Model Framework for the Evolving Shape of Compacting Soot Aggregates

PAYTON BEELER, Joel Corbin, Laura Fierce, Pacific Northwest National Laboratory

     Abstract Number: 349
     Working Group: Carbonaceous Aerosols

Abstract
Black carbon (BC) particles formed via combustion processes are emitted as fractal aggregates. Condensation of co-emitted species onto the surface of BC aggregates soon after emission can lead to restructuring of the aggregates into more compact shapes. Compaction of BC aggregates impacts their optical properties and could subsequently affect the overall radiative impact of BC. Several laboratory studies have measured the evolving shape of BC aggregates as coatings are applied to the BC surface. However, the results of these studies cannot be easily incorporated into chemical transport and Earth system models, which largely assume that BC particles exist as spheres. The most recent attempts to account for BC compaction rely on a threshold coating amount, beyond which BC particles are considered spherical. These approaches do not account for the continuously evolving shape of BC aggregates throughout the entire coating process. Here, we combine data from laboratory measurements of BC compaction with bottom-up simulations of BC aggregation to develop an empirical model that predicts the collapse of BC aggregates given their composition and initial size. The empirical model approximately reproduces the measured trends of BC compaction as a function of coating-to-BC volume ratio and number of primary particles in the aggregate and additionally predicts the fractal dimension of BC aggregates, an important input for BC optics models. The developed framework provides a continuous and differentiable relationship that connects particle size and composition to the shape of both the overall particle and the BC core. Incorporation of our results into future versions of chemical transport and Earth system models will allow for approximation of BC particle shape given properties already tracked in these models.