Computational Analysis and Design Optimization of Convective PCR Devices

Polymerase Chain Reaction (PCR) is a relatively novel technique to amplify a few copies of DNA to a detectable level. PCR has already become common in biomedical research, criminal forensics, molecular archaeology, and so on. Many have attempted to develop PCR devices in numerous types for the purpose of the lab-on-chip (LOC) or point-of-care (POC). To use PCR devices for POC lab testing, the price must be lower, and the performance should be comparable to the lab devices. For current practices with the existing methods, the price is pushed up higher partially due to too much dependence on numerous developmental experiments. Our proposition herein is that the computational methods can make it possible to design the device at lower cost and less time, and even improved performance. In the present dissertation, a convective PCR, that is the required flow circulation is driven by the buoyancy forces, is researched towards the use in POC testing. Computational Fluid Dynamics (CFD) is employed to solve the nonlinear equations for the conjugate momentum and heat transfer model and the species transport model. The first application of the models considers four reactors in contact with two separate heaters, but with different heights. Computational analyses are carried out to study the nature of buoyancy-driven flow for DNA amplification and the effect of the capillary heights on the performance. The reactor performance is quantified by the doubling time of DNA and the results are experimentally verified. The second application includes a novel design wherein a reactor is heated up by a single heater. A process is established for low-developmental cost and high-performance design. The best is searched for and found by evaluating the performance for all possible candidates. The third application focuses on the analysis of the performance of single-heater reactors affected by positions of a capillary tube: (1) horizontal, and (2) vertical. In the last application, numerous double-heater reactor designs are considered to find the one that assure the optimal performance. Artificial Neural Network (ANN) is employed to approximate the CFD results for optimization. In summary, through the four segments of our studies, the results show significant possibilities of increasing the performance and reducing the developmental cost and time. It is also demonstrated that the proposed methodology is advantageous for the development of cPCR reactors for the purpose of POC applications