WO2010097634A1

WO2010097634A1 - Filter for filtering particulate matter from exhaust gas emitted from a positive ignition engine

Info

Publication number: WO2010097634A1
Application number: PCT/GB2010/050334
Authority: WO
Inventors: Louise Clare Arnold; Robert James Brisley; David Robert Greenwell; Christopher Gough Morgan
Original assignee: Johnson Matthey PLC
Current assignee: Johnson Matthey PLC
Priority date: 2009-02-26
Filing date: 2010-02-26
Publication date: 2010-09-02
Anticipated expiration: 2011-08-26
Also published as: GB2485260A; KR101899919B1; BRPI1013401B1; GB2497440B; EP3320964B1; EP2401056B1; CN105642116A; GB201118038D0; RU2527462C2; US20100239478A1; GB2497441A; GB2497442B; KR102072505B1; US8211393B2; GB2468210B; BRPI1013395A2; GB2497442A; GB201301893D0; DE102010002425B4; JP2012518537A

Abstract

A filter for filtering particulate matter (PM) from exhaust gas emitted from a positive ignition engine, which filter comprising a porous substrate having inlet surfaces and outlet surfaces, wherein the inlet surfaces are separated from the outlet surfaces by a porous structure containing pores of a first mean pore size, wherein the porous substrate is coated with a wash coat comprising a plurality of solid particles wherein the porous structure of the wash coated porous substrate contains pores of a second mean pore size, and wherein the second mean pore size is less than the first mean pore size.

Description

FILTER FOR FILTERING PARTICULATE MATTER FROM EXHAUST GAS EMITTED FROM A POSITIVE IGNITION ENGINE

The present invention relates to a filter for use in treating particulate matter (PM) in exhaust gas derived from any combustion process where it is not possible to remove PM from the exhaust gas by build-up of PM (so-called "cake filtration") or by a combination of depth filtration and cake filtration. The combustion process is typically that of a vehicular engine. In particular, the invention relates to a filter for use in treating PM derived from a vehicular positive ignition engine, particularly stoichiometrically operated positive ignition engines but also lean- burn positive ignition engines.

Positive ignition engines cause combustion of a hydrocarbon and air mixture using spark ignition. Contrastingly, compression ignition engines cause combustion of a hydrocarbon by injecting the hydrocarbon into compressed air. Positive ignition engines can be fuelled by gasoline fuel, gasoline fuel blended with oxygenates including methanol and/or ethanol, liquid petroleum gas or compressed natural gas.

Ambient PM is divided by most authors into the following categories based on their aerodynamic diameter (the aerodynamic diameter is defined as the diameter of a 1 g/cm³ density sphere of the same settling velocity in air as the measured particle):

(i) PM-IO - particles of an aerodynamic diameter of less than 10 μm;

(ii) Fine particles of diameters below 2.5 μm (PM-2.5);

(iii) Ultrafme particles of diameters below 0.1 μm (or 100 nm); and

(iv) Nanoparticles, characterised by diameters of less than 50 nm.

Since the mid- 1990 's, particle size distributions of particulates exhausted from internal combustion engines have received increasing attention due to possible adverse health effects of fine and ultrafme particles. Concentrations of PM-10 particulates in ambient air are regulated by law in the USA. A new, additional ambient air quality standard for PM-2.5 was introduced in the USA in 1997 as a result of health studies that indicated a strong correlation between human mortality and the concentration of fine particles below 2.5 μm.

Interest has now shifted towards nanoparticles generated by diesel and gasoline engines because they are understood to penetrate more deeply into human lungs than particulates of

AA 1864 PCT greater size and consequently they are believed to be more harmful than larger particles, extrapolated from the findings of studies into particulates in the 2.5-10.0 μm range.

Size distributions of diesel particulates have a well-established bimodal character that correspond to the particle nucleation and agglomeration mechanisms, with the corresponding particle types referred to as the nuclei mode and the accumulation mode respectively (see Figure

1). As can be seen from Figure 1, in the nuclei mode, diesel PM is composed of numerous small particles holding very little mass. Nearly all diesel particulates have sizes of significantly less than 1 μm, i.e. they comprise a mixture of fine, i.e. falling under the 1997 US law, ultrafme and nanoparticles.

Nuclei mode particles are believed to be composed mostly of volatile condensates (hydrocarbons, sulfuric acid, nitric acid etc.) and contain little solid material, such as ash and carbon. Accumulation mode particles are understood to comprise solids (carbon, metallic ash etc.) intermixed with condensates and adsorbed material (heavy hydrocarbons, sulfur species, nitrogen oxide derivatives etc.) Coarse mode particles are not believed to be generated in the diesel combustion process and may be formed through mechanisms such as deposition and subsequent re-entrainment of particulate material from the walls of an engine cylinder, exhaust system, or the particulate sampling system. The relationship between these modes is shown in Figure 1.

The composition of nucleating particles may change with engine operating conditions, environmental condition (particularly temperature and humidity), dilution and sampling system conditions. Laboratory work and theory have shown that most of the nuclei mode formation and growth occur in the low dilution ratio range. In this range, gas to particle conversion of volatile particle precursors, like heavy hydrocarbons and sulfuric acid, leads to simultaneous nucleation and growth of the nuclei mode and adsorption onto existing particles in the accumulation mode. Laboratory tests (see e.g. SAE 980525 and SAE 2001-01-0201) have shown that nuclei mode formation increases strongly with decreasing air dilution temperature but there is conflicting evidence on whether humidity has an influence.

Generally, low temperature, low dilution ratios, high humidity and long residence times favour nanoparticles formation and growth. Studies have shown that nanoparticles consist

AA 1864 PCT mainly of volatile material like heavy hydrocarbons and sulfuric acid with evidence of solid fraction only at very high loads.

Contrastingly, engine-out size distributions of gasoline particulates in steady state operation show a unimodal distribution with a peak of about 60-80nm (see e.g. Figure 4 in SAE 1999-01-3530). By comparison with diesel size distribution, gasoline PM is predominantly ultrafine with negligible accumulation and coarse mode.

Particulate collection of diesel particulates in a diesel particulate filter is based on the principle of separating gas-borne particulates from the gas phase using a porous barrier. Diesel filters can be defined as deep-bed filters and/or surface-type filters. In deep-bed filters, the mean pore size of filter media is bigger than the mean diameter of collected particles. The particles are deposited on the media through a combination of depth filtration mechanisms, including diffusional deposition (Brownian motion), inertial deposition (impaction) and flow-line interception (Brownian motion or inertia).

In surface-type filters, the pore diameter of the filter media is less than the diameter of the PM, so PM is separated by sieving. Separation is done by a build-up of collected diesel PM itself, which build-up is commonly referred to as "filtration cake" and the process as "cake filtration".

It is understood that diesel particulate filters, such as ceramic wallflow monoliths, may work through a combination of depth and surface filtration: a filtration cake develops at higher soot loads when the depth filtration capacity is saturated and a particulate layer starts covering the filtration surface. Depth filtration is characterized by somewhat lower filtration efficiency and lower pressure drop than the cake filtration.

WO 03/011437 discloses a gasoline engine having an exhaust system comprising means for trapping PM from the exhaust gas and a catalyst for catalysing the oxidation of the PM by carbon dioxide and/or water in the exhaust gas, which catalyst comprising a supported alkali metal. The means for trapping PM is suitable for trapping PM of particle range 10-100 nm, and can be a wallflow filter made from a ceramic material of appropriate pore size such as cordierite coated with the catalyst, a metal oxide foam supporting the catalyst, a wire mesh, a diesel

AA 1864 PCT wallflow filter designed for diesel applications, an electrophoretic trap or a thermophoretic trap (see e.g. GB-A-2350804).

WO 2008/136232 Al discloses a honeycomb filter having a cell wall composed of a porous cell wall base material and, provided on its inflow side only or on its inflow and outflow sides, a surface layer and satisfying the following requirements (1) to (5) is used as a diesel particulate filter: (1) the peak pore diameter of the surface layer is identical with or smaller than the average pore diameter of the cell wall base material, and the porosity of the surface layer is larger than that of the cell wall base material; (2) with respect to the surface layer, the peak pore diameter is from 0.3 to less than 20 μm, and the porosity is from 60 to less than 95% (measured by mercury penetration method); (3) the thickness (Ll) of the surface layer is from 0.5 to less than 30% of the thickness (L2) of the cell wall; (4) the mass of the surface layer per filtration area is from 0.01 to less than 6 mg/cm²; and (5) with respect to the cell wall base material, the average pore diameter is from 10 to less than 60 μm, and the porosity is from 40 to less than 65%. See also SAE paper no. 2009-01-0292.

Other techniques suggested in the art for separating gasoline PM from the gas phase include vortex recovery.

Emission legislation in Europe from 1^st September 2014 (Euro 6) requires control of the number of particles emitted from both diesel and gasoline (positive ignition) passenger cars. For gasoline EU light duty vehicles the allowable limits are: 1000mg/km carbon monoxide; 60mg/km nitrogen oxides (NO_x); lOOmg/km total hydrocarbons (of which < 68mg/km are non- methane hydrocarbons); and 4.5mg/km particulate matter ((PM) for direct injection engines only). Although the authorities have not set the PM number standard for Euro 6 yet, it is widely understood that it will be set at 6.0 x 10¹¹ per km. The present specification is based on the assumption that this number will be adopted in due course.

In the United States, no similar emission standards have been set. However, the State of California Air Resources Board (CARB) recently published a paper entitled "Preliminary

Discussion Paper - Amendments to California's Low-Emission Vehicle [LEV] Regulations for Criteria Pollutants - LEV III" (release date 8^th February 2010) in which a new PM standard of between 2 and 4 mg PM/mile (1.25-2.50 mg PM/km (currently lOmg PM/mile (6.25 mg PM/km))) is proposed, the paper commenting that: "Staff has received input from a number of

AA 1864 PCT manufacturers suggesting that a standard of 3 mg PM/mile (1.88 mg PM/km) can be met for gasoline direct injection engines without requiring the use of particulate filters." Additionally, the paper states that since the PM mass and count emissions appear to be correlated: "Although a mandatory number standard is not being considered at this time, an optional PM number standard of about 10¹² particles/mile [6.25 ⁿ particles/km] is being considered (which could be chosen by manufacturers instead of the PM mass standard)". However, since neither the PM standard nor the PM number standard has been set by CARB yet, it is too soon to know whether particulate filtration will be necessary for the Californian or US vehicle market generally. It is nevertheless possible that certain vehicle manufacturers will choose filters in order to provide a margin of safety on any positive ignition engine design options selected to meet whatever standards are eventually set.

The new Euro 6 emission standard presents a number of challenging design problems for meeting gasoline emission standards. In particular, how to design a filter, or an exhaust system including a filter, for reducing the number of PM gasoline (positive ignition) emissions, yet at the same time meeting the emission standards for non-PM pollutants such as one or more of oxides of nitrogen (NO_x), carbon monoxide (CO) and unburned hydrocarbons (HC), all at an acceptable back pressure, e.g. as measured by maximum on-cycle backpressure on the EU drive cycle.

It is envisaged that a minimum of particle reduction for a three-way catalysed particulate filter to meet the Euro 6 PM number standard relative to an equivalent flowthrough catalyst is >50%. Additionally, while some backpressure increase for a three-way catalysed wallflow filter relative to an equivalent flowthrough catalyst is inevitable, in our experience peak backpressure over the MVEG-B drive cycle (average over three tests from "fresh") for a majority of passenger vehicles should be limited to <200 mbar, such as <180 mbar, <150 mbar and preferably <120 mbar e.g. <100 mbar.

PM generated by positive ignition engines has a significantly higher proportion of ultrafme, with negligible accumulation and coarse mode compared with that produced by diesel (compression ignition) engines, and this presents challenges to removing it from positive ignition engine exhaust gas in order to prevent its emission to atmosphere. In particular, since a majority of PM derived from a positive ignition engine is relatively small compared with the size distribution for diesel PM, it is not practically possible to use a filter substrate that promotes

AA 1864 PCT positive ignition PM surface-type cake filtration because the relatively low mean pore size of the filter substrate that would be required would produce unpractically high backpressure in the system.

Furthermore, generally it is not possible to use a conventional wallflow filter, designed for trapping diesel PM, for promoting surface-type filtration of PM from a positive ignition engine in order to meet relevant emission standards because there is generally less PM in positive ignition exhaust gas, so formation of a soot cake is less likely; and positive ignition exhaust gas temperatures are generally higher, which can lead to faster removal of PM by oxidation, thus preventing increased PM removal by cake filtration. Depth filtration of positive ignition PM in a conventional diesel wallflow filter is also difficult because the PM is significantly smaller than the pore size of the filter medium. Hence, in normal operation, an uncoated conventional diesel wallflow filter will have a lower filtration efficiency when used with a positive ignition engine than a compression ignition engine.

Another difficulty is combining filtration efficiency with a washcoat loading, e.g. of catalyst for meeting emission standards for non-PM pollutants, at acceptable backpressures. Diesel wallflow particulate filters in commercially available vehicles today have a mean pore size of about 13μm. However, we have found that washcoating a filter of this type at a sufficient catalyst loading such as is described in US 2006/0133969 to achieve required gasoline (positive ignition) emission standards can cause unacceptable backpressure.

In order to reduce filter backpressure it is possible to reduce the length of the substrate. However, there is a finite level below which the backpressure increases as the filter length is reduced. Suitable filter lengths for filters according to the present invention are from 2-12 inches long, preferably 3-6 inches long. Cross sections can be circular and in our development work we have used 4.66 and 5.66 inch diameter filters. However, cross-section can also be dictated by space on a vehicle into which the filter is required to fit. So for filters located in the so-called close coupled position, e.g. within 50 cm of the engine exhaust manifold where space is at a premium, elliptical or oval filter cross sections can be contemplated. As would be expected, backpressure also increases with washcoat loading and soot loading.

There have been a number of recent efforts to combine three-way catalysts with filters for meeting the Euro 6 emission standards.

AA 1864 PCT US 2009/0193796 discloses a three-way conversion catalyst coated onto a particulate trap. The Examples disclose e.g. a soot filter having a catalytic material prepared using two coats: an inlet coat and an outlet coat. The mean pore size of the soot filter substrate used is not mentioned. The inlet coat contains alumina, an oxygen storage component (OSC) and rhodium all at a total loading of 0.17 g in^"3; the outlet coat includes alumina, an OSC and palladium, all at a total loading of 0.42 g in^"3. However, we believe that the three-way catalyst washcoat loading of <0.5 g in^"3 provides insufficient three-way activity to meet the required emission standards alone, i.e. the claimed filter appears to be designed for inclusion in a system for location downstream of a three-way catalyst comprising a flowthrough substrate monolith.

WO 2009/043390 discloses a catalytically active particulate filter comprising a filter element and a catalytically active coating composed of two layers. The first layer is in contact with the in- flowing exhaust gas while the second layer is in contact with the out-flowing exhaust gas. Both layers contain aluminium oxide. The first layer contains palladium, the second layer contains an oxygen-storing mixed cerium/zirconium oxide in addition to rhodium. In Examples, a wallflow filter substrate of unspecified mean pore size is coated with a first layer at a loading of approximately 31 g/1 and a second layer at a loading of approximately 30 g/1. That is, the washcoat loading is less than 1.00 g in^" . For a majority of vehicle applications, this coated filter is unlikely to be able to meet the required emission standards alone.

We have now discovered, very surprisingly, that it is possible to adapt a relatively porous particulate filter - such as a particulate filter adapted for a diesel application - so that it can be used to trap e.g. ultrafme positive ignition PM at an acceptable pressure drop and backpressure. In particular, our inventors have determined that a washcoat that hinders access of the PM to a porous structure of a filter substrate can beneficially promote surface filtration substantially at the expense of depth filtration to the extent that cake filtration of PM derived from a positive ignition engine is promoted or enhanced.

Early indications suggest that positive ignition PM combusts in oxygen at lower temperatures than diesel PM. Investigations are continuing, but the invention makes use of this observation by providing means for trapping the positive ignition PM for combustion in oxygen.

AA 1864 PCT According to one aspect, the invention provides a filter for filtering particulate matter

(PM) from exhaust gas emitted from an engine, such as a positive ignition engine, e.g. a vehicular positive ignition engine such as a stoichiometrically-operated positive ignition engine or a lean burn positive ignition engine, which filter comprising a porous substrate having inlet surfaces and outlet surfaces, wherein the inlet surfaces are separated from the outlet surfaces by a porous structure containing pores, e.g. surface pores, of a first mean pore size, wherein the porous substrate is coated with a washcoat comprising a plurality of solid particles wherein the porous structure of the washcoated porous substrate contains pores of a second mean pore size, and wherein the second mean pore size is less than the first mean pore size.

Early indications are that the present invention is capable of reducing positive ignition engine particle number emissions by >30% such as >50% e.g. >80% or even >90% at acceptable backpressure.

Mean pore size can be determined by mercury porosimetry.

It will be understood that the benefit of the invention is substantially independent of the porosity of the substrate. Porosity is a measure of the percentage of void space in a porous substrate and is related to backpressure in an exhaust system: generally, the lower the porosity, the higher the backpressure. However, the porosity of filters for use in the present invention are typically >40% or >50% and porosities of 45-75% such as 50-65% or 55-60% can be used with advantage. The mean pore size of the washcoated porous substrate is important for filtration. So, it is possible to have a porous substrate of relatively high porosity that is a poor filter because the mean pore size is also relatively high.

The porous substrate can be a metal, such as a sintered metal, or a ceramic, e.g. silicon carbide, cordierite, aluminium nitride, silicon nitride, aluminium titanate, alumina, cordierite, mullite e.g., acicular mullite (see e.g. WO 01/16050), pollucite, a thermet such as Al₂OsZFe, Al₂O3/Ni or B₄CZFe, or composites comprising segments of any two or more thereof. In a preferred embodiment, the filter is a wallflow filter comprising a ceramic porous filter substrate having a plurality of inlet channels and a plurality of outlet channels, wherein each inlet channel and each outlet channel is defined in part by a ceramic wall of porous structure, wherein each inlet channel is separated from an outlet channel by a ceramic wall of porous structure. This filter arrangement is also disclosed in SAE 810114, and reference can be made to this document

AA 1864 PCT for further details. Alternatively, the filter can be a foam, or a so-called partial filter, such as those disclosed in EP 1057519 or WO 01/080978.

Reasons motivating the coating of a wallflow filter for a diesel application are typically different from that of the present invention. In diesel applications, a washcoat is employed to introduce catalytic components to the filter substrate, e.g. catalysts for oxidising NO to NO₂, yet a significant problem is to avoid backpressure issues as soot is accumulated. Accordingly, a balance is struck between the desired catalytic activity and acceptable backpressure.

Contrastingly, a primary motivating factor for washcoating a porous substrate for use in the present invention is to achieve both a desired filtration efficiency and catalytic activity.

In one embodiment, the first mean pore size e.g. of surface pores of the porous structure of the porous filter substrate is from 8 to 45 μm, for example 8 to 25 μm, 10 to 20μm or 10 to 15μm. In particular embodiments, the first mean pore size is >18μm such as from 15 to 45 μm, 20 to 45μm e.g. 20 to 30μm, or 25 to 45μm.

In embodiments, the filter has a washcoat loading of >0.25 g in^"3, such as >0.5g in^"3 or

>0.80 g in^"3, e.g. 0.80 to 3.00 g in^"3. In preferred embodiments, the washcoat loading is >1.00 g in^"3 such as >1.2 g in^"3, >1.5 g in^"3, >1.6 g in^"3 or >2.00 g in^"3 or for example 1.6 to 2.4 g in^"3. In particular combinations of filter mean pore size and washcoat loading the filter combines a desirable level of particulate filtration and catalytic activity at acceptable backpressure.

In a first, preferred embodiment, the filter comprises a surface washcoat, wherein a washcoat layer substantially covers surface pores of the porous structure and the pores of the washcoated porous substrate are defined in part by spaces between the particles (interparticle pores) in the washcoat. That is, substantially no washcoat enters the porous structure of the porous substrate. Methods of making surface coated porous filter substrates include introducing a polymer, e.g. poly vinyl alcohol (PVA), into the porous structure, applying a washcoat to the porous filter substrate including the polymer and drying, then calcining the coated substrate to burn out the polymer. A schematic representation of the first embodiment is shown in Figure 2A.

Methods of coating porous filter substrates are known to the skilled person and include, without limitation, the method disclosed in WO 99/47260, i.e. a method of coating a monolithic

AA 1864 PCT support, comprising the steps of (a) locating a containment means on top of a support, (b) dosing a pre-determined quantity of a liquid component into said containment means, either in the order (a) then (b) or (b) then (a), and (c) by applying pressure or vacuum, drawing said liquid component into at least a portion of the support, and retaining substantially all of said quantity within the support. Such process steps can be repeated from another end of the monolithic support following drying of the first coating with optional firing/calcination.

In this first embodiment, an average interparticle pore size of the porous washcoat is 5.0nm to 5.0μm, such as 0.1-l.Oμm.

A D90 of solid washcoat particles in this first, surface coating embodiment can be greater than the mean pore size of the porous filter substrate and can be in the range 10 to 40μm, such as 15 to 30μm or 12 to 25μm. "D90" as used herein defines the particle size distribution in a washcoat wherein 90% of the particles present have a diameter within the range specified. Alternatively, in embodiments, the mean size of the solid washcoat particles is in the range 1 to 20 μm. It will be understood that the broader the range of particle sizes in the washcoat, the more likely that washcoat may enter the porous structure of the porous substrate. The term "substantially no washcoat enters the porous structure of the substrate" should therefore be interpreted accordingly.

According to a second embodiment, the washcoat can be coated on inlet and/or outlet surfaces and also within the porous structure of the porous substrate. We believe that a surface coating around a pore opening at the inlet and/or outlet surfaces, thereby narrowing the e.g. surface pore size of a bare filter substrate, promotes interaction of the gas phase including PM without substantially restricting the pore volume, so not giving rise to significant increases in back pressure. That is, the pores at a surface of the porous structure comprise a pore opening and the washcoat causes a narrowing of substantially all the pore openings. A schematic representation of the second embodiment is shown in Figure 2B.

Methods of making a filter according to the second embodiment can involve appropriate formulation of the washcoat known to the person skilled in the art including adjusting viscosity and surface wetting characteristics and application of an appropriate vacuum following coating of the porous substrate (see also WO 99/47260).

AA 1864 PCT In the first and second embodiments, wherein at least part of the washcoat is coated on inlet and/or outlet surfaces of the porous substrate, the washcoat can be coated on the inlet surfaces, the outlet surfaces or on both the inlet and the outlet surfaces. Additionally either one or both of the inlet and outlet surfaces can include a plurality of washcoat layers, wherein each washcoat layer within the plurality of layers can be the same or different, e.g. the mean pore size in a first layer can be different from that of a second layer. In embodiments, washcoat intended for coating on outlet surfaces is not necessarily the same as for inlet surfaces.

Where both inlet and outlet surfaces are coated, the washcoat formulations can be the same or different. Where both the inlet and the outlet surfaces are washcoated, the mean pore size of washcoat on the inlet surfaces can be different from the mean pore size of washcoat on the outlet surfaces. For example, the mean pore size of washcoat on the inlet surfaces can be less than the mean pore size of washcoat on the outlet surfaces. In the latter case, a mean pore size of washcoat on the outlet surfaces can be greater than a mean pore size of the porous substrate.

Whilst it is possible for the mean pore size of a washcoat applied to inlet surfaces to be greater than the mean pore size of the porous substrate, it is advantageous to have washcoat having smaller pores than the porous substrate in washcoat on inlet surfaces to prevent or reduce any combustion ash or debris entering the porous structure.

According to a third embodiment, the washcoat sits substantially within, i.e. permeates, the porous structure of the porous substrate. A schematic representation of this third embodiment is shown in Figure 2C. Methods of making a filter according to the third embodiment include the appropriate formulation of the washcoat known to the person skilled in the art including viscosity adjustment, selection of low wetting characteristics and application of an appropriate vacuum following washcoating of the porous substrate (see also WO 99/47260). Alternatively, the porous substrate can be soaked in an appropriate solution of salts and the resulting product dried and calcined.

EP 1663458 discloses a SCR filter, wherein the filter is a wallflow monolith and wherein an SCR catalyst composition permeates walls of the wallflow monolith. The specification discloses generally that the walls of the wallflow filter can contain thereon or therein (i.e. not both) one or more catalytic materials. According to the disclosure, "permeate", when used to

AA 1864 PCT describe the dispersion of a catalyst slurry on the wallflow monolith substrate, means the catalyst composition is dispersed throughout the wall of the substrate.

In the second and third embodiments, wherein at least part of the washcoat is in the porous structure, a size, e.g. a mean size, of the solid washcoat particles can be less than the mean pore size of the porous filter substrate for example in the range 0.1 to 20μm, such as 1 to

18μm, 1 to 16μm, 2 to 15μm or 3 to 12μm. In particular embodiments, the abovementioned size of the solid washcoat particles is a D90 instead of a mean size.

In further particular embodiments, the surface porosity of the washcoat is increased by including voids therein. Exhaust gas catalysts having such features are disclosed, e.g. in our WO 2006/040842 and WO 2007/116881.

By "voids" in the washcoat layer herein, we mean that a space exists in the layer defined by solid washcoat material. Voids can include any vacancy, fine pore, tunnel- state (cylinder, prismatic column), slit etc., and can be introduced by including in a washcoat composition for coating on the filter substrate a material that is combusted during calcination of a coated filter substrate, e.g. chopped cotton or materials to give rise to pores made by formation of gas on decomposition or combustion. Where voids are present, voids are different from, and therefore should not be counted towards determination of, the average interparticle pore size of the porous washcoat.

The average void ratio of the washcoat can be from 5-80%, whereas the average diameter of the voids can be from 0.2 to 500μm, such as 10 to 250μm.

The washcoat for use in the filter of the invention is preferably a catalytic washcoat, and in embodiments is selected from the group consisting of a hydrocarbon trap, a three-way catalyst (TWC), a NO_x absorber, an oxidation catalyst, a selective catalytic reduction (SCR) catalyst, a lean NO_x catalyst and combinations of any two or more thereof. For example, in preferred embodiments, inlet surfaces are coated with a TWC washcoat or NO_x absorber composition and the outlet surfaces are coated with SCR washcoat. In this arrangement, intermittent rich running of the engine, e.g. to regenerate the NO_x absorption capacity of the NO_x absorber, can generate ammonia in situ on the TWC or NO_x absorber for use in reducing NO_x on SCR catalyst on the

AA 1864 PCT outlet surfaces. Similarly, an oxidation catalyst can include hydrocarbon trap functionality. In one embodiment, the inlet surfaces are not coated with SCR catalyst.

The catalytic washcoat, such as the TWC, NO_x absorber, oxidation catalyst, hydrocarbon trap and the lean NO_x catalyst, can contain one or more platinum group metals, particularly those selected from the group consisting of platinum, palladium and rhodium.

TWCs are intended to catalyse three simultaneous reactions: (i) oxidation of carbon monoxide to carbon dioxide, (ii) oxidation of unburned hydrocarbons to carbon dioxide and water; and (iii) reduction of nitrogen oxides to nitrogen and oxygen. These three reactions occur most efficiently when the TWC receives exhaust from an engine running at or about the stoichiometric point. As is well known in the art, the quantity of carbon monoxide (CO), unburned hydrocarbons (HC) and nitrogen oxides (NO_x) emitted when gasoline fuel is combusted in a positive ignition (e.g. spark-ignited) internal combustion engine is influenced predominantly by the air-to-fuel ratio in the combustion cylinder. An exhaust gas having a stoichio metrically balanced composition is one in which the concentrations of oxidising gases (NO_x and O₂) and reducing gases (HC and CO) are substantially matched. The air-to-fuel ratio that produces the stoichiometrically balanced exhaust gas composition is typically given as 14.7:1.

Theoretically, it should be possible to achieve complete conversion of O₂, NO_x, CO and HC in a stoichiometrically balanced exhaust gas composition to CO₂, H₂O and N₂ and this is the duty of the three-way catalyst. Ideally, therefore, the engine should be operated in such a way that the air-to-fuel ratio of the combustion mixture produces the stoichiometrically balanced exhaust gas composition.

A way of defining the compositional balance between oxidising gases and reducing gases of the exhaust gas is the lambda (λ) value of the exhaust gas, which can be defined according to equation (1) as:

Actual engine air-to-fuel ratio/Stoichiometric engine air-to-fuel ratio, (1)

wherein a lambda value of 1 represents a stoichiometrically balanced (or stoichiometric) exhaust gas composition, wherein a lambda value of >1 represents an excess of O₂ and NO_x and the

AA 1864 PCT composition is described as "lean" and wherein a lambda value of <1 represents an excess of HC and CO and the composition is described as "rich". It is also common in the art to refer to the air-to-fuel ratio at which the engine operates as "stoichiometric", "lean" or "rich", depending on the exhaust gas composition which the air-to-fuel ratio generates: hence stoichiometrically- operated gasoline engine or lean-burn gasoline engine.

It should be appreciated that the reduction of NO_x to N₂ using a TWC is less efficient when the exhaust gas composition is lean of stoichiometric. Equally, the TWC is less able to oxidise CO and HC when the exhaust gas composition is rich. The challenge, therefore, is to maintain the composition of the exhaust gas flowing into the TWC at as close to the stoichiometric composition as possible.

Of course, when the engine is in steady state it is relatively easy to ensure that the air-to- fuel ratio is stoichiometric. However, when the engine is used to propel a vehicle, the quantity of fuel required changes transiently depending upon the load demand placed on the engine by the driver. This makes controlling the air-to-fuel ratio so that a stoichiometric exhaust gas is generated for three-way conversion particularly difficult. In practice, the air-to-fuel ratio is controlled by an engine control unit, which receives information about the exhaust gas composition from an exhaust gas oxygen (EGO) (or lambda) sensor: a so-called closed loop feedback system. A feature of such a system is that the air-to-fuel ratio oscillates (or perturbates) between slightly rich of the stoichiometric (or control set) point and slightly lean, because there is a time lag associated with adjusting air-to-fuel ratio. This perturbation is characterised by the amplitude of the air-to-fuel ratio and the response frequency (Hz).

The active components in a typical TWC comprise one or both of platinum and palladium in combination with rhodium, or even palladium only (no rhodium), supported on a high surface area oxide, and an oxygen storage component.

When the exhaust gas composition is slightly rich of the set point, there is a need for a small amount of oxygen to consume the unreacted CO and HC, i.e. to make the reaction more stoichiometric. Conversely, when the exhaust gas goes slightly lean, the excess oxygen needs to be consumed. This was achieved by the development of the oxygen storage component that liberates or absorbs oxygen during the perturbations. The most commonly used oxygen storage

AA 1864 PCT component (OSC) in modern TWCs is cerium oxide (CeO₂) or a mixed oxide containing cerium, e.g. a Ce/Zr mixed oxide.

NO_x absorber catalysts (NACs) are known e.g. from US patent no. 5,473,887 and are designed to adsorb nitrogen oxides (NO_x) from lean exhaust gas (lambda >1) and to desorb the NO_x when the oxygen concentration in the exhaust gas is decreased. Desorbed NO_x may be reduced to N₂ with a suitable reductant, e.g. gasoline fuel, promoted by a catalyst component, such as rhodium, of the NAC itself or located downstream of the NAC. In practice, control of oxygen concentration can be adjusted to a desired redox composition intermittently in response to a calculated remaining NO_x adsorption capacity of the NAC, e.g. richer than normal engine running operation (but still lean of stoichiometric or lambda = 1 composition), stoichiometric or rich of stoichiometric (lambda <1). The oxygen concentration can be adjusted by a number of means, e.g. throttling, injection of additional hydrocarbon fuel into an engine cylinder such as during the exhaust stroke or injecting hydrocarbon fuel directly into exhaust gas downstream of an engine manifold.

A typical NAC formulation includes a catalytic oxidation component, such as platinum, a significant quantity, i.e. substantially more than is required for use as a promoter such as a promoter in a TWC, of a NO_x-storage component, such as barium, and a reduction catalyst, e.g. rhodium. One mechanism commonly given for NO_x-storage from a lean exhaust gas for this formulation is:

NO + V₂ O₂ → NO₂ (2); and

BaO + NO₂ + ¹Z₂ O₂ → Ba(NO₃)₂ (3),

wherein in reaction (2), the nitric oxide reacts with oxygen on active oxidation sites on the platinum to form NO₂. Reaction (3) involves adsorption of the NO₂ by the storage material in the form of an inorganic nitrate.

At lower oxygen concentrations and/or at elevated temperatures, the nitrate species become thermodynamically unstable and decompose, producing NO or NO₂ according to reaction (4) below. In the presence of a suitable reductant, these nitrogen oxides are

AA 1864 PCT subsequently reduced by carbon monoxide, hydrogen and hydrocarbons to N₂, which can take place over the reduction catalyst (see reaction (5)).

Ba(NOs)₂ → BaO + 2NO + ³/₂ O₂ or Ba(NO₃)₂ → BaO + 2NO₂ + ¹Z₂ O₂ (4); and

NO + CO → ¹A N₂ + CO₂ (5);

(Other reactions include Ba(NO₃)₂ +8H₂ → BaO +2NH₃ +5H₂O followed by NH₃ + NO_x → N₂ + yH₂0 or 2NH₃ + 2O₂ + CO → N₂ + 3H₂O + CO₂ etc.).

In the reactions of (2)-(5) above, the reactive barium species is given as the oxide. However, it is understood that in the presence of air most of the barium is in the form of the carbonate or possibly the hydroxide. The skilled person can adapt the above reaction schemes accordingly for species of barium other than the oxide and sequence of catalytic coatings in the exhaust stream.

Oxidation catalysts promote the oxidation of carbon monoxide to carbon dioxide and unburned hydrocarbons to carbon dioxide to water. Typical oxidation catalysts include platinum and/or palladium on a high surface area support.

Hydrocarbon traps typically include molecular sieves and may also be catalysed e.g. with a platinum group metal such as platinum or a combination of both platinum and palladium.

SCR catalysts can be selected from the group consisting of at least one of Cu, Hf, La, Au, In, V, lanthanides and Group VIII transition metals, such as Fe, supported on a refractory oxide or molecular sieve. Suitable refractory oxides include Al₂O₃, TiO₂, CeO₂, SiO₂, ZrO₂ and mixed oxides containing two or more thereof. The non-zeolite catalyst can also include tungsten oxide, e.g. V₂O₅/WO₃/TiO₂.

Lean NO_x catalysts, sometimes also called hydrocarbon-SCR catalysts, DeNO_x catalysts or even non-selective catalytic reduction catalysts, include PtMl₂O₃, Cu- Pt-, Fe-, Co- or Ir- exchanged ZSM-5, protonated zeolites such as H-ZSM-5 or H-Y zeolites, perovskites and Ag/ Al₂O₃. In selective catalytic reduction (SCR) by hydrocarbons (HC), HC react with NO_x, rather than with O₂, to form nitrogen, CO₂ and water according to equation (6):

AA 1864 PCT {HC} + NO_X → N₂ + CO₂ + H₂O (6)

The competitive, non-selective reaction with oxygen is given by Equation (7):

{HC} + O₂ → CO₂ + H₂O (7)

Therefore, good HC-SCR catalysts are more selective for reaction (6) than reaction (7).

In particular embodiments, the washcoat comprises at least one molecular sieve, such as an aluminosilicate zeolite or a SAPO, for trapping positive ignition PM. The at least one molecular sieve can be a small, a medium or a large pore molecular sieve, for example. By "small pore molecular sieve" herein we mean molecular sieves containing a maximum ring size of 8, such as CHA; by "medium pore molecular sieve" herein we mean a molecular sieve containing a maximum ring size of 10, such as ZSM-5; and by "large pore molecular sieve" herein we mean a molecular sieve having a maximum ring size of 12, such as beta. Small pore molecular sieves are potentially advantageous for use in SCR catalysts - see for example WO 2008/132452.

Particular molecular sieves with application in the present invention are selected from the group consisting of AEI, ZSM-5, ZSM-20, ERI including ZSM-34, mordenite, ferrierite, BEA including Beta, Y, CHA, LEV including Nu-3, MCM-22 and EU-I.

In embodiments, the molecular sieves can be un-metallised or metallised with at least one metal selected from the group consisting of groups IB, HB, IIIA, IIIB, VB, VIB, VIB and VIII of the periodic table. Where metallised, the metal can be selected from the group consisting of Cr, Co, Cu, Fe, Hf, La, Ce, In, V, Mn, Ni, Zn, Ga and the precious metals Ag, Au, Pt, Pd and Rh. Such metallised molecular sieves can be used in a process for selectively catalysing the reduction of nitrogen oxides in positive ignition exhaust gas using a reductant. By "metallised" herein we mean to include molecular sieves including one or more metals incorporated into a framework of the molecular sieve e.g. Fe in-framework Beta and Cu in-framework CHA. As mentioned above, where the reductant is a hydrocarbon, the process is sometimes called "hydrocarbon selective catalytic reduction (HC-SCR)", "lean NO_x catalysis" or "DeNO_x catalysis", and particular metals for this application include Cu, Pt, Mn, Fe, Co, Ni, Zn, Ag, Ce, Ga.

AA 1864 PCT Hydrocarbon reductant can either be introduced into exhaust gas by engine management techniques, e.g. late post injection or early post injection (so-called "after injection").

Where the reductant is a nitrogenous reductant (so-called "NH3-SCR"), metals of particular interest are selected from the group consisting of Ce, Fe and Cu. Suitable nitrogenous reductants include ammonia. Ammonia can be generated in situ e.g. during rich regeneration of a NAC disposed upstream of the filter or by contacting a TWC with engine-derived rich exhaust gas (see the alternatives to reactions (4) and (5) hereinabove). Alternatively, the nitrogenous reductant or a precursor thereof can be injected directly into the exhaust gas. Suitable precursors include ammonium formate, urea and ammonium carbamate. Decomposition of the precursor to ammonia and other by-products can be by hydrothermal or catalytic hydrolysis.

The cell density of diesel wallflow filters in practical use can be different from wallflow filters for use in the present invention in that the cell density of diesel wallflow filters is generally 300 cells per square inch (cpsi) or less, e.g. 100 or 200 cpsi, so that the relatively larger diesel PM components can enter inlet channels of the filter without becoming impacted on the solid frontal area of the diesel particulate filter, thereby caking and fouling access to the open channels, whereas wallflow filters for use in the present invention can be up to 300 cpsi or greater, such as 350 cpsi, 400, cpsi, 600 cpsi, 900 cpsi or even 1200 cpsi.

An advantage of using higher cell densities is that the filter can have a reduced cross- section, e.g. diameter, than diesel particulate filters, which is a useful practical advantage that increases design options for locating exhaust systems on a vehicle.

According to a further aspect, the invention provides an exhaust system for a positive ignition engine, which system comprising a filter according to the invention. Positive ignition engines for use in this aspect of the invention can be fuelled by gasoline fuel, gasoline fuel blended with oxygenates including methanol and/or ethanol, liquid petroleum gas or compressed natural gas.

In one embodiment, the exhaust system comprises means for injecting a reductant fluid, e.g. a hydrocarbon or nitrogenous reductant or a precursor thereof, into exhaust gas upstream of the filter. In a particular embodiment, the reductant fluid is a nitrogenous compound.

AA 1864 PCT In a particular embodiment, the injector and filter are both located downstream of a TWC.

In another aspect, the invention provides a positive ignition engine comprising an exhaust system according to the invention and to a vehicle comprising such a positive ignition engine. In a preferred embodiment, the positive ignition engine is a direct injection positive ignition engine.

In a further aspect, the invention provides a method of trapping particulate matter (PM) from exhaust gas emitted from a positive ignition engine by depth filtration, which method comprising contacting exhaust gas containing the PM with a filter comprising a porous substrate having inlet and outlet surfaces, wherein the inlet surfaces are separated from the outlet surfaces by a porous structure containing pores of a first mean pore size, wherein the porous substrate is coated with a washcoat comprising a plurality of solid particles wherein the porous structure of the washcoated porous substrate contains pores of a second mean pore size, and wherein the second mean pore size is less than the first mean pore size.

In order that the invention may be more fully understood, reference is made to the accompanying drawings wherein:

Figure 1 is a graph showing the size distributions of PM in the exhaust gas of a diesel engine. For comparison, a gasoline size distribution is shown at Figure 4 of SAE 1999-01-3530;

Figures 2A-C show schematic drawings of three embodiments of washcoated porous filter substrates according to the invention;

Figure 3 is a schematic graph of mercury porosimetry relating the pore size distribution of a porous filter substrate, a porous washcoat layer and a porous filter substrate including a porous surface washcoat layer; and Figure 4 is a Table setting out a matrix of wallflow filter substrate pore size vs. washcoat loading indicating the suitability of the coated wallflow filter for use in a vehicular gasoline exhaust gas aftertreatment system.

Figures 2A-C show a cross-section through a porous filter substrate 10 comprising a surface pore 12. Figure 2 A shows a first embodiment, featuring a porous surface washcoat layer 14 comprised of solid washcoat particles, the spaces between which particles define pores

AA 1864 PCT (interparticle pores). It can be seen that the washcoat layer 14 substantially covers the pore 12 of the porous structure and that a mean pore size of the interparticle pores 16 is less than the mean pore size 12 of the porous filter substrate 10.

Figure 2B shows a second embodiment comprising a washcoat that is coated on an inlet surface 16 and additionally within a porous structure 12 of the porous substrate 10. It can be seen that the washcoat layer 14 causes a narrowing of a pore openings of surface pore 12, such that a mean pore size 18 of the coated porous substrate is less than the mean pore size 12 of the porous filter substrate 10.

Figure 2C shows a third embodiment wherein the washcoat 14 sits substantially within, i.e. permeates, the porous 12 structure of the porous substrate 10.

Figure 3 shows an illustration of a graph relating pore size to pore number for a porous filter substrate 20, a porous washcoat layer 22 and a porous diesel filter substrate including a surface washcoat layer 24. It can be seen that the filter substrate has a mean pore size of the order of about 15μm. The washcoat layer has a bimodal distribution comprised of intraparticle pores 22A (at the nanometre end of the range) and interparticle pores 22B towards the micrometer end of the scale. It can also be seen that by coating the porous filter substrate with a washcoat according to the invention that the pore distribution of the bare filter substrate is shifted in the direction of the interparticle washcoat pore size (see arrow).

Figure 4 sets out a matrix showing preliminary results for a washcoat loading study for a three-way catalyst washcoat on three wallflow filters having different mean pore sizes. In conclusion, there is a band of acceptable backpressure and filtration starting with a combination of 13μm mean pore size wallflow filter and relatively low washcoat loading (0.4 g in^" ) through the 20μm and 13μm pore size substrates having 0.8 g in^"3 to the 1.6 and 2.4 g in^"3 loadings on the 38μm and 20μm mean pore size substrates.

However, overlying this matrix for three-way catalyst use is that washcoat loadings of

>1.6 g in^"3 are preferred for acceptable three-way catalyst activity in a stand-alone product. The invention allows a combination of sufficient three-way catalyst activity and PM filtration to be achieved without a significant increase in backpressure. Increased washcoat loadings on lower mean pore size wallflow filter substrates can only be used in applications that can tolerate

AA 1864 PCT increased backpressure. With reference to Figure 4, whilst in certain applications where backpressure increases can be tolerated, a 13μm mean pore size wallflow filter substrate can be used in combination with >1.6 g in^" washcoat loading, we presently prefer to use a mean pore size of >20μm for >1.6 g in^" loadings to achieve a desirable balance between catalyst activity, filtration and backpressure. A benefit of the invention is that a state-of-the-art three-way catalyst comprising a flow-through monolith substrate that is typically located on a vehicle in either an underfloor or close-coupled location can be replaced with a filter according to the invention to provide sufficient three-way activity to meet legislative requirements for gaseous HC, CO and NO_x emissions, while also meeting particle number standards as required by e.g. Euro 6 standards.

The filter according to the invention could obviously be used in combination with other exhaust system aftertreatment components to provide a full exhaust system aftertreatment apparatus, e.g. a low thermal mass TWC upstream of the filter and/or downstream catalytic elements, e.g. NO_x trap or SCR catalyst, according to specific requirements. So, in vehicular positive ignition applications producing relatively cool on-drive cycle exhaust gas temperatures, we contemplate using a low thermal mass TWC disposed upstream of the filter according to the invention. For vehicular lean-burn positive ignition applications, we envisage using a filter according to the invention upstream or downstream of a NO_x trap. In vehicular stoichiometrically-operated positive ignition engines, we believe that the filter according to the present invention can be used as a standalone catalytic exhaust system aftertreatment component. That is, in certain applications the filter according to the present invention is adjacent and in direct fluid communication with the engine without intervening catalysts therebetween; and/or an exit to atmosphere from an exhaust gas aftertreatment system is adjacent to and in direct fluid communication with the filter according to the present invention without intervening catalysts therebetween.

An additional requirement of a TWC is a need to provide a diagnosis function for its useful life, so called "on-board diagnostics" or OBD. A problem in OBD arises where there is insufficient oxygen storage capacity in the TWC, because OBD processes for TWCs use remaining oxygen storage capacity to diagnose remaining catalyst function. However, if insufficient washcoat is loaded on the filter such as in the specific Examples disclosed in US 2009/0193796 and WO 2009/043390, there may not be enough OSC present to provide an accurate OSC "delta" for OBD purposes. Since the present invention enables washcoat loadings

AA 1864 PCT approaching current state-of-the-art TWCs, the filters for use in the present invention can be used with advantage in current OBD processes.

In order that the invention may be more fully understood the following Examples are provided by way of illustration only. The washcoat loadings quoted in the Examples were obtained using the method disclosed in WO 99/47260 described hereinabove by coating half of the washcoat from one end and the remaining half of the washcoat from the other end, i.e. the entire washcoat was not coated only on the inlet or outlet channels of the filter, but on both the inlet and outlet channels of the filter.

Example 1

Two three-way catalyst (TWC) coatings were prepared at a washcoat loading of 2.4 g/in and a precious metal loading of 85 g/ft³ (Pd:Rh 16:1); one was milled to a small particle size (d90 <5 μm) that would be expected to pass into the pore structure of a wallflow filter ("in- wall"), while the other was less milled (d90 <17 μm) so that it would be expected preferentially to locate more at the surface of a wallflow filter wall ("on-wall"). The coatings were applied to 4.66 x 4.5 inch 300 cells per square inch cordierite wallflow filter substrates having 12 thousandths of an inch wall thickness ("300/12") with a nominal average pore size of 20 micrometers (hereinafter "microns") (62% porosity). Each filter was hydrothermally oven-aged at 980⁰C for 4 hours and installed in a close-coupled position on a Euro 5 passenger car with a 1.4L direct injection gasoline engine. Each filter was evaluated over a minimum of three MVEG-B drive cycles, measuring the reduction in particle number emissions relative to a reference catalyst, wherein the filter was exchanged for a TWC coated onto a flowthrough substrate monolith at the same washcoat and precious metal loadings - and the backpressure differential was determined between sensors mounted upstream and downstream of the filter (or reference catalyst).

In Europe, since the year 2000 (Euro 3 emission standard) emissions are tested over the New European Driving Cycle (NEDC). This consists of four repeats of the previous ECE 15 driving cycle plus one Extra Urban Driving Cycle (EUDC) with no 40 second warm-up period before beginning emission sampling. This modified cold start test is also referred to as the

"MVEG-B" drive cycle. All emissions are expressed in g/km.

AA 1864 PCT The Euro 5/6 implementing legislation introduces a new PM mass emission measurement method developed by the UN/ECE Particulate Measurement Programme (PMP) which adjusts the PM mass emission limits to account for differences in results using old and the new methods. The Euro 5/6 legislation also introduces a particle number emission limit (PMP method), in addition to the mass-based limits.

The results in Table 1 demonstrate that the filters prepared with the larger particle size "on-wall" washcoat have significantly improved particle number reduction than the filters prepared with the smaller particle size "in-wall" washcoat, with a small, but acceptable, increase in peak backpressure.

Table 1. Effect of washcoat location within filter on particle number reduction and backpressure

(BP)

Example 2

5.66 x 3 inch cordierite wallflow filter substrates with a cell density of 300 cells per square inch and a wall thickness of 12 thousandths of an inch (approximately 0.3mm) were coated with a three-way catalyst (TWC) coating at a washcoat loading of 0.8 g/in³ and a palladium loading of 80 g/ft³. Three pore structures were compared: a nominal average pore size of 38 microns at 65% porosity, a nominal average pore size of 20 microns at 62% porosity and a nominal average pore size of 15 microns at 52% porosity. Each filter was hydrothermally oven- aged at 980⁰C for 4 hours and installed in the underfloor position on a Euro 4 passenger car with a 1.4L direct injection gasoline engine, with a fully formulated three-way catalyst coated on a flowthrough substrate monolith located in the close-coupled position, i.e. upstream of the filter. Each filter was evaluated over a minimum of three MVEG-B drive cycles, measuring the

AA 1864 PCT reduction in particle number emissions relative to a reference system, wherein the underfloor filter was exchanged for a TWC coated on a flowthrough substrate monolith at identical washcoat and palladium loadings and the backpressure differential was determined between sensors mounted upstream of the close-coupled TWC and downstream of the filter (or reference catalyst). The peak backpressure results given in Table 2 are the backpressure reading on the third repeat of the MVEG-B cycle.

The results in Table 2 demonstrate that the 38 micron filter had significantly lower levels of particle number removal (insufficient for this vehicle application), albeit with the lowest backpressure. The 20 micron filter gave acceptable levels of particle number reduction with a moderate increase in backpressure. The 15 micron filter was most effective at reducing particle number emissions but had significantly higher backpressure than the 20 micron filter embodiment.

Table 2. Comparison of particle number reduction and backpressure (BP) for different pore size filters

Example 3

4.66 x 4.5 inch, 300/12 cordierite wallflow filter substrates with a nominal average pore size of 20 microns and porosity of 62% were coated with a three-way catalyst coating at washcoat loadings of 0.8, 1.6 and 2.4 g/in³ respectively. Each sample had a precious metal loading of 85 g/ft³ (Pd:Rh 16:1). Each filter was hydrothermally oven-aged at 980⁰C for 4 hours and installed in a close-coupled position on a Euro 4 passenger car with a 1.4L direct injection gasoline engine. Each filter was evaluated over a minimum of three MVEG-B drive cycles,

AA 1864 PCT measuring the reduction in particle number emissions relative to a reference catalyst, wherein the close-coupled filter was exchanged for a TWC coated on a flowthrough substrate monolith at an identical washcoat and precious metal loading, the backpressure differential and the conversion efficiency for gaseous HC, CO and NO_x emissions were determined between sensors mounted upstream and downstream of the filter (or reference catalyst).. Only non-methane hydrocarbons (NMHC) conversion is reported in Table 3 (the NMHC for Euro 6 is 68mg/km within a total hydrocarbon emission limit of 100mg/km).

The results in Table 3 demonstrate that the filter prepared with a washcoat loading of 0.8 g/in had significantly lower levels of particle number removal and the lowest NMHC conversion efficiency. Such TWC performance would not be sufficient to meet Euro 6 gaseous emissions limits for a typical passenger car. Increasing the washcoat loading to 1.6 and 2.4 g/in gave greater reductions in particle number emissions, albeit at increasing, but acceptable, backpressure. TWC activity (as represented in Table 3 by NMHC performance) was also significantly improved with the higher washcoat loadings.

Table 3. Comparison of particle number reduction, backpressure (BP) and TWC activity at different washcoat loadings

* The "engineering target" is often used by the vehicle manufacturers and represents a percentage of the legislated emissions. For the purposes of these Examples, we have used an engineering target of 80%. Since the Euro 6 NMHC standard is 68mg/km, the engineering target is 54mg/km. The calculated percentage of this number is used to assess the reduction in NMHC results achieved over the MVEG-B drive cycle. This gives values above and below 100% that relate well to the acceptable three-way catalyst activity.

AA 1864 PCT Example 4

4.66 x 4.5 inch, 300/12 cordierite wallflow filter substrates with a cell density of 300 cells per square inch and a wall thickness of approximately 0.3mm were coated with a three-way catalyst coating at a washcoat loading of 1.6 g/in³ and a precious metal loading of 85 g/ft³ (Pd:Rh 16:1). Two pore structures were compared: a nominal average pore size of 38 microns at 65% porosity and a nominal average pore size of 20 microns at 62% porosity. A smaller pore sample was not evaluated because, from the results obtained from the Example 2, the backpressure was expected to be too great for the Euro 4 passenger car in this test. Each filter was hydrothermally oven aged at 980⁰C for 4 hours and installed in a close-coupled position on a Euro 4 passenger car with a 1.4L direct injection gasoline engine. Each filter was evaluated over a minimum of three MVEG-B drive cycles, measuring the reduction in particle number emissions relative to a reference catalyst, wherein the close-coupled filter was exchanged for a TWC coated on a flowthrough substrate monolith at an identical washcoat and precious metal loading , the backpressure differential and the conversion efficiency for gaseous HC, CO and NO_x emissions were determined between sensors mounted upstream and downstream of the filter (or reference catalyst). Only non-methane hydrocarbons (NMHC) conversion is reported in Table 4.

The results in Table 4 demonstrate that the 38 micron filter had significantly lower levels of particle number removal (insufficient for this vehicle application) and lower backpressure, which may be acceptable in other vehicle applications. The 20 micron filter gave good levels of particle number reduction with a moderate increase in backpressure. Both samples had good TWC activity at a washcoat loading of 1.6 g/in³.

AA 1864 PCT Table 4. Comparison of particle number reduction, backpressure (BP) and TWC activity for different pore size filters

^ See footnote to Table 3.

Example 5

4.66 x 4.5 inch, 300/12 cordierite wallflow filter substrates with a cell density of 300 cells per square inch and a wall thickness of approximately 0.3mm were coated with a three-way catalyst coating at a washcoat loading of 2.4 g/in³ and a precious metal loading of 85 g/ft³ (Pd:Rh 16:1). Two pore structures were compared: a nominal average pore size of 38 microns at 65% porosity and a nominal average pore size of 20 microns at 62% porosity. A smaller pore sample was not evaluated because, from the results obtained from the Example 2, the backpressure was expected to be too great for the Euro 5 passenger car in this test. Each filter was hydrothermally oven-aged at 980⁰C for 4 hours and installed in a close-coupled position on a Euro 5 passenger car with a 1.4L direct injection gasoline engine. The filters were evaluated over a minimum of three MVEG-B drive cycles, measuring the reduction in particle number emissions relative to a reference catalyst, wherein the close-coupled filter was exchanged for a TWC coated on a flowthrough substrate monolith at an identical washcoat and precious metal loading, the backpressure differential and the conversion efficiency for gaseous HC, CO and NO_x emissions were determined between sensors mounted upstream and downstream of the filter (or reference catalyst). Only non-methane hydrocarbons (NMHC) conversion is reported in Table 5.

The results in Table 5 demonstrate that the 38 micron filter had significantly lower levels of particle number removal (borderline for this vehicle application) and lower backpressure than the 20 micron filter embodiment. The 20 micron filter gave good levels of particle number reduction with a moderate increase in backpressure. Both samples had good TWC activity at a washcoat loading of 2.4 g/in . Both samples displayed greater particle number reduction and increased backpressure compared to the 1.6 g/in³ samples described in Example 4.

AA 1864 PCT

Table 5. Comparison of particle number reduction, backpressure (BP) and TWC activity for different pore size filters

t See footnote to Table 3.

Example 6

118 x 60mm, 360 cells per square inch cordierite wallflow filter substrates having 5 thousandths of an inch cell wall thickness (360/5) with a nominal average pore size of 13 microns and porosity of 48% were coated with a three-way catalyst coating at washcoat loadings of 0.4 and 0.8 g/in³. Each sample had a precious metal loading of 85 g/ft³ (Pd:Rh 16:1). Higher washcoat loadings were not evaluated because the resulting backpressure was expected to be too great for the Euro 4 passenger car in this test. A fresh (i.e. un-aged) filter was installed in a close-coupled position on a Euro 4 passenger car with a 1.4L direct injection gasoline engine.

Each filter was evaluated over a minimum of three MVEG-B drive cycles, measuring the reduction in particle number emissions relative to a reference catalyst, wherein the close-coupled filter was exchanged for a TWC coated on a flowthrough substrate monolith at an identical washcoat and precious metal loading and the backpressure differential and the conversion efficiency for gaseous HC, CO and NO_x emissions were determined between sensors mounted upstream and downstream of the filter (or reference catalyst). Only non-methane hydrocarbons

(NMHC) conversion is reported in Table 6.

The results in Table 6 demonstrate that the 13 micron filter prepared with a washcoat loading of 0.8 g/in gave moderate levels of particle number removal (borderline for this vehicle application) but had extremely high backpressure. Reducing the washcoat loading to 0.4 g/in gave more acceptable backpressure but a smaller reduction in particle number emissions. Such low washcoat levels would not be expected to give sufficient three-way catalyst activity to meet Euro 6 emission standards.

AA 1864 PCT

Table 6. Comparison of particle number reduction and backpressure (BP) at different washcoat loadings

Example 7

A Euro 5 passenger car with a 2.0L direct injection gasoline engine equipped with a fully formulated three-way catalyst coated on a flowthrough substrate monolith in the close-coupled position was tested over the MVEG-B and FTP (Federal Test Procedure) 75 drive cycles. The number of particles emitted over the MVEG-B drive cycle was measured according to the PMP methodology. The mass of particulate matter emitted over the FTP 75 drive cycle was measured following standard protocols. A 125 x 120 mm, 300/12 cordierite wallflow filter with a nominal average pore size of 12 microns and porosity of 55% coated with a three-way catalyst coating at a washcoat loading of 0.8 g/in³ and a precious metal loading of 20 g/ft³ (Pd:Rh 3:1) was then fitted in the underfloor position, i.e. downstream of the flowthrough substrate monolith. Particulate mass and number emissions measurements were repeated.

The results in Table 7 demonstrate that fitment of the additional coated filter reduced particle number emissions over the MVEG-B cycle by -99% and reduced the particulate mass emitted over the FTP 75 cycle by -75% relative to the flowthrough TWC-only system.

Depending what CARB PM emission standard is adopted, the 2.7 mg PM/mile figure could fail that standard.

AA 1864 PCT

Table 7. Effect of filter fitment on particulate number and mass emissions

For the avoidance of any doubt, the entire contents of all prior art documents cited herein is incorporated herein by reference.

AA 1864 PCT

Claims

CLAIMS:

1. A filter for filtering particulate matter (PM) from exhaust gas emitted from a positive ignition engine, which filter comprising a porous substrate having inlet surfaces and outlet surfaces, wherein the inlet surfaces are separated from the outlet surfaces by a porous structure containing pores of a first mean pore size, wherein the porous substrate is coated with a washcoat comprising a plurality of solid particles wherein the porous structure of the washcoated porous substrate contains pores of a second mean pore size, and wherein the second mean pore size is less than the first mean pore size.

2. A filter according to claim 1, wherein a first mean pore size of the porous structure of the porous substrate is from 8 to 45 μm.

3. A filter according to claim 1 or 2, wherein the washcoat loading is >0.50 g in^"3.

4. A filter according to claim 3, wherein the washcoat loading is > 1.00 g in^"3.

5. A filter according to claim 1, 2, 3 or 4, comprising a surface washcoat, wherein a washcoat layer substantially covers surface pores of the porous structure and the pores of the washcoated porous substrate are defined in part by spaces between the particles (interparticle pores) in the washcoat.

6. A filter according to claim 5, wherein a mean interparticle pore size of the porous washcoat is from 5.0nm to 5.0μm.

7. A filter according to any preceding claim, wherein a mean size of solid washcoat particles is greater than the first mean pore size.

8. A filter according to claim 7, wherein the mean size of the solid washcoat particles is in the range 1 to 40μm.

9. A filter according to claim 1, 2, 3 or 4, wherein the pores at a surface of the porous structure comprise a pore opening and the washcoat causes a narrowing of substantially all the surface pore openings.

AA 1864 PCT

10. A filter according to claim 1, 2, 3 or 4, wherein the washcoat sits substantially within the porous structure of the porous substrate.

11. A filter according to claim 9 or 10, wherein a mean size of solid washcoat particles is less than a mean pore size of the porous substrate.

12. A filter according to claim 11, wherein a mean size of solid washcoat particles is in the range 0.1 to 20μm.

13. A filter according to claim 11, wherein a D90 of solid washcoat particles is in the range 0.1 to 20μm.

14. A filter according to any preceding claim, wherein the washcoat is coated on inlet surfaces, outlet surfaces or both the inlet and the outlet surfaces.

15. A filter according to claim 14, wherein both the inlet and the outlet surfaces are washcoated and wherein a mean pore size of washcoat on the inlet surfaces is different from a mean pore size of washcoat on the outlet surfaces.

16. A filter according to claim 15, wherein the mean pore size of washcoat on the inlet surfaces is less than the mean pore size of washcoat on the outlet surfaces.

17. A filter according to claim 16, wherein a mean pore size of washcoat on the outlet surfaces is greater than a mean pore size of the porous substrate.

18. A filter according to any preceding claim, wherein the porous substrate is a ceramic wallflow filter, a metal filter or a ceramic foam.

19. A filter according to claim 18, wherein the metal filter is a sintered metal filter, a partial filter and/or comprises a wire mesh.

20. A filter according to any preceding claim, wherein the washcoat is a catalytic washcoat.

AA 1864 PCT

21. A filter according to claim 20, wherein the catalytic washcoat is selected from the group consisting of a hydrocarbon trap, a three-way catalyst, a NO_x absorber, an oxidation catalyst, a selective catalytic reduction (SCR) catalyst and a lean NO_x catalyst.

22. A filter according to claim 20 or 21, wherein the catalytic washcoat comprises at least one molecular sieve.

23. A filter according to claim 22, wherein the at least one molecular sieve is a small, medium or large pore molecular sieve.

24. A filter according to claim 22 or 23, wherein the at least one molecular sieve is selected from the group consisting of AEI, ZSM-5, ZSM-20, ERI, LEV, mordenite, BEA, Y, CHA, MCM-22 and EU-I.

25. A filter according to claim 22, 23 or 24, wherein the molecular sieve is un-metallised or is metallised with at least one metal selected from the group consisting of groups IB, HB, IIIA, IIIB, VB, VIB, VIB and VIII of the periodic table.

26. A filter according to claim 25, wherein the metal is selected from the group consisting of Cr, Co, Cu, Fe, Hf, La, Ce, In, V, Mn, Ni, Zn, Ga and the precious metals Ag, Au, Pt, Pd and Rh.

27. A filter according to claim 25 or 26, wherein the metal is selected from the group consisting of Cu, Pt, Mn, Fe, Co, Ni, Zn, Ag, Ce and Ga.

28. A filter according to claim 25, 26 or 27, wherein the metal is selected from the group consisting of Ce, Fe and Cu.

29. An exhaust system for a positive ignition engine, which system comprising a filter according to any preceding claim.

30. An exhaust system according to claim 29, comprising means for injecting a reductant fluid into exhaust gas upstream of the filter.

AA 1864 PCT

31. An exhaust system according to claim 30, wherein the reductant fluid is a nitrogenous compound.

32. A positive ignition engine comprising an exhaust system according to claim 29, 30 or 31.

33. A positive ignition engine according to claim 31, fuelled with a hydrocarbon fuel selected from the group consisting of gasoline, gasoline blended with methanol and/or ethanol, liquid petroleum gas and compressed natural gas.

34. A method of trapping particulate matter (PM) from exhaust gas emitted from a positive ignition engine by depth filtration, which method comprising contacting exhaust gas containing the PM with a filter comprising a porous substrate having inlet and outlet surfaces, wherein the inlet surfaces are separated from the outlet surfaces by a porous structure containing pores of a first mean pore size, wherein the porous substrate is coated with a washcoat comprising a plurality of solid particles wherein the porous structure of the washcoated porous substrate contains pores of a second mean pore size, and wherein the second mean pore size is less than the first mean pore size.