Experimental Validation of Particulate Matter (PM) Capture in Open Substrates Jonas Sjö blom,* Henrik Strö m, Ananda Subramani Kannan, and Houman Ojagh † Department of Applied Mechanics, Chalmers University of Technology, SE 412 96 Gothenburg, Sweden ABSTRACT: The capture of engine-borne particulate matter (PM) in after-treatment systems is a complex process. Because of the intrinsic heterogenic nature of the PM, the particles undergo transformations that make it very difficult to isolate their motion and deposition in experiments. In a previous study, a model for hydrocarbons (HC) evaporation of the particles during the capture process was proposed to explain experimental results that showed a significant increase in the capture efficiency when compared to those predicted from theoretical models [J. Sjö blom and H. Strö m, Ind. Eng. Chem. Res. 2013, 52, 8373]. In this work, inert NaCl particles were fed to an open substrate (cordierite monolith). It was demonstrated that the capture efficiency can be experimentally observed, isolated from other experimental phenomena and uncertainties, if the particles are truly inert. Consequently, the previously proposed model for HC evaporation is a valid starting point for development of comprehensive models for PM motion and transformations. ■ INTRODUCTION The emission of particulate matter (PM) poses a severe threat to human health and the environment. 1 In most countries, the legislation puts great demand on PM removal, especially for particulate number emissions. For diesel vehicles, this results in the implementation of diesel particulate filters (DPFs) designed as wall-flow filters. In the DPF, every second channel is plugged and the exhaust flow is forced through the permeable channel wall, which enables a very high capture efficiency (CE). However, the high CE is accompanied with a high-pressure drop and together with the need for periodic regeneration, the use of a DPF results in a fuel penalty in the order of 2%-3%. 2 In order to enable optimization of DPF design and operation, a detailed understanding of the capture process is necessary. This also includes the processes taking place upstream to the DPF, such as those occurring in the diesel oxidation catalyst (DOC) commonly placed upstream to produce NO 2 and to remove hydrocarbons (HC) and CO. The DOC is commonly designed as an open substrate (e.g., a cordierite monolith with square channels) that may have a profound effect on the PM capture in the downstream DPF, as it will change the PM character- istics. These changes include PM capture, creation of sulfate PM as well as transformation due to HC evaporation. To study the capture phenomena in open substrates, an experimental campaign was conducted using a passenger-car diesel engine connected to an exhaust after-treatment system (EATS). 3 To reduce the inherent correlation (e.g., space velocity and temperature) in exhaust properties, the EATS is designed to deliver independent variation of flow and temperature, as well as enable the addition of air to the exhaust stream. 4 This setup, together with the use of Design of Experiments (DoE), enabled the experimental study of isolated changes in flow parameters that could not be accomplished by changes to the engine operation alone, and thus improved the interpretation of the results. In the experiments, the flow (and thus the channel Reynolds number) was low in order to get a significant CE, despite the open channel structure, since the CE of a monolith substrate is otherwise much lower than that of a DPF, because of the insufficient diffusive transport of PM toward the wall during the retention time in the channel. 5 However, the CE was much higher than predicted from theory (Brownian motion/ diffusion) alone. 6 Figure 1 shows an example of such poor agreement between the measured CE of automotive PM and the theoretically predicted CE of inert particles of identical size. This deviation between experiments and theory was attributed to HC evaporation from the surface of the PM. Because of the short diffusion distance in the monolith channel and the rapid adsorption of the HC on the channel walls, the channel becomes a strong “HC sink” and thus drives off HCs from the PM. The evaporation process was incorporated into a tanks-in-series model and could be used to explain the experimental findings. 3 The tanks-in-series model assumed a bulk concentration of HC equal to zero (i.e., assuming that the diffusion process of HC to the wall was faster than the Received: November 29, 2013 Revised: February 14, 2014 Accepted: February 14, 2014 Published: February 14, 2014 Figure 1. Example of experimental CE (solid lines), compared to CE predicted from theory (dashed lines). 3 Research Note pubs.acs.org/IECR © 2014 American Chemical Society 3749 dx.doi.org/10.1021/ie404046y | Ind. Eng. Chem. Res. 2014, 53, 3749-3752