High Frequency Carbon Nanotube/Graphene Transistor




Introduction: High Frequency Carbon Nanotube/Graphene Transistor

Transistors made from nanocarbon materials have the potential to revolutionize electronics. Carbon nanotubes are perfect semiconducting or metallic 1-dimensional wires and graphene sheets are showing record breaking electron mobility.

The challenge in building high performance electronic devices from carbon nanotubes or graphene lies in integrating them into circuits without destroying their unique properties. In particular, when creating nanocarbon transistors for wireless applications making an efficient gate electrode has proven difficult.

One way to improve performance is to create freestanding 3-dimensional gate structures. Unfortunately, this is quite difficult using conventional photolithography or electron beam lithography. This instructable will show a fast and easy way how to create a suspended gate electrode for high frequency nanoelectronics.

Step 1: Preparations

What you need to get started:
- Chunk of graphite
- Scotch tape
- Silicon wafer
- Microscopes: optical, AFM, SEM
- Clean room facilities
- Focused ion beam (handy but not absolutely necessary)

Making graphene is fairly straightforward: take a piece of graphite and rub it over the silicon wafer. Then use the scotch tape to peel of layers until you end up with a single layer left on the substrate. You can check the process with an optical microscope: if the layer is only barely visible anymore you have (probably) succeeded.

Step 2: Lithography

The source/drain and the gate electrodes can be patterned in a resist layer either by optical lithography or electron beam lithography. In any case after developing the resist you should end up with structure similar to the one shown below.

Step 3: Metallization

The sample with the developed resist mask is covered with a metal layer. Typical several different metals are sputtered or evaporated in a vacuum apparatus (e.g. first a 5 nm thick titanium adhesion layer and on top a 50 nm gold layer).

Step 4: Bridging

The gate electrode is going to be formed from the metal covering the resist on top of the graphene sheet. In order to support it a thick bridge like structure needs to be created. This can be done conveniently by focused ion beam (FIB) induced metal deposition:

A organic precursor gas is decomposed by the focused Ga-ion beam resulting in local deposition of a conducting compound. The metal covered resist layer protects the sensitive graphene sheet from any gallium ion implantation or irradiation damage.

(In principle it is also possible, yet more difficult, to achieve this in a second lithography process.)

Step 5: Release

In order to create a well defined gate electrode the to-be-suspended part of the metal layer is separated from the adjacent parts. This is done using the high energy ion beam to mill cuts through the metal layer.

Step 6: Lift-off

The resist and surplus metal is removed using a suitable solvent; e.g. acetone (so called [http://en.wikipedia.org/wiki/Lift-off_(microtechnology) lift-off] process). The result is a graphene field effect transistor (FET) for radio frequency applications.

- Special precaution stepsprecaution steps might be necessary to prevent capillary forces during drying from collapsing the bridge
- Performance can be increased further when the graphene sheet itself is subsequently suspended by etching away the sacrificial layer

Step 7: Proof of Concept

The micrographs shows a suspended structure demonstrating the feasibility of the method. It is notable, that the structure withstood dicing and lift-off without any special precautions.



    • Fix It! Contest

      Fix It! Contest
    • Creative Misuse Contest

      Creative Misuse Contest
    • Metalworking Contest

      Metalworking Contest

    12 Discussions

    Hi recently I have read an article in NY Times about the advantages of graphene outweighing the silion. Now having your steps in mind, I can feel sure the prosperous future of graphene and nanotech. This instructable shows how to create a suspended gate electrode in an interesting way and indictates a future of high-tech. It is attractive, so I shall start following your steps right away.

    good instructable. dont see how you put the parts there if its only 100nm wide. Relating to Killerjackalope's comment. there is a real problem because they are making things so small that the Quantum effect starts happening. they make a wire 4 atoms thick, and the probability that they al move in the same direction at the same time it vastly more than a table say. so in your chip a wire could just randomly move.

    1 reply

    I'll probably make an instructable on electron beam lithography. And I will add more micrographs as I go along with the project. So hopefully you can later see how to build things that small.

    This is very interesting, but I have read books and articles in the past that deal with how IC chips are made. I have also read about unijunction transistors. From the photograph in the last step I see that the scale is graduated in micro-microns. These are obviously very small. When you say these will revolutionize electronics, I would assume that means more transistors on CPUs resulting in faster speeds and more computing power, as well as all sorts of other types of miniaturization of devices we have not yet even conceived. Thanks for posting this. Obviously not many of us are going to run home and make one.

    4 replies

    Strangely we're apparently near the end of what we can do by making them smaller, I was reading up on this stuff and apparently some processors are now getting to the point where light speed isn't enough, the pulse still take time to cross the board and they can be ready to fire more before it gets across, which is insane... The next thing is apparently pointing towards better processing structures etc. Now I'm not an expert but that's what I've taken from some recent stuff, obviously if we make everything smaller it solves the problem to some extent so in that case... Nice 'ible lorenz, atypical but interesting and informative, by the way if I'm off with what I just said anybody feel free to dispute, it's not my area of expertise but it's an issue that seems to exist... *realizes that 3.21GHZ is 321000000hz, that many pulses a second, and my computer's not the fastest in the world, single core... The issue actually seems obvious in that respect...

    You are right on. In one clock cycle of your computer light can travel only roughly 10 cm (4 inches). But that is only one of many problems...


    From what I see there's a huge amount of research going in to better structure and program development, if I could find the article again, it was really interesting how they were going about making the processor work faster without brute force clock speeds... I suppose if transistors were halfed in size it could potentially allow for up to 6GHZ of amazingly well laid out processing but it'd have to be smaller and of strange shaped, probably be an absolute nightmare to cool... After that we'd be a bit stuck however if you had that much clock speed to play with I'd say the computer would be acceptable for some time to come..

    Thanks for commenting, Phil. I just started last week with this idea and the images don't show a real device yet. It is merely a 5 micron test structure to see whether the principle works. The real thing will be roughly 50 times smaller, around 100 nanometers. This is still large compared to 45 nm feature size in commercial microprocessors but we doing fundamental research.
    The performance improvement comes from the new material properties (e.g. ultrahigh electron mobility). The scaling down is then later a task for industrial development teams.

    Not your typical instructable, but very informative. (But hey, if you figure out how to make one in your garage from liquid dish soap, scrap soda bottles, and a single LED, be sure to let us know. ;) )

    1 reply

    I'm not exactly sure what you just said, but 5 stars just for the awesomeness of what ever you did.