(401a) Formation and Removal of Ba-Carbonates/Carboxylates On Pt/Ba/Al2O3 Lean NOx Traps

Authors: 
Kispersky, V., Purdue University
Chaugule, S. S., Purdue University
Yezerets, A., Cummins Inc.
Currier, N. W., Cummins Inc.
Ribeiro, F. H., Purdue University
Delgass, W. N., Purdue University


NOx
adsorption and removal on Pt/BaO/Al2O3 lean NOx
traps (LNT) has been a major focus of study in the exhaust aftertreatment
community.  The practically important effects of CO2 and H2O
on NOx adsorption have been studied in our laboratory using NOx
breakthrough curves obtained from flow reactor experiments [1]
as well as by other research groups [2].
In our work, we focused on samples with 4, 8 and 20 wt.% Ba to investigate the
effects of the state of Ba at the beginning of NOx adsorption on the
amounts of NOx stored before slip occurs.  Each of our experiments
contained a sequence of NOx capture and regeneration with CO2
and H2O in the feed followed by purge in inert and finally NOx
capture and regeneration without CO2 and H2O at 300°C.
The amount of pre-adsorbed CO2 released during NOx capture
upon the introduction of NO2 correlated well with available Ba sites
for NOx storage and the amount of NOx stored on the monolithic
trap with 20Ba but not on the traps with lower Ba loadings. Subsequent DRIFTS
experiments in the form of sequential adsorption of CO2, NO2
+ O2, H2 or NO2 + O2, CO2,
H2 at 300°C were completed.  At low loadings of Ba (8Ba and 4Ba) the
CO2 adsorbed mainly in the form of carboxylates, presumably on highly
dispersed Ba sites.  On 20Ba samples, where the BaO phase is known to form in
large, bulk-like structures, CO2 was found to adsorb only as a
carbonate.  Competitive adsorption between CO2 and NOx
encouraged replacement of the pre-adsorbed species by the one in the gas phase.
 Given enough time, NOx was able to completely displace all CO2
and subsequent regeneration with H2 returned the catalyst to its original
clean state, i.e. BaO.  When pre-adsorbed NOx was partially
displaced by carbonates/carboxylates, regeneration by H2 was able to
remove all NOx adsorbates and any carboxylates, but not the
carbonates.  The implications of this adsorbate behavior played out in the
following manner.  During regeneration in the flow reactor with H2, and
with CO2 and H2O also present, adsorbed nitrates were removed,
producing Ba sites available for CO2 storage.  On the 20Ba sample,
CO2 adsorbed as carbonate and remained on the storage component
during regeneration and purge.  On the 8Ba and 4Ba samples, which
preferentially formed carboxylates, the instability of the adsorbed
carboxylates in H2 ensured that the majority of available Ba sites
remained open (as BaO), thus explaining the loss of correlation between the
amount of pre-adsorbed CO2 and NOx storage capacity
mentioned above.  Furthermore, the low Ba loadings exhibited greater use of
total Ba for fast NOx storage, defined as the amount of Ba used to
store NOx until 1% of inlet NOx breakthrough occurs in
the flow reactor, which were 9, 8 and 4% for the 8Ba, 4Ba and 20Ba,
respectively.  The diminished use of Ba on the 20Ba sample is explained by
carbonates blocking available Ba sites vicinal to Pt centers which facilitate
fast NOx adsorption by supplying the surrounding Ba with spilled
over oxygen.  It has been shown that NOx will adsorb first on BaO,
then Ba(OH)2 and finally BaCO3 sites when all sites are
present in the catalyst [3]
as was the case in our experiments.  Carbonates adsorbed on Ba vicinal to Pt
lead to slowed NOx adsorption kinetics which cause the reduced
amount of Ba used for fast NOx storage on the 20Ba catalysts.  Thus
the initial state of the Ba trapping phase becomes greatly important and a
properly chosen Ba loading can lead to improved Ba usage.  Different Pt
loadings were also investigated, and on the 20Ba samples they primarily
affected the carbonate/nitrate exchange through supply of the oxidant for
formation of nitrates leading to higher amounts of fast NOx storage
with increased Pt loading. Implications of these chemical findings on NSR
catalyst optimization will be discussed.

References

[1]
S.S. Chaugule, A. Yezerets, N.W. Currier, F.H. Ribeiro, W.N. Delgass, Catal.
Today, In Press, Corrected Proof, 2010, doi: 10.1016/j.cattod.2010.02.024

[2]
W.S. Epling, G.C. Campbell, J.E. Parks, Catal Lett, 90 (2003) 45-56.

[3]
I. Nova, L. Castoldi, L. Lietti, E. Tronconi, P. Forzatti, Catal. Today, 75
(2002) 431-437.