Client Name: Coordinating Research Council
`CRC Project No.: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`
`LITERATURE SURVEY TO ASSESS THE STATE-OF-THE-
`ART OF SELECTIVE CATALYTIC REDUCTION OF VEHICLE
`NOx EMISSIONS
`
`
`CRC Project No. AVFL-7
`
`Ricardo Inc.
`7850 Grant Street
`Burr Ridge, Illinois, 60527
`TEL: (630) 789-0003
`FAX: (630) 789-0127
`
`Coordinating Research Council, Inc.
`3650 Mansell Road, Suite 140
`Alpharetta, Georgia, 30022
`
`F0529
`C02 -2814
`21 JUN 2002
`
`Brad Adelman
`
`Renaud Rohe
`Roland Christopher
`
`
`
`Prepared for:
`
`
`
`
`
`Project:
`Archive:
`Date:
`
`
`
`Author:
`
`Collaborators:
`
`
`
`
`
`
`
`
`
`
`
`Technical Approval:
`
`
`
` Dr. Brad Adelman, Ph. D.
`
`
`
` Technical Specialist / Performance And Emissions
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`
`21 JUNE 2002
`
`
`
`BASF-2004.001
`
`

`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`
`
`Final Report for Project AVFL-7
`
`LITERATURE SURVEY TO ASSESS THE STATE-OF-THE-ART OF SELECTIVE CATALYTIC
`REDUCTION OF VEHICLE NOx EMISSIONS
`
`
`
`EXECUTIVE SUMMARY
`
`
`
`Ricardo has been commissioned by Coordinating Research Council to perform a
`comprehensive search of several scientific databases in order to assess the state-of-the-art of
`selective catalytic reduction of NOx using hydrocarbons as the reductant source. The objective
`is to identify potential catalyst formulations which show promise as emission control
`technologies to be used in LDD applications. To this end, Ricardo Powerlink™, Compendex,
`INSPEC, NTIS, CAB Abstracts and CHEMWEB – Catalysis Forum have been searched using
`the following key words: Selective Catalytic Reduction, SCR, NOx Reduction, Diesel Exhaust,
`Lean, DeNOx, Non urea/ammonia. From this search, 289 papers have been identified from
`which 122 have been selected for detailed reading. A catalyst formulation is defined as a
`potential ECT if the peak NOx conversion occurs below 300°C and the conversion levels are
`greater than 70%. The subsequent reading lead to the following observations.
`
`Most of the literature available for SCR-HC omits key exhaust gas components, namely
`water vapor and SO2. The most frequently sited catalyst, Cu/ZSM-5, is irreversibly deactivated
`upon extended exposure to water vapor and SO2. Most of the ion-exchanged zeolite supported
`catalysts are not potential ECTs as they either suffer from one or more of the following:
`water/SO2 inhibition, inferior conversion levels, elevated conversion temperature window. An
`Fe/ZSM-5 catalyst synthesized via a solid-gas exchange between FeCl3 and H/ZSM-5 yields a
`very active catalyst which is water/SO2 tolerant. The limitation of this formulation is the difficult
`procedure required for synthesis.
`
`Supported platinum catalysts on many materials (e.g., Al2O3, SiO2, zeolites, mixed metal
`oxides) demonstrate NOx conversion levels within the desired temperature range. NOx
`conversion levels vary from 30-95% between 200°C to 300°C. Conversion depends on metal
`loading, reductant employed, support material and presence of other metals. Zeolites and
`alumina are generally selective supports when the platinum loading is 1wt%. Light paraffins are
`nonselective reductants while light olefins are selective. Light hydrocarbons yield the highest
`N2O:N2 ratio of all reductants. Heavier hydrocarbons as well as oxygenates favor N2 formation
`though most reductants still form some quantity of N2O. Supported platinum catalysts are
`tolerant to water vapor and SO2 in the exhaust gases though the oxidation of SO2 to SO3 occurs
`simultaneously with NOx reduction. A catalyst technology for the concomitant removal of N2O is
`required in order for supported platinum catalysts to achieve greater potential.
`
`Alumina supported silver catalysts are active for NOx reduction over a large temperature
`window. Selective reductants include higher paraffins and oxygenated organics (excluding
`methanol). These catalysts are resistant to water inhibition and SO2 though the effects of SO2
`are less documented. Catalyst preparation methods include incipient wetness and hydrolysis of
`aluminum alkoxide; i.e., silver salt gels with the hydrolysis method allowing improved dispersion
`at higher silver loadings. A method for producing the oxygenates on-board is required to avoid
`the need for an additional tank.
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 2 of 53
`
`
`
`
`
`BASF-2004.002
`
`

`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`Recently, non-thermal plasmas have received much attention. This technology converts
`
`NO to NO2 and partially oxidizes olefins to aldehydes. When a catalyst such as Ba/Y zeolite is
`downstream of a dielectric barrier discharge devise, significant NOx removal is observed at
`200°C. This technology is tolerant to water vapor and SO2; moreover, the oxidation of SO2 to
`SO3 does not occur. There are limitations to NTP. The catalysts tend to create significant
`quantities of surface deposits at low temperatures and it is possible that after extended low
`temperature operation the catalysts could become deactivated. The DBD devise requires
`energy input (20-60 J/L) which, in turn, will lower fuel economy. The durability of the DBD
`devises remains to be proven. Toxic byproducts such as CH2O and HCN are produced within
`the NTP plumes. This requires the use of an oxidation catalyst which in turn has the potential to
`create N2O.
`
`Despite the large body of data for SCR-HC catalysts, very few formulations possess
`adequate NOx conversion levels at temperatures observed with LDD applications. Moreover,
`most of the formulations tested have been done so in the absence of water vapor and SO2.
`When these components of real diesel exhausts are added to the feed gases, many
`formulations which had been proven active are irreversibly deactivated. Cu/ZSM-5 is a classic
`example of this. Supported platinum catalysts are active but have a propensity to form N2O.
`Fe/ZSM-5 is very active but its synthesis method is complex. Alumina supported silver catalysts
`hold promise. These formulation can be augmented by the use of non-thermal plasma. Plasmas
`are also quite active with Ba/Y zeolite and Na/Y zeolite catalysts.
`Suggested areas of future research:
`•
`All future research endeavors to include realistic diesel exhaust conditions:
`water vapor and SO2
`Pt-based formulations which do not form N2O
`Novel catalyst formulations which decompose/reduce N2O below 300°C
`Facile synthesis routes to form Fe/ZSM-5 with equivalent performance and
`durability as those formed by the solid-gas or anaerobic aqueous exchange
`of FeC2O4
`On-board routes to form oxygenated reductants for silver-based catalysts
`Continued investigation into non-thermal plasma technologies
`Engine thermal management techniques to minimize exhaust conditions
`which are below 180°C – maintain catalyst within peak operating temperature
`window
`Techniques for storing NOx emissions during cool exhaust conditions
`followed by re-injection of the stored NOx when the ECT has achieved light-
`off conditions.
`
`•
`•
`•
`
`•
`•
`•
`
`•
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 3 of 53
`
`
`
`
`
`BASF-2004.003
`
`

`
`Page
`
`2
`
`4
`
`8
`
`8
`9
`
`10
`
`11
`
`11
`15
`17
`17
`17
`18
`
`19
`
`19
`20
`21
`
`21
`
`21
`22
`22
`
`23
`
`24
`
`24
`
`
`Page 4 of 53
`
`
`
`
`
`TABLE OF CONTENTS
`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`
`
`EXECUTIVE SUMMARY
`
`TABLE OF CONTENTS
`
`INTRODUCTION
`
`1.1 Alternate Technologies To TWC
`1.2 Requirements For Light-Duty Diesel (LDD) Applications
`
`
`
`LITERATURE IDENTIFICATION
`
`
`
`ROLE OF THE SUPPORT MATERIAL
`
`3.1 Zeolite-Based Catalysts
`3.2 Alumina-Based Catalysts
`3.3 Mixed-Metal Oxide (MMO) Based Catalysts
`3.4 Reaction Mechanisms
`3.4.1 Organic Surface Intermediates Pathway
`3.4.2 NOx Decomposition Pathway
`
`
`
`ACTIVE METALS – SUMMARY
`
`4.1 Supported Platinum Catalysts
`4.2 Supported Silver Catalysts
`4.3 Supported Copper Catalysts
`
`
`
`EFFECT OF THE REDUCTANT – SUMMARY
`
`5.1 Light Olefins And Paraffins
`5.2 Methane
`5.3 Oxygenated Reductants
`
`
`
`POTENTIAL OF NON-THERMAL PLASMA (NTP)
`
`
`
`ENGINE PARAMETERS AND OTHER VARIABLES
`
`
`
`COMPARISON TO SELECTIVE CATALYTIC REDUCTION WITH NH3
`21 JUNE 2002
`© 2002 by Ricardo, Inc
`
`
`
`
`
`1.
`
`2.
`
`3.
`
`4.
`
`5.
`
`6.
`
`7.
`
`8.
`
`BASF-2004.004
`
`

`
`25
`
`28
`
`38
`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`
`
`SUMMARY AND PROPOSAL FOR FUTURE WORK
`
`
`
`APPENDICES
`
`
`
`GRAPHS, FIGURES AND TABLES
`
`9.
`
`
`
`
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 5 of 53
`
`
`
`
`
`BASF-2004.005
`
`

`
`LIST OF FIGURES
`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`
`
`
`
`Figure 1: Simulated Thermal Trace versus Engine to Mass Ratio
`
`Figure 2: Peak NOx Conversion versus Temperature for ZSM-5 Supported Catalysts
`
`Figure 2A: Peak NOx Conversion versus Temperature for ZSM-5 Supported Pt Catalysts
`
`Figure 3: Peak NOx Conversion versus Temperature for Other Zeolite Supported Catalysts 40
`
`Figure 4: Peak NOx Conversion versus Temperature for Alumina Supported Catalysts
`
`Figure 4A: Peak NOx Conversion versus Temperature for Alumina Supported Pt Catalysts 41
`
`Figure 4B: Peak NOx Conversion versus Temperature for Alumina Supported Ag Catalysts 41
`
`Figure 5: Peak NOx Conversion versus Temperature for Mixed-Metal Oxide Supported Catalysts
`
`Figure 6: Peak NOx Conversion versus Temperature for Supported Platinum Catalysts
`
`Figure 7: Peak NOx Conversion versus Temperature for Supported Silver Catalysts
`
`Figure 8: Peak NOx Conversion versus Temperature with Methane as the Reductant
`
`Figure 9: Peak NOx Conversion versus Temperature with Oxygenated Reductants
`
`Page
`38
`
`39
`
`39
`
`40
`
`42
`
`42
`
`43
`
`43
`
`44
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 6 of 53
`
`
`
`
`
`BASF-2004.006
`
`

`
`PAGE
`
`45
`
`46
`
`47
`
`48
`
`49
`
`50
`
`51
`
`52
`
`53
`
`LIST OF TABLES
`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`
`
`
`
`
`Table 1: NOx Conversion on ZSM-5 Supported Catalysts
`
`Table 2: NOx Conversion on Zeolite (non ZSM-5) Supported Catalysts
`
`Table 3: NOx Conversion on Alumina Supported Catalysts
`
`Table 4: NOx Conversion on Mixed-Metal Oxide Supported Catalysts
`
`Table 5: NOx Conversion on Supported Platinum Catalysts
`
`Table 6: NOx Conversion on Supported Silver Catalysts
`
`Table 7: NOx Conversion with Methane as the Reductant
`
`Table 8: NOx Conversion with Oxygenated Reductants
`
`Table 9: References for the Data Points
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 7 of 53
`
`
`
`
`
`BASF-2004.007
`
`

`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`INTRODUCTION:
`
`
`
`1.
`
`
`
`As the effects of anthropogenic emission on the environment are becoming better
`
`understood, governments have acted to minimize this impact. Current legislation imposes limits
`on the emissions from internal combustion engines. These limits have been set for the release
`of carbon oxides (CO, directly, and CO2, indirectly by fuel economy limits), hydrocarbons and
`nitrogen oxides (NO and NO2 referred to as NOx). For diesel engines, the mass of particulate
`matter is also legislated. Sulfur emissions, though not directly legislated, are limited so that H2S
`is not formed as it has a strong, pungent odor. The formation of sulfates and SO3 is also limited
`as they readily adsorb onto the particulate matter and thus impart an increase in total particulate
`mass. Nitrous oxide is not currently a regulated emission. It is, however, a potent greenhouse
`gas and may become a regulated compound if vehicle emissions of N2O are deemed to be
`significant. Of all these emissions, the most difficult to remove from vehicle emissions is NOx.
`This task becomes even more difficult for compression ignition vehicles (diesel) than spark
`ignition (gasoline).
`
`For several decades, spark Ignition (SI) engines have employed three-way catalytic
`converters (TWC). The technology is called “three-way” since hydrocarbons and carbon
`monoxide are oxidized to CO2 and H2O while NOx is simultaneously reduced to N2. These
`catalysts have been proved to yield high conversion efficiencies at temperatures above 250°C
`provided that the exhaust gas is at or near the stoichiometric point (enough oxygen present to
`combust all reductants to CO2 and H2O). As lambda (a measure of stoichiometry) deviates from
`stoichiometry (λ = 1), the efficiency of the TWC rapidly decreases. Decreasing lambda (exhaust
`gas becomes net rich) permits significant hydrocarbon and CO emissions as not enough oxygen
`is present for combustion. Increasing lambda (exhaust gas becomes net lean) permits
`significant NOx emissions as the catalyst sites responsible for NOx reduction are saturated with
`oxygen preventing dissociative adsorption of NOx, the first step in NOx reduction on a TWC. In
`order to maintain exhaust gas streams at stoichiometry, engine control units monitor the
`exhaust gas lambda values and constantly adjust fueling and air flow to the engine. Moreover,
`TWCs also employ an oxygen storage component (OSC) to the catalyst formulation. The OSC
`is capable of buffering the exhaust gas composition. When excess O2 is present, the OSC
`stores oxygen. When excess reductants are present, the OSC releases oxygen.
`
`Unfortunately for compression ignition (CI) technologies such as diesel, it is not possible
`to buffer the exhaust gas so that the catalyst technology operates near the stoichiometric point.
`This would require an excessive amount of fuel and thus remove a primary advantage of CI
`engines; i.e., significantly improved fuel economy. Therefore, CI engines require an alternate
`technology.
`
`1.1 Alternate Technologies To TWC
`
`
`Several technologies have demonstrated promise. Lean NOx Traps (LNT) remove the
`NOx from the gas stream and store them as metal nitrates such as barium nitrate. Once the
`barium sites for NOx adsorption are saturated, these catalytic sites must be regenerated by
`converting the exhaust stream from net lean to net rich. During the regeneration cycle, the
`nitrate is reduced and NO is released. The resultant NO is further reduced to N2 over a metal
`site (preferably Rh). By limiting the duration and frequency of the rich regeneration, the fuel
`penalty caused by creating the rich gas stream can be minimized. This tends to result in large
`catalyst volumes which can be several times greater than the cylinder displacement of the
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 8 of 53
`
`
`
`
`
`BASF-2004.008
`
`

`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`engine. Another drawback to LNT is that they are sulfur intolerant. Since the NOx adsorption
`reaction is an acid-base reaction, materials capable of NOx adsorption also demonstrate
`significant sulfur uptake in the form of surface sulfate groups. These sulfate groups are
`significantly more stable than the adsorbed nitrates. While the sites present as nitrates can
`readily be regenerated (< 400°C), those present as sulfates require higher temperatures
`(>550°C) and longer regeneration events (minutes).
`
`Selective Catalytic Reduction of NOx with ammonia (SCR-NH3) is currently employed as
`a NOx removal technology for stationary power sources. SCH-NH3 is based on the reaction of
`NOx from the engine reacting with injected NH3 to produce N2 and H2O. The injected quantity of
`NH3 must be controlled to prevent excess ammonia from being released into the environment.
`This is called “NH3-slip”. Contrarily, insufficient NH3 will result in NOx release into the
`environment. The use of a ‘clean-up’ catalyst downstream of the SCR-NH3 catalyst prevents
`NH3-slip by oxidizing any excess NH3, though this yields NOx. Conversion efficiencies for SCR-
`NH3 have been reported to be above 90%. In order to apply SCR-NH3 to mobile sources, NH3
`must be stored on the vehicle or produced onboard. Owing to the health and safety issues of
`storing NH3, it will need to be produced onboard. Since NH3 is not a significant engine emission,
`the source for producing NH3 will need to be stored in the vehicle. There are two main
`compounds that can be used for NH3 formation: urea and ammonium carbamate. Urea, present
`as a ∼35% wt solution, readily hydrolyzes to form NH3 and H2O at temperatures above 180°C.
`There are a few limitations to using urea as the NH3 source. One is that 65% of the urea
`solution is water and therefore, does not add to performance. This becomes wasted mass.
`There currently is no infrastructure present in North America for urea distribution. Moreover,
`since the urea tank will require frequent refilling, no mechanism is in place on the vehicle which
`would ensure that the tank never becomes empty. A concern among government regulator is
`that without this precaution it would be possible to defeat the SCR-NH3 by not refilling the urea
`tank. Added to this are concerns about the various byproducts (cyanuric acid, ammelide,
`ammeline, etc.) formed during urea pyrolysis. Ammonium carbamate, present as a solid, has
`cost, distribution and handling issues which makes it an unattractive source for producing
`ammonia on board.
`Selective Catalytic Reduction of NOx with hydrocarbons (SCR-HC) has the advantage
`that the catalyst utilizes hydrocarbon species present in the exhaust stream for NOx reduction.
`The reaction can be supplemented with additional hydrocarbons either via secondary injection
`of fuel in the cylinder or direct injection into the exhaust stream. The former has the advantage
`that diesel fuel will partially combust to yield lighter hydrocarbons which tend to have
`performance benefits over raw diesel fuel injected over the catalyst. SCR-HC, also known as
`active deNOx, requires minimal hardware and uses the diesel fuel as the reductant. This avoids
`the need for an additional reagent tank. As a result of its simplicity, SCR-HC would be the
`desired technology to attain future legislated NOx emission limits.
`
`1.2 Requirement For Light-Duty Diesel (LDD) Applications
`
`
`Within the publicly available information, there exists a large body of data concerning the
`performance of SCR-HC technologies. The breadth of catalyst formulations, test operating
`conditions and reductant used make it difficult to compare these data and assess the potential
`for SCR-HC. It is the purpose of this report to assess SCR-HC as a feasible emission control
`technology (ECT) for light-duty applications. The legislated emission target is US TIER 2 Bin 5
`with a NOx emission limit of 0.07g NOx/mi. Catalyst activity over the FTP cycle directly affects
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 9 of 53
`
`
`
`
`
`BASF-2004.009
`
`

`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`the engine out NOx limits. For example, a catalyst which converts 90% of NOx emissions during
`the FTP cycle can tolerate engine out levels of 0.70 g NOx/mi. and still achieve the Bin 5 target
`of 0.07g/mi. A catalyst with 50% efficiency requires engine out NOx to be no greater than 0.14
`g/mi. For the purpose of this discussion a conversion efficiency target of 70% has been used.
`This corresponds to engine out NOx levels no greater than 0.23 g/mi. Based on literature
`values, these engine out targets are somewhat aggressive but achievable.
`The actual temperature trace for the FTP cycle is engine dependant and varies with the
`relationship between engine size and vehicle mass. The larger the engine is with respect to the
`vehicle, the lower the inlet temperature to the catalyst will be. Figure 1 gives a simulated
`temperature trace for two engine/mass scenarios: the higher engine-to-mass ratio yields the
`lower thermal trace. Based on the two thermal traces, effective SCR-HC catalysts for LDD
`application would need to have peak performance between 200-300°C. Catalysts that
`demonstrate significant NOx reduction capabilities only at higher temperatures are not likely to
`be effective, as exhaust gases for LDD applications rarely exceed 350°C. These traces
`demonstrate that engine parameters will play a large role in achieving the legislated emission
`targets.
`Additionally, the amount of hydrocarbon required for effective (>70% removal) will need
`to be low in order to minimize the fuel penalty. For the purpose of this study, an attempt to
`correlate hydrocarbon used to an exact fuel penalty has not been performed. It is assumed that
`the lower the C/NOx ratio (normalized value of the hydrocarbon based on equivalent C1 units
`versus the quantity of NOx) the lower the fuel penalty. None of the data points were excluded
`based on an elevated C/NOx ratio.
`
`LITERATURE IDENTIFICATION
`The selection of the papers presented in this study was performed as follows.
`
`Six different databases, which cover the wide range of the Selective Catalytic Reduction
`domains, were chosen. Their particularities are described in the Appendix. A search was
`conducted using the following keywords:
`• Selective Catalytic Reduction (SCR)
`• NOx Reduction
`• Diesel exhaust
`Lean
`•
`• DeNOx
`• Non urea/ammonia
`A pre-selection, based on the article title and abstract, has been applied to eliminate the papers
`which were not related to the subject. From this search, 289 papers have been selected.
`
`The abstract papers were studied and the papers were classified on an excel database. The
`classification fields were
`• Type of papers (General, Laboratory test, Modeling, Review…)
`• Catalyst
`• Reductant
`• Application
`Interest for this study
`•
`
`122 papers were ordered for detailed reading. These papers are listed in the Appendix .
`
`2
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 10 of 53
`
`
`
`
`
`BASF-2004.010
`
`

`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`ROLE OF THE SUPPORT MATERIAL
`
`The bulk of this study will be divided into three sections: the role of the support, the role
`of the active metal center and the role of the reductant. In each section the representative data
`will be presented and discussed with respect to the section topic. The first topic is the role of the
`support material. Several support materials have been shown to be active supports for SCR-
`HC. The most prominently studied supports are zeolites. Zeolites are crystalline arrays of
`oxygen-bridged silica tetrahedra. When some of the silica (Si4+) centers are replaced with
`alumina (Al3+) the result is a net negative charge to the framework of the zeolite. The negative
`charge is compensated with a cation. Upon synthesis the cation is generally sodium or
`ammonium. This cation can readily be exchanged with other cations, thus imparting many
`opportunities for new catalyst formulations. The first section will list some of the more actively
`studied zeolite-based formulations along with those zeolite-based catalysts which show the
`greatest potential for application in LDD emission control technologies.
`
`In addition to zeolites, amorphous oxide supports have also been demonstrated to be
`active supports. The best of these is alumina. Amorphous alumina has high surface area which
`is important in dispersing active metal sites. Moreover, it has greater hydrothermal stability than
`zeolites. Alumina can also possess acidic or basic properties based on the preparation method.
`For most of the references, the alumina supports were provided from an industrial source. As a
`result, little characterization data are presented. Attempts will be made to identify differences in
`performance of similar catalyst formulations based on the synthesis method of the support
`material or catalyst.
`
`Similar to alumina, several other oxide supports have been proven active supports for
`SCR-HC. These supports can be mixed metal oxides (MMO) such as pillared clays or oxides
`such as silica and zirconia. Catalyst preparation with these types of supports can include
`incipient wetness where the minimum amount of water is used in order to cover the support
`surface. This tends to form bulk oxides. As this preparation employs dilute metal solutions, it
`limits the metal loading achievable. Co-precipitation is another catalyst preparation method.
`Unlike incipient wetness, co-precipitation uses a solution containing the support material mixed
`and the active metal. The solution pH is gradually increased with a combustible base such as
`urea or ammonia. As the pH rises, the solution forms a gel. Upon calcination, the resultant
`powder will posses highly dispersed active metal cites.
`When possible, these differences will be assessed with respect to catalytic performance.
`
`
`Zeolite-Based Catalysts
`
`The first SCR-HC catalyst of significant interest is Cu/ZSM-5. In independent studies
`Iwamoto and Held demonstrated enhanced NOx reduction of copper ion-exchanged zeolites
`when an excess of oxygen is present. NOx reduction begins around 300°C and reaches a
`maximum between 400-450°C. The shape of the conversion graphs have been shown to be
`dependant on reductant type and concentration as well as on space velocity. The experimental
`conditions did not include H2O or SO2 which are ubiquitous in real diesel exhausts. When these
`components are included in the feed, NOx reduction is inhibited. For short duration, the
`inhibiting effect of H2O is reversible while extended exposure to water vapor irreversibly
`deactivates Cu/ZSM-5 catalysts. This effect is accelerated at elevated temperatures. The
`Cu/ZSM-5 system remains a popular catalyst for academic studies and continues to elucidate
`new features of heterogeneous catalysis. It will, however, never meet the rigors for industrial
`application in LDD emission control technologies.
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 11 of 53
`
`
`
`
`
`3
`
`3.1
`
`
`BASF-2004.011
`
`

`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`An almost Edisonial approach has been employed to screen potential catalyst
`formulations. Figures 2 and 3 show a very small sample of the various exchange metals tested
`for SCR-HC activity. Since ZSM-5 has been the most actively studied support, those
`formulations using ZSM-5 have been presented in a separate graph (Figure 2) in an attempt to
`create a more facile comparison. Though the number of points is small, a good representative
`sample has been presented.
`As evidenced in Figure 2 most of the catalysts possess peak NOx conversion either at
`temperatures which are higher than expected for LDD applications or at conversion levels too
`low to meet future legislative standards. There are four samples which possess sufficient
`conversion at the proper temperatures. Two of the catalysts are supported platinum catalysts
`and the other two are supported cerium catalysts. Another set of six catalysts posses sufficient
`conversion but the peak temperatures are slightly higher than desired. Of these, four are
`supported copper catalysts. The other two are supported iron catalysts, one which is promoted
`by lanthanum. Since the focus of this report is to assess the potential of various catalyst
`formulations as emission control technologies on LDD applications, all of the other samples will
`not be considered except in the case when these inferior catalysts are of similar composition to
`those which have displayed adequate performance. It will be critical to reconcile these
`differences in order to fully understand the potential of the adequate technologies.
`Of all the references selected for this project, the most frequent formulation studied is
`Cu/ZSM-5. These results can be summarized as follows. Supported copper catalysts have high
`NOx conversion levels usually at temperatures above 350°C when the reductant is C3 or higher
`hydrocarbon. For ethene, moderate conversion occurs between 200-300°C though it is not
`significant enough (~40%).It is interesting to note that for lighter hydrocarbons (C6 or lower), the
`presence of water vapor leads to significant catalytic inhibition while heavier hydrocarbons such
`as C8 show moderate water vapor tolerance though peak NOx conversion occurs at 350°C with
`normal octane and 500°C with iso-octane. Most of the differences in Cu/ZSM-5 performance
`can be traced to the hydrocarbon employed. Though heavier hydrocarbons maintain catalytic
`activity over Cu/ZSM-5 even in the presence of water vapor, the formulation on the whole
`possesses flaws which prevent its use in LDD applications. Extended hydrothermal exposure
`leads to irreversible loss of catalytic activity. The presence of SO2 also leads to permanent
`deactivation. Though it may be possible to avoid SO2 with the use of synthetic diesel (Fischer
`Tropsch), water and heat present in diesel exhaust are enough to deactivate Cu/ZSM-5
`eventually.
`As has been readily discussed in the literature, Cu/ZSM-5 remains as a reference
`catalyst even though there is no potential for this formulation to have industrial success.
` There are seven conversion levels reported for supported Pt catalysts in Figure 2. Of
`the seven, only two are deemed suitable with respect to conversion level and temperature of
`maximum conversion. All of these catalysts are represented in Figure 2A and are summarized
`as follows. Supported platinum catalysts have been shown to be ineffective when ethane is
`used as the reductant. At high space velocities, the conversion peak is diminished and shifted to
`higher temperatures. Ion-exchanged samples are less effective than bulk impregnated samples.
`This difference implies that larger platinum ensembles are required for effective NOx removal
`and not intimate interaction between Pt and the zeolite pores. Of the seven samples, only the
`ion exchanged ones were tested in the presence of water vapor or SO2 but the same
`generalization can be applied to all Pt/ZSM-5 catalysts. There is only a mild inhibition (<5%)
`from the presence of water vapor. In the presence of SO2 the conversion window is shifted
`slightly to higher temperatures (+25°C) and is also slightly broader (+50°C). Therefore, under
`
`21 JUNE 2002
`
`
`
`© 2002 by Ricardo, Inc
`
`
`
`
`Page 12 of 53
`
`
`
`
`
`BASF-2004.012
`
`

`
`
`Client Name: Coordinating Research Council
`CRC Project No: AVFL-7
`Project No.:
`F0529
`Archive:
`C02-2814
`
`
`
`
`realistic conditions, Pt/ZSM-5 has a high likelihood of meeting the future emission legislation.
`Unfortunately, Pt/ZSM-5 possesses a serious limitation. For all of the catalysts, a significant
`fraction of NOx reduced resulted in the formation of N2O. Though N2O is not currently a
`regulated emission, it is a strong greenhouse gas. The use of an emission control technology
`which produces N2O in significant quantities would not be welcomed and would most likely
`prompt government legislation. In order for Pt/ZSM-5 to achieve industrial interest, the discovery
`of a Pt formulation that does not form N2O or catalyst that removes N2O concomitantly with NOx
`reduction is required. Currently no such catalyst exists though Rh on ZSM-5 or Al2O3 show
`promise for N2O decomposition, attaining light-off between 250°C and 300°C.
`The remaining two catalyst formulations of interest are supported iron catalysts. The iron
`only parent catalyst shows 75% NOx conversion at 350°C when iso-butane is the reductant. It is
`remarkable to note that the presence of 10% water vapor does not inhibit conversion. By adding
`0.9wt% lanthanum to the parent catalyst, the conversion increases to 83% and has a larger
`conversion window though shifted to higher temperatures. The promotional effect of adding
`0.9wt% lanthanum is not inhibited by the presence of water vapor. For most of the supported
`iron catalysts, water vapor actually increases catalytic performance at lower temperatures. This
`presumably is a result in removal of coke precursors which would otherwise block catalytically
`active sites. Unlike the Pt/ZSM-5 catalysts, Fe/ZSM-5 yields N2 only. Moreover, after 100 hours
`time-on-stream, conversion only decreases by 10%. This effect is reversible by removing water
`vapor from the f