11 April 2023
KAUST integrates 2D h-BN on 180nm CMOS to create high-speed, low-energy-consumption memristors
Exploiting the excellent electronic properties of two-dimensional (2D) materials to fabricate advanced electronic circuits is a major goal for the semiconductors industry. However, despite multiple studies that have reported prototype devices with promising properties for sensing and driving electrical current, their technology readiness level is still very low. This is because the studies mostly use synthesis and processing methods that are incompatible with industry, create isolated large (>1µm2) devices on unfunctional SiO2/Si substrates, and present poor variability and yield.
Now, a team at Saudi Arabia’s King Abdullah University of Science and Technology (KAUST) led by Dr Mario Lanza, associate professor of materials science and engineering, has integrated two-dimensional materials on silicon microchips, and achieved what is said to be excellent integration density, electronic performance and yield, according to the paper ‘Hybrid 2D/CMOS microchips for memristive applications’ published in Nature (www.nature.com/articles/s41586-023-05973-1). “In the future, most microchips will exploit some of the many outstanding electronic and thermal properties of these materials,” believes Lanza.
In particular, Lanza’s team used a sheet of multi-layer hexagonal boron nitride (h-BN, about 6nm thick) as a two-dimensional insulating material, and transferred it onto the back-end-of-line (BEOL) interconnections of silicon chips bearing 180nm-node complementary metal-oxide-semiconductor (CMOS) transistors. The circuits were then completed by patterning the top electrodes and interconnections.
The resulting hybrid 2D/CMOS chips exhibit high durability and electronic properties that enable the fabrication of artificial neural networks with very low power consumption. They can successfully compute spiking neural networks (SNN) — electrical stresses applied during a very short time — a key component of existing artificial intelligence (AI) systems that are increasing in demand. Most existing devices are not suitable for implementing this type of neural network, and there is a market need to find new approaches. KAUST’s CMOS transistors provide outstanding control over the currents across the h-BN memristors, allowing endurances of ~5 million cycles in memristors as small as ~0.053µm2.
KAUST says that the research has attracted the interest of leading semiconductor companies such as Taiwan Semiconductor Manufacturing Corp (TSMC) and Advanced Semiconductor Materials Lithography (ASML) and could help other companies reduce processing costs and energy. Most companies in the field of microchip manufacturing and artificial intelligence are aiming to create new hardware that reduces data processing time and energy consumption but have not yet found a suitable device.
IBM has attempted to integrate graphene into transistors for radio-frequency applications, but the devices were not able to store or process information. In contrast, the devices created by Lanza’s team measure only 260nm and could be made much smaller very easily if more advanced microchips were available.
This work is exciting for the field of nanoelectronics and semiconductors because of the high performance of the devices and circuits produced and the potential for far-reaching industry applications.