SSCI-VIDE+CARE+ABO:CGO ; International audience ; Chamber studies on the formation of secondary aerosols are mostly performed with purified air as matrix, it is of wide concern in what extent they might be different from the situations in ambient air, where a variety of gaseous and particulate components preexist. Here we compared the photo-oxidation of "toluene + NOx + SO2" combinations in a smog chamber in real urban ambient air matrix with that in purified air matrix. The secondary organic aerosols (SOA) mass concentrations and yields from toluene in the ambient air matrix, after subtracted ambient air background primary and secondary organic aerosols, were 9.0-34.0 and 5.6-12.9 times, respectively, greater than those in purified air matrix. Both homogeneous and heterogeneous oxidation of SO2 were enhanced in ambient air matrix experiments with observed 2.0-7.5 times higher SO2 degradation rates and 2.6-6.8 times faster sulfate formation than that in purified air matrix, resulting in higher in-situ particle acidity and consequently promoting acid catalyzed SOA formation. In the ambient air experiments although averaged OH radical levels were elevated probably due to heterogeneous formation of OH on particle surface and/or ozonolysis of alkenes, non-OH oxidation pathways of SO2 became even more dominating. Under the same organic aerosol mass concentration, the SOA yields of toluene in purified air matrix experiments matched very well with the two-product model curve by Ng et al. (2007), yet the yields in ambient air on average was over two times larger. The results however were much near the best fit curve by Hildebrandt et al. (2009) with the volatility basis set (VBS) approach. (C) 2016 Published by Elsevier Ltd.