Device to Boost AI and Machine Learning Algorithms Sustainably
Even if a gadget were to last forever, the reality is that technology evolves so quickly that many devices soon find themselves outdated and discarded. This constant churn contributes to overflowing landfills and to tackle this growing problem, scientists from Khalifa University and other institutions have developed a sustainable memristor – a device made from organic materials synthesized via green synthetic approach, which acts like a memory chip that can remember electrical signals, enhancing AI applications in technology ranging from self-driving cars to smartphone cameras.
The research was published in a titled ‘Energy Efficient Memristor Based on Green-Synthesized 2D Carbonyl-Decorated Organic Polymer and Application in Image Denoising and Edge Detection: Toward Sustainable AI’ in Advanced Science, a top 1% journal. The Khalifa University efforts are led by Dr. Dinesh Shetty, Department of Chemistry and theme leader in the Center for Catalysis and Separations (CeCaS), and team includes Dr. Abdul Khayum Mohammed, Ruba Al-Ajeil, Department of Chemistry, Dr. Ammar Nayfeh, Dr. Ayman Rezk, Department of Electrical Engineering. The work was achieved in collaboration with the external research team led by Dr. Nazek El-Atab, King Abdullah University of Science and Technology (KAUST) as well as Dr. Pratibha Pal, and Dr. Hanrui Li, Dr. Georgian Melinte.
According to the , around 53 million metric tons of e-waste is produced globally every year, with most of it going unprocessed, with traditional memory chips utilizing nonrecyclable or even toxic materials. The memristor developed by the team is constructed from a biocompatible polymer, which is chemically stable and environmentally friendly. It can regulate and remember the flow of electricity, making them incredibly efficient and eco-friendly. Moreover, the device successfully operated over 1,000 cycles, repeatedly switching between different electrical states without losing performance. Such high level of endurance, reliability and stability, without significant degradation over time make the device a promising option for long-term use in technology.
“We have developed a sustainable memory chip from organic materials to help improve AI applications in technologies ranging from self-driving cars to smartphone cameras.”
— Dr. Dinesh Shetty, Associate Professor, Khalifa University
Dr. Dinesh Shetty said: “The good electronic properties, environmentally friendly nature, and ingenious structure render the carbonyl-rich, two-dimensional organic polymer a suitable candidate for neuromorphic computing. Such characteristics of the device confirm its ability to mimic synaptic functions of the biological brain. More importantly, the combination of a green synthesized and biocompatible switching layer with electrodes that can be potentially fully recycled, in addition to the high energy efficiency, can contribute to sustainable applications.”
Alisha Roy
Science Writer
18 October 2024
