Abstract
Recent advances in flow-based microfluidics have led to the emergence of biochemistry-on-a-chip as a new paradigm in clinical diagnostics and biomolecular recognition. However, a potential roadblock in the deployment of microfluidic biochips is the lack of test techniques to screen defective devices before they are used for biochemical analysis. Defective chips lead to repetition of experiments, which is undesirable due to high reagent cost and limited availability of samples. Prior work on fault detection in biochips has been limited to digital ('droplet') microfluidics and other electrode-based technology platforms. We propose the first approach for automated testing of flow-based microfluidic biochips that are designed using membrane-based valves for flow control. The proposed test technique is based on a behavioral abstraction of physical defects in microchannels and valves. The flow paths and flow control in the microfluidic device are modeled as a logic circuit composed of Boolean gates, which allows us to carry out test generation using standard ATPG tools. The tests derived using the logic circuit model are then mapped to fluidic operations involving pumps and pressure meters in the biochip. Feedback from pressure meters can be compared to expected responses based on the logic circuit model, whereby the types and positions of defects are identified. We show how a fabricated biochip can be tested using the proposed method, and we achieve 100% coverage of faults that model defects in channels and valves. © 2013 IEEE.