Cyber-physical Integration for Digital Microfluidic Biochips

Abstract:

The goal of this project is to integrate digital microfluidics systems with thin-film photodetectors in the top plate to realize biochemical target sensing using fluorescence. System control, adaptation, and reconfiguration through software will lead to a general-purpose lab-on-chip computing platform, in the same way as programmable computing devices allow multifunctional capabilities via software on a hardware platform. This level of integration, decision, and controlled reconfigurability will be a significant step forward in clinical diagnostics. Research is underway on integrating the physical components of this technologythe microfluidic platform and miniaturized sensors, and the cyber componentssoftware for control, decision-making, and adaptation. In particular, hardware/software co-design methods are being developed to ensure that biochips are as versatile as the macro-labs that they are intended to replace. Specific tasks for this project include:

• Hardware platform development: Research into silicon-based digital microfluidics and integration of optical sensors with digital microfluidics.
• Cyberphysical system control: Closed-loop operation under software control; error recovery and controller/microfluidics integration; control software for run-time optimization.
• Decision-tree architectures: Smart decision-making during run-time; adaptive reconfiguration; multi-step bio-molecular recognition strategies.
• System demonstration: The complete cyberphysical testbed will be demonstrated for the problem of nucleic acid identification on a fabricated chip with detection sites.

During Year 2, on-chip DNA purification and quantification have been demonstrated. DNA purification is realized by Immiscible Phase Filtration (IPF). Currently, the detection limit is approximately 243 copies of Alu Yb8 per 92 nl PCR mix, and the dynamic range is 103. An integrated thin-film resistive heater has been developed and tested for on-chip PCR.Annular thin-film photodectors (PDs) have been designed, fabricated, and integrated into the top plate of a digital microfluidic system. The PDs have a hole in the center to input optical power into the system to pump the fluorophore. The fabricated PDs have been characterized, and have excellent dark current, responsivity and SNR, comparable to commercial thick Si PDs. We developed a hardware-assisted error-recovery method that relies on an error dictionary for rapid error recovery. Error recovery solutions are stored as a dictionary in the memory on FPGA before the execution of the bioassay. The dictionary has a decision-tree structure and is organized as a finite-state machine (FSM). The cost, the size and response time of the system is reduced with this error-recovery method. A new biochemistry operation-interdependency-aware synthesis method has also been developed. When there exists completion time-uncertainty, the proposed synthesis method can reduce completion time and achieve higher accuracy for the execution of operations. As the next step, we are now gearing up for experiments and cyberphysical demonstration using fabricated biochips and integrated sensors. These experiments will involve more intelligent sensor-driven response compared to the experiments reported in Year 1.