Analysis and Modeling of Atmospheric Particles to Assess Effects on Climate

Summary

Atmospheric aerosol particles often have compositions and shapes that cannot be described as a single material phase within a simplified geometric form such as a sphere. We analyze individual atmospheric particles and create 3-dimensional spatial models to determine how particles with complex compositions and shapes interact with solar energy and, therefore, affect climate.

Description

The upper figure shows 3-D spatial models of various urban dust particles with multiple material phases in each particle. Darker-hued phases are more light-absorbing, lighter-hued phases are more light-scattering. Red material is strongly light-absorbing soot. The lower figure shows how the optical extinction efficiencies of various geometric-shaped particles compare with an actual Asian dust particle. In this case, the tetrahedron is closest to the actual particle.

Spectral data from satellite and ground-based remote sensing instruments are used to monitor the effects of atmospheric aerosols on climate. Determinations of how aerosols interact with sunlight or infrared radiation from Earth’s surface rely upon an aerosol model that approximates the sizes, shapes, and compositions of the particles. Particle composition and shape are typically modeled as a single material within a simplified geometric form such as a sphere or spheroid. In this research, we use scanning electron microscopy and 3-dimensional spatial modeling of individual particles to show how the optical behavior of actual particles in the atmosphere compares with aerosol models that utilize homogeneous geometric shapes for particles.

Important technologies used in this research are focused ion-beam (FIB) tomography for generating particle 3-D spatial models and the discrete dipole approximation method [B.T. Draine and P.J. Flatau, 1994, J. Opt. Soc. Am., 11, 1491-1499] for calculating particle optical properties. FIB tomography is performed with a scanning electron microscope equipped with an ion-beam column. The ion beam slices through the particle incrementally as the electron beam images each slice. Element maps of the particle may be acquired with energy-dispersive x-ray spectroscopy. The images and maps are used to create the 3-D spatial model, from which the discrete dipole approximation software [Draine and Flatau, User Guide for the Discrete Dipole Approximation Code DDSCAT 7.3, 2013] is used to calculate optical extinction, single scattering albedo (SSA, ratio of scattering to extinction), asymmetry parameter, and the phase function.

Aerosol remote sensing techniques for monitoring climate typically separate particles into two size classes: <1 μm and >1 μm. Examples of particles <1 μm are tailpipe soot and secondary organic aerosol formed by oxidation of gas-phase organic carbon within the atmosphere. Examples of particles >1 μm are mineral dust from arid regions and urban road dust.

We investigate dust particles selected from samples collected at urban sites (e.g., Los Angeles and Seattle) and at the Mauna Loa Observatory (MLO) in Hawaii. Particles from the MLO samples are identified as Asian dust. For the urban and Asian dust samples, we compare the optical properties for the actual particles with the particles as simple geometric shapes, which include spheres, ellipsoids, cubes, and tetrahedra. The geometric particles are volume- and compositionally-equivalent to the original particles.

This research is intended to help atmospheric scientists who utilize aerosol models to derive aerosol optical properties from remote-sensing spectra. We show how remote-sensing aerosol models may be improved by incorporating additional geometric particle-shape information that accounts for composition and morphological features of different types of atmospheric dust. Improved particle composition and shape information may then be used to improve atmosphere-ocean circulation models for creating accurate climate change scenarios.

Related Publications

J. M. Conny and D.L. Ortiz-Montalvo, “Effect of Heterogeneity and Shape on Optical Properties of Urban Dust Based on 3-Dimensional Modeling of Individual Particles,” J. Geophys. Res.- Atmospheres, in review.

J. M. Conny, S.M. Collins, A.A. Herzing, “Qualitative Multiplatform Microanalysis of Individual Heterogeneous Atmospheric Particles from High-Volume Air Samples,” Anal. Chem. 2014, 86, 9709-9716.

J. M. Conny, “Internal Composition of Atmospheric Dust Particles from Focused Ion-Beam Scanning Electron Microscopy,” Environ. Sci. Technol. 2013, 47, 8575-8581.

J. M. Conny and G.A. Norris, “Scanning Electron Microanalysis and Analytical Challenges of Mapping Elements in Urban Atmospheric Particles,” Environ. Sci. Technol. 2011, 45, 7380-7386.

Major Accomplishments

Models of dust particles as geometric shapes generally exhibit lower extinction efficiencies than the actual particles. This may be due to the exclusion of voids and surface features in geometric models. (See above figure.)

Particle shape is a more important factor for modeling extinction efficiency than composition heterogeneity. This was indicated from the comparison of geometric models that included heterogeneity with models that incorporated the actual particle shape but had a homogeneous composition.

For particles that had loosely-held phases and widely-varying refractive indexes, only geometric models that account of heterogeneity may suffice for determining SSA.

Tetrahedral geometric models generally perform better than spherical and cubic models for the extinction efficiency and SSA. (See above figure.)

For the phase function, ellipsoid geometric models were closest to that for the actual particles.

Initial work on soot-associated mineral particles from the Asian dust samples shows that the presence of 1.7 % soot by volume in the particle resulted in a 5.7% decrease in SSA.

Additional Technical Details

For the urban dust particles, a series of 3-D spatial models consisted of test models and a reference model. The reference model for each particle exhibited the size, shape, void structure, and heterogeneous composition of the actual particle. If the particle contained a light-absorbing carbon or iron phase based on microanalysis, different compositions for these phases, for example, hematite, goethite, or limonite for the iron phase, were included to assess the how optical properties would vary with different compositions. Thus, each particle had multiple composition treatments. Optical properties were then calculated for the different composition treatments.

For the test 3-D spatial models of each particle, two model types were generated:

1) Homogeneous spatial models included particle size, actual shape, and void spaces. However, for each particle the material phases were treated as a single material by combining the complex refractive indexes of the individual phases using the Maxwell-Garnett dielectric function for the effective medium approximation. Thus, for these test models the particles were treated as homogeneous particles.

2) Geometric models approximated the particle’s size, shape, and composition with spheres, cubes, and tetrahedra. To represent a particle’s composition heterogeneity, the particle was modeled as a collection of spheres with each sphere corresponding to a separate material phase or group of phases. Particles were also modeled as a single sphere, cube, and tetrahedron with a homogeneous composition using the Maxwell-Garnett effective medium approximation.

Aerosol samples collected at MLO during March and April 2011 consisted of a 72-hour integrated collection during daytime (6am to 6pm) and the same during nighttime (6pm to 6am). Analysis of particle populations by SEM was performed by Robert D. Willis at EPA. Classification of particles based on element composition by SEM-EDX resulted in 12 particle classes. Two classes were identified as Asian dust: 1) particles enriched in calcium and magnesium, identified as the mineral dolomite (CaMg(CO₃)₂), and 2) particles enriched in only calcium, identified as calcite (CaCO₃). The Asian dust impacted sampling at MLO during the first weeks of sampling. Air mass back trajectories and global maps of aerosol types from the (U.S.) Navy Aerosol Analysis and Prediction System confirmed that Asian dust likely impacted MLO on days when the number concentrations of dolomite and calcite particles were found to be highest. A third particle class, exhibiting enriched calcium and sulfur, identified as gypsum, was likely part of marine background aerosol.

Similar to the urban dust particles, 3-D spatial models were generated that consisted of a reference model, which depicted the actual shape and internal structure of the particles, and test models consisting of particles as homogeneous geometric shapes. The following geometric shapes were included for the Asian dust: 1) sphere, 2) 3-axis ellipsoid, 3) cube, 4) 2-axis rectangular prism, and 5) tetrahedron. Optical properties were then calculated for the reference and test models of each particle.