Fast Scanning Electrical Mobility Spectrometry: Theory, Experiments, and Inversion Methodologies.

摘要

Ultrafine particles (diameter less than 100 nm) play a critical role in determining local air quality and from a human health perspective (Oberdorster et al. 2004, 2005). The environmental/health effects of these particles are often associated with their surface area, and, hence, measurements of particle number concentrations are the most appropriate to characterize the impact of these particles on their environment. The entire range of ultrafine particles, from ~ 2 nm to 100 nm, can be characterized in several minutes with available techniques. These measurement times may, however, be too slow when conditions of fast particle dynamics affect particle populations. The objective of this work is to develop instruments and methodologies for near real-time particle size distribution measurements of ultrafine particles. For the particle size distribution measurements of these particles, the proposed work aims to build on the past developments of the DMA (Differential Mobility Analyzer; Knutson and Whitby, 1975) because of the accuracy of these instruments and their use for downstream processing measurements with a tandem-DMA setup. The DMAs, however, have a slow time response, making them only useful in applications where a relatively long measurement time is acceptable. A key constraint in the deployment of DMAs for faster measurements is the lack of knowledge of their classification capability, or transfer function, under fast scanning. Also, the quality of particle size distributions calculated under fast-scan operation will be strongly affected by the low signal-to-noise ratio that results from such operation and the ill-conditioning of the Kernel function because of the degradation of transfer function resolution.;As part of this thesis, a new approach to determine scanning DMA transfer functions is introduced. The trajectory-based approach of Dubey and Dhaniyala (2008) is used in combination with Monte Carlo simulations to determine mobility-based transfer functions of diffusional particles in scanning DMAs. With the knowledge of the ATF (Arrival time transfer function) area and the resolution, near real-time calculation of the instrument transfer function is possible. Estimation of particle size distributions from SEMS measurements, particularly under fast scan operation, will require the solution of the discretized form of the Fredholm integral equation of the first kind, requiring inversion of a kernel function. A new Multiscale Expectation-Maximization approach for SEMS particle size distribution calculation is introduced here. This approach is seen to provide a robust solution for a range of test particle size distributions (narrow or smooth) and is recommended for use with SEMS measurements when the solution characteristics are unknown.;To address the practical aspect of fast particle size distribution measurements, a new High-flow Dual Channel Differential mobility analyzer is designed, fabricated, and calibrated. The HD-DMA has a classifier section with a large radius that permits the operation of the instrument at high flow rates, while maintaining low Reynolds number in the classifier section. In order to cover a broad size range, the HD-DMA is designed with two sample ports. The upper port is used to sample particles in the size range of 2-80 nm and the lower port can sample particles in a size range of 5 to 200nm, for a large sample flow rate of 15 LPM. Preliminary measurements suggest that the combination of the HD-DMA, the new scanning DMA transfer functions, and the use of appropriate inversion algorithms, permits very fast particle size distribution measurements in ~ 5s.