= iances that would be obtained for the distribution depicted in figure 5 and then have inverted these results on the assumption that the system was a logarithmic distribution. This gave a modal size a 4.75 instead of 4.90 and σ = 0.02. Thus, the smaller class of particles affects the estimate only slightly, and if there were not some a priori information concerning them, there would be no reason, on the basis of the light-scattering data, to suspect their presence. Accordingly, light scattering from the assembly of particles cannot be used to detect the presence of these smaller particles. We also determined the size distribution of this latex in the conventional manner using the liquid suspension. The results for the average diameter are compared in table 1 with determinations in other laboratories. Table 1. Comparison of average (or modal) diameters The differences are probably within the experimental errors of the techniques. The principal advantage of the single particle techniques (both light scattering and electron microscopy) is that no a priori assumptions are required regarding the form of the distribution function. Indeed, the negative skew of the main population group in figure 5 and the presence of the two classes of smaller particles could not predicted nor can their presence be confirmed using conventional lightscattering techniques. be The failure heretofore to observe the smaller particles calls for further investigation, using both light scattering and electron microscopy. Also, much work remains to be done to assure the lack of bias in the sampling procedure for the single particle light scattering; otherwise, the question of the skewness of the distribution cannot be resolved definitively. In any case, the prospect for applying light scattering to size distributions broader than any heretofore amenable to light-scattering analysis appears to be a reasonable prospect. III. RESPONSE CALCULATIONS FOR LIGHT-SCATTERING We have recently completed response calculations for five commercial light-scattering aerosol particle counters which take into account the emissive power of the light source, the spectral sensitivity of the phototube, and the specific geometrical factors for each instrument. Earlier calculations had been published for two of these instruments, but these earlier calculations had not considered each of these factors appropriately. The results will be published elsewhere [3]. Here we wish only to state that there is a strong dependence of the response upon both the real and the imaginary part of the refractive index and, for a given refractive index, a multivalued response in the submicrometer range for three of the five instruments. Certainly these results indicate that particle size determination with any of these instruments will be highly precarious whenever the refractive index of the particles is unknown. Furthermore, the hazards of venturing to extend the use of these instruments for particle size analysis of nonspherical particles are quite unpredictable. REFERENCES [1] Kerker, M., "The Scattering of Light and Other Electromagnetic Radiation," Academic Press, New York, N.Y., 1969. [2] Cooke, D. D. and Kerker, M., J. Colloid Interface Sci. 42, 150 (1973). [3] Cooke, D. D. and Kerker, M., Appl. Optics. Submitted for publication. NATIONAL BUREAU OF STANDARDS SPECIAL PUBLICATION 412 LIGHT SCATTERING METHODS FOR THE CHARACTERIZATION OF C. C. Gravatt National Bureau of Standards ABSTRACT This paper presents a brief overview of new light scattering methods for the rapid characterization of particulate matter in air. An instrument has been developed which determines the size distribution of particulate matter in air in essentially real time by a forward lobe light scattering method. The basic concept involves the simultaneous measurement of the intensity of light scattered by a single particle at two small scattering angles. The ratio of the two intensities is a direct measure of the size and is fairly independent of the index of refraction of the particle. Numerical solutions of the Mie equations for spheres have indicated that the sizing error by this method is greater than ±15% for the range of particle sizes from 0.2 to 4 um for essentially all possible indices of refraction. In addition, it appears feasible to extend the lower limit of size determination to 0.05 um by the measurement of a single forward lobe intensity. Also, technique has been developed which permits some degree of chemical characterization of particles, and which has been employed in a a smoke detector capable of distinguishing between fire-produced and nonfireproduced aerosols. not Key words: aerosol sizing; aerosol spectrometer; chemical characterization of particles; fire produced particles; laser light scattering by aerosols; particle size measurements; particulates refractive index; smoke detector. I. INTRODUCTION The National Bureau of Standards, as part of its Measures for Air Quality Program (MAQ), has been investigating rapid, light scattering methods for the characterization of particulate matter. These investigations have primarily dealt with particles in air with diameters in the respirable size range (0.05 μm to 5 μm) although the techniques described below can be modified to operate over other size ranges and in other media, such as water. A single particle light scattering instrument has been developed which is based on a forward-lobe-ratio technique and which covers the particle diameter range from 0.2 μm to 4 um. Theoretical calculations for spheres have shown that the particle size error, due to variations in index of refraction, is approximately +15% for essentially all possible materials. This technique has been incorporated into a field compatible instrument constructed for and partially supported by the EPA, which will be part of the Regional Air Pollution Study (RAPS) to get comparative data among various particle sizing instruments. In addition, a method has been developed which permits some degree of chemical characterization of particles, specifically, a determination of whether a particle is carbon like, metal like, or something else. This method has been incorporated into a smoke detector which can distinguish between fire-produced and nonfireproduced aerosols. Light scattering methods provide a useful way of sizing particulate matter in air in essentially real time and without the necessity of collection of the sample. In addition, the size range of most interest for lung interaction, 0.05 μm to 5. μm, is relatively easily detected by light scattering if moderate power lasers lasers are employed as sources. However, the scattered intensity is not only a function of the particle size but depends rather strongly on the particle index of refraction. As a result, in those instances where a wide and unknown range of indices of refraction may exist in the sample, the light scattering method may give erroneous size values. It is found that errors the order of factors of 2 or greater are not uncommon when sizing particles which have indices of refraction differing from that of the calibration species [1]. It was first pointed out by Hodkinson [2] that the forward lobe region of the particle scattering spectra is least affected by the index of refraction and provides a convenient method of sizing particles to a moderately high degree of accuracy. II. PARTICLE SIZING An instrument has has been been developed which limits the error due to unknown particle index of refraction to approximately ±15% for essentially all possible types of aerosols. The basic concept of this method involves the measurement of the intensity of light scattered by a single particle at two small scattering angles and the calculation of the ratio of these two intensities. The intensity ratio is directly related to the size of the particle and is essentially independent of the index of refraction of the particle. The intensity ratio method also attractive since it eliminates or minimizes complications resulting from a number of experimental phenomena, such as source intensity variations and nonuniform particle velocities. Since this instrument and its theoretical evaluation has been discussed in several publications only a brief summary is given below [1,3,4]. is Figure 1 provides an introduction to the concepts used for both particle sizing and chemical characterization. This figure shows the scattered intensity, in arbitrary units, as a function of scattering angle, 0° to 180°, for a particle diameter of 1.0 μm and an incident wavelength, λ, of 514.5 nm. The incident beam is assumed to be linearly polarized and the scattered intensity is measured in the plane of Figure 1. Scattered intensity as function of scattering polarization. The two particles are one with an index of refraction of n=1.55-0.01, representing a nonabsorbing material with an index similar to many earth dusts, and the second a carbon particle, with index of 1.96 - 0.66 i. It can be seen that the forward lobe for these two materials are very similar in shape even though the intensity differs by a factor of two or so. The similarity of shape of the forward lobe for widely varying indices of refraction is the phenomenon employed for particle sizing. Specifically, a ratio of two intensities in the forward lobe is measured. The angular ratio of 20°/10° is effective over the particle range 0.1 μm to 1.5 μm, the ratio 10°/5° is effective over the range 0.2 μm to 3.5 μm and the ratio 5°/2.5° is effective over the range 0.5 μm to 10 um if the incident wavelength is approximately 500 nm. Shorter and longer wavelengths decrease and increase these ranges respectively. Also, it can be seen that the region from 30 to 90° differs by about by about one order of magnitude for these two materials. This effect is found to be generally true of absorbing and nonabsorbing materials and permits the partial chemical characterization of particles. 23 |