[University home]

Centre for Atmospheric Science

ACES Project - Development of secondary organic aerosol (SOA) formation mechanisms


The formation of SOA in atmospheric models has most commonly been treated using empirical approaches based on aerosols yields measured in conventional chamber studies of the oxidation of selected hydrocarbon species. This approach typically assumes the oxidation of a given hydrocarbon into a limited number of notional oxygenated products (usually two) which partition into the aerosol phase, with the yields and partitioning coefficients of the notional products obtained by fitting to the chamber SOA yields. Although such an approach may capture some aspects of the SOA formation process, and has provided a useful basis for interpreting chamber experiments, it is becoming clear from more recent experimental studies and interpretative studies, that chamber aerosol yields are highly dependent on conditions: and it is apparent that any derived lumped product yields and partitioning parameters would also vary significantly with experimental conditions. Given that atmospheric conditions differ considerably from those of conventional chamber studies, and are themselves highly variable, it is highly unlikely that such parameters derived from the chamber datasets give a reliable representation of SOA formation for the range of atmospherically-relevant conditions (i.e. they only really apply to the experimental conditions for which they were derived).

In view of these limitations, recent studies have begun to consider more explicit representations of SOA formation, which take greater account of the details of the oxidation mechanisms. For example, such an approach, based on the highly detailed description of VOC degradation in the Master Chemical Mechanism (MCM), has been applied with some success to interpretation of OA formation in the 2003 heat-wave in the southern UK as part of the NERC TORCH project. The ACES mechanism development work aims to extend that work to a comprehensive and systematic consideration of the relevant BVOC systems, and to use that information to generate validated detailed and reduced representations of SOA formation for application in a range of atmospheric models.

The work described below will be conducted by Atmospheric Chemistry Services (ACS) and The University of Manchester (UM).


Development of detailed gas-phase degradation mechanisms (ACS)

Explicit gas-phase degradation mechanisms will be constructed for a series of BVOC emitted by tropical forests. The selected species will significantly extend the BVOC which are treated in the MCM which are currently isoprene, 2- methylbut-3-ene-2-ol, α-pinene and β-pinene to include key species representative of other monoterpene and sesquiterpene classes. The current version of the MCM has been shown to provide a good basis for the simulation of oxidant and (anthropogenic and biogenic) SOA formation for the conditions appropriate to the boundary layer over NW Europe, carried out as part of the TORCH project. It is the most comprehensive gas-phase degradation mechanism available, providing a direct link to laboratory studies of the kinetics and mechanisms of elementary chemical reactions. For the current work, the construction method will follow an updated MCM protocol, with guidance and input from an international steering committee co-ordinated within the EU ACCENT network. This will ensure that the developed mechanisms will provide a rigorous description of the detailed degradation chemistry for a wide range of conditions, including the comparatively low NOx conditions relevant to the field study location, and will thus provide a robust description of the distributions of organic oxygenated species, and the variety of functional groups they contain. The new mechanisms will be incorporated into a future release of the MCM.


Property estimation and development of gas-aerosol partitioning code (ACS/UM)

Previous investigations of secondary organic aerosol formation have established that there are broadly three mechanisms responsible for the transfer of organic material from the gaseous to the aerosol phase (i) Partitioning into the condensed organic phase; (ii) Partitioning into the condensed aqueous phase; and (iii) Enhancement of mechanisms (i) and (ii) driven by accretion reactions in the condensed phase, leading to the formation of oligomeric compounds of lower volatility. Previous studies based on the MCM considered mechanism "(i)" explicitly, with mechanisms "(ii)" and "(iii)" notionally accounted for by empirically-derived scaling factors.

In the current work, the participation of all three mechanisms identified above will be considered. Condensation coefficients will be estimated for the complete series (probably > 1000) of closed-shell oxygenated organic products generated from the gas phase chemistry of the relevant biogenic presursors. A corresponding code will be developed to represent bulk partitioning of the "monomeric" oxygenated products into the condensed organic and aqueous phases, allowing a basis for representation of total mass concentrations present in the gaseous and aerosol phases. The code will be further refined to provide an explicit description of the formation of specific dimeric and oligomeric species in the condensed organic phase. Recent theoretical assessments of the thermodynamics of accretion reactions for specific oxygenate classes will also be taken into account in the construction of the refined mechanisms.


Evaluation using chamber data (ACS/UM)

The detailed chemical mechanism and the associated gas-aerosol partitioning code described above will be evaluated using the results of the chamber studies on individual VOC systems and VOC mixtures. The mechanisms will be refined, if necessary, to take account of new insights into the reaction mechanisms resulting from the chamber studies. Particular attention will be given to new information relating to the identities of oligomeric species in the condensed phase and to representing plausible routes to their formation. An explicit coupled model of photochemistry and aerosol microphysics has only once been applied to predict the size-dependent partitioning of organic species in chamber experiments. The approach used was to simplify the oxidation scheme to consider only one involatile product of α-pinene, myrcene or sabinene oxidation. Since the condensing product was assumed insoluble, no explicit vapour/liquid/solution thermodynamic treatment was used. In the current work the chamber experiments will also be simulated using an existing combined model of photochemistry, microphysics and aerosol partitioning, updated on the basis of the explicit oxidation scheme developed in this project, the partitioning model based on property estimation activities described above and using an explicit phase-partitioning thermodynamic model. This approach will enable consideration of the full range of BVOC oxidation products reversibly condensing to a wet or dry, inorganic or organic seed aerosol population.


Mechanism reduction (ACS)

The detailed chemical mechanism and the associated gas-aerosol partitioning code will be applied in an "emissions-driven" two-layer box model designed to represent processes occurring in the canopy and above. This will allow the performance of the mechanism to be tested in broad relation to the campaign observations for oxidant species and aerosol. The model will be used to identify the major precursors to SOA formation under the simulated atmospheric conditions, and to establish the major organic oxygenated components of which the SOA is comprised. This information will be used to define a reduced series of emitted VOC and SOA components for representation in a reduced chemical mechanism. The performance of the mechanism will be tested and optimised by comparison with the detailed mechanism using the two-layer atmospheric box model for a range of conditions. This programme of mechanism development work will thus yield a comprehensive reduced description of SOA formation which is traceable to the validated detailed mechanisms and chamber results, and which can be applied in the scaling-up modelling activities which also form part of the ACES project.