Fractal Morphology Analysis of Combustion-Generated Aggregates Using Angular Light Scattering and Electron Microscope Images

Abstract

Experimental studies of the fractal morphology of flame-generated aggregates are described here, considering not only the fractal dimension, Df, but also the fractal prefactor (lacunarity), kg, both of which are shown to be needed to fully characterize aggregates. Measurements were made using angular light scattering (ALS) and thermophoretic sampling followed by transmission electron microscopy (TEM) for soot aerosols found in laminar and turbulent flame environments. Df and the prefactor kg were simultaneously inferred from ALS measurements using the optical properties of aggregates composed of small primary particles. TEM-based inferences of these fractal properties involve analysis of aggregateprojected images from which the actual morphologies are obtained by correlating the radius of gyration to the outer radius (half of the maximum length) of an aggregate. Both of our procedures for determining the detailed morphology of aggregates yield Df = 1.7 ± 0.15 and kg = 2.4 ± 0.4 for carbonaceous soot, in good agreement with earlier TEM measurements involving multiple angle images. Furthermore, we show that these values also seem to be valid for other materials such as alumina aggregates, suggesting that not only the fractal dimension but also the fractal prefactor are universal properties of aggregates found in combustion environments due to similar mechanisms of aggregation. This universality for Df and kg is observed at various positions in four different flame types with nine various gaseous and liquid fuels for aggregates with mean primary particle radii and number of primary particles in aggregates in the range 12-26 nm and 2-104, respectively.

The Effect of Chemical Composition on the Fractal-Like Structure of Combustion-Generated Inorganic Aerosols

“…In contrast, fractal exponents of approximately 2.5 are expected when growth is dominated by the diffusion of primary particles (monomers) to an existing aggregate (Schleicher et al 1995). The observations reported herein are also in the range of fractal exponents reported for soot aggregates (1.6-1.9) obtained in both laminar and turbulent diffusion ames summarized by Koylu et al (1995). The experimental results generated herein therefore suggest that the morphology of ame generated aggregates is largely controlled by cluster-cluster aggregation of smaller aggregates formed in the early stage of the ame and is invariant with changes in chemical composition and mass density of particulate in the ame for the range of conditions considered herein.…”

“…If a system were to encounter an increasing temperature gradient, or a prolonged period at elevated temperature, both internal rearrangement and modi cation of the fractal exponent would be expected. This has been demonstrated by Schmidt-Ott (1988) for silver aerosols and has been more recently demonstrated for alumina in a laboratory counter ow diffusion ame (Xing et al 1997;Koylu et al 1995).…”

“…Particle fractal dimension D f and fractal prefactor k g were calculated using the methods speci ed by Megaridis and Dobbins (1990) and Koylu et al (1995) for two-dimensional images, valid for samples with a fractal exponent D f of 2 or below. This rst required determination of the number of primary particles per aggregate, N, from…”

Structural Properties of Diesel Exhaust Particles Measured by Transmission Electron Microscopy (TEM): Relationships to Particle Mass and Mobility

“…By means of numerical simulations, they found that ka and a fall in the following ranges: a = 1.08-1.19 and ka = 1.00-1.81. Koylu et al (1995b) suggested a = 1.09 and ka = 1.15 by their extensive analysis of particle morphology. Oh and Sorensen ( 1997) took the overlapping of primary particles into account and found that a = 1.19 and ka = 1.81 when the overlap parameter, which is the ratio of the primary particle diameter to the distance between two touching particles, is 2.…”

“…They showed that in their simulated agglomerates, accounting for overlap leads to more primary particles per agglomerate, and therefore to agglomerates that are more dense and that have a higher fractal dimension and prefactor (D 1 = 2.01, kg = 3.0) compared to values when overlapping is ignored (D 1 = 1.80, kg = 1.30). In the present study, the values of a and ka suggested by Koylu et al (1995b; a = 1.09, ka = 1.15) and Oh and Sorensen (1997; a = 1.19, ka = 1.81) are employed when using Equation (3) to determine the number of primary particles in diesel agglomerates.…”