Abstract :
Integrated photonic biosensors take advantage of the increased sensitivity of microstructures, and their detection principle is based on the modulation of light properties such as intensity, phase or wavelength. Detection by evanescent coupling of microparticles selectively trapped on the surface of an optical waveguide is a common solution. However, these sensors generally require potentially harmful and fragile chemical treatments of the detection zone, which severely limits the sensor’s cleanability and regeneration. This thesis presents an integrated photonic device on glass dedicated to the detection of bacterial viability in aqueous media. The intended applicationis the design of a water pollution sensor exploiting bacteria as sentinels of water toxicity. To overcome the limitations of surface functionalization, we propose a solution in which the sorting and trapping of alive and dead bacteria are based on the force of dielectrophoresis, induced by planar aluminum castellated electrodes, integrated on a glass substrate. Detection is achieved by a single-mode surface optical waveguide, located between the electrodes, whose guided light is disturbed by particles collected on its surface. A model has been developed to calculate Mie scattering by particles on the surface of such a structure. This semi-analytical modeling is based on a complex angle approach to evaluate the interaction between a surface-guided mode and dielectric spheres. In this approach, which is faster than 3D simulations, an evanescent incident wave is described as a plane wave undergoing geometric rotation through a complex angle. This model, which can be used to evaluate waveguide losses induced by dielectric microparticles placed close to its surface, provides a physical understanding of the behavior of the propagated optical signal. On the other hand, measurements carried out with polystyrene microbeads with a diameter of 1 μm show that the device can provide fast discrimination of the order of a minute for concentrations ranging from 1.36×10^8 to 4.55×10^8 particles/μL and spaced at around 4.55×10^7 particles/μL. Initial tests were carried out with alive and dead E. coli bacteria. They proved that the device is effective in discriminating bacteria according to their viability, as they exhibit different switching frequencies with respect to the dielectrophoresis regime. Optical detection of trapped bacteria was also demonstrated for a solution of 1.5×10^6 bacteria/μL in deionized water. The analytical model is then compared with these experimental results.
Jury members :
Elise GHIBAUDO , PROFESSOR of UNIVERSITIES - Université Grenoble Alpes : Supervisor
Christelle MONAT, PROFESSOR of UNIVERSITIES - Ecole Centrale de Lyon : Reviewer
Luiz POFFO, PROFESSOR of UNIVERSITIES - Université de Limoges : Reviewer
Marie FRENEA-ROBIN, PROFESSOR of UNIVERSITIES - Université Claude Bernard Lyon 1 : Examiner
Julien POETTE, PROFESSOR of UNIVERSITIES - Grenoble INP - UGA : Examiner