LOW DIMENSIONAL SYSTEMS The effect of a nonmonotonic potential profile on edge magnetic states E. B. Gorokhov, D. A. Romanov, S. A. Studenikin, and V. A. Tkachenko Institute of Semiconductor Physics, Russian Academy of Sciences (Siberian Department), 630090 Novosibirsk, Russia O. A. Tkachenko Novosibirsk State University, 630090 Novosibirsk, Russia Submitted May 5, 1997; accepted for publication May 20, 1997 Fiz. Tekh. Poluprovodn. 32, 1083–1088 September 1998 The dispersion law for electrons moving along a specularly reflecting boundary of a two- dimensional electron gas in the presence of a near-boundary potential well and a weak magnetic field is investigated theoretically. Numerical simulation is used to identify a number of features of the density of edge magnetic states that can be observed by magnetotransport and magnetooptics investigations. Ways to fabricate structures for studying these states are discussed. It is demonstrated that perfect-crystal interterrace boundaries can be created for a two- dimensional electron gas by introducing oblique slip planes into the heterostructure. © 1998 American Institute of Physics. S1063-78269801509-9 I. INTRODUCTION Edge magnetic states EMShave recently been success- fully used in recent times to explain various experiments on quasi- two-dimensional 2Delectron gases in strong mag- netic fields, in particular the integer-valued quantum Hall effect and Shubnikov-de Haas oscillations in semiconductor heterostructures. 1 In fact, EMS are the only current carrying states in the quantum-Hall-effect regime; therefore, they are responsible for all the phenomena of electron transport in quantizing magnetic fields. During the earlier stages of investigating EMS, the edge of the region where an electron gas exists was treated simply as a geometrical boundary. This was followed by inclusion of the smooth electrostatic barrier of the depletion layer, and features of electronic screening in the quantum-Hall-effect regime were discussed extensively in Refs. 2, 3, and 4. Until, there have been no further discussions of more complex na- tures or shapes of the potential near the boundary. This is not surprising, since at low temperatures and strong magnetic fields the current-carrying quasi-1D quasiparticles are pressed tightly against the boundary and can avoid scattering by fluctuation-induced obstacles in the potential around them. In the one-electron picture, regardless of the behavior of the potential near the boundary, the current in the final analysis is determined by the number of Landau levels below the Fermi level far from the boundary, where the potential and electron gas are assumed to be uniform, the simplest regime from a theoretical point of view. The situation changes radically in weaker magnetic fields, where scattering can no longer be ignored. In this case we can find the current along the boundary by solving the kinetic Boltzmann equation, in which we should use the dis- persion relation for carriers in edge magnetic states. This dispersion relation, in turn, is determined by solving the Schro ¨ dinger equation with a specific profile for the near- boundary potential. The roadmap to solving this problem taking into account collisions and the combined action of magnetic, electric, and high-frequency fields on an electron as applied to EMS has not yet been fully identified. How- ever, a number of useful conclusions from the point of view of experiment can be derived directly from analyzing the form of the dispersion relation, as we will show below by numerically simulating the edge current states. Here a relevant analogy with magnetic surface levels for three-dimensional crystals is worth mentioning. Magnetic surface levels, which were first observed as peculiar reso- nance peaks in the rf impedance of metals at very weak magnetic fields, 5 were later identified from their contribu- tions to a number of static effects in semiconductors. 6,7 As was shown, these latter effects can significantly affect the character of the surface potential. 8 Edge magnetic states in a two-dimensional electron gas have considerable advantages compared to the three-dimensional case with regard to fur- ther investigations and possible applicationsof these phe- nomena. Contemporary methods for growing GaAs/AlGaAs structures make it possible to obtain mean-free pathS for an electron in the 2D gas greater than 10 m. Methods available today for lithography allow us to create various shapes of the potential wells on scales smaller than the mean-free path of an electron. For example, by using a system of gate elec- trodes we can vary the depth and width of the potential well for a 2D electron gas in the lateral direction. This creates unique possibilities for controlling the dispersion relation of edge magnetic states by using a nonmonotonic profile for the edge potential. In other words, it is possible to ‘‘prepare’’ SEMICONDUCTORS VOLUME 32, NUMBER 9 SEPTEMBER 1998 970 1063-7826/98/32(9)/5/$15.00 © 1998 American Institute of Physics