No abstract

A modified Young's double-slit experiment is realized for electrons in a high-mobility two-dimensional electron gas (2DEG). The observed quantum interference is employed to study dephasing by electronelectron interaction on length scales shorter than the elastic mean free path. It is found that the measured phase-breaking length agrees very well with theoretical calculations of the e-e mean free path in an ideal 2DEG. In contrast to the diffusive regime, dephasing occurs via e-e scattering events with an energy exchange on the order of the carrier excess energy.PACS numbers: 73.40.-c Electronic interference phenomena in solid-state systems and their destruction by electron-electron interaction have been extensively studied both theoretically and experimentally in recent years. Examples include weak localization, Aharonov-Bohm oscillations in small rings, universal conductance fluctuations in mesoscopic conductors, and the magnetoresistance of narrow wires. All these studies, however, were concerned with diffusive electronic motion where the elastic mean free path l e was the shortest relevant length in the system. In particular, in all of these experiments and corresponding theories, the e-e mean free path l e . e , as well as the phase-breaking length / 0 , has been much longer than l e . Here we are concerned with the opposite limit, the ballistic one (as far as phase breaking is concerned), where l e >: /^. Macroscopic transport in this case is expected to be purely classical and the experiment presented below was therefore carried out on length scales smaller than or comparable to both l e and 1$.In Fermi-liquid theory, the dephasing time is a key parameter that determines the quasiparticle lifetime. An experimental determination of this quantity, and its dependence on different parameters in the various regimes, is therefore of great importance in testing our theoretical understanding of an interacting two-dimensional electron gas (2DEG). A more practical motivation originates from the need to predict the constraints on operation of electronic devices based on quantum interference phenomena. For the diffusive case, l^le, it was shown by Altshuler, Aronov, and Khmelnitzkii l that for dimensionality equal to or lower than 2, and low enough temperatures, phase is lost due to e-e scattering events characterized self-consistently by an energy exchange on the order of h/r^ (T^, is the phase-breaking time). For 2D conductors, this energy is smaller than the temperature ksTby a factor kfle/lnikfle)^ 1 (k F is the Fermi wave vector). The quasiparticle energy is therefore a well-defined quantity as required for a quasiparticle description of a Fermi liquid. The resulting phase-breaking time is considerably shorter than the en-ergy relaxation time, which is governed by scattering events with an energy exchange on the order of kgT.The importance of small-energy scattering events for dephasing in ID wires was confirmed experimentally by Wind et at. 2 For the ballistic case, we are not aware of any detailed discussion o...

Graphene is currently at the forefront of cutting-edge science and technology due to exceptional electronic, optical, mechanical, and thermal properties. However, the absence of a sizeable band gap in graphene has been a major obstacle for application. To open and control a band gap in functionalized graphene, several gapping strategies have been developed. In particular, hydrogen plasma treatment has triggered a great scientific interest, because it has been known to be an efficient way to modify the surface of single-layered graphene and to apply for standard wafer-scale fabrication. Here we show a monolayer chemical-vapour-deposited graphene hydrogenated by indirect hydrogen plasma without structural defect and we demonstrate that a band gap can be tuned as wide as 3.9 eV by varying hydrogen coverage. We also show a hydrogenated graphene field-effect transistor, showing that on/off ratio changes over three orders of magnitude at room temperature.

We present a systematic investigation of the influence of cross geometry on the Hall effect in narrow ballistic wires. Various differently shaped cross regions have been fabricated, which demonstrate that near zero magnetic field the Hall resistance can be quenched, enhanced over its classical value, or even negative. A "last plateau" is seen in all devices, proving that its cause is not intimately linked to the quenching. A simple physical picture is presented showing how these effects come about from the scattering of electrons in such geometries. PACS numbers: 72.20.My, 73.50.Jt, 73.60.Br The quenching of the Hall effect is an intriguing result observed in narrow high-mobility devices by Roukes et al x and others. 2,3 As the magnetic field B is reduced, the Hall resistance RH (measured at voltage probes on either side of a narrow channel) forms a plateaulike feature (the "last plateau") and then drops sharply below its classical value so as to be close to zero ("quenched") in some finite region around B-0. This phenomenon has received considerable theoretical attention but it still lacks a definitive explanation. It has not been clear whether or not the last plateau and the quench are related, or indeed whether the effects are explicable solely in terms of single-particle transport theory. In this Letter, we describe measurements on several different cross geometries which yield alternatively quenched, enhanced, or negative RH-The last plateau, however, appears in all samples.It has been proposed that RH would be quenched if the channel width were less than the size of the edge states formed by the Lorentz force, 4 but the quantitative agreement with experiment was limited. 2 Very recently such narrow channels have been modeled quantum mechanically by a direct numerical solution of the Schrodinger equation together with the multiprobe resistance formula 5 to calculate RH from the scattering coefficients. 6 "" 9 A perfect cross with an infinite squarewell potential ("hard walls") quenches only for very specific values of the Fermi energy Ef, 1 contrary to experimental results, 2 which quench for a wide range of Ef. A similar model with a parabolic potential obtains ranges of E F where realistic-looking quenches occur. For more than two subbands, however, the quench has almost disappeared, whereas early devices 1,2 probably had between three and nine subbands populated when quenching was observed.New theoretical work 9 obtains generic quenching over wide ranges of E F and width by modifying the geometry and by energy averaging. The important change incorporated into the new model is collimation of the electrons as they enter the cross region. This is accomplished by adiabatic widening of the wire near the junction. Because the widening of the wire is gradual, the electrons at the Fermi surface do not populate the extra subbands available in the wider region, and because the subbands are depressed in the wider region, the ratio of the longitudinal momentum to the transverse momentum increases: The electrons ar...

Time-resolved and cw photoluminescence of excitons in coupled quantum wells with an applied electric field is modeled using a Fermi-Dirac distribution. This distribution can result from inhomogeneous broadening due to interface roughness and the strong, short-range electric dipole repulsion between excitons. The model quantitatively explains the striking temperature dependence of the luminescence linewidth and peak position previously interpreted as a phase transition to an ordered state [T. Fukuzawa, E. E. Mendez, and J. M. Hong, Phys. Rev. Lett. 64, 3066 (1990)]. At very low temperatures (;S6 K), the excitons are in a metastable distribution.PACS numbers: 73.20.Dx, 78.47.+p, 78.55.Cr, 78.65.Fa An exciton in a bulk semiconductor or a quantum structure is an electron and a hole which pair up because of their mutual Coulomb attraction. Because both electron and hole are fermions, the exciton is bosonlike. Recently, several experiments have been interpreted in terms of a possible condensation of excitons, in bulk CU2O (Refs. 1 and 2) and in coupled quantum wells. 3 In a symmetric coupled quantum well with an electric field applied along the growth direction, the electrons and holes separate. This separation reduces the overlap between their respective wave functions, and greatly increases the exciton radiative lifetime. 4 ' 5 Because of the longer lifetime, it has been suggested that there would be sufficient time for the excitons to come to a condensed, ordered phase. 6 The recent observation 3 of a sharply narrower linewidth of excitons in a coupled quantum well as the temperature was decreased from 12 to 6 K was interpreted as possible evidence for this condensation.On the other hand, the quantum-well structures grown by modern epitaxial techniques do not result in microscopically smooth heterointerfaces. 7,s It has been known for some time that this roughness leads to an inhomogeneous broadening of the excitonic absorption. 9 Here we show that the excitonic occupation of this inhomogeneous line can be modeled with a Fermi-Dirac distribution. (Using a Fermi-Dirac distribution does not imply that the excitons are fermions. For example, the equilibrium occupation of a set of impurity states follows a Fermi-Dirac distribution if each impurity can trap only a single particle.) Since isolated excitons have boson characteristics, the Fermi-Dirac distribution apparently results from a strong repulsion of the excitons. The repulsion must be large enough that two excitons do not occupy the same spatial position, yet short ranged enough to be neglected for spatially separated excitons. Excitons in coupled quantum wells in an electric field have an electric dipole along the growth direction, and we will show that the repulsion between the dipoles can satisfy these conditions. The p-i-n diode sample used here consists of ten undoped double-quantum-well units separated by 200 A of Alo.3Gao.7As. Each unit consists of two 50-A-wide GaAs quantum wells separated by a 40-A-thick Alo.3Gao.7As barrier. The layers were grown on a...

Techniques that can produce patterns with nanoscale details on surfaces have a central role in the development of new electronic, optical and magnetic devices and systems. High-energy ion irradiation can produce nanoscale patterns on ferromagnetic films by destroying the structure of layers or interfaces, but this approach can damage the film and introduce unwanted defects. Moreover, ferromagnetic nanostructures that have been patterned by ion irradiation often interfere with unpatterned regions through exchange interactions, which results in a loss of control over magnetization switching. Here, we demonstrate that low-energy proton irradiation can pattern an array of 100-nm-wide single ferromagnetic domains by reducing [Co(3)O(4)/Pd](10) (a paramagnetic oxide) to produce [Co/Pd](10) (a ferromagnetic metal). Moreover, there are no exchange interactions in the final superlattice, and the ions have a minimal impact on the overall structure, so the interfaces between alternate layers of cobalt (which are 0.6 nm thick) and palladium (1.0 nm) remain intact. This allows the reduced [Co/Pd](10) superlattice to produce a perpendicular magnetic anisotropy that is stronger than that observed in the metallic [Co/Pd](10) superlattices we prepared for reference. We also demonstrate that our non-destructive approach can reduce CoFe(2)O(4) to metallic CoFe.

No abstract

We have observed localized surface states (Tamm states) intentionally introduced in AlGaAs/GaAs superlattices by a terminating layer of AlAs. The formation of these states is manifested by excitonic interband transitions in photoluminescence excitation spectra. Critical confirmation is provided by photocurrent experiments under an electric field that show additional transitions as well as anticrossing interactions between the Tamm states and the Stark-ladder states associated with the superlattice.