Abstract. The morphology and effective density of externally mixed black carbon (extBC) aerosols, important factors affecting the radiative forcing of black carbon, were studied using a tandem technique coupling a differential mobility analyzer (DMA) with a single-particle soot photometer (SP2). The study extended the mass–mobility relationship to large extBC particles with a mobility diameter (dmob) larger than 350 nm, a size range seldom included in previous tandem measurements of BC aggregates in the atmosphere. The experiment was conducted at an urban site in Beijing during a 19 d winter period from 23 January to 10 February 2018. Ambient dry particles were selected by the DMA, and the size-resolved extBC particles were distinguished from particles with a thick coating (internally mixed) according to the time delay between the incandescence signal peak and the scattering peak detected by the SP2. The masses of the extBC particles were then quantified. The time differences between the DMA size selection and the SP2 measurement were processed previously. The normalized number size distributions were investigated at the prescribed dmob sizes in the range of 140–750 nm to provide the typical mass of extBC at each dmob. On this basis, the mass–mobility relationship of the ambient extBC was established, inferring a mass–mobility scaling exponent (Dfm) (an important quantity for characterizing the morphology of fractal-like BC aggregates) with a value of 2.34±0.03 in the mobility range investigated in this study. This value is comparable with those of diesel exhaust particles, implying a predominant contribution of vehicle emissions to the ambient extBC in urban Beijing. Compared to the clean period, a higher Dfm value was observed in the polluted episode, indicating a more compact BC aggregate structure than that in the clean period. The effective densities (ρeff) of the extBC in the same dmob range were also derived, with values gradually decreasing from 0.46 g cm−3 at 140 nm mobility to 0.14 g cm−3 at 750 nm mobility. The ρeff values were slightly lower than those measured using the DMA–aerosol particle mass analyzer (APM) system. The difference in ρeff values was likely due to the lower BC masses determined by the SP2 compared to those measured by the APM at the same mobility, since the SP2 measured the refractory BC (rBC) mass instead of the total mass of the BC aggregate, which consists of both rBC and a possible fraction of nonrefractory components measured by the APM. The ρeff values in the 280–350 nm dmob range were much closer to the values for soot aggregates reported in the literature. It might be related to the more compact structure of BC aggregates in this range, resulting from the reconstruction effect by volatile and/or semivolatile components in the atmosphere. The reconstruction effect might also result in a hiatus in the increased dynamic shape factor in the range of 200–350 nm, which presented an overall increase from 2.16 to 2.93 in the 140–750 nm dmob range.