By better understanding the Jekyll-and-Hyde nature of a semiconductor alternative—one that goes from an electrically resistant insulator to a current-conducting metal—Nebraska’s Xia Hong and colleagues may have opened a new path towards smaller, more efficient digital devices. The team reports its findings in the journal Natural communications.
The semiconductor’s ability to conduct electricity in the Goldilocks region (poorer than a metal, better than an insulator) made it the ideal choice for engineers looking to build transistors, those tiny switches on-off which encode the 1s and 0s of binary signals. . Apply a certain voltage to the control button called the gate insulator, and the semiconductor channel allows electric current to flow (1); remove it and this flow stops (0).
Millions of these nanoscopic semiconductor-based transistors now cover modern microchips, turning on and off to collectively process or store data. But as tiny as transistors are, consumer and competitive demands continue to push electrical engineers to shrink them even further, either in an effort to pack in more features or to reduce the size of the devices that house them.
Unfortunately, these engineers now come up against the practical, even fundamental, limits of semiconductor size.
Researchers, in turn, have begun to look beyond not just industry-favorite silicon, but also semiconductors as a whole. More than twenty years ago, some began using a class of materials called Mott insulation. If the semiconductor is the happy medium that has led to decades of steady progress, the Mott insulator is more of a two-faced joker whose ambivalence is the source of both its appeal and its frustration .
A long-standing theory of conductivity says that a material with the electronic characteristics of a Mott insulator should generally be classified as a metal. Unlike the electrons of a metal or semiconductor, those of a Mott insulator do not behave as independent particles.
Instead, they interact in a way that confines them to localized sites and prevents them from moving freely through a material. Yet certain conditions (higher temperatures, introduction of more electrons) can overwhelm these forces, ultimately freeing the electrons and essentially turning the Mott insulator into a conductive metal.
“So you (traditionally) have either roaming electrons or localized electrons,” said Hong, a physics professor at the University of Nebraska-Lincoln. “It’s very clearly defined.”
“But in the case of a Mott insulator, electronic interactions cannot be ignored. Because of this correlation, it is difficult to define it as simply a metal or an insulator. If you can tune the strength of the interaction, it could be a metal, or an insulator. it could be an insulator.
By topping a Mott insulator with a gate insulator made of a so-called ferroelectric material, then using voltage to reverse the latter’s polarization, or alignment of positive and negative charges, the researchers realized they could direct the Mott transition from insulation to the metal and vice versa. In this way, the behavior of the pairing and its most promising function followed that of the semiconductor.
Yet the metallic phase of the Mott insulator gives it a vital advantage over its long-standing counterpart: it carries a number and density of electrical charges that far exceed what a semiconductor could ever do.
The higher this density, the less space charged electrons need to filter electric fields that would otherwise reverse the polarization of the ferroelectric and prevent the transistor from maintaining an “off” state. And the shorter the length of shielding these charged particles need, the smaller a transistor can be – potentially smaller than any of its semiconductor predecessors.
The problem? This same reduction density also increases the difficulty of passing the Mott channel from the insulator to the metal, or vice versa, through the ferroelectric material above.
Engineers often measure the technological viability of a transistor in terms of its on-off ratio: the amount of current it carries when a voltage is applied compared to the amount, ideally close to zero, that flows when voltage is removed. The higher the ratio, the greater the margin for error when processing and storing data. Minimizing current in the “off” state also saves power, while maximizing it when “on” can increase processing speeds.
In 2018, a year after Hong, PhD advisor Yifei Hao and postdoc Xuegang Chen first tackled the problem, another research team reported an on-off ratio of 11, the highest to date in a Mott-ferroelectric pairing at room temperature. Through some experimentation, the Husker-led team eventually pushed that score to 17, better, but still way too low.
Eventually, Hong and his colleagues decided to try adding another layer below the Mott Canal. For this third underlying layer, the team chose a material that could not support the same density of charges as the Mott material above, but which, more importantly, could allow the charges to migrate from the Mott downwards. , as they tend to do when they encounter a sparser region.
The team had in fact preserved the Jekyll and domesticated the Hyde. The space-saving benefits of the high-density region remained, but as the overall density decreased (thanks to the additional bottom layer), the team also retained greater control over the insulator-to-metal transition. This benefit manifested itself in the form of a record on/off ratio, which stood at 385, more than 20 times higher than anything reported so far.
That figure, Hong said, may well exceed the ceiling of what can be achieved with Mott’s ferroelectric approach.
Also in his favor? Ferroelectrics are non-volatile, meaning they can maintain their 1s and 0s without constant power. And the fact that they only need a few sips of voltage to reverse their polarization makes Mott-ferroelectric coupling more energy efficient than non-volatile but magnetism-based memory, including MRAM.
“For me, in terms of technology development, this is a big deal, because it shows that it’s possible,” Hong said. “We can have high-performance devices, retaining many of the manufacturing processes of conventional semiconductors and overcoming some of their fundamental limitations.”
According to Hong, Mott-based transistors could end up in these devices sooner rather than later.
“I think it’s ready,” she said of the concept. “It’s really competitive with other non-volatile memory technologies. I think someone with the right mindset can take the concept and use it.”
More information:
Yifei Hao et al, Record resistance switching at high ambient temperature in ferroelectrically gated Mott transistors unlocked by interfacial charge engineering, Natural communications (2023). DOI: 10.1038/s41467-023-44036-x
Provided by University of Nebraska-Lincoln
Quote: New hardware design for transistors could reduce next-generation technology (February 2, 2024) retrieved February 2, 2024 from
This document is subject to copyright. Except for fair use for private study or research purposes, no part may be reproduced without written permission. The content is provided for information only.