ARTICLE pubs.acs.org/ac
XPS Analysis of Combustion Aerosols for Chemical Composition, Surface Chemistry, and Carbon Chemical State Randy L. Vander Wal,*,† Vicky M. Bryg,‡ and Michael D. Hays§ †
The Department of Energy and Mineral Engineering and The EMS Energy Institute, Penn State University, University Park, Pennsylvania 16802, United States ‡ The Universities Space Research Association (USRA), The NASA-Glenn Research Center, Cleveland, Ohio 44135, United States § National Risk Management Research Laboratory, United States Environmental Protection Agency, Research Triangle Park, North Carolina 27711, United States
bS Supporting Information ABSTRACT: Carbonaceous aerosols can vary in elemental content, surface chemistry, and carbon nanostructure. Each of these properties is related to the details of soot formation. Fuel source, combustion process (affecting formation and growth conditions), and postcombustion exhaust where oxidation occurs all contribute to the physical structure and surface chemistry of soot. Traditionally such physical and chemical parameters have been measured separately by various techniques. Presented here is the unified measurement of these characteristics using X-ray photoelectron spectroscopy (XPS). In the present study, XPS is applied to combustion soot collected from a diesel engine (running biodiesel and pump-grade fuels); jet engine; and institutional, plant, and residential oil-fired boilers. Elemental composition is mapped by a survey scan over a broad energy range. Surface chemistry and carbon nanostructure are quantified by deconvolution of high-resolution scans over the C1s region. This combination of parameters forms a distinct matrix of identifiers for the soots from these sources.
A
wide range of natural and anthropogenic sources emit organic gases and submicrometer aerosol particles to the atmosphere. Primary source emissions play a role in regulating atmospheric aerosol nucleation and oxidation.1 They also affect particle growth by condensation and the photochemical production of ozone and secondary organic aerosols (SOA).2 Knowledge of how these processes evolve and the composition and mass of aerosol they yield is essential for predicting the direct and indirect effects of atmospheric aerosol particles on the Earth’s climate system and for assessing the human health effects of aerosols. The surface composition of aerosol particle emissions is known to modulate particle oxidation and SOA yield: surfaces of acidic aerosol emissions from fossil fuel combustion heterogeneously catalyze and increase production of SOA,2 and combustion particle oxidation rates in the high troposphere can be surface-limited.1a Moreover, aerosol surface composition influences particle size and hygroscopicity, in turn influencing cloud microphysics and formation of cloud nuclei and fog droplets. Anthropogenic and biomass burning sources account for greater than 60% of primary carbonaceous aerosol emissions—estimated at 67-127 Tg/yr globally3—and much of the organic matter in these aerosols is concentrated potentially at or near the particle surface.4 Yet, there is scant information about the atomic composition, carbon chemical state, and nature of chemical bonding at the surfaces of emissions aerosols from combustion. With so many implications, there is growing interest in developing analytical techniques that characterize combustion source and atmospheric particles. For example, Murphy5 reports r 2011 American Chemical Society
that greater than 20 mass spectrometer instruments have been developed for the purpose of real-time identification of chemical species in aerosols. These instruments have identified molecular fragments of organic aerosol ensembles6 and simultaneously measured particle size and the mixing states and elemental compositions of single particles.7 However, the fragmentation of organic compounds and particle ablation resulting from these techniques interferes with determining the particle surface chemistry and the identity of surface functional groups. Energy-dispersive X-ray analysis within an electron microscope is commonly utilized for elemental analysis.8 However, the spectra from incident electron beams of such high energy only provide information about elements with atomic number >12 and cannot elucidate surface functional groups. Electron energy loss spectroscopy (EELS) can differentiate σ and π carbon bonds, but two primary difficulties exist.9 First, the spectra are obtained using a very high energy electron beam (∼200 keV), blurring the relatively minor energy difference between the elemental response and a rising background. Response peaks are thus difficult to extract, integrate, and model. Second, the technique cannot differentiate between organic and elemental carbon of the graphene lamella within combustion generated soots, as even their bonding will contain both σ and π components if significant curvature is present in the nanostructure. Knauer et al.10 have Received: September 5, 2010 Accepted: January 19, 2011 Published: February 15, 2011 1924
dx.doi.org/10.1021/ac102365s | Anal. Chem. 2011, 83, 1924–1930
Analytical Chemistry shown how Raman microspectroscopy and transmission electron microscopy can be combined to interpret the complex molecular structure of diesel combustion soot and use this tool to estimate the surface functionality of the soot, albeit indirectly. X-ray imaging techniques can offer chemical information while transverse spatial resolution of a TEM for single particle analysis is provided. Braun11 used scanning transmission X-ray (STX) analysis to measure organic functional group abundance and morphology of atmospheric aerosols. Yet STX and NEXAFS are X-ray absorption techniques that form a cumulative spectrum in the C1s region when applied to carbon with energy resolution limited by a high-incident beam energy, much like EELS. Thus, spectral decomposition is required but hindered by the scanned incident beam energy. Russell et al.12 used soft X-ray microscopy for aerosol characterization, providing detailed composition maps showing organic coatings of carboxylic acids surrounding particles containing inorganic ions. As a transmission technique, it provides rich two-dimensional (2-D) maps of individual particle compositions, using hundreds of scans at different energy levels to spatially resolve elemental composition. Cutting across the stacks of energyfiltered scan images necessary to generate spatially resolved images creates spectra that identify spatially resolved functional groups. The strength of the technique is that it provides detailed information on particle microphysics. The deficit is that because the STEM spatially details only individual particles, average (bulk) aerosol composition must be built from individual single-particle statistics, a rather an impractical approach. In contrast, XPS uses a monochromatic X-ray beam and measures the kinetic energy of ejected electrons to distinguish different surface groups. Moreover, XPS does not require a synchrotron radiation source. XPS is complementary in that it provides a surface analysis (∼1 nm for carbon), providing global measures of elemental and functional group compositions within a (relatively large) sampled area as determined by a beam size of ∼100-300 μm diameter.13 Comprehensive elemental content is provided by “survey” scans over a broad energy region, while functional groups are identified by high-resolution scans over core-shell energy levels of specific elements. Common to both techniques is that the stacks of energy-filtered scan images or direct scans of photoelectron energy are taken over an energy region corresponding to an element core-shell energy level of interest. In summary, X-ray spectromicroscopy as a transmission technique provides spatially resolved 2-D maps of composition and functional groups for individual particles, volumetrically averaged, while XPS provides sample averaged, surface specific values of composition and functional groups. Determination of particle surface composition by XPS is likely to improve estimates of (i) the sulfur-catalyzed SOA production, owing to anthropogenic or pyrogenic particle emissions; (ii) the extent of any atmospheric oxidation likely to occur on combustion particle matter (PM); and (iii) the particle surface functional groups and elements associated with acute near-emissions exposures. This study addresses how fresh soot surfaces generated from anthropogenic combustion and biomass burning differ compositionally.
’ EXPERIMENTAL SECTION A total of seven aerosol particle samples were collected from anthropogenic combustion or biomass burning emissions sources and analyzed for surface chemistry using XPS. Unless noted otherwise, the aerosols were collected on prefired (550 °C,
ARTICLE
12 h), high-volume quartz filters using a modified version of the dilution sampling system (DSS) described by Hildemann et al.14 A dilution ratio near 1:45 was typical for the source tests conducted with the DSS. The emissions sources examined here include (i) utility (USOB)-, institutional (ISOB)-, and residential-scale (ROB) oil boilers; (ii) jet aircraft (JE); (iii) heavy-duty diesel (and biodiesel) engine exhaust; and (iv) wildfire (WF). Two diesel fuel types were evaluated, a locally available pump-grade diesel fuel (PD) with sulfur content of 401 ppm and a biodiesel blend [BD; 20% vegetable oil, 80% low S diesel (
