Tuning Wiki

Petrol and diesel engines burn the same sort of hydrocarbon fuel, but they do not make the same exhaust. A petrol engine normally mixes fuel and air before the spark, and a modern one is held near the chemically balanced mixture so that one catalyst can treat carbon monoxide, hydrocarbons and nitrogen oxides together. A diesel engine injects fuel into hot compressed air and burns it as a spray, with air left over. That leaves oxygen in the exhaust, which stops a three-way catalyst from removing nitrogen oxides, and it leaves soot from the parts of the spray that are locally rich. The equipment that follows — catalysts, exhaust-gas recirculation, secondary air and particulate filters — is there because of that difference. Figures below are taken from the cited regulation, test manual or technical paper. They describe those documents, not a target for every engine.

This page explains how the emissions are formed and how the control systems work. It does not describe how to remove or defeat them. The MOT inspection manual treats manufacturer-fitted emissions equipment that is missing, obviously modified or obviously defective as a defect on the vehicles it specifies, and it has a separate defect for a diesel particulate filter that shows evidence of tampering.

What each engine emits

The regulated tailpipe species for European passenger cars are carbon monoxide (CO), hydrocarbons, oxides of nitrogen (NOx), particulate mass (PM) and, from Euro 5 and 6, particle number (PN). Carbon dioxide is measured as well, but it is the product of complete combustion rather than a pollutant the catalyst is asked to destroy. In Regulation (EC) 715/2007, as consolidated on 1 September 2020, positive ignition means a spark-ignition engine and compression ignition means a diesel. For a category M car the Euro 6 limits in that text are:

Euro 6 limit, category MPositive ignition (petrol)Compression ignition (diesel)
Carbon monoxide1,000 mg/km500 mg/km
Total hydrocarbons100 mg/kmNo separate THC limit
Non-methane hydrocarbons68 mg/kmNo separate NMHC limit
NOx60 mg/km80 mg/km
Hydrocarbons plus NOxNo combined limit170 mg/km
Particulate mass4.5 mg/km, direct injection only4.5 mg/km
Particle number6.0 × 1011 per km, direct injection only6.0 × 1011 per km

Other vehicle classes in the same table have higher gaseous limits. The regulation states that the positive-ignition particle limits apply only to direct-injection engines, and it records a temporary manufacturer option of 6.0 × 1012 particles per km for those engines. Later standards exist; this table is the Euro 6 text, not a statement of every limit in force for a new type-approval. Regulation (EC) 715/2007, consolidated text of 1 September 2020, Annex I, Table 2.

Those limits already show the split. A diesel has the harder NOx and particulate problem, so its CO limit is tighter in the table while hydrocarbons are controlled together with NOx. A port-injected petrol engine has no particle-number limit in that table; a direct-injection petrol engine does, because injecting petrol into the cylinder raises the number of particles.

Petrol

CO is carbon that did not find enough oxygen. It rises when the mixture is rich, when a cold engine is deliberately enriched, or when a catalyst is not yet working. Hydrocarbons are fuel that did not burn, or that burned only part of the way: crevices, a wet intake port, a misfire, or a very lean pocket that the flame does not cross. NOx, mostly nitric oxide with some nitrogen dioxide, forms when nitrogen and oxygen in the air sit together at high temperature. A stoichiometric petrol engine gives a three-way catalyst the mixture it needs to oxidise CO and hydrocarbons and reduce NOx at the same time. AECC, Emissions Control Technologies, section 5.2.

A premixed petrol charge produces little soot, because the fuel has evaporated and mixed before the spark. Direct injection changes that. Fuel that lands on the piston or the liner, and fuel that burns as a diffusion flame rather than as a premixed flame, produces particulates. AECC identifies the higher particle number of gasoline direct injection, compared with port fuel injection, as the reason gasoline particulate filters were developed. The same paper notes that those particle numbers can be higher on the road than on the regulatory cycle.

Diesel

A diesel does not throttle the air to match the fuel. Load is set by how much fuel is injected into a cylinder already full of hot air. The US Department of Energy’s diesel primer describes the result: the engine does not need one uniform mixture, combustion starts by compression ignition, and the exhaust of a lean diesel still contains oxygen. Soot comes from the rich core of the burning spray. NOx comes from the hot regions where oxygen remains. The primer also says that diesel exhaust therefore contains more oxygen, and less CO2 and water, than petrol exhaust, which is why a diesel needs a larger share of recirculated exhaust if EGR is the tool being used to cut NOx. DOE, Diesel Power: Clean Vehicles for Tomorrow, July 2010.

CO and hydrocarbons are still there — a cold diesel, a late injection, or oil on the cylinder wall will make them — but they are not the species that dominates the aftertreatment. AECC says the oxygen in diesel exhaust rules out a three-way catalyst. A diesel oxidation catalyst oxidises CO and hydrocarbons, and it also oxidises some of the hydrocarbons stuck to the soot, which lowers the measured particulate mass. NOx has to be dealt with by EGR, a NOx storage catalyst, selective catalytic reduction, or a combination.

What a test can actually see

Type approval and a workshop test do not look for the same things. Euro 6 limits above are cycle results, in milligrams or particles per kilometre, measured with laboratory instruments. A UK MOT is an in-service inspection. The MOT manual, section 8.2, separates the two engines.

On a spark-ignition engine the gaseous test measures CO and hydrocarbons, and on a catalyst test it also checks lambda, the air–fuel ratio relative to stoichiometric. The manual’s basic emissions test uses these default limits: CO not above 0.2% and hydrocarbons not above 200 ppm at a fast idle of 2,500 to 3,000 rpm, lambda between 0.97 and 1.03 at that fast idle, and CO not above 0.3% at idle. A full catalyst test uses the vehicle’s own limits where they are known, or the default limits where they are not. The manual’s example of the older non-catalyst limits is CO of 3.5% and hydrocarbons of 1,200 ppm. It also says some carburettor and fuel-injection designs give unstable readings, and that a limit has to be exceeded constantly for at least five seconds before a fail. Dense blue or clearly visible black smoke is a separate defect. The manual does not make NOx or particle number part of this gaseous test.

On a compression-ignition engine the metered test is smoke opacity, from a diesel smoke meter, on vehicles first used on or after 1 January 1980. The result is a light-absorption coefficient. The limit is the manufacturer’s plate value where there is one. The manual also tells the tester to look at visible equipment: a diesel oxidation catalyst, a particulate filter, an EGR valve and selective catalytic reduction equipment. A particulate filter that has been cut open and rewelded is rejected unless there is evidence of a valid reason, such as cleaning. Opacity is not a particle-number count, and it is not a NOx reading. A diesel can be dirty on NOx and still pass a smoke test, and a petrol car can be dirty on NOx and still pass a CO and hydrocarbon test.

Workshop gas analysers and type-approval benches are different instruments as well. Bosch BEA equipment covers petrol and gas analysis, diesel smoke opacity, and, in the BEA 090, a particle counter. A particle counter is a different measurement from an opacity meter. A four-gas analyser reports CO, CO2, hydrocarbons and oxygen, and it calculates lambda from that sample. It does not replace the flame-ionisation, chemiluminescence and particle-counting methods used for a type-approval result.

Carburettors

A carburettor meters petrol into the air entering the manifold. The mixture moves when the venturi airflow, the fuel level, the temperature and the throttle all move, and cold starting and acceleration are handled by enrichment: a choke, and usually an accelerator pump. The MOT manual treats a choke as an enrichment device that must not be operating during the idle check. While that enrichment is in, there is not enough oxygen for the carbon in the fuel, so CO in the sample rises. Fuel that lands on the manifold wall, or that is pulled through in a lump when the throttle opens, leaves the cylinder as hydrocarbons.

A plain carburettor has no exhaust oxygen sensor and no loop that can hold lambda at 1. The mixture wanders either side of stoichiometric, and that is where a three-way catalyst loses one of its jobs. Rich of the balance point, CO and hydrocarbons get through because there is not enough oxygen to finish oxidising them. Lean of it, oxygen is left over and the catalyst cannot reduce NOx. AECC records that European standards from 1993 effectively required catalysts on petrol cars, and that unleaded fuel was what made the three-way catalyst usable, because lead poisons it. Feedback carburettors and early injection systems were the bridge. They could hold the mixture near the catalyst window in a way a fixed jet could not. The manual’s non-catalyst limits, several times the catalyst-test defaults, are the practical trace of that older metering.

Port injection

Port injection, or manifold injection, puts the petrol in at the intake port instead of in a venturi. Bosch L-Jetronic is the usual reference point. An airflow meter and the engine speed tell the controller how much air is entering. The controller opens a solenoid injector for a calculated time. Bosch’s description of the conventional system fires the injectors together, twice per four-stroke cycle, so fuel sprayed against a closed inlet valve sits in the port until that valve opens. The time in the port is useful: the fuel evaporates and mixes with the air before it burns, which is why a port-injected engine makes far fewer particles than a direct-injected one. AECC’s comparison is explicit on that point.

Where an oxygen sensor is fitted, the same pulse width can be trimmed from the exhaust. AECC describes the three-way catalyst as a closed-loop system: the lambda sensor holds the air–fuel ratio where the catalyst can oxidise CO and hydrocarbons to CO2 and water and reduce NOx to nitrogen. The coating contains alumina, ceria and other oxides, with platinum, palladium and rhodium. During a lean swing the ceria stores oxygen; during a rich swing it gives that oxygen back, so CO and hydrocarbons can still be oxidised. AECC describes these oxygen-storage materials as what widens the air–fuel window for three-way operation. The MOT lambda band of 0.97 to 1.03 is an in-service check that this loop, and the catalyst, are still in that region at fast idle. It is not a map of the whole operating range.

Port injection did not remove every hydrocarbon path. Fuel still wets the port wall, a cold engine still needs enrichment, and the catalyst still does nothing until it is hot. What it did remove was the carburettor’s inability to meter cylinder by cylinder and to correct itself from the exhaust. CO and hydrocarbons at the tailpipe of a healthy port-injected car are mostly a cold-start and catalyst question, not a steady-cruise question. NOx is handled by the catalyst for as long as the mixture stays near stoichiometric. Particle number stays low enough that the Euro 6 positive-ignition particle limits were not applied to these engines.

Direct injection

Direct injection puts the petrol into the cylinder. The timing of that injection decides what kind of charge the spark sees. An early injection, during the intake stroke, gives the fuel time to evaporate and mix. The charge is then broadly homogeneous, and the engine can be run at lambda 1 so that the three-way catalyst still works. A late injection, near the end of compression, does not have that time. The fuel stays in a cloud. That is stratified charge: the cloud near the spark plug is ignitable, and the rest of the cylinder is much leaner than stoichiometric.

Bosch’s gasoline-systems description of DI-Motronic, given by Rolf Leonhard in 2003, said the stratified-charge system had reduced fuel consumption by about 10% compared with port injection. The saving is real in principle. A lean, less-throttled charge wastes less energy pulling air past a nearly closed throttle. The exhaust chemistry changes with it. Overall lambda is above 1, so oxygen remains in the exhaust and a three-way catalyst can still oxidise CO and hydrocarbons but cannot reduce NOx. AECC treats lean-burn direct-injection petrol in the same NOx chapter as diesel: both need something other than a three-way catalyst, such as a NOx storage catalyst, once they leave stoichiometric operation.

The particle penalty comes from the same late, poorly mixed injection. A review of gasoline direct-injection particle numbers distinguishes two ways of aiming the spray. In a wall-guided system the spray is pointed at a bowl in the piston, which is supposed to turn the fuel up towards the plug. Fuel films on the piston crown and the liner are a normal result, and those films emit unburned hydrocarbons and particulates. In a spray-guided system the injector and the plug are close together and the spray itself is aimed at the plug, so the fuel need not be bounced off the piston. That review reports lower particle numbers for spray-guided engines than for wall-guided ones. A chassis-dynamometer comparison of one spray-guided and one wall-guided car found the same split: the wall-guided car produced the higher particulate mass and number. Anders Johansson’s Chalmers thesis on spray-guided stratified engines found that stratified operation can improve efficiency and still does not remove the particulate problem. Particulate mass and number rose with the amount of non-premixed, diffusion-flame combustion, and with fuel that struck the piston. The thesis also cites stratified particulate emissions that may be more than a hundred times those of homogeneous operation on a direct-injection engine. Raza, Chen, Leach and Ding, Energies, 2018; Karavalakis and others, Energy, 2015; Johansson, Chalmers, 2017.

Homogeneous stoichiometric direct injection became the common production compromise. The 2015 chassis-dynamometer paper describes stoichiometric homogeneous wall-guided engines as the dominant share of direct-injection cars then on the market, and notes that stratified spray-guided engines were mostly a European question because lean operation needs its own NOx aftertreatment. That does not mean the particle problem went away. Early injection still wets the bore when the piston is low and the spray is long, and a cold combustion chamber makes the film worse. Euro 6 is why the particle limits in the table apply to direct-injection petrol and not to port injection. The filter that answers them is the petrol particulate filter, below.

What stratification changes

  • Fuel use at light load. Johansson lists four reasons a stratified charge uses less fuel than a homogeneous one: less throttling and therefore less pumping work, less heat lost to the walls because the burned cloud is surrounded by air or exhaust, more air in the cylinder because the fuel can be injected after the inlet valve has shut, and charge cooling as that fuel evaporates. Leonhard’s figure for DI-Motronic was about 10% against port injection; it is his comparison, not a result for every later engine.
  • NOx. The burned cloud is hot and the surrounding charge still contains oxygen. NOx forms, and the exhaust is too lean for a three-way catalyst to reduce it. EGR can cool and dilute the combustion. A NOx storage catalyst can hold the NOx and release it during a short rich interval. Either way, stratified running is no longer a mixture-only problem.
  • Hydrocarbons. Mixture at the edge of the cloud can be too lean to burn. Fuel on the piston does not evaporate in time. Both leave the cylinder as hydrocarbons, and the exhaust temperature of a lean, unthrottled engine can be too low for the catalyst to finish them.
  • Particles. Where the cloud is locally rich, combustion is a diffusion flame and it makes soot. The Chalmers stratified engines showed that link directly: more non-premixed flame, more particulate mass and number. Higher injection pressure and a spray that misses the piston reduce it. They do not make a stratified charge behave like a port-injected one.
  • CO. Overall lean running gives CO less chance than a rich carburettor mixture. CO is not the reason stratified engines grew a second aftertreatment system. NOx and particles are.

Common-rail diesel and post-injection

A common-rail system separates the making of fuel pressure from the moment of injection. The pump holds the rail at pressure, and each injector takes fuel from that rail when the controller opens it. Bosch’s current description of the piezo common-rail system puts the practical point in one sentence: injection timing and quantity are calculated per cylinder, and the system can use several pre-injections and post-injections with short gaps between them. Pressure and timing are no longer locked together by a cam, which is what a distributor pump or a unit injector had to live with. Bosch Mobility, common-rail system with piezo injectors.

Ulrich Dohle, then of Bosch diesel systems, described what the extra injections are for. One or two pre-injections reduce white and blue smoke after a cold start, and they reduce combustion noise. A supplementary injection immediately after the main injection lowers soot. A further injection can regenerate a particulate filter. A 2006 Bosch paper for the US Department of Energy’s DEER conference splits the same jobs by operating region: pilot injection for NOx reduction at higher load and for combustion noise at part load, and a coupled post-injection for soot reduction. The air side of that paper is cooled high-pressure EGR, at rates up to 30% at full load in the duty it was describing, followed by a particulate filter with active regeneration and by SCR. Dohle, quoted in AI Online, November 2003; Parche, DEER 2006.

The two post-injections are not the same event.

  • Close post-injection. This follows the main injection while combustion is still under way. Dohle’s account is that it lowers soot. The extra fuel burns late enough to oxidise soot already formed, and early enough that it still burns in the cylinder rather than being pushed out as raw fuel. Parche calls this the coupled post-injection. It is an engine-out soot control, not a filter control.
  • Late post-injection. This is the regeneration injection. It is late enough that the fuel is not there to make torque. Hydrocarbons leave the cylinder, an oxidation catalyst burns them, and the heat is used to raise the filter temperature until the trapped soot will burn in the remaining oxygen. AECC describes the same step as increasing exhaust temperature through post-combustion so that particulates burn with oxygen, as distinct from the lower-temperature passive reaction with nitrogen dioxide. An injector fitted in the exhaust, upstream of the filter, is another way to supply those hydrocarbons without putting the fuel through the cylinder.

Pilot injection changes the noise and NOx trade-off because it changes the ignition delay. A small early dose starts to burn before the main quantity arrives, so the main combustion pressure rise is less abrupt. Less delay also means less fuel sitting in the chamber when ignition finally happens, which is why Parche associates pilot injection with NOx at high load and with noise at part load. None of these events has one crank angle that suits every engine. The Audi EDC17 examples already on this wiki show how far the hardware diverges: piezo injectors and 1,800 bar in one training system, solenoid injectors and 2,000 bar in another, with different EGR and aftertreatment around them. EDC17 Audi 2.0 TDI system examples.

Particulate filters

Diesel particulate filter

A diesel particulate filter is a wall-flow filter. The channels are blocked at alternate ends, so the exhaust has to pass through the porous wall, and the soot collects as a cake on that wall. AECC lists cordierite, silicon carbide and aluminium titanate as the ceramics used. The same source gives filtration above 95% by mass and above 99% by number, including particles under 100 nanometres, across a wide range of operating conditions. Euro 5 made the filter standard equipment on new diesel cars in Europe; Euro 5, Euro 6 and Euro VI limits are written in both mass and number so that the fine particles are controlled even when the mass looks small.

The filter will block if the soot is only stored. Regeneration burns it. AECC separates two chemical routes. Nitrogen dioxide, made by an upstream oxidation catalyst that raises the NO2 share of the NOx, oxidises soot at ordinary exhaust temperatures. That is passive regeneration. Oxygen oxidises soot only when the exhaust is made hotter for a period. That is active regeneration, and late post-injection is one of the ways to produce the heat. Other methods in the same review are a catalytic coating on the filter itself, a fuel-borne catalyst, an exhaust-mounted fuel injector, and electrical heating.

A fuel-borne catalyst, ceria or an iron additive metered into the fuel, lowers the temperature at which the soot will burn. AECC quotes about 350°C instead of 650°C, and a burn of a few minutes rather than a long active event. The solid residue of that additive stays in the filter as ash. Ash is the part regeneration does not remove. Oil ash and ash-forming fuel additives add to it as well; AECC notes that the filter and its ash capacity have to be designed for them. Soot loading is followed from the pressure difference across the filter and from a model of how the wall’s permeability changes. A pressure reading on its own does not say whether the filter is full of soot, full of ash, or simply seeing a different exhaust flow.

Active regeneration has a thermal limit. AECC flags the drop-to-idle case: soot begins to burn at high temperature and high flow, then the driver lifts off, the flow collapses, and the heat has nowhere to go. The controller’s job during regeneration is to keep that excursion inside what the filter can survive. A regeneration that does not complete is not, by itself, a blocked filter. The Audi training examples distinguish a soot model based on driving and temperature from one based on flow resistance, and they treat a missing sensor or an unsuitable operating state as something that can stop regeneration without proving the filter has failed.

Petrol particulate filter

A petrol particulate filter is the same idea on a spark-ignition engine. German literature calls it an Ottopartikelfilter (OPF); English technical literature usually says GPF. AECC’s account is that it was developed from diesel-filter experience specifically because direct-injection petrol engines produce a higher particle number than port-injected engines. The European Commission’s Joint Research Centre preference, as AECC reports it, is a relatively small filter close to the engine. Petrol exhaust is hot enough that this close-coupled filter regenerates passively most of the time. The penalty is thermal mass in front of, or combined with, the catalyst, which can delay light-off. An underfloor filter is cooler and is more likely to need an active regeneration in urban and stop-start use.

AECC’s 2017 gasoline-filter summary gives the temperatures that matter. At a stoichiometric mixture the filter core temperature that ignites the soot is about 650°C. With a leaner mixture, and therefore more oxygen, that ignition temperature falls to about 500°C and the pressure drop comes down faster. A catalytic coating also helps. Active regeneration, where it is needed, is not a copy of the diesel late post-injection. For an engine that normally runs stoichiometric, AECC expects retarded spark and split injection. For an engine running lean, it expects a post injection. Both are methods already used to heat a catalyst after a cold start. The longer-term arrangement in that paper is a coated filter that does the three-way job as well, with the same precious-metal loading as the catalyst it replaces.

A port-injected engine was left outside the Euro 6 particle limits because its particle numbers are low. Fitting a filter to a direct-injection engine does not change the combustion that made the particles. It catches them, and it has to be able to burn them off in the temperatures and oxygen levels that engine actually produces.

Exhaust-gas recirculation

EGR puts exhaust back into the intake. The DOE primer’s account of why it cuts NOx is the one that matches both engines. The recirculated gas is mostly nitrogen, CO2 and water. It dilutes the charge and raises the heat capacity of what is in the cylinder, so the same heat release produces a smaller temperature rise. The primer describes NOx formation as being held down once peak combustion temperature stays below about 3,100°F. That threshold is the primer’s way of stating the temperature dependence; it is not a switch with a universal set point.

On a petrol engine the exhaust is the leftover of a near-stoichiometric burn, so it is poor in oxygen. A modest EGR rate cools combustion and cuts NOx, which is why the primer can say EGR had already been used on spark-ignition engines for more than twenty-five years by 2010. On a diesel the exhaust still contains oxygen. Dilution only happens if the EGR rate is much higher, particularly at light load where the air excess is largest. Cooling the EGR before it enters the intake raises the mass that can be put through and lowers the combustion temperature again. The primer says most diesels of that period already used cooled EGR.

There are two places to take it from. High-pressure EGR is taken before the turbine and returned to the intake manifold. It responds quickly, and it is hot unless it is cooled. Low-pressure EGR is taken after the turbine and, on a filtered engine, after the particulate filter, then returned upstream of the compressor. The gas is cleaner, which matters for the compressor and the intercooler, and the turbocharger sees the exhaust flow that the high-pressure loop would have stolen. A 2009 dual-loop study on a 2-litre diesel, with cooled high-pressure and cooled low-pressure loops, reported charge temperatures more than 30°C lower and a 24% NOx reduction on the NEDC at constant particulate mass. That is one engine and one test, not a universal gain. The Audi 2.0 TDI training material shows the production version of the low-pressure path: exhaust taken downstream of the particulate filter and routed through the turbocharger. Czarnowski, DEER 2009; Audi EDC17 examples.

Too much EGR, or EGR that is too cold at light load, moves the problem. Combustion slows, hydrocarbons and CO rise, and a diesel makes more soot because the oxygen available to burn it has been diluted away. A. M. Greaney’s Ricardo paper at DEER 2004 says light-load EGR temperature has to be controlled to avoid hydrocarbon, CO and fuel-consumption penalties, which is why cooled EGR systems have a cooler bypass. EGR is a trade, not a valve that is better the more it opens. Greaney, Ricardo, DEER 2004.

Secondary air injection

Secondary air injection is the pump, or the timed valve, that puts fresh air into the exhaust ports or the manifold during a cold start. Workshop language calls the belt-driven American version a smog pump. “Desmog” is the workshop word for removing that pump and its plumbing. The system exists because a cold catalyst does not convert anything, and a cold petrol engine is run rich, so the exhaust leaving the ports is full of CO and hydrocarbons and short of oxygen.

HELLA’s description matches the Bosch cold-start strategy. On a stoichiometric petrol engine a three-way catalyst can convert more than 90% once it is working, but a large share of the emissions of a driving cycle — HELLA says up to 80% — leaves during the cold start, before the catalyst is at about 300 to 350°C. Air added to that rich exhaust lets CO and hydrocarbons react to CO2 and water in the manifold. The reaction is exothermic, so the catalyst reaches light-off sooner. On a car with lambda control, HELLA says the secondary air is switched off after about 100 seconds. A passive system uses the exhaust pulses to draw air through a valve. An active system uses a pump, which is the arrangement that can be controlled rather than merely hoped for. HELLA, secondary air injection.

An MIT study under John Heywood measured the same path from the exhaust port to the catalyst outlet. Secondary air cut hydrocarbons and CO before the catalyst by oxidising them, and the hotter gas shortened catalyst light-off. In that engine the fastest light-off was a mixture 20% rich of stoichiometric with secondary air equal to the engine’s own airflow, and the post-catalyst hydrocarbons were best when the mixed exhaust was at about lambda 1.3. Those ratios are the result of that experiment. They are not a setting for every pump. The useful point is the one the experiment was built to test: the air has to arrive while the exhaust is still hot enough to burn, and it has to be mixed, or it cools the gas instead of burning it. Lee, MIT, 2010.

Once the catalyst is lit and lambda control is closed, the pump is a liability if it keeps running. Extra air in the exhaust looks like a lean mixture to the oxygen sensor, and the controller will richen the engine to chase it. That is why the system is a cold-start device with a shut-off, not a permanent air bleed. Taking it off does not change the full-warm CO and hydrocarbon numbers the MOT measures at 2,500 rpm. It changes the cold-start emissions the catalyst never sees, and on an engine that monitors the system it sets a diagnostic fault. AECC notes that close-coupled catalysts and low-thermal-mass substrates have cut light-off from a minute or two to a few seconds, which is why some later engines do not need a pump. Where the pump is still fitted, it is doing the heating the catalyst cannot yet do for itself.

Oxidation catalysts, NOx storage and urea

The three-way catalyst covers stoichiometric petrol. Everything that runs lean needs a different NOx method, and both fuels need somewhere for the particles if the combustion produces them.

A diesel oxidation catalyst is the older oxidation catalyst in diesel exhaust. It converts CO and hydrocarbons and has little effect on NOx, which is what the oxygen in the exhaust allows. It does two further jobs AECC lists as the reason it is still fitted upstream of everything else. It oxidises hydrocarbons adsorbed on the soot, so the particulate mass falls even before the filter. And it oxidises NO to NO2, which is the oxidant for passive filter regeneration and one of the two nitrogen oxides the fast SCR reaction wants.

A NOx storage catalyst, or lean NOx trap, stores NOx as nitrates while the exhaust is lean. Alkaline-earth materials do the storing. When the store is full, the controller makes a short rich interval and the nitrates release. The released NOx is reduced to nitrogen by CO on the precious metal in the same brick. AECC lists the ways that rich interval is made: intake throttling, EGR, late ignition, injection upstream of the trap, and post-combustion fuel injection. Sulphur occupies the same storage sites. Removing it needs a hotter, reducing desulphation. Ultra-low-sulphur fuel is what made the trap practical; AECC says sulphate storage and release was brought under control by diesel below 10 ppm sulphur, mandatory for European road fuels from 2011.

Selective catalytic reduction uses ammonia, in an exhaust that still contains oxygen. The ammonia is carried as a 32.5% urea solution, AdBlue or diesel exhaust fluid, sprayed into the exhaust and hydrolysed before the catalyst. AECC gives the conversion, once the system is warm and the dose is right, as over 70% and up to 95% of the NO and NO2. The fast reaction uses NO and NO2 together at about one to one, which is why the oxidation catalyst’s NO2 output matters below about 200°C. Vanadium catalysts tolerate sulphur but not the temperatures of a filter regeneration. Copper zeolites work better when cold; iron zeolites work better when hot. The urea dose has to match the NOx. Too little and NOx remains. Too much and ammonia leaves the brick, which is why an ammonia-slip catalyst is coated on the outlet of some systems. AECC’s heavy-duty consumption figures are 5 to 7% of the fuel for a Euro V engine and 2.5 to 6% for Euro VI, depending on load and road. A passenger-car figure depends on how the manufacturer chose to split the work between EGR and SCR, so it is not one percentage.

The arrangements are mixed to suit the engine. A stoichiometric direct-injection petrol engine can use a three-way catalyst and a petrol particulate filter, or a coated filter doing both. A stratified or otherwise lean petrol engine adds a NOx trap or SCR because the three-way brick cannot reduce NOx in lean exhaust. A modern diesel commonly has an oxidation catalyst, a filter, and either a NOx trap, SCR, or both, with high-pressure EGR, low-pressure EGR, or the two together. AECC notes SCR coated onto the filter, close to the engine, as a way to keep the SCR hot in urban driving. Which of those parts is on a particular car is an equipment question. The engine code and the emissions level decide it; the combustion does not.

Sources