Reliably unreliable nanotechnologies
 

Objectives of the project and expected results

The twentieth century has witnessed an exceptional technological progress that has mainly been driven by the invention of the transistor and integrated circuits. Chemistry and materials science have played a pivotal role in this ICT evolution by enabling the development of active devices with distinct and reliable properties that over the past 60 years have been following Moore’s scaling trend. CMOS technology is however approaching the nanoscale floor, with devices attaining comparable dimensions to their constituting atoms, confronting us with challenges associated with the performance, reliability and manufacturability of conventional analogue and digital circuits. We have nowadays reached the point where it is imperative to substantiate “beyond CMOS technologies”, based upon inherently unreliable information processing and memory elements; fulfilling Feynman’s vision on fabricating and operating emerging devices by “manipulating a few atoms”.

The realization of a nanoscale memory-resistor (memristor) by Hewlett Packard (HP) in 2008 came almost 40 years after its theoretical inception and manifests a prominent paradigm for extending CMOS beyond its current physical limits, both in terms of memory and computation. When compared to existing (flash) and other emerging memory technologies such as ferroelectric, magneto-resistive and phase-change random-access memory (FeRAM, MRAM and PCRAM), resistive RAM (ReRAM) elements are considered advantageous due to their infinitesimal dimensions, their capacity to store multiple bits of information per element and the miniscule energy required to write distinct states. Yet, most of their impact is anticipated through the realisation of bio-inspired/-mimetic systems that can support unconventional computation formalisms, due to the devices’ capacity to compute and store information locally.

Thus far however, we have not been able to harness their full potential, due to the following technological barriers:

  • There is no convincing indication yet of what the underlying physical switching mechanism is, with researchers trying deciphering the phenomenon at the device level. This restrains the programmability and/or controllability as well as the further scaling of the devices.
  • Memristors have been traditionally researched as emerging memory elements, with engineers focusing on enhancing non-volatility. This approach seeks diminishing mechanisms leading into unstable equilibrium conditions that can cause state restoration; essentially suppressing the devices’ intrinsic rich-dynamics.
  • As ReRAM cells are scaled substantially, intrinsic defects and conduction channels become quantum confined, making them inherently unreliable. This aggravates the devices’ variability that prohibits their deployment with conventional deterministic circuit paradigms for realising practical applications.

To address these issues it is imperative to use the strengths of many. This fellowship leverages the fellow’s strong expertise on nanotechnologies, electron devices (particularly memristors) and bio-inspired systems, in addition to an interdisciplinary research team that altogether is capable to coordinate aggressive technological developments within the ICT remit and beyond (materials and chemistry). During this programme we will:

  • Fabricate, characterise and model phase transitions of functional-oxides; from atomic-scale to devices.
  • Engineer volatile switching mechanisms into solid-state devices to devise dynamic computation elements.
  • Formulate circuits/systems/architectures that are capable of harnessing the devices’ rich internal dynamics for solving computation-demanding spatial-temporal problems.