(736a) Experimental and Modeling Study of Passive NOx Adsorption: Pd-Exchanged-ZSM-5

Authors: 
Ambast, M., University of Houston
Harold, M., University of Houston
Karinshak, K., University of Houston

Experimental
and Modeling Study of Passive NOx Adsorption: Pd-Exchanged-ZSM-5

 

Mugdha Ambast, Kyle
Karinshak, and Michael P. Harold

Dept. of Chemical
and Biomolecular Engineering, University of Houston,

 Houston, TX
77204-4004, USA

mharold@uh.edu; mugdhaambast@gmail.com

 

Introduction

With the announcement by the EPA to
impose more stringent regulations for emissions of nitrogen oxides (NOx) with a
target-year implementation of 2024, reduction of NOx emissions has become a
great challenge for the vehicle industry [1]. Advanced technologies
like selective catalytic reduction (SCR) and NOx storage and reduction (NSR)
have been implemented for NOx reduction in diesel and lean gasoline emission
control. However, their effectiveness at temperatures below 2000C is
limited. Thus, NOx emissions during the vehicle cold start complicate the
meeting of the aforementioned emission rules. The passive NOx adsorber (PNA) have
been gaining attention recently as a technology that has the potential to abate
cold-start NOx emissions. Catalysts used in PNAs are generally noble metals [2][3]
like Pd or Pt supported by rare earth oxides and zeolites such as ZSM-5, BEA
and SSZ-13. These materials adsorb NOand NO2 at low
temperatures and release the species at high temperatures, enabling their
reduction by downstream SCR.

Here we report a combined
experimental and modeling study to understand and predict the effects of
various operating parameters on a model PNA material 1%Pd-ZSM-5. Specifically,
the effects of temperature, flowrate, O2 and H2O are
investigated with the objective to develop a predictive reactor model
containing mechanistic based kinetics.

Catalyst
preparation and Experimental setup

Two different catalysts were used for the experiments: 1% Pd-ZSM-5 and HZSM-5. The
incipient wet impregnation technique was used for the Pd containing materials. A
cordierite monolith of 400 cpsi and 4cm in length was washcoated with the
catalyst solution by dip-coating method until the desired loading was achieved.
The experiments were conducted in a bench flow reactor system. The washcoated
monolith was contained in a quartz tube contained within a
temperature-controlled furnace. Experiments were carried out at space velocities
in the range 14-23k hr-1. The catalysts were pretreated in 5% O2
and balance Argon at 5500C for 30 minutes and then cooled to the
adsorption temperature (50 0C, 80 0C or 150 0C).
The experiment involved exposure to NO (400ppm)/ O2 (2%)/Ar at the
prescribed temperature and duration, followed by exposure to O2
(2%)/Ar, under a temperature ramp to 5500C at 200C/min. For
experiments with water, 7% water was added throughout the experiment.
Downstream of the reactor a FTIR measured the concentration of effluent gases.
A syringe pump and heated injection syringe system was used to add water into
the feed.

Results
and Discussions


Experiments were
carried out with the two samples over a range of feed temperatures and
flowrates. Selected results are highlighted here. The amount of NOx adsorbed
decreased with temperature (Figure1). The NOx uptake decreased with increasing
temperature, suggesting kinetic limitations. When water was added, the amount
of NOx adsorbed further decreased as a result of site competition. Both NO and
NO2 appeared during breakthrough. The desorption data for NO and NO2
revealed the existence of multiple site types on both the zeolite (in absence
of H2O) and exchanged Pd and PdO. Experiments done with different
concentrations of O2 indicated its importance in amount of NOx
adsorbed as well as generation of NO2.

A continuous stirred-tank adsorber (CSTA) model was developed to predict the
TPD profiles of NO and NO2. The model considers that adsorption of NO
and NO2 on 1%Pd-ZSM-5 occurs on two different types of sites corresponding
to zeolite and Pd(Figure2). Estimation of the kinetic parameters was done by
fitting the experimental TPD data with a trust-region-reflective scheme. The steps
used for modeling the TPD curves are shown in table below.

S.NO.

Steps

1

Estimation of total site concentration by integrating area under the TPD curve.

2

Coupled differential and algebraic equations were solved by MATLAB routine ODE23s

3

Trust-region-reflective least squares algorithm of MATLAB was coupled with the ODE file to estimate kinetic parameters

A more detailed 1+1 D, two-phase
transient monolith model with a more complex and mechanistically-consistent
reaction scheme was developed. This model enables an evaluation of the
underlying mechanism, impact of transport processes, washcoat thickness, among
many other parameters. The parameter estimates of selected steps are utilized
in the tuning of the monolith model. The combined experimental and modeling
study is useful in identifying rate limiting processes and improving catalysts
and operating strategies.

Significance


Current automobile
after-treatment system designs do not meet the forecast 2024 NMOG regulations.
So, further experimental and computational research is necessary for PNA as it may
have the potential to significantly reduce NOx concentration during cold start
emissions.

References

[1]
Environmental Protection Agency, “MEMORANDUM IN RESPONSE TO PETITION FOR
RULEMAKING TO ADOPT ULTRA-LOW-NOX STANDARDS FOR ON-HIGHWAY HEAVY DUTY TRUCKS
AND ENGINES,’ EPA, Washington DC, 2016.

[2]
J. Thesis, “An assessment of Pt and Pd model catalysts for low temperature NOx
adsorption,” Catalysis Today, vol.267, pp. 93-109, 2016.

[3]
C. Descorme, P. Gelin, M. Primet and C. Lecuyer, “Infrared study of nitrogen
monoxide adsorption on palladium ion-exchanged ZSM-5 catalysts,” Catalysis
Letters,
vol.41, pp.

133-138,
1996.