Optimization of injection molded polymer lab-on-a-chip for acoustic blood plasma separation using virtual design of experiment

Abstract

This study presents the development towards the optimization and manufacturing of a novel cost-effective fully polymer-based ultrasound actuated chip as a Point-of-care (POC) diagnostic platform for blood plasma separation and analysis. The chip was designed following multi-physic functional simulation to obtain optimal dimensions with respect to blood separation enhanced capacity. Initial dimensional guidelines showed most favorable performance to be achieved with design dimensions of 150 µm height, 375 µm width and 34 mm length. The chip encompasses a micro-fluidic channel with a relatively long length compared to the height and the width. As such, main manufacturability challenge is represented by the capability to maintain the channel dimensional conformance throughout the entire length. In order to tackle the challenge and optimize the results, developing a feedback loop was initiated for the device process development where the data of the preliminary micro-injection molding (µIM) production was exploited to implement a simulation-based full-factorial virtual design of experiment (DOE) for all the four materials of interest: polymethyl methacrylate (PMMA), cyclic olefin copolymer (COC), polycarbonate (PC) and polystyrene (PS). The four essential µIM parameters namely, mold temperature, melt temperature, injection velocity and packing pressure were used as factors. Overall, simulations were performed based on the virtual DOE inputs for the four aforementioned materials. The input data for the utilized process parameters in the virtual DOE was provided by our first production batch. Ultimately, in the case of PMMA, parts were molded and characterized based on the process settings that had demonstrated the lowest linear shrinkage in our entire virtual DOE responses. For the linear shrinkage of these parts, a respective width and height difference of 1% and 3.3% between the simulations and production results were obtained. Moreover, as a secondary study sink marks were also inspected and minimized through simulations. Initially, sink marks as deep as 80 µm were spotted on the chip. After the use of optimization cycle the value decreased to as low as 10 µm. The results confirm the positive impact of the virtual experiments for future rapid assessment of potential new chip designs.