Microfluidic detection of waterborne pathogen through light scattering of particle immunoassays

This dissertation focused on detecting waterborne pathogens in a microfluidic biosensing system which enables point-of-care, real-time monitoring. Within this framework, I have been addressing three objectives.;The first objective was to enhance mixing of particles in a microfluidic device. To this end, SDS (sodium dodecyl sulfate) or Tween 80 (polyethylene sorbitol ester) was added to the antibody-conjugated polystyrene microparticle suspension. Both surfactants showed non-specific binding (with SDS) or very poor diffusion (with Tween 80). As an alternative approach, highly carboxylated polystyrene microparticles showed very low non-specific binding comparable to that with Tween 80 and good diffusional mixing equivalent to that with SDS. This work was published in Appendix A (© 2008 Elsevier).;The second objective was to detect E. coli K-12 using the microfluidic-based system with low detection limit in Appendix B (© 2008 Elsevier). This method was essentially one-step and requires no sample pre-treatment or cell culturing. Conventional immunoassay using polyclonal antibody detects not only viable cells, but also dead cell and free antigens. In order to reduce false positive signal originated from dead cells and free antigens, target solution was washed three times. The detection limit was as low as 40 cfu ml-1 or 4 cfu per device (viable cells only), or <10 cfu ml-1 or <1 cfu per device (including dead cells and free antigens).;Our final objective was to develop real-time, high sensitive method for detecting waterborne pathogens in a water distribution system in Appendix C. Detection of Escherichia coli (E. coli) in a single straight pipe was demonstrated using a microfluidic 10 system utilizing light scattering detection of latex immunoagglutination assay. Assay time is <5 min per assay and detection limit is 10 cfu ml-1. Optical signals are compared with viable E. coli counts (not real time) and salt tracer experiments. Laminar (Re = 1,102) and turbulent (Re = 6,144) flows are used to simulate the flow regimes in a real water distribution system.