`
`177
`
`Urea-SCR in automotive applications
`
`Pa¨ r L.T. Gabrielsson*
`
`Haldor Topsøe A/S, Nymollevej 55, DK-2800 Lyngby, Denmark
`
`Stricter emission legislation for diesel vehicles will make exhaust after-treatment mandatory in the near future. Urea-SCR has
`been chosen for NOx reduction on trucks and busses. Since its effectiveness in reducing NOx is great, and since there is a trade-off
`between NOx emissions and fuel consumption, engines can be adjusted on fuel efficiency. This review article deals with subjects such
`as the description of SCR catalysts, catalytic systems, system performance, choice of reducing agent and durability.
`
`KEY WORDS: urea; SCR; deNox; truck; nox; vanadium; zeolites; urea hydrolysis; modelling.
`
`1. Introduction
`
`Nitrogen oxide is always emitted if fossil or other
`fuels are combusted and vehicles are no exception. NOx
`from gasoline engines is reduced by means of three-way
`catalyst. However, this technology cannot be used on
`diesel engines due to its lean mode of operation. SCR
`has for more than a decade been mentioned as a
`promising technology to reduce NOx on diesel engines in
`automotive applications. The
`legislation has now
`become stringent to the extent that engine management
`alone is not enough to make engines compliant with the
`regulations, and exhaust gas after-treatment in one form
`or another must be introduced. The feature with Urea-
`SCR is summarised in a publication by ACEA [1], the
`European Automobile Manufacturers Association. The
`conclusions drawn are that there is a possibility that fuel
`optimised engines combined with a Urea SCR system
`could reduce fuel consumption by 7% and running cost
`by 3% compared to today’s engine technology. The
`benefit compared to other alternative technologies is
`5%. In a global perspective, transportation could be
`carried out with less CO2 emitted, which is what many
`countries in Europe strive for in order to comply with
`the CO2 levels they signed up to in the Kyoto protocol.
`The unique feature with a Urea-SCR after-treatment
`system is that, since it allows the engine to be optimised
`on fuel consumption, it will have a pay back time on
`a heavy-duty truck. A cost assessment made by Warren
`[2] showed that the payback time for a Urea-SCR
`system could be as low as one to two years in a long
`haul application. This is of course dependent on the
`fuel price, urea price, device price and the mileage
`travelled.
`The technology was firstly developed for NOx reduc-
`tion in stationary application such as boilers, incinera-
`tors and stationary diesel engines. One of the first
`articles describing such a system for automotive use was
`
`* E-mail: pg@topsoe.dk
`
`published by Held et al. [3], who suggested that urea
`should be used as a reducing agent. The technology is
`rather complex and requires a precise control of the
`reducing agent, urea. However, under transient opera-
`tion, it has over the years developed to be mature enough
`to be implemented onboard vehicles. The challenges
`ahead are to reduce size and to develop the low
`temperature performance of the current catalyst types.
`
`2. Technologies and system performance
`
`There are two different routes for applying SCR
`catalysts. The first route is to use one single full catalyst,
`which has a channel wall made of catalytic material. The
`other route is to use wash-coated metallic or cordierite
`substrate. These substrates carry less active amount of
`catalysts and, therefore, possess less low-temperature
`performance than that of full catalysts. However, the
`low temperature performance can be enhanced by
`increasing the relation between NO and NO2 to about
`50/50 over a base metal type of catalyst as reported by
`Anderson [4] and over a zeolitic catalyst as reported by
`Brandin et al. [5].
`Gieshoff et al. [6] compared two systems based on
`coated cordierite substrates. The basic system consisted
`of a hydrolysis catalyst, and an SCR catalyst with or
`without pre-oxidation catalyst. The SCR catalyst had a
`volume of 9.2 L, the pre-oxidation catalyst 2 L and the
`hydrolysis catalyst 4 L. On a 4-L diesel engine they
`found that the activity was enhanced when a pre-
`oxidation catalyst, containing 90 g Pt/ft3, was used.
`Results from ESC test cycles showed that the overall
`conversion efficiency could be increased by 30%, when a
`pre-oxidation catalyst was used with the same SCR
`catalyst volume.
`Walker et al. [7] demonstrated a system consisting of
`a pre-catalyst, an SCR-catalyst and an ammonia slip
`oxidation catalyst, but also combined with a soot filter.
`The total volume of the system active in the SCR
`reaction was in the range of 36 L. The system showed a
`
`1022-5528/04/0400–0177/0 Ó 2004 Plenum Publishing Corporation
`
`Exhibit 2026.001
`
`
`
`178
`
`P.L.T. Gabrielsson/Urea-SCR in automotive applications
`
`rather good low-temperature performance, enhanced by
`the pre-oxidation catalyst.
`Gekas et al. [8] compared the performance over three
`different catalytic systems, pure SCR systems, with slip
`oxidation catalyst and with a pre-oxidation catalyst. By
`using a pre-oxidation catalyst,
`they measured an
`enhanced low-temperature performance of the catalyst,
`but the effect was limited, and by replacing the pre-
`oxidation catalyst with the same volume of SCR
`catalyst, they reached similar conversion efficiencies,
`but with a less complex system. Gekas et al. [8] also
`showed an improved catalyst performance when the cell
`density was increased from 130 to 300 cpsi leading to the
`fact that the catalyst volume could be reduced by 1/3 for
`the same conversion efficiency.
`Gekas et al. [9] demonstrated that a 12-L class, 400
`horsepower engine, which was optimised on low fuel
`consumption and low soot particle emissions, could be
`compliant with future emission legislation, with 20 L of
`a 300 cpsi SCR catalyst alone (see Table 1).
`Lambert et al. [10] tested an SCR system, consisting of
`a combination of base metals and zeolite type of catalyst,
`on a passenger car engine and found that the SCR system
`could reduce, on average, 83% NOx in a US federal test
`cycle. This was carried out with an ammonia slip, which
`actually was lower than that for gasoline cars with three-
`way catalysts reported in other literature.
`A graphical description of the different catalytic set-
`ups is seen in figure 1.
`
`2.1. Reducing agents
`
`Different ammonia precursors or reducing agents
`have been suggested. Liquefied NH3 has been widely
`used in larger stationary installations, and has also been
`suggested for the use on vehicles as described by Funk
`et al. [12]. However, the safety is an issue when such a
`reducing agent is handled onboard vehicles, and lethal
`damages cannot be excluded since a rather large amount
`will need to be transported around to gas stations and
`also onboard the vehicles, where it is finally used.
`Another suggested reducing agent
`is ammonium
`carbamate NH4NH2COO, which upon heating decom-
`poses into urea (NH2)2CO and H2O. In support of
`ammonium carbamate is that it is a solid which reduces
`the volume of the reducing agent which is needed
`onboard a vehicle compared to urea which is normally
`dissolved in water to give a 32.5% urea solution. The
`disadvantage of a solid reducing agent is that the
`injection system tends to be rather complex. It was
`suggested by Stieger and Weisweiler [13] that ammo-
`nium carbamate is dosed as powder in a bath of oil, in
`which it decomposes first to urea and later to NH3. The
`ammonia gas was led into the exhaust gas manifold
`upstream of the SCR catalyst. Weisweiler and Buchholz
`[14] even suggested the use of solid urea, which is
`decomposed to reactive species by heat. There are a
`couple of drawbacks with such a system, one is that
`solid urea is hygroscopic and needs to be protected
`
`Table 1
`Urea-SCR system conversion efficiency over European test cycles, ETC (European transient test cycle) and ESC (European stationary test cycle),
`with a 12 litre class 400 hp heavy-duty diesel engine and 20 litres of Urea-SCR catalyst [9]. ‘‘EU V’’ referrers to emission legislation which comes
`in to force in 2008
`
`Particulate matter (g/kWh)
`
`Test-cycle
`
`NOx (g/kWh)
`
`HC (g/kWh)
`
`CO (g/kWh)
`
`Insoluble
`
`Lube oil
`
`Fuel oil
`
`BSFC (g/kWh)
`
`EU V limit
`
`ESC w/o SCR
`ESC w SCR
`ETC w/o SCR
`ETC w SCR
`
`2
`
`9.0
`1.4
`8.5
`1.5
`
`0.28
`0.05
`0.26
`0.04
`
`0.30
`0.48
`0.38
`0.61
`
`0.02 ESC
`0.03 ETC
`0.007
`0.009
`0.015
`0.013
`
`0.020
`0.003
`0.014
`0.002
`
`0.010
`0.004
`0.010
`0.005
`
`194
`194
`198
`197
`
`Urea injection
`
`Urea injection
`
`Flow
`
`a
`
`SCR-
`cat
`
`c
`
`Flow
`
`Ox-
`cat
`
`SCR-
`cat
`
`Slip-
`cat
`
`Urea injection
`
`Urea injection
`
`b
`
`Flow
`
`SCR-
`cat
`
`Slip-
`cat
`
`d
`
`Flow
`
`Ox-
`cat
`
`UH-
`cat
`
`SCR-
`cat
`
`Slip-
`cat
`
`
`
`Figure 1. Different catalytic set ups: (a) single SCR catalyst [9]; (b) SCR and NH3 slip oxidation catalyst [9]; (c) pre-oxidation, SCR and NH3 slip
`oxidation catalyst [7]; (d) pre-oxidation, urea-hydrolysis, SCR and NH3 slip oxidation catalyst [6,11]; Slip cat ¼ ammonia oxidation catalyst, Ox-
`cat ¼ oxidation catalyst, UH-cat ¼ urea hydrolysis catalyst.
`
`Exhibit 2026.002
`
`
`
`P.L.T. Gabrielsson/Urea-SCR in automovite applications
`
`179
`
`against humidity, another is that an even distribution of
`powder in the exhaust manifold could be difficult.
`Currently, the preferred choice is a solution of 32.5%
`urea in water, as described in a recent DIN norm. This is
`going to be the standard in Europe, when an infrastruc-
`ture for the reducing agent is developed according to
`ACEA [15] which is also about to standardise the urea
`filling nozzle. Fang and DaCosta [16] studied the urea
`decomposition and found by mixing catalyst with
`urea and heating the mix above the melting point of
`urea, 135 °C, that it was possible to decompose urea
`with a higher rate when the urea catalyst mix was used.
`According to Fang the decomposition can be made in
`two stages of which the first stage starts after the melting
`point and the second above 300 °C. Melamine com-
`plexes are formed in the second stage, which could
`inhibit the catalytic reaction. Fang et al. also found that
`the nature of the urea spray had an influence on
`deposition in the front of the catalyst. A finer atomisa-
`tion gave less deposition in the front of the catalyst.
`Larrubia et al. [17] studied the adsorption of urea on a
`V2O5–MoO3–TiO2 catalyst by FTIR and proposed a
`mechanism for the decomposition on the surface.
`According to them, the mechanism should be adsorp-
`tion of an anionic urea species which further decom-
`poses to ammonia and ammonium species and cyanate
`)
`anions (N@C@O)
`. The cyanate anion can be further
`decomposed to ammonia and CO2 by hydrolysis with
`water. Ball [18] investigated the potential toxicity of urea
`and its degradation products and found no reasons for
`concern. However, since data for some compounds are
`poor, further investigations were recommended. Koebel
`et al. [19] compared performance for different reducing
`agents and found differences in the ability to reduce
`NOx at the same stoichoimetric ratio. The differences
`were most pronounced at high space velocities, where
`liquefied ammonia showed the highest activity.
`
`2.2. Catalysts
`
`2.2.1. Different types of monoliths for SCR catalysts
`As mentioned above there are two different families
`of SCR monoliths, full or coated. The full catalyst could
`either be an extruded or corrugated structure. The
`extruded catalyst normally consists of about 700–1000 g/
`L of active material, while the corrugated structure is in
`
`the range of 450–550 g/L. A coated monolith is based on
`a cordierite or a metallic substrate, which is washcoated
`with about 150–200 g of the active phase.
`Koebel et al. [20] compared extruded catalysts, 300
`cpsi, with coated catalysts of 400 cpsi, and found that
`a coated catalyst with 3% V2O5 on a cordierite
`substrate could have higher than or similar activity to
`an extruded catalyst. He also found that an extruded
`catalyst stored more ammonia than a coated catalyst
`at low temperatures, while it stored similar or less at
`high temperatures.
`Kleeman et al. [20] investigated the ammonia adsorp-
`tion capacity over an extruded and coated catalyst based
`on the same catalytic material, and found that the
`adsorption capacity was fairly similar for all coatings.
`However, there is a tremendous difference in the amount
`of active material in the two different types of catalysts.
`Since they carry different amounts of active material
`they also have different properties, see table 2.
`The benefit of a high ammonia adsorption capacity is
`that it has a buffering property, it continues to reduce
`NOx for several minutes after the injection is switched
`off as described by Kleeman et al. [21]. This means that
`the requirement for precision of reducing agent injection
`is lower than if a coated catalyst is used. However, the
`disadvantage of an extruded catalyst is that it is more
`difficult to control the ammonia slip if there is a sudden
`temperature increase in the system, because more
`ammonia is desorbed from the surface and this causes
`an ammonia slip.
`
`2.2.2. Activity comparison
`The different catalysts described above possess dif-
`ferent activity in different temperature areas. The coated
`substrate has normally lower activity in the low tem-
`perature region from 170 to 250 °C than that of the full
`catalysts. The reason is that the reaction is more or less
`kinetically controlled, and therefore the activity is more
`or less dependent on the amount of catalyst. There are
`also differences between extruded and corrugated full
`catalysts. The extruded catalyst has a denser structure,
`which originates from the extrusion process, compared
`to the corrugated structure. This means that in an
`extruded catalyst the reaction is more limited by pore
`diffusion. However, in the high temperature area from
`300 °C and upward, the reaction is more controlled by
`
`Table 2
`NH3 storage capacity for different types of SCR catalyst [21]
`
`Catalyst type
`
`Cpsi
`
`Catalyst WO3
`
`V2O5
`
`NH3 capacity (mg/g)
`
`NH3 capacity (mg/L)
`
`(g/L)
`
`(%)
`
`(%)
`
`200 °C
`
`350 °C
`
`200 °C
`
`350 °C
`
`Coated
`Coated
`Extruded
`
`400
`400
`300
`
`178
`203
`955
`
`8
`0
`10
`
`3
`4
`3
`
`2.02
`2.00
`1.78
`
`0.78
`0.75
`0.46
`
`360
`406
`1700
`
`139
`152
`439
`
`Exhibit 2026.003
`
`
`
`180
`
`P.L.T. Gabrielsson/Urea-SCR in automotive applications
`
`film mass transfer, and is therefore more dependent on
`the exposed surface area in channels within the mono-
`lith.
`It is sometimes difficult to compare different catalysts.
`The reason is that the experimental conditions differ.
`Sometimes monolithic materials are used and other
`times fine powder in a packed bed of the same type of
`catalyst is used. Between those two materials, there is an
`immense difference in,
`for
`instance, mass
`transfer
`constants, while the powder catalyst more or less has
`an efficiency factor close to 1. The kinetics in a monolith
`are normally more or less outer film mass transfer
`limited at temperatures above 300 °C, while a powder is
`more or less kinetically controlled or controlled by pore
`diffusion.
`An easy way to scale and compare the activity over a
`monolithic catalyst is to calculate a first order rate
`constant based on space velocity, KNHSV.
`
`KNHSV ¼ lnð1 xÞ F
`Vcat
`where x is the NOx conversion, F is the flow in N m3/h
`and Vcat the catalyst volume in m3. KNHSV is expressed
`)1.
`in the unit, N h
`In table 3, the first order rate constant has been
`calculated. The pattern emerging from the data is that
`coated catalyst, as mentioned above, possesses a lower
`low-temperature performance and needs a higher level
`of NO2 in the gas to reach the activity levels of the full
`catalysts. It is also possible to observe a pattern on
`which level the rate constant should be at the two
`temperatures, 200 and 350 °C, in order to reduce NOx in
`a test cycle with a reasonable volume of catalyst. The
`)1
`rate constants should in this case be above 25,000 N h
`)1at 350 °C. Apply-
`at 200 °C and above 130–150,000 N h
`ing a pre-oxidation catalyst upfront of the urea-SCR
`catalyst and its urea injector enhances the rate constant
`
`of the coated type of catalysts. With rate constants
`above these levels it will be possible to achieve about
`80% NOx conversion in a European Transient Test
`Cycle, which is used for certification in the emerging
`European market. However, it should be pointed out
`that there are other markets, such as in the US, where
`certification test cycles run at lower temperatures and,
`therefore, these might need an even higher conversion
`efficiency at lower temperatures than described above.
`The above calculations are a textbook example that
`should not be forgotten. It is apparent that many
`authors do not know the boundaries of the applicability
`of a catalyst, and that this should be used as guideline
`for what to look for in their research and to judge the
`importance of their results.
`
`2.2.2.1. Experimental conditions
`Having this opportunity one would also like to urge
`experimentalists in this area to be very careful with their
`experimental conditions. It is for instance quite surpris-
`ing to see that some authors are running experiments
`without water. Such results can only be used for relative
`comparisons within the same research and it is difficult
`to estimate their usefulness in real life. An example of a
`real exhaust gas composition is given by Roudit et al.
`[26] and it could be regarded as an input for exper-
`imental conditions, see table 4.
`The influence of SO2 and CO2 on the catalyst could be
`debated. A more reasonable goal for a synthetic gas mix
`could have less SO2 in the gas, 1–2 ppm seems more
`reasonable, and would be in line with the emissions from
`future engines in Europe. However, one should bear in
`mind that future vehicles occasionally might need to go to
`a country where low sulphur fuel is unavailable, and
`therefore the catalyst might see considerably higher levels
`of sulphur than it normally does. This, sulphur tolerance,
`could be an issue, which at least for very new and
`
`Table 3
`Comparison of first order rate constant between different catalyst types
`
`)1)
`KNHSV (N h
`
`200 °C
`NO
`
`~51,300
`
`~3580
`
`~15,000
`
`~23,200
`
`~4200
`
`~3500
`
`~8200
`
`200 °C
`NO/NO2
`
`~98,200
`
`~64,500
`
`~28,300
`
`N/A
`
`~34,300
`
`N/A
`
`350 °C
`NO
`
`~33,700
`
`~109,537
`
`~182,000
`
`~182,341
`
`N/A
`
`~64,377
`
`N/A
`
`References
`
`Blakeman et al.
`[22]
`Blakeman et al.
`[22]
`Koebel et al.
`[23]
`Koebel et al.
`[23]
`Gieshof et al.
`[6]
`Winkler et al.
`[24]
`Bukart et al.
`[25]
`
`Material
`
`Active phase
`Monolith type
`
`‘‘Low temp.’’
`400 cpsi, coated
`‘‘High temp.’’
`400 cpsi, coated
`V2O5–WO3/TiO2
`400 cpsi, coated
`V2O5–WO3/TiO2
`300 cpsi, extruded
`V2O5/WO3/TiO2
`400 cpsi, coated
`Not specified
`400 cpsi, coated
`4.6% V2O5/TiO2
`400 cpsi coated
`
`Exhibit 2026.004
`
`
`
`P.L.T. Gabrielsson/Urea-SCR in automovite applications
`
`181
`
`unknown materials should be investigated. Another gas
`component, of which the presence is not negotiable, is
`water because the activity and selectivity are affected by it.
`Odenbrand et al. [27] made experiments with and
`without water and found that the selectivity of formation
`of N2O was heavily influenced by the presence of water.
`The temperature window for NOx reduction changed to
`higher temperatures when 1% water was added to the
`stream, having as a result that the temperature had to be
`increased by 15–20 °C in order to obtain the same NOx
`conversion with water and without.
`Hydrocarbons are another component with which
`one should be careful. Gieshoff et al. [6] tried different
`levels of n-decane in the gas mix and found that the
`activity window below 300 °C was changed by 50 °C to
`the higher end, when 10 ppm of n-decane was used
`which is 100 ppm C1. Since different hydrocarbons have
`different effects, it is suggested that the hydrocarbons
`are excluded in laboratory experiments, or that at least a
`reference comparison with and without hydrocarbons is
`made.
`
`2.2.3. Catalytic material
`Before starting to work in the field of SCR, one has to
`read an article by Bosch and Jansen [28]. The article is
`an extensive review of the area of SCR. Among other
`things it deals with topics like formation of NOx,
`catalyst preparation and testing and kinetic mecha-
`nisms. It is also a source of inspiration not only about
`how to carry out experiments and prepare catalyst but
`also because many different catalytic materials are listed
`and referenced.
`
`2.2.3.1. Base metal oxide catalysts
`Base metal oxide catalysts are still the most common
`commercial types of catalyst for SCR and the predom-
`inant active ingredient is vanadium on a carrier of
`titanium oxide. Tungsten or molybdenum is used to
`increase the catalyst acidity. For the temperature sta-
`bility a catalyst with WO3 or MoO3 holds a higher
`activity and is active in a broader temperature window,
`according to Forzatti et al. [29], due to the formation of
`more acid Brønsted sites and OH-groups on the surface.
`These are the sites with which NH3 coordinates and
`becomes activated before it reacts with NOx from the
`gas phase. Koebel et al. [30] compared three different
`vanadium-based catalysts, extruded 200 and 300 cpsi
`and coated 400 cpsi monoliths and found that the 300
`
`and 400 cpsi catalysts exhibited the best activities. He
`also found that the 400 cpsi catalyst had a sharper
`boundary for the appearance of ammonia slip. Madia
`et al. [31] tried the thermal stability of coated catalyst
`with 173 g catalyst/L on a 370 cpsi substrate. The
`catalyst was based on a WO3–TiO2 carrier, which was
`impregnated with 1, 2 and 3% V2O5. Three ageing
`‘‘100 h @ 550 °C’’,
`procedures
`were
`tested,
`‘‘100 h @ 550 °C + 30 h @ 600 °C’’ and ‘‘100 h @ 550
`°C + 30 h @ 600 °C + 15 h @ 650 °C’’. They found that
`the 3% V2O5 catalyst was most sensitive to high
`temperature, and its conversion efficiency went down
`during all the different ageing procedures. However, the
`2% V2O5 catalyst remained stable up to 600 °C but lost
`activity going to 650 °C. The 1% V2O5 catalyst gained
`activity in the two first ageing procedures, but its initial
`activity was low in relation to the other two catalysts.
`The more dramatic effect on the 3% V2O5 catalyst at
`650 °C was due to loss of anatase surface in the TiO2
`structure, which was induced by the vanadium loading.
`The vanadium-based system is not very active below
`200 °C. A catalyst with low temperature performance
`that could be of interest is the MnOx–CeO2 system,
`which recently was described by Qi et al. [32]. Except for
`its low-temperature activity, the catalyst has robustness
`to high temperatures and is reversibly deactivated, when
`sulphur is used in the gas. The drawback with low-
`temperature active catalysts is that they are too active at
`higher temperatures and therefore burn off the ammonia
`instead of reducing NOx.
`
`2.2.3.2. Zeolitic catalysts
`There are only few articles on zeolites and urea SCR.
`Two of those describe the use of zeolites in an engine
`application, Gieshoff et al. [33] and Lambert et al. [34],
`where the catalyst shows a good performance. However,
`the long-term stability and the sulphur tolerance are not
`discussed in these papers.
`Sullivan et al. [35] compared a Cu-ZSM5 zeolite with
`traditional vanadium-based catalyst and found that the
`Cu-ZSM5 catalyst was active in a broader temperature
`window. One can have some concern about the exper-
`imental conditions, since they were carried out with
`0.6% NH3 and NOx in the gas, which could be regarded
`as rather high. They also saw a promoting effect when
`water was used, which goes in the opposite direction to
`what is expected. The conversion efficiency was en-
`hanced at low temperatures when water was added to
`
`Table 4
`Diesel exhaust gas composition according to Roudit et al. [26]
`
`P
`
`(kW)
`
`890
`
`NO
`
`NO2
`
`HC1
`
`CO
`
`(ppm)
`
`(ppm)
`
`(ppm)
`
`(ppm)
`
`1020
`
`115
`
`30
`
`490
`
`CO2
`
`(%)
`
`7,1
`
`O2
`
`(%)
`
`10,5
`
`H2O
`
`(%)
`
`4–5
`
`SO2
`
`(ppm)
`
`N2
`
`30
`
`Bal.
`
`Exhibit 2026.005
`
`
`
`182
`
`P.L.T. Gabrielsson/Urea-SCR in automotive applications
`
`the gas. This might be due to the massive NH3
`concentration. The durability of the catalyst is not
`discussed. Lifeng et al. [36] investigated a Cu-type of
`zeolite. The structure is not specified since it has been
`produced by one of the established catalyst producers in
`the car industry. The catalyst, without a pre-oxidation
`catalyst, possesses a rather modest activity at 200 °C;
`)1 although the activity rises
`KNHSV of about 4500 h
`about
`four times when a pre-oxidation catalyst
`is
`applied up in front of the SCR catalyst. Lifeng et al.
`also found out that the reaction was inhibited by NO
`adsorption on the Cu sites. The effect was most
`pronounced below 250 °C and a sophisticated engine
`operating strategy was needed to reduce its influence on
`the overall activity. The catalysts were shown to be
`sensitive to sulphur, but the effect was less pronounced
`when a pre-oxidation catalyst was used. It was found
`that the zeolite possessed a rather large NH3 adsorption
`capacity, which according to Liefeng et al. would have a
`buffering effect in transient operation. Piehl et al. [37]
`prepared and tested a VO-ZSM catalyst, which relative
`to a European reference catalyst possessed a rather high
`activity. The catalyst was tested for durability in both
`wet and sulphur containing gases, but even water had a
`negative long-term effect on the catalyst and its appli-
`cability must be regarded as rather low.
`One can be quite pessimistic about the use of zeolites
`for several reasons:
`
`(i) The catalyst types have been tested in the SCR
`reaction since the eighties without a commercial
`breakthrough.
`(ii) Zeolites have been extensively tested in the so-called
`HC-SCR reaction with insufficiently long-term
`stability.
`types of
`(iii) The industrial availability of different
`zeolites is quite poor. There are only a few com-
`mercialised zeolites on the market as it is.
`
`2.3. Durability
`
`The SCR catalyst based on V2O5 on TiO2 is known
`for its robustness and it has been used for more than two
`
`decades in stationary applications in quite harsh envi-
`ronments such as with 2000 ppm SO2 in the gas and a lot
`of ash in the flue gas. Use of up to 50,000 h is not
`uncommon in stationary applications. One could say
`that the vehicle application offers a milder environment
`than for instance a boiler application. However, the
`vehicle application offers more vibrations and, further-
`more, more starts and stops, which could affect the long-
`term stability. Therefore, it is very important to dem-
`onstrate a full catalyst life in a vehicle. Ammon and
`Keefe [38] followed more than 20 vehicles in a test in
`which some of the vehicles accumulated more than
`300,000 miles or 480,000 km. In table 5 results are shown
`from tests with fresh and used SCR catalysts. It shows
`that the overall conversion efficiency in the engine test
`cycles used for certification of engines on the European
`market, is similar after more than 200,000 km of use.
`The emission warranty of an SCR catalyst is suggested
`by the European Commission to be 500,000 km and
`table 4 shows performance after half
`this life with
`encouraging results.
`Ammon and Keefe [38] also studied the long-term
`stability using a bench scale reactor and samples aged on
`the road (see figure 2). What they found was that the
`conversion efficiency decreased by about 10%, com-
`pared to the fresh catalysts over more than 300,000 miles
`or 480,000 km.
`Fritz et al. [39] made micro reactor tests of catalysts
`aged for 94,000 miles on the road and these showed only
`a modest reduction in conversion efficiency over time.
`
`3. Modelling
`
`3.1. Different model types
`
`Roudit et al.
`[26] developed a three-dimensional
`model, which takes into account not only the influence
`of the turbulent inlet flow, hydrodynamic effect, but also
`the role of a real exhaust. The use of real exhaust in the
`exhaust makes the model applicable to real exhaust gas
`conditions and it was stated that the reaction rate was
`enhanced, but others have come to the opposite con-
`clusion. The model was developed around an extruded
`
`Table 5
`NOx reduction performance of fresh and aged SCR catalysts in Engine Test Bench measurement using ETC and ESC
`test cycles (Ammon and Keefe [38])
`
`Test cycle
`
`System no.
`
`Mileage (km)
`
`NOx in (g/kW h)
`
`NOx out (g/kW h)
`
`NOx red. (%)
`
`ESC
`
`ETC
`
`DC B2
`DC B2
`DC B3
`DC B3
`
`DC B2
`DC B2
`DC B3
`DC B3
`
`0
`240,000
`0
`272,000
`
`0
`240,000
`0
`272,000
`
`9.37
`9.63
`9.74
`9.63
`
`8.58
`7.94
`7.18
`8.27
`
`3.60
`3.77
`3.53
`3.62
`
`2.55
`2.38
`2.01
`2.25
`
`61.6
`60.8
`63.7
`62.4
`
`70.3
`70.0
`72.0
`72.8
`
`Exhibit 2026.006
`
`
`
`P.L.T. Gabrielsson/Urea-SCR in automovite applications
`
`183
`
`is
`it
`coated substrate and a pre-oxidation catalyst
`possible to comply with future emission legislation.
`Urea will be used as reducing agent in Europe and urea
`injection control is a key issue when urea-SCR is applied
`on a diesel vehicle. Urea control can be made by real
`time kinetic calculation or by urea injection maps. Urea
`decomposes in the gas phase or on the catalyst forming
`cyanic anions, which can be further hydrolysed into
`ammonia and CO2, or ammonia. Durability has been
`demonstrated for more than half the full
`life of a
`catalyst.
`
`References
`
`[1] ACEA Report on Selective Catalytic Reduction, The most
`promising Technology to comply with imminent Euro IV and
`Euro V emissions standards for HD engines, Final Report, 23/6/
`2003.
`[2] J.P. Warren, SAE Technical Paper 2001-01-1948, 2001.
`[3] W. Held, A. Koenig and T. Richter, SAE Technical Papers 900496.
`[4] Lars H. Andersson, Licentiate Thesis 1989, Selektiv katalytisk
`reduktion av kva¨ veoxider, University of Lund, LUTKDH/
`TKKT-1008/1-65/1989.
`[5] Jan G.M. Brandin, Lars A.H. Andersson and C.U.I. Odenbrand,
`Catal.Today 4 (1989) 187.
`[6] J. Gieshoff, A. Sha¨ fer-Sindlinger, P.C. Spurk, J.A.A. van den
`Tillaart and C. Garr, SAE Technical paper 2000-01-0189.
`[7] A.P. Walker, R. Allansson, P.G Blakeman, M. Lavenius, S.
`Erkfeld, H. Landa¨ lv, B. Ball, P. Harrod, D. Manninger and L.
`Bernegger, SAE Technical paper 2003-01-0778.
`[8] I. Gekas, P. Gabrielsson, K. Johansen, L. Nyengaarda and T.
`Lund, SAE Technical paper 2002-01-0289.
`[9] I. Gekas, P. Gabrielsson, K. Johansen I. Bjørn, J. Husted-Kjæer,
`W. Reczek and W. Cartellieri, SAE Technical paper 2002-01-
`2885.
`[10] C. Lambert, J. Vanderslice, R. Hammerdale and R. Belaire. SAE
`Technical paper 2001-01-3623.
`[11] E. Jacob, Berichte zur Energie und Verfahrenstechnik, Heft 99.1,
`1999, 76.
`[12] A. Funk, B. Mauer and G. Huethwohl, EP1215373.
`[13] D. Stieger and W. Weisweiler, Chem. Ing. Tech. 73 (1–2) (2001)
`123.
`[14] W. Weisweiler and F. Buchholz, Chem. Ing. Tech. 73 (7) (2001)
`882.
`[15] ACEA Statement on the Adoption of SCR Technology to Reduce
`Emissions Levels of Heavy-Duty Vehicles 30/06/2003.
`[16] H.L. Fang and H.F.M. DaCosta, Appl. Catal. B: Environ., In
`press.
`[17] Larrubia et al. App. Catal. B: Environ. 27 (2000) L145.
`[18] J.C. Ball, SAE Technical paper 2001-01-3621.
`[19] M. Koebel, M. Elsner and M Kleeman, Catal. Today 59 (2000)
`335.
`[20] M. Koebel, M. Elsener and G. Madia, SAE Technical paper 2001-
`01-3625.
`[21] M. Kleemann, M. Elsner, M. Koebel and A. Wokaun, Appl.
`Catal. B: Environ. 27 (2000) 231.
`[22] P.G. Blakeman, G.R. Chandler, G.A. John and A.J.J. Wilkins,
`SAE Technical paper 2001-01-3624.
`[23] M. Koebel, M. Elsener and G. Madia, SAE Technical paper 2001-
`01-3625.
`[24] C. Winkler, P. Flo¨ rchinger, M.D. Patil, J. Gieshoff, P. Spurk and
`M. Pfeifer, SAE Technical paper 2003-01-0845.
`[25] A. Burkardt, W. Veisweiler, J.A.A. Van den Tillart and A. Scha¨ fer-
`Sindlinger, E.S. Lox, Top. Catal., Vols. 16/17 (1–4), 2001, 369.
`
`0
`
`100000
`
`200000
`Mileage (miles)
`
`300000
`
`400000
`
`1
`
`0.9
`
`0.8
`
`0.7
`
`0.6
`
`0.5
`
`Relative NOx conversion activity
`
`Figure 2. On road aged SCR catalyst checked for activity on a bench
`scale reactor at 320 °C after different driving mileages [38].
`
`SCR catalyst with a wall thickness of 0.78 mm. With
`such a large wall thickness, a three-dimensional model
`becomes more important. Gieshoff and co-workers [24]
`developed a less complicated one-dimensional model.
`As model catalyst they used a coated 400 cpsi and
`4 ml substrate, with a washcoat amount of about
`180 g/L. Kinetic parameters from the literature are used
`and fitted for the model. The model gives a reasonable
`fit for different space velocities. However, it overesti-
`mates the reaction rate in the temperature range 225–
`350 °C at higher space velocities, where the deviation in
`the first order rate constant is as much as 15–20%. The
`reason why this simplified model gives an accurate result
`is that the washcoat layer is rather thin, about 50 lm,
`compared to the extruded catalyst above, which means
`that the influence from the diffusion resistance within
`the wall
`is rather low on the overall reaction rate.
`Another more pragmatic approach was used by An-
`dersson et al. [40] who, in order to develop a control
`strategy for injection of the reducing agent, used a
`structured model with lumped kinetic parameters. The
`included four reactions, NOx reduction, NH3
`model
`adsorption, NH3 desorption and NH3 oxidation. The
`kinetic parameters were fitted by multiple non-linear
`regression. NH3 adsorption and desorption parameters
`were fitted using results from TPD experime