Advanced Computing in the Age of AI | Friday, March 29, 2024

Graphene Comes to Flash Memory 

<img style="float: left;" src="http://media2.hpcwire.com/dmr/flashstorage.jpg" alt="" width="95" height="71" />The single-atom thick material known has graphene has been at the center of both research and the Nobel Prize recently thanks to its ultra-thin build and unique electronic properties, spanning applications from water desalination to solar cells. But graphene's usefulness is not exclusive to these applications, as demonstrated by a new type of flash storage being developed with graphene at its core.

The single-atom thick material known has graphene has been at the center of both research and the Nobel Prize recently thanks to its ultra-thin build and unique electronic properties, spanning applications from water desalination to solar cells. But graphene's usefulness is not exclusive to these applications, as demonstrated by a new type of flash storage being developed with graphene at its core.

As one of very few known atom-thin materials, graphene is especially promising because electrons are able to pass through the material as if they had no mass, making it the closest thing we have to a room-temperature superconductor. This means that electrons can more easily move across graphene's surface, leading to data storage that should require less energy and produce less heat.

Taking advantage of these properties are scientists at the École Polytechnique Fédérale de Lausanne's Laboratory of Nanometer Electronics and Structures (LANES), who recently detailed the results of this research in the journal ACS Nano.

The LANES team first made headlines two years ago with the announcement of a molybdenite (MoS2) chip that is expected to surpass the capabilities of silicon-based chips and electronic transistors. Now they've combined MoS2 with graphene to take combine both materials' electronic properties while taking advantage of this single-atom thick dimensions. Compared to silicon memory, the prototype shoes great potential for miniaturization and mechanical flexibility.

The transistor prototype is best-pictured as a sandwich: a thin layer of MoS2 channels electrons between layers of graphene, which are responsible for transmit electricity to the MoS2 layer and capture electric charge.

At the moment, the graphene components are several layers thick, so the design will likely shrink down in upcoming versions. However, the manufacturing process currently involves finding individual graphene layers by hand using an optical microscope, leaving room for improvement for the production process.

Nonetheless, researchers were able to assemble the ultra-thin devices, and once they did, the memory worked. As intended, applying positive voltage to the control electrode meant that charge accumulated in the graphene layers. Reversing the voltage then erased the devices.

The prototypes functioned for all 120 write/erase cycles over which they were tested. Afterwards, the researchers left the device unpowered to measure the storage decay over time, estimating that after ten years, 30 percent of the original charge would still be present.

Not only did charge decay very slowly, the difference between the clear and charged states was extremely large – roughly 1 to 104 difference in current flowing through the device. This could allow graphene-based flash to encode multiple distinct states to store multiple bits, just as multilevel flash does now.

But going forward, researchers seem confident in the devices' ability to be mass-produced into existing electronics. However, it's also clear that the technology still has a ways to go before it will come off the assembly line.

Full story at Ars Technica

EnterpriseAI