New method gets better performance than atomically thin transistors – Ars Technica

Atomically thin materials such as graphene are single molecules in which all chemical bonds are oriented so that the resulting molecule forms a sheet. These often have distinct electronic properties that can enable the production of electronics with incredibly small features only two atoms thick. And there have been a number of examples of functional devices being built from these two-dimensional materials.

But nearly all of the examples so far have used detailed construction, sometimes involving researchers manipulating individual sheets of material by hand. So we’re not at the point where we can make complex electronics out of these materials. But a research paper released today describes a method for producing wafer-scale transistors based on two-dimensional materials. The resulting transistors perform more consistently than those made using more traditional manufacturing methods.

Better manufacturing

Most efforts to facilitate the production of electronics based on atomically thin materials have involved incorporating these materials into conventional semiconductor manufacturing techniques. This makes sense because these technologies allow us to perform incredibly precise processing of materials in large quantities. Usually, this means that much of the metal wire needed for electronics is put into place by conventional manufacturing. The 2D material is then placed on top of the metal, and additional processing is performed to form functional transistors.

“Additional processing” often involves placing a metal layer over the 2D material. Researchers are of the opinion that this is probably not the best way to do things. Metal deposition can damage the 2D material, and some of the individual metal atoms can diffuse into the 2D material, creating small short circuits within the larger feature. All of this degrades the performance of any circuit built using this technology.

So the team came up with a way to machine all the individual parts of the circuit separately and put them together under gentle conditions. The simplest part was forming the gates of the transistors, which were simply engraved onto a solid substrate and then anodized.

Separately, the team formed a uniform sheet of atomically thin material (molybdenum disulfide) on top of a silicon dioxide surface by chemical vapor deposition. That plate was then lifted and transferred over the aluminum oxide, creating an atomically thin layer of semiconductor over the gate. To make a transistor, the researchers were only missing the source and drain electrodes.

Manufactured completely discretely by shaping each wire onto a solid surface. The wires were then embedded in a polymer, and everything was peeled off the surface, creating a slab of polymer with the wires embedded on its bottom surface. This polymer on its own is flexible enough that it can stretch or deform, and so the wires will not line up with gates, as is required to form functional circuits. To reduce these distortions, the researchers bonded the polymer to a sheet of quartz before stamping it onto the wafer covered with gate electrodes. This led to the deposition of wires directly on top of molybdenum disulfide, completing the formation of functional transistors.

Once everything is in place, the polymer can be removed under mild conditions, and any excess material can be cut off using plasma etching. The result was an array of transistors where the semiconductor connection to the source and drain electrodes was formed simply from material placed physically next to each other. This limits the potential for damage to the atomically thin semiconductor material.

better performance

While all the processing required here is much gentler than typical semiconductor manufacturing, this manufacturing simplifies things by configuring all the features that are needed in the end. For this approach to work, the source and drain electrodes are fabricated separately from the gate and must be dropped into place afterwards. For circuits with small features, this requires incredibly precise alignment.

This… didn’t always work out. There have been a number of cases where an entire set of electrodes have ended up out of alignment, usually due to a slight misalignment as they were dropped into place. This is something that could be improved, but it will likely still be a challenge.

The good news is that when it worked, it worked really well; The performance of the devices has been much more consistent than those produced using more typical techniques. And by most measures, they did much better. The voltage at entry and exit differed by nine orders of magnitude. Out-of-state dropout was also very low.

Overall, the approach worked. The researchers were able to build functional circuits across an entire 2-inch wafer, including half-snake modules, an essential component of computational hardware. So while this is obviously still in the demo stage, the demo is more about what hardware it can be used for.

This is not to say that molybdenum disulfide is on the fast track to replace silicon. Decades of experience have made it possible to do some incredibly complex things with silicon circuits. But it does mean that people are beginning to develop toolkits that may one day make 2D materials a viable competitor to silicon.