PNNL

Abstract

More stringent emissions regulations are continually being proposed to mitigate adverse human health and environmental impacts of internal combustion. With that in mind, it has been proposed that vehicular particulate matter (PM) emissions should be regulated based on particle number in addition to particle mass. One aspect of this project is to study different sample handling methods for number based aerosol measurements, specifically, two different methods for removing volatile organic compounds (VOCs) from an aerosol sample. One method is a Dekati Thermodenuder (TD) and the other is an evaporative chamber/diluter (EvCh). These sample handling methods have been implemented for this project in an engine test cell built around a direct injection spark ignited (DISI) engine. The engine was designed for stoichiometric, homogeneous combustion. Direct injection is of particular interest for improved fuel efficiency but this comes with the production of a significant amount of (PM) and may therefore be subject to the proposed number based regulation. Another aspect of this project is to characterize PM from this engine in terms of particle number and composition. The first interesting observation is that PM number distributions, acquired using a TSI SMPS, have a large accumulation mode (30-294 nm) but a very small nuclei mode (8-30 nm). This is understood to represent a lack of condensation particles meaning that neither the exhaust conditions nor the sample handling conditions are conducive to condensation. This lack of nuclei mode does not, however, represent a lack of VOCs in the sample. It has been observed, using mass spectral analysis (limited to PM>50 nm), that PM from the DISI engine has approximately 40% organic content through varying operating conditions. This begs the question of how effective different sample handling methods are at removing these VOCs. For one specific operating condition, called Cold Start, the un-treated PM was 40% organic. The TD reduced this by 7% while the EvCh reduced it by 13%. For other operating conditions, PM treated for volatile removal actually exhibited an increase in organic fraction on the order of 5%. This addition appears to be sensitive to the gaseous hydrocarbon concentrations in the exhaust although a precise correlation has not yet been derived. It has been concluded that VOCs are tightly bound to the PM carbon core and thus are not effectively removed by either treatment method.