The reliable chemical characterization of non-exhaust emissions generated by brakes is of fundamental importance in order to provide correct information for source apportionment studies as well as for their toxicological and environmental assessment. Nowadays, the best option to obtain samples of PM10 emissions composed only by material worn from the tribological interface, i.e. the braking disc (BD) and the friction material (FM) rubbing surfaces, is to sample them on suitable collection filters at a dedicated dyno-bench, during a standard braking test cycle. In particular, the use of enclosed dyno-bench is necessary for excluding other spurious contributions from the environment, while defined test cycles are necessary to simulate standard driving conditions. Nevertheless, different braking cycles are usually characterized by different overall temperature profiles or energy parameters, which in the end have significant influence on the wear and the oxidation of the materials involved in the tribological event. Thus, when collection is performed at the dyno-bench, the chemical composition of the PM10 emissions is expected to be influenced by the temperature and the test cycle. At this regard, the work investigates the difference in terms of chemical composition of PM10 samples generated by two tribo-couples (standard cast iron BD coupled respectively with standard ECE R90 FM and prototypal FM based on inorganic binder) during the 3h-LACT and the WLTP-Brake test cycles. More in detail, a wide physico-chemical characterization performed by SEM/EDXS and XRD analysis is used to unveil how braking cycles influence both the elemental and phase composition of the generated emissions. This will provide useful insights regarding the correlation between driving conditions, relative consumption of the tribological interface and chemistry of the emitted particulates.