Development of innovative multi-physics solutions to optimize biological measure performance

Abstract: Paper strips are becoming widely used for blood testing and particularly with embedded immunoassays. Among the so called "Point of Care" devices, the association of LFIA tests with a reading technology able to capture the strip and analyze its composition to output the biological result is advantageous in many ways.         In this thesis, a biosensor composed of a LFIA strip embedded on a reading device is investigated. The variability of the optical system used in the device is assessed, and the dynamic of the capillary flow in the LFIA strip is investigated in order to build a model of the flow field to understand the test physically and further optimize it.         Statistical methods based on available database as well as experimental data are used in order to assess the optical system variability. The camera variability in positioning, zoom and sharpness is investigated, as well as LED lightning system reproducibility and repeatability. The camera shows a zoom variability of 6% when computed on the database of devices available, and around 40% for sharpness variability. It reveals that sharpness is the most sensible camera parameter, because images of different sharpness gives varying results in the signal processing measurement method used to output the biological result. The lightning system has a variability below 1%, a reproducibility of 4%, and a repeatability below 1%, which are low results and it comforts the reliability of the lightning system during the measurement process. The different quantitative results for the optical system performances give precious insight on the further signal processing methods used to analyze the strip composition.         The paper strip is composed of four different substrips placed horizontally in series with different length and cross section. The method consists in setting up a model for the flow field by combining Darcy's law and mass conservation at the junction between strips. A relation between the fluid velocity and travelling distance is deduced. The four paper parameters are determined experimentally and input into the model equations. The flow field model fits favorably to experimental result of the flow in the strip. The model presents velocity jumps at the junction between strips and the importance of paper parameters such as cross sections, porosity, permeability and capillary pressure to govern the flow field is stressed. The flow field microfluidic model demonstrates that Darcy's law has to be modified in order to account for the pad superposition.         This project lead to a quantification of optical system variability sources which help choosing the signal processing methods embedded in the software to read and analyse the LFIA strip test. They will also be further used for designing device quality tests to ensure that the device is able to output a reliable biological result. The microfluidic model built and combined with experimental flow data helps to understand and optimize the flow in the strip, which is responsible for bringing the analyte of interest towards the reaction zone on the strip. This model could be enhanced if linked to the biochemical reaction kinetic model, and could further be integrated into a more general biosensor model to predict its behavior physically and biologically.