Soot particles are the most strongly light-absorbing particles commonly found in the atmosphere. They are major contributors to the radiative budget of the Earth and to the toxicity of atmospheric pollution. Atmospheric aging of soot may change its health- and climate-relevant properties by oxidizing the primary black carbon (BC) or organic particulate matter (OM) which, together with ash, comprise soot. This atmospheric aging, which entails the condensation of secondary particulate matter as well as the oxidation of the primary OM and BC emissions, is currently poorly understood. In this study, atmospheric aging of wood-stove soot aerosols was simulated in a continuous-flow reactor. The composition of fresh and aged soot particles was measured in real time by a dual-vaporizer aerosol-particle mass spectrometer (SP-AMS). The dual-vaporizer SP-AMS provided information on the OM and BC components of the soot as well as on refractory components internally mixed with BC. By switching the SP-AMS laser vaporizer off and using only the AMS thermal vaporizer (at 600 °C), information on the OM component only was obtained. In both modes, OM appeared to be generated largely by cellulose and/or hemicellulose pyrolysis and was only present in large amounts when new wood was added to the stove. In SP-AMS mode, BC signals otherwise dominated the mass spectrum. These signals consisted of ions related to refractory BC (rBC, C 1-5 + ), oxygenated carbonaceous ions (CO 1-2 + ), potassium (K + ), and water (H 2 O + and related fragments). The C 4 + : C 3 + ratio, but not the C 1 + : C 3 + ratio, was consistent with the BC-structure trends of Corbin et al. (2015c). The CO 1-2 + signals likely originated from BC surface groups: upon aging, both CO + and CO 2 + increased relative to C 1-3 + while CO 2 + simultaneously increased relative to CO + . Factor analysis (positive matrix factorization) of SP-AMS and AMS data, using a modified error model to address peak-integration uncertainties, indicated that the surface composition of the BC was approximately constant across all stages of combustion for both fresh and aged samples. These results represent the first time-resolved measurements of in situ BC surface aging and suggest that the surface of beech-wood BC may be modelled as a single chemical species.