B. Pereda-Ayo et al. / Journal of Catalysis 285 (2012) 177–186
179
2.3.2. Moving to shorter regeneration periods by increasing H2
concentration
species and consequently it can be observed that NH3 is only de-
tected after 40 s into the regeneration.
In order to more closely reflect real conditions, the NOx storage
and reduction experiments were carried out limiting the length of
the storage and reduction periods to 2 min. During the regenera-
tion, the higher H2 concentration allowed us to shorten the regen-
eration time of the catalysts to a few seconds. The hydrogen
concentration was varied from 0.46% to 3.83%, and the experi-
ments were run with either 14NO or 15NO present during the
regeneration.
The presence of NO during regeneration increases the total
amounts of N2, NH3 and H2O significantly, whereas the formation
of N2O remains practically unchanged, suggesting that stored ni-
trates are the only source for N2O formation and that incoming
NO does not participate in its formation. As expected, the addition
of NO during the rich period increases the time needed to obtain
the complete regeneration of the trap. This can be understood be-
cause apart from stored nitrates, hydrogen also has to reduce the
incoming gas phase NO. Consequently, the hydrogen breakthrough
is delayed for 40 s as can be observed in Fig. 1b. The complete con-
sumption of H2 during the initial period of the regeneration to-
gether with the rectangular shape of the H2O formation curve
indicates a ‘‘plug flow’’ type of mechanism. As several authors have
already reported [9,14,15,25], the hydrogen front travels through
the catalyst bed with complete regeneration of the trapping sites
as it propagates down the bed. Those reactions, leading to the final
regeneration of the trap, are exothermic, increasing the tempera-
ture of the catalyst bed from 189 °C to a maximum of 196 °C, which
is coincident in time with the hydrogen breakthrough, or the end of
the regeneration. After complete regeneration has been obtained,
the catalyst temperature slowly decreases until the initial temper-
ature is recovered.
In the absence of NO during regeneration, the signal corre-
sponding to N2, NH3, N2O and H2O decreased slowly back to the
baseline once the regeneration of the trap was finished. However,
in the presence of NO, the signals corresponding to N2, NH3 and
H2O remained constant at values higher than the baseline. Thus,
it can be deduced that after catalyst regeneration, when no nitrates
are present in the catalyst surface, the reaction between NO and H2
led to the formation of those products, with the production of NH3
being more favourable than N2 due to the low NO/H2 ratio, as re-
ported by Mulla et al. [15]. The reaction between NO and H2 once
the trap was regenerated led to a total reduction in NO, as the (m/
e = 30) signal was maintained below 20 a.u. Consequently, a partial
consumption of H2 was observed, as compared with the (m/e = 2)
H2 signal in the absence of NO (Fig. 1b).
3. Results and discussion
3.1. Effect of the presence of NO during LNT regeneration
As indicated in Introduction, most work dealing with mechanis-
tic aspects of NOx storage and reduction has omitted to include NO
during the regeneration part of the experiments. It is important to
investigate whether or not the inclusion of NO during regeneration
changes the product profiles. For these experiments, the lean gas
mixture composed of 800 ppm NO, 6% O2 and Ar to balance was
fed to the Pt–Ba/Al2O3 catalyst until saturation of NOx. As the pre-
vious studies reported [11], the NOx storage capacity shows a vol-
cano-type dependence on the temperature, with a maximum
typically occurring at 350 °C. Therefore, 15 min was needed to sat-
urate the catalyst when the reaction was carried out at 190 °C,
whereas 20 min was needed at 340 °C. After saturation, the feed-
stream was switched to rich conditions, and the catalyst regenera-
tion was carried out in the presence or absence of 800 ppm NO
along with 0.46% H2 and Ar to balance.
During the regeneration, the evolution of different mass signals
which can be assigned to different products was recorded and
these are plotted in Fig. 1. Fig. 1a shows the evolution of NO (m/
e = 30), N2 (m/e = 28), NH3 and OH (m/e = 17) and NH (m/e = 15)
signals in the presence (red1 points) and in the absence (black
points) of NO when the regeneration was carried out at 190 °C. Sim-
ilarly, Fig. 1b shows the evolution of N2O (m/e = 44), H2O (m/e = 18),
H2 (m/e = 2) and the exit temperature profile. The Y axis depicted in
Figs. 1 and 2 is shown in arbitrary units except in case of tempera-
ture. The same procedure was used to investigate the NOx reduction
performance when the reaction was carried out at 340 °C (Fig. 2a and
b).
As can be observed in Fig. 1a, when the rich feed is admitted to
the catalyst (t = 0), the NO signal drops from the saturation level
(300 a.u.) to below 20 a.u. in the presence and in the absence of
NO. However, in the presence of NO during regeneration, a sudden
release (known as ‘‘NOx puff’’) is recorded at the beginning of the
rich period, which is not observed in the absence of NO. Therefore,
it can be concluded that when NO is admitted during regeneration,
the supply of reductant is not sufficient to reduce the NO being re-
leased from the catalyst surface as well as that present in the gas
stream. Simultaneously, the reduction in stored nitrates with the
incoming H2 led to the formation of N2, NH3, N2O and H2O. As
can be observed in Fig. 1a and b, the formation of N2 and N2O is de-
tected immediately after the rich period starts, whereas the detec-
tion of H2O and NH3 is delayed to some extent. The (m/e = 18)
signal corresponding to H2O is detected after 20 s of regeneration.
At the same time, together with H2O breakthrough, the (m/e = 17)
signal breakthrough is detected due to the OH (m/e = 17) fragmen-
tation coming from H2O. However, if the monitoring of ammonia is
carried out using the NH fragment (m/e = 15), it is possible to
remove any contributions from mass fragmentation from other
Several differences can be observed when the NOx storage and
reduction is carried out at 340 °C (Fig. 2). First, a longer lean period
time was needed to saturate the catalyst, which means that higher
quantities of NOx were stored and consequently the formation of
larger quantities of products is expected. In fact, the areas under
the N2 and H2O curves increased by 50% and 62%, respectively. In
contrast, the formation of N2O was negligible and NH3 formation
was only detected after 120 s of regeneration.
The results obtained in the absence of NO during regeneration
are in agreement with mechanistic aspects of the regeneration al-
ready reported [13–15,18]. The reduction in stored nitrates with
hydrogen has been reported to occur by the following reactions:
BaðNO3Þ þ 8H2 ! 2NH3 þ BaO þ 5H2O
ð1Þ
ð2Þ
BaðNO3Þ2 þ 5H2 ! N2 þ BaO þ 5H2O
2
Lietti et al. [18] reported that during reduction in stored nitrates
at 100 °C, reaction (1) accounted for almost all the H2 consump-
tions, demonstrating that stored nitrates were reduced efficiently
and selectively (>90%) to ammonia. On increasing the reduction
temperature, nitrogen formation was promoted due to reaction
(3) where the ammonia formed continues to react further with
stored nitrates to form nitrogen.
3BaðNO3Þ2 þ 10NH3 ! 8N2 þ 3BaO þ 15H2O
ð3Þ
Thus, nitrogen formation involves a two-step pathway: the fast
formation of ammonia by reaction of nitrates with H2 (reaction (1))
and the subsequent reaction of the ammonia formed with stored
1
For interpretation of colour in Figs. 1–5 and 7–9, the reader is referred to the web
version of this article.