Polycyclic aromatic hydrocarbons (PAHs) are ubiquitous compounds emitted by all combustion sources. They are of major health concern because of their toxic properties and are therefore, regulated pollutants in ambient air. In the atmosphere, PAH oxidation through homogeneous and heterogeneous reactions may lead to the formation of oxy- and nitro-PAHs. These latter species are also emitted concomitantly with PAHs during incomplete combustion processes. Oxy- and nitro-PAHs are potentially more toxic than their parent PAHs. The identification of the origins of oxy- and nitro-PAHs is challenging, due to the coexistence of their primary and/or secondary sources. These species are also of prime interest because they are, typically part of the secondary organic aerosol (SOA) generated from gaseous PAH oxidation, which is significant in urban environments. Based on ambient air field observations, the objective of this work is to better understand the chemical processes involved in the formation of nitro- and oxy-PAHs, their sources and to identify specific molecules that could further be used as molecular markers of PAH oxidation and SOA formation. Field measurements were performed at both locations in France (Grenoble, 2013 and at the SIRTA station, 2015 (25 km SW from Paris)) over 1 year with samplings, every third day, of the gaseous and particulate (PM10) phases. Intensive observations at SIRTA have been also performed with PM10 samples collected every 4-hour during a period of severe PM pollution event (PM>50 µg m-3 for several days) in March 2015, concomitantly with online measurements (e.g. ACSM, 7λ-Aethalometer). The study of the occurrence of nitro-, -oxy and parent PAHs in the atmosphere, the seasonal and diurnal variations of their concentrations and substance patterns and the assessment of the cancer risk induced by these compounds have been performed. Based on these observations combined with literature knowledge and an extended aerosol chemical characterization, specific molecules of PAH oxidation have been identified. These substances were then used in source-receptor models such as positive matrix factorization (PMF) to apportion the SOA contribution from PAH oxidation on PM10 mass.