Warm clouds play a key role in cooling Earth’s climate by significantly reflecting shortwave radiation to space (L’Ecuyer et al., 2019). It has been a longstanding question of how aerosol particles influence the formation of clouds, their radiative effect, lifetime, and precipitation, exerting large uncertainty in climate assessment of the anthropogenic radiative forcing in recent Intergovernmental Panel on Climate Change (IPCC) reports (Forster et al., 2021; Myhre et al., 2013).
Aerosol particles can act as cloud condensation nuclei (CCN) and activate to cloud droplets through water condensation and form warm clouds when the air parcels are lifted in the atmosphere. Traditionally, Köhler theory (1936) is used to predict the CCN activation of aerosol particles. That is, as the saturation ratio of water vapour (Sw) increases with adiabatic cooling, aerosol particles consisting of or containing non-volatile hygroscopic constituents take up water and grow in size to remain in equilibrium with the environment. When Sw reaches the critical supersaturation ratio, which depends on aerosol size and chemical composition, aerosol particles activate and grow spontaneously to cloud droplets until the surrounding water vapour is sufficiently depleted (Köhler, 1936). However, other condensable vapours (e.g. NH3, HNO3, HCl, organic compounds) also can co-condense on aerosol particles together with water vapour hence enhancing hygroscopic growth, modifying the Köhler curve (Laaksonen et al., 1998) and influencing cloud formation, referred to as “co-condensation effect”.
The co-condensation effect may significantly influence local climate as semi-volatile compounds are ubiquitous in the atmosphere, but this effect is poorly constrained currently. It is a big challenge to assess hygroscopic growth and CCN activation while accounting for co-condensing semi-volatile compounds, because of unidentified evaporation losses of semi-volatile compounds during drying or heating processes in traditional instruments (Hu et al., 2018), such as HTDMAs (e.g., Duplissy et al., 2009) and CCN counters (e.g., Roberts and Nenes, 2005). Only a few studies investigated co-condensation of nitric acid (Rudolf et al., 2001) and simple organic surrogates (Rudolf et al., 1991; Hu et al., 2018) using modified instruments. Wang and Chen (2019) used an optics-based method and retrieved the hygroscopic growth factor of aerosol particles including co-condensation in ambient conditions in Delhi (India). This method was further combined with a thermodynamic model and revealed that co-condensation of HCl in Delhi contributed by 50 % to the visibility reduction during haze events in winter and can halve the critical supersaturation for CCN activation (Gunthe et al., 2021). A similar co-condensation effect was found in Beijing but with HNO3, which forms particulate nitrate in aerosol particles (Wang et al., 2020). Makkonen et al. (2012) included a nitric acid co-condensation scheme developed by Romakkaniemi et al. (2005) in a global climate model and found that including nitric acid co-condensation can increase the CDNC by 7 % globally. It is challenging to assess the co-condensation effect of organics as these comprise more than tens of thousands of species and the total amount of ambient organic species in the gas and particulate phase cannot be fully detected.
To simulate co-condensation, models based on Raoult’s law combined with the ideal gas equation have been developed in mass (e.g., Donahue et al., 2006) and molar concentrations (e.g., Pankow, 1994; Barley et al., 2009) with complexity ranging from a single inorganic co-condensing species (HNO3) (Kulmala et al., 1993) to considering 2727 organic compounds (Topping and McFiggans, 2012). To account for organic compounds with volatilities differing by orders of magnitudes and aiming for simplicity in computation, Topping et al. (2013) assigned aerosols from different sources to 10 logarithmically spaced volatility bins and found a considerable enhancement of CDNC due to co-condensation in a cloud parcel simulation with different aerosol size distributions, organic aerosol types, and updraft velocities. The volatility basis sets were based on the thermodenuder measurement results by Cappa and Jimenez (2010), which include bins up to log(C*) of 3 and cover only partially intermediate-volatility organic compounds (IVOC) and volatile organic compounds (VOC). Heikkinen et al. (2024) further extended volatility distributions up to log(C*) of 7 in their cloud parcel model by constructing them from the new FIGAERO-I-CIMS (Filter Inlet for Gases and AEROsols coupled to an iodide-adduct chemical ionization mass spectrometer) technique (Lopez-Hilfiker et al., 2014) to compute co-condensation of organic compounds in Hyytiälä, Finland. They found that IVOC and VOC (log(C*) >3) could significantly enhance the co-condensation effect, and that the aerosol size distribution has a large influence on the magnitude of the effect with a nascent ultrafine mode enhancing it. However, the mass spectral peaks from FIGAERO-I-CIMS need to be calibrated with standard species and interpreted in terms of functional groups for an assignment to volatility bins, which could result in biases in volatility distributions (Heikkinen et al., 2024; Voliotis et al., 2022; Peng et al., 2023; Gkatzelis et al., 2021).
The studies performed by Topping et al. (2013) and Heikkinen et al. (2024) both based their conclusions on the assumption of ideal mixing of the constituents in the condensed phase, which provides an upper estimate of the co-condensation effect and may be realistic in the case of an aerosol consisting of highly oxidised organic species. For a more reliable estimate of a mixed organic-inorganic aerosol, volatility and hygroscopicity information need to be combined. However, to the best of the authors’ knowledge, no single study provides information on the volatility distribution of substances over a large vapour pressure range together with hygroscopicity information as required to simulate gas-particle partitioning taking solution non-ideality into account.
Here, we set up two different VBS by integrating information from different experimental and modelling studies to investigate the sensitivity of the combined co-condensation effect of organic and inorganic species on VBS distribution. To simulate the co-condensation effect, we introduced new features into our model. Firstly, we recalculated the condensed mass of organics by accounting for the evaporation loss of semi-volatile compounds during sampling to estimate the mass loss and to provide a more accurate total organic mass for model initialisation. Secondly, we developed our model to include solution non-ideality of organic-inorganic aqueous mixtures. With these new developments, we investigate the sensitivity of the co-condensation effect of organics and inorganics to volatility distribution, updraft velocity and non-ideality for a boreal forest site using field data from Hyytiälä, Finland.