Search

We studied narrow (submicron) constrictions in the layers of ferromagnetic semiconductor (Ga, Mn)As. We have demonstrated a contribution of the quantum localization effects to the magnetoresistance of the constricted samples. We have also found a negative contribution of a domain wall trapped in the constriction to the resistance, due presumably to the erasing of the localization effects by the domain wall.

Two modes of electron gas injection in resonant tunnelling through GaAs/AlGaAs double-barrier heterostructures were revealed while studying their current—voltage characteristics. Examining peculiarities of the characteristics within the temperature range 4-350 K and under a high magnetic field, we were able to distinguish the contribution to resonant tunnelling of ballistic electrons mjected from a three-dimensional electron gas in the emitter contact and that of electrons injected from a two-dimensional electron gas in the accumulation layer formed near the emitter barrier.

Fabrication of GaAs-based structures on Si substrates gives possibilities of easy integration of silicon electronics with optical devices. A major obstacle for a success of such technology are lattice defects, e.g. antiphase boundaries, twins, and dislocations, which are generated in GaAs by the 1attice mismatch to Si [1]. A substantial contribution of charged misfit dislocations to electronic vertical transport in those structures is then expected. In this work we investigated a system consisting of GaAs layer grown by molecular beam epitaxy on p+-type Si (for growth details see Ref. [2]). Formation of antiphase domains in that polar-on-nonpolar system was suppressed by simple misorienting the Si (001)-oriented substrate slightly toward the [011] direction. The nominally undoped GaAs epitaxial layer of 2 m thickness was n-type with Si concentration of 10 15 -10 16 cm-3 . Typically, as verified by secondary ίon mass spectroscopy (SIMS), the Si atoms move into the GaAs layer, and Ga atoms diffuse into the Si substrate. Moreover, a site exchange mechanism can occur by which Si donors on Ga sites move to As vacancies, thereby creating acceptor sites. This process is tremendously enhanced near the interface by the high density of lattice defects in heteroepitaxial layer. As a result the p-n junction is shifted to GaAs side with respect to the Si/GaAs interface [3]. The lattice constant of GaAs is about 4% larger than that of Si. This mismatch leads to the formation of a twodimensional network of misfit dislocations at

Here we present the appearance of the quantum Aharonov-Bohm (ΑB) interference [1] in a macroscopic system: semiconductor heterostructure containing an array of misfit dislocations. Let us consider a straight-line dislocation threading a heavily-doped semiconductor crystal. Localized states of the dislocation core (dangling bonds) accept the majority charge carriers from the bulk of material thus making the dislocation to be surrounded by a screening carrier-free cylinder. In GaAs with the carrier concentration of 1017 cm-3 the cylinder diameter is of the order of 100 nm that is presumably smaller than the phase-coherence length of the carrier wave function at liquid-helium temperatures. The charge carrier which meets the dislocation has to pass around the cylinder either clockwise or counter-clockwise. When a magnetic field, B, is applied parallel to the dislocation axis, these two alternative paths are non-equivalent since the vector potential of the magnetic field alters the wave-function phase differently on each path. As a result, the wave function behind the cylinder, which results from the interference of partial waves on these two paths, becomes a periodic function of the magnetic flux enclosed inside the carrier-free cylinder, so that the squared modulus of the wave function will exhibit the period Φ0 = h/e. The

Apart from usual quantization steps in the conductance of 1D systems, equal to G0 = 2e /h [1], an additional plateau-like feature appears at a fraction of about 0.7 below the rst conductance step in GaAs-based quantum point contacts (QPCs) [2, 3]. Despite a huge number of research works devoted to this anomalous feature, its origin remains still unclear [4]. Here, an unique explanation of this anomaly is proposed, which relies on the fundamental principles of quantum mechanics: superposition and interference of (spin) states.

A double-barrier resonant tunneling stucture (DBRTS) grown by molecular beam epitaxy at UMIST [1] was investigated. It consisted of two 8-nm Al 0 . 4 Ga0.6As barriers separated by a 4-nm GaAs well and had incorporated in the emitter region an n-doped three-period superlattice. Measurements were performed on mesa diodes of the area (50 μm)2 fabricated from such a structure. A 5 Ω resistor was connected in parallel with the device in a close proximity to the investigated structure. We measured current-voltage I(V) and photocurrent-voltage characteristics of the diodes in the temperature range 80-300 K. The photocurrent was excited by overall illumination of the stucture with a monochromatic light (λ ≈750 nm) chopped at a low frequency. Figure 1a shows a part of the 1(V) characteristic at 80 K. A "double-step" feature exhibiting a broad shoulder behind the resonance peak is seen in this figure. Such a feature has been commonly observed in characteristics of DBRTS's [2, 3], but its origin has not been satissfatorily explained until now. Foster et al. [4] have argued that it results from time-averaged current oscillations generated in the negative-differential-resistivity region of a diode.

The question: how much time it takes for an electron to tunnel through a barrier, is a basic problem in quantum mechanics, being today also of technological importance. Since 1932 [1] many different expressions for the tunnelng time have been derived [2]. Because time has not a status of a dynamical variable in quantum mechanics, and there is no operator for time, the definition of the tunneling time is not unequivocal. Usually one defines the tunneling time by following the peak of a wave packet transmitted through the barrier region [3]. However, such a definition leads to a physically unacceptable result: when the barrier width tends to infinity, an average speed of tunneling electron tends also to infinity. This is likely a manifestation of the fact that the transmitted packet cannot represent the tunneling electron by itself [4]. We proposed a novel definition of the tunneling time that does not use the concept of a wave packet [5]. In this short note the definition has to be, of necessity, introduced in a heuristic way and the aim of this note may only be to show that in the simplest case it leads to intuitively acceptable consequences. it should be emphasized, however, that both the foundation of this approach and its consequences are by no means so trivial as it might appear from this presentation. Instead, they may be helpful to understand how to picture electrons when they are tunneling through a barrier. Let us take into account a monoenergetic beam of electrons, having an energy E and a wave number k, tunneling through a rectangular potential barrier. Suppose that at some moment of time the barrier height, V0, is instantaneously changed by Δ V. If a change in the probability current from j to j + δi at the front wall of the barrier is followed by the same change at the back wall, but occurring at a time interval τ later, it means that a current carrier had to spend the time τ on traveling through the barrier region, and thus this τ has a meaning of the transit