Chip Gallery
Group only

Origami Biomedical Implants
Team members: Matthew Loh, Manuel Monge and Abhinav Agarwal

Origami implant design is a 3D integration technique which addresses the size and cost constraints in biomedical implants. Large systems can be split into multiple chips and connected using 3D integration techniques to be folded compactly for implantation and unfolded inside the body. Electronics can be partitioned into functional blocks for mass-production and customs implants can be assembled from these relatively cheap modules.

Retinal prostheses can benefit from this approach. Instead of one large chip, multiple smaller chips can be distributed over a flexible substrate that can be folded up before implantation and then opened up to a curved shape inside the eye. Such an implant will conform to the shape of the retina to improve electrode contact and make stimulation more effective. Thus, our origami approach will enable the possibility of making an extended planar system to be folded up into a compact structure for minimally-invasive surgery, and then deployed into its operating configuration. With a system that conformed to the curvature of the eye, the location of the chips and electrodes could be optimized through the design of the origami structure.

Capacitive Proximity Communication with Distributed Alignment Sensing
Team members: Matthew Loh

Since the Origami implant is deployed in-body, the alignment between the communicating chips is poorly controlled. Furthermore, the alignment will change over time due to patient movement and tissue growth. As a result, the proximity communication system needs to periodically sense its alignment and adapt itself to maximize power efficiency for a given data rate. Alignment sensing is also useful for providing feedback on the deployment status of the implant. Due to the tight power constraints on implants in sensitive organs such as the eye, this alignment-and-adaptation operation needs to be energy efficient and computationally simple.

The proximity interconnect is formed by capacitive coupling between plates in the pad-level metal of the two chips. Since only one side of the link needs to be able to sense alignment, two different array types, sensor and target, are used. The target array contains only a transmitter and receiver, while the sensor array adds alignment sensing blocks. The sensor array embeds a distributed TDC-based chip-to-chip alignment sensor that provides direct information about link quality, enabling straightforward adaptation of the link to changing alignment conditions. The sensor and transceiver share functional blocks, saving power and area. Data rates from 10-60 Mbps are achieved over 4-12 μm of parylene-C, with efficiencies up to 0.180 pJ/bit.


Related Publications:

 © California Institute of Technology