Electric Fi Elds Drive Spins

Spin is the only internal degree of freedom of an electron, and that raises two interesting, closely related questions. First, what kind of new physics emerges when the processes traditionally attributed separately to electron charge and electron spin become closely interconnected? And second, to what sort of new electronic devices might this physics lead us? Success in applications has the potential to add a whole new branch of technology to traditional solid-state electronics — already given the name ‘spintronics’. On page 195 of this issue, Duckheim and Loss1 tackle the fi rst question, and raise implications for the second, with their theory of electron spin resonance driven by a time-dependent electric fi eld. Th is ‘electric dipole spin resonance’, or EDSR, is essentially diff erent from the traditional electron paramagnetic resonance (EPR) that is driven by a magnetic fi eld of the same frequency. Textbook physics teaches us that an electron possesses charge and angular momentum, called spin. Th e magnetic moment associated with the spin is known as the Bohr magneton. Electric fi elds couple to the charge and control electron orbital dynamics; magnetic fi elds couple to the magnetic moment and control spin dynamics. For slow electrons in a vacuum, corrections to this simple picture have no major consequences. However, for an electron in a crystal environment the pattern is rather diff erent. Even when the mean velocity of an electron is small, its motion includes orbiting around nuclei, and this motion may be relativistic. As a result, orbital and spin motions become coupled. Th e correlation of these motions is known as spin–orbit coupling. A simple estimate shows how strong the changes in spin dynamics are expected to be. In a vacuum, the only characteristic energy is the Dirac gap for creating electron–positron pairs, which is as large as 106 eV, and spin–orbit coupling is inversely proportional to this quantity. In crystals, typical gaps for creating electron–hole pairs are only about 1 eV. Typical spin–orbit level splittings for elements from the middle and lower parts of the periodic table are scaled with the same energy. As a result, the coupling of the two motions can be enhanced by six orders of magnitude. Th is giant enhancement of spin–orbit coupling makes the electrical operation of electron spins in solids practical, and a useful playground to test the eff ect is found in quantum wells — narrow semiconductor layers with thicknesses of only about 15 nm. Electrons confi ned inside a quantum well behave as nearly free two-dimensional particles. Th e properties of these particles can be controlled by growth conditions and external fi elds, and for many of them no analogy exists in the physics of free electrons in a vacuum. In such systems, the application of a driving inplane electric fi eld results in homogeneous spin polarization in the bulk, and in accumulating spin populations near the fl anks of the sample; the latter phenomenon is known as the spin Hall eff ect2. a b