Soot particles are the most efficient light absorbing aerosol species in the atmosphere, playing an important role as a driver of global warming. Their climate effects strongly depend on their mixing state, which significantly changes their light absorbing capability and cloud condensation nu-clei (CCN) activity. Therefore, knowledge about the mixing state of soot and its aging mechanism becomes an important topic in the atmospheric sciences. The size-resolved (30–320 nm diameter) mixing state of soot particles in polluted megacity air was measured at a sub-urban site (Yufa) during the CAREBeijing 2006 campaign in Beijing, using a volatility tandem differential mobility ana-lyzer (VTDMA). Particles in this size range with non-volatile residuals at 300 • C were considered to be soot particles. On average, the number fraction of internally mixed soot in total soot particles (F in), decreased from 0.80 to 0.57 when ini-tial D p increased from 30 to 320 nm. Further analysis reveals that: (1) F in was well correlated with the aerosol hygroscopic mixing state measured by a CCN counter. More externally mixed soot particles were observed when particles showed more heterogeneous features with regard to hygroscopicity. (2) F in had pronounced diurnal cycles. For particles in the accumulation mode (D p at 100–320 nm), largest F in were ob-served at noon time, with "apparent" turnover rates (k ex→in) up to 7.8 % h −1 . (3) F in was subject to competing effects of both aging and emissions. While aging increases F in by con-verting externally mixed soot particles into internally mixed ones, emissions tend to reduce F in by emitting more fresh and externally mixed soot particles. Similar competing ef-fects were also found with air mass age indicators. (4) Un-der the estimated emission intensities, actual turnover rates of soot (k ex→in) up to 20 % h −1 were derived, which showed a pronounced diurnal cycle peaking around noon time. This result confirms that (soot) particles are undergoing fast ag-ing/coating with the existing high levels of condensable va-pors in the megacity Beijing. (5) Diurnal cycles of F in were different between Aitken and accumulation mode particles, which could be explained by the faster growth of smaller Aitken mode particles into larger size bins. To improve the F in prediction in regional/global models, we suggest parameterizing F in by an air mass aging indica-tor, i.e., F in = a + bx, where a and b are empirical coeffi-cients determined from observations, and x is the value of an air mass age indicator. At the Yufa site in the North China Plain, fitted coefficients (a, b) were determined as (0.57, 0.21), (0.47, 0.21), and (0.52, 0.0088) for x (indicators) Published by Copernicus Publications on behalf of the European Geosciences Union. 4478 Y. F. Cheng et al.: Measurement of the mixing state of soot in the megacity Beijing as [NO z ]/[NO y ], [E]/[X] ([ethylbenzene]/[m,p-xylene]) and ([IM] + [OM])/[EC] ([inorganic + organic matter]/[elemental carbon]), respectively. Such a parameterization consumes lit-tle additional computing time, but yields a more realistic de-scription of F in compared with the simple treatment of soot mixing state in regional/global models.