Researchers at TU Braunschweig successfully test micro-LED technology for the artificial intelligence of tomorrow
Known for their energy efficiency, light-emitting diodes (LEDs) also open up completely new possibilities for applications beyond lighting: with a neuron network of microscopic LEDs for the artificial intelligence (AI) of tomorrow, a research group at the Nitride Technology Center (NTC) at the Technical University of Braunschweig wants to make computers of the future more powerful and energy-efficient.
Miniaturization, scalability and energy efficiency are crucial for the development of more powerful hardware for AI applications. The NTC research group at TU Braunschweig is using micro-LED technology to take a completely new approach to building computers. The researchers are miniaturizing and scaling the energy-efficient micro-LEDs in a way that creates a neuromorphic computer. The joint team from TU Braunschweig, Ostfalia University and ams OSRAM explains in a study in the Journal of Physics Photonics how such a computer can take AI applications to a higher level.
“Our optical neuromorphic computing replicates the functioning of biological neural networks such as 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 Center. “This avoids the weaknesses of conventional digital computer technology, which lead to immense energy requirements in the massively parallel information processing for AI applications,” adds Professor Christian Werner from Ostfalia University. It is expected that in 10 years’ time, around a third of the electrical energy generated worldwide will be used for supercomputers and their cooling.
The scientists are focusing on gallium nitride (GaN) for the use of micro-LED technology. On the one hand, this semiconductor is becoming increasingly popular in power electronics as it offers a higher power density and better efficiency than traditional silicon semiconductors. On the other hand, unlike silicon, GaN is optically active and therefore the basic building block of blue LEDs. The Nitride Technology Center (NTC) at TU Braunschweig is driving forward 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 grids with hundreds of thousands of micro-LEDs, which are also used in the QuantumFrontiers Cluster of Excellence and the cluster4future Quantum Valley Lower Saxony (QVLS). “The special properties of gallium nitride are ideal for micro-LEDs with dimensions of one micrometer and smaller,” says Waag.
The research group also sees great potential in GaN-based micro-LED technology to reduce the power consumption caused by the enormous “energy hunger” of AI systems by a factor of up to 10,000. The micro-LEDs take on the task that would otherwise be performed by silicon transistors. Parallel in-memory processing combined with efficient photon production and detection creates hardware that physically maps the various levels of neural networks and enables the parallel flow of information.
However, intensive research work is still required before an “artificial brain” based on this new technology can be realized, but it also promises enormous energy savings. However, the research group at the NTC has already succeeded in developing a macroscopic optical micro-LED demonstrator with 1,000 neurons. It has also already passed a standard AI pattern recognition test: It identifies messily written digits from zero to nine, some of which are difficult even for a human to decipher.
CONTACT
Prof. Dr. Andreas Waag
Nitride Technology Center (Sprecher)
Technische Universität Braunschweig
Institut für Halbleitertechnik
Hans-Sommer-Straße 66
38106 Braunschweig, Germany
Tel.: +49 531 391-3774
Email: a.waag@tu-braunschweig.de
www.tu-braunschweig.de/iht
www.tu-braunschweig.de/iht/forschung/nitride-technology-center
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