Item

Abstract

Organic aerosol constitutes a major fraction of
the global aerosol burden and is predominantly formed as
secondary organic aerosol (SOA). Environmental chambers
have been used extensively to study aerosol formation and
evolution under controlled conditions similar to the atmo-
sphere, but quantitative prediction of the outcome of these
experiments is generally not achieved, which signifies our
lack in understanding of these results and limits their porta-
bility to large-scale models. In general, kinetic models em-
ploying state-of-the-art explicit chemical mechanisms fail to
describe the mass concentration and composition of SOA ob-
tained from chamber experiments. Specifically, chemical re-
actions including the nitrate radical (NO3) are a source of
major uncertainty for assessing the chemical and physical
properties of oxidation products. Here, we introduce a ki-
netic model that treats gas-phase chemistry, gas–particle par-
titioning, particle-phase oligomerization, and chamber va-
por wall loss and use it to describe the oxidation of the
monoterpenes α-pinene and limonene with NO3. The model
can reproduce aerosol mass and nitration degrees in experi-
ments using either pure precursors or their mixtures and in-
fers volatility distributions of products, branching ratios of
reactive intermediates and particle-phase reaction rates. The
gas-phase chemistry in the model is based on the Master
Chemical Mechanism (MCM) but trades speciation of sin-
gle compounds for the overall ability of quantitatively de-
scribing SOA formation by using a lumped chemical mech-
anism. The complex branching into a multitude of individ-
ual products in MCM is replaced in this model with product
volatility distributions and detailed peroxy (RO2) and alkoxy
(RO) radical chemistry as well as amended by a particle-
phase oligomerization scheme. The kinetic parameters ob-
tained in this study are constrained by a set of SOA forma-
tion and evaporation experiments conducted in the Georgia
Tech Environmental Chamber (GTEC) facility. For both pre-
cursors, we present volatility distributions of nitrated and
non-nitrated reaction products that are obtained by fitting
the kinetic model systematically to the experimental data
using a global optimization method, the Monte Carlo ge-
netic algorithm (MCGA). The results presented here provide
new mechanistic insight into the processes leading to for-
mation and evaporation of SOA. Most notably, the model
suggests that the observed slow evaporation of SOA could
be due to reversible oligomerization reactions in the particle
phase. However, the observed non-linear behavior of precur-
sor mixtures points towards a complex interplay of reversible
oligomerization and kinetic limitations of mass transport in
the particle phase, which is explored in a model sensitivity
study. The methodologies described in this work provide a
basis for quantitative analysis of multi-source data from en-
vironmental chamber experiments but also show that a large
data pool is needed to fully resolve uncertainties in model
parameters.