New types of computer memory can greatly reduce energy use and improve performance

(Nanowerk News) Researchers have developed a new design for computer memory that could greatly improve performance and reduce the energy demands of internet and communications technologies, which are expected to consume nearly a third of global electricity in the next ten years.

The researchers, led by the University of Cambridge, developed a device that processes data in the same way as the synapses in the human brain. The device is based on hafnium oxide, a material already used in the semiconductor industry, and a small, self-assembled barrier, which can be raised or lowered to allow electrons to pass.

This method of changing the electrical resistance in computer memory devices, and allowing information processing and memory to exist in the same place, could lead to the development of computer memory devices with much greater densities, higher performance, and lower energy consumption. The results are reported in a journal Science Advances (“The thin film design of the amorphous hafnium oxide nanocomposite enables strong interfacial resistive switching uniformity”).

Our data-hungry world has led to ballooning demand for energy, making it even more difficult to reduce carbon emissions. In the next few years, artificial intelligence, internet use, algorithms and other data-driven technologies are expected to consume more than 30% of global electricity.

“To a large extent, this explosion in energy demand is due to a shortage of current computer memory technology,” said first author Dr Markus Hellenbrand, from Cambridge’s Department of Metallurgical and Materials Sciences. “In conventional computing, there’s memory on one end and processing on the other, and data shuffled back between the two, which wastes energy and time.”

One potential solution to the problem of inefficient computer memory is a new type of technology known as resistive switching memory. Conventional memory devices can have two states: one or zero. A functioning resistive switching memory device, however, would be capable of a continuous range of states – computer memory devices based on this principle would be capable of much greater density and speed.

“A typical USB stick based on continuous range would be capable of storing between ten and 100 times as much information, for example,” says Hellenbrand.

Hellenbrand and his colleagues developed a prototype device based on hafnium oxide, an insulating material already used in the semiconductor industry. The problem with using this material for resistive switching memory applications is known as the uniformity problem. At the atomic level, hafnium oxide is structureless, with hafnium and oxygen atoms mixing randomly, making it difficult to use for memory applications.

However, the researchers found that by adding barium to a thin film of hafnium oxide, some unusual structures started to form, perpendicular to the plane of the hafnium oxide, in the composite material.

These vertical barium-rich ‘bridges’ are highly structured, and allow electrons to pass through them, while the surrounding hafnium oxide remains unstructured. At the point where this bridge meets the device contacts, an energy barrier is created, which electrons can cross. The researchers were able to control the height of this barrier, which in turn changed the electrical resistance of the composite material.

“This allows multiple states to exist in matter, unlike conventional memory which only has two states,” Hellenbrand said.

Unlike other composite materials, which require expensive high-temperature manufacturing methods, hafnium oxide composites can be self-assembled at low temperatures. Composite materials exhibit high levels of performance and uniformity, making them very promising for next-generation memory applications.

A patent on the technology has been filed by Cambridge Enterprise, the University’s commercialization unit.

“What’s really exciting about these materials is they can work like synapses in the brain: they can store and process information in the same place, as our brains, making it very promising for the fast-growing fields of AI and machine learning,” Hellenbrand said.

The researchers are now working with industry to carry out a larger feasibility study on the material, to understand more clearly how the high-performance structure forms. Since hafnium oxide is a material already used in the semiconductor industry, the researchers said it would not be difficult to integrate it into existing manufacturing processes.