#077 Scrutinizing the Future of Transistors

Assistant Professor UEDA Akiko, Faculty of Pure and Applied Sciences

Integrated circuits constitute the heart of such IT devices as personal computers and mobile phones. Within these devices, transistors serve as switches that turn electrical signals on and off to perform all manner of operations. Through transistor miniaturization, CPU performance has made mind-boggling advances. However, miniaturization may be approaching a limit. In order to search for a new type of transistor that can move past such limits, simulations of the conduction of electrons in various materials and structures are being performed.

In the world of semiconductors, there is a famous empirical relationship known as "Moore's law". It was proposed in 1965 by Gordon Moore, one of the founders of Intel, and was a prediction of technology progress, stating that the performance (the number of incorporated transistors) of semiconductor integrated circuits would double every 18 months. By miniaturizing transistors, the number that can be incorporated into a given area increases. As a result, the performance of a computer or other device using the integrated circuit increases as well. And in fact, up until a few years ago, the processing performance of integrated circuits continued to rise as predicted by this law. The size of a single transistor, which in 1970 or so had been 10 μm, today has shrunk to a little over 10 nm (where 1 nm = 0.001 μm = 0.000001 mm). However, the distance between atoms in silicon, which is the main material of transistors, is about 0.3 nm. Hence it is in principle impossible to make transistors smaller than this. In order to further boost CPU performance, it will be essential to develop a transistor that operates according to some mechanism fundamentally different from that of conventional semiconductor devices. But what kinds of materials and structures are suitable for such new transistors? And how will the transistor characteristics change depending on the material used? These are the questions that Prof. Ueda is exploring through mathematical simulations.

In nanometer-size materials, electrons show the behavior as wave that is different from the world we can see with our own eyes. This is called quantum effects, and as a consequence, phenomena occur that cannot be explained by ordinary transistor theory. Such phenomena can be observed experimentally, but materials cannot be modified at the nano-level in order to explore the reasons for this behavior. Hence the need for simulations. A theoretical model is constructed which takes quantum effects into account, and when the model is applied to calculate the states of a material, the meaning and causes of phenomena that appear to be abnormal can be comprehended. The aim is to feed back these insights in an effort to improve material properties.

It is easy enough to use the term "simulation" for what is done in this research, but the properties of the materials themselves, their electrical conduction characteristics, and other behavior span an extraordinarily broad range, and the fields of specialty are all different. There are also diverse levels to the simulations, which run the gamut from simple ones that can be calculated immediately on a single general-use personal computer, to simulations that must be run on a supercomputer, such as the K computer, and take an entire week to complete. But it is not necessarily the case that calculation using a supercomputer yields higher-precision results. When trying to capture features intrinsic to matter, complex quantum mechanical calculations become necessary, and when attempting to grasp the changes in characteristics of a transistor or other device arising from its structure or other parameters, it turns out that simple calculations are sufficient. In either case, the various properties of matter, which is to say, physics fundamentals, must be properly understood.

Explaining the urgency of research on next-generation transistors

Prof. Ueda is currently conducting research on semiconductors for use in transistors and other next-generation devices. She says it's not because she was originally particularly interested in electronic devices. As a child, she was interested in living things, and wanted to know such things as how organisms moved, and why they are intelligent. But she reasoned that this line of inquiry would ultimately end up addressing the behavior of molecules and atoms, and so in college she majored in physics. And she explains that while studying quantum mechanics, which deals with the fundamental properties of matter, she undertook research on the flow of electrons, and a demand for this expertise arose in the field of device R&D. It may appear that she has entered a completely different field, but as it happens, transistors and living things are not completely unrelated. For example, the transfer of information by nerve cells relies on electric currents, and there are organisms that themselves generate electricity. Seen from the vantage points of quantum mechanics and devices, it may even be possible to grasp new aspects of living things.

It is widely understood in this field that transistors are currently approaching a limit. However, there is as yet no consensus as to what might be suitable as a replacement. Candidates being considered at present include transistors that utilize an electron tunneling phenomenon (tunnel FETs), and transistors using a semiconductor other than silicon, such as graphene (in which carbon atoms bond in a honeycomb structure to form a single layer) and other layered materials. However, when considering various practical aspects of use as devices, such as compatibility with conventional transistors, reliability, manufacturing efficiency, cost, and so on, at present there are not yet any highly promising contenders. That said, it is probably only a matter of time before a new type of transistor emerges. And it is anticipated that this will in turn bring about major changes to a variety of devices and systems. Through the world of vanishingly tiny electrons, one might be able to glimpse an image of society in ten or twenty years' time, or beyond.

During her studies as a foreign student in Israel at Ben-Gurion University of the Negev, in the middle of the Negev desert.
One pleasant memory is a countdown to the new year in the laboratory with post-doc friends from overseas.

Article by Science Communicator at the Office of Public Relations