Title et al. 2016) and introduce isoprene/aqueous chemistry


a      M. Surdu
Earth Observatory of Singapore
Nanyang Technological University
50 Nanyang Ave, Singapore
E-mail: [email protected]


Abstract: Abstract
Text, 800-1000 characters.

Things to do

-talk more about underestimation of SOA from models
(Hodzic et al. 2016) and introduce isoprene/aqueous chemistry after

-mention isoprene as a new source of SOA, with some history (i.e.
experiments in the 80s/90s showed no SOA forming potential but new studies show
tremendous yields), Shrivastava_et_al-2017, Carlton et al. 2009

aqueous chemistry as a new source of SOA (Lim et al 2010)

-talk about wall effects in chambers!

1. Introduction

The Earth’s climate
is affected by atmospheric aerosols, liquid or solid particles suspended in
air, from a multitude of sources, both biogenic and anthropogenic. Aerosols can
be classified in two divisions, specifically primary aerosols and secondary
aerosols. Examples of biogenic sources of primary particles include volcanic
aerosol and desert dust whereas primary anthropogenic aerosols are a result of
industrial emissions, biomass burning and fossil fuel combustion, amongst
others. Secondary aerosol particles are produced by oxidation of gas to form
low volatility compounds that nucleate or partition on other existing aerosol
particles in a process of gas-to-particle conversion. 1 The
atmosphere could be thought of as a large oxidative chemical reactor, where
oxidants (OH radicals, O3, NO3 and others) react with
volatile organic compounds (VOCs) to form semivolatile or low-volatility
organic compounds.27

This review
concentrates on organic aerosol (OA), in particular secondary organic aerosol
(SOA), whose composition, concentration and formation is generally less well
understood than that of primary organic aerosol (POA). There is currently
increased interest in SOA related to environmental concerns about the role
these components have on climate and human health. It has recently been
determined that there is a discrepancy between current modelled SOA formation
and removal rates and the real atmospheric rates, affecting the global SOA
budget. 26 Sources of SOA not previously discussed are also being identified
in current literature from research in isoprene – a biogenic volatile organic
carbon – and aqueous aerosol chemistry. 2728 In recent years, it has been
observed that previous chemical transport models consistently underpredicted
ambient atmospheric OA concentrations 7 and that the concentration of
oxidized OA is in fact higher than that of POA outside urban areas 8,
establishing SOA processes as a popular topic of study in laboratories across
the world.

Atmospheric aerosols
have been shown to play a key role in many environmental processes by
scattering and absorbing solar radiation, affecting the formation of clouds and
participating in heterogeneous chemical reactions in the atmosphere. 2 There
is evidence that SOA, which makes up an important part of fine particles less
than 2.5 µm in diameter (PM2.5), has a great impact on human health,
exerting negative effects on the pulmonary and cardiovascular systems. 4 It
has been reported that secondary particles in particular increase the
likelihood of contracting myocardial infarction, even when compared to other
sources of PM2.5. 25 They have long been shown to affect global climate
change through significant contributions to radiative forcing, applying a
cooling effect on the planet. 3. Thus, the study of SOA origins, formation
and processes is essential for understanding, predicting and ultimately
preventing the impact of SOA on health and climate. This review aims to discuss
the methods

used to study SOA and advances in their
development over time.

simulation chambers (“smog” chambers or environmental chambers) have been
around since the late 1960s, with the Bureau of Mines Laboratories, Oklahoma
(1966) and the University of North Carolina (1972) chambers being some of the
first physical models for learning about atmospheric chemistry 9. Such
chambers are typically large (usual range from 2 up to 300 m3) PTFE (Polytetrafluoroethylene)
or FEP (Fluorinated ethylene propylene) bags and simulate SOA formation and
oxidation with residence times of several hours, with experiments carried out
in batch mode. The concentration of oxidants in environmental chambers is similar
to that in the atmosphere with OH concentrations in the range 106-107
molecules cm-3 and ozone concentrations of several hundred ppbv.
Therefore, the aim of environmental chambers is to mimic the ambient atmosphere
as closely as possible.

Nevertheless, the relatively
low oxidant concentrations in environmental chambers are, amongst others, a
major limitation of chamber techniques. This has led to the development of
aerosol oxidation flow reactors in the mid-2000s, most notably the Potential
Aerosol Mass (PAM) reactor 10, enabling the rapid study of SOA formation and
The PAM reactor has been used in a considerable amount of experiments since,
including SOA formation from urban ambient air 22, biomass burning 13 and
ambient pine-forest air 23. Results obtained by oxidation with the PAM
reactor have been shown to agree reasonably well with environmental chamber
data. 15

Using results from
recent developments in SOA studies, this review analyses the various methods in
which SOA can be formed in the laboratory by discussing their advantages and

2. History of Organic Aerosol

Atmospheric organic
compounds were mostly unstudied before the 1950s with very little being known
about their composition beyond the fact that they contain methane and
formaldehyde. However, with the increase in haze, smog and air pollution
especially in the greater Los Angeles area, research slowly started to be
diverted to the topic of organic aerosols in the atmosphere. Mader et al. published
the first detailed report on the chemical composition of atmospheric organic
compounds in 1952, the same year that Haagen-Smit observed the photochemical
reaction between VOCs and nitrogen oxides (NOx) to form ozone.

The number of
organic compounds identified in the atmosphere has risen exponentially, from a
reported 606 compounds in 1978 to 2857 in 1986 to Goldstein and Galbally’s
estimate of 104-105 atmospheric organic compounds in
2007. Likewise,

atmospheric aerosol rapidly became a popular
research subject after the 1980s as it received increased attention in
atmospheric science literature (Figure 1).

            Despite the considerable advance in
understading SOA formation and atmospheric processes, many current climate
models do not include all recently discovered processes. 27 In fact,  a recent intercomparison of organic aerosol
models (Tsigaridis et al. 2014) identifies variances of a whole order of
magnitude in estimates of annual SOA production rate (range 13 – 119 Tg y-1).
30 A better understanding and inclusion of these developments is needed to
more accurately predict climate forcing as well as estimate the influence of
anthropogenic emissions and land use changes to the global aerosol budget.

            Isoprene (2-methyl-1,3-butadiene, C5H8)
is a biogenic VOC with the largest global atmospheric emissions of all VOCs
except methane, estimated at ~600 Tg y-1. Although isoprene has been
studied in outdoor environmental chamber studies as early as 1982 (Kamens et
al. 1982) and 1991 (Pandis et al. 1991), significant SOA formation was not
observed at the time. Thus, isoprene was not considered to be a noteworthy SOA
precursor. However, more recently there has been increased interest in SOA formation
from isoprene and it is now suggested that isoprene SOA formation yields are in
fact significant. 27 The early isoprene chamber studies did not show
significant SOA production since they used experimental conditions not suitable
for SOA formation (low OH concentration, low seed aerosol concentration, high
temperature). Even with a relatively minor yield (e.g. 1%), the overall SOA
formation contribution of isoprene could be significant (e.g. 6 Tg y-1)
due to the large global atmospheric emissions of isoprene. 32

            Another explanation for the
discrepancy between atmospheric observations and climate models is proposed by
Lim et al. (2010) by considering aqueous aerosol chemisrtry. They propose that
SOA can form through reactions in atmospheric waters such as clouds, fogs and
aerosol water. 33

Figure 1. Number of papers
having atmospheric science published every year in the refereed literature
(Figure from Fuzzi et al. 2015). 28



As discussed in section 2, there is a discrepancy between ambient
atmospheric SOA data and SOA data predicted by models. Model predictions of SOA
are consistently substantially lower than ambient observations. Since the vast
majority of SOA models are based on SOA formation studied in environmental chamber
experiments in the laboratory, it has been proposed that this underestimation
is due to losses of semi volatile vapours to chamber walls. 31

In order to be able
to compare SOA formation processes in oxidation chambers and flow reactors to
the atmosphere, some assumptions have to be made.       Firstly,             it
must be assumed that the kinetics of processes in the laboratory – occurring at
high oxidant concentration and low exposure time- can be extrapolated to
atmospheric processes occurring at low oxidant concentration and long exposure
time. Moreover, the assumption that the lower exposure time in laboratory
experiments does not limit the nucleation and phase partitioning of SOA must
also be made. 5

chambers are affected by both wall loss of oxidized vapours and particles. This
wall loss causes residence times in chambers to be limited to several hours.
6 Moreover, while loss of particles to the walls are typically accounted for
in laboratory experiments, vapour wall loss does not routinely come into
consideration. 31

The PAM flow reactor
produces OH radicals in the absence of NOx, through the reaction

 followed by

, where O3
is produced by irradiation of O2 with an external mercury lamp
(?=185 nm). O(1D) was generated by the UV photolysis of O3
inside the PAM reactor using four mercury lamps (?=254 nm). 6 In
environmental chambers, OH radicals are typically produced by UV photolysis
(?=350 nm) of hydrogen peroxide (H2O2) in the absence of
NOx or by UV photolysis of either nitrous acid (HONO) or methyl nitrite
(CH3ONO) in the presence of NOx.

However, these comparisons need to be extended over a
wider range of reactants and experimental conditions than are currently
available2 .

While smog chambers
are designed to match atmospheric conditions as closely as possible in order to
more accurately simulate SOA processes in the real atmosphere, including their
nonlinearities and other complicating factors, aerosol flow reactors are
typically designed to only focus on isolating a particular aspect of SOA
chemistry at a time. Instead of mimicking the ambient atmosphere, flow reactors
are usually employed to tackle a specific feature of SOA chemistry.


The uncertainty in
calculating the total amount of atmospheric SOA is large, with Volkameer et al.
(2006) finding that the proportion of anthropogenic atmospheric SOA may be as
high as 33% as contrasted to the values of approximately 10% from previous
studies. 24 Typically, the total amount of atmospheric SOA is calculated by
using SOA yields obtained from laboratory data for known SOA precursors as well
as their emission factors and profiles. This method is a cause of uncertainty
as not all of the SOA precursors are taken into consideration and their yields
may not be known – VOC data is less commonly available.

A different method
was proposed by Kang et al. 2007 10 which is useful both in measuring
potential aerosol mass (PAM) in the atmosphere as well as examining SOA
processes in a laboratory environment. Unlike the previous procedure of
estimating the amount of potential SOA from known precursors, Kang et al. 2007
propose the rapid oxidation of all precursor gases with extreme amounts of
oxidants to low volatility compounds in order to form aerosols. This SOA
formation aims to simulate all of the photo-oxidation processes in the
atmosphere but on a much quicker timescale due to the extreme oxidant


Some shortcomings of
typical flow reactors and especially environmental chambers include the long
residence time (around 100s in flow reactors and several hours in environmental
chambers), making it difficult to study transient processes. Moreover, smog
chambers and large flow reactors are stationary, deeming them inconvenient for
use on the study of emissions from large moving sources such as aircraft or
ship engines. Nevertheless, advancements have recently been made to either
mobilize medium-sized smog chambers or develop smaller more portable flow
reactors.  Presto et al. (2011) have
characterised primary particle and gaseous emissions from an in-use aircraft
engine using a medium-sized chamber in the form of a 7m3 Teflon bag.
20 Platt et al. (2013) have adopted a similar chamber, the Paul Scherrer
Institute (PSI) mobile smog chamber, consisting of a 9-12m3 Teflon
bag in order to study SOA formation of gasoline vehicle emissions. 21 The PSI
mobile chamber features a modular design allowing flexibility of installation
and increased transportability. An oxidation flow reactor developed in order to
study SOA formation from vehicle exhaust during a transient driving cycle is
the TSAR (TUT Secondary Aerosol Reactor). The TSAR is a 3.3L quartz glass
cylinder with OH radicals produced from the photolysis of ozone at 245 nm UV
radiation from two mercury lamps. 11 
Potential advantages of the TSAR and other oxidation flow reactors as
opposed to large environmental chambers could include short residence time
(around 40s for the TSAR), allowing the monitoring of constantly changing
situations due to increased time resolution, higher degree of oxidation and

5. Problems and Limitations

Renbaum and Smith
(2011) use a kinetic study to suggest that flow tube reactor conditions (high
oxidant concentration but short exposure times) can be extrapolated to ambient
atmospheric conditions (low oxidant concentration, long exposure times).  It has been observed that the oxidant
concentration and reaction time can be considered to be interchangeable
parameters as the oxidant exposure – defined as the integral of the oxidant
species concentration and the sample residence time- is conserved. 166

However, flow
reactors are not free of experimental artefacts, the most significant of which
may be the nonreactive absorption of O3 on the surface of particles
due to the high OH concentrations. This absorption leads to active sites on the
surface of particles to be blocked and hence unavailable for heterogeneous
oxidation by OH radicals. 17. Renbaum and Smith (2011) suggest that O3
absorption leads to a ~30% decrease in the rate of heterogeneous oxidation by OH

Laboratory chambers have been shown to be
substantially affected by wall loss of semi-volatile compounds which could lead
to a considerable underestimation of organic compound and SOA yields. 18

As laboratory SOA data is of key importance in
simulating SOA formation and processes in the atmosphere as well as climate
models it is necessary for laboratory conditions to be representative of those
in the atmosphere. When comparing laboratory SOA results to ambient organic
aerosol it has been found that laboratory SOA is more similar to semi-volatile
oxidized organic aerosol (SV-OOA) and does not usually become as oxidized as
ambient low-volatility oxidized organic aerosol (LV-OOA). This observation has
been attributed to the higher SOA precursor concentrations (loadings) used in
laboratory experiments and/or limited oxidant exposure in smog chamber
experiments. 19 3 

One advantage of the PAM
reactor is that wall losses are reduced relative to smog chambers.

6. Summary and Outlook

“Therefore, utilizing flow tubes and smog chamber reactors with different
designs can complement each other, making it possible to extend studies over a
range of parameters unattainable by either method individually, and ultimately
lead towards a better understanding of atmospheric aerosol processes. The
results of laboratory aerosol experiments are used as inputs to climate models.
Therefore, the evaluation of experimental uncertainties associated with
measurements is needed for reliable application. The characterization of
different reactor designs is important to establish the reliability of the
experimental techniques.”4 

More info from Lambe et al.

Lambe et al. 2015. Rephrase.
Comparisons between reactors/chambers/ambient SOA.

This is for chambers, flow
reactors seem to be ok. Check lambe et al. 2015

Lambe et al. 2011