资源环境科技发展态势分析平台

Optical properties of flame-generated black carbon (BC) containing soot particles were quantified at multiple wavelengths for particles produced using two different flames: a methane diffusion flame and an ethylene premixed flame. Measurements were made for (i) nascent soot particles, (ii) thermally denuded nascent particles, and (iii) particles that were coated and then thermally denuded, leading to the collapse of the initially lacy, fractal-like morphology. The measured mass absorption coefficients (MACs) depended on soot maturity and generation but were similar between flames for similar conditions. For mature soot, here corresponding to particles with volume-equivalent diameters >similar to 160 nm, the MAC and absorption Angstrom exponent (AAE) values were independent of particle collapse while the single-scatter albedo increased. The MAC values for these larger particles were also size-independent. The mean MAC value at 532 nm for larger particles was 9.1 +/- 1.1m(2) g(-1), about 17% higher than that recommended by Bond and Bergstrom (2006), and the AAE was close to unity. Effective, theory-specific complex refractive index (RI) values are derived from the observations with two widely used methods: Lorenz-Mie theory and the Rayleigh-Debye-Gans (RDG) approximation. Mie theory systematically underpredicts the observed absorption cross sections at all wavelengths for larger particles (with x > 0.9) independent of the complex RI used, while RDG provides good agreement. (The dimensionless size parameter x = pi d(p)/lambda, where d(p) is particle diameter and lambda is wavelength.) Importantly, this implies that the use of Mie theory within air quality and climate models, as is common, likely leads to underpredictions in the absorption by BC, with the extent of underprediction depending on the assumed BC size distribution and complex RI used. We suggest that it is more appropriate to assume a constant, size-independent (but wavelength-specific) MAC to represent absorption by uncoated BC particles within models.