Environmental contextChemical mechanisms are an important component of predictive air quality models that are developed using smog chambers. In smog chamber experiments, UV lamps are often used to simulate sunlight, and the choice of lamp can influence the obtained data, leading to differences in model predictions. We investigate the effect of various UV lamps on the prediction accuracy of a key mechanism in atmospheric chemistry. AbstractA new smog chamber was constructed at CSIRO following the decommissioning of the previous facility. The new chamber has updated instrumentation, is 35 % larger, and has been designed for chemical mechanism and aerosol formation studies. To validate its performance, characterisation experiments were conducted to determine wall loss and radical formation under irradiation by UV lamps. Two different types of blacklights commonly used in indoor chambers are used as light sources, and the results using these different lamps are investigated. Gas-phase results were compared against predictions from the latest version of the SAPRC chemical mechanism. The SAPRC mechanism gave accurate results for hydrocarbon reaction and oxidation formation for propene and o-xylene experiments, regardless of the light source used, with variations in ozone concentrations between experiment and modelled results typically less than 10 % over 6-h irradiation. The SAPRC predictions for p-xylene photooxidation showed overprediction in the rate of oxidation, although no major variations were determined in mechanism results for different blacklight sources. Additionally, no significant differences in the yields of aerosol arising from new particle formation were discernible regardless of the light source used under these conditions.