Robust In-Field Testing of Digital Microfluidic Biochips

Abstract

Microfluidic technology offers vast promise for implementing biochemistry-on-chip with diverse applications to clinical diagnosis, genome analysis, drug design, and point-of-care testing. Among various types of fluid-chips, droplet-based digital microfluidic biochips (DMFBs), which consist of a patterned array of controllable electrodes, provide the advantage of programmability, ease of fluidic operations, and versatile droplet mobility. However, because of manufacturing or field defects, electrode degradation, or dielectric breakdown, these chips may suffer from incorrect fluidic behavior. Reliability of fluidic operations is of utmost concern in DMFBs that are used to perform safety-critical bio-protocols. Various methods are deployed to test these devices, either offline or being overlapped with bioassay operations (termed as concurrent or in-field testing). The main challenge of in-field testing lies in the fact that the test must run concurrently with the execution of the normal assay without hampering the correctness of the latter. In prior work, optimal testing for droplet mobility over all electrodes was formulated in terms of finding either a Hamiltonian path or a Eulerian path in an undirected graph that represents the electrode-adjacency structure. Although these models have been studied for offline testing, no such effort was made in the area of concurrent testing. In this work, we propose, for in-field application, an SAT-based modeling and solution approach to find an optimal test plan that can be used to check droplet movement across the boundary between every pair of adjacent electrodes, which is visited by the droplets of the ongoing assay. The proposed method is robust and determines a test solution successfully regardless of the cover assay that is being executed concurrently. Experiments on several real-life assays and other test cases demonstrate the effectiveness of the method with respect to test completion time.