At the heart of all modern gadgets, like mobile phones and computer chips, are tiny transistors that carry electrical signals at close to the speed of light. Their design and the way they are placed decide how efficiently a gadget can perform its operations.
My group at the Indian Institute of Science, Bangalore, has invented and demonstrated a transistor design which can lead to a breakthrough in future chip technology. The proposed transistor has significantly better performance and scalability as compared to the technology used in current computer and cellphone chips. The invention has been recently patented by the IISc and the results were subsequently published by our group.
Over the years, transistors have become smaller and smaller, which has resulted in denser, cheaper, sleeker and high-performing chips, as predicted by Moore’s law in 1970. Intel co-founder Gordon Moore had then predicted that the number of transistors and other such components in an integrated circuit would roughly double every year. The law has held till now and the chip industry has planned its development according to this prediction. We see consistent improvement in speed and power consumption of wireless and handheld devices.
The current generation of transistors, widely adopted by semiconductor giants since 2011, is called FinFETs, mainly because they resemble the fin of a fish. They measured about 22 nanometres when they came about. A hydrogen or helium atom is roughly one tenth of a nanometre. The FinFETs are expected to shrink further to the size of about 10 nanometres (nm) by 2018. However, a below-10nm size can adversely affect FinFET’s performance. At this scale, the movement of electrons, which carry the signals and perform the electronic tasks, does not obey the switch that is meant to regulate their movement. The transistor starts behaving like a wire without a switch.
An out-of-the-box transistor design is the need of the day for consistent improvement in computer chips and mobile phones.
Because the chip designing industry was staring at stagnation beyond 10nm, a new transistor design needed to be worked out. What was important was that this new design needed to be supported by the existing manufacturing processes in order to facilitate a smooth transition. A ‘disruptive technology’ that calls for an entirely new way to put things together takes years to be adopted and requires fresh investment.
Since it takes time for a new technology to go from prototype to mass production, the semiconductor industry feels that the newer design must be in line with current-day manufacturing technology. In the multi-billion-dollar semiconductor industry, any abrupt change in structure and manufacturing process of devices is prohibitively expensive. The solution we need for the future is to have a fundamentally different mechanism of operation but with a manufacturing process that is similar to current FinFET technology. This will ensure a smooth and economically viable transition to the new solution. The proposed transistor design keeps all this in mind.
This has been achieved through a process called quantum tunnelling. The device can make use of existing-FinFET based chip manufacturing facilities and process; however, unlike FinFETs, it modulates vertical tunnelling across hetro junction to control electron flow. Published results depict that this breakthrough transistor technology, beside higher speed, consumes less power and will heat up less as compared to the existing transistors. This will help in extending the battery power of next-generation electronics.