Aerosol number density distributions

Table VIII. Physical Processes Affecting the Evolution of Aerosol Number Density Distributions...

Fig. 7-3. Average volume size distributions for continental aerosols. [Adapted from Whitby and Sverdrup (1980).] The measurement data were smoothed and idealized by fitting to them additive log-normal distributions. (1) Background aerosol, very clean (2) normal background aerosol (3) background aerosol disturbed by an urban plume (these data from measurements at Goldstone, California). (4) Average urban aerosol (from data taken at Minneapolis, Minnesota, Denver, Colorado, and various locations in California). The dashed curve gives the volume distribution resulting from the number density distribution for the rural continental aerosol shown in Fig. 7-1. The integrated volumina, given by the area underneath each curve, are shown in the insert.

Fig. 1. Number density distribution of maritime and continental aerosol in the surface boundary layer. (Straight Line power law distribution with exponent s = 3.4, see text.)...

Particle Measurements. A variety of instruments is available for measuring the number density and size distribution of particles sampled from airborne platforms. This discussion is restricted to instruments that measure particles smaller than 50 xm (cloud droplets and aerosol particles) because these particles are of most interest to atmospheric chemists. 

The additional channels in the lower portion of the size spectrum overlap the sizes measured by the ASASP and the PCASP probes and have allowed intercomparisons between aerosol size distributions measured by these two probes (90). Dye et al. have evaluated the capabilities of this probe for measuring the number density and size distribution of atmospheric aerosols (98). 

Accomplishment of the complex observational experiment LACE-98 made it possible to obtain extensive information about atmospheric aerosol (aircraft measurements of the size distribution and number density of fine aerosols, coefficients of aerosol absorption, backscattering and depolarization, chemical composition of aerosol, as well as surface observations of the spectral optical thickness of the atmosphere, coefficients of extinction and backscattering). Fiebig et al. (2002) compared the observational data on optical parameters obtained from the results of numerical modeling for total H2S04 aerosol near the tropopause as well as for the ammonium sulfate/soot mixture in the remainder of the air column (Osborne et al., 2004). 

Barr et al. (2003) performed an analysis of the impact of phytogenic aerosol (PhA) which is defined as forming mainly due to monoterpene oxidation (primarily, a- and /3-pinenes), on the radiative regime of the ABL over the forest in the eastern part of Canada. In the forest ecosystem the level of emissions to the atmosphere of biogenic hydrocarbons is moderate, with the concentration of a- and /3-pinenes constituting about 1.6 ppb. NMHC oxidation resulted in the formation of PhA at a number density of particles of about 5 108 cm 3. For a given concentration and size distribution of aerosol, its impact on the short-wave radiation transfer in the ABL was assessed. 

A basic understanding of the nebulizer function and the types of nebulizers is necessary to successfully interface CE to the ICP-MS. Nebulization, as previously described, is the process to form an aerosol, i.e., to suspend a liquid sample into a gas in the form of a cloud of droplets. The quality of any nebulizer is based on many different parameters including mean droplet diameter, droplet size distribution, span of droplet size distribution, droplet number density, and droplet mean velocity. There are numerous nebulizers commercially available for the use with ICP-MS systems, and their detailed description can be found elsewhere.Pneumatic designs, both concentric and cross flow, are the most popular for CE interfaces with the occasional use of the ultrasonic nebulizer (USN). Figure 2 shows some typical nebulizers. The pneumatic nebulizer is either a concentric design (Fig. 2A), where both the gas stream and the liquid flow in... 

A size distribution may refer to number density, volume, mass, or any other property of the aerosol that varies with particle size. The left-hand side of Fig. 7-1 shows somewhat idealized size spectra for the number densities associated with the marine and the rural continental aerosols. The distribution of sea-salt particles that contribute to the marine aerosol is added for comparison. Particle radii and number densities range over many orders of magnitude. In view of the need for a logarithmic representation of the data, it has been found convenient to use the decadic logarithm of the radius as a variable and to define the distribution function by /(log r) = dN/d( log r). A conversion of /(log r) to the more conventional form /(r), if needed, can be done by means of the rule d(log r) = dr/ r In 10) with In 10 = 2.302, which yields... 

Fig. 7-2. Model size distributions of the marine background aerosol (a) particle number density, (b) surface area, (c) volume. The contribution of sea salt to the volume distribution is indicated by the shaded area, and arrows indicate the appropriate scale. By integration one obtains a total number density N =290 particles/cm3, a total surface area A = 5.8 x 10 7 cm2/cm3, and a total volume V= 1.1 x 10 " cm3/cm3. For an average density of 103 kg/m3, the mass concentration is 11 pig/m3 (5 pig/m3 of sea salt). The dashed curve gives the distribution of the surface area that is effective in collisions with gas molecules. For larger particles the collision rate is lowered by the rate of diffusion.

Fig. 7-8. Influence on cloud nuclei formation of the mass fraction e (water-soluble material/particle dry mass). Left Critical supersaturation of aerosol particles as a function of particle dry radius. Right Cloud nuclei spectra calculated for e = 0.1 and 1 on the basis of two size distributions each for continental and maritime aerosols (solid and dashed curves, respectively). [Adapted from Junge and McLaren (1971).] The curves for the maritime cloud nuclei spectra are displaced downward from the original data to normalize the total number density to 300 cm-3 instead of 600 cm-3 used originally. The curves for e = 1 give qualitatively the cumulative aerosol size distributions starting from larger toward smaller particles (sk = 10 4 corresponds to r0 0.26 p.m, sk = 3 x 10 3 to rs 0.025 Atn). Similar results were subsequently obtained by Fitzgerald (1973, 1974). The hatched areas indicate the ranges of cloud nuclei concentrations observed in cloud diffusion chambers with material sampled mainly by aircraft [see the summary of data by Junge and McLaren (1971)] the bar represents the maximum number density of cloud nuclei observed by Twomey (1963) in Australia.

Figure 7-11 compares number density size distributions for soil and aerosol particles at several locations in North Africa to show the dominance of submicrometer particles in all populations studied. The maximum number density was found near 0.1 jtm radius in all samples, and particles as small as 0.02 (xm were detected. 

Fig. 1.2. Normalised frequency plots of number, surface, and volume (particle volume times particle density) distributions for the grand average 1969 Pasadena, California smog aerosol. Note the bimodal distribution of mass. Each weighting shows features of the distribution not shown by the other plots. From Whitby (1975, p. II-ll) in NRC (1979).

Figure 4 (A, B) Number-frequency distribution and (C) cumulative number distribution of an aerosol of unit-density spheres. Indicated are the count median diameter (CMD), the surface median diameter (SMD), and the mass median diameter (MMD) of the number-frequency distribution. The 16, 50, and 84% size cut of the cumulative number distribution are shown. For further explanation, see text.

In practice, when one measures the size distributions of aerosols using techniques discussed in Chapter 11, one normally measures one parameter, for example, number or mass, as a function of size. For example, impactor data usually give the mass of particles by size interval. From such data, one can obtain the geometric mass mean diameter (which applies only to the mass distribution), and crg, which, as discussed, is the same for all types of log-normal distributions for this one sample. Given the geometric mass mean diameter (/) ,) in this case and crg, an important question is whether the other types of mean diameters (i.e., number, surface, and volume) can be determined from these data or if separate experimental measurements are required. The answer is that these other types of mean diameters can indeed be calculated for smooth spheres whose density is independent of diameter. The conversions are carried out using equations developed for fine-particle technology in 1929 by Hatch and Choate. 

Important issue for urban dispersion modelling is the characteristics of the release, e.g., radiochemical composition, density for gases, size distribution for aerosols, etc. For radioactive aerosols the particle size distribution (e.g., number of modes, distribution type, average diameter and standard deviation for each mode, density, and nuclides) varies significantly for different release types and from one nuclide to another. The particle size spectrum could be very broad, e.g. 0.001-200 fim. 

---