[38] Hathaway, Aerosol Age, 20, Sept. 1973. [39] Stigliani, D. J. et al., J. Opt. Soc. America, 60 (8), 1059, 1970. [40] Herman, B. M. et al., J. Atmos. Sci., 28, 763, 1971. [41] Pasqualucci, F., Proc. Ill., Oct. 1972. 15th Conf. Radar Meteorology, Champaign, NATIONAL BUREAU OF STANDARDS SPECIAL PUBLICATION 412 LIGHT SCATTERING BY SINGLE AEROSOL PARTICLES Milton Kerker Clarkson College of Technology ABSTRACT Although angular light scattering is a sensitive measure of particle size, sensitivity is rapidly lost when a mixture of particulates covers a range of particle sizes. However, instrumentation for studying scattering by single particles can overcome this limitation. This will be illustrated with reference to scattering by single fibers and spheres. Key words: aerosol fibers; light scattering; particle size measurements; particulates; refractive index. I shall touch upon three aspects of the art of particle size analysis by light scattering. These are (i) difficulties associated with the inversion of light scattered by dispersions, (ii) our experience with scattering by single particles, and (iii) some recent calculations of the response curves curves for commercial particle counters. Background material may be found in reference 1. I. THE PROBLEM OF INVERSION Light scattering by particles which are comparable in size to the magnitude of the wavelength of light is strongly dependent upon particle size, and it is for this reason that light scattering can provide a sensitive technique for particle size analysis. However, for a dispersion, the light signals are averaged over the particle size distribution. This has the effect of washing out the information and makes inversion of the light scattering-data to obtain the particle size distribution increasingly difficult as the distribution broadens. This effect is illustrated in figure 1, which is based upon calculations for refractive index m = 1.43. The ordinate p 1.43. The ordinate p represents the ratio of the intensity of the parallel polarized component of the 2.6 Electrical Mobility Methods These methods determine the charge-weight relationship of the droplets. Mobility analyzers precipitate particles in an electric field of varying voltage, and measure the current due to the charges on the particle surfaces. The various instruments are discussed in references 4 and 7. Commercial instruments include those produced by ThermoSystems [7]. 2.7 Optical Methods Optical methods that have been used to measure droplet size have included: When a light beam passes through a monodispersed aerosol which is viewed at an angle to the incident light against a dark background, the presence of particles is observed by the occurance of Tyndall spectra due to the scattered light. The Owl is an instrument that uses this principle to measure size [30]. With this device, the spectral colors are observed using a rotating telescope. With white light, size can deduced from the number of red bands that appear in the spectrum. be Light scattering from single particles has been very widely used, and over the last 20 years, many variations variations in particle counting instruments have been developed. These are discussed in detail in references 4 and 5. Very recently, Gravatt [31] at NBS developed a light scattering counter for the EPA in which a simultaneous measurement of measurement of the light intensity at two small angles was performed. The ratio of these intensities was then related to size. Use of the intensity ratio method at two angles is also made with multi-scattering particle systems. With the the Bryce-Phoenix light scattering photometer, this technique, termed the polarization ratio method, is applied. Other photometer systems using multiple particle scattering have been built and tested by Ranade [32] and Cooke and Kerker [33]. These systems were designed to measure the kinetic behavior of aerosol particles in flowing gas streams. The above light scattering devices use conventional optics, but other systems have been developed that measure the scattering from particles in a laser beam. Schleusener [34] developed a gas laser for sizing particles; Harris and Morse [35] developed a laser system for measuring fog droplet sizes. A recent device in this class was developed by Gucker et al. [36] and is the subject of the paper by Parmenter in this symposium. Knollenberg also later discusses laser scattering devices developed at Particle Measuring Systems Inc. in Boulder, Colorado. Laser holographic methods have also found application in sizing droplet particles. Horn et al. [37] described a spray droplet analyzer, and offers a laser holographic service through Laser Holography Inc. This afternoon, Hotham describes the use of this system for sizing sprays from deodorant, hairspray, and pesticide spray cans. Optronics International offers a similar service, as do Laser Photographic Laboratories in Chicago. Hathaway [38] recently published data taken by this latter system on deodorant spray cans, and showed the effect of evaporation on the size distribution at distances of 1", 6", and 12" from the nozzle. The use of laser holography for sizing water drops was explored by Stigliani [39], who developed a forward scatter system at the Illinois Water Survey. For meteorological studies, laser-radar methods have found application. Herman et al. [40] have reported on a system for measuring atmospheric aerosols, and Pasqualucci [41] has described the use of Doppler radar systems for rainfall measurement. Laser heterodyne spectroscopy, which measures a distribution of Doppler shifts, has been used by Gollub et al. [42] to measure the size of falling water droplets in a diffusion cloud chamber. The principle of this system and some size data are reported by Chabay in this symposium. The Doppler shift principle is also described today by Yanta, who has used it to measure size distributions of particles accelerated through a nozzle. In this work, the particle velocities were determined using the scattered light spectrum from the particles passing through an interference pattern. 3. COMPARISON OF RESULTS FROM DIFFERENT DEVICES To give some estimate of the problems associated with droplet sizing, a comparative study on stearic acid aerosols was performed at IIT Research Institute. Some of the results are given in figure 4. For this comparison, the methods using Tyndall spectra, single particle scattering, electrical mobility, particle collection, and image reconstruction were employed. The data shows that a mean size of 0.9 μm was recorded by the Owl from Tyndall spectral measurements. This agreed fairly well with both the value of 0.84 μm obtained by the collection, microscope measurement and particle reconstruction method, and the value of 0.8 μm recorded by the mobility analyzer. The optical particle counters showed greater variability, which was mainly attributed to a 7 Figure 4. Comparative data taken on optical particle coincidence problem. With 30,000 particles/cubic feet, the Bausch and Lomb, Climet and Royco counters gave mean sizes ranging from 1.22-1.4 μm. In contrast, the IITRI counter, operating with a 10:1 dilution system, gave a value of 1.08 μm. In comparing these results, some correlation with the response time of the sizing methods should be reported. For example, the Owl was situated on the exit of the aerosol generator, and measured the aerosol at time = 0. In contrast, the optical particle counters, the Whitby Aerosol Analyzer, and the electrostatic precipitator were connected to an 800 cu ft aerosol chamber, and measured the aerosol several minutes after generation. Allowing 5 minutes for homogenization of the aerosol after the completion of aerosol generation, all the instruments began their sampling period, at a time, t, = 5 min. The optical particle counters then recorded the size distribution during a 1 minute sampling period, the Whitby Aerosol Analyzer recorded the size distribution during a 4 minute sampling period, and the electrostatic precipitator collected aerosol over a 20 minute sampling period. It is important to note that the collection and the mobility analyzer results at t = 25 and t 9 minutes, respectively, both indicated lower mean values than the Owl at time = 0. As the standard deviation of the distribution by microscope was less than 1.2, the aerosol can be termed monodispersed, and the Owl data should be reliable. This suggests that the analyses at time = 9 min and time = 25 min shows indication of slight size contraction of the aerosol. Though not considered an evaporative aerosol, stearic acid will undergo slight |