Intelligent sensing and actuation systems in microfluidics

In recent years, the technology of microfluidic devices has spread in multiple scientific disciplines, including chemistry and biology. In these scopes, microfluidics gave birth to Lab-on-Chips, powerful platforms able to integrate laboratory facilities within a few micrometres. Particularly, when applied to a human organ mimicking, Lab-on-Chips are usually known as Organ-on-Chips. However, the possibility of introducing multiple degrees of freedom in the platform can be limited if the scientist does not have complete control and understanding of what is happening during the experiment, particularly in the Organ-on-Chips. For this reason, it is crucial to work on introducing sensors and actuators into microfluidic devices, permitting us to monitor the physical and chemical quantities inside the experiment and act on it. In addition, the rise of machine learning algorithms and their interaction with sensing microfluidic systems can be the link that can permit these platforms to act independently, boosting the diffusion of this technology into the industrial and commercial world. This thesis project proposes two techniques the researchers can exploit to integrate these instruments into the Lab-on-Chip platform. The thesis will demonstrate a technique based on Laser-Induced Graphene to deliver low-cost conductive electrodes onto biocompatible and transparent plastic substrates. The transferred electrodes were highly characterised and then utilized to deliver a dielectrophoretic force to polystyrene beads. After that, similar electrodes were exploited to electroporate an adherent glioblastoma cell line, monitoring the effect of the electric field in a label-free manner using a custom machine-learning algorithm. Next, focusing more on sensor integration, we developed a low-cost optical setup and protocol to generate hydrogel microstructures inside a Lab-On-Chip environment. This technology was used in combination with an inexpensive multispectral source and a dedicated machine-learning algorithm to extract the pH trends from colourimetric optical sensors integrated into 3D cell culture. Finally, the described protocol was empowered using DNA scaffold, an extremely recent technology used here to generate a stable and reliable ratiometric sensing system.