Micro-LEDs for neuromorphic computing Researchers successfully test micro-LED technology for tomorrow's artificial intelligence
Known for their energy efficiency, light-emitting diodes (LEDs) are also opening up completely new possibilities for applications beyond lighting: by using a neuron network of microscopic LEDs for the artificial intelligence (AI) of tomorrow, a research group at Nitride Technology Centre (NTC) at TU Braunschweig aims to make future computers more powerful and energy-efficient.
Professor Andreas Waag (TU Braunschweig) and Professor Christian Werner (Ostfalia) at the demonstrator for an LED-based neuromorphic computer. Photo credit: Laurenz Kötter/TU Braunschweig
Miniaturisation, scalability and energy efficiency are crucial for the development of more powerful hardware for AI applications. The NTC research group at TU Braunschweig is taking a completely new approach to building computers using micro-LED technology. The researchers are miniaturising and scaling the energy-efficient micro-LEDs in a way that makes a neuromorphic computer possible. In a study published in the Journal of Physics Photonics, the team from TU Braunschweig, Ostfalia University of Applied Sciences and ams OSRAM explain how such a computer could take AI applications to a higher level.
“Our optical neuromorphic computing mimics the functioning of biological neural networks, such as those in the human brain, by using electronic circuits or photonic components,” says Professor Andreas Waag from the Institute of Semiconductor Technology at TU Braunschweig and also spokesperson for the Nitride Technology Centre. “This avoids the weaknesses of conventional digital computer technology, which lead to immense energy demands in massively parallel information processing for AI applications,” adds Professor Christian Werner from Ostfalia University of Applied Sciences. It is expected that in 10 years’ time around a third of the world’s electrical energy will be used for supercomputers and their cooling.
Gallium nitride (GaN) is the semiconductor of choice for micro-LED technology. This semiconductor is increasingly used in power electronics because it offers higher power density and better efficiency than traditional silicon semiconductors. However, unlike silicon, GaN is optically active, making it the basic building block for blue LEDs. The Nitride Technology Centre (NTC) at TU Braunschweig is driving the development of nitride semiconductor technology as the second pillar of microelectronics.
The researchers are combining GaN components with conventional silicon microelectronics to open up completely new fields of application – such as highly integrated arrays with hundreds of thousands of micro-LEDs, which are also being used in the QuantumFrontiers cluster of excellence and in the Quantum Valley Lower Saxony (QVLS). “The special properties of gallium nitride are ideal for micro-LEDs with dimensions of one micrometre and smaller,” says Waag.
Technology reduces energy consumption
The research group also sees great potential in GaN-based micro-LED technology for reducing the power consumption caused by AI systems with their enormous “hunger for energy” by a factor of 10,000. The micro-LEDs perform the task that would otherwise be performed by silicon transistors. Parallel in-memory processing combined with efficient photon production and detection creates a hardware that physically maps the different levels of neural networks and enables parallel information flow.
Much research is needed before an “artificial brain” based on this technology can become a reality, but it promises enormous energy savings. The NTC research group has already developed a macroscopic optical micro-LED demonstrator with 1,000 neurons. The demonstrator has already passed a standard AI pattern recognition test: it identifies numbers from zero to nine written in a jumbled fashion, some of which are difficult for a human to decipher.
Original publication
R Kraneis et al: MicroLEDs for optical neuromorphic computing—application potential and present challenges, 2024 J. Phys. Photonics 6 04LT01 DOI: 10.1088/2515-7647/ad8615
