Speciation of organic fraction does matter for source apportionment. Part 1: A one-year campaign in Grenoble (France)

Highlights

• Source apportionment using key primary and secondary organic molecular markers

• Uncommon resolved sources: plant debris, fungal spores, biogenic & anthropogenic SOA

• High contribution of anthropogenic SOA (PAH SOA) noticed in winter PM pollution event

• Investigation of HuLiS origins: both, primary and secondary contributions highlighted

Abstract

PM₁₀ source apportionment was performed by positive matrix factorization (PMF) using specific primary and secondary organic molecular markers on samples collected over a one year period (2013) at an urban station in Grenoble (France). The results provided a 9-factor optimum solution, including sources rarely apportioned in the literature, such as two types of primary biogenic organic aerosols (fungal spores and plant debris), as well as specific biogenic and anthropogenic secondary organic aerosols (SOA). These sources were identified thanks to the use of key organic markers, namely, polyols, odd number higher alkanes, and several SOA markers related to the oxidation of isoprene, α-pinene, toluene and polycyclic aromatic hydrocarbons (PAHs). Primary and secondary biogenic contributions together accounted for at least 68% of the total organic carbon (OC) in the summer, while anthropogenic primary and secondary sources represented at least 71% of OC during wintertime. A very significant contribution of anthropogenic SOA was estimated in the winter during an intense PM pollution event (PM₁₀ > 50 μg m^− 3 for several days; 18% of PM₁₀ and 42% of OC). Specific meteorological conditions with a stagnation of pollutants over 10 days and possibly Fenton-like chemistry and self-amplification cycle of SOA formation could explain such high anthropogenic SOA concentrations during this period. Finally, PMF outputs were also used to investigate the origins of humic-like substances (HuLiS), which represented 16% of OC on an annual average basis. The results indicated that HuLiS were mainly associated with biomass burning (22%), secondary inorganic (22%), mineral dust (15%) and biogenic SOA (14%) factors. This study is probably the first to state that HuLiS are significantly associated with mineral dust.

Introduction

Airborne particles (particulate matter, PM) are a major concern of current research in atmospheric science due to their impact on both climate (Boucher et al., 2013) and air quality (Heal et al., 2012). Elucidating their emission sources and transformation processes constitutes a crucial step for the elaboration of efficient and cost-effective abatement strategies.

Organic matter (OM) is a major PM component. Organic aerosols (OA) are categorized into either primary organic aerosol (POA), directly emitted from anthropogenic and natural sources, or secondary organic aerosol (SOA), formed in the atmosphere notably via gas-particle conversion processes such as nucleation, condensation and heterogeneous multiphase chemical reactions involving (semi-) volatile compounds (VOCs or SVOCs) (Carlton et al., 2009, Ziemann and Atkinson, 2012). Due to the multiplicity of sources and of transformation mechanisms, the apportionment of the relative contribution of each of the different primary and secondary OA fractions is still fairly uncertain.

Specific organic compounds can provide insight into OA sources (Schauer et al., 1996). They are commonly referred to as molecular markers (tracers), such as levoglucosan for biomass burning (Simoneit et al., 1999a) or α-methylglyceric acid for SOA from isoprene oxidation (Carlton et al., 2009). Source-receptor models, such as positive matrix factorization (PMF), have been widely implemented using traditional aerosol chemical speciation, such as elemental carbon (EC), organic carbon (OC), major ions, and metals. The inclusion of a comprehensive set of organic molecular markers potentially offers a closer link between factors and sources, but it has been rarely applied in PMF studies because it requires large datasets and intensive lab-work (Jaeckels et al., 2007, Laing et al., 2015, Schembari et al., 2014, Shrivastava et al., 2007, Srimuruganandam and Shiva Nagendra, 2012, Waked et al., 2014, Wang et al., 2012, Zhang et al., 2009).

Source apportionment studies based on the use of source-receptor models assume that organic molecular markers are chemically stable in the atmosphere (defined as tracer compounds) (Schauer et al., 1996). However, these compounds can react in the atmosphere by photochemical processes involving sunlight and atmospheric oxidants such as O₃, NO_x, radicals OH, NO_3… For instance, levoglucosan is usually assumed to be very stable (Simoneit et al., 1999b) but recent studied have shown its significant atmospheric chemical degradation (Hennigan et al., 2010, Kessler et al., 2010, Mochida et al., 2010, Zhao et al., 2014). For most of these compounds, experimental data about their stability or atmospheric lifetimes are very scarce or not available. They are usually based on empirical calculations like for SOA markers (Nozière et al., 2015). If some markers have tendency to undergo a rapid decay in the atmosphere, so short lifetime, their use may cause a bias in the source apportionment results.

The main objective of this work is to apportion specific primary and secondary OA fractions using various and distinctive molecular markers in a PMF model. The present paper is based on results obtained from a year-long campaign conducted in an Alpine city, while a following paper will be dedicated to the use of a similar approach in the Paris region during a 3-week intensive sampling campaign, with a higher time resolution for filter samplings (every 4 h) through an intense PM pollution event (Srivastava et al., in preparation). A focus has been put here on usually unresolved PM sources, such as primary biogenic sources and secondary sources such as biogenic SOA formed from pinene or isoprene oxidation, and anthropogenic SOA formed from the oxidation of polycyclic aromatic hydrocarbons (PAHs), toluene and phenol. In addition, this work provides insight into the sources of total OC and of humic-like substances (HuLiS), a significant fraction of OM which plays an important role in the atmosphere (Graber and Rudich, 2006) and has not been extensively explored.