The Internet of Things (IoT) is widely predicted to cause a profound technological and societal transformation, on a par with the technological innovations of microelectronics and the internet that will underpin it. The number of IoT devices increased 31% year-on-year to reach 8.4 billion by 2017, and it is estimated that there will be in excess of 20 billion such devices by 2020. The IoT, along with artificial intelligence and automated (robotic) production are expected to add $3.75 trillion to the World economy by 2025, whilst at the same time addressing the UN Sustainable Development Goals.
However, there are considerable technological, societal and ethical challenges associated with the IoT. Examples include communication bandwidth (capacity), privacy and data security. Other, more fundamental and practical challenges exist at the level of the individual sensor and associated hardware embedded in the ‘Things’. In particular, many of these sensors may be electrically isolated (autonomous), either having their own power source (battery) and/or be required to harvest or scavenge energy from their environment, for example from solar, wind, thermal or vibrational sources. Harvesting provides very little energy, so such autonomous sensors are required to have the lowest possible power consumption. They may even need to remain in a quiescent state slowly collecting or storing data for extended periods, before bursting into life when sufficient energy is available to connect to the outside world. A key component in achieving this is the devices’ memory. Such a memory should work at low voltages, use as little energy as possible when being programmed, and be capable of robustly storing the data almost indefinitely with no power.
In this project, we will develop a potentially ground-breaking type of novel compound-semiconductor memory, ULTRARAM, for use in autonomous IoT sensors. ULTRARAM is predicted to have an intrinsic storage time that exceeds the age of the Universe, a single bit switching energy (per unit area) that is orders of magnitude lower than any conventional or emerging memory and ultra-fast programming speeds. ULTRARAM should be compatible with silicon or compound semiconductor platforms, allowing it to be integrated into a broad range of sensor types.
The objective of the project is to design, fabricate and test small (2×2) arrays of ULTRARAM memory cells on silicon wafers.