The design and characterization of thin-film ribbon cables as electrical interconnects for implanted neural stimulation and recording devices are reported. Our goal is to develop flexible and extensible ribbon cables that integrate with thin-film, cortical penetrating microelectrode arrays (MEAs). Amorphous silicon carbide (a-SiC) and polyimide were employed as the structural elements of the ribbon cables and multilayer titanium/gold thin films as electrical traces. Using photolithography and thin-film processing, ribbon cables with linear and serpentine electrical traces were investigated. A cable design with an open lattice geometry was also investigated as a means of achieving high levels of extensibility while preserving the electrical function of the cables. Multichannel ribbon cables were fabricated with 50 mm lengths and metallization trace widths of 2-12 μm. The ribbon cables tolerate flexural bending to a radius of 50 μm with no change in trace impedance but tolerate less than 5% tensile elongation without trace failure. Ribbon cables with a lattice structure exhibit 300% elongation without failure. The high elongation tolerance is attributed to a lattice design that results in an out-of-plane displacement that avoids fracture or plastic deformation. Extensible ribbon cables underwent up to 50,000 tensile elongation cycles to 45% extension without failure. An electrical interconnect process using through-holes in the distal gold bond pads of the ribbon cables was used to connect to an a-SiC-based MEA. The electrical connection was created by stenciling a conductive epoxy into the through-holes, bridging metallization between the traces, and MEA. The interconnect was tested using a ribbon cable connected to an a-SiC MEA implanted acutely in rat cortex and used to record neuronal activity. These highly flexible and extensible ribbon cables are expected to accommodate large extensions and facilitate cable routing during surgical implantation. They may also reduce tethering forces on implanted electrode arrays, potentially improving chronic neural recording performance.
Read full abstract