Review
Chemistry of secondary organic aerosol: Formation and evolution of low-volatility organics in the atmosphere

https://doi.org/10.1016/j.atmosenv.2008.01.003Get rights and content

Abstract

Secondary organic aerosol (SOA), particulate matter composed of compounds formed from the atmospheric transformation of organic species, accounts for a substantial fraction of tropospheric aerosol. The formation of low-volatility (semivolatile and possibly nonvolatile) compounds that make up SOA is governed by a complex series of reactions of a large number of organic species, so the experimental characterization and theoretical description of SOA formation presents a substantial challenge. In this review we outline what is known about the chemistry of formation and continuing transformation of low-volatility species in the atmosphere. The primary focus is chemical processes that can change the volatility of organic compounds: (1) oxidation reactions in the gas phase, (2) reactions in the particle phase, and (3) continuing chemistry (in either phase) over several generations. Gas-phase oxidation reactions can reduce volatility by the addition of polar functional groups or increase it by the cleavage of carbon–carbon bonds; key branch points that control volatility are the initial attack of the oxidant, reactions of alkylperoxy (RO2) radicals, and reactions of alkoxy (RO) radicals. Reactions in the particle phase include oxidation reactions as well as accretion reactions, non-oxidative processes leading to the formation of high-molecular-weight species. Organic carbon in the atmosphere is continually subject to reactions in the gas and particle phases throughout its atmospheric lifetime (until lost by physical deposition or oxidized to CO or CO2), implying continual changes in volatility over the timescales of several days. The volatility changes arising from these chemical reactions must be parameterized and included in models in order to gain a quantitative and predictive understanding of SOA formation.

Introduction

Organic compounds make up a substantial fraction of atmospheric fine particulate matter, accounting for 20–90% of aerosol mass in the lower troposphere (Kanakidou et al., 2005). A full understanding of the health, climate, and visibility effects of atmospheric particulate matter thus requires the detailed characterization of the sources and fates of organic aerosols, so that their loading in the atmosphere can be accurately modeled. Given the large number and variability of chemical constituents, sources, and possible chemical transformations of organic aerosol, such characterization presents a major challenge for both experiments and models. The controlling factors and effects of organic particulate matter thus remain highly uncertain, and are the subject of a number of recent review articles (Jacobson et al., 2000; Turpin et al., 2000; Seinfeld and Pankow, 2003; Kanakidou et al., 2005; Pöschl, 2005; Fuzzi et al., 2007; Goldstein and Galbally, 2007; Rudich et al., 2007).

Probably most uncertain is the formation and evolution of secondary organic aerosol (SOA), particulate matter formed by the chemical transformation of atmospheric organic compounds. The most commonly studied (and probably most atmospherically important) mechanism of SOA formation is the oxidation of volatile organic compounds (VOCs), forming products of lower volatility that subsequently partition into the condensed phase. However, reactions of less-volatile organics may lead to the formation of particulate matter as well, so SOA may also be formed from chemical reactions of organic compounds emitted originally in the condensed phase (Robinson et al., 2007).

SOA accounts for a large, and often dominant, fraction of total organic particulate mass, based on several complementary measurements of ambient aerosol, such as the ratios of organic mass (OM) to organic carbon (OC) (e.g., Turpin and Huntzicker, 1995; Lim and Turpin, 2002), loadings of water-soluble OC (Sullivan et al., 2006), and level of oxidation from online aerosol mass spectrometry (AMS) (e.g., Zhang et al., 2005b; Lanz et al., 2007). As a result, a major focus of laboratory studies of organic aerosol has been the quantification of SOA formation from individual precursors for integration into atmospheric chemical transport models. The standard view is that SOA formation is dominated by a few classes of VOCs (mostly monoterpenes and aromatic compounds) that form aerosol with yields readily measured in laboratory chamber studies. However, models informed by such chamber measurements do not always capture the variability of observed SOA loadings (Heald et al., 2006), and often predict far less SOA than is observed (de Gouw et al., 2005; Heald et al., 2005; Volkamer et al., 2006). This underestimation of SOA strongly suggests the importance of additional pathways of SOA formation not typically studied in experiments or included in models.

The identification of the most important SOA-forming reactions, and hence the accurate prediction of atmospheric SOA, requires an understanding of which pathways form low-volatility organics, compounds of sufficiently low vapor pressures to be present in the condensed phase. This includes compounds present entirely in the condensed phase (nonvolatile organics) as well as those that may be present in appreciable amounts in both the gas and particle phases (semivolatile organics). This definition of semivolatile organics is quite broad, involving saturation vapor pressures spanning at least seven orders of magnitude (Donahue et al., 2006), and so includes a significant fraction of atmospheric organics.

The key concept underlying modern treatments of SOA is that it is composed predominantly of semivolatile organics (Pankow, 1994a, Pankow, 1994b; Odum et al., 1996), allowing for the description of SOA formation in terms of gas–particle partitioning. The absorptive partitioning of semivolatiles is described by the theory of Pankow, 1994a, Pankow, 1994b, defining an equilibrium partitioning coefficient Kp:Kp=PGMin which G is the mass concentration (mass per volume air, e.g. μg m−3) of the semivolatile species in the gas phase, P is the mass concentration (μg m−3) of the semivolatile species in the particle phase, and M is the mass concentration (μg m−3) of the total absorbing particle phase. The partitioning coefficient Kp (m3 μg−1) is thus inversely proportional to the saturation vapor pressure (c*) of the pure semivolatile compound. M refers only to the particulate matter participating in absorptive partitioning (organic aerosol into which semivolatiles can partition, and possibly aqueous aerosol in the case of highly water-soluble organics). By Eq. (1), as long as any absorbing mass is present, some fraction of a given semivolatile compound can partition into the particle phase, even if its gas-phase concentration is below its saturation vapor pressure. The fraction F of a semivolatile compound in the particle phase is given byF=PP+G=MKp1+MKp=11+c*/MHence as the amount of absorbing material (M) increases, compounds of higher volatility (higher c*, lower Kp) will increasingly partition into the condensed phase.

The dependence of F on absorbing mass M and vapor pressure of the semivolatile species c* is illustrated in Fig. 1. When c* is equal to M, half of the semivolatile mass resides in the particle phase. If c*⪡M, essentially all of the semivolatile species are in the particle phase; conversely, if c*⪢M, its fraction in the particle phase approaches zero. Fig. 1 shows the nature of F both as a function of c* at a fixed aerosol loading M (left panel) and as a function of M for a single semivolatile compound of vapor pressure c* (right panel).

Odum et al. (1996) showed that SOA yield Y (defined as ΔM/ΔHC, the mass of aerosol formed per mass of hydrocarbon reacted) can be expressed in terms of the formation of a collection of semivolatile compounds:Y=ΔMΔHC=MiαiKp,i1+MKp,iin which Kp,i and αi are the partitioning coefficient and mass yield, respectively, of compound i. SOA yield from a given precursor is therefore not a stoichiometric quantity, but rather increases with increasing total organic particulate loading, consistent with a wide range of experimental results (e.g., Odum et al., 1996; Seinfeld and Pankow, 2003). The SOA-forming potential of a given reaction is determined by the semivolatile product yields (α's) and volatilities (Kp's), which together make up a “volatility distribution” of the reaction products.

In principle, aerosol formation can be calculated by carrying out the summation in Eq. (3) over all semivolatile compounds formed in a given reaction, but this degree of detail is generally infeasible owing to the large number of products formed and the difficulty in measuring all individual semivolatile compounds. Instead, two surrogate products (i=2) have traditionally been used to express the volatility distribution from SOA-forming reactions. This “two-product model” of SOA formation, shown in Fig. 2, generally represents laboratory SOA growth data well (e.g., Seinfeld and Pankow, 2003; Keywood et al., 2004b), and can be incorporated into atmospheric chemistry models in a straightforward manner (e.g., Chung and Seinfeld, 2002; Koo et al., 2003; Tsigaridis and Kanakidou, 2003; Heald et al., 2005; Henze and Seinfeld, 2006).

Recently, Donahue and coworkers (Donahue et al., 2006; Presto and Donahue, 2006; Pathak et al., 2007) demonstrated that partitioning over a broad range of organic aerosol loadings (including typical ambient levels, 0.1–20 μg m−3) is more accurately represented by a larger number of products (typically i=10) spanning a wide range of volatilities. In this approach the volatility distribution is represented by binning all organics by volatility, with the bins defined by a set of prescribed vapor pressures (the “volatility basis set”) (Donahue et al., 2006). Fig. 3 shows how the fractional distribution in Fig. 1 is represented in terms of both the “two-product model” and the “volatility basis set”, and how partitioning in each model is affected by a change in absorbing aerosol loading M.

The description of SOA formation in terms of semivolatile partitioning (Eqs. (1), (2), (3)) is the primary focus of an earlier review (Seinfeld and Pankow, 2003). The subject of the present review is the detailed chemistry of the formation and evolution of semivolatile organics, which controls their amount and volatility, and remains poorly understood. In particular, we focus on the three primary factors that determine the volatility, and hence the SOA-forming potential, of organic compounds in the atmosphere:

  • (1)

    oxidation reactions of gas-phase organic species, which lower volatility by addition of functional groups but can also increase volatility by cleavage of carbon–carbon bonds;

  • (2)

    reactions in the particle (condensed) phase, which can change volatility either by oxidation or formation of high-molecular-weight species; and

  • (3)

    the extent to which these reactions occur, as the volatility distribution of oxidation products will continually evolve as a result of ongoing chemistry.

These correspond to three major developments arising from recent studies of SOA formation: the description of SOA production in terms of known VOC oxidation mechanisms, the characterization of complex chemistry occurring within particles, and an improved understanding of SOA formation kinetics. In the following three sections of this review, each of these factors will be discussed individually. Together they suggest possible explanations for discrepancies between modeled and measured ambient aerosol, and suggest areas of future research, discussed in the final section.

Because of the focus on the chemistry of the formation and evolution of low-volatility organics, this review is not intended to be a comprehensive literature review of all studies of SOA formation. Many important aspects of the field, such as laboratory techniques, ambient measurements, model simulations, aerosol properties, and new particle formation are not discussed in detail. Laboratory studies of SOA formation under carefully controlled conditions represent the foundation from which our understanding of the chemistry of SOA formation is derived. This review relies heavily on what has been learned from such studies.

Section snippets

Gas-phase oxidation

Gas-phase oxidation, initiated by reaction with species such as the hydroxyl radical (OH), nitrate radical (NO3), and ozone (O3), is the primary process by which the volatilities of organic species in the atmosphere evolve. Oxidation of a VOC can produce species of sufficiently low vapor pressure to be condensable, leading to the formation of SOA; products of higher volatility than the parent VOC (such as CO2, CH2O, etc.) may be formed as well. As a result of the chemical complexity of these

Particle-phase reactions

Organic compounds may also undergo chemical reactions in the condensed phase, affecting their chemical properties and volatility. As first suggested by Haagen-Smit (1952), and inferred by the ambient measurements of Ellis et al. (1984), these reactions may form products of low volatility. Particle-phase reactions, which include both heterogeneous and multiphase reactions (Ravishankara, 1997), are expected to be significant if they occur on timescales shorter than the lifetimes of tropospheric

Multigenerational chemistry

All organic compounds in the atmosphere, whether in the gas phase or the particle phase, are susceptible to oxidation. Thus, even after a particular oxidation reaction has gone to completion, and/or SOA is formed, the organic products will continue to evolve chemically. As discussed in the previous two sections, this chemical evolution can occur in the gas or particle phases, involving increases or decreases in organic volatility. As a result, over the course of their atmospheric lifetimes,

Conclusions

Since the volatility of an organic compound in the atmosphere can change by reactions in the gas phase (Section 2) and reactions in the particle phase (Section 3), over the course of several generations of oxidation (Section 4), the reaction scheme underlying SOA formation is probably something like that shown in Fig. 12. Note that this scheme shows only a single product from each reaction; in most cases there is likely to be substantially more branching, with each step forming a number of

Acknowledgments

The authors gratefully acknowledge the Department of Energy and the US Environmental Protection Agency for support; N.M. Donahue, N.L. Ng, D.R. Worsnop, and P.J. Ziemann for helpful discussions; and A.W.H. Chan for assistance in the preparation of several figures.

References (254)

  • J.J. Chen et al.

    Modeling secondary organic aerosol formation from oxidation of α-pinene, β-pinene, and d-limonene

    Atmospheric Environment

    (2005)
  • S.H. Chu

    PM2.5 episodes as observed in the speciation trends network

    Atmospheric Environment

    (2004)
  • M. Claeys et al.

    Formation of secondary organic aerosols from isoprene and its gas-phase oxidation products through reaction with hydrogen peroxide

    Atmospheric Environment

    (2004)
  • C.J. Colville et al.

    The roles of individual oxidants in secondary organic aerosol formation from Δ3-carene: 2. SOA formation and oxidant contribution

    Atmospheric Environment

    (2004)
  • K.K. Crahan et al.

    An exploration of aqueous oxalic acid production in the coastal marine atmosphere

    Atmospheric Environment

    (2004)
  • N.M. Czoschke et al.

    Markers of heterogeneous reaction products in α-pinene ozone secondary organic aerosol

    Atmospheric Environment

    (2006)
  • N.M. Czoschke et al.

    Acidity effects on the formation of α-pinene ozone SOA in the presence of inorganic seed

    Atmospheric Environment

    (2006)
  • N.M. Czoschke et al.

    Effect of acidic seed on biogenic secondary organic aerosol growth

    Atmospheric Environment

    (2003)
  • S. Decesari et al.

    Water soluble organic compounds formed by oxidation of soot

    Atmospheric Environment

    (2002)
  • E.O. Edney et al.

    Polar organic oxygenates in PM2.5 at a southeastern site in the United States

    Atmospheric Environment

    (2003)
  • E.O. Edney et al.

    Formation of 2-methyl tetrols and 2-methylglyceric acid in secondary organic aerosol from laboratory irradiated isoprene/NOx/SO2/air mixtures and their detection in ambient PM2.5 samples collected in the eastern United States

    Atmospheric Environment

    (2005)
  • T.L. Eliason et al.

    Oxidation of organic films relevant to atmospheric aerosols

    Atmospheric Environment

    (2004)
  • E.C. Ellis et al.

    Thermal characterization of organic aerosols

    Science of the Total Environment

    (1984)
  • H.J.L. Forstner et al.

    Molecular speciation of secondary organic aerosol from photooxidation of the higher alkenes: 1-octene and 1-decene

    Atmospheric Environment

    (1997)
  • R.M. Garland et al.

    Acid-catalyzed reactions of hexanal on sulfuric acid particles: identification of reaction products

    Atmospheric Environment

    (2006)
  • D. Grosjean

    Atmospheric reactions of ortho cresol: gas phase and aerosol products

    Atmospheric Environment

    (1984)
  • M. Hanke et al.

    Atmospheric peroxy radicals: ROXMAS, a new mass-spectrometric methodology for speciated measurements of HO2 and ΣRO2 and first results

    International Journal of Mass Spectrometry

    (2002)
  • M.R. Alfarra et al.

    A mass spectrometric study of secondary organic aerosols formed from the photooxidation of anthropogenic and biogenic precursors in a reaction chamber

    Atmospheric Chemistry and Physics

    (2006)
  • K.E. Altieri et al.

    Evidence for oligomer formation in clouds: reactions of isoprene oxidation products

    Environmental Science and Technology

    (2006)
  • J. Arey et al.

    Alkyl nitrate, hydroxyalkyl nitrate, and hydroxycarbonyl formation from the NOx–air photooxidations of C5–C8 n-alkanes

    Journal of Physical Chemistry A

    (2001)
  • R. Atkinson et al.

    Atmospheric degradation of volatile organic compounds

    Chemical Reviews

    (2003)
  • R. Bahreini et al.

    Measurements of secondary organic aerosol from oxidation of cycloalkenes, terpenes, and m-xylene using an aerodyne aerosol mass spectrometer

    Environmental Science and Technology

    (2005)
  • U. Baltensperger et al.

    Secondary organic aerosols from anthropogenic and biogenic precursors

    Faraday Discussions

    (2005)
  • I. Barnes et al.

    Kinetics and products of the reactions of NO3 with monoalkenes, dialkenes, and monoterpenes

    Journal of Physical Chemistry

    (1990)
  • G. Bench et al.

    Relative contributions of fossil and contemporary carbon sources to PM 2.5 aerosols at nine Interagency Monitoring for Protection of Visual Environments (IMPROVE) network sites

    Journal of Geophysical Research—Atmospheres

    (2007)
  • J.D. Blando et al.

    Secondary formation and the Smoky Mountain organic aerosol: an examination of aerosol polarity and functional group composition during SEAVS

    Environmental Science and Technology

    (1998)
  • B. Bonn et al.

    New particle formation during α- and β-pinene oxidation by O3, OH and NO3, and the influence of water vapour: particle size distribution studies

    Atmospheric Chemistry and Physics

    (2002)
  • B. Bonn et al.

    Influence of water vapor on the process of new particle formation during monoterpene ozonolysis

    Journal of Physical Chemistry A

    (2002)
  • B. Bonn et al.

    High contribution of biogenic hydroperoxides to secondary organic aerosol formation

    Geophysical Research Letters

    (2004)
  • C.A. Brock et al.

    Particle growth in urban and industrial plumes in Texas

    Journal of Geophysical Research—Atmospheres

    (2003)
  • N.I. Butkovskaya et al.

    Mechanism of the OH-initiated oxidation of hydroxyacetone over the temperature range 236–298 K

    Journal of Physical Chemistry A

    (2006)
  • X.Y. Cai et al.

    Secondary aerosol formation from the oxidation of biogenic hydrocarbons by chlorine atoms

    Journal of Geophysical Research—Atmospheres

    (2006)
  • M. Camredon et al.

    The SOA/VOC/NOx system: an explicit model of secondary organic aerosol formation

    Atmospheric Chemistry and Physics

    (2007)
  • A.G. Carlton et al.

    Link between isoprene and secondary organic aerosol (SOA): pyruvic acid oxidation yields low volatility organic acids in clouds

    Geophysical Research Letters

    (2006)
  • A.G. Carlton et al.

    Atmospheric oxalic acid and SOA production from glyoxal: results of aqueous photooxidation experiments

    Atmospheric Environment

    (2007)
  • A.W.H. Chan et al.

    Kinetic modeling of secondary organic aerosol formation: effects of particle- and gas-phase reactions of semivolatile products

    Atmospheric Chemistry and Physics

    (2007)
  • S.H. Chung et al.

    Global distribution and climate forcing of carbonaceous aerosols

    Journal of Geophysical Research—Atmospheres

    (2002)
  • B. Chuong et al.

    Cycloalkene ozonolysis: collisionally mediated mechanistic branching

    Journal of the American Chemical Society

    (2004)
  • J.A. de Gouw et al.

    Reactive uptake of ozone by liquid organic compounds

    Geophysical Research Letters

    (1998)
  • J.A. de Gouw et al.

    Budget of organic carbon in a polluted atmosphere: results from the New England Air Quality Study in 2002

    Journal of Geophysical Research—Atmospheres

    (2005)
  • Cited by (0)

    View full text