A new type of energy- and data-efficient nano-electronic system that integrates input/output, computation and data storage capabilities in a single three-dimensional chip is reported in Nature this week. The system is compatible with existing silicon-based circuitry, and the work could support efforts to overcome a major bottleneck in computing: the need to transfer data between off-chip memory and on-chip logic circuits. To demonstrate their device, the authors show that the system can sense different gases and vapours (including pure nitrogen and the vapours of lemon juice, vodka and beer), store the sensor data in more than one million RRAM cells and classify them using carbon-nanotube-based logic.
Computers have grown consistently more powerful over the past few decades, but a key challenge, which hinders further advance in performance, is that current technologies require large amounts of data to be processed and at ever-faster rates.
Max Shulaker and colleagues describe a prototype of a three-dimensional nano-electronic system comprising several emerging technologies. The system combines carbon-nanotube sensors and transistors with an emerging memory technology called resistive random-access memory (RRAM), and conventional silicon circuitry. It can thus capture massive amounts of data from the outside world, store and process the captured data and output highly processed information, all on the same chip.
Writing in an accompanying News & Views, Sherief Reda notes that the work could eventually aid the development of applications such as “embedded smart cameras that have high-performance artificial-intelligence capabilities, intelligent robots that swim and deliver drugs through the bloodstream, and artificial retinas.”
Technology: Slim display could enable holographic videos on mobile devicesNature Communications
Planetary science: Jupiter’s moon Europa may glow in the darkNature Astronomy
Materials: Making strong bio-based replacements for plasticsNature Communications
Biotechnology: ‘Porcupine’ system tags objects with DNANature Communications