Scientists Expand the Ability of 2D Materials to Refine Technology

Ever since the discovery of graphene in 2004, scientists have been fascinated by the elasticity, flexibility, and exceptional electronic properties of two-dimensional (2D) materials that are as thin as a single layer of atoms.

Some of these materials can be particularly vulnerable to variations in their material properties as they are stretched and pulled. When exposed to strain, they have been predicted to endure phase transitions as disparate as superconducting in one moment to non-conducting the next moment, or optically opaque in one moment to transparent in the following moment.

Now, scientists from the University of Rochester have integrated 2D materials with oxide materials in a novel way, using a transistor-scale device platform, to completely investigate the capabilities of these changeable 2D materials to change optics, electronics, computing and a multitude of other technologies.

"We're opening up a new direction of study," says Stephen Wu, assistant professor of electrical and computer engineering and physics. "There's a huge number of 2D materials with different properties — and if you stretch them, they will do all sorts of things."

The platform created in Wu's lab, constructed quite like traditional transistors, permits a small flake of a 2D material to be placed on a ferroelectric material. Voltage applied to the ferroelectric — serves like a transistor's third terminal, or gate-strains the 2D material by the piezoelectric effect, making it stretch. That, consecutively, stimulates a phase change that can fully alter the way the material acts. When the voltage is switched off the material retains its phase until an opposite polarity voltage is supplied, causing the material to get back to its primary phase.

The ultimate goal of two-dimensional straintronics is to take all of the things that you couldn't control before, like the topological, superconducting, magnetic, and optical properties of these materials, and now be able to control them, just by stretching the material on a chip.  If you do this with topological materials you could impact quantum computers, or if you do it with superconducting materials you can impact superconducting electronics.

Stephen Wu, Assistant Professor of Electrical and Computer Engineering and Physics, University of Rochester

In a research paper in Nature Nanotechnology, Wu and his students explain using a thin film of 2D molybdenum ditelluride (MoTe₂) in the device platform. When stretched and unstretched, the MoTe₂ turns from a low conductivity semiconductor material to an extremely conductive semi-metallic material and back again.

It operates just like a field effect transistor. You just have to put a voltage on that third terminal, and the MoTe₂ will stretch a little bit in one direction and become something that's conducting. Then you stretch it back in another direction, and all of a sudden you have something that has low conductivity.

Stephen Wu, Assistant Professor of Electrical and Computer Engineering and Physics, University of Rochester

The process functions at room temperature, he adds, and, extraordinarily, "requires only a small amount of strain — we're stretching the MoTe₂ by only 0.4 percent to see these changes."

Moore's law notably predicts that the amount of transistors in a dense integrated circuit doubles approximately every two years.

However, as technology gets closer to the boundaries at which traditional transistors can be scaled down in size — as the end of Moore's law is reached— the technology created in Wu's lab could have far-reaching effects in moving past these confines as the hunt for ever more powerful, faster computing endures.

Wu's platform has the prospect to carry out the same functions as a transistor with a lot less power consumption since power is not required to retain the conductivity state. Furthermore, it reduces the leakage of electrical current because of the steep slope at which the device alters conductivity with applied gate voltage. Both of these matters — high power consumption and leakage of electrical current — have hindered the performance of traditional transistors at the nanoscale.

"This is the first demonstration," Wu adds. "Now it's up to researchers to figure out how far it goes."

One benefit of Wu's system is that it is arranged quite like a traditional transistor, rendering it easier to ultimately adapt to existing electronics. However, more work is required before the platform gets to that stage. Presently the device can work only 70 to 100 times in the lab before the device fails. While the durability of other non-volatile memories, like flash, is a lot higher they also work a lot slower than the ultimate potential of the strain-based devices being designed in Wu's lab.

"Do I think it's a challenge that can be overcome? Absolutely," says Wu, who will be working on the issue with Hesam Askari, an assistant professor of mechanical engineering at Rochester, also the paper’s co-author. "It's a materials engineering problem that we can solve as we move forward in our understanding how this concept works."

They will also investigate how much strain can be applied to different 2D materials without breaking them. Establishing the decisive limit of the concept will help guide scientists to other phase-change materials as the technology progresses.

Wu, who finished his PhD in physics at the University of California, Berkeley, was a postdoctoral scholar in the Materials Science Division at Argonne National Laboratory before he entered the University of Rochester as an assistant professor in the Department of Electrical and Computer Engineering and the Department of Physics in 2017.

He began with a single undergraduate student in his lab — Arfan Sewaket '19, who was spending the summer as a Xerox Research Fellow. She assisted Wu in setting up a provisional lab, then was the first to try out the device idea and the first to show its viability.

Since then, four graduate students in Wu's lab — lead author Wenhui Hou, Ahmad Azizimanesh, Tara Pena, and Carla Watson "have done so much work" to document the device's properties and improve it, developing around 200 different types to this point, Wu says. All are listed with Sewaket as co-authors, together with Askari and Ming Liu of Xi'an Jiaotong University in China.