As part of the U.S. Department of Homeland Security Detect-to-Protect program, a multilab [Sandia National Laboratories (SNL), Lawrence Livermore National Laboratories (LLNL), Pacific Northwest National Laboratory (PNNL), Oak Ridge National Laboratory (ORNL), and Los Alamos National Laboratory (LANL)] effort is addressing the need for useable detect-to-warn bioaerosol sensors for public facility protection. Towards this end, the SNL team is employing rapid fluorogenic staining to infer the protein content of bioaerosols. This is being implemented in a flow cytometry platform wherein each particle detected generates coincident signals of forward scatter, side scatter, and fluorescence. Several thousand such coincident signal sets are typically collected to generate a probability distribution over the scattering and fluorescence values. A linear unmixing analysis is performed to differentiate components in the mixture. After forming a library of pure component distributions from measured pure material samples, the distribution of an unknown mixture of particles is treated as a linear combination of the pure component distributions. The scattering/fluorescence probability distribution data vector a is considered the product of two vectors, the fractional profile f and the scattering/fluorescence distributions from pure components P. A least squares procedure minimizes the magnitude of the residual vector e in the expression a = fP^T + e. The profile f designates a weighting fraction for each particle type included in the set of pure components, providing the composition of the unknown mixture. We discuss testing of this analysis approach and steps we have taken to evaluate the effect of interferents, both known and unknown.

Date Published: 15 April 2008
PDF: 8 pages
Proc. SPIE 6945, Optics and Photonics in Global Homeland Security IV, 69450R (15 April 2008); doi: 10.1117/12.777407

Published in SPIE Proceedings Vol. 6945:
Optics and Photonics in Global Homeland Security IV
Craig S. Halvorson; Daniel Lehrfeld; Theodore T. Saito, Editor(s)