Penn State researchers build DNA-perovskite memristor using 100 times less power
This digest was compiled by AI from multiple sources — links to the originals are below.
Penn State researchers build a DNA-perovskite memristor that uses 100 times less power. The work appears in Advanced Functional Materials. The device stores and processes information in the same location after power is removed, mimicking neuron function and offering a route to low-power neuromorphic computing for AI.
Key Facts
- The DNA-based memristor uses 100 times less power than conventional memory devices.
- A single gram of DNA can hold about 215 million gigabytes of data.
- The memristor combines synthetic DNA with crystalline perovskite, a semiconductor used in solar cells and lasers.
- The device stores and processes information in the same location and retains previous electrical activity after power is removed.
- The work was published in Advanced Functional Materials and is the subject of a patent application.
The Bio-Hybrid Memristor
Penn State researchers built a memory resistor, or memristor, from synthetic DNA and crystalline perovskite. Synthetic DNA was made from commercially available, chemically engineered molecules arranged into short genetic sequences tailored for electronic requirements. Crystalline perovskite is already used in solar cells, lasers and data storage devices. The team published the work in Advanced Functional Materials and filed a patent application.
Low-Power Memory Operation
The device uses 100 times less power than conventional memory devices. Unlike ordinary resistors, memristors preserve a record of previous electrical activity after power is removed. The device stores and processes information in the same location, resembling how neurons function in the brain. Kavya S. Keremane, co-corresponding author, said the bio-hybrid system fundamentally changes how low-power memory devices can be designed.
AI and Neuromorphic Computing
Bed Poudel, co-corresponding author, said AI and other emerging technologies will increasingly depend on neuromorphic computing. Neuromorphic computing is designed to operate more like the human brain. A single gram of DNA can hold about 215 million gigabytes of data. The team said the bio-hybrid approach addresses the demand for low-power, high-storage devices driven by AI.
1 source
Penn State researchers build DNA-perovskite memristor using 100 times less power
