Initial transmission spectra are taken of aerosol samples levitated in an ultrasonic field. Small amounts of the sample are levitated and subsequently collected for size distribution analysis.The whole experiment is conducted under a neutral dry nitrogen atmosphere and the experiment set-up means we have the capabilitiy to heat (100°C) and cool (-50°C) the atmosphere. Samples so far investigated are: soot, sand and two volcanic ashes, one being a basalt (low percentage of silicate: approx.50%) and the other being a Rhyolite (High percentage of silicate, approx. 75%). The natural samples have been sieved into four size bands: 32µm, 32-45µm, 45-63µm and 63-90µm; covering the spectrum of what is classed as an atmospheric aerosol. Figure 1 is an example of spectra obtained.
These Initial spectra look unlike other typical aerosols transmission spectra as the ultrasound has caused coalescence of the smaller particles. In certain ambient conditions ultrasound will preferantially cause some sizes in the distribution to coalesce. In these experiments the optimum coalescence occurs for particles approx. 10um in diameter. The size distribution is then shifted towards larger diameters and so the scattering part of the signal has shifted towards lower wavenumbers. In future experiments we will manipulate the ambient conditions to minimise coalescence of these particles in order to get spectra of different size ranges
We have been able to determine the size distribution of the natural samples from Scanning Electron Microscope (SEM) pictures. This method will be used to find the size distribution of the sieved samples and the samples retireved after levitattion in each run. An example of the SEM pictures of the Rhyolitic volcanic ash are shown in figure 2.
Initially specialist software was used to automatically count the size distribution but it was found that th edeliniation routines were not appropriate for the texture of our samples and so (unfortunatley!!!) manual counting was found to be the next best method. The resulting histogram of the size distribution of the basaltic sample is shown in figure 3
There were two resonant trapping nodes present in the trap during the course of each experiment although only one is in the line of the infrared beam. Videos, taken of the trap during each experiment, have been used to determine the relative amount of dust present in each node. Several frames were extracted from the videos and, close-up snap shots of each node extracted.The method is illustrated in figure 4
The pixel colour that determines the edge each of the nodes are chosen by and the number of pixels within the node area counted, thankfully not by hand!! The number of pixels represents the area projected by the node, from this the ratio of the volumes of the nodes are calculated assuming an ellipsoid. The average volume ratio of the nodes for several frames was taken. The ratio of the volumes range from 0.3-0.5. Figure 5 shows a histogram of pixel colour within a close up snap-shot of a node
This should form the complete set of data needed to retrieve the complex refractive index of our samples.
SBSL