Near-wall dispersion, deposition and transformation of particles in automotive exhaust gas aftertreatment systems

Combustion-generated nanoparticles are present in exhaust gas aftertreatment systems in the approximate size range 1-1000 nm. Successful optimization of aftertreatment systems for pollution control relies on the existence of numerical tools to predict the momentum, heat and mass transfer between these types of particles and the surrounding gas phase. Such tools can only be readily obtained if our fundamental understanding of the phenomena pertaining to the particle behavior is correct. The small nano-particulates in automotive exhaust gas aftertreatment systems are typically described as spherical and inert, closely following the gas phase streamlines apart from a superimposed Brownian motion. However, for real particulate matter, produced by an internal combustion engine, we show that the deposition in an automotive catalyst substrate cannot in general be well described by the aforementioned modelling approach, as particle transformations become active inside the substrate channels, altering the apparent deposition efficiency. A conceptual model is proposed that is able to explain the initially observed discrepancies between measurements and simulations, by describing the particulate matter as a mixture of three different types of particles: truly inert particles, semi-volatile particles and completely volatile particles. The conceptual model is corroborated by experimental and numerical investigations into the behavior of truly inert particles in automotive catalyst substrates. Finally, the possibility to use the model for in-situ characterization of particulate matter is demonstrated. For the first time, data is presented to support the hypothesis that differences in particle properties, as characterized in this way, have a meaningful correlation to particle reactivity in e.g. oxidation experiments. In other words, particle reactivity may be assessed indirectly by investigations of particle mobility. It is also shown how the pressure drop through a bed of deposited nanoparticles will differ depending on the properties of the deposited particles, as characterized by the conceptual model.