As conventional electronic components approach the physical limits of miniaturization, scientists are looking for ways to exploit the quantum-mechanical properties of subatomic particles to increase storage density and processing power. One intriguing possibility is a new type of circuitry that is based on a property of electrons called spin.
Conceptually, you can think of an electron as a tiny, electrically charged sphere that spins about an axis. Because the spinning sphere carries a charge, it forms a current loop that gives rise to a magnetic field with poles at either end of the spin axis. (In reality, electrons are not tiny spinning spheres, and the exact nature of spin is too complicated to discuss here.) Interestingly, all electrons have exactly the same amount of spin, which is restricted to a discrete value by the laws of quantum mechanics.
Spin is traditionally depicted by assigning a direction to the axis: if you view the “rotation” as going west to east, the axis points north or up, as with the Earth. An external magnetic field affects the electron's energy according to the relative orientation of the field and spin axis.
In ordinary electrical current, the electrons' spins point in random directions and have no effect on the operation of the circuit. However, circuits in which the spin direction plays a role offer some significant advantages in performance and power consumption; such circuits are known as spin-based electronics, or spintronics. Perhaps the simplest example is a current passing through a magnetized ferromagnetic metal, such as iron or cobalt, which tends to impede electrons in all but one spin direction, resulting in a spin-polarized current.
That is the basis of a magnetic tunnel junction, in which two ferromagnetic layers are separated by an insulating layer. One of the ferromagnetic layers is permanently magnetized in a specific direction, while the other one's magnetic orientation can be changed at will. Electrons entering the fixed, or pinned, layer are spin-polarized, and if the magnetic field of the variable layer is parallel to that of the pinned layer, some of the electrons tunnel through the insulator (a fascinating quantum effect in itself). If the magnetic fields of the two ferromagnetic layers are antiparallel, the electrons do not tunnel through the insulator.
A single magnetic tunnel junction can be used to store one bit of data within a larger structure that is called magnetic random-access memory (MRAM), which retains its data whether the power is on or off and offers switching rates and rewritability comparable to that of conventional RAM. Motorola has fabricated MRAM chips with capacities of one megabit (see Fig. 1), and commercial products should become available in the next couple of years.
A similar idea might be applied to field-effect transistors (FETs), as proposed in 1990 by Supriyo Datta and Biswajit A. Das, then of Purdue University. In the Datta-Das spin FET, a ferromagnetic source electrode injects spin-polarized electrons into a semiconductor channel that connects the source with a ferromagnetic drain electrode. If a voltage is applied to the gate electrode directly above the channel, the resulting electric field changes the spin direction of the electrons, causing them to be rejected by the drain's magnetic field; otherwise, they pass through the drain unimpeded.
Changing the spins in this manner takes much less energy and time than pushing electrons out of the channel as in a conventional FET. In addition, it might be possible to change the magnetic orientation of the source and/or drain to alter the logic gate's function on the fly. As yet, no one has constructed a spin FET, but recent experiments with ferromagnetic semiconductors hold great promise for the future of spintronic computers and the musical applications they might serve.