature. The reduction was accomplished with
CO (PCO = 10 mbar) and was performed at: (i) 350 °C, at increasing
exposure times or (ii) increasing temperature in the range 150–
350 °C.
The interaction of CO at 350 °C with the NOx species stored on
Ba/Al2O3 catalyst was also followed by FTIR spectroscopy and the
results are presented in Fig. 2: (i) curve a is the spectrum of NOx
species previously stored at 350 °C, i.e. mainly ionic nitrates
(related bands at 1410, 1320 and 1030 cmꢁ1
,
the splitted
(NO3)sym modes, respectively) and minor
amounts of bidentate nitrates (
(N@O) mode at 1550 cmꢁ1, the
m(NO3)asym and the m
m
only one visible) [25,26]; (ii) curves b–e correspond to the spectra
recorded during the interaction with CO at increasing exposure
times at 350 °C. Carbonate species (the related bands expected at
1560, 1350 and 1060 cmꢁ1, corresponding to
m
(C@O),
m(OCO)asym
and
m(OCO)sym, respectively [24,27]) along with small amounts of
isocyanate and/or cyanate species (weak bands at 2164 and
2222 cmꢁ1) [28] are detected. Owing to the strong overlapping of
the bands related to nitrates and carbonates, it is not easy to eval-
uate the extent of nitrates reduction. However, even after 1 h of
exposure (curve e), the reduction of nitrates is not complete; in-
deed, the broad and asymmetric band at 1345 cmꢁ1 is due to the
superposition of the two
m(NO3)asym modes of the ionic nitrates
and the (OCO)asym mode of the carbonates.
m
NO (Praxair, purity P99.0%) was freshly distilled before use.
Conversely, NO2 (Praxair, purity P99.5%), O2 (Praxair, purity
P99.999%) and CO (Praxair, purity P99.9%) were directly used.
In line with CO-TPSR and CO-ISC data, FTIR data show that the
reduction by CO of nitrates stored onto the Ba/Al2O3 is very slow at
350 °C and occurs to a rather limited extent. The reaction leads to
the formation of mainly carbonate species and of minor amounts of
isocyanate and/or cyanate species.
3. Results and discussion
It is worth noticing that isocyanates/cyanates present after 1 h
of exposure, as well as residual nitrates and carbonates, are stable
after 10 min of evacuation at 350 °C (not shown).
3.1. Reduction of NOx stored onto Ba/Al2O3
3.1.1. CO-TPSR experiments
The reactivity toward CO of NOx species stored at 350 °C over
the Ba/Al2O3 sample was investigated at first by performing a
CO-TPSR experiment. The results are presented in Fig. 1 and show
that CO is consumed to a very limited extent above ꢀ330 °C
(minimum concentration near 1800 ppm) and that CO2 and NO
are simultaneously formed along with trace amounts of N2 and
NO2. The concentrations of the products increase up to 400 °C
and then slowly decrease to 0 ppm while the system is hold at this
temperature (400 °C); in the meantime, the concentration of CO
slowly recovers the inlet value of 2000 ppm.
The moles of CO consumed and of the different products formed
during CO-TPSR over Ba/Al2O3 were calculated from Fig. 1 and are
reported in Table 1 together with those obtained during CO-TPSR
over PtABa/Al2O3 that will be discussed later on. It appears that
only a fraction of the NOx originally stored as nitrates onto the cat-
alyst is removed during the CO-TPSR experiment (2.91 ꢂ 10ꢁ4 gmo-
les/gcat of NOx released as N2 + NO + NO2 vs. 13.1 ꢂ 10ꢁ4 gmoles/gcat
stored NOx, which corresponds to ꢀ22% of the initial NOx loading).
It is worth noticing that CO is able to react at ca. 330 °C with
NOx stored at 350 °C. This temperature value is near to the onset
temperature for nitrates thermal decomposition (ꢀ360 °C) [25]
3.2. Reduction of NOx stored onto PtABa/Al2O3
3.2.1. CO-TPSR experiment and FTIR study
The reactivity toward CO of NOx species previously stored at
350 °C over the PtABa/Al2O3 sample was investigated by perform-
ing a CO-TPSR experiment and the results are presented in Fig. 3.
CO consumption is seen starting from ꢀ180 °C and is complete in
the temperature range 270–320 °C. The formation of CO2 is de-
tected simultaneously with the consumption of CO, showing a
maximum of ꢀ1500 ppm in the temperature range where CO con-
sumption is complete. Very small amounts of NO are observed at
low temperatures but well above 180 °C, namely at 215–250 °C.
At 250 °C, the concentration of NO diminishes to zero. At the same
time, nitrogen is formed and its concentration shows a sharp in-
crease from 200 to 250 ppm at 250 °C, it keeps at this level up to
290 °C and then slowly decreases. Above ꢀ320 °C, the CO concen-
tration increases and tends to recover the inlet value of 2000 ppm
while non negligible quantities of ammonia are formed (with an
upper concentration of 50–60 ppm). From ꢀ350 °C, when the
nitrogen concentration decreases to zero, hydrogen is detected in
the gaseous products together with a shoulder in the CO2 trace

166
P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
2200
2000
CO
1800
300
200
100
0
CO2
Tads
NO
N2
NO2
100
200
300
400
500
600
700
Temperature (°C)
Fig. 1. CO-TPSR (2000 ppm CO in He) after NO2 adsorption at 350 °C over the Ba/Al2O3 catalyst.
centered at T = 400 °C. The presence of ammonia and H2 among the
reaction products is likely associated with the presence of trace
amounts of water in the feed and/or of surface hydroxyls (vide
infra).
Table 1
Moles/gcat of CO consumed and of products formed during CO-TPSR over Ba/Al2O3 and
PtABa/Al2O3.
Catalyst/reagent – product
species
Temperature range
It is worth noticing that over the PtABa/Al2O3 sample, the con-
sumption of CO occurs at temperatures well below that of NOx
thermal desorption over the same catalyst. Indeed, the He-TPD
experiment performed after NOx adsorption at 350 °C, already pub-
lished elsewhere [4] and shown in the insert of Fig. 3 for compar-
ison purpose, indicates that nitrates are stable up to 350 °C, i.e. up
to the NOx adsorption temperature. Accordingly, the CO-TPSR data
clearly prove that CO is able to reduce the nitrate species stored
onto PtABa/Al2O3 catalyst at temperature well below that of their
thermal decomposition (ꢀ180 °C vs. 350 °C).
RT-250 °C
RT-400 °C with hold at
400 °C
Ba/Al2O3; NOx stored: 13.1 ꢂ 10ꢁ4
CO consumed
CO2 formed
N2 formed
NO formed
–
–
–
–
–
4.62 ꢂ 10ꢁ4
3.59 ꢂ 10ꢁ4
0.464 ꢂ 10ꢁ4
1.82 ꢂ 10ꢁ4
0.159 ꢂ 10ꢁ4
NO2 formed
PtABa/Al2O3; NOx stored: 5.76 ꢂ 10ꢁ4
CO consumed
CO2 formed
N2 formed
1.43 ꢂ 10ꢁ4
20.7 ꢂ 10ꢁ4
14.8 ꢂ 10ꢁ4
0.935 ꢂ 10ꢁ4
0.427 ꢂ 10ꢁ4
The moles of CO consumed and of the different products formed
during CO-TPSR up to 400 °C with hold at this temperature for
30 min over PtABa/Al2O3 were calculated from Fig. 3 and are
reported in Table 1. In total, 2.46 ꢂ 10ꢁ4 gmoles/gcat of NOx are
released in the gas phase as N2 + NO + NH3 while 5.76 ꢂ
10ꢁ4 gmoles/gcat of NOx species have been stored previously onto
the catalyst. The much higher amount of NOx stored over Ba/
Al2O3 starting from NO2 (13.1 ꢂ 10ꢁ4 gmoles/gcat) with respect to
that measured over the PtABa/Al2O3 from NO/O2 mixture
(5.76 ꢂ 10ꢁ4 gmoles/gcat) can be explained by considering that the
higher adsorption capability of NO2 if compared to NO/O2. In any
case, over the PtABa/Al2O3 sample only ꢀ43% of the stored nitrates
are reduced to gaseous N-containing products and consequently N-
containing species are still present at the catalyst surface at the end
of the CO-TPSR experiment. Besides, the C mass balance between
COx at the inlet and outlet does not close since 2.07 ꢂ 10ꢁ3 moles/
gcat of CO are consumed and only 1.48 ꢂ 10ꢁ3 moles/gcat of CO2
are produced. This indicates that C-containing species are also
stored onto the catalyst surface at the end of the CO-TPSR
experiment.
–
NO formed
NH3 formed
H2 formed
0.0301 ꢂ 10ꢁ4 0.086 ꢂ 10ꢁ4
–
–
0.507 ꢂ 10ꢁ4
0.475 ꢂ 10ꢁ4
1.2
1.0
0.8
0.6
0.4
0.2
0.0
e
a
Worth of note, the C balance is not satisfied also in the low-tem-
perature region, below 250 °C (1.43 ꢂ 10ꢁ4 moles/gcat of CO con-
sumed vs. 0.427 ꢂ 10ꢁ4 moles/gcat of CO2 produced), where only
a very small quantity of N-containing species is released in the
gas phase (0.031 ꢂ 10ꢁ4 moles/gcat of NO). This points out that in
this temperature range, the stored nitrates are reduced with for-
mation of C-containing species which remain adsorbed on the sur-
face. This is indeed confirmed by the results of FTIR measurements,
2200 1800
1600
1400
1200
1000
Wavenumbers (cm-1)
Fig. 2. FTIR spectra of NOx reduction at 350 °C in CO (5 mbar) at increasing
exposure time over the Ba/Al2O3 catalyst. Curve a, spectrum of NOx stored at 350 °C
by NO2 adsorption; curves b–e, spectra after reduction for 2 min, 5 min, 10 min and
1 h, respectively.

P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
167
2000
1500
Tads
CO
CO2
TPD
400
200
0
800
600
400
200
0
NO
T ads
O2
(x 10)
NO
2
100 200 300 400 500 600
N2
NH3 H2
NO
100
200
300
400
500
Temperature (°C)
Fig. 3. CO-TPSR (2000 ppm CO in He) after NO/O2 adsorption at 350 °C over the PtABa/Al2O3 catalyst. The results of the TPD run in He after NO/O2 adsorption at 350 °C over
the same catalyst is shown in the insert.
where the reduction by CO of NOx species previously stored at
1.2
1.0
0.8
0.6
0.4
0.2
0.0
350 °C onto the PtABa/Al2O3 catalyst has been investigated
(Fig. 4A). In this figure, curve a is the spectrum of NOx species pre-
viously stored at 350 °C and cooled down to 150 °C. The stored ni-
trates are not perturbed by interaction with CO at 150 °C (not
shown); curves b–e correspond to the spectra recorded during
interaction with CO at increasing temperature in the 200–350 °C
range after 10 min of exposure at each temperature. The bands of
the stored nitrates slightly decrease at 200 °C (spectrum b) and al-
most completely disappear in the range 300–350 °C (curves d and
e). Starting from 200 °C, very small amounts of NCO species are
formed, as revealed by the two related bands at 2230 and
2170 cmꢁ1, whose intensity increases markedly with temperature,
A
2222
a
b
2164
d=e
c
d
c
e
b
a
while the position shifts to lower frequency (2222 and 2164 cmꢁ1
,
respectively). Simultaneously, surface carbonates are formed,
whose related bands (1555, 1350 an 1060 cmꢁ1) are well evident
only starting from 300 °C (spectra d and e), when the intensity of
the bands related to nitrates is strongly decreased. From this tem-
perature, a shoulder at about 1250 cmꢁ1 indicates the presence of
small amounts of nitrites. Actually, the partial superposition of the
bands related to nitrates, nitrites and carbonates makes difficult to
track the contemporaneous NOx removal and carbonates forma-
tion. In any case, the results point out that nitrates are reduced
by CO to nitrites and to NCO species, while CO2 produced during
the reduction is adsorbed at the catalyst surface forming carbonate
species. The two bands at 2222 and 2164 cmꢁ1 have been assigned
in the literature to isocyanate species coordinated to Al3+ in tetra-
hedral coordination and on the barium phase, respectively
[19,21,29]. However, our results obtained from experiments per-
formed for assignment purposes do not match with literature.
We found that both the bands are related to isocyanate/cyanate
species on the Ba-phase surface [30,31]. We have also observed
that carbonates and isocyanates/cyanates present after reduction
at 350 °C are stable under 10 min of evacuation at 350 °C (not
shown).
2250 2100
1600
1400
1200
1000
Wavenumbers (cm-1)
0.14
0.12
0.10
0.08
0.06
0.04
0.02
0.00
B
2075
2055
e
2078
2084
d
c
b
a
2115
2130
2100
2070
2040
2010
Wavenumbers (cm-1)
Fig. 4. FTIR spectra of NOx reduction in CO (5 mbar) at increasing temperature over
the PtABa(16)/Al2O3 catalyst. Curve a, spectrum of NOx stored at 350 °C by NO/O2
adsorption; curves b–e, spectra after reduction at 200, 250, 300 and 350 °C,
respectively. Each spectrum was run after 10 min of exposure to CO. Section B is an
expansion of the spectra in the 2130–2010 cmꢁ1 region.
In line with TPSR data, FTIR experiments confirm that nitrates
stored onto PtABa/Al2O3 catalyst are reduced by CO at tempera-
tures much lower than that of their thermal decomposition
(200 °C vs. 350 °C).
It is also worth noticing that, after CO interaction at 200 °C, FTIR
spectrum showed two weak bands at 2110 cmꢁ1 and at 2084 cmꢁ1
related to the stretching modes of linear carbonyls adsorbed on Pt+
and Ptd+ sites, indicating that the platinum is only partially reduced
at this temperature. At higher temperatures only a band shifting

168
P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
2000
COin
COin
2000
1500
1000
500
0
150
100
50
N2
CO
CO2
NH3
NO
0
0
250 500 750 1000 1250 1500
Time (s)
N2
NH3
NO
0
250
500
750
1000 1250 1500 1750
Time (s)
Fig. 5. ISC experiments with CO (2000 ppm in He) at 350 °C over the PtABa/Al2O3 catalyst after NO/O2 adsorption at the same temperature.
from 2078 to 2055 cmꢁ1 is present (see Fig. 4B). Actually, the vibra-
tion frequency is strictly related to the extent of the reduction of
the Pt phase and to the CO coverage of the Pt sites [32,33]. In par-
ticular, the frequency shifts to lower value upon increasing the
reduction extent and, on the reduced Pt particles, upon decreasing
the coverage, i.e. increasing the interaction temperature.
lean phase equals the amounts of N-containing species evolved
in the entire lean-rich cycle. This occurs in correspondence to the
third or fourth lean/rich cycle, as will be discussed later on.
Fig. 5 shows the results of a CO-ISC experiment over the PtABa/
Al2O3 catalyst by admitting in a stepwise manner 2000 ppm of CO
at 350 °C after NOx adsorption at the same temperature. Upon
admission of CO, ꢀ150 ppm of N2 and ꢀ1550 ppm of CO2 are
immediately produced, together with negligible amounts of NO
(see insert). CO is completely consumed for about 200 s; afterward
CO is detected at the reactor outlet and its concentration increases
and reaches the inlet value within less than 1000 s. After CO break-
through, the N2 concentration decreases, and the evolution of very
small amounts of ammonia is observed, while the CO2 concentra-
tion decreases. The presence of small amounts of ammonia among
the reaction products is likely associated with the presence of trace
amounts of water in the feed and/or of surface hydroxyls (vide in-
fra). It is worth of note that a negligible increase in the catalyst
temperature of 62–3 °C is measured upon CO admission, so that
the reduction is performed under nearly isothermal conditions.
When the CO feed is stopped (at 1250 s), a small amount of nitro-
gen is produced, along with CO2 (not shown in the insert of Fig. 5).
The moles of CO consumed and of the different products formed
during CO-ISC experiment over PtABa/Al2O3 were calculated from
Fig. 5 and are listed in Table 2. It appears that also during the
CO-ISC experiment the C and N balances in the gas phase are not
satisfied: in fact, upon CO admission, 7.74 ꢂ 10ꢁ4 moles/gcat of
CO are consumed while 6.15 ꢂ 10ꢁ4 moles/gcat of CO2 are formed.
Also, a total amount of 0.88 ꢂ 10ꢁ4 gmoles/gcat of NOx are released
in the gas phase as N2 + NO + NH3 while 1.53 ꢂ 10ꢁ4 gmoles/gcat of
NOx species were stored onto the catalyst before reduction (see
Table 3): accordingly both N- and C-containing species are present
3.2.2. CO-ISC experiment and FTIR study
The reactivity toward CO of NOx stored at 350 °C over the
PtABa/Al2O3 sample was also investigated by performing isother-
mal reactivity experiments at the same temperature.
The results of a CO-ISC experiment run at 350 °C reported and
discussed in the following are collected over a sample that showed
stable cyclic behavior, i.e. after performing several lean-rich cycles
(i.e. NOx storage and reduction by CO with an inert purge in be-
tween) at 350 °C until the amounts of NO consumed during the
Table 2
Moles/gcat of CO consumed and of different products formed during CO-ISC carried
out at 350 °C over PtABa/Al2O3: 2000 ppm stepwise CO admission (t = 0–1250 s);
2000 ppm stepwise CO shut off (t = 1250–1750 s).
Reagent-product species
CO admission
CO shut off
(t = 0–1250 s)
(t = 1250–1750 s)
NOx stored: 1.53 ꢂ 10ꢁ4
CO consumed
CO2 formed
N2 formed
NO formed
7.74 ꢂ 10ꢁ4
6.15 ꢂ 10ꢁ4
3.96 ꢂ 10ꢁ5
1.87 ꢂ 10ꢁ6
7.43 ꢂ 10ꢁ6
–
*
2.5 ꢂ 10ꢁ6
8.18 ꢂ 10ꢁ6
–
–
NH3 formed
*
The contribution of the CO2 tailing in Fig. 5 has been subtracted.
Table 3
Moles/gcat of NO consumed and different species released in the NOx storage phase during a lean-rich cycle after the reduction phase
presented in Fig. 5 over a fully conditioned catalyst: 3% O2 stepwise addition (t = 0–500 s); 1000 ppm NO stepwise admission
(t = 500–650 s); 1000 ppm NO stepwise removal (t = 650–900 s).
Reagent-product
species
3% O2 stepwise addition
(t = 0–500 s)
1000 ppm NO stepwise admission
(t = 500–650 s)
1000 ppm NO stepwise removal
(t = 650–900 s)
NO stored
–
1.53 ꢂ 10ꢁ4
5.52 ꢂ 10ꢁ6
1.54 ꢂ 10ꢁ5
4.94 ꢂ 10ꢁ6
2.7 ꢂ 10ꢁ5
ꢁ2 ꢂ 10ꢁ6
–
N2 released
NO evolved
NO2 evolved
CO2 released
1.7 ꢂ 10ꢁ5
1.35 ꢂ 10ꢁ6
2.06 ꢂ 10ꢁ6
3.39 ꢂ 10ꢁ5
1.5 ꢂ 10ꢁ6
3.3 ꢂ 10ꢁ7

P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
169
at the catalyst surface at the end of the CO-ISC experiment (i.e.
after 1250 s from CO admission). It is noted that in the CO-ISC
experiment, the amount of stored NOx is much lower than that
measured over the same catalyst up to saturation during the CO-
TPSR experiment (1.53 ꢂ 10ꢁ4 vs. 5.76 ꢂ 10ꢁ4 gmoles/gcat) due to
the shorter adsorption time adopted during the CO-ISC experiment
(150 s vs. 1000 s).
Actually, the band at 1555 cmꢁ1 due to bidentate nitrates, if pres-
ent, is always very weak.
On one hand, we observe that (i) for PtABa/Al2O3, the ratio be-
tween the amounts of nitrates originally stored onto the catalyst
and the amounts of isocyanates/cyanates formed after 2 min of
exposure to CO, INO3/INCO, is 1.39; (ii) for Ba/Al2O3, the INO3/INCO ra-
tio estimated after the same exposure time to CO is 100 and 22
after 1 h of exposure to CO, i.e. still very high. On the other hand,
(i) for PtABa/Al2O3, the ratio between the amounts of nitrates orig-
inally stored onto the catalyst and the amounts of carbonate spe-
cies formed after 2 min of exposure to CO, INO3/ICO3, is 6.3; (ii) for
Ba/Al2O3 the INO3/ICO3 ratio is 9 after 2 min of exposure to CO
(and 2.5 after 1 h of exposure to CO). It is worth to note that for
2 min of exposure to CO, the surface nitrates are completely disap-
peared for the PtABa/Al2O3 catalyst, while they prevail on the car-
bonates on the Ba/Al2O3 catalyst and are still present after 1 h of
exposure. The previous observations show that the reduction of
stored NOx is very effective in the presence of Pt leading to both
isocyanates/cyanates and carbonates, the ratio of their amounts
(INCO/ICO3) being 4.5 at the end of the reduction process after ca.
2 min of exposure. In the absence of Pt, the reduction of the stored
nitrates is still incomplete after 1 h of exposure to CO at 350 °C, and
again, both isocyanates/cyanates and carbonates are formed. How-
ever, the ratio of their amounts (INCO/ICO3) is now very low, i.e. 0.12.
This indicates that the reduction leading to isocyanates/cyanates is
catalyzed by Pt.
Finally, upon CO shut off, 8.18 ꢂ 10ꢁ6 moles/gcat of N2 are re-
leased together with 7.7 ꢂ 10ꢁ6 moles/gcat of CO2 (vide infra).
The interaction of CO at 350 °C with the NOx species stored on
PtABa/Al2O3 catalyst was also followed by FTIR spectroscopy, and
the results are presented in Fig. 6A: curve a is the spectrum of
NOx species previously stored at 350 °C; curves b–d correspond
to the spectra recorded during interaction with CO at increasing
exposure times at 350 °C.
The reduction of nitrates by CO is very fast since the intensity of
the nitrate bands at 1340, 1420 and 1550 cmꢁ1 is reduced mark-
edly after 30 s of CO contact (curve b) and even more after 1 min
of CO contact (curve c). Simultaneously, the carbonate species ap-
pear, due to the adsorption of CO2 produced during the reduction,
along with NCO species. After 2 min, nitrate species have almost
disappeared (spectrum d) as mostly carbonate bands are evident.
A careful inspection of Fig. 6A also reveals the presence of nitrite
species, as it can be appreciated from a weak shoulder at 1250
cmꢁ1 in curves c and d. This confirms that also in this case nitrates
are reduced by CO to give nitrites and NCO species. Furthermore,
FTIR spectra also show, after CO interaction, the presence of a weak
band at 2070 cmꢁ1 related to the presence of linear carbonyls onto
Pt sites.
3.3. Reaction network in the reduction by CO of nitrates stored onto
PtABa/Al2O3
The prolonged outgassing at 350 °C (see Fig. 6B) indicates that
NCO species are very stable at this temperature; notably, a fraction
of these species is still present after 2 h 30 min of outgassing (curve
c).
The results collected during CO-TPSR over Ba/Al2O3 (Fig. 1) and
over PtABa/Al2O3 (Fig. 3) highlight that in the presence of Pt the
reduction of nitrate species by CO is faster considering that the
temperature of onset of the reaction is much lower (ꢀ180 °C vs.
330 °C), and the conversion of CO is much higher and in fact is lim-
ited by the concentration of the reducing agent. Besides, it does not
require as first step the thermal release of NOx in the gas phase
since the reaction is seen at temperatures well below that of the
thermal decomposition of adsorbed NOx (350 °C). The results col-
lected during CO-ISC experiment over Ba/Al2O3 and over PtABa/
Al2O3 (Fig. 5) are in line with the results of CO-TPSR experiments
and further indicate that the reduction at 350 °C by CO of NOx
stored onto the PtABa/Al2O3 catalyst is very fast and controlled
by the concentration of CO, while it is almost negligible over Ba/
Al2O3. It is possible to conclude that the reduction by CO under
It is now of interest to make a comparison between the results
described in Fig. 6A with those obtained by interaction of CO at
350 °C with the NOx species stored on Ba/Al2O3 catalyst, shown
in Fig. 2. For this reason, we can take as a measure of the amounts
of the nitrate species stored the integrated intensity (INO3) of the
bands at 1555, 1420, 1310 cmꢁ1 and, for the isocyanate/cyanate
species, the integrated intensity (INCO) of the bands at 2222 and
2164 cmꢁ1. The same should be made for the carbonate species,
the related bands being at 1560, 1350 and 1060 cmꢁ1, but owing
to the superposition of these bands to those of nitrates, if still pres-
ent, we will take only the integrated intensity of the band at
1560 cmꢁ1, ICO3, as measure of the amounts of the carbonates.
1.5
d
B
A
a
c
1.0
b
a
b
0.5
0.0
b
c
1250
1200
c
d
a
2300 2200 2100 2000
2250 2100
1600
1400
-1
Wavenumbers (cm )
Fig. 6. (A) FTIR spectra of NOx reduction in CO (5 mbar) at 350 °C at increasing exposure time over the PtABa/Al2O3 catalyst. Curve a, spectrum of NOx stored at 350 °C by NO/
O2 adsorption; curves b–d, spectra after reduction for 30 s, 1 min and 2 min, respectively. (B) FTIR investigation on the stability of NCO species under evacuation at 350 °C.
Curve a, the same spectrum reported as curve d in section A; curves b and c, spectra after evacuation for 30 min and 2 h 30 min, respectively.

170
P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
The formation of N2 (and of CO2) is also apparent in CO-ISC exper-
iment of Fig. 5 upon CO shut off and cannot be explained by reaction
(3) due to the absence of CO. In this case, a reaction between NCO
adsorbed species with residual amounts of nitrates is suggested, fol-
lowing the stoichiometry of the overall reaction (5):
nearly isothermal conditions of nitrates stored onto PtABa/Al2O3
catalyst at high temperature (i.e. 350 °C) occurs according to a
Pt-catalyzed pathway. Such a pathway does not involve as first
step the thermal decomposition of stored NOx species with release
of NOx in the gas phase and is effective already at low temperature.
As it is apparent from Figs. 3 and 5, this catalytic route of nitrates
reduction by CO leads mainly to nitrogen and CO2 in the gas phase.
It is worth noticing that a low-temperature Pt-catalyzed path-
way has already been demonstrated for the reduction by H2 of
NOx stored over the same catalyst used in this study [4], leading
mostly to nitrogen; hydrogen is more effective than CO since the
onset temperature of the reduction is lower (140 °C vs. 180 °C).
This suggests that the reduction of stored NOx with H2 and with
CO might present analogies that are worth to be investigated, the
onset temperature of the process depending from the reactivity
of the reducing agent toward the stored NOx species.
5BaðNCOÞ2 þ 3BaðNO3Þ2 ! 8N2 þ 8BaCO3 þ 2CO2
ð5Þ
It is noted that the amount of CO2 formed in correspondence with the
CO shut off and reported in Table 2 (2.5 ꢂ 10ꢁ6 moles/gcat) well
corresponds to that expected from reaction (5) based on the
amount of N2 formed (i.e. 8.18 ꢂ 10ꢁ6 ꢂ 2/8 moles/gcat = 2.045 ꢂ
10ꢁ6 moles/gcat).
Notably, also formation of N2 by global reaction (3) during CO-
TPSR and CO-ISC may involve the reaction of NCO species formed
by reaction (2) with adsorbed NOx species according to reaction
(5). NCO species are not completely consumed during the rich
phase of the CO-ISC experiment; their presence at the surface
was pointed out by their reaction toward gaseous oxygen and
NOx during the subsequent lean phase of a lean-fuel rich cycle, as
will be shown in the next paragraph. This fraction of NCO species
accumulated on the catalyst surface do not react with nitrates
either because they are not reactive being not close to nitrate spe-
cies and/or to vacant Pt sites or because residual nitrates are few
[11,14].
In our opinion, the occurrence of reaction (5) is pointed out by
the N2 formation at the shut off of the CO feed (insert of Fig. 5), due
to the availability of vacant Pt sites which were partially poisoned
by CO during the CO feed, i.e. in the presence of gaseous CO (self-
inhibition by CO). As a matter of fact, CO-ISC and pulse experi-
ments performed at lower temperatures and reported elsewhere
[31,34] showed that the nitrogen peak is always visible upon the
CO shut off, and that the peak is greater at lower temperatures.
This is in line with the hypothesis of a self-inhibition of CO due
to its strong adsorption on the Pt active sites.
Finally to explain the formation above 300 °C of ammonia
(5.07 ꢂ 10ꢁ5 moles/gcat in Table 1) and of hydrogen (4.75 ꢂ
10ꢁ5 moles/gcat in Table 1) during CO-TPSR as well as the formation
of very small amounts of ammonia during the CO-ISC experiment
(7.43 ꢂ 10ꢁ6 moles/gcat in Table 2), the presence of small amounts
of water in the feed and/or of surface hydroxyl species must be in-
voked. The formation of ammonia is likely due to hydrolysis of sur-
face NCO species [21], reaction (6):
Besides the results of CO-TPSR and CO-ISC experiments over
PtABa/Al2O3 (Figs. 3 and 5) and of the complementary FTIR study,
(Figs. 4 and 6) indicate that the reduction by CO of nitrates stored
onto the PtABa/Al2O3 catalyst occurs according to the stoichiome-
try of the following overall reactions:
BaðNO3Þ þ 2CO ! BaðNO2Þ2 þ 2CO2
BaðNO3Þ2 þ 8CO ! BaðNCOÞ þ 6CO2
ð1Þ
ð2Þ
ð3Þ
BaðNO3Þ2 þ 5CO ! BaCO3 þ2N2 þ 4CO2
2
Reactions (1)–(3) account for the consumption of CO and of nitrates,
and for the formation of CO2. In particular, reaction (1) represents
the reduction of the nitrates to nitrites; this reaction is not clearly
apparent from FTIR data when nitrate species are still abundant at
the surface, but it is evident both by Figs. 4A and 6A when nitrates
are almost disappeared. Reactions (2) and (3) account for the con-
sumption of nitrates with formation of cyanates/isocyanates, N2
and CO2; they also explain the fact that the N balance and the C bal-
ance in the gas phase do not close at the end of both CO-TPSR and
CO-ISC experiments. Reaction (3) also accounts for the formation of
carbonates at the catalyst surface, as revealed by FTIR. Notice that
reaction (1) is incorporated into reactions (2) and (3). Also, note that
in reactions (1)–(3) the participation of Ba bulky species has been
considered for simplicity, but the involvement of surface species
is most likely. The same comment applies for the other reactions
discussed in the following.
The consumption of nitrates and the formation of NCO species,
according to reaction (2), are revealed by the FTIR spectra which
have been collected at T P 200 °C, and is in line with the fact that
the C and N balances in the gas phase are not closed in the overall
range of low temperatures examined (180–250 °C). This suggests
that reaction (2), i.e. formation of NCO species, is effective in the
180–250 °C temperature range; in the same temperature range
reaction (3) does not occur since N2 is not detected. Our results
compare well with those previously reported by Szailer et al.
[19]: these authors observed the formation of surface NCO species
near 150 °C and an appreciable reduction of nitrates above 200 °C
upon storage of NOx at RT onto a PtABa/Al2O3 catalyst and using
CO as the reducing agent. The formation of NCO species during
the reduction of NOx by CO was previously reported also by Unland
[18] and Lesage et al. [20].
It is worth mentioning that in reaction (3) the formation of
BaCO3 has been envisaged upon reduction of the stored nitrates
with formation of N2. This is consistent with FTIR data shown in
Fig. 4 (where carbonate species are detected already in small
amounts by reduction at 250 °C) and in Fig. 6 (by reduction at
350 °C). However, BaO can also be present due to incomplete reac-
tion with CO2 (reaction 4):
BaðNCOÞ2 þ 3H2O ! BaO þ 2NH3 þ 2CO2
ð6Þ
Notice that the temperature of onset of ammonia formation in Fig. 3
may be slightly lower than 300 °C due to a delay in the dynamic re-
sponse of ammonia in our apparatus. Such a delay has been ob-
served during dedicated ammonia transient experiments and is
likely due to ammonia adsorption–desorption in the capillary of
the MS, maintained at 90 °C.
On the other hand, the formation of hydrogen is likely due to
the WGS reaction, reaction (7):
CO þ H2O ! CO2 þ H2
ð7Þ
ð8Þ
and/or to ammonia decomposition above 350 °C, reaction (8):
NH3 ! 1=2N2 þ 3=2H2
Indeed the occurrence of the WGS reaction was experimentally ob-
served over this catalyst from a CO/H2O mixture. Also, the decom-
position of ammonia and the associated formation of N2 and H2 in
a 1/3 ratio were observed during NH3–TPSR of NOx stored onto
PtABa/Al2O3 above 350 °C [7] and the formation of H2 in Fig. 3 is
seen together with a shoulder in the CO2 trace centered at
T = 400 °C.
BaO þ CO2 ! BaCO3
ð4Þ

P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
171
3.4. Lean phase over PtABa/Al2O3 catalyst
2164
0.8
0.6
0.4
0.2
0.0
1.2
0.8
0.4
0.0
Fig. 7 displays the results of O2 interaction after the reduction
phase (Fig. 5); the subsequent NO admission (NOx storage phase)
is also shown. Upon admission of 3% v/v O2 (Fig. 7, t = 0 s), the evo-
lution of CO2 with maximum concentration of ꢀ850 ppm and of
nitrogen with maximum concentration of ꢀ450 ppm is observed,
along with that of NO and NO2 in minor amounts with maximum
concentration of ꢀ80 and ꢀ50 ppm respectively. These species
are produced by the oxidation of surface NCO groups that remain
adsorbed at the catalyst surface after reduction with CO, as dis-
cussed earlier.
The oxidation of NCO species by oxygen was also investigated
by means of FTIR. After NOx storage, the NCO species were accumu-
lated on PtABa/Al2O3 by contacting the sample for 5 min with CO
at 350 °C: after that the catalyst was outgassed for few minutes
and cooled down to RT. Then, the NCO species were contacted with
an excess of oxygen (10 mbar) and the temperature was progres-
sively increased. The corresponding FTIR spectra are shown in
Fig. 8. No changes are observed till 150 °C, when the bands corre-
sponding to NCO species begin to decrease; by rising the tempera-
ture the oxidation proceeds, and after 10 min at 350 °C of contact
the bands at 2222 and 2164 cmꢁ1 completely disappear. In parallel,
the bands in the region of 1700–1000 cmꢁ1 reveal the formation of
mainly carbonate (bands at 1555 and 1350 cmꢁ1) and nitrite spe-
cies (band at 1230 cmꢁ1) in the 150–300 °C range and, mainly ni-
trate species at 350 °C. As a result, the spectrum recorded at
350 °C after interaction of NCO with oxygen (spectrum e) perfectly
matches in the region 1700–1000 cmꢁ1 the spectrum recorded at
the same temperature during the reduction of the stored NOx by
CO 1 min after its admission (spectrum f, see insert of Fig. 8). This
indicates that in the presence of oxygen the NCO species can be
oxidized to nitrate and carbonate species; besides nitrates forma-
tion involves nitrites as intermediate.
f
a
b
e
f
c
e
2250 2100
1600 1400 1200 1000
Wavenumbers (cm-1)
d
1350
e
d
c
e
a
1400
2250 2100
1600
1200
1000
Wavenumbers (cm-1)
Fig. 8. FTIR investigation into the reactivity of NCO species toward oxygen. Curve a,
spectrum of NCO species formed at 350 °C and cooled down to RT; curves b–e:
spectra after admission of O2 and heating for 10 min at 150, 250, 300 and 350 °C,
respectively. In the inset curve e is compared with the spectrum of NOx reduction
with CO at 350 °C after 30 s of contact (curve f).
reported by Lesage et al. over a PtARh/Ba/Al2O3 catalyst [20], by
Scholz et al. over a PtABa/Al2O3 catalyst [22] and by Bion et al. over
a Ag/Al2O3 catalyst [21]. Reaction (10) is incorporated into reaction
(11); reactions (10) and (11) account for the formation of nitrite and
nitrate species revealed by FTIR. Formation of N2 according to reac-
tion (9) may occur via the interaction of nitrates or nitrites with
residual isocyanate/cyanate species according with reaction (5)
and (12), respectively:
5BaðNCOÞ2 þ 3BaðNO3Þ2 ! 8N2 þ 8BaCO3 þ 2CO2
ð5Þ
The results shown in Figs. 7 and 8 can be explained by the fol-
lowing overall reactions (9)–(11):
BaðNCOÞ2 þ BaðNO2Þ2 ! 2N2 þ 2BaCO3
ð12Þ
The amounts of the different species released upon the oxygen
addition were calculated from Fig. 7 and are presented in Table 3.
Upon oxygen addition, 3.39 ꢂ 10ꢁ5 moles/gcat of CO2 are released,
along with 1.7 ꢂ 10ꢁ5 moles/gcat of N2. This is in nice agreement
with the occurrence of reaction (9). Formation of BaO according
to reaction (9) is documented by the results of subsequent NO step
addition at t = 500 s (Fig. 7). Indeed after the NO addition CO2 evo-
lution is observed, with a delay of about 15 s. As already reported in
[17], over a fully carbonated surface CO2 evolution is observed with
no delay with respect to NO admission. The death time here ob-
served in the CO2 evolution is due to the adsorption of NOx at
BaO sites still present at the surface after the oxidation phase. From
the delay in the CO2 evolution it is calculated that the number of
BaO sites involved in the storage of NOx to give Ba(NO2)2 or
Ba(NO3)2 is 1.86 ꢂ 10ꢁ5 gmoles/gcat, corresponding to 0.76% of the
Ba catalyst loading. This number compares well to 1.7 ꢂ 10ꢁ5 gmo-
les/gcat of N2 generated in the previous oxidation phase of NCO spe-
cies by oxygen (see Table 3), that according to reaction (9) should be
equal to the gmoles/gcat of BaO.
Notably, reactions (9)–(11) do not account for the release of
small quantities of NOx upon oxygen addition reported in Table 3
(3.41 ꢂ 10ꢁ6 moles/gcat). However, this NOx quantity is of minor
importance since the N2 selectivity during the O2 addition calcu-
lated from Table 3 is ꢀ91% (3.41 ꢂ 10ꢁ6 moles/gcat of NO + NO2
vs. 1.7 ꢂ 10ꢁ5 moles/gcat of N2). Considering that nitrogen is
formed only by reaction (9), it is possible to estimate that
1.7 ꢂ 10ꢁ5 moles/gcat of Ba(NCO)2 are decomposed upon admission
of oxygen to give N2.
BaðNCOÞ þ 3=2O2 ! N2 þ BaO þ 2CO2
BaðNCOÞ2 þ 3O2 ! BaðNO2Þ þ 2CO2
BaðNCOÞ2 þ 4O2 ! BaðNO3Þ2 þ 2CO2
ð9Þ
ð10Þ
ð11Þ
2
2
Reactions (9)–(11) account for the consumption of NCO species
with oxygen and for the formation of CO2. Reaction (9) also ac-
counts for the evolution of N2 upon oxygen admission. The reaction
of isocyanates/cyanates with O2 to give N2 and CO2 was previously
1000
O2 admission
CO2
NOin
900
800
700
600
500
400
300
200
100
0
CO2
NO
N2
N2
NO2
NO
0
200
400
Time (s)
600
800
Fig. 7. Lean phase at 350 °C over the PtABa/Al2O3 catalyst following NOx adsorption
at 350 °C and reduction with CO (2000 ppm in He) at the same temperature. t = 0:
The oxidation of NCO species with O2 at 350 °C does not lead to
the consumption of all NCO species present at the catalyst surface.
O
2 admission (3% v/v in He); t = 500 s: NO admission (1000 ppm in He + 3% v/v O2);
t = 650 s: NO shut off.

172
P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
This is demonstrated by the results also displayed in Fig. 7 showing
that, upon the subsequent admission of NO to the reactor at
t = 500 s, small amounts of N2 are formed, with a time delay of
about 15 s. In fact the formation of N2 during NOx storage under
lean conditions can be explained by the oxidation of residual
NCO species present at the catalyst surface by means of surface ni-
trites, nitrates, NO/O2 and/or NO2, according to reactions (5) and
(12)–(14):
cycle and accordingly a reproducible cyclic behavior is attained
with the closure of the N balance.
These results indicate that the changes in the catalyst behavior
during CO pulses are primarily associated to the buildup of surface
NCO species during reduction of stored NOx by CO. These species
are subsequently released or evolved upon oxygen addition, upon
NO addition and upon the shut off of the NO feed.
A quantitative analysis of NCO species still present onto the cat-
alyst at the end of the reduction phase was performed in a dedi-
cated experiment. It is well known [20] that cyanates and
isocyanates decompose to ammonia and CO2 through reaction with
water, according to the following reaction:
5BaðNCOÞ2 þ 3BaðNO3Þ2 ! 8N2 þ 8BaCO3 þ 2CO2
ð5Þ
BaðNCOÞ þ BaðNO2Þ2 ! 2BaCO3 þ 2N2
ð12Þ
ð13Þ
ð14Þ
BaðNCOÞ2 þ 3=2NO þ 3=2O2 ! 5=2N2 þ BaCO3 þ CO2
BaðNCOÞ2 þ 4H2O ! BaðOHÞ2 þ 2NH3 þ 2CO2
ð15Þ
BaðNCOÞ2 þ 3=2NO2 ! 5=2N2 þ BaCO3 þ CO2
Accordingly, water was admitted to the reactor following the reduc-
tion of nitrates with CO at 350 °C. Ammonia and CO2 production
was observed and calculations confirmed that ammonia and CO2
production was in line with the reaction (15). The amounts of
NCO species present onto the catalyst evaluated from the produced
ammonia were in line with what estimated by N2 formation during
the lean phase.
2
The moles of NOx stored and of the gaseous species released or
evolved during the NOx storage phase (t = 500–650 s in Fig. 7) were
calculated from Fig. 7 and are shown in Table 3. It appears that
5.52 ꢂ 10ꢁ6 moles/gcat of N2 are released so that the number of
Ba(NCO)2 species oxidized according to reactions (5) and (12)–
(14) is between 2.21 ꢂ 10ꢁ6 and 3.45 ꢂ 10ꢁ6 moles/gcat; these
quantities are calculated in case only reaction (13) (or reaction
(14)) or only reaction (5) is effective, respectively. These data con-
firm that surface NCO species are still present after oxygen addition
and are oxidized during the subsequent NOx storage phase because
of the superior oxidizing capability of NO + O2, NO2, surface nitrites
and/or surface nitrates as compared to O2 alone. The oxidation of
isocyanates/cyanates over Pd/Al2O3 at low temperature (150 °C) fol-
lowing switch to a flow containing 500 ppm NO + 5% O2 was previ-
ously documented by Macleod and Lambert [23].
In fact, the final situation detected during the FTIR study carried
out on the oxidation of NCO species by O2 (Fig. 8) is slightly differ-
ent from that observed in Fig. 7 and discussed earlier because NCO
species completely disappear in the former case whereas they are
still present after oxygen addition during the lean phase of Fig. 7. It
is speculated that the formation of nitrites/nitrates together with
the total consumption of NCO species in the FTIR study is likely
due to the different and more severe conditions employed in this
case (e.g. the longer time of contact with oxygen). This is also con-
sistent with the role of reactive intermediates of nitrite/nitrate spe-
cies in reactions (12) and (5).
3.5. Reaction pathway and mechanism in the reduction by CO of
nitrates stored onto PtABa/Al2O3
3.5.1. Reaction pathway
Based on the previous discussion, the reaction pathway for the
reduction by CO of nitrates stored over PtABa/Al2O3 under dry con-
ditions can be suggested as shown in Fig. 10. It is likely that the
reduction of the stored nitrates proceeds according to an in series
reaction scheme through the subsequent formation of nitrites
(path a) and of NCO species (path b), followed by the reaction of
these species with nitrates/nitrites and/or O2, NO + O2 and NO2 to
give N2, CO2 and BaCO3 (path c).
The conversion of Ba(NO3)2 into Ba(NO2)2 and Ba(NCO)2 during
reduction by CO with the corresponding evolution of CO2 (paths a
and b) has been derived from the analysis of CO-TPSR and CO-ISC
experiments over PtABa/Al2O3 and from the complementary FTIR
investigation. These reactions were observed to a much lower ex-
tent over Ba/Al2O3 and accordingly they are catalyzed by Pt, but
they also require the presence of Ba.
On the basis of the results previously discussed, it is worth of
note that when CO is used as reducing agent, the stored NOx spe-
cies are reduced to N2 both during the lean and the rich phases.
This is well illustrated in Fig. 9, which compares the gmoles/gcat
of NO consumed from the gas phase during NO storage (lean
phase) with the gmoles/gcat of gaseous N-containing species
(mostly N2) released or evolved during both the lean and the rich
phases in a series of four consecutive lean-rich cycles, starting from
a lean phase over a conditioned catalyst sample (i.e. a catalyst sam-
ple cycled using H2 as reductant, see experimental). As shown in
Fig. 9, the amount of NO consumed during the lean phase at
350 °C (open triangles) is greater during the first cycle (1.75 ꢂ
The oxidation of NCO species in the presence of O2, NO/O2 and/
or NO2 to give N2 and CO2 in the gas phase and Ba(NO2)2, Ba(NO3)2
1,8x10-4
consumed NO
1,5x10-4
1,2x10-4
total N released
9,0x10-5
N released
10ꢁ4 gmoles/gcat
)
and decreases to the value of 1.53 ꢂ 10ꢁ4
in Rich phase
6,0x10-5
gmoles/gcat when the catalyst reaches reproducible cyclic opera-
tion. Besides during the first lean/rich cycle only a fraction (roughly
50%) of the NO consumed during the lean phase is released as gas-
eous N-containing species (full circles), almost entirely as N2 dur-
ing the rich phase. Indeed starting from a catalyst that has been
previously conditioned with H2, no N2 is observed both upon O2
admission and during the NOx storage. In the subsequent cycles,
the amount of released N-containing products increases since N2
is formed during the lean phase as well. At the fourth cycle, the
amount of NO consumed from the gas phase in the lean phase
equals that of the evolved N-containing species during the whole
3,0x10-5
N released
in Lean phase
0,0
1
2
3
4
Cycle
Fig. 9. Integral amounts of NO consumed from the gas phase (open triangles) and of
N released into the gas phase (as N2, NH3, etc., filled circles) during four lean/rich
cycles. Filled columns represent the total N released during the lean phase; empty
columns show the amounts of N released during the rich phase.

P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
173
CO
CO2
CO
CO2
a
e
b
d
c
Ba(NO3)2
Ba(NO2)2
Ba(NCO)2
BaCO3
O2,
N2
CO2
NO/O2,
Ba(NO2)2,
NO2,
CO2
O2
O2
Ba(NO3)2
Fig. 10. Reaction pathway in the reduction by CO of nitrates stored onto PtABa/Al2O3 catalyst.
and BaCO3 at the catalyst surface (paths c, d, e) has been revealed
during the experiment over PtABa/Al2O3 shown in Fig. 7 upon
admission of O2 and of NO/O2 and by the complementary FTIR
study. The oxidation of NCO species by means of Ba(NO3)2 and/or
Ba(NO2)2 (again path c) has also been invoked to explain the re-
lease of nitrogen during the CO-ISC experiment upon the shut off
of CO. All these reactions have been observed only in the presence
of Pt.
as reductant, vacant Pt sites (Ptꢃ) are formed by reaction of hydro-
gen with adsorbed oxygen atom (Ptꢄ ꢄ ꢄOa) left on the Pt surface
forming water. The water thus formed destabilizes surface nitrates
which could then decompose on the Pt surface with release of NOx
in the gas phase.
Vacant Pt sites are also generated upon reaction of adsorbed
oxygen with CO. Although the destabilization of surface nitrates
in the case of only CO and under dry conditions is not discussed
by the authors, it could be explained assuming that at first nitrates
spill over from the Ba particles to the Pt particles and are decom-
posed in the presence of CO with release of NOx species in the
gas phase, which then are dissociated on vacant Pt particles to form
adsorbed nitrogen atoms (Ptꢄ ꢄ ꢄNa). At low temperature, the pri-
mary reaction with CO is the removal of adsorbed oxygen atoms
from the Pt particles to form CO2. The vacant Pt sites become avail-
able for the adsorption of CO, which can then react with adsorbed
nitrogen atoms to form NCO species (PtANCO).
NCO formed on the Pt particles spill over to the oxide compo-
nent of the catalyst. The reaction to form nitrites and eventually
NCO species is faster than the reaction of adsorbed nitrogen atoms
to give N2 (2Ptꢄ ꢄ ꢄNa ? 2Ptꢃ + N2). Accordingly, at low temperature
the formation of nitrites, NCO species and CO2 but not of N2 is ob-
served (viz. at 180–250 °C during CO-TPSR in Fig. 3).
At higher temperatures, these NCO species can readily react
with the NOx species stored by the catalyst (NCOa + NOx ?
N2 + CO2). It is not clear whether this reaction does involve ad-
sorbed or gas-phase NOx; however, the authors observed that the
onset of this reaction corresponds to the onset of nitrate decompo-
sition, suggesting a possible intermediation of gaseous NOx.
This mechanistic proposal appears consistent with the typical
behavior of LNT systems, characterized by long lean phases and
short rich periods during which the stored nitrates (and nitrites)
decompose to gaseous NOx due to the relevant exothermicity that
takes place upon lean/rich switch. This mechanism is also consis-
tent with several features of the reaction pathway depicted in
Fig. 10 and of the transient reactivity data presented in this paper.
However, it does not seem fully consistent with all the features
pointed out in the present study, and in particular:
The CO-TPSR experiment over PtABa/Al2O3 has shown that the
reaction of surface nitrates to give nitrogen is slower than that
responsible for the initial reduction of nitrates to give either ni-
trites or NCO species. In fact, the formation of nitrogen is observed
only above 250 °C while evidence for the formation of NCO species
and of nitrites from 180 °C is provided by the consumption of CO
and by the simultaneous evolution in the gas phase of CO2 with
no N-containing species. In fact both the C and N balances do not
close in the temperature interval 180–250 °C (Fig. 3 and Table 1).
The formation of nitrogen via the reduction of nitrates by CO
has been revealed during the CO-ISC experiment in Fig. 5 and the
complementary FTIR study in Fig. 6. In fact under the CO-ISC con-
ditions, this reaction is explained as the sum of the reduction of ni-
trates by CO to give surface NCO species [reaction (2)] and the
subsequent oxidation of these species by surface nitrates to give
nitrogen [reaction (5)], while this last reaction is controlled by
the presence of reactive nitrate species as the reduction of nitrates
proceeds to a significant extent. The results in Fig. 7 and the com-
plementary FTIR study in Fig. 8 have demonstrated that NCO spe-
cies are left at the catalyst surface after reduction of CO; these
species can be oxidized to give N2 during the subsequent lean
phase either by oxygen or by NO + O2, NO2, surface nitrites and/
or surface nitrates. Finally, the oxidation of surface NCO species ob-
served upon O2 addition in Fig. 7 with formation of N2 is also quan-
titatively explained by the reaction between nitrates (or nitrites),
obtained by oxidation of NCO species, with residual NCO species
[reaction (5) or reaction (12)]. Indeed the FTIR study has also
shown that surface nitrites and nitrates are reformed by oxidation
of NCO species (paths d and e). Thus, it appears that the formation
of nitrogen either during the reduction of stored NOx by CO, or dur-
ing the oxidation of NCO species upon oxygen addition or during
the storage of NOx upon admission of NO in the presence of excess
oxygen is always explained by the reaction between NCO species
and surface nitrates (or nitrites), where NCO species, nitrites and
nitrates can be transformed ones into the others via oxidation or
reduction.
(i) all the nitrogen containing species involved in the reaction
pathway in Fig. 10 are adsorbed at the catalyst surface so
that one could expect that the mechanism responsible for
the formation of nitrogen proceeds entirely through
adsorbed species. It is worth noting that in the present study
the reduction of stored NOx is accomplished at 350 °C under
near isothermal conditions with limited concentration of CO
to prevent thermal decomposition of nitrates with release of
NOx in the gas phase; accordingly there is no need to invoke,
as a preliminary step, the decomposition of nitrates or
nitrites with release of NOx into the gas phase which are
3.5.2. Mechanistic considerations
The mechanism of the reduction of stored NOx by H2 and by CO
over PtABa/Al2O3 catalyst under dry conditions has been previ-
ously discussed by Szailer et al. [19] and also reported by other
authors [22]. Szailer et al. proposed that when hydrogen is used

174
P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
then re-adsorbed onto the catalyst with decomposition and
further rearrangement to give nitrogen (via surface NCO
species);
Reaction (III) accounts for the subsequent reduction step to
form oxyisocyanate ions. The trimer of hydroxamic acid
HOAN@C@O was synthesized and characterized in solid state by
IR, MS, and TG analysis and in DMSO-d6 solutions by NMR and in
aqueous solution by IR as well to prove its chemical structure, to
(ii) along similar lines the assistance of CO in the decomposition
of nitrates which results in the release of NOx species in the
gas phase must account for the very large decrease of the
onset temperature of the reduction by CO of stored NOx from
350 °C (seen during thermal desorption) down to 180 °C
(observed during CO-TPSR). An even larger effect is seen in
the case of H2 [7], with a decrease of the onset temperature
from 350 °C down to 70 °C;
reveal
a number of physico-chemical properties related to
hydroxamate moieties and to indicate hydroxylisocyanate as a
possible degradation product in [36]. Theoretical studies on the
oxyisocyanate negative ions are also available in the literature [37].
Reaction (IV) describes the successive reduction of oxyisocya-
nate ions to give NCO species.
(iii) N2 is formed primarily according to an in series two steps
process where NCO species are formed first and then are
converted to nitrogen upon reaction with NOx,a species.
Accordingly, the alternative route to form nitrogen based
on the reaction of adsorbed nitrogen atoms to give N2 is
believed to be poorly active not only at relatively low tem-
perature but also at higher temperature, which seems not
in line with literature reports on the NO reduction by hydro-
gen over PtABa/Al2O3 [14];
(iv) the mechanism responsible for the formation of nitrogen
through oxidation of NCO species upon admission of oxygen
during the lean phase of the cycle implies that NCO species
that are formed by reduction of stored NOx can undergo re-
oxidation to give nitrogen. This feature, along with the reac-
tion of NCO species with either NO/O2, NO2 or nitrites/
nitrates to form nitrogen, is not consistent with the mecha-
nism proposed by Szailer et al. [19].
All reactions II–IV, catalyzed by Pt, require the spillover of the
species from the Ba component onto the Pt particles at the PtABaO
interface. According to these scheme nitrates are reduced to NCO
species in a stepwise manner, so that the formal oxidation number
of nitrogen is reduced progressively from +5 (in nitrate ad-species)
to ꢁ3 (for NCO ad-species) according to a process which involves
the transfer of two electrons in each step. Notably, the intermedi-
ates involved in reactions I–IV have been spectroscopically de-
tected or synthesized previously. In particular, nitrates, nitrites
and NCO species have been observed in the present study by FTIR.
Besides CO2NOꢁ ions have been synthesized at very low tempera-
ture (11 K) and under dilute conditions while hydroxyisocyanate
has been observed only upon decomposition of the more stable
cyclic trihydroxamic acid. Accordingly, these intermediates are
expectedly very reactive and indeed they are not detected in the
present study. As a consequence, the reduction of nitrates to give
NCO species is likely controlled by the reduction of nitrates to give
nitrites. On the other hand, considering that the steps involved in
the reduction of nitrites to give NCO species are fast they can pro-
ceeds through reactive intermediates different from those pro-
posed in reactions II and III, but generated via the same or a
similar chemical route.
A possible alternative mechanism is presented in the following
for the reduction of surface nitrates by CO to give nitrogen over
PtABa/Al2O3 under dry conditions.
The first step in the reaction scheme is represented by the
reduction of nitrates by CO adsorbed at a Pt site (Ptꢄ ꢄ ꢄCOa) to give
nitrites with evolution of CO2 (reaction I):
The formation of NCO species is then followed by their reaction
with residual nitrites to give nitrogen, reaction (V):
ABaAONO2 þ Pt ꢄ ꢄ ꢄ COa ! ABaAONO þ CO2 þ Ptꢃ
ðIÞ
ABaANCO þ ABaAONO ! ABaAOABaA þ N2 þ CO2
ðVÞ
This reaction has been demonstrated by the FTIR data in Figs. 4 and
6, where the reduction of nitrates upon contact with CO has been
studied at increasing temperature and at increasing time of contact
and constant temperature. It has also been proved that the reaction
is catalyzed by Pt.
The present study has shown that surface nitrites are then re-
duced by CO to give surface NCO species and that this reaction is
much faster over PtABa/Al2O3 than over Ba/Al2O3 and therefore
is catalyzed by Pt as well. Due to the large variation in the oxida-
tion state of N associated with the overall reaction from nitrites
to NCO species (i.e., from formal oxidation state +3 to ꢁ3), it is
most likely that this reaction proceeds in a stepwise manner.
Accordingly, the following stepwise reduction of nitrites to give
NCO species is speculated that seems consistent on a chemical
ground and where only intermediate species already detected or
synthesized are formed (reactions II–IV):
Reaction (V) is consistent with the data in Figs. 3 and 4 and in Figs. 5
and 6 showing that the reduction of nitrates to give N2 during CO-
TPSR and during the CO-ISC experiment always proceeds with for-
mation of nitrites. It is also consistent with the data in Figs. 7 and
8 which show that NCO species are oxidized to give N2 upon oxygen
admission and subsequent admission of NO in the presence of ex-
cess oxygen and that the oxidation of NCO species leads to the for-
mation of nitrites as intermediate species.
Indeed once NCO species are formed they are very susceptible
to nucleophilic attack and therefore they can react with nitrite ions
with intra-molecular rearrangement, internal redox process and
CO2 loss to give dinitrogen molecule as described by reaction (V).
An example of this chemistry is provided by the reaction of aryl
isocyanates with nitrite ion, as reported in [38].
This reaction has also strict analogies with that of amide species
with NO to give nitrosoamide [39,40] or that of ammonium ion
with nitrite ion to give ammonium nitrite [41,42]. Both nitrosa-
mide and ammonium nitrite have been proposed as the key inter-
mediates responsible for the production of nitrogen in the SCR of
NOx by ammonia over V- and W-based oxide catalysts [40,43]
and Fe- and or Cu-exchanged zeolites [44–46] indeed are known
to decompose easily and with full selectivity to nitrogen and water
[47].
ABaAONO þ Pt ꢄ ꢄ ꢄ COa ! ABaAOCONO þ Ptꢃ
ABaAOCONO þ Pt ꢄ ꢄ ꢄ COa ! ABaAOCNO þ CO2 þ Ptꢃ
ABaAOCNO þ Pt ꢄ ꢄ ꢄ COa ! ABaANCO þ CO2 þ Ptꢃ
ðIIÞ
ðIIIÞ
ðIVÞ
Reaction (II) describes the reduction of nitrites to give a CO2NOꢁ
ions by CO insertion. The experimental observation of the covalent
bonded CO2NOꢁ ion produced by co-condensation of laser-ablated
metal atoms and electrons with NO and CO2 mixtures in excess
Ar and at 11 K has been reported in [35]. Matrix isolation FTIR com-
bined with theoretical calculations has been employed to identify
the anion. The anion assignment was further supported by photol-
ysis and CCl4 doping experiments.
4. Conclusions
In this paper, the reactivity of NOx stored at 350 °C onto PtABa/
Al2O3 LNT catalyst in the reduction by CO under dry conditions was

P. Forzatti et al. / Journal of Catalysis 274 (2010) 163–175
175
[8] P. Forzatti, L. Lietti, I. Nova, Energy Environ. Sci. 1 (2008) 236.
[9] J.A. Pihl, J.E. Parks, C.S. Daw, T.W. Root, SAE Technical Paper 2006-01-3441,
2006.
investigated by means of transient experiments (CO-TPSR and iso-
thermal CO-ISC) and complementary FTIR study.
It was shown that the reduction by CO of nitrates stored onto
PtABa/Al2O3 at high temperature under nearly isothermal condi-
tions occurs through a Pt-catalyzed pathway, which does not in-
volve the thermal decomposition of stored NOx with release of
NOx in the gas phase and leads mainly to nitrogen and CO2.
Transient experiments performed both under temperature pro-
gramming (CO-TPSR) and under isothermal conditions (CO-ISC)
and the complementary FTIR study have suggested that (i) the
reduction of nitrates by CO occurs through a consecutive reaction
scheme which includes the reduction of nitrates into nitrites and
then into NCO species followed by the reaction of these species
with residual nitrates/nitrites to give nitrogen; (ii) the reaction of
surface nitrates/nitrites with NCO species to give nitrogen is slower
than the reduction of nitrates to give either nitrites or NCO species;
(iii) the formation of nitrogen that is, observed during the reduc-
tion of stored NOx by CO and during the oxidation of NCO species
upon oxygen/NO + O2 addition, can be explained in all cases by the
reaction between NCO species and surface nitrates and nitrites.
A mechanism was proposed for the reduction of surface nitrates
by CO under dry conditions, which proceeds through a stepwise
reduction of nitrates in the order into nitrites, CO2NOꢁ ions, oxyis-
ocyanate ions and finally NCO ions and involves the transfer of two
electrons in each step. As opposite to other mechanisms previously
proposed in the literature, the suggested scheme involves only ad-
sorbed species.
[10] L. Cumaranatunge, S.S. Mulla, A. Yezerets, N.W. Currier, W.N. Delgass, F.H.
Ribeiro, J. Catal. 246 (2007) 29.
[11] H. Mahzoul, J.F. Brilhac, P. Gilot, Appl. Catal. B: Environ. 20 (1999) 47.
[12] A.J. Paterson, D.J. Rosenberg, J.A. Anderson, Stud. Surf. Sci. Catal. 138 (2001)
429.
[13] Y. Sakamoto, K. Okumura, Y. Kizaki, S. Matsunuga, N. Takahashi, H. Shinjoh, J.
Catal. 238 (2006) 361.
[14] L. Castoldi, I. Nova, L. Lietti, L.P. Forzatti, Catal. Today 96 (2004) 43.
[15] R.D. Clayton, M.P. Harold, V. Balakotaiah, Appl. Catal. B 84 (2008) 616.
[16] D. Bhatia, R.D. Clayton, M.P. Harold, V. Balakotaiah, Catal. Today 147 (2009)
250.
[17] L. Lietti, P. Forzatti, I. Nova, E. Tronconi, J. Catal. 204 (2001) 175.
[18] M.L. Unland, J. Phys. Chem. 77 (1973) 1952.
[19] T. Szailer, J.H. Kwak, D.H. Kim, J.C. Hanson, C.H.F. Peden, J. Szanyi, J. Catal. 239
(2006) 51.
[20] T. Lesage, C. Verrier, P. Bazin, J. Saussey, M. Daturi, Phys. Chem. Chem. Phys. 5
(2003) 4435.
[21] N. Bion, J. Saussey, M. Haneda, M. Daturi, J. Catal. 217 (2003) 47.
[22] C.M.L. Scholz, B.H.W. Maes, M.H.J.M. de Croon, J.C. Schoulten, Appl. Catal. A:
Gen. 332 (2007) 1.
[23] N. Macleod, R.M. Lambert, Appl. Catal. B: Environ. 46 (2000) 483.
[24] F. Frola, M. Manzoli, F. Prinetto, G. Ghiotti, L. Castoldi, L. Lietti, J. Phys. Chem. C
112 (2008) 12869.
[25] I. Nova, L. Castoldi, L. Lietti, E. Tronconi, P. Forzatti, F. Prinetto, G. Ghiotti, J.
Catal. 222 (2004) 377.
[26] Prinetto, G. Ghiotti, I. Nova, L. Lietti, E. Tronconi, P. Forzatti, J. Phys. Chem. B
105 (2001) 12732.
[27] F. Frola, F. Prinetto, G. Ghiotti, L. Castoldi, I. Nova, L. Lietti, P. Forzatti, Catal.
Today 126 (2007) 81.
[28] G. Socrates, Infrared Characteristic Group Frequencies, John Wiley & Sons,
Great Britain, 1980.
[29] N. Bion, J. Saussey, C. Hedouin, T. Seguelong, M. Daturi, Phys. Chem. Chem.
Phys. 3 (2001) 4811.
[30] F. Frola, Ph.D. Thesis, University of Turin, 2007.
[31] I. Nova, L. Lietti, P. Forzatti, F. Frola, F. Prinetto, G. Ghiotti, Top. Catal. 52 (2009)
1757–1761.
[32] A. Lokhov, A.A. Davydov, Kinet. Catal. 21 (1980) 1523.
[33] P. Hollins, Surf. Sci. Rep. 16 (1992) 51.
It was noted that NCO species, nitrites and nitrates can be trans-
formed into the others upon oxidation or reduction. This provides a
unifying view to explain the production of nitrogen in the reduc-
tion by CO of NOx stored as nitrites and/or nitrates, as well as in
the reduction by isocyanates and in the oxidation of NCO species
under dry conditions.
[34] I. Nova, L. Lietti, P. Forzatti, F. Prinetto, G. Ghiotti, Catal. Today 151 (2010) 330.
[35] M. Zhou, L. Zhang, Q. Qin, JACS 122 (2000) 4483.
[36] M.J.-M. Tarak, I. Kos, M. Birus, I. Butula, M. Gabricevic, J. Mol. Struct. 782 (2006)
24.
[37] R.J. Waugh, P.C.H. Eichinger, R.A.J. O’Hair, J.C. Sheldon, J.H. Bowie, R.N. Hayes,
Rapid Commun. Mass Spectrom. 3 (5) (1989) 151.
[38] N.P. Botting, B.C. Challis, J. Chem. Soc. Chem. Commun. 1585 (1989).
[39] L.J. Stief, W.D. Brobst, D.F. Nava, R.P. Borkowski, J.V. Michael, J. Chem. Soc.
Faraday Trans, II (1982) 1391.
References
[1] N. Takahashi, H. Shinjoh, T. Iijima, T. Suzuki, K. Yamazaki, K. Yokota, H. Suzuki,
N. Miyoshi, S. Matsumoto, T. Tanizawa, T. Tanaka, S. Tateishi, K. Kasahara,
Catal. Today 27 (1996) 63.
[40] G. Ramis, G. Busca, F. Bregani, P. Forzatti, Appl. Catal. 64 (1990) 259.
[41] P. Forzatti, Appl. Catal. A: Gen. 222 (2001) 221.
[42] M. Li, Y. Yeom, E. Weitz, W.M.H. Sachtler, J. Catal. 235 (2005) 201.
[43] L. Lietti, J. Svachula, P. Forzatti, G. Busca, G. Ramis, F. Bregani, Catal. Today 17
(1993) 131.
[44] I. Nova, C. Ciardelli, E. Tronconi, D. Chatterjee, B. Blandl-Konrad, Catal. Today
114 (2006) 3.
[45] A. Grossale, I. Nova, E. Tronconi, D. Chatterjee, M. Weibel, J. Catal. 256 (2008)
312.
[2] W.S. Epling, L.E. Campbell, A. Yezerets, N.W. Currier, J.E. Parks, Catal, Rev. Sci.
Eng. 46 (2004) 163.
[3] H. Shinojh, N. Takahashi, K. Yokota, M. Sugiura, Appl. Catal. B: Environ. 15
(1998) 189.
[4] I. Nova, L. Lietti, L. Castoldi, E. Tronconi, P. Forzatti, J. Catal. 239 (2006) 244.
[5] L. Castoldi, L. Lietti, I. Nova, R. Matarrese, P. Forzatti, F. Vindigni, S. Morandi, F.
Prinetto, G. Ghiotti, Alkaline and alkaline-earth oxides based Lean NOx Traps:
effect of the storage component on the catalytic reactivity, Chem. Eng. J.
(2009), doi:10.1016/j.cej.2009.10.065.
[46] Y.H. Yeom, J. Henao, M.J. Li, M.H. Sachtler, E. Weitz, J. Catal. 231 (2005) 181.
[47] R. Schwartz, H. Giese, Chem. Ber. 67 (1934) 1111.
[6] I. Nova, L. Lietti, P. Forzatti, Catal. Today 136 (2008) 128.
[7] L. Lietti, I. Nova, P. Forzatti, J. Catal. 257 (2008) 270.