(82a) Experimental and Modeling Studies of Cycle Frequency, Reductant Type and Non-Isothermal Effect on the Performance of a Lean NOx Trap

Experimental and
Modeling Studies of Cycle Frequency, Reductant Type and Non-isothermal Effect
on the Performance of a Lean NOx Trap

Allen Wei-Lun
Ting, Michael P. Harold and Vemuri Balakotaiah

Department of Chemical &
Biomolecular Engineering, University of Houston, Houston, TX 77204, USA

Abstract

The
¡§Di-Air¡¨ system proposed by Toyota researchers shows an overall increase in NOx
conversion when using propene as the reductant under fast cycling [1]. In order
to clarify this mechanism, a combined experimental and modeling study of fast
cycling NOx storage and reduction for emissions control of lean burn gasoline
and diesel vehicles is conducted to provide mechanistic insight. NOx generation
pathways are analyzed by using H₂ and C₃H₆ as
reductants, under different cycle frequencies, and under near-isothermal and
non-isothermal conditions. LNT (lean NOx trap) catalysts containing Pt/BaO/CeO₂/Al₂O₃
are used in the monolithic reactor experiments. A 1-D two phase model that
includes washcoat diffusional effects is used to simulate the cyclic
experimental results to provide more details such as the spatio-temporal
concentration, site saturation, and temperature profiles.

The
mechanism of Di-Air system has been receiving great attention for its NOx
conversion enhancement at both low and high feed temperature only by increasing
the cycle frequency. The HC-intermediate pathway from which short-lived HC
intermediates generated is thought to play important role in this novel system.
The essence of this pathway is as follows: The intermediates form from reaction
between NOx and propylene and are able to survive through the rich period even at
high temperature and reduce NOx in the early lean period. Therefore, the goal
of this work is to investigate how this HC-intermediate pathway enhances the
NOx conversion.

By
feeding H₂ or C₃H₆ as the reductant agents by
same reductivity in stoichiometry, the NOx regeneration pathway and the HC NOx
regeneration pathway can be compared. Due to the different molecular
characteristics of H₂ and C₃H₆, the thermal
effect under a non-isothermal environment (aerobic rich feed condition) is
studied. The much larger molecular diffusivity of H₂ compared to C₃H₆
shortens the reaction zone. This is observed by measuring the temperature in the
front quarter and third quarter of the monolith reactor. Fig. 1a shows a
sharper temperature rise and larger axial temperature gradient when using H₂
than C₃H₆ under slow cycling. A 60 ^oC higher temperature
rise (250~280 ^oC for H₂) is measured in the first quarter
of the monolith under slow cycling when using H₂. This reduces the
NOx storage capacity during the early lean period and results in a 10 to 15 %
lower NOx conversion, as shown in Fig. 2.

Fast
cycling when using H₂ as the reductants has been thoroughly studied
in our previous work, exhibiting a better usage of storage sites and a near-steady-state
axial temperature profile comparing to slow cycling. However, under fast
cycling these is no difference when using H₂ and C₃H₆
in axial temperature profiles (Fig. 1b), so the previously mentioned thermal
effect doesn¡¦t exist under fast cycling. Only a 5 ¡V 10 % NOx conversion
enhancement can be found when using C₃H₆. This may be contributed
from the HC-intermediate mechanism.

In
addition to the thermal effect, steam reforming of propylene may also play an
important role in NOx regeneration pathways because it converts C₃H₆
to CO and H₂ through H₂O, and oxidation reactions in such
non-isothermal environment produces a lot of water.

Therefore,
experiments under near-isothermal condition (anaerobic rich feed) with no H₂O
in feed were conducted. Fig. 3 shows that the NOx conversions are very similar under
this near-isothermal condition at temperatures higher than the ignition temperature.
Under fast cycling the enhancement of NOx conversion using C₃H₆
is less than 5 %.

This
concludes that the NOx conversion enhancement of Di-air system mainly results
from the thermal effect, and HC-intermediate mechanism is relatively minor.

Reference:

[1] Y. Bisaiji, K. Yoshida, M. Inoue, N. Takagi, T.
Fukuma, Development of Di-Air  - A New Diesel deNOx System by Adsorbed
Intermediate Reductants, SAE Int. J. SAE Int. J. Fuels Lubr. 5 (2012) 380¡V388.

Fig. 1 The
temperature measured at first quarter (solid line) and third quarter (dashed
line) of the catalyst under 305 ^oC feed temperature for H₂
(red) and C₃H₆ case (black) under slow cycling (a, 60/10
s) and fast cycling (b, 6/1 s). Feed condition: Lean: 5 % O₂, Rich:
2.5% O₂, 9 % H₂/ 1 % C₃H₆, Both: 5%
CO₂.

Fig. 2 Cycle-averaged
NOx conversion with different cycle-averaged catalyst temperature, cycling
frequency using H₂ and C₃H₆ as the reductant
under non-isothermal condition. Feed condition: Both: 5 % CO₂, 500
ppm NO. Lean: 5 % O₂, Rich: 9 % H₂/ 1 % C₃H₆,
2.5 % O₂.

Fig. 3 Cycle-averaged
NOx conversion with different cycle-averaged catalyst temperature, cycling
frequency using H₂ and C₃H₆ as the reductant
under near-isothermal condition. Feed condition: Both: 5 % CO₂, 500
ppm NO. Lean: 5 % O₂, Rich: 4 % H₂/ 4445 ppm C₃H₆.