sources and additional reading: https://pubs.acs.org/doi/pdf/10.1021/es034928d
https://phys.org/news/2017-05-carcinogenic-soot-particles-pe...
https://www.energy.gov/sites/prod/files/2014/03/f9/2002_deer...
Second paper only considers fine-size (<20 nm) particulates (and not larger sizes like 100nm+) of direct injection cars, and is a bit misleading by comparing six cars to a single carefully-selected diesel car with particulate filter. Further, in reality here in a London, people are cutting open their Diesel particulate filters to avoid cost of replacing clogged filters and still passing MOT exhaust tests, which apparently don’t/can’t check for filter tampering.
Third paper is way back in 2002 and considers only cars, not heavy trucks and buses, which are way worse in urban settings.
Diesels are only good for limited purposes, like buses and trucks, that need the higher efficiency.