Building a Single-molecule Transistor from Scratch

Alexander Hellemans

An international team of researchers has demonstrated for the first time that a single molecule can operate as a field-effect transistor when surrounded by charged atoms that operate as the gate. The team published its results in the August 2015 issue of the journal Nature Physics.

The experiments were performed in Berlin at the Paul-Drude-Institut für Festkörperelektronik (PDI), in collaboration with researchers at the Free University of Berlin (FUB), the NTT Basic Research Laboratories (NTT-BRL) in Japan, and the U.S. Naval Research Laboratory (NRL) in Washington, D.C.

The researchers used a technique first demonstrated by researchers at IBM in 1990 when they created the letters I, B, and M by moving single atoms around on a metal surface with a scanning tunneling microscope (STM). In order for the molecule to function as a transistor, the researchers had to deposit it—as well as the charged indium atoms that surround it, forming the gate—on a semiconductor surface (in this case, indium arsenide) instead of a metal.

Doing so was expected to be more difficult, because molecules on semiconductor surfaces usually attach themselves by covalent bonds, which are very strong and make it difficult to move the atoms around with the STM tip. But the molecule they used, a dye called copper phthalocyanine, is attached to the semiconductor surface by van der Waals forces, explains Stefan Fölsch at PDI who led the research. These forces are much weaker, and allow the molecule to be moved around easily. The weak bond also allows a current to flow from the tip of the STM through the molecule. “You have sequential tunneling between the tip, the molecule and the surface,” says Fölsch. In this way, the molecule functions as the channel, with the tip and the semiconductor substrate behaving as the ‘source’ and ‘drain’ electrodes.

In a normal transistor, the current through the channel is controlled by modulating the gate voltage. Clearly, this is not possible with this setup, where the gating atoms have fixed charges. Instead, it’s possible to mimic the modulation of the electric field by varying the distance between the channel and the gate.

“We created a certain electrostatic potential ‘landscape’ on the surface by placing [charged] atoms in a certain geometry through which we are moving the molecule on a fixed line,” says Fölsch. “In each new position, the molecule feels a different electrostatic potential created by these atomic-scale gates.”

The gating of the channel is different from what we are used to with conventional transistors; the mechanism of how the intensity of electrostatic field controls the current through the channels is just as unfamiliar, in that it changes the quantum state of the molecule. The gate controls the charge state in the molecule, which in turn controls the ability of electrons that tunnel between the gate and the STM tip to hop via close molecular orbitals in the molecule.

The results of the experiments are still far from finding applications in real-world devices, not only because of the complexity of the experiments, but also because much of the physics involved is still not fully understood. “These are very basic experiments where we have very ‘ideal’ systems, and it is important  to gain a detailed understanding of what is going on,” says Fölsch, who adds that without theoretical work performed at the NRL and FUB, the experiments wouldn’t have gone far. Still, the path from their work, performed with a high-precision STM at a temperature of 4 degrees Kelvin under ultra high-vacuum conditions, to a practical device will involve a lot of engineering, admits Fölsch. “It is a long way to go,” he says.

Leave a Reply