A team of the researchers from Sungkyunkwan University and Hanyang University in South Korea, together with their colleagues from the Italian Institute of Technology and Shanghai University presented a comprehensive analysis of the prospects of the so-called two-dimensional electronics. We are talking about materials with the thickness of a single atom that can replace silicon in microchips. For example, molybdenum disulfide and tungsten disulfide are semiconductor compounds based on transition metals and sulfur that exhibit stable electronic properties at the nanoscale. Transition to such materials can extend Moore’s Law, that is to say, maintain the growth of computing power even after the capabilities of silicon have been exhausted. The main advantage of these materials is that transistors based on them consume less power, generate less heat, and can be arranged not only in a plane but also in multiple layers, one above another.
Currently, silicon electronics faces a number of limitations: it is hardly possible to make transistors smaller, without losing efficiency. This leads to leakage current and other effects degrading a chip performance. Unlike silicon, two-dimensional materials are devoid of a number of these problems – their electrical properties are more stable even at very small sizes. However, the transition to this technology requires addressing several challenges: first, learn how to grow large, uniform films of such materials directly on silicon wafers without overheating the materials; second, ensure high-quality contact with the metals to allow the current flow without unnecessary loss; third, create ultra-thin insulating layers that will control transistor operation.
The researchers have analyzed various production methods for such materials and drew a conclusion that metal-organic chemical vapor deposition is the most promising. It allows for the formation of high-quality films at relatively low temperatures. A molybdenum disulfide monolayer has already been grown on a large substrate at a temperature of approximately 350°C, which brings technology closer to industrial application. To reduce contact losses, the scientists propose to use semimetals such as bismuth and forming contacts at the edge of the atomic layer, which significantly reduces transistor resistance and improves their efficiency.
The study places particular emphasis on thermal effects. In 3D microchips, the upper layers heat up more because they are located further from the cooling substrate. However, calculations show that due to their structure, 2D materials are able to dissipate heat more efficiently. As a result, such chips could be not only more productive but also more energy efficient, reducing the requirements for the cooling systems.
In addition to the transistors themselves, the researchers also considered the issue of interconnects – the metal lines connecting billions of elements on the chip. As their size decreases, traditional copper conductors with protective layers begin to take up too much space and increase resistance. 2D materials can act as ultra-thin barriers without significant reduction of the conductive cross-section. This allows for simultaneous improvements in the conductivity and reliability of microchips.
An important result of the study was demonstration of the working vertical structures where one transistor is positioned directly above another. These devices were created with high precision and stability, demonstrating the technology’s potential for practical application. At the same time, the scientists note that significant challenges still remain, such as ensuring the reliability of multilayer structures and power management at different levels. To address them, they propose to use atomic layer etching for precise element formation, as well as new architectures for power delivery from the back of the chip, freeing up space for active layers.
Although 2D computing systems remain a future goal, it is already clear that the transition to such materials opens the way for further development of microelectronics beyond the capabilities of silicon.
