A scalable bonding technique for the development of nextgeneration brain-machine interfaces * .

Brain-machine interfaces (BMIs) have demonstrated potential both for neuroscience studies and for neural-prosthetic or therapeutic devices. Given the high density of brain neurons, engineering a dense array of neural recording sites across a large brain region is important for BMIs with significantly improved capabilities such as high-dexterity motor control. In our development of a high-channel-count and high-density BMI system using complementary metal-oxide semiconductor (CMOS) arrays mated with massively parallel microelectrodes, we determined that the connection between the microelectrode and chip interface (MCI) is a crucial process for device performance and scalability. Here we report our results extending flip chip bonding technique to establish the electrical connections en masse at the MCI. Key parameters affecting bonding quality were identified and optimized, and the quality of bonding was evaluated by electrochemical impedance spectroscopy (EIS) in phosphate buffered saline (PBS). With proper packaging, the bonding technique can be directly transferred to the fabrication of high-channel-count BMIs and standardized for broader applications where interconnection between massively parallel interfaces is required.