BIO27: Automated PDMS Engraving and Assembly of a Prototype Microfluidic Artificial Lung

Abstract

Purpose: The ability to cost-effectively produce high-surface area microfluidic devices would bring many small-scale technologies from the realm of research to clinical and commercial applications. In particular microfluidic artificial lungs (µALs) promise improved gas exchange efficiency through micron-scale feature sizes. However, most µALs have only a single layer which limits their rated flows. Efforts to scale up these devices, such as by stacking multiple flat µALs have been labor intensive and resulted in bulky devices. Here, we present an automated roll-to-roll (R2R) manufacturing system and a cylindrical multi-layer µAL. Methods: The R2R system (Fig 1A) begins with a roll of biocompatible silicone rubber (Silpuran 2030). A heavily-modified CO2 laser engraver (Omtech K40 with custom vacuum engraving bed, N2 purge, and custom motherboard) creates patterns in the silicone sheet (Fig 1B). The sheet is then plasma-activated using two CeraPlas F-series cold-plasma wands to bond the PDMS to itself in an irreversible bond. Finally, it is rerolled into a processed device around a 30 mm PDMS core. This entire process from unrolling to rerolling is performed simultaneously and synchronously, allowing the creation of very large area microfluidic devices. Further, it is automated (Arduino Mega 1680), eliminating human error over this repetitive process. Results: Using our automated manufacturing system, we fabricated a 27-layer cylindrical prototype µAL, consisting of 1) an 8-layer base, 2) 11 layers of alternating blood and gas channels, and 3) an 8-layer protective cover. Laser-engraved fluidic channels were triangular, 60 µm deep and 180 µm wide (Fig 1B). If unrolled, the microfluidic part of this system would stretch 1.1 m, and the total length would stretch 2.7 m. This was placed into a custom 3D printed housing which routes blood and gas to their respective microfluidic channels (Fig 2A,B; filled with dyed water for visualization). Empirical pressure drop data were gathered for flows from 1 mL/min to 10 mL/min, which were similar to theoretical values (Fig 3). While the R2R manufacturing system successfully created the prototype 27-layer µAL, the device exhibited some internal fluid pooling and layer-to-layer leaks during fluidic testing. Both issues are believed to be caused by inadequate localized layer-to-layer adhesion. We plan to remedy these issues in the coming months. Conclusion: This roll-to-roll manufacturing system is designed to manufacture the first large surface-area microfluidic devices that are impractical to manufacture by conventional means. This can be employed for large-scale membrane gas exchange such as human-scale µALs.Figure 1. (A) An automated roll to roll manufacturing system for silicone. The path of the silicone is shown in green. The silicone comes on a protective PET backing, which is removed and highlighted in cyan. (B) Confocal scan of engraved Silpuran 2030. Channels were 60 µm deep, and 180 µm wide.Figure 2. (A) Finished device, with custom housing (B) Device with colored water being routed through. Red arrows depict blood connections, and blue arrows depict gas connections. Two of the blood outlets filled with blue dye, and there is a dark blue region indicating imperfect layer-to-layer bonding resulting in leaking.Figure 3. Experimental vs theoretical pressure drop across the μAL (water)