Detailed multi-technique analyses were performed on particles generated from a single cylinder engine. For the first time it was possible to identify key molecular markers and discriminate particles by source, size and engine operation regime. Particles were identified according to gasoline-specific bonds, lubricant- specific bonds and engine-specific bonds. The measurements clearly and unambiguously showed the effectiveness of the catalytic stripper used to remove the semi-volatile particles. A large amount of data was collected and used to support the development of MGA.
MGA combines physico-chemical simulations with statistical algorithms to simulate the formation of particulate emissions. During the project, the MGA was applied to a single cylinder engine, a multi-cylinder test engine with exhaust aftertreatment, and to simulate the emissions from vehicles in various drive cycles, including real driving emissions (RDE). The physico-chemical model was extended to include not only particles from incomplete combustion, but also from other important sources (e.g. fuel impingement on cylinder walls in the engine, ash formation from lubrication oil and condensation of semi-volatiles in the exhaust). The MGA results matched the physical characteristics from the single cylinder engine studies.
Three prototype measurement devices were developed and they were optimized using a two-step approach. In a first step, the condensation particle counters were optimized and re-calibrated, including a temperature adjustment for the saturator and condenser. In the second step, the configuration was optimized to reduce particle losses. The highest losses occur in the catalytic stripper, where a trade-off is made between reducing losses and removing semi-volatiles. The optimized catalytic stripper achieved a penetration efficiency of over 60% (i.e. losses less than 40%) for 8 nm particles, and a semi-volatile removal efficiency of over 99% for particles composed of tetracontane.
An extensive testing campaign was used to validate the performance of the prototype measurement devices. Three test configurations were used: chassis dyno tests with the measurement equipment located at the tailpipe, chassis dyno tests with the measurement equipment located after the CVS tunnel, and real road tests. The results show that sub-23 nm particles are mainly generated at the engine start and during acceleration phases. There was a high consistency between different real driving emission (RDE) tests and good repeatability: the differences are all between 15% and 35%, so well below 50% (the current legislative limit).
The PEMs4Nano consortium participated in 47 conferences, including the major international aerosol meetings and invited presentations at the TRA2018 in Vienna and TRA2020 in Helsinki. Seven peer-reviewed articles have been published, and five more are under review. The project was highly visible, with two articles making it to the journal cover. It has won 4 prizes, including the Trojan Horse award at the ETH Particle Meeting 2019, a paper selected as one of the best papers at the SAE 2019 conference, and the Prix Jean Bricard 2020 prize. Finally, a patent on a semi-volatile mass detector was filed.
