Highly conformable chip-in-foil implants for neural applications

Current neurotechnology trends focus on mechanically compliant, miniaturized implants with high integration density and decades-long lifespans to avoid replacements. Numerous neural implants (brain pacemakers, retina implants, brain-computer interfaces) aim to restore lost functions and improve patient quality of life. Advances require cross-disciplinary work in neuroscience, biotechnology, material science, and engineering. Sophisticated silicon (Si)-based integrated circuits (ICs) offer high complexity and spatial resolution but their stiffness causes a foreign body response, leading to tissue encapsulation and compromised functionality. Matching the implant's mechanical properties to the host tissue and enhancing conformability minimizes this response. Polyimides (PIs) are ideal substrate materials due to their chemical stability, biocompatibility, mechanical properties, low water uptake, high electric resistivity, and dielectric strength. The combination of compliant PI substrates and small Si-based circuits offers complex functionality and good adaptation to host tissue biomechanics, allowing coverage of larger brain areas. While intrinsically flexible electronics using organic transistors are developing, they currently lack the integration density and complexity of Si-based technology. Modular circuit arrangements are beneficial for adapting implant designs to various applications (monitoring large cortical areas, organ healing, peripheral nervous system bridging). Implant design must consider target tissue mechanical and geometrical conditions (e.g., brain softness and curvature). Conformability, the ability of a foil to attach and conform to a body's shape, ensures close electrode-cortex contact. Hybrid bioelectronic devices—sophisticated ICs modularly distributed on miniaturized flexible substrates—offer a promising approach to combine complex functionality with longevity in the biological environment. This work focuses on design rules for implant geometries and thin-film routing to minimize mechanical stress and ensure system integrity and good electrode-tissue contact. The goal is to prevent delamination of dice from the substrate or interconnect damage.

The introduction extensively reviews the existing literature on neural implants, highlighting the limitations of current technologies and the need for more biocompatible and conformable devices. It discusses the advantages of silicon-based integrated circuits and polyimide substrates, citing relevant research papers that support the choice of these materials. The review also touches upon the challenges of integrating complex functionality with long-term stability in the biological environment and the need for modular designs adaptable to diverse applications.

The study employed a two-pronged approach: finite element modeling (FEM) and microfabrication. The FEM simulation modeled the mechanical behavior of the hybrid implant conforming to a rat's somatosensory cortex (approximated as a cylinder with a 3 mm radius of curvature). The simulation assessed the impact of die geometry (specifically, edge fillet radius) on stress distribution at the die-substrate interface. Delamination height was used as an indicator of die-substrate integrity. The simulation also investigated the effect of interconnect routing proximity to die corners and the placement of contact pads on dice. The microfabrication process involved transferring and aligning multiple Si-based dice onto flexible PI-based substrates and establishing electrical interconnections. The process allowed for arbitrary die shapes and sizes at independent target positions. The process steps included placing dice into a carrier substrate, structuring the PI substrate with interconnects and contact pads, thinning the die backside, and releasing the finished devices. The developed process enabled batch transfer of multiple dice, offering scalability and facilitating the creation of prototypes.

FEM simulations revealed that providing edge fillets in the die base shape significantly improved die-substrate integrity and increased the available area for contact pads. A larger edge fillet radius (50 µm) resulted in a smaller maximum delamination height compared to a smaller radius (20 µm). The effective die area for a 50 µm fillet was 92.3% of the total die area, compared to 87.8% for a 20 µm fillet. The simulations also indicated that interconnect routing should be avoided in the immediate vicinity of die corners to minimize stress concentration in the substrate. Contact pads on dice should be placed with sufficient clearance from the die rim to prevent delamination during conformation to a curved surface. The microfabrication process successfully demonstrated the transfer, alignment, and interconnection of multiple dice onto the flexible substrate, enabling the creation of functional prototype implants. SEM images confirmed the integrity of the interconnections and the absence of voids at the PI-die interface.

The findings provide crucial design rules for creating highly conformable chip-in-foil neural implants. The optimized die geometry (with edge fillets) and interconnect routing strategies minimize mechanical stress and improve implant longevity. The successful microfabrication process demonstrates the feasibility of producing these complex devices at scale. The use of polyimide as a substrate material contributes significantly to biocompatibility and the ability to conform to curved brain surfaces. These results address key challenges in developing long-lasting, high-resolution neural interfaces capable of covering large brain areas. The modularity of the design enables adaptation to different applications and anatomical requirements.

This study successfully developed design rules and a microfabrication process for creating highly conformable chip-in-foil neural implants. Optimized die geometry, interconnect routing, and contact pad placement improve implant stability and longevity. The modular design enables adaptation to various applications. Future research could explore the in vivo performance of these implants and further optimize the design for improved biointegration and functionality.

The FEM simulations simplified the implant model (e.g., neglecting thin-film interconnects and substrate fenestrations). In vivo studies are needed to validate the long-term performance and biocompatibility of the implants. The current study focused on rat brain curvature; further optimization might be needed for different species or brain regions.