You are most likely reading this article on a computer, and as you scroll down the page, you may decide to briefly switch over to Facebook or Twitter to type a quick status update. We usually do not stop to think that the ease with which we can do this is due to the seamless flow of charged electrons in our computers. Thus far, utilizing charged electrons to make computers has been endlessly fruitful, allowing us to build smaller and faster computer chips. Unfortunately, we cannot continue improving technology simply by scaling down to smaller sizes because we will eventually reach atomic sizes where our devices will no longer function. As we look ahead into the not-too-distant future, we will need to explore new, innovative technologies that go beyond utilizing electron charge – one such exciting new direction is the field of spintronics.
A Brief History of Transistors
Our tech-savvy age began in 1948 with the invention of the transistor at Bell Labs. A transistor is a device that allows us to amplify a signal by taking a small input current of electrons and translating it into a large output current. Transistors also allow us to turn a signal off and on, giving computers a code of “0”s and “1”s to store information (memory) and perform instructions (logic).
At the most basic level, transistors work by creating a current, or flow, of electrons. Electrons are subatomic particles that carry a “negative charge” (a term coined by Benjamin Franklin) and are attracted to positively charged subatomic particles such as protons. Within a transistor, we can we create an area of net positive charge that attracts electrons, and this allows current to flow to that part of the transistor. When the net charge is the same everywhere within the transistor, current does not flow and the device is off. In this way we can control if the transistor is on or off. We can use the state of the transistor to store information in the computer and execute instructions.
The Problem with Moore’s Law
In the last fifty years, we have observed a trend in technology known as Moore’s Law, which describes the doubling of the amount of transistors that can fit onto a single computer chip every two years. This has allowed us to continue making computers smaller, lighter, and more energy efficient. In fact, transistors were the size of a human blood cell in the 1970s, and shrank down to the size of a virus by 2000 [1]. To give you a better idea, if transistors were the size of a basketball in the 1970s, transistors in the year 2000 would be equivalent to the size of a pea. If this trend continues, five years from now these building blocks of computers will be just over one molecule thick! Once we reach these atomic scales, however, it will become impossible to make transistors both thin and smooth enough to conduct charge properly. To solve this issue, we need to think of more innovative ways to improve transistors.
Although the computers we use today all utilize electron charge to create current, electrons have many other interesting properties beyond just their charge. Turning to quantum physics, scientists are now looking at the idea of using the property of “electron spin” to encode classical information, which has given rise to the field of “spintronics.”
The Potential of Spintronics
Electrons have a quantum mechanical property called “spin.” Because spin is quantized, each electron has exactly two possible spin states, either spin up or spin down. In spintronics, engineers make use of this property to make the 0’s and 1’s needed for computer memory and logic.
One way we can harness this property of electron spin is to use magnetic fields and electric currents to switch the spins of nanoscale magnetic materials. Certain metallic elements such as copper, nickel, and iron have an odd number of electrons circling their atomic nucleus, all but one of which are tightly paired. This extra unpaired electron has a spin that is free to be either up or down. In addition, there is a strong quantum-mechanical interaction between the spins of neighboring atoms. Thus, these extra electrons can collectively align their spins with the direction of a magnetic field. Even when the magnetic field is taken away, these extra electrons maintain their original spin orientation [2].
Unlike electron charge, electron spin exhibits “collective behavior” when switching between on and off states. When we move electrons using charge, each electron acts independently and does not communicate with other electrons. However, when we apply a magnetic field or a current to a nanoscale magnet, the spins of the electrons can quickly and collectively jump to align themselves with the external stimulus. Therefore, the energy needed to switch a magnetic transistor is potentially much lower than that needed for charge-based transistors, leading some to dub this field “green electronics.” Lower energy means less heat loss, which would also allow us to improve computer performance and stability. Unsurprisingly, companies like IBM, Toshiba, and Samsung are actively researching this promising field.
Of course, making the jump to encoding information using spin rather than charge is no simple task. One obstacle is that mass production of electronics is already optimized for the fabrication of charged-based integrated circuits, and it will be expensive to develop similar supporting technologies for spin. Furthermore, some of the physics of collective electron spin behavior is still not fully understood [3], such as how electron spins interact with electron currents. We are a few years away from realizing the full potential of this new technology. Scientists researching nanotechnology are diligently working to come up with the best ways to implement “magnetic logic”, whether it involves using arrays of nanoscale magnetic “dots” or using quantum mechanical tunneling effects in magnetic tunnel junctions [4]. Although we do not yet know what specific design of magnetic spintronics will win out, with further study and application of engineering ingenuity, we can look forward to a day when our computers will be powered by electron spin. Hopefully, this will allow Moore’s Law to continue to hold true for decades to come.
Jean Anne Currivan, Harvard Physics graduate student
References
1. International Technology Roadmap for Semiconductors: http://public.itrs.net/
2. O’Handley, Robert. Modern Magnetic Materials. New York: John Wiley and Sons, 2000.
3. Beach, G.S.D., M. Tsoi, and J.L. Erskine. “Current-Induced Domain Wall Motion.” Journal of Magnetism and Magnetic Materials 320, no. 7 (2008): doi:10.1016/j.jmmm.2007.12.021.
4. Bandyopadhyay, S. and M. Cahay. “Electronic spin for classical information processing: a brief survey of spin-based logic devices, gates and circuits.” Nanotechnology 20, 1-34 (2009).
5. Nuedeck, George W. Modular Series on Solid State Devices. New Jersey: Prentice Hall, 1989.
6. Coclaser, R.A. Microelectronics – Processing and Device Design. New York: John Wiley and Sons, 1980.
7. Pierret, R.F. Semiconductor Device Fundamentals. Massachusetts: Addison Wesley, 1996.