(465f) Impact of Hydrocarbon Trapping on Temperature Programmed Oxidation   over a Pt/Pd/BEA Monolith Catalyst

Impact of hydrocarbon trapping on temperature
programmed oxidation over a Pt/Pd/BEA monolith catalyst

Po-Yu Peng, Michael P. Harold*, Dan Luss**

Department of
Chemical and Biomolecular Engineering, University of Houston, Houston, TX 77204

*mharold@uh.edu, **dluss@uh.edu

The function of the diesel
oxidation catalyst (DOC) comprising a washcoat of Platinum (Pt) and Palladium
(Pd) (PGM = precious group metals) is to catalyze the oxidation of diesel
effluent mixture of CO and hydrocarbons (HCs). A DOC operates at very high CO
and HC conversion (>90%) after a sufficiently high operating temperature has
been reached. However, during the “cold start” a fraction of the exhaust
hydrocarbons may escape unreacted [1, 2]. A large-pore zeolite such as Beta
(BEA) is added to the DOC to trap diesel exhaust hydrocarbons during the cold
start. High molecular weight hydrocarbons are especially prone to trapping in
BEA-modified PGM DOCs. Although extensive research has been conducted on the
BEA storage capacity, there is a need to gain insight on how BEA affects the
performance of HC light-off (LO) and of the HC oxidation transient performance [3-7].
We conducted experiments to gain a deeper understanding of functioning of the
HC trap and to enable improved trap design and operating strategy. The insight gained
from this study will help optimize the multi-functional and component catalyst
formulation, reactor design and operating strategy for achieving high
conversion of multiple hydrocarbon species.

In a typical experiment dodecane (C₁₂)
is trapped on a 0.8 in. diameter by 3-in. long mixed washcoated Pt-Pd/BEA/Al₂O₃
400 cpsi cordierite monolith. The catalyst contains 60 g/ft³ PGM in
a Pt:Pd ratio of 2:1 by weight (~1:1 atomic ratio) on a 0.7 g/in³
load of γ-Al₂O₃ support and incorporated with the
BEA content (0 and 0.5 g/in³) into the washcoat. The C₁₂
and CO₂ gas concentrations were measured by spatially resolved mass
spectrometry (SpaciMS, QMS-Omnistar GSD 300) and a FTIR. Spatial resolved
temperature profiles were measured by a coherent optical frequency domain
reflectometer (c-OFDR with a 1.26 m long gold-coated optical fiber placed in the
center of monolith channel. The trapping experiments were conducted by flowing
160 ppm C₁₂, 10% O₂ and the balance Ar at 60 ^oC
until the catalyst was saturated. This was followed by ~10 ^oC/min
inlet temperature ramp. The total gas flowrate was 8 L/min (GHSV= 12,438 h^-1).
During the dynamic hysteresis analysis, the saturated catalyst was heated up by
1 ^oC/min from 110 ^oC to 170 ^oC inlet
temperature and then cooled by -1 ^oC/min from 170 ^oC to
110 ^oC. The total gas flowrate varies from 6 L/min (GHSV= 10,262 h^-1)
and 8 L/min (GHSV= 12,438 h^-1). Adiabatic behavior was confirmed by
measuring temperature profiles under non-reaction conditions.

Unusual CO₂ formation
spikes were observed during the oxidation of HC that was pre-stored on the
catalyst before exposure to the temperature ramp. Fig. 1 shows the effluent C₁₂
concentrations along with the effluent generated CO₂ concentration
during a linear feed temperature ramp. The C₁₂ is completely trapped
by C₁₂ breakthrough with a slow approach to the feed concentration
value at uniform temperature of ~60 ^oC. However, during the
subsequent temperature ramp, in addition to the release of C₁₂ at
the 48 minute mark, two impressive spikes of CO₂ appeared between 50
and 57 minutes. Corresponding to the first two spikes are measureable spikes in
the outlet temperature.

The c-OFDR measurements enabled
construction of a 3-D contour plot of temperature versus time and position.
(Figure 2) The temperature spikes correlate with the measured CO₂
yield spikes. The feed temperature increased linearly during the ramp. The
local temperature sequentially increased and decreased three times at a nearly
fixed position in the monolith. During the first and second cooling period, the
temperature went back close to the inlet temperature. This behavior suggests that
the oxidation stops abrubtly as the generated heat is quickly removed by the
flowing gas stream. The drop in the CO₂ concentration between the
first and second peaks in Fig. 1 also supports the stoppage of the HC oxidation.
A mass balance analysis indicates that the fed C₁₂ during this time period
is released and re-trapped by the catalyst. The C₁₂ continues to accumulate
until that the catalyst saturates. This intermittent HC oxidation behavior
continues until the HC trapping capacity becomes smaller at high temperature. The
rapid propagation of the maximum temperature from the midstream to the
downstream during the first peak also supports high heat convection by the gas
flow. This rapid temperature peak propagation greatly accelerates the HC
oxidation downstream.

A slow heating/cooling rate leads
to a significant counter-clockwise hysteresis behavior, underscoring the competition
between HC trapping and oxidation. Figure 3 shows the C₁₂ and CO₂
concentrations during the hysteresis under the slow heating/cooling rate of ±1 ^oC/min.
In the ramp-up period, the C₁₂ (CO₂) gradually decreases
(increases), indicating a delicate balance between trapping and oxidation.
After ignition at ~159 ^oC, more impressive peaks occur. The
occurrence of multiple peaks suggests that the C₁₂ trapping occurs
before the C₁₂ oxidation. After a certain amount of C₁₂ is
trapped, the C₁₂ oxidation reaction then triggers the release of stored-C₁₂.
The HC trapping capability becomes smaller at high temperatures. During the
ramp-down period, a high C₁₂ conversion remains until the feed temperature
is below ~136 ^oC. The reactor extinguishes at ~23 ^oC below
the ignition, showing a strong ability to reduce HC emission when the engine
stalls.

The non-isothermal nature of this adiabatic
system suggests that HC trapping and oxidation dominates in the different
regions in the catalyst. Since the downstream is much hotter than the upstream,
the HC is re-trapped by BEA at the upstream and the HC is drastically oxidizes at
the downstream. Both events greatly consume the C₁₂, helping to
reduce C₁₂ emission, and lead to a much lower extinction
temperature. Even at such low heating/cooling rate of ±1 ^oC/min, a significant
counter-clockwise hysteresis behavior still exists. This study is the first to
provide insight into the spatiotemporal features of transient behavior of HC
trapping and oxidation of BEA/DOC catalysts, and also the first discovery of
multiple HC LO phenomena.

Figure 1. Dodecane and CO₂ concentration measured during
adsorption followed by TPO with BEA-zeolite added
Pt/Pd/Al₂O₃ monolith catalyst. Condition during
saturation: 160 ppm dodecane and 10% O₂ balanced with Ar. Condition
during desorption and TPO: 10% O₂ balanced with Ar.

Figure 2. The 3-D plot of the spatially-resolved temperature versus
time in feed of 10% O₂ balanced with Ar. (Catalyst: D-05)

Figure 3. Experimental C₁₂ and CO₂
concentration versus inlet fluid temperature in a dynamic hysteresis analysis. (C₁₂
= 160 ppm, 10% O₂; space velocity of 12,438 h^-1;
temperature ramp = ±1 ^oC/min).

References

1.            Mukai,
K., et al., Adsorption and desorption characteristics of the adsorber to
control the HC emission from a gasoline engine. 2004, SAE Technical Paper.

2.            AL‐Harbi, M., et al., Competitive no,
co and hydrocarbon oxidation reactions over a diesel oxidation catalyst.
The Canadian Journal of Chemical Engineering, 2012. 90(6): p. 1527-1538.

3.            Li, H.-X., et al., Application of zeolites
as hydrocarbon traps in automotive emission controls. Studies in Surface
Science and Catalysis, 2005. 158: p. 1375-1382.

4.            Chang, H., et al., Gasoline Cold Start
Concept (gCSC™) Technology for Low Temperature Emission Control. SAE Int.
J. Fuels Lubr, 2014. 7(2): p. 480-488.

5.            Murakami, K., et al., Development of a high
performance catalyzed hydrocarbon trap using Ag-zeolite. 2004, SAE
Technical Paper.

6.            Elangovan, S., et al., A comparative study
of zeolites SSZ-33 and MCM-68 for hydrocarbon trap applications.
Microporous and mesoporous materials, 2006. 96(1): p. 210-215.

7.            Elangovan, S., et al., Silicoaluminophosphate
molecular sieves as a hydrocarbon trap. Applied Catalysis B: Environmental,
2005. 57(1): p. 31-36.