Future Semiconductors Could be 'Photocopied' Using Graphene

Semiconductor chips of tomorrow could one day be manufactured by using a "photocopying" process on super-thin sheets of graphene, rather than using traditional silicon, according to research being conducted at the Massachusetts Institute of Technology (MIT).
The MIT research involves graphene, which are single-atom-thin sheets of graphite that are being used to copy intricate crystalline patterns from an underlying silicon wafer onto to an adjacent top layer of an identical material.

The idea is seen as a promising way to reduce the cost of existing semiconductor wafer design and construction, while also allowing the use of more exotic materials which have better conductivity properties compared to silicon, according to the school.

Yunjo Kim, a graduate student at MIT and a co-author of the research so far, told ITPro that many similar experiments by other researchers have involved graphene over the years but it hasn't yet demonstrated industrial scalability for production.

"The idea of using graphene in electronics has always been a challenge," he said, with some researchers trying flaked graphene and the MIT team using large-scale single crystal-based graphene. At MIT, the researchers are using the material to seek new ways of improving and lowering the costs of chip production.

"Currently, most industries are fabricating silicon-based devices on silicon wafers because silicon is the cheapest semiconductor in the world," said Kim. "However, certain properties of silicon fall short compared to other semiconductors."

Existing semiconductors are made today by "growing," or layering material atom by atom onto a wafer, similar to how Lego blocks are stacked on top of each other and attached to a base piece, said Kim. The wafer is an important part of the process because it tells the incoming atoms how to arrange themselves and allows the semiconductor to be built.

The MIT experiments are placing an atomically-thin sheet of graphene on top of the wafer and then growing the semiconductor on the graphene, atop the wafer. The graphene sheet is a million times thinner than a hair strand, which is why the atoms are still able to stack on top of the wafer through the graphene sheet, he said.

The technique "raises new questions on how atomic interactions work" because the "atoms were able to arrange themselves on top of graphene without making a chemical bond with the wafer," said Kim.

And because the graphene is extremely slippery and extremely strong, the researchers are finding they can reuse it to repeat the process many times to make thin film semiconductors using a single wafer below the graphene layer, he said. That's how they are able to create the "photocopying" capabilities which let them grow the semiconductors on top of the graphene, peel them away and then grow new semiconductors on top of the graphene again.

The process is called "Two-Dimensional Material based Layer Transfer," or 2DLT.

So far, the MIT researchers have been able to demonstrate the process at the research level, but are now working to find ways of scaling it for industrial use, said Kim.

Ultimately, the team wants to try a variety of different compounds other than silicon for semiconductor production, including Gallium Arsenide, Indium Phosphide and Indium Gallium Arsenide, according to the research team. But because those materials are more expensive than silicon, the use of graphene as a copy machine to fabricate multiple thin film semiconductors using a single wafer became their road map to future development.

Many hurdles remain, said Kim, including that semiconductor companies today are very comfortable using silicon wafers, which would likely take a lot of effort to transition their well-established procedures to move to a new process even if it saved them money.

"In order to make the transition easier, we can transfer these thin films on top of a cheaper substrate such as glass wafers or even silicon wafers," he said. "By offering cheap wafers with a thin film of rare semiconductor on top of it, industries can still fabricate on these thin-film semiconductors as if they were fabricating on a silicon wafer. This would allow for a fast, easy and cost-saving transition that would make it faster to realize extremely high-performance devices in the near future."

Kim said he estimates that some semiconductor companies could use these techniques in a few years to emulate the processes at the industrial scale. The MIT team is already "well underway to work with industries using this technology."

Graphene, which was discovered in 2004, has amazing electronic properties, high mechanical strength (200 times stronger than steel), efficient heat conductivity, optical transparency and more, but its wide use in industry is still in the future, said Kim.

"While this project greatly benefits the semiconductor industry, it is also a window of opportunity for graphene to see large scale industrial application," he said.

The MIT research is being led by Assistant Professor Jeehwan Kim of the Mechanical Engineering and Materials Science and Engineering departments, and also includes graduate students Samuel Cruz, Babatunde Alawonde, Chris Heidelberger, Yi Song, and Kuan Qiao. Also involved are post-doctorate students Kyusang Lee, Shinhyun Choi and Wei Kong; as well as several others.