Scientists Fuse Synthetic DNA With a Semiconductor to Build a 100x More Efficient Memory Chip

Started by FinnBalor99, Today at 11:13 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

Topic: Scientists Fuse Synthetic DNA With a Semiconductor to Build a 100x More Efficient Memory Chip   Views(Read 23 times)
Active members in this topic:
FinnBalor99(1)

FinnBalor99

Penn State researchers have combined synthetic DNA with a crystalline perovskite semiconductor to create a memristor, a memory device that stores and processes information in the same physical location much like a neuron does. The resulting device operates at ultra low voltages under 0.1 volts and reportedly uses roughly 100 times less power than equivalent traditional memory storage while offering higher storage density.

DNA's appeal for data storage has been obvious for a while now on paper, since a single gram of it can theoretically hold about 215 million gigabytes of information, an almost absurd density compared to any conventional storage medium. The challenge has always been getting that biological material to actually function reliably alongside inorganic electronic components rather than existing as a purely theoretical curiosity confined to a chemistry lab.

The team's approach doped short, custom synthesized DNA sequences with silver nanoparticles, which let the DNA conduct electricity while also orienting its structure more precisely at the nanoscale. That doped DNA was then integrated with thin films of the perovskite semiconductor, the same broad family of material already used in solar cells and certain lasers, to create a working bio hybrid electronic system rather than two separate materials merely sitting next to each other.

The device reportedly remained stable across a wide temperature range and continued operating reliably for six straight weeks at room temperature, which is a meaningfully longer stability window than a lot of earlier experimental bio hybrid electronics have managed to demonstrate. Researchers see potential applications in energy hungry fields like AI and neuromorphic computing specifically, where systems designed to store and process information simultaneously could meaningfully cut the enormous power draw of current data centers.

Still very early stage lab work with a patent application filed rather than an actual product anywhere near shipping, but it is a genuinely creative direction for tackling the AI industry's massive and growing energy appetite

Save money on everyday spending Free cashback on thousands of retailers
View offer