Spin Properties of Confined Electrons in Semiconductors

Recently we found indirect evidence for a most efficient way of exciting spin Rabi oscillations of electrons in Si quantum wells by an electric current. The underlying mechanism is the current induced Bychkov-Rashba (BR) field which is a consequence of spin-orbit interaction in systems with lower than mirror symmetry. In the case of one- sided modulation doped Si quantum wells, the electric field due to the charged donors is responsible for this effect which was shown by us to rule also spin decoherence and relaxation. The BR field (i) shifts the g-factor when a dc current is applied and (ii) it causes spin transitions when microwave currents are excited in the sample. The latter is expected to require a by many orders of magnitude lower microwave power than the usual magnetic dipole excitation. In this project we plan to quantify this effect, to exploit it to investigate spin resonance in Si nanostructures, and to extend it also to other materials (GaN, graphene). In addition to classical ESR investigations where a microwave resonator is used, we intend to apply the microwave field via striplines and contacts to the sample. Alternatively, microresonators will also be considered. Making use of electrical spin resonance detection via the low frequency conductivity of the sample, also nanostructures, like quantum wires can be investigated, since the limitation due to the integral sensitivity of conventional ESR spectrometers can be avoided. The possibility to excite spin resonance in quantum dots by electric dipole transitions will be also investigated. In all cases, spin echo experiments are planned in order to demonstrate full spin manipulation control and to investigate spin decoherence and relaxation. With this new technique at hand, we can extend the ESR investigations also to low temperatures (< 2K) and variable frequency. Both types of experiments are very difficult with conventional ESR systems because of bulky resonators designed for a single frequency. The frequency dependence will allow to investigate the transition regime from low to high mobility (small to large ct ) in order to verify our model for the earlier observed Dysonian ESR lineshape. Low temperatures will be needed to investigate the effect of spin polarization on the so- called 0.7 anomaly in 1d conductors and also the valley-orbit resonance in Si quantum structures.

In addition to their charge, electrons have also a magnetic moment (connected to their "spin") in that respect they constitute the smallest conceivable kind of a compass needle. Due to their quantum nature, the latter can be stably oriented only parallel or antiparallel to a magnetic field. This property suggests them for the realization of a binary unit a logical bit. Among the advantages could be that the electron does not have to be moved between two pockets when its logical value is changed, in contrast to the classical charge-based electronics: moving the electron costs time and more energy. This is one of the motivations to think about a spin-based electronics, called spintronics. There is, however, a long way towards the practical realization of spin-based devices. One of the first questions is how the spin can be adjusted in an efficient way and how long the input information can be stored. These questions were treated in this project. It turned out that exposure of an electron in a semiconductor layer to a high frequency electric field can be by orders of magnitude faster and more efficient than the traditional use of high frequency magnetic fields. The origin of this lies in the so-called Rashba effect which appears in samples without mirror symmetry. In such samples, the spin and the velocity of the electron are not independent and thus the spin can be affected by moving the electron back and forth. A similar coupling occurs also for a very fast moving "relativistic" electron a situation recognized by Erwin Schrödinger already in 1930, shortly after Dirac had published his famous relativistic wave equation. Schrödinger concluded that such an electron exerts an oscillatory motion transverse to its velocity (in a way the origin of the magnetic moment) even without any force, but the frequency of this "Zitterbewegung" is far too high and the amplitude by far to small for observation. In this project it is shown that in semiconductor layers an analogous situation appears but the frequency is strongly reduced and the amplitude enhanced so that this effect finally became observable. Other results include the following. Carbon nanotubes become ferromagnetic (i) at low temperatures which is unexpected since they do not contain metal impurities. Conduction electrons stabilize this ferromagnetic phase. Ferromagnetic resonance (FMR) an effect similar to electron spin resonance - can be used (ii) to detect and to identify tiny ferromagnetic precipitates - an important tool in the process of searching for ferromagnetic semiconductors. In a mixture of a ferromagnetic semiconductor and a ferroelectric one, Ge1-x Mn x Te, we found that a ferroelectric distortion changes the orientation of the magnetization (iii) indicating that the magnetic moment can be reoriented directly by an electric field. We found a new method to position (magnetic) nanocrystals (iv) with nm precision. Using a magnetic force microscope, we could detect the magnetization of these nanocrystals and we demonstrated thus that a magnetic memory with more than 10 Terrabit/inch2 can be realized.