Article
`
`pubs.acs.org/IECR
`
`Integrated Selective Catalytic Reduction−Diesel Particulate Filter
`Aftertreatment: Insights into Pressure Drop, NOx Conversion, and
`Passive Soot Oxidation Behavior
`Kenneth G. Rappé*
`
`Pacific Northwest National Laboratory, P.O. Box 999, Richland, Washington 99354, United States
`
`ABSTRACT: Integrating urea-selective catalytic reduction (SCR) and diesel particulate filter (DPF) technologies into a single
`device has the potential to reduce the complexity of current diesel aftertreatment strategies. Fundamental studies were performed
`to shed light on the pressure drop and reaction behavior of integrated SCR and DPF systems. Details of SCR washcoat amount
`and location were investigated for effect on pressure drop during soot filtration. The SCR catalyst primarily impacted depth
`filtration of soot, promoted by increased catalyst located within the upstream portion of the porous filter wall. This effect is
`believed to be related to the nature of the porous filter substrate and pore network and changing of the rate at which pores plug
`in the presence of catalyst. SCR catalyst on the wall of the inlet filter channel also had an effect on the pressure rise during cake
`filtration of soot. NOx reduction efficiency measurements were performed to determine the nature and magnitude of the effect of
`soot on SCR performance. The effect of soot on the SCR performance is primarily attributed to the contribution of passive soot
`oxidation, and the propensity for soot oxidation to shift the NO2/NOx fraction relative to 0.5. SCR performance at NO2/NOx <
`0.5 is adversely affected by the presence of soot oxidation by increasing the SCR dependency on standard (NO only) SCR
`reactions; conversely, at NO2/NOx > 0.5, the SCR performance is positively impacted by a decreased dependency on NO2-only
`SCR reactions. Temperature-programmed oxidation studies were performed to evaluate the impact of SCR on passive soot
`oxidation. SCR adversely impacts soot oxidation performance via NO2 diffusive effects, decreasing NO2 concentration in the inlet
`channel. This impact can be minimized or recovered at higher NO2 concentration and NO2/NOx fractions >0.5.
`
`1. INTRODUCTION
`The diesel engine is currently the primary engine employed in
`the long- and short-haul commercial trucking industry because
`of its attractive wear characteristics, increased engine durability,
`and ability to deliver power efficiently under high load
`conditions. Manufacturers are planning much more widespread
`usage than historically employed, motivated by reduced fossil
`fuel consumption, reduced CO2 emissions, and the ability to
`operate on biologically derived fuels. However,
`for such a
`paradigm shift to take place, there are hurdles to overcome.
`Diesel close-coupled injection, compression, and ignition result
`in a heterogeneous two-phase burn that takes place largely at
`fuel droplet edges where the fuel and the air are mixing. The
`result of this is a near-stoichiometric burn that is very hot and
`comparatively long in duration, leading to elevated levels of
`NOx and PM formation. Engine controls and advanced fuel
`injection and combustion strategies have made significant
`strides in lowering diesel engine emission levels. However,
`persistently high particulate matter (PM) and nitrogen oxides
`(NOx) emissions dictate that exhaust aftertreatment will be a
`necessity.
`Exhaust aftertreatment reduces pollutants from the engine
`exhaust by treating the exhaust gas in a thermally and
`chemically controlled environment. Widespread introduction
`of advanced diesel exhaust aftertreatment
`is necessary for
`simultaneous compliance with NOx and particulate matter
`emission standards. Aftertreatment may lessen the trade-offs
`involved in controlling both NOx and PM emissions while
`minimizing losses in fuel efficiency.
`
`Several existing emission control technologies have proven
`effective at controlling emissions individually. A diesel oxidation
`catalyst (DOC) effectively oxidizes NO, CO, and hydrocarbon
`(HC) emissions, and the diesel particulate filter effectively
`removes particulate matter from the exhaust stream. The wall-
`flow diesel particulate filter
`is a honeycomb monolithic
`structure made from porous ceramic that has channels
`alternately plugged at the ends so that exhaust gas is forced
`through the porous channel wall. As engine exhaust is passed
`through a DPF, particulates are retained on the upstream
`portion of the filter wall and accumulate over time. Soot
`filtration first occurs by depth filtration, which is the collection
`of soot within the porous microstructure of the filter wall. This
`is typically accompanied by a sharp increase in the pressure
`drop across the filter. However, filtration will usually quickly
`transition to cake filtration, which is when soot is no longer
`accumulating within the wall microstructure, but rather, on the
`upstream surface of the filter wall. In this instance, the soot cake
`present on the upstream filter wall serves as the filter, and this is
`typically accompanied by a smaller increase in pressure drop
`across the filter. Together, these lead to an increase in pressure
`drop across DPF and, thus, an increase in the back pressure on
`the engine that adversely affects the engine operation and fuel
`consumption.
`
`Received:
`July 16, 2014
`Revised: October 9, 2014
`Accepted: October 20, 2014
`Published: October 20, 2014
`
`© 2014 American Chemical Society
`
`17547
`
`dx.doi.org/10.1021/ie502832f | Ind. Eng. Chem. Res. 2014, 53, 17547−17557
`
`BASF-2015.001
`
`

`
`Industrial & Engineering Chemistry Research
`
`Article
`
`HNCO
`
`(1)
`
`(2)
`
`Selective catalytic reduction (SCR) has been developed as an
`effective technology at reducing NOx emissions under lean
`burn conditions. SCR catalysts selectively reduce NOx to N2
`and H2O through the use of NH3 as the reductant. NH3 is
`supplied via a urea injection system and subsequent thermal
`decomposition to isocyanic acid and 1 equiv of NH3, as shown
`in eq 1. A second equivalent of NH3 is produced from the
`surface-catalyzed decomposition of isocyanic acid, as shown in
`eq 2.
`+
`→
`−
`−
`NH
`NH CO NH
`3
`2
`2
`+
`→
`+
`CO
`NH
`HNCO H O
`2
`3
`2
`The SCR process begins with reversible ammonia chem-
`isorption on catalytic Brønsted acid sites, where it becomes
`−SCR reactions
`available for SCR reaction. The main NH3
`promoted by Fe- and Cu-exchanged zeolites are shown in eqs
`3−5 and consist of the standard SCR reaction (NO only), the
`fast SCR reaction (equimolar NO and NO2), and the NO2-only
`SCR reaction, respectively. Zeolite-based catalysts (such as
`vanadium systems) exhibit maximum NOx reduction activity at
`NO2/NOx = 0.5 as a result of the prevalence of the fast SCR
`reaction. The standard SCR reaction participates at NO2/NOx
`< 0.5 because of the relative abundance of NO; similarly, the
`NO2-only SCR reaction participates at NO2/NOx > 0.5 because
`of the relative abundance of NO2.
`+
`+ →
`+
`4NH
`4NO O
`4N
`2
`3
`2
`+
`+
`→
`+
`NO NO
`2NH
`2N
`3
`2
`2
`+
`→
`+
`3NO
`3.5N
`6H O
`4NH
`3
`2
`2
`2
`Current focus in the U.S. has been on development of Fe-
`and Cu-based zeolite formulations as effective SCR catalysts for
`meeting NOx emission regulations.1,2 Early zeolite catalysts
`favored Fe-exchanged versions (Fe−ZSM-5) with medium-
`sized pores (∼5.5 Å) for heavy-duty diesel (HDD) applications
`for their improved activity above 350 °C and superior thermal
`stability as compared with Cu-exchanged versions, which had
`previously been confined to light-duty diesel (LDD)
`applications. Fe zeolites exhibit less sensitivity to NO2/NOx >
`0.5 because of the efficiency of the NO2-only SCR reaction on
`these catalysts;3,4 however,
`recent advancements
`in Cu-
`exchanged small pore zeolites have made them more much
`attractive for both HDD and LDD applications.5 Specifically,
`the small pore Cu-exchanged chabazite-type zeolites Cu-SSZ-
`13 and Cu-SAPO-34 have recently exhibited good activity over
`a broad temperature range, good thermal stability, and good
`resistance to hydrocarbon deactivation.6−10 These Cu zeolites
`are shown to exhibit less sensitivity to NO2/NOx < 0.511 and
`have also demonstrated improved SCR selectivity to N2 versus
`undesirable products, including N2O.7
`Integration of catalytic SCR and DPF technologies is a
`relatively new area in exhaust aftertreatment systems. Integrated
`reduction and soot
`SCR/DPF technology combines NOx
`filtration in a single 2-way device. Its development is motivated
`by emission compliance in a manner that reduces aftertreat-
`ment
`system volume and cost and increases packaging
`flexibility. As such,
`it is currently being pursued for both
`LDD and HDD diesel applications.
`Figure 1 shows a comparison of wall-flow and flow-through
`substrates as well as a schematic of an integrated wall-flow 2-
`−SCR
`way device.12 The technology consists of coating an NH3
`catalyst within the porous wall of a wall-flow filter (i.e., DPF).
`
`6H O
`2
`3H O
`2
`
`(3)
`
`(4)
`
`(5)
`
`Figure 1. Schematic comparison of a wall-flow DPF, flow-through
`SCR and integrated wall-flow 2-way DPF/SCR. Reproduced with
`permission from ref 12. Copyright 2011 Sage Publications.
`
`An integrated SCR/DPF device combines soot filtration with
`an SCR catalyst, achieving simultaneous soot filtration and NOx
`reduction within a single device.13
`For current conventional aftertreatment strategies, placement
`of SCR and DPF functionalities in the exhaust stream is
`governed by SCR activity and soot management. Regeneration
`of filtered soot requires both heat and oxidant; major oxidants
`in a diesel application are O2 and NO2, with NO2 oxidizing soot
`at lower temperature. When exhaust gas temperatures do not
`reach levels necessary for soot oxidation, active regeneration is
`necessary; this is facilitated by the injection of fuel in front of
`the DOC and subsequent oxidation, heating the downstream
`DPF via the ensuing exotherm to temperature necessary for
`soot oxidation. For HDD applications, exhaust temperature is
`comparatively high, and oxidation of soot with NO2 (i.e.,
`passive soot oxidation) is highly desirable. This motivates
`placement of the SCR catalyst downstream of the DPF to avoid
`depletion of NO2 prior to the DPF; however, SCR catalyst
`performance is adversely affected by the upstream DPF because
`of thermal inertia; Koltsakis and co-workers predicted as high as
`10% reduced NOx reduction efficiency due to an upstream
`DPF.14 For LDD applications,
`in which the exhaust gas
`temperature is comparatively low, resulting in low passive soot
`oxidation potential, the SCR catalyst is placed upstream of the
`DPF for improved activity and cold-start performance.
`Assuming adequate soot filtration efficiency is achievable
`with the wall-flow substrate employed, the requirements for
`successful SCR/DPF integration are maximizing NOx reduction
`efficiency while demonstrating acceptable pressure drop across
`the filter. Maximizing NOx reduction efficiency will be a
`function of total SCR catalyst washcoat volume and optimum
`catalyst coating techniques for maximum SCR efficiency.
`Demonstrating acceptable pressure drop across the filter will
`be a function of filter characteristics (e.g., total porosity),
`catalyst washcoat, and soot management. Thus, high porosity
`filter materials development
`is a technology facilitator for
`integrated SCR/DPF systems for integration of larger catalyst
`volumes while retaining an acceptable pressure drop.15−18 The
`washcoating technique is also a technology facilitator for similar
`reasons and, thus,
`is an area of continued optimization by
`catalyst manufacturers.
`focus for
`This leaves soot management as an area of
`successful deployment of integrated SCR/DPF technology. For
`LDD applications that must rely predominantly on active soot
`
`17548
`
`dx.doi.org/10.1021/ie502832f | Ind. Eng. Chem. Res. 2014, 53, 17547−17557
`
`BASF-2015.002
`
`

`
`Industrial & Engineering Chemistry Research
`
`regeneration, soot loading and time between regenerations
`must be closely controlled because of the exotherm from soot
`burn and possible adverse effects on SCR catalyst deactivation.
`In contrast,
`for HDD applications that rely on significant
`passive soot oxidation, SCR/DPF integration is more complex.
`Passive soot oxidation is NO2-dependent, and current emission
`control strategies employ oxidation functionality in DPFs for
`enhanced NO2 production and improved passive soot oxidation
`capacity. Integrated SCR/DPF systems will likely not possess
`oxidation functionality because of
`its adverse impact on
`reductant usage (i.e., NH3 oxidation).
`In addition, SCR
`processes will consume NO2 competitively with soot oxidation.
`Thus, the incorporation of SCR/DPF for HDD applications
`must be developed and controlled in such a manner that
`optimizes NO2 availability for passive soot oxidation to avoid
`(or minimize) the fuel penalty associated with the increased
`active regeneration frequency.
`The motivation for integration of DPF and SCR function-
`alities into a single device resides in both performance and
`reduced aftertreatment volume. The integration of DPF and
`SCR functionalities into a single device has the potential to
`reduce the number of aftertreatment “bricks” required, leading
`to reduced volume, reduced cost, and increased simplicity. With
`regard to performance, development of advanced wall-flow
`DPF substrates have allowed engine manufacturers to focus
`their attention on reducing the amount of NOx emitted from
`tailpipes. Integration of DPF and SCR provides a pathway for
`NOx emission reduction with greater flexibility. By integrating
`SCR and DPF functionalities, the adverse effect of upstream
`DPF thermal
`inertia on SCR catalyst performance can be
`largely mitigated,
`thereby improving the cold-start NOx
`performance.14 Increased SCR catalyst washcoat volumes are
`feasible through different configurations, which would provide
`increased NOx conversion capability that could facilitate fuel
`savings by allowing engines to be calibrated to higher engine-
`out NOx levels; however, there is uncertainty surrounding the
`effect of SCR catalyst on passive soot oxidation in integrated
`SCR/DPF solutions. The adverse impact of upstream SCR on
`DPF passive soot oxidation is well documented; the extent to
`which this can be minimized is a topic of significant current
`interest which this paper discusses.
`A number of early works have been performed to evaluate
`the feasibility of simultaneous NOx and PM aftertreatment with
`integrated SCR/DPF systems.19−23 These early works success-
`fully demonstrated the feasibility of combined NOx and PM
`reduction and identified key barriers to its effective deployment.
`Some of these key technology hurdles include
`(cid:129) Increased back pressure resulting from high SCR
`washcoat volumes
`(cid:129) Necessity of
`improved washcoating technology for
`maximizing SCR catalyst volume and gas contact
`(cid:129) Soot oxidation challenges
`Washcoat optimization is a topic to which many SCR/DPF
`investigators have alluded;24 however, little has been mentioned
`of the details of the SCR catalyst integration within the wall-
`flow filter. This work attempts to provide insight into the
`impact of SCR catalyst location on soot-loading characteristics
`of the filter. The intent is to improve the level of understanding
`for maximizing SCR catalyst washcoat volumes while
`minimizing the back pressure that results from soot collection.
`The impact of soot on SCR reactor performance has been
`reported by many with conflicting results in some cases. This
`
`Article
`
`work will report on the impact of soot on NOx conversion
`performance to develop a more fundamental understanding of
`the nature of
`the interaction.
`In addition, as previously
`mentioned, the adverse effect of SCR reaction on passive
`soot oxidation (with NO2) is a recognized barrier to successful
`SCR/DPF integration for HDD. However, detailed under-
`standing of the nature of this interaction and the primary
`reaction mechanisms that govern system performance are
`lacking. This work will report on the impact of SCR reaction on
`passive soot oxidation in an attempt
`to develop a more
`fundamental understanding of the primary reactive drivers that
`govern soot oxidation feasibility in the integrated system.
`
`2. EXPERIMENTAL METHODS
`High porosity cordierite wall-flow filters (i.e., DPFs; 25 mm o.d.
`× 40 mm long) were coated with SCR catalyst for the purpose
`of this investigation. The SCR catalyst employed was a Cu-
`chabazite catalyst loaded by an unnamed industrial partner to
`varying catalyst density. Upon receipt, samples were degreened
`under lean conditions at 10% O2, 5% H2O, balance N2.
`Soot loading and reactive interrogation of the SCR/DPF
`samples occurred separately in iterative fashion. The samples
`were loaded with soot employing the exhaust from a 2003 VW
`Jetta TDI following a DOC placed upstream of the filters.
`During soot loading, the exhaust was continually measured with
`a smoke meter to determine post-DOC particulate density. A
`slip stream of the exhaust was pulled through the SCR/DPF
`sample with the use of a vacuum pump, with the slip stream
`subsequently conditioned and measured for flow rate. Flow rate
`was electronically controlled to target 55 000 GHSV during
`loading; measured particulate density and flow rate were
`combined to target a desired total soot loading, typically 4 g/L.
`Pressure drop was measured at ports just upstream and
`downstream of
`the SCR/DPF sample holding apparatus.
`Subsequent to soot loading, the samples were removed and
`weighed at elevated temperature for redundant measurement of
`soot mass uptake. Estimated soot loading from smoke meter
`and mass measurements were in good agreement with typically
`less than 10% variance.
`Reactive measurements were performed on a dedicated test
`bench designed to accommodate the soot-loaded filter and
`housing. The simulated exhaust composition, consisting of 9%
`O2, 9% CO2, 8% H2O, 500−1500 ppm of NOx (NO and NO2),
`and balance N2, was blended together to simulate post-DOC
`heavy-duty lean burn diesel exhaust. For measurements
`including SCR reaction, NH3 was included at 500−1500 ppm.
`Dry gases were supplied by electronically controlled MKS
`mass flow controllers. Air, N2, and CO2 were combined as dry
`gases and preheated to the desired temperature with an
`electronically controlled tube furnace. Simultaneous
`to
`preheating, the feed stream was humidified to the desired
`level with an electronically controlled water vaporizer employ-
`ing a N2 sweep flow. Following preheating, NO, NO2, and NH3
`were supplied separately as dry gas feeds to the exhaust flow in
`desired quantities. The full simulated exhaust was then fed to
`the reactor assembly consisting of a quartz tube containing the
`SCR/DPF sample held in place by high temperature alumina
`paper. An electronically controlled tube furnace controlled the
`reactor temperature to the desired level. Temperature measure-
`ments were made employing Watlow type K thermocouples
`placed just upstream and downstream of the DPF/SCR sample.
`Pressure drop measurement was made via ports located just
`upstream and downstream of the sample holder. Gas analysis
`
`17549
`
`dx.doi.org/10.1021/ie502832f | Ind. Eng. Chem. Res. 2014, 53, 17547−17557
`
`BASF-2015.003
`
`

`
`Article
`
`alternating ends on the filter, channels in opposing corners of
`each subfigure in Figure 2 are common to a respective end of
`the filter; direction of exhaust flow through the samples is
`reversed by simply rotating the filter axially with respect to flow.
`With the SCR/DPF sample loaded to 60 g/L catalyst density
`(B), SCR catalyst is deposited solely within the filter wall
`microstructure and not on the channel walls. Catalyst is located
`heavily to one side of the filter wall and penetrating, on average,
`40−60% across the width of the wall. In the SCR/DPF sample
`loaded to 90 g/L catalyst density (C), SCR catalyst is similarly
`deposited predominantly within the filter microstructure and
`also heavily to one side of the filter wall. Periodic occurrences
`of the catalyst penetrating the full width of the porous filter wall
`are evident, and with little or no catalyst present on either set of
`the channel walls. Comparatively in the SCR/DPF sample
`loaded to 150 g/L catalyst density (D), there are notable
`differences to the previous two samples. First, catalyst of
`varying amounts is present on one set of channel walls, outside
`of
`the wall microstructure. And second,
`the SCR catalyst
`consistently penetrates across the full width of the filter wall;
`thus, catalyst is in intimate contact with both filter walls. It is
`worth noting that the SEM imaging was conducted at four
`locations within the coated samples, with the images shown
`above selected randomly from the samples analyzed. The
`results discussed in the text with regard to catalyst location
`were consistently reflected in all samples analyzed.
`Figure 3 shows the corresponding results of Hg porosimetry
`analysis of the DPF and SCR/DPF samples, presented as
`
`Industrial & Engineering Chemistry Research
`
`was accomplished with a Nicolet 6700 FTIR employing a metal
`gas cell heated to 190 °C. Sample gas transfer lines to the FTIR
`were heated to 200 °C to avoid NH4NO3 formation. FTIR gas
`analysis was conducted at 100 Torr through the use of a
`vacuum pump located downstream from the FTIR gas cell and
`an electronically controlled MKS pressure controller. Redun-
`dant NOx measurement was made via a heated California
`Analytics chemiluminescent NOx meter. Control of the entire
`assembly, including mass flow controllers, all heated zones, and
`data acquisition, was facilitated by a custom designed Visual
`Basic control system.
`Bench-scale reactive measurements were performed in
`pseudosteady state fashion by employing slow thermal ramping,
`∼2 °C/min, from ∼200 to 550 °C. Reactive measurements
`included SCR performance for NOx conversion efficiency and
`temperature-programmed oxidation (TPO) for soot oxidation
`characterization.
`Scanning electron microscopy (SEM) was performed with a
`JEOL 5900 scanning electron microscope. The sample was cut
`perpendicular to the cordierite channels with a wet diamond
`wafer saw, potted in LR white resin, and spun at 2000 rpm for
`10 min in a standard centrifuge and subsequently evacuated to
`remove gases from the filter and resin. Following polymer-
`ization and hardening of the resin, the sample was cross-
`sectioned perpendicular
`to the channel axis, ground and
`polished with a series of diamond grits (25, 9, 3, and 1 μm),
`and final polished with 0.05 μm γ-alumina.
`Pore characteristics (porosity, pore size distribution) were
`measured via Hg intrusion employing a Micromeritics
`AutoPore IV 9500 series mercury porosimeter.
`
`3. RESULTS AND DISCUSSION
`SEM imaging and porosity characteristics were compared with
`soot collection characteristics in an attempt to determine the
`combined effect of catalyst loading and location on filtration
`behavior. Figure 2 shows the SEM imaging of the uncoated
`DPF substrate (A) and SCR/DPF samples coated with ∼60
`(B), ∼90 (C), and ∼150 g/L (D) SCR catalyst loading density.
`In each of the subfigures, using A as a reference, black is
`background or void space, (off-)white is cordierite substrate,
`and gray is SCR catalyst. Since channels are plugged at
`
`Figure 3. Hg porosimetry analysis of DPF (no catalyst) and SCR/DPF
`samples loaded to 60, 90, and 150 g/L SCR catalyst.
`
`incremental pore volume (mL/g) versus pore diameter (mm).
`The uncoated DPF measured 63% porosity. The addition of 60
`g/L of SCR catalyst decreased the total porosity to 56%, with
`similar peak pore diameters for both the DPF and 60 g/L SCR/
`DPF sample. This, coupled with SEM imaging, provides good
`indication that the 60 g/L of SCR catalyst was depositing solely
`within the filter wall microstructure. Similar peak pore
`diameters support the observation that a large portion of the
`porous filter wall remained largely void of catalyst.
`With an additional 30 g/L catalyst (i.e., 90 g/L total SCR
`catalyst), the total porosity decreased to 52% with a slight shift
`in the peak pore diameter to a smaller maximum. The shift in
`the peak pore diameter suggests that the less porous filter wall
`is void of catalyst versus the 60 g/L sample. These results
`
`17550
`
`dx.doi.org/10.1021/ie502832f | Ind. Eng. Chem. Res. 2014, 53, 17547−17557
`
`Figure 2. SEM imaging of clean DPF (A) and SCR/DPF samples
`coated with 60 (B), 90 (C), and 150 (D) g/L SCR catalyst loading.
`
`BASF-2015.004
`
`

`
`Industrial & Engineering Chemistry Research
`
`Article
`
`Figure 4. Soot-loading characteristics of 90 (left) and 150 g/L (right) SCR/DPF samples configured such that catalyst was present predominantly
`on the upstream portion of the filter microstructure and on the inlet channel wall (for the 150 g/L sample).
`
`Figure 5. Soot-loading characteristics of 90 g/L (left) and 150 g/L (right) SCR/DPF samples configured such that catalyst was present
`predominantly on the downstream portion of the filter microstructure and on the outlet channel wall (150 g/L sample).
`
`coupled with SEM indicate that 90 g/L of SCR catalyst
`continued to deposit predominantly within the filter wall
`microstructure.
`With an additional 60 g/L catalyst (i.e., 150 g/L total SCR
`catalyst), the total porosity did not decrease significantly but,
`rather, remained fairly constant at 51%, with no change in the
`peak pore diameter (versus 90 g/L). The SCR catalyst filled a
`small number of very large pores >20 mm, leading to a slight
`asymmetric narrowing of the pore distribution. One difference
`from the previous two samples is that a significant amount of
`catalyst did not penetrate the wall microstructure but, rather,
`remained coated on the filter channel.
`In summary, total porosity and peak pore characteristics,
`coupled with SEM imaging, support the notion that up to 90 g/
`L SCR catalyst was loaded into the filter wall microstructure,
`penetrating the width of the porous wall to varying degrees but
`loaded predominantly heavy to one side of the filter wall.
`Comparatively, the results suggest that >90 g/L SCR catalyst
`was largely deposited as a coating on the filter channel, with the
`catalyst deposited typically across the full width of the porous
`filter wall.
`3.1. Soot-Loading Behavior. The 90 and 150 g/L SCR/
`DPF samples were loaded with soot to determine the effect of
`catalyst location on the soot loading characteristics of the
`samples. Figures 4 and 5 show on-engine backpressure (i.e.,
`pressure drop) resulting from the 90 g/L sample (left) and the
`150 g/L sample (right) as they are loaded with soot. In Figure
`4, samples are configured with the catalyst on the upstream
`
`portion of the filter (in close proximity to the collected soot).
`Figure 5 shows analogous backpressure measurements with the
`samples configured in the opposite direction (i.e., catalyst on
`the downstream portion of the filter). The data presented here
`were very reproducible, with subsequent pressure drop (versus
`collected soot) traces in close agreement.
`The pressure drop of the clean SCR/DPF samples was
`comparatively low: 0.35 and 0.44 kPa for 90 and 150 g/L SCR
`catalyst loading, respectively, under the conditions tested. Thus,
`collected soot affected filter permeability much more
`significantly than the catalyst washcoat; however, both the
`catalyst concentration and location impacted the magnitude of
`pressure drop resulting from the collection of soot. The most
`prominent effect occurred during depth filtration of soot; this is
`not surprising, given the knowledge that the catalyst is largely
`located within the wall microstructure. With the catalyst on the
`upstream portion of
`the filter (Figure 4),
`the system
`demonstrated 11.5 and 20.5 kPa back pressure for the 90 and
`150 g/L SCR catalyst densities,
`respectively, under
`the
`conditions tested when it transitioned from depth to cake
`filtration mode. Comparatively, with the catalyst on the
`downstream portion of
`the filter (Figure 5),
`the system
`demonstrated 3.2 and 7.4 kPa back pressure for the 90 and 150
`g/L SCR catalyst densities, respectively, under the conditions
`tested when it transitioned from depth to cake filtration mode.
`The pressure drop during depth filtration is affected by the
`amount of SCR catalyst located on the upstream portion of the
`filter wall microstructure. The magnitude of pressure drop does
`
`17551
`
`dx.doi.org/10.1021/ie502832f | Ind. Eng. Chem. Res. 2014, 53, 17547−17557
`
`BASF-2015.005
`
`

`
`Industrial & Engineering Chemistry Research
`
`Article
`
`Figure 6. SCR/DPF NOx reduction efficiency at 35 000 GHSV, NH3/NOx = 1, NO2 = 250 ppm, and varying NO2/NOx fraction; NO2/NOx = 0.5
`and less shown on the left, NO2/NOx = 0.5 and greater shown on the right.
`
`not correlate solely with the mass of soot collected, but appears
`to also be a function of SCR catalyst amount and location. Two
`feasible contributing factors to this would be (1) relative
`density, and thus permeability, of soot deposits, or (2) relative
`impact on the flow field in the porous network. Evaluating the
`first in more detail,
`if back pressure is governed (solely or
`predominantly) by the density of soot deposits collected during
`depth filtration, then one would expect a closer quantitative
`relationship between the total mass of soot collected, the
`amount of SCR catalyst present on the upstream portion of the
`filter wall, and the magnitude of resulting back pressure.
`Although there appears to be a very general trend, it is not
`consistent; neither is it quantitative. For example,
`in both
`figures, the 90 and 150 g/L catalyst samples filtered similar
`masses of soot at the transition from depth to cake filtration but
`exhibited distinctively different back pressure as a result.
`These results suggest that SCR catalyst on the upstream
`portion of the filter wall directly affects the impact of soot on
`the porous network flow field. This brings to light the concept
`of pore throats. Pore throats are constriction points in the
`three-dimensional channels of wall-flow monolith pores, and
`they play a significant role in permeability.25 They are physically
`manifested a number of ways in the wall microstructure,
`ranging from simple constriction points along a pore channel to
`a directional change in the flow field passing through the
`porous network. Soot collected in depth filtration primarily
`impedes exhaust flow at pore throats26 and results in increased
`pressure drop during depth filtration; the extent is governed by
`the nature and density of those pore throats. Because the
`difference in clean pressure drop between the samples (0.35
`versus 0.44 kPa) was small compared with the effect of soot, the
`density of pore throats is not expected to be significantly
`different. Thus, it appears that the nature of a large fraction of
`pore throats (and their surrounding volumes) is affected by the
`presence of SCR catalyst. The result is an increased extent of
`plugging as a result depth filtration of soot in regions of
`elevated SCR catalyst concentration.
`Comparing the 150 g/L catalyst sample with the 90 g/L
`sample, the samples collected similar masses of soot during
`depth filtration in both configurations; however, the difference
`in back pressure between the samples was significantly different,
`depending on flow configuration. With the catalyst on the
`downstream portion of the filter wall, the 150 g/L sample
`
`measured 4.2 kPa greater back pressure versus the 90 g/L
`sample. This can be attributed to more SCR catalyst located
`across the full width of the filter wall for the 150 g/L sample
`and in close contact with collected soot near the upstream
`filter-channel wall. However, with the catalyst on the upstream
`portion of the filter wall, the 150 g/L sample exhibited 9.0 kPa
`greater back pressure compared with the 90 g/L sample. This
`can largely be attributed to significantly more catalyst present
`on the inlet channel wall for the 150 g/L sample versus the 90
`g/L sample. This suggests that catalyst on the inlet channel wall
`has further impact on the extent of plugging during depth
`filtration and is hypothesized to be a magnification of the pore
`throat effect in the proximity of the upstream channel-wall
`interface.
`Once the system transitioned to cake filtration, additional
`discrepancies are apparent. Without catalyst on the inlet filter
`channel walls (Figure 4, left, and Figure 5, left and right), cake
`filtration behavior was similar in all three samples, exhibiting a
`pressure-drop rise in the range of 1.5−1.7 kPa/(g/L soot)
`under the conditions tested. However, with significant SCR
`catalyst on the inlet filter channel walls (Figure 4, right), cake
`filtration exhibited a pressure-drop increase of 3.4 kPa/(g/L
`soot). This is double the mass-specific effect of cake-filtered
`soot on pressure drop relative to the case with little or no
`catalyst present on the inlet channel.
`It
`is evident
`that
`permeability of the filter during cake filtration is adversely
`affected by the presence of catalyst on the inlet channel wall.
`Contributing to this effect is believed to be a combination of
`reduced open volume in the inlet channel and effective
`filtration area on the upstream channel wall surface, combined
`with the effect of the catalyst on flow dynamics at the wall and
`the