By employing the particle sizing instrument described above and adding to it an optical system which collects the light scattered by each particle over the angular region of 40° to 70° in the plane of polarization it should be possible to classify those particles which are larger than 0.3 um as to whether they are nonabsorbing, carbon-like, or metallic. In somes cases, it also appears that some discrimination to the type of metal might be possible. as In order to test the ability to distinguish between absorbing and nonabsorbing particles, a smoke detector based on a measurement of the intensity in the polarization plane from 40° to 70° was constructed. The primary method now employed for the detection of fire-produced aerosols involves monitoring a signal related to the total particle density. This technique has a disadvantage in that certain nonfire produced aerosols may have total particle densities high enough to cause confusion. These aerosols might be cigarette smoke, industrial dusts or fumes, or "hair spray" aerosols. The smoke detector instrument contained a fairly straightforward optical system. Light from an an incandescent source was polarized, collimated and impinged on an aerosol. Light scattered by the aerosol was monitored over two regions. One of these, I,,, was over the angular range from 40° to 70° in the plane of polarization. This quantity should be relatively large for nonabsorbing aerosols and should be less for aerosols containing absorbing materials. The second region was perpendicular to the polarization plane and over the angular range 80° to 100°. This intensity, I can be shown to be approximately proportional to the total number of particles in the beam regardless of whether they are absorbing or not. For display purposes, these two outputs have been applied to the vertical and horizontal axes of an oscilloscope, and the traces for a series of materials shown in figure 5. The length of the line is arbitrary and is related to how many particles are in the beam at any one time. As can be seen, I1, is large relative to I for nonabsorbing materials such such as steam, cigarette smoke, and freon containing aerosols. However, for metal dusts and carbon containing materials I is small relative to I. Thus, smoke detector which went into alarm on the slope of the I/I curve rather than the total particle density could distinguish between fire-produced and nonfire-produced aerosols. REFERENCES [1] Gravatt, C. C., "Real time measurement of the size distribution of particulate matter in air by a light scattering method," J. Air Pollution Control Association, 23:12, 1035-1038 (Dec. 1973). [2] Hodkinson, J. R., "Particle sizing by means of the forward scattering lobe," Appl. Opt.5:5, 839-844 (May 1966). Figure 5. Smoke detector response. I, and I are defined in [3] Gravatt, C. C., and Allegrini, Ivo, A new Light Scattering Method for the Determination of the Size Distribution of Particulate Matter in Air, Proceedings of the 3rd International Clean Air Congress, p. C3 (1973). [4] Greenough, M. L., and Gravatt, C. C., EAP Final Report on Particle Sizing Instrument (in preparation). [5] Gravatt, C. C., and Allegrini, Ivo, Particle Sizing by Forward Lobe Techniques (in preparation). [6] Gravatt, C. C., and Allegrini, Ivo, The Characterization as to Their Optical Absorbing Properties Particulates of (in preparation). DISCUSSION GEORGE SINNOTT: In using two wavelengths to decrease the size do you mean that one would take the two wavelength scatterings separately? GRAVATT: No. The laser is in the all-line mode so there will be some eight lines emitted by the laser--the three prominent ones are the 448.0, 514.5 and 640.0 nm. These calculations are done weighing these intensities appropriately. It is It is assumed that all the light is collected and it results in a smoothing of the ripples of the response curve shown for a single wavelength. RONALD NELSON: Going back to your calculation of the ratio as a function of the particle size in the envelope, my question is, when it gets down to the practical business of calibrating, how does one choose the center of that envelope for the real particle calibration? GRAVATT: Let's go back to figure 3 and it will probably be easier to see what is involved. For instance, we have calculated a large number of these curves to make up the envelope. There is one for polystyrene latex, particles the index of those is 1.59. Using that curve and and the instrument output for a number of polystyrene latex samples it is possible to transform the data from a relative scale to an absolute size scale. It is a problem to generate monodisperse aerosols, however. There is not a large number of well standardized or well calibrated latex or aerosol aerosol standards. So calibrating any particle sizing instrument is going to be somewhat difficult. NELSON: I think that this ought to be considered as to how does one define this. I think that maybe we ought to advocate using the full width of this envelope as opposed to maybe half-width because if you are going to pick half it implies you know where the middle of it is and that is not necessarily the case. GRAVATT: Well, as far as as stating errors is concerned, one can report the full width or the half width. I do not feel that it is difficult to define where the middle of the curve is. It does not seem to me to be a problem that the particular calibration samples employed do not lie on the center line. However, since neither NBS or EPA have done an extensive calibration of the instrument I will have to withhold final comment until we get some experience and data. I personally feel that the most important problem in calibration is that of generating monodisperse aerosols. ROBERT KNOLLENBERG: Did you say the mean particular size is 0.3 micrometers? GRAVATT: The minimum resolvable particle size for a ratio if 10° to 5° is .3 μm. You can detect particles down to about 0.1 μm. With 3 1/3 orders of magnitude gain in the electronics, if the large particle cutoff is set at approximately 3 μm. KNOLLENBERG: Do you know roughly what percentage of the total light scattering you collect? GRAVATT: No, I don't know it exactly although I can calculate it. I would estimate that we collect several percent. KNOLLENBERG: You are collecting a large percentage then. GRAVATT: The EPA instrument has 1° aperatures and this probably collects a few percent of the total scattered intensity. NELSON: I guess it is about 1 percent. GRAVATT: It may be 1 percent. WILLIAM DORKO: Is this calculation dependent upon the structure of the particle such as spherical? GRAVATT: All calculations are for spherical particles. It is stated that the forward lobe is least sensitive to particle shape. Dr. Kerker could probably make a better comment to that than I can. Hodgkinson also felt that way, but we have not done any calculations. We've tested nonspherical particles in the instrument, and they fit within the calibration curve. That's not really proof, it's only a suggestion. This is certainly an area where much more work is needed. NATIONAL BUREAU OF STANDARDS SPECIAL PUBLICATION 412 FLOW APPARATUS FOR THE CHARACTERIZATION OF AEROSOLS Madhav B. Ranade Fine Particles Research ABSTRACT A flow apparatus has been developed to enable the size measurement of liquid aerosols to be measured under defined conditions with the help of a light scattering technique. This apparatus was used to study the evaporative behavior of submicron aqueous aerosols during transportation in air. The apparatus is currently being used to characterize large (~ 50 μm) aerosol particles under sedimentation. Portable versions of the flow apparatus are presently under development to characterize acid mist droplets, and to investigate the characteristics of therapeutic aerosols. Key words: aerosol size measurements; aerosol spectrometer; aerosol sprays; condensation on aerosol droplets; evaporation of aerosol droplets; laser light scattering by aerosols; therapeutic aerosols. I. INTRODUCTION An aerosol system is characterized by the size distribution of the particles, concentration of the particles, and the physical nature of the particles. Physical properties of aerosols are strongly dependent on the particle size characteristics. Sampling and measurement techniques for liquid aerosol systems must be chosen with a regard to physical conditions. In order to obtain a representative sample from an aerosol system, changes in the temperature, pressure, and flow field during sampling must be minimized. Volatile liquid aerosols pose a special problem, as the particle size may change significantly during sampling. Examples of such aerosols are the therapeutic aerosols and aerosol can sprays. In the therapeutic applications, the drug is dispersed in the form of aqueous solution. In the aerosol cans, the product is dispersed by spraying with a volatile solvent in the form of droplets up to 100 μm in diameter. The particle size of the final residue may vary from submicronic to a few micrometers. |