Atomically thin materials like graphene are single molecules with all chemical bonds oriented and the resulting molecules form sheets. They have unique electronic properties that enable the fabrication of electronics with incredibly small features, often just a few atoms thick. And there are many examples of functional hardware built from these two-dimensional materials.
But almost all of the examples so far have used custom-made structures, and sometimes researchers have manually manipulated individual flakes of the material. It is not at the production stage. However, a paper published today describes a method for wafer-scale fabrication of transistors based on two-dimensional materials. And the resulting transistors have more consistent performance than those made with conventional manufacturing methods.
better manufacturing
Most of the efforts made to facilitate the production of electronics based on atomically thin materials involve integrating these materials into conventional semiconductor manufacturing techniques. This makes sense because these techniques allow for very granular manipulation of large amounts of materials. Typically, this means that much of the metal wiring required for electronics is laid out using traditional manufacturing methods. The 2D material is then layered on top of the metal and subjected to additional processing to form a functioning transistor.
Often that “additional processing” involves layering metal on top of the 2D material. This method is probably not the best one, claims the researchers behind the study. Depositing metal can damage the 2D material and some individual metal atoms can diffuse into his 2D material causing small short circuits within large features. All of which degrade the performance of circuits built using this technique.
So the team figured out how to form all the individual parts of the circuit separately and put them together under mild conditions. The simplest part was forming the gates of the transistors. It was simply patterned on a solid substrate and coated with aluminum oxide.
Separately, the team formed a uniform sheet of atomically thin material (molybdenum disulfide) on a silicon dioxide surface by chemical vapor deposition. The sheet was then lifted and transferred onto aluminum oxide, placing an atomically thin layer of semiconductor over the gate. To form a transistor, researchers only lacked source and drain electrodes.
They were made completely separately by forming all the wiring on a solid surface. The wires were then embedded in the polymer and the whole peeled away from the surface to create a sheet of polymer with the wires embedded on the underside. This polymer is flexible enough on its own that the wiring does not match the gates as is required to form a functional circuit. To limit these distortions, the researchers linked the polymer to a sheet of quartz before pressing it onto the gate electrode-covered wafer. This deposited the wiring directly on top of the molybdenum disulfide, completing the formation of the functional transistor.
Once everything was in place, we were able to remove the polymer under mild conditions and use plasma etching to cut away the excess material. You now have a collection of transistors simply formed by the materials placed in the . This limits the possibility of damaging atomically thin semiconductor materials.
better performance
All the processing required here is much gentler than normal semiconductor manufacturing, but that manufacturing simplifies the problem by forming all the functions that are ultimately required. To make it work, the source and drain electrodes must be made separately from the gate and dropped into place afterwards. Circuits with small features require incredibly precise alignment.
It… didn’t always work. There have been many cases where the entire electrode collection has been misaligned due to a slight twist when the electrodes are dropped into place. This could be improved, but may remain a challenge.
The good news is that when it worked it worked very well. Devices performed much more consistently than those manufactured using more conventional techniques. And, in most respects, their performance improved significantly. The on-state and off-state voltages differed by nine orders of magnitude. Leakage in the off state was also very low.
More generally, this approach worked. Researchers were able to build a functional circuit across a 2-inch wafer, including the half-adder unit, an integral component of the computational hardware. So this is clearly still in the demonstration stage, but the demonstration is potentially usable hardware.
This does not mean that molybdenum disulfide is about to replace silicon. Decades of experience have made incredibly sophisticated things possible with silicon circuits. However, this means that one day he is beginning to develop a toolkit that will make 2D materials a viable competitor to silicon.
nature nanotechnology, 2023 Doi: 10.1038/s41565-023-01342-1