DISCUSSION MADHAV RANADE: Is that 20 seconds your measurement time or total? CHABAY: Twenty seconds is the total time to accumulate 128 averages. A single scan and display of a spectrum requires 150 milliseconds. The minimum period of observation of the photocurrent depends on the frequency range, i.e., on the bandwidth of the spectrum analyzer. RANADE: Isn't it true that within the time of observation you will have evaporation or growth of the water droplets in your cloud chamber? CHABAY: Yes, but the point is that this is a steady state chamber. As the particles fall through the chamber, you have a steady state dition at each height. That is, if growth occurs through the chamber at known conditions of supersaturation as a function of height, the droplet size distribution is constant at each height. MILTON KERKER: You pointed to the limitation that can be caused by the Brownian motion. What about convection? Isn't this a problem? CHABAY : It's very definitely a problem. It's something that we struggled with in our first set of experiments. I think that it can be eliminated, certainly for particles of over a 1/2 micrometer. I haven't tried to work with particles smaller than that. My guess is that it can be sufficiently minimized to allow work with even smaller particles, where one would determine a net diffusion constant, rather than the size distribution itself. One can measure the diffusion constant as an average quantity when Brownian motion is predominant over settling velocity. I think convection can be controlled well enough to do that by proper insulation of the chamber. RONALD NELSON: How would you adapt this method to measurements of aerosol in aerosol cans? CHABAY: The system is essentially the same. Basically one needs to have a very slow air flow into which the aerosol particles are introduced. A very simplified version would have a large scattering chamber with aerosol can on top. You push the button of the can and aerosol particles enter the chamber, fall through it, and are removed at the bottom. Let me show you the present apparatus (fig. 2). Water saturated air flows in through an annulus at the top next to the heated upper block, then down through the chamber to the exhaust annulus at the bottom. The wire screens help make the air and particle flow more uniform. If the entire chamber were kept at constant temperature and aerosol were sprayed in at the top, the same kind of data could be collected for aerosols as has been for water droplets. Figure 2. Schematic of cloud chamber for producing aerosols and proving scattering volume. NELSON: Can you get absolute concentrations this way, with the essentially unknown flow-rate? CHABAY: The flow-rate is something that one can determine independently but the absolute concentration is not something that we were trying to measure. We were really looking at the relative distribution of sizes of particles rather than absolute numbers. KERKER: Are you familiar with Joe Katz's chamber? CHABAY: Yes. KERKER: Could this technique be used to get a finer handle on the rate of nucleation? CHABAY : That was included included in our original plans, but as yet we haven't had time to pursue it. One could change the chemical nature of the nuclei nuclei introduced with the inlet air, then observe the growth process as a function of the type of nuclei. This could complement the work described in a recent paper by Mirabel and Katz on binary homogeneous nucleation and aerosol formation. KERKER: We are getting a little off the topic, of course, but are these people actually counting drops? CHABAY: Right. KERKER: But are you actually able to follow the rate of growth? KERKER: Howard Reiss of UCLA found that by using his cloud chamber and using the theory of nucleation by Byer and Greissers he could study kinetics of formation of droplets. What I'm thinking is that this heterodyne technique could give you a handle on growth rates. CHABAY: I think so. We are just now publishing some growth curves from the data that we have collected, and I think that this could be extended to a very wide range of systems. It gives you very clearly the growth characteristics of the material. KNOLLENBERG: In studying the nucleation of particles down to the order of 0.02 micrometer, you are talking about a nucleation technique with particles that would be limited to the order of one micrometer. CHABAY: The point is that with this technique one cannot study directly the nucleation process for particles less than 1 micrometer, let's say. However, what one can see is the effect of growth as a function of the much smaller nuclei. KERKER: Reiss is counting the total number of grown particles, and he is getting a rate of nucleation from that. He (Chabay) can go further back into the growth stage and actually follow the growth--not all the way back, no. CHABAY: No, I can only start at about 1/2 micrometer size particle and work from there up to very large particles. RANADE: seconds. They will have grown to a few micrometers in a matter of CHABAY: You see the point is that in this kind of chamber the time scale for that actually to occur is irrelevant. I do not have to resolve the time scale, I only have to resolve the height. How well I can resolve a given time difference is a question of the size of the droplets, how fast they are falling, and how well I can manipulate the beam in terms of height--the resolution in height of this arrangement. I should mention in regard to this, that at the moment we were using a beam of about 2 mm diameter as a matter of convenience. We just used the output beam directly. One could focus that down to a much smaller beam, and get a higher resolution of the distribution and also greater detail in terms of the growth process. CARY GRAVATT: Do you get instant readout on the particle distribution and how long does it take? CHABAY: What you get is the spectra such as I showed you, which are power spectra. In order to get a size distribution out of that, you have to divide by the Mie scattering curve which is a fairly trivial arithmetic operation. A better way to do it, which we didn't do because didn't have facilities to do it directly, but I hope to do it while I'm here at the Bureau, is to do processing directly on-line with with the 71 computer. That is, is, one could do this whole thing on the computer and immediately process the data so it would be a matter of fractions of second after you have taken the number of scans desired that the size distribution would appear. SINNOTT: Another technique very similar to this is to set up an interference pattern in the scattering region and then watch the particle scattered light fluctuations. that. CHABAY: Right. I think Dr. Yanta is going to be talking about SINNOTT: techniques? Do you have any remarks on the relationship of these two I CHABAY: Maybe it's appropriate to wait until Dr. Yanta's talk. think generally the techniques are very similar. I know of no no other study which has put together both the intensity and the falling velocity information to get the size distribution. You can certainly get the velocity of the droplet by the laser Doppler velocimeter measurements. The basic physics of the technique is the same. In our case, we do image a real fringe field in space and look at the particle moving through the light and dark regions to determine the rate rate at which it goes through the fringes. The process of heterodyning gives the same information on velocity, though. not NATIONAL BUREAU OF STANDARDS SPECIAL PUBLICATION 412 MEASUREMENTS OF AEROSOL SIZE DISTRIBUTIONS WITH A William J. Yanta Naval Ordnance Laboratory ABSTRACT A miniature wind tunnel has been built which together with the Laser Doppler Velocimeter (LDV) has been used to determine aerosol size distributions. In principle the LDV was used to measure the particle lag of individual aerosol particles as they were accelerated through a small supersonic nozzle. The measured velocity lag was then used in conjunction with numerical predictions to determine the particle size. An optical owl was used to determine the mean of the size distributions. The LDV measurements were in good agreement with the owl measurements. Key words: aerosol sizing; aerosol spectrometer; aerosol sprays; Doppler measurements of particle size; droplet sizing; interferometer; laser light scattering by aerosols; particle size measurements; particle velocity measurements. INTRODUCTION The Laser Doppler Velocimeter has been used for many years in measuring flow velocities. The primary requirement for the LDV is that micrometer size particles be present in the flow. The LDV then measures the velocity of these particles. If one assumes the particles are moving at the same velocity as the fluid (either gas or liquid), then the fluid velocity is inferred directly from the particle velocity. However, in air, particles greater than one micrometer may not follow the flow precisely, that is, the particle may lag the flow. This is especially true in supersonic flow [3]. It is demonstrated in reference 3 that the particle lag can be predicted if the particle size is known. Conversely, one can predict particle size if the particle lag is known. This particle lag can be determined from velocimeter results (particle speed) in a calibrated or known flow field from which the particle size can be deduced. The particle lag can be induced by placing a particle in a rapidly accelerating or decelerating flow field. A convenient method of generating this type of flow field is with a supersonic nozzle. Using the computational procedure described in reference 3, one can readily predict a particle's velocity in a supersonic nozzle. *This paper is a unification of references [1] and [2]. |