Search

We propose a new modeling strategy to build efficient neural network representations of chemical kinetics. Instead of fitting the logarithm of rates, we embed the hyperbolic sine function into neural networks and fit the actual rates. We demonstrate this approach on two detailed surface mechanisms: the preferential oxidation of CO in the presence of H2 and the ammonia oxidation under industrially relevant conditions of the Ostwald process. Implementing the surrogate models into reactor simulations shows accurate results with a speed-up of 100 000. Overall, the approach proposed in this work will significantly facilitate the application of detailed mechanistic knowledge to the simulation-based design of realistic catalytic systems. We foresee that combining this approach with neural ordinary differential equations will allow building machine learning representations of chemical kinetics directly from experimental data.

Gasoline particulate filters (GPFs) are an effective means of reducing particle emissions from gasoline direct injection (GDI) engines. This paper explores the filtration efficiency and pressure drop of three GPFs with varying in-wall washcoat volumes through three parameters: particle size, flow rate of gas through the GPF, and amount of soot accumulation. The three GPFs are tested on a custom filtration efficiency test rig with a miniature inverted soot generator used as a particle source at three space velocities of: 32, 000, 97, 000, and 161, 000 h−1. A scanning mobility particle sizer (SMPS) is used to measure the filtration efficiency and amount of soot accumulated on the GPF. The GPFs are loaded with soot until the average filtration efficiency reached > ~ 98%. Results show that the filtration efficiency is a function of particle size, soot loading, and space velocity. With increasing soot loading on the GPF, the filtration efficiencies increase significantly and become less sensitive to mobility diameter. Filtration efficiency is generally found to decrease with increasing space velocity. For clean filters, the GPF without washcoat has the same or higher filtration efficiency than the GPFs with washcoat. At high amounts of soot accumulation, the GPFs with a washcoat have the highest filtration efficiency. The coated filters have a much higher rate of filtration efficiency and pressure drop increase compared to the bare filter. The mechanisms leading to the observed trends in filtration efficiency and pressure drop are also discussed.

In the present study, the pore space of a mesoporous cerium oxide material is investigated, which forms by the self-assembly of primary particles into a spherical secondary structure possessing a disordered mesopore space. The material under study exhibits quite stable mesoporosity upon aging at high temperatures (800 °C) and is, thus, of potential interest in high-temperature catalysis. Here, different characterization techniques were applied to elucidate the structural evolution taking place between heat treatment at 400 °C and aging at 800 °C, i.e., in a water-containing atmosphere, which is usually detrimental to nanoscaled porosity. The changes in the mesoporosity were monitored by advanced physisorption experiments, including hysteresis scanning, and electron tomography analysis coupled with a 3D reconstruction of the mesopore space. These methods indicate that the 3D spatial arrangement of the primary particles during the synthesis under hydrothermal conditions via thermal hydrolysis is related to the thermal stability of the hierarchical mesopore structure. The assembly of the primary CeO2 particles (∼4 nm in size) results in an interparticulate space constituting an open 3D mesopore network, as revealed by skeleton analysis of tomography data, being in conformity with hysteresis scanning. At elevated temperatures (800 °C), sinter processes occur resulting in the growth of the primary particles, but the 3D mesopore network and the spherical secondary structure are preserved.

Both a mechanistic model, which accounts for ammonia, nitrate, and ammonium nitrate storage, and a global model, which only accounts for ammonia storage and no nitrates, were calibrated to the same data collected across a Cu-CHA catalyst. Once calibrated, the models were directly applied to simulate various driving cycles (cold WHTC, hot WHTC, FTP, NRTC, ETC, and WNTE), with different catalyst layouts (i.e., washcoat loading and cpsi) as part of the model validation process. It was demonstrated that the mechanistic model provides a notable improvement in the prediction of the NOx conversions for colder cycles (i.e., cold WHTC, hot WHTC, and FTP) due to its inclusion of nitrates and ammonium nitrate. The quality of prediction of the hotter cycles (i.e., NRTC, ETC, and WNTE) was close to the same for both models, since no ammonium nitrate accumulates at these higher temperatures. Despite the additional reactions and surface species included in the mechanistic model, no significant increase in simulation time is observed compared to the global model. Finally, the mechanistic model is applied to calibrate a simple NH3 dosing strategy. The dosing strategy is compared to a constant NH3/NOx approach, where it was shown through simulations, and confirmed experimentally, that the dosing approach allows for an increase in NOx conversion. Overall, this demonstrated that the model accurately predicts the increase in NOx conversion and that the model is a valuable tool to establish the catalyst’s true potential during driving cycles.

This paper presents a detailed study of the fluid dynamics inside a wall-flow filter and proposes a new pressure drop model. A 3D channel scale computational model of a filter validated with experiments is used. A detailed description of the pressure drop for flow entering, passing through and leaving the filter is provided. The computational grid is extensively analyzed, and it is found that wall-flow is very insensitive to the grid quality, opposite to the local pressure, which is very sensitive. Several flow rates and wall permeability are analyzed. The most critical assumptions commonly found in current models are discussed based on the results. It is found that the friction factor of the channels is non-constant, it is different for the inlet and the outlet channels, and both differ from that for pipes with non-porous walls. A new criterion to determine the flow inside the filter as fully developed is also presented. The wall-flow along the perimeter of a cross-section is observed to be variable, consistently for many flow rates and wall permeability. The results are also used to develop a comprehensive, physically based, pressure drop model that shows very good agreement with experimental data. It is found that the propagation error when using the model to back-calculate physical parameters is strongly sensitive to the experimental conditions; hence, guidelines to minimize it in further experiments are provided.

Advances in engine technologies are placing additional demands on emission control catalysts, which must now perform at lower temperatures, but at the same time be robust enough to survive harsh conditions encountered in engine exhaust. In this Review, we explore some of the materials concepts that could revolutionize the technology of emission control systems. These include single-atom catalysts, two-dimensional materials, three-dimensional architectures, core@shell nanoparticles derived via atomic layer deposition and via colloidal synthesis methods, and microporous oxides. While these materials provide enhanced performance, they will need to overcome many challenges before they can be deployed for treating exhaust from cars and trucks. We assess the state of the art for catalysing reactions related to emission control and also consider radical breakthroughs that could potentially completely transform this field. Exhaust emissions catalysts can be used for the removal of harmful pollutants. This Review explores synthesis routes and materials for advanced catalysts, and identifies grand challenges for the transformation of pollutants.

Stoichiometric cerium oxide nanoparticles with different sizes, ranging between 2 and 12 nm, were prepared by a rational design of the synthetic concept, based on a special hydrothermal procedure. ...

Despite significant effort spent on the investigation of catalytic ammonia oxidation at the molecular scale, there is surprisingly little work that investigates the behavior of the established reaction mechanisms under industrial conditions in the presence of mass transfer limitations. This paper presents reactive flow simulations of ammonia oxidation on platinum gauzes under industrial operating conditions, combining a mechanistic description of the surface chemistry with the computation of the flow-, temperature and concentration fields around the platinum wires. Overall, the simulations yield temperature- and concentration fields, as well as integral N2O selectivity in line with industrial experience and (limited available) experimental data. In particular, the simulations predict the experimentally observed increase of the integral N2O selectivity with increasing flow velocity, decreasing wire diameter and wire-to-wire distance, and increased surface area due to surface reconstruction. The main result of the paper is that the local interaction of flow and surface chemistry leads to a variation in the local N2O selectivity across the gauze:The N2O and N2 selectivity is higher on the front side of a wire than on the rear side. A reduced N2O selectivity is observed where one wire is shadowed by another wire. Increased N2O selectivity is observed at stagnation points where upstream wires direct the flow so that it hits a downstream wire with higher velocity. These examples show that - through the flow directing effect of the upstream wires- the selectivity on an individual wire is influenced by the presence of other wires. This observation provides a mechanistic explanation for the industrial observation that optimized gauze geometries can lead to reduced N2O formation.

Spatially resolved operando HERFD-XANES (high energy resolution fluorescence detected X-ray absorption near edge structure) complemented by CO concentration gradient profiles along the catalyst bed...

Machine learning based surrogate models that interpolate precomputed solutions of the rate equations can greatly accelerate simulations of catalytic systems. Established schemes for the construction of surrogate models depend on a logarithmic scaling of the source terms, which limits their application to cases where species are exclusively consumed or produced under all conditions of interest. We propose a new approach based on interpolating the forward- and reverse rates of the rate-determining reactions and thus overcoming this limitation. The new scheme is demonstrated using a surface reaction mechanism describing the oxidation of CO and H2 as well as the water gas shift reaction including 5 gas species, 9 surface species and 18 reversible reactions. Multivariate spline interpolation is used for a first evaluation of our approach. With the splines, the new method reproduces the source terms of CO with an error of 0.5 % which is 50 to 100 times more accurate than the established approaches of either mapping the source terms directly (49.9 % error) or mapping adsorption/desorption rates (23.7 % error). The true potential of the new approach develops in combination with machine learning techniques like neural networks. Even very small neural networks with a single hidden layer of 20 nodes yield an error of 0.35 %. This is about 200 times more accurate than the same neural networks used with the established approaches. Increasing the network size to still moderate 30 nodes in two hidden layers (2342 parameters in total) reduces the error to 0.0058 %. Besides the increased accuracy, the neural networks also outperform spline interpolation by at least an order of magnitude with respect to interpolation time, storage space and required amount of training data. Due to the extremely low errors achieved with moderate size neural networks and small training data sets, the proposed method based on mapping the rate-determining steps shows promise to scale to much larger and more complex reaction systems in catalysis and other fields like combustion, atmospheric chemistry or systems biology.

fff; International audience; Reductive treatments with pulses of CO-rich atmosphere have been used to increase and maintain the low temperature activity of a Pt/CeO2-based oxidation catalyst. A combination of operando infrared thermography and spatiotemporal-resolved quick scanning extended X-ray absorption fine structure spectroscopy on a fixed bed microreactor unraveled that, apart from the pulse length, the reaction atmosphere, and the reactor temperature, also the emerging reaction heat during such activating pulses has a strong influence on the structure and catalytic performance of CO and propylene conversion in the axial direction of a fixed-bed and a monolithic reactor. The reductive pulse activation led to an increase of the integral catalyst activity as well as to the generation of zones of different particle sizes along the catalyst bed. In the case of an activation temperature between 250 and 350 °C and pulse lengths between 5 and 30 s, a hotspot of more than 80 K was observed at the beginning of the catalyst bed. Spatially resolved X-ray absorption spectroscopy indicates that larger and more reduced Pt particles are formed particularly at the beginning of the catalyst bed, whereas its subsequent part is less affected. Both the length of the reductive pulses and activation temperature have a distinct influence on the noble metal particle size. On the basis of these results, a Pt/CeO2 based honeycomb shaped substrate was activated in a similar manner. Spatially resolved gas phase profiling showed different reaction rates at the beginning of the reactor, which indicates that the concept can be transferred also to industrially relevant catalysts. In the future, such an activation procedure might open up the door to a new class of operation strategies, by which individual zones generated in the catalyst bed could be assigned for removal of specific pollutants in the exhaust stream

The impact of catalyst pretreatment and its role in diesel oxidation catalysts was systematically studied by comparing the performance, CO adsorption characteristics, oxidation state, and local str...

The formation of CeO2 colloidal particles upon heating an aqueous solution of (NH4)2Ce(NO3)6 to 100 °C was investigated by time-resolved in situ SAXS analysis using synchrotron radiation, providing absolute intensity data. In particular, the experiments were performed by applying different temperatures between room temperature and 100 °C as well as under variation of the ionic strength and concentration. Using validated SAXS evaluation tools (SASfit and McSAS software), the analyses revealed the presence of two types of particle populations possessing average dimensions of ca. 2 nm and 5-15 nm, with the latter being agglomerates of the 2 nm particles rather than single crystallites. The analysis revealed not only the changes in the size, but also the relative volume fractions of these two CeO2 particle populations as a function of the aforementioned parameters. Increasing the temperature increases the number of the 5-15 nm agglomerates on one hand by the enhanced nucleation rate of the primary particles. On the other hand, especially at high temperatures (90 and 100 °C) the larger agglomerate particles precipitate, resulting in interesting trends in the fractions of the two populations as a function of time, temperature, ionic strength, and precursor concentration. The experimental studies are complemented by calculating colloidal interaction energies based on classical DLVO theory. Thereby, this study provides detailed insight into the nucleation, growth, and agglomeration of CeO2 nanoparticles. The primary objective of this study is to provide a better understanding of the nucleation and growth of particles by the hydrolysis of the tetravalent cerium ion in aqueous solutions.

Mechanistic proposals for the different SCR subreactions are integrated into one surface reaction mechanism that describes the main SCR reactions (Standard SCR, Fast SCR, NO2 SCR), transient effects due to nitrate storage, as well as the production of the side product N2O over a copper chabazite catalyst. The mechanism is parameterised to steady state and transient experiments, and is shown to predict the behaviour of the catalyst during a driving cycle, without any refitting of kinetic parameters. A dual site approach is used, where site 1 accounts for the adsorbed ammonia that forms on the Bronsted acid sites and copper ions, while site 2 is a copper ion (Cu2+-OH) where nitrites and nitrates are adsorbed. All main SCR reactions proceed via a reaction between ammonia and nitrites (ammonium nitrite pathway) to produce nitrogen; nitrites are also the linking species between the Standard SCR and NO oxidation reactions. Reactions between nitrates and ammonia to produce ammonium nitrate are also included, along with ammonium nitrate decomposition pathways (i.e., via NO addition to feed). Additionally, a global reaction taking place between adsorbed ammonia and gaseous NO2 to produce N2 at low temperatures ( The mechanism was used to analyse the importance of nitrate formation during a standard driving cycle. Surprisingly, although a significant amount of inhibitive ammonium nitrate is modelled to form during low temperature Fast and NO2 SCR steady state experiments, almost no ammonium nitrate is predicted to form during the driving cycle, thus allowing for a higher reaction activity than predicted based on steady state data. From a modelling and catalyst testing perspective, this shows the importance of capturing the catalyst’s transient behaviour rather than only steady state conditions, since steady state is not necessarily reached during practical driving scenarios.

Monodisperse packed beds have long been used as surrogate models to predict the filtration performance of particulate filters using analytical methods. In recent years, however, polydisperse packed beds have received special attention as they have the potential to better represent the microstructure of porous filter walls. In this paper, an analytical model for the filtration performance of clean polydisperse packed beds is derived based on the well-proven classical packed bed filtration theory. Predictions of the newly developed model were compared to the results of numerical simulations of the filtration performance of two polydisperse packed beds. The proposed filtration model was in considerably better agreement with the simulations than previous analytical models.

By applying different transient test protocols in a lab reactor, it is demonstrated that in the presence of C 3 H 6 a Pt/Pd based oxidation catalyst experiences a persistent deactivation that can be attributed to the accumulation of propene partial oxidation products on the catalyst surface. The deactivation can be reversed by treating the catalyst at higher temperature in a hydrocarbon free atmosphere. The three most important scenarios giving rise to the activation/deactivation phenomenon are: (1) Consecutive heat-up/cool-down cycles: If several consecutive heat-up/cool-down cycles are performed with a catalyst pretreated in a propene free oxidative atmosphere, a lower ignition temperature is observed during the first heating ramp than during the following cycles. The effect can be explained by a reversible blocking of active sites by hydrocarbon (HC) intermediates, which are formed during the ignition branch, then entirely are removed after the HC ignition is completed and finally are formed again during the extinction branch, providing a deactivated state of the catalyst for the following ignition. The deactivation can be completely reversed either by a conditioning in an oxidizing HC free gas mixture at temperatures above 110 °C or a conditioning in N 2 above 140 °C. The deactivation effect is more pronounced under conditions that lead to a low light-off temperature and completely disappears at conditions leading to higher light-off temperatures, i.e. high CO concentrations or high space velocities. (2) Light-off tests with varying temperature ramp rate: For faster ramp rates, CO/propene ignition is observed at lower inlet temperatures. This is the opposite behavior than would be expected based on thermal considerations and can be explained by a higher accumulation of HC intermediates during the slower ramp. (3) Transient observation of the deactivation effect: If an initially clean catalyst is operated at a constant temperature in the light-off range, a transient decrease in CO/HC conversion is observed that can be attributed to the gradual build-up of the blocking HC intermediates. A kinetic model is presented which considers not only the standard CO and hydrocarbon oxidation kinetics, but also the formation of site blocking intermediates by partial oxidation of propene and the removal of these blocking intermediates either by total oxidation with O 2 or by thermal desorption. The model nearly quantitatively reproduces the experimental results of the different test protocols.

Common system analysis and controller design methods are based on state-space models. However, most physical-chemical catalyst models are not represented in a state-space. In this paper, we present an approach to transform simplified physical–chemical models into the state-space, and hereby, we introduce the concept of weighting functions. The weighting functions represent the catalyst model in a compact form and are a convenient way for system analysis. The weighting functions will be analyzed from the system theoretical and chemical point of view. The generalization of this approach will be demonstrated. Exemplarily, three reaction mechanisms will be transformed into nonlinear state-space models. Based on the first model, the nonlinear state-space model will be linearized and a linear quadratic controller will be designed. Contrarily to the LQ controllers in the literature, the weighting matrix for the state vector is chosen in the model-based form. The only few design parameters are intuitive to choose.

A titration method is presented that measures the redox thermodynamics of oxygen storage in three way catalysts under typical operating conditions. The titration scheme starts with an oxygen pulse to completely oxidize the catalyst. The catalyst is then brought into equilibrium with an H2/H2O mixture and the resulting oxidation state is determined from the amount of H2 consumed to establish the equilibrium. Measurements were performed in the temperature range of 300 °C–500 °C for two commercial three-way catalysts containing a doped ceria-zirconia solid solution of the composition Ce0.404Zr0.552Nd0.044O1.978. The results confirm that under typical automotive operating conditions the oxygen available from the oxygen storage is limited by the equilibrium thermodynamics. Even under very rich conditions (8000 ppm H2/10% H2O) at 500 °C only about 20% of the theoretically available oxygen can be used. At the typical steady state operating point of a three-way catalyst (lambda sensor voltage of 650 mV, 70 ppm H2), at 500 °C only about 6% of the ceria is reduced. Virtually identical oxygen storage performance is obtained for two catalysts with identical oxide composition but different precious metal loadings of 100 g/ft3 and 10 g/ft3. In a second part of the study, the dynamical oxygen breakthrough is studied experimentally in the H2/H2O/O2 system. First, the catalyst is brought into the typical steady state oxidation state and then it is exposed to an oxygen pulse of varying length. If the thermodynamic data is implemented in a kinetic model, the H2 and O2 signals during the dynamic oxygen breakthrough experiments are very well described. A sensitivity study shows that the H2 response to small O2 pulses, where no O2 breakthrough is observed yet, is completely controlled by the thermodynamics of the oxygen storage. As the O2 pulses get longer and O2 breakthrough is observed, besides thermodynamics also kinetics and internal as well as external mass transfer become important.

Advances in engine technologies are placing additional demands on emission control catalysts, which must now perform at lower temperatures, but at the same time be robust enough to survive harsh conditions encountered in engine exhaust. In this Review, we explore some of the materials concepts that could revolutionize the technology of emission control systems. These include single-atom catalysts, two-dimensional materials, three-dimensional architectures, core@shell nanoparticles derived via atomic layer deposition and via colloidal synthesis methods, and microporous oxides. While these materials provide enhanced performance, they will need to overcome many challenges before they can be deployed for treating exhaust from cars and trucks. We assess the state of the art for catalysing reactions related to emission control and also consider radical breakthroughs that could potentially completely transform this field.

Spatially resolved

This paper examines the activity of a platinum catalyst for the oxidation of CO under conditions of temperature cycling. Each cycle consists of an ignition and extinction curve obtained with a linear temperature ramp. It is shown that the catalyst activity is higher for the first ignition cycle than for the subsequent ones regarding CO conversion. The activity for the extinction portion of each cycle is the same for all cycles. It is proposed that the loss of catalyst activity results from the deposition of hydrocarbon residues during the extinction part of the cycle. The catalyst activity can be completely restored by heating the catalyst to 350 °C in the presence of oxygen and nitrogen.

A catalytically active particulate filter with a first order catalytic reaction taking place inside the filter walls is investigated by numerical simulation. The conversion efficiency for different channel geometries and operating conditions is systematically studied as a function of the governing dimensionless parameters. It is found that the conversion efficiency of a catalytically coated wall flow filter is very close to that of an ideal plug flow reactor over the full range of realistic operating conditions. Only in a range of intermediate residence times, the filter reactor shows some diffusion limitation which leads to conversion efficiencies slightly below that of the plug flow reactor. In all cases, these deviations from ideal conversion behaviour are below 15%. If the filter and the open monolith are compared at identical operating conditions and channel geometries, for fast reactions, the filter reactor shows higher conversion efficiency than the open monolith, since in this case the open monolith becomes strongly mass transfer limited. However, there can be a small range of conditions where the monolith is slightly more efficient than the filter. The reason for this effect is that due to the thinner washcoat layer the onset of mass transfer limitation in the coated monolith is shifted to higher reaction rates, compared to the filter reactor.

A new numerically efficient simulator for dual layer catalysts is implemented and its performance is demonstrated using the example of the dual layer ammonia oxidation catalyst. For the solution of the radial mass balances, each washcoat layer is represented by a single volume element and the diffusive mass fluxes into and within the washcoat are computed based on the concept of the internal mass transfer coefficients. The performance of the new simulator is compared against a reference simulator that fully resolves the concentration profiles in the washcoat. For a steady state test case, the error introduced by the new solution scheme is below 2 % for all relevant exhaust components. For a transient test case, the deviations become more significant. For the simulation of a WHTC, the maximum deviation in the NH3 outlet concentration is 23.3 %, whereas the cumulated NH3 emissions deviate by 10.6 %.With the new simulator, the 1800 s of the transient WHTC cycle can be simulated within 53 s on a standard laptop, 30 times faster than real time.

A new experimental setup is introduced, enabling a preheated exhaust gas mixture to abruptly heat up an initially cold monolith catalyst and, thus, reproduce transient catalyst light-off behavior. With this setup, cold start experiments have been carried out on an aged monolithic Pd three-way catalyst (TWC). This study aims at systematically investigating the effect of diverse hydrocarbons, including methane, acetylene, ethene, ethanol, acetaldehyde, propane, propene, toluene, and xylene in TWC exhaust. It was shown that the commonly used model components propane and propene do not represent the behavior of aromatics and other hydrocarbons present in a typical TWC exhaust very well, since they do not account for their variability in inhibition strength and reactivity. Among the studied hydrocarbons, acetylene proved to be the strongest inhibitor; even at low concentrations of 70 ppm (a typical value found in TWC exhaust), acetylene inhibition increases CO cold-start emissions by 40 %. Furthermore, aromatic and liquid hydrocarbon species (like ethanol) first condense on the initially cold catalyst surface and evaporate again as the catalyst heats up, resulting in a desorption peak at the catalyst outlet.

Knowledge of the actual catalyst state plays a key role in automotive exhaust gas aftertreatment. The oxygen loading degree of three-way catalysts (TWC), the amount of stored ammonia in selective reduction catalysts (SCR), or the NOx loading degree in NOx storage catalysts (NSC) are important parameters. Today, they are determined indirectly and/or model-based, applying models that are typically calibrated by gas sensors installed up- and/or downstream of the catalysts. A novel approach to determine directly the catalyst state by microwaves (radio frequencies, rf) emerged recently. For this method, the catalyst housing serves as an electrical cavity resonator. As “sensor”, one or two simple antennas are mounted in the canning. The electrical properties of the honeycomb incl. coating change with gas loading, affecting either the resonance frequencies or the power transmission. Such contactless-obtained information is strongly correlated with the catalyst state as will be discussed here for TWC and SCR systems. This contribution reviews the progress in the past 3 years that exceeds by far the status of initial studies.

Zur effizienten Abgasnachbehandlung bei Ottomotoren hat sich der Drei-WegeKatalysator seit mehr als drei Jahrzehnten bewahrt. Seinen Namen verdankt er der Eigenschaft, die drei Hauptschadstoffe im Abgas von Ottomoren, namlich Kohlenwasserstoffe (HC), Kohlenstoffmonoxid (CO) und Stickoxide (NOx), gleichzeitig in ungiftiges Wasser (H2O), Kohlenstoffdioxid (CO2) und Stickstoff (N2) umzuwandeln. Nach dem aktuellen Stand der Technik bleibt das dynamische Verhalten des Katalysators beim Reglerentwurf weitgehend unberucksichtigt [1]. Der Wunsch nach steigender Effizienz beim Betrieb und bei der Herstellung des Katalysators erfordert allerdings den Einsatz von modellbasierten Verfahren zur Systemanalyse und zum Reglerentwurf.

The impact of catalyst distribution within the porous wall of a coated gasoline particulate filter on its catalytic performance is investigated. Sections of catalyzed filter wall with different amounts of washcoat and different levels of uniformity are digitally reconstructed from tomography data. A coupled problem involving the exhaust gas flow and the transport and reaction of a dilute species in the reconstructed structures is solved using the lattice Boltzmann method. The resulting model is verified and its accuracy is assessed for first-order reactions and those with inhibition kinetics. Computed washcoat effectiveness values agree well with the Thiele modulus theory. Improvements in catalytic conversion and washcoat effectiveness of uniform washcoat distribution are observed with increased washcoat loading. These effects are quantified and the benefits are found to be limited to the transitional regime. The uniform washcoat distribution is shown to be beneficial for both filtration and conversion performance.

The validation of a numerical model used to predict the permeability of the microporous wall of a coated gasoline particulate filer is arduous due to multiscale effects. To circumvent this, an experimental “magnified twin” approach based on kinematic similarity is proposed. It consists of (1) printing magnified porous wall sections by fused filament additive manufacturing; and (2) characterizing their permeability using a modified falling head permeability measurement device and testing procedure. A detailed approach is also proposed to select the appropriate scaling factor to produce the samples by 3D printing. Comparison between numerical, semi-analytical, and experimental predictions of the impact of coating amount and degree of uniformity on wall permeability is presented. Despite the inherent sensitivity of permeability to pore space characteristics, a very good agreement is reported between simulations and experiments with a 19.3% mean discrepancy for the seven magnified structures tested. The proposed procedure is a relatively easy-to-implement and inexpensive method that may find many applications in the field of porous media.

A three-step numerical model sheds light on the impact of three-way catalyst washcoat distribution within the cordierite porous wall of a clean gasoline particulate filter on its filtration performance. The model relies on (1) the numerical reconstruction of porous wall sections with various washcoat distributions and coating amounts, generated using a novel set of morphological image processing operations applied on segmented X-ray computed tomography data, (2) the computation of the flow field using the lattice Boltzmann method, and (3) the prediction of soot capture efficiency by solving the Langevin equation. The impact of washcoat distribution and amount on the pressure drop, permeability, filtration efficiency, and filter quality factor is systematically investigated. For a non-uniform washcoat distribution, an unexpected decrease in filtration efficiency with an increase in washcoat amount is explained and this highlights the complexity of the effects generated by the deposition of washcoat within the porous wall of the filter.

Currently, the literature on ammonium nitrate formation across SCR catalysts mainly sees ammonium nitrate formation as an undesired problem because it can inhibit the catalytic activity. This paper shows that, besides the inhibition effects, ammonium nitrate formation on the catalyst can also positively influence the catalytic activity because it acts as an NO2 buffer that stores NO2 under NO2-excess conditions and releases NO2 under NO2-deficient conditions. To demonstrate the working principle of the ammonium nitrate buffer, in a first step the ammonium nitrate loading is determined as a function of the NO2/NOx ratio in the feed, using a titration scheme recently proposed in the literature. Then, step tests are performed where at otherwise constant feed conditions, the NO2/NOx ratio is switched between 75 % NO2/NOx and 25 % NO2/NOx. During the slow frequency switching (3 min per phase), an increased NOx conversion is observed after both the transition to NO2-rich and NO2-lean. This can be explained by a build-up of ammonium nitrate during NO2-rich and the consumption of ammonium nitrate during NO2-lean conditions owing to the enhanced SCR reaction. When the frequency of the NO2/NOx switches is increased to 30 s, the ammonium nitrate storage loading can no longer follow the fluctuations and stays at an intermediate level. During this scenario, the catalyst shows a constant high conversion close to the Fast SCR reaction. Finally, it is demonstrated that the ammonium nitrate buffer contributes to the overall NOx conversion during standard driving cycles. In particular, the NO2 emissions are predicted significantly better by a model that accounts for the ammonium nitrate storage equilibrium.

This paper presents a model extension to dual-layer monolithic catalysts from a one-dimensional (1D) ( Balakotaiah, 2008 , Joshi et al., 2009 ) mathematical model used for single-layer monolithic catalysts. The novel transient non-isothermal model computes average concentrations inside the bulk flow and in each layer of washcoat at every computational grid point along the monolith channel. The model has the capability that physical (e.g. diffusivity, pore radius, and thickness) and chemical (reaction kinetics) properties of the two layers of the washcoat can be different from each other. The relatively low computational cost offers real-time simulations of multi-layer monolithic catalysts. Analytically derived expressions for transverse mass transfer coefficients between the bulk flow and washcoat layers are presented for single-layer and dual-layer catalysts with circular cross-section, which can be easily extended for multi-layer catalysts. In addition, three expressions are presented for calculation of three internal Sherwood numbers, which are used in the dual-layer model, based on the values of Thiele moduli of the washcoat layers. These expressions are especially useful for calculation of the mass transfer coefficients for dual-layer catalysts with non-circular cross-sections. Application limitations of the 1D multi-layer model with respect to physical and chemical operating conditions are discussed. The derived model is applied for an automotive aftertreatment catalyst and the results of this model are compared with the results of a detailed 2D model (selective catalytic reduction/platinum group metals dual-layer catalysts).

This paper presents laboratory reactor measurements of the steady-state conversion of the pollutants CO, hydrocarbons and NO over a three-way catalyst containing ceria as an oxygen-storage component. It is shown that the presence of the saturated hydrocarbons C 3 H 8 or CH 4 causes a shift in the stoichiometry of optimal conversion (CO–NO crossover) to fuel-rich compositions. The shift was more pronounced at lower temperatures and can be explained by the kinetic limitation of the hydrocarbon oxidation reaction. A microwave cavity perturbation technique was used to measure in situ the oxidation state of the ceria. In a first step, titration experiments were performed. The oxygen-storage level was adjusted to a predefined level by equilibration with a H 2 /H 2 O mixture. The experiments showed that for a given temperature, the microwave-derived signal (here the resonance frequency) correlates well with the oxidation state of ceria. The microwave cavity perturbation technique was then applied simultaneously to steady-state performance tests in the presence of different hydrocarbons. It was found that if the exhaust stoichiometry is ramped from lean to fuel rich, the CO–NO crossover point in all cases coincides with a strong decrease in the oxidation state of ceria. The correlation between the oxidation state of the catalyst and the catalytic performance is found to be independent of the catalyst temperature and the nature of the hydrocarbon. The CO–NO crossover point can be precisely determined by the microwave cavity perturbation technique. The results suggest that a microwave-based measurement may, under some circumstances, allow for a more precise control of the catalyst performance than the conventional control by lambda sensors.

In today’s vehicle applications, SCR ammonia dosing is completed using complex control algorithms that need to be parameterized for the individual catalyst technology. Owing to the complexity of the parameterization procedures, ammonia dosing is frequently completed using either oversimplified dosing strategies (e.g., constant NH 3 /NO x ratio) or strategies previously optimized for different catalysts during the early design phase’s simulation and laboratory studies. This paper presents a practical, model-based approach that allows for simulation and laboratory studies to be performed with an individually adapted dosing profile. The procedure is based on conventional feedback control strategies, where the average ammonia storage level is controlled using a temperature dependent setpoint (the desired storage level), which is interpolated from a look-up table. The approach of this paper is to optimize numerically the entries of this look-up table for a given test cycle such that the NO x conversion is maximized while the ammonia slip is kept below a defined level. Although the optimization is performed for a particular cycle, it is demonstrated that the dosing strategies obtained in this way also work reasonably well for other test cycles. The proposed procedure was applied in a catalyst screening simulation study that compared the performance of an Fe– and a Cu–zeolite catalyst for a given test cycle. It was demonstrated that each of the two catalysts appeared to have a better performance than the other catalyst when the specified catalyst’s optimized dosing profile was applied to both catalysts. The best catalyst was only identified when comparing both catalysts using their own respective optimized dosing profile. A second simulation study used the proposed dosing procedure to compare catalysts with different ammonia storage capacities. Again, the relative ranking of the catalysts’ performance was shown to depend on the dosing profile. Overall it is demonstrated that the true potential of a catalyst can only be determined when an individually adapted dosing strategy is used.

Pt-CeO2-Al2O3 catalysts play an important role in diesel oxidation and three-way catalysis. In this study, the fast structural dynamics of both platinum and ceria in a 1 wt %Pt/5 wt %CeO2-Al2O3 catalyst prepared by flame spray pyrolysis have been systematically investigated under reducing and oxidizing conditions to elucidate the role of the Pt–CeO2 interface for CO oxidation and fast oxygen storage/release of ceria. The catalyst showed enhanced catalytic activity, particularly after application of a reducing/oxidizing conditioning step at 250 °C, with a pronounced dependence on the reducing agent (C3H6 < H2 < CO). In situ time-resolved X-ray absorption spectroscopy (XAS) at the Ce L3-edge unraveled a dependence of the reduction extent of ceria during temperature-programmed reduction on the noble metal constituent and the applied reducing agent. Dynamic reducing/oxidizing cycling (2% H2 ↔ 10% O2 or 2% CO ↔ 10% O2) at various temperatures (150, 250, and 350 °C) showed that the reducibility of ceria increas...