Modelling the impact of viscosity on Secondary Organic Aerosols formation with the SOAP thermodynamic model inside a 3D air quality model
Résumé
Numerous studies (e.g. Virtanen et al., 2010) show that the partitioning between the gas and particle phases of Semi-Volatile Organic Compounds (SVOC) can differ significantly from thermodynamic equilibrium due to the high viscosity of the organicphase. However, to our knowledge the impact of organic-phase viscosity on Secondary Organic Aerosol formation has never been investigated in 3D air quality models, which usually assume that organic aerosols are not viscous. The impact of the particle viscosity on secondary organic aerosols (SOA) is studied here by coupling an air quality model to the Secondary Organic Aerosol Processor (SOAP Couvidat and Sartelet, 2015) thermodynamic model. SOAP can compute the partitioning of SVOC concentrations either by assuming thermodynamic equilibrium or by computing the dynamic evolution of concentrations according to the kinetics of condensation/evaporation and particle-phase diffusion (as a function of the organic-phase coefficient diffusion). To compute the evolution of the SVOC partitioning, organic particles are separated into several layers with the external layer at the gas/particle interface and the internal layer at the center of the particle. To accurately solve the equations of diffusion of SVOCs inside particles, a high number of layers would be needed (around 100). However, this discretization would lead to a system that is too complex to be implemented in an air quality model. In SOAP, a simplified representation was designed to represent implicitly the diffusion of SVOC inside the particle with a low number of layers (between 2 and 5). A comparison between the explicit representation and the implicit representation of diffusion is shown in Figure 1. SOAP was implemented in the 3D air quality model Polyphemus and SOA concentrations were simulated over Europe during July 2012. Concentrations were simulated assuming either that particles are not viscous (the condensation/evaporation is not limited by the diffusion inside the particle) or that particles are infinitely viscous (the diffusion inside the particle is very slow and compounds do not diffuse inside the particle). Assuming infinitely viscous particles leads to a slight increase of SOA concentrations (SVOC concentrations in the particle phase). Less volatile compounds appear to not be affected by the viscosity of particles, but the concentrations of more volatile compounds increased for viscous particles and could therefore exceed concentrations at equilibrium. Indeed, for volatile organic compounds, in the infinitely viscous simulation the compounds could condense even if the compounds would evaporate for a non-viscous aerosol.