Self-Assembly Trick Makes Transistors and Diodes

Using liquid metal, scientists have devised a new way to make electronics that assemble themselves. With prototypes including nanoscale to microscale transistors and diodes, the researchers suggest their research might help greatly simplify electronics production.

Existing chip manufacturing techniques require many steps and depend on extremely complex technologies, which makes fabrication costly and time-consuming, says Martin Thuo, a professor of material science and engineering at North Carolina State University.

As such, for decades, scientists have sought to develop self-assembling electronics. “Self-assembly is the default approach in nature—the brain is self-assembled,” Thuo says. By avoiding the use of advanced manufacturing tools self-assembly “lowers the capital investment and the level of trained manpower we need in manufacturing.” And multi-step processes, such as the creation of a field-effect transistor could be done in a single step, he says.

The Long Road to Self-Assembly

Previous research explored many avenues toward self-assembly, such as trying to build computers from molecules or using DNA or other compounds to assemble components.

Two key problems these approaches have faced are keeping contaminants infiltrating the final products and building components at multiple scales, Thuo says. To overcome these challenges, Thuo and his colleagues built an interdisciplinary team. The resulting expertise in chemistry, materials science, fluid dynamics, and electrical engineering “is what allowed us to stand out,” he says.

The technique they developed starts with particles of liquid metal, such as Field’s metal, an alloy of indium, bismuth, and tin that is liquid at a balmy 62 degrees C. These roughly 2-micrometer-wide particles are placed on one side of a silicone rubber mold, which the researchers can make in any pattern or size.

The scientists next poured on a vinegary solution, which harvested metal ions from the particles’ surfaces. Inside the mold, ion-bearing molecules in the solution , called ligands, assemble themselves into 3-D structures, while the liquid part of the solution starts to evaporate, packing the 3-D structures closer and closer together. The mold helps these structures form predictable, symmetric arrays as the liquid dries.

The researchers then remove the mold and heat the arrays at 600 degrees C, breaking up the ligand molecules and freeing up carbon and oxygen atoms. The oxygen interacts with the metal ions to form semiconductor metal oxides, while the carbon atoms form graphene sheets.

Using their new technique, the scientists fabricated wires 44 nanometers to 1 micron wide, and transistors and diodes nanoscale to microscale in size. (Transistors were made by using the silicon substrate as the device’s gate and attaching gold electrodes to a semiconductor wire. Diodes were done by exploiting a natural asymmetric conductance in a wire.) Ultimately, they can generate patterns millimeters to centimeters across, “so scalability is not a challenge,” Thuo says.

The researchers can control the properties of the semiconductor structures by manipulating the kind of liquid used in the solution, the dimensions of the mold, and the solution’s rate of evaporation. In addition, the bismuth in the liquid metal particles makes the resulting arrays responsive to light, meaning this new technique can help create optoelectronic devices.

Uses of Self-Assembly

“Self-assembled electronics have been a long-term dream, as they promise to simplify fabrication and associated costs to meet the ever-increasing demand for more complex electronics,” says Chris Nijhuis, a professor of hybrid materials for optoelectronics at the University of Twente in the Netherlands, who did not take part in this research. “To see now such concepts being used to self-assemble electronic and optically active devices starting from super-cooled liquid metals is astounding.”

The first applications for this new technique may be microelectromechanical systems (MEMS) and related sensors, Thuo says. “We hope to then move this method to manufacture some predicted but yet to be commercialized transistor architectures such as BiSFETs [bilayer pseudospin field-effect transistor], where interfaces and 2D materials play a major role... it would be very interesting to see whether the concepts can be further developed in adaptive circuits, or multifunctional or even 3-D electronics.”

“In addition, this new strategy to create wires with such exquisite control may be important in applications where it is difficult to form interconnects otherwise,” says Nijhuis. He cautions that the high temperatures needed to create structures with this new technique may limit potential applications, but “there is room for improvement.”

The scientists are now developing a startup to move their work forward. In addition, Thuo notes that he and his colleagues are part of a National Science Foundation Innovation Corps program to help them reach out to industry. “Self-assembly processes are readily adaptable, but they need to be aligned with a specific need,” Thuo says. “That is why we are talking to semiconductor companies.”

Thuo and his colleagues detailed their findings online 25 November in the journal Materials Horizons.