Analysis and characterization of multi-scale scattering: Application to bacteria colonies

Eui Won Bae, Purdue University

Abstract

Many problems in science and engineering involve physical processes and a range of dimensions which can be categorized as micro, meso, or macro-scale. This thesis focuses on bacterial colonies and their rapid characterization and identification using light scattering as an important problem in multi-scale physics. Biological samples can be roughly categorized as single bacteria (micro-scale), bacteria clusters (meso-scale), and bacteria colonies (macro-scale). Light scattering approaches, to be effective in determining physical properties such as morphology and material composition, must comprehend this spectrum of length scales. The discrete dipole approximation (DDA), a powerful modeling tool for rigorous 3-D vector scattering, has shown its capability to predict the light scattering from micro-scale objects. To be able to accommodate meso-scale objects, we need to extend the computational limits of the DDA method such that it could compute object sizes of 10 - 103 λ. To accomplish this, an analysis of the DDA method was performed for meso-scale cases both in semiconductor and biological applications. This study included analysis on the Sommerfeld integration path, new iterative algorithm, and preconditioning techniques. For the macro-scale bacteria colonies, Rayleigh-Sommerfeld diffraction theory was implemented whereby the bacterial colony was modeled as an amplitude/phase modulator. To harness the full power of polarization characteristics of incident and scattered light, the Mueller matrix approach was employed in DDA for both free space and surface cases and compared to other existing theories. Mueller matrix techniques are applicable to all 3 size domains. This powerful modeling framework was then focused on the measurement challenge: early identification of bacteria colonies. The BActeria Rapid Detection using Optical Scattering Technology (BARDOT) system was designed and constructed with higher resolution CCD image sensor and 2 dimensional stages for sample scanning. Experiments were performed which revealed the time dependent nature of scattering patterns of bacterial colonies. Confocal microscopy was performed to reveal not only the internal structure but also the topographical information of the bacteria colonies. Based on the physical parameters obtained from the experiment, Rayleigh-Sommerfeld diffraction theory was applied to 3 different species (Listeria innocua, monocytogenes, and ivannovii) and simulated scattering images were compared with pattern captured from BARDOT system. The physics-based models adequately predicted the scattering characteristics from these very similar and difficult-to-separate Listeria species, and provided the long term validity of light scattering as a rapid, non-contact characterization tool for bacteria.

Degree

Ph.D.

Advisors

Hirleman, Purdue University.

Subject Area

Mechanical engineering

COinS