Future of Nano-Electronics in the New Era of Computing

For more than fifty years, miniaturisation of transistors, which are the building block of all electronic chips, has been driving the semiconductor industry. This miniaturisation has been following a trend of doubling of the transistor density in a chip every two years and is popularly known as Moore’s law. In general, reducing transistor sizes facilitates realising electronic products with a higher speed and lower power consumption at lower cost

Asustained technological innovation and economic pulls from varied applications have allowed unabated Moore’s law driven growth of the semiconductor industry. Over the years, the sizes of the transistors have reduced from more than a micrometer in 1980s, to less than ten nanometres, in the current bleeding-edge technologies.

With the transistors going into nanoscale dimensions, there are several technological challenges. There are fundamental physical limits that make further reducing the dimensions challenging. In the state- of- the- art transistors, electrons easily leak way and it is challenging to switch them off. Even when a transistor is controlled by a three-sided structure known as gate, the transistor still leaks away current when kept idle and increases the power consumption. This necessitates the search of alternative nano-electronic devices that can be better controlled than the state-of-the-art transistors.

We have now entered into a new era of computing. The applications such as artificial intelligence, autonomous systems, big data, Internet of Things, 5G etc. have taken the center stage of computing. These applications consume and generate tremendous amount of data and demand unprecedented high computational power.

The research on nano-electronics is being undertaken on several fronts. Lower dimension materials such as nanowires are being explored. A nano-wire is a cylindrical structure with a diameter of a few nanometers. In a transistor fabricated using nano-wires, the gate can surround its

current carrying part all-around and largely stop the leakage of unwanted electrons. Consequently, the transistors and circuits implemented using nano-wires can be more energy-efficient and can replace the state-of-the-art transistors in future.

Though silicon, the most widely used semiconductor, can be used to fabricate nanowire- based transistors, other materials such as III-V semiconductors, carbon nanotubes, graphene etc. are being considered for implementation of transistors. These novel materials offer better speed in contrast to silicon; nevertheless, it is challenging to integrate these materials in the well-established integrated circuit fabrication technology.

In addition to researching new materials for transistor applications, there is a rapid advancement in exploratory nanoelectronic devices with new operating principles. Among these exploratory devices, tunnel transistors seem to be the most promising in replacing the stateof-the-art transistors. Tunnel transistors operate on the principle of quantum tunneling, a phenomenon in which an electron can cross a barrier even though it has energy less than the barrier.

The tunnelling phenomenon is similar to throwing a ball against a hard wall and the ball reaching the other side of the wall without any apparent sign of hole in the wall. Such phenomenon seems weird from our daily experience, but in the realms of electrons tunneling phenomenon do occur. In tunnel transistors, by providing a high barrier, the leaky electrons are stopped.

By exploiting the tunneling phenomenon, a tunnel transistor switcheson even when a very small voltage is applied. Consequently, tunnel transistors can be ten times more energy efficient than the state-of-the-art transistors, and are dubbed as “green transistors”. However, due to high obstruction to current conduction, tunnel transistors exhibit low speed. A great deal of research is being undertaken to make tunnel transistors faster and capable of replacing the state-of-the-art transistors.

There are other exquisite nanoelectronic devices, such as a molecular transistor. It consists of a single molecule for current conduction and represents the ultimate limit of miniaturization. Though, currently, the performance of a molecular transistor is inferior to the state-of-the-art transistor, with more research molecular transistors can be quite interesting. Moreover, researchers are trying to exploit another property of an electron known as “spin”.

Traditionally, transistors utilize the “charge” of an electron for information processing. However, the spin of an electron can also be manipulated in a transistor and be utilized for information processing. It is expected that in some scenarios spin-based transistors can be faster, more energy-efficient and compact than the traditional chargebased devices.

Another trend that is driven by the energy- efficient data- intensive applications is to move away from highprecision computing to novel computing paradigms such as probabilistic computing, quantum computing, approximate computing and bio-inspired computing. These novel computing paradigms, in general, exploit massive parallelism and increased tolerance to errors to gain energy-efficiency. These computing systems show unique characteristics that even the current bleeding-edge transistors cannot deliver. However, these new computing paradigms require innovation at both the system level and at the device level, which the research programmes in nano- electronics is expected to deliver.

Thus, with the advent of new era of computing, we have entered into a new phase of research and innovation. Though the applications such as artificial intelligence, big data and internet of things, and their impact on our daily life is becoming evident, there are several technologies which need to be developed to fully empower these applications. Among them, nano-electronics are one of the most critical enablers. In the times to come, the research on nanoelectronics will become crucial, as the easy returns obtained from Moore’s law driven miniaturization will soon perish.

“It is now evident that the state-of-the-art transistors, even when scaled down to very small dimensions, cannot fulfil the requirements of the new era computing. The demand for energy-efficient computation from the newer applications has made the research on nano-electronic devices necessary” — Dr. Sneh Saurabh, Professor, Indraprastha Institute of Information Technology (IIIT) Delhi