International audience

Dinitrogen pentoxide (N₂O₅) is an important nighttime reservoir for nitrogen oxides (NO_x). Heterogeneous hydrolysis on aerosol particles is a major loss pathway for N₂O₅, removing NO_x from the atmosphere. Current models use the bulk composition of the particle population to calculate the N₂O₅ reaction probability (γN2O5). While this is appropriate when the aerosol is internally mixed, it remains an open question of how large the error is when the aerosol has a more complex mixing state, which is common in the real atmosphere. To better understand the role of the mixing state in calculating γN2O5, we used the stochastic particle-resolved model PartMC-MOSAIC to generate thousands of populations covering a wide range of mixing states. Our scenario library focused on submicron aerosols in urban environments, specifically examining mixtures of carbonaceous primary aerosols, secondary organic aerosols, and the inorganic species sulfate, nitrate, and ammonium. We excluded primary emissions of mineral dust or sea salt aerosols and focused on the processes in the residual layer during nighttime. To evaluate γN2O5, we compared three commonly used parameterizations: Riemer et al. (R09), Davis et al. (D08), and Bertram and Thornton (BT09). The results were qualitatively similar and highlighted that the internal-mixing assumption could lead to either over- or underestimation of γN2O5. The largest range of relative errors was obtained for the BT09 parameterization, ranging from −100% to 57%. We present a detailed process analysis that explains the reasons for the observed differences. Errors of this magnitude in reaction probability result in errors in the prediction of N₂O₅ in the range of ±10% and of total nitrate in the range of ±3%. The errors in ozone concentration were negligible, at less than 1%. Copyright © 2025 American Association for Aerosol Research

A marine coccolithophore (Emiliania huxleyi) and coccolithovirus (EhV-207) were grown together in a marine aerosol reference tank (MART) to investigate how the viral lysis of phytoplankton affects the formation of immersion mode ice nucleating particles (INPs) in sea spray aerosol (SSA). The mean ice nucleation temperatures of SSA produced during viral infection were slightly lower (–28.5 °C) than pre-viral infection (–27.5 °C). Ice nucleation temperatures were relatively low, indicating that organic matter from E. huxleyi is less effective as an INP than phytoplankton examined in previous studies. E. huxleyi is covered in calcium carbonate coccoliths and contains high intracellular concentrations of dimethylsulfoniopropionate (DMSP). The ice nucleation efficiencies of purified components were measured to better understand this model system. Purified coccoliths were moderately effective INPs (–25.3 ± 0.4 °C at 5 x 10² mg L^-1 (mean ±pooled SD)) that showed a concentration effect, with lower freezing temperatures at lower concentrations. Coccoliths and DMSP were both weakly efficient INPs. In comparison, purified phytoplankton viruses (EhV-207 and CtenRNAV-01) infecting coccolithophores and diatoms, respectively, did not affect freezing at temperatures warmer than the procedural blank. Our results suggest that E. huxleyi, in contrast to diatoms and cyanobacteria, is not a significant source of immersion mode INPs to the marine atmosphere, despite its broad distribution in the global ocean and large-scale bloom formation.