Figure 2 is a block diagram of the optical and electrical sytem used in the original instrument. The EPA instrument differs slightly in ways which are given below. Block diagram of forward-lobe-ratio particle sizing Figure 2. The light from a CW laser, either argon ion or helium-neon, is focused to a small point by a lens, and the sample blown through this point. A flow rate is chosen such that a particle takes approximately 10 us to pass through the beam. When a 2 mW He-Ne laser is used as the source, it is found that the smallest particle which can be sized effectively is approximately 0.4 um in diameter due to insufficient scattered intensity for smaller particles. However, when the source is an argon laser of 10 mW or higher power, particles can be sized to the limit of resolution of the ratio method, 0.2 μm for the angles of 10° and 5°. Smaller particles than this are "seen" and, therefore, counted by the system, but the ratio is relatively insensitive to particle sizes below 0.2 μm. The EPA instrument employs an air cooled 10 mW argon ion laser as source. The light scattered by a particle is measured by the light collector consisting of three parts and operating as a fixed, two angle, annular iris detector. The innermost section has a hole in its center to allow the main beam to pass through and an outer conical shape at an angle of 5° relative to the optical center. The second section has an inner conical shape of 5° and an outer conical shape of 10°. The outer most section has an inner conical shape of 10°. Optical fibers are positioned in the rear of the collector system and these fibers are grouped in two bundles, one each for the scattering angles of 50 and The output ends of the fiber optic bundles are placed close to the photocathodes of two photomultiplier tubes. When a particle passes 10°. through the light beam a pulse will be produced in the output of each photomultiplier. The height of these pulses is proportional to the scattered light intensity and the width is the time of flight of the particle through the beam. The optical system of the EPA instrument is similar except the fiber optics have been replaced by a replaced by a series of lenses, mirrors, and baffles. The active optical region is defined by the diameter of the laser beam, 0.14 mm, and the diameter of the sample stream 0.1 mm, and it has a volume of 1.5 x 10-5 cm3. Therefore, concentrations as high as 2 x 104 particles/cm3 can be measured without coincidence effects being larger than 2% of the total counts. At higher particle densities the sample stream can be diluted to reduce the concentration to an acceptable level and thus prevent coincidence losses caused by two or more particles being in the beam at the same time. The scattered light is detected by the two photomultipliers, and level detectors in a coincidence circuit are employed in each output channel to define that period of time during which a particle is "seen" by both detectors. During this time the scattered intensity at each angle is integrated in analog circuits (in the EPA instrument the peak intensity is detected). At the completion of the passage of the particle through the beam an output pulse is generated whose height is proportional to the ratio of the two integrated intensities. These pulses are counted either in a multichannel analyzer or a minicomputer to determine the particle size distribution. It is important to note that by taking a ratio of the two scattered intensities, most experimental complications are eliminated. There is no uncertainty caused by particles taking different times to pass through the light beam, or by passing through different parts of the beam, or by variation in intensity of the laser. in intensity of the laser. In addition, it is possible to design the logic and timing system so that in the event that two particles are in the beam at the same time, no data are taken. The performance of this instrument has been tested on a number of particulate samples. The primary calibration has been performed using polystyrene latex sphere samples of accurately known diameter. In addition to the latex sphere samples, this system has been tested with a variety of materials having wide ranges of indices of refraction and nonspherical shapes. These samples include zinc cadmium sulfate (fluorescent), tungsten, iron oxide, carbon black, titanium dioxide, puff ball spores, kaolin, and silica. Since these materials have not been accurately characterized as to their size distribution by an independent method, it is impossible to reach quantitative conclusions, but from the reported data on mean size or size range it has been found that this ratio instrument is not making large errors due to index of refraction or shape. It is estimated that the total size error, assuming a calibration performed with polystyrene latex spheres, is greater than 20%. Numerical solutions of the Mie equations for spheres have been performed to evaluate theoretically the optimal parameters and the inherent error in the forward lobe ratio method [4,5]. These calculations were performed for essentially all possible indices of refraction that might be expected in ambient air and other special types of particulate samples such as fire-produced aerosols. Specifically, the real component of the index of refraction was investigated over the range 1.33 to 2.0 and the imaginary component investigated over the range 0.0 to 4.0. This range of indices will cover most inorganic and organic materials, metal ores, earth dusts, carbons, and the metals. The results of these calculations have shown that the inherent error in the particle size, due to index of refraction variation in the sample, is greater than 15% for the following combination of particle size ranges and angular ratios, assuming the particles are spherical. The angular ratio of 20°/10° is effective over the particle range 0.1 um to 1.5 μm, the ratio of 10°/5° over the range 0.2 μm to 3 um and the ratio 5°/2.5° over the range 0.5 μm to 10 μm if the incident wavelength is approximately 500 nm. Shorter and longer wavelengths decrease and increase these ranges respectively. The effect of angular aperture was found to be small, and moderate apertures of approximately 0.5° are advantageous in that they increase the scattered intensity and do not degrade the size sensitivity. Various ratios were investigated, and it was found that those with a value of 2 are as effective as any other ratio and lead to the most symmetrical calibration curve. Figure 3 shows the theoretical error envelope for the intensity ratio of 110/15 for unpolarized incident radiation. Also shown is the response curve for particles with index of refraction of 1.50 -0.01. The error envelope incloses all response curves and can be used to determine the theoretical error of forward lobe ratio instruments. In general, the individual response curves are not monotonic functions, and therefore, in those situations in which the particle index of refraction is known, the forward lobe ratio technique may not be the most accurate method to employ. Also note that beyond a diameter of 3.0 μm (α = ПD/ = 19) the ratio envelope increases. To prevent substantially higher errors due to this effect, an intensity cutoff is employed to prevent the recording of a ratio for particles larger than this cutoff. We have investigated theoretically several ways of decreasing the size error. The most promising of these is to use a multiwavelength laser as a source, especially one that has two widely-spaced lines. For example, an argon-krypton laser, with blue, green, and red wavelengths, would have an error of approximately 3% rather than the 15% error for a single wavelength. It appears that it is important that the source contain two or more widely spaced lines rather than having a continuous wavelength distribution as in incandescent sources. An extension of the theoretical studies described above [4,5] has indicated that particles in the diameter range 0.05 μm to 0.2 μm which are below the resolution of the ratio method (although they can be detected by the system) can be sized by a single forward lobe intensity or by a difference of two forward lobe intensities. The theoretical Figure 3. Error envelope for particle sizing instrument. error of these techniques is approximately 25% for all possible indices of refraction. However, the use of a single intensity or a difference of intensities will produce experimental errors which are not of concern in the ratio method. One possible way of eliminating the majority of the experimental errors is to normalize the intensity data by the time of flight of the particle through the laser beam. The time of flight is relatively easily determined and serves as a measure of the main beam intensity incident on each particle. III. CHEMICAL CHARACTERIZATIONS In many situations some degree of chemical characterization of particulates would be advantageous, especially if if it could be accomplished without collection and subsequent chemical analysis of a sample. As an outgrowth of the theoretical studies described above, which are concerned with a study of particle sizing methods, a technique was discovered which seems to offer a way of determining information related to the chemical composition of each particle as it passes through the laser beam. The information that can be obtained is a function of the imaginary component of the particle's index of refraction. some Figure 4 shows how chemical information can be deduced. The quantity I, represented as the integral over 40° to 70° of the scattered intensity in the plane of polarization is shown as a function of particle size for a series of absorbing and nonabsorbing materials. I can also be considered as the output of an intensity detector in the polarization plane with an aperture from 40° to 70°. For particles smaller than 0.3 μm, it is found that I is insensitive to chemical character but for larger particles the discrimination is quite good. The band represented by k=0 is the envelope region for all values of the real component of n but with no absorption character. A fairly complete analysis of this effect angular region from 40° to 70° is the most effective for discrimination among materials [6]. 10% 70 I = √ i(0)d0 40 has shown that the region to employ I 103 Figure 4. Detection of absorbing and nonabsorbing particles. |