Flash Memory To Be Based on 2D Materials — A Single Atom Thick?

Memory cells are one of the most important building blocks utilized in digital electronics.  The explosion in growth of mobile devices has fueled the continuing miniaturization of components.  Some technologies eventually reach the point of their smallest possible functional versions being achieved.  Any further attempts at size reduction run smack up against limitations that often seem insurmountable.  This sets the stage for moving to an entirely new technology or production breakthrough.  What if  I showed you that nonvolatile memory cells  can be created utilizing materials in layers that are a single atom thick?  Memory cells so thin that they are virtually only two-dimensional?

dk backgrd flash

Engineers in Switzerland at the Electrical Engineering Institute Ecole Polytechnique Federale De Lausanne have achieved success creating nonvolatile memory cells that open up some highly scalable and adaptable possibilities.  With current flash memory technology, some significant hurdles would have to be overcome to further continue miniaturization to ever-smaller architectures.  Continued vertical scaling results in reduction of program/erase voltages, but can’t get past the charge retention requirement of the floating gate.  The thickness of the floating gate would need to be reduced to mitigate some of the negative effects, such as capacitive interference between neighboring cells.  This led the Swiss engineering team to investigate utilizing 2D materials — these can exist in a stable state in a layer only a single atom thick.  The team’s aim was to investigate the use of such materials to replace traditional choices for semiconducting channels, interconnects, and charge trapping layers in a memory cell.


Graphene, Boron Nitride and Molybdenum Sulfide are the thinnest known two-dimensional materials  with electronic properties that can be utilized in a broad range of nanotechnology applications.  At layers as thin as a single atom, this would represent the ultimate limit in vertical scaling.  These three materials could be all that is needed to build electronic circuits: the boron nitride (BN) could act as the substrate and gate dielectric barrier, graphene as the interconnect layer, and molybdenum sulfide (MoS2) as the semiconducting channel.  Graphene being semimetallic , it could create an ideal contact to the 2D semiconductors, and would be capable of supporting large current densities.  Since the graphene’s work function can be chemically or electrostatically tuned, it could adapt to a wide array of 2D semiconductors with different work functions and band gaps.  Utilizing layers of these materials results in a new heterostructure capable of operating as a memory cell.


The Swiss team has demonstrated that it is possible to design memory devices using 2D building blocks, including contacts, floating gate, and the semiconductor channel.  Also, the excellent mechanical properties of molybdenum sulfide (and other 2D semiconductors) can be utilized with flexible substrates.  These devices could be produced inexpensively and on a massive scale utilizing liquid-scale processing and roll-to-roll printing of CVD-grown material.  You can view the Swiss engineering team’s complete report here.  Could this be the future of flash memory technology in its infancy?