Abstract A high resolution air quality monitoring campaign (PM, CO2 and CO) was conducted on differently designed station platforms in the Barcelona subway system under: (a) normal forced tunnel ventilation, and (b) with daytime tunnel ventilation systems shut down. PM concentrations are highly variable (6–128 μgPM1 m−3, 16–314 μgPM3 m−3, and 33–332 μgPM10 m−3, 15-min averages) depending on ventilation conditions and station design. Narrow platforms served by single-track tunnels are heavily dependent on forced tunnel ventilation and cannot rely on the train piston effect alone to reduce platform PM concentrations. In contrast PM levels in stations with spacious double-track tunnels are not greatly affected when tunnel ventilation is switched off, offering the possibility of significant energy savings without damaging air quality. Sampling at different positions along the platform reveals considerable lateral variation, with the greatest accumulation of particulates occurring at one end of the platform. Passenger accesses can dilute PM concentrations by introducing cleaner outside air, although lateral down-platform accesses are less effective than those positioned at the train entry point. CO concentrations on the platform are very low (≤1 ppm) and probably controlled by ingress of traffic-contaminated street-level air. CO2 averages range from 371 to 569 ppm, changing during the build-up and exchange of passengers with each passing train.