Search

We consider the stationary characteristics and high-frequency conductance of dual-gate field-effect transistors with a substantially inhomogeneous capacitive connection between gates and a current-conducting channel for an electron enhancement control mode. In most cases, such transistors are quasihomogeneous because their stationary JV-characteristics are not different from characteristics of perfectly homogeneous devices, which can be described by a certain effective channel length. But a nonstationary small-signal conductance of such transistors with the same stationary characteristics can be noticeably different since it is dependent on detailed space distributions of the capacitive inhomogeneity. In particular, a high-frequency gate current can depend on the stationary current direction in channels with symmetrical stationary characteristics. We discuss in detail the cases when a specific channel-gate capacitance increases or decreases monotonically along the channel current direction. Then high-frequency gate currents depend substantially on the stationary channel current direction (in spite of the same stationary JV-characteristics for both directions). [ABSTRACT FROM AUTHOR]

A very short-channel field-effect transistor (FET), which is controlled by a nonplanar prismoidal gate, is proposed and its stationary JV characteristics are derived. The prismoidal gate effectively depletes a thin electron channel in a very narrow strip around the gate apex. Interaction of the channel electrons with a controlling gate potential is additionally focused by a selected two-layer dielectric separating the channel and gate. It is shown that in the considered FETs there exist both “standard” regimes characterized by quasisaturation of the drain current jD, with an increase in the drain-source voltage uDS, and nonstandard regimes, with drooping jDuDS characteristics and a “self-blocking” current jD. We consider the dependence of the channel conductance on the gate curvature near the apex. [ABSTRACT FROM AUTHOR]

The first subject considered in the article is a donor center embedded in a thin heterostructural barrier separating a semiconductor medium into two halves. As a result of the small thickness of this barrier, the wave function of an electron bound by the donor center shifts almost completely into both halves of the surrounding semiconductor medium. The ground and first excited electron states of such a donor center are separated from each other by a narrow energy gap determined by the symmetric-antisymmetric tunnel splitting. Such structures can be implemented in both GaAs/AlXGa1-XAs and Si/GeXSi1-X material systems. The second considered subject is an exciton formed in analogous heterostructures when the staggered band alignment takes place between the heterobarrier and semiconductor medium. As a result of such band alignment, the hole participating in the exciton creation is located in the formed quantum well and the electron, which is the hole’s opponent, is separated into halves (on different sides of the quantum well) as before. Unlike the donor center, the exciton can be shifted and localized in arbitrary positions along the staggered “barrier-well” boundary by inhomogeneous electric fields of external controlling gates. [ABSTRACT FROM AUTHOR]

Electron wave tunneling through a rectangular heterostructural emitter barrier is considered in the case of a homogeneous high-frequency (hf) alternating electric field directed normal to the barrier interfaces. This hf field leads not only to the well-known increase in a stationary tunnel current through the emitter barrier, which is proportional to EB2 (where EB is the electric-field amplitude) but also to a linear (∼EB) increase in an alternating current (ac) through this barrier with the same frequency ω as the electric-field frequency. The ac is a sharp function of ω, which grows significantly with an increase in ω (typically in the terahertz range). In a certain intermediate current and frequency region, the above-mentioned increase in the ac is the dominating effect of the alternating field. Such an effect can be used to optimize tunnel barrier emitters for ballistic transit-time terahertz-range oscillators. [ABSTRACT FROM AUTHOR]

A very thin positively charged metallic filament separated from a surface of a semiconductor (S) by a thin nontunneling potential barrier (B) induces a quantum wire (QWr) in the semiconductor at the B/S interface. Single-electron quantum states of this QWr are controlled by a potential (and a charge) of the metallic filament. Two close parallel metallic filaments placed over such a B/S interface form a double-quantum wire with the ground and the first excited electron states, which appear as a result of a symmetric–antisymmetric splitting of the ground electron state in the single QWr. Two crossed metallic filaments, which are parallel to the B/S interface, form a quantum dot with completely localized electron states under the crossing point of the metallic filaments. The analogous crossing of a metallic filament by a pair of close metallic filaments forms a double-quantum dot (DQD). The latter can serve as a two-level qubit cell. Such qubits can be controlled by potentials of three independent metallic filaments inducing the above-mentioned DQD. Besides this “outside” metallic wire control, the DQDs can be connected to each other across the “inside” quantum wires, which have formed these DQDs by crossing. [ABSTRACT FROM AUTHOR]

A plane electron wave incident on a tunnel-transparent potential barrier formed by the potential V(x,t)=V0(x)+V1(x)cos ωt generates, in addition to the usual stationary transmitted and reflected stationary waves, also “transmitted” and “reflected” electron waves oscillating with the same frequency ω. The transmitted oscillating wave can serve as the basis for transit-time microwave generators oscillating in the terahertz range. (Such oscillators are ballistic analogs of the tunnel-emission transit-time diode oscillators suggested almost half a century ago.) In the special case of a rectangular potential barrier, we describe the dependence of a small transmitted oscillating wave amplitude on the frequency co and the value of V1(x). We consider two forms of V1(x): (1) homogeneous oscillation of the height of the rectangular barrier and (2) V1(x)-aδ(x-x1) [where δ(x) is the Dirac delta function and 0

We develop an analytic theory based on an earlier model of the admittance of a ballistic transit time diode terahertz oscillator with tunnel emission of electrons into a transit space. The focus of this work is on the actual case when electrons are injected with high enough energy to move from the start with maximal (saturated) ballistic velocity (∼1×10[sup 8] to 2×10[sup 8] cm/s). On the one hand, such diodes have maximal oscillation frequencies and, on the other hand, a simple analytic theory describes them and allows us to avoid a cumbersome numerical procedure, which characterizes the general case. Such a description is analogous to the description of oscillatory diodes with diffusive transport and saturated drift velocity. We have also considered a special case when a small part of the ballistic electrons crossing the transit space scatter into a diffusive subsystem with a small drift velocity. The appearance of such slow-drifting electrons substantially increases space charge in the transit space and influences the static JV-characteristic but the high-frequency admittance is almost invariable. © 2004 American Institute of Physics. [ABSTRACT FROM AUTHOR]

We have considered interactions between ballistic (or quasiballistic) electrons accelerated by a dc electric field in an undoped transit space (T space) and a small ultrahigh frequency ac electric field and have calculated the linear admittance of the T space. Electrons in the T space have a conventional, nonparabolic dispersion relation. After consideration of the simplest specific case when the current is limited by the space charge of the emitted electrons, we turned to an actual case when the current is limited by a heterostructural tunnel barrier (B barrier) separating the heavily doped cathode contact and the T space. We assumed that the B barrier is much thinner than the T space and both dc and ac voltages drop mainly across the T space. The emission tunnel current through the B barrier is determined by the electric field E(0) in the T space at the boundary B barrier/T space. The more substantial is, the tunnel current limitation the higher the electric field E(0) becomes. We have shown that for a space-charge limited current the change from parabolic dispersion to the nonparabolic branch induces narrowing and closing of the frequency windows of transit-time negative conductance starting with the lowest-frequency windows. These narrowing and closing frequency windows become effective only for very high voltages U across the T space: U>mV[sub S][sup 2]/2e, where m is the effective mass for the parabolic branch and V[sub S] is the saturated velocity for the nonparabolic branch. For moderate voltages U, the effects of nonparabolicity are not very substantial. The tunnel current limitation decreases the space-charge effects in the T space and diminishes the role of the detailed electron dispersion relation. As a result, restoration of the frequency windows of transit-time negative conductance and an increase in the value of this negative conductance occur. The implementation of the considered tunnel injection transit time oscillator diode promises to lead to... [ABSTRACT FROM AUTHOR]

A dispersion relation for an electron in a two-layer (and also multilayer) quantum well (QW) is formed as a result of a certain combination of initial dispersion relations for each of the forming layers. Such a combination can be used to engineer new dispersion relations with desirable properties. The same relates to a two-dimensional electron gas (2DEG) induced in a multilayer medium. In this study, we consider first such a 2DEG in a specific two-layer structure where a superlattice (SL) plays the role of the second half-infinite layer, and electrons with large wave numbers along the SL vector spread from the first ordinary QW layer to this SL. As a result of such a quantum (dynamic) real-space transfer, electrons become heavier, and the dispersion relation achieves an additional negative effective mass (NEM) section. Such NEM dispersion relations were studied for several different material systems, including the two most interesting three-material systems: (1) an isomorphic Al[sub 0.15]Ga[sub 0.85]As//GaAs/Al[sub 0.5]Ga[sub 0.5]As structure and (2) a strained In[sub 0.5]Ga[sub 0.47]As//In[sub x]Ga[sub 1-x]As/In[sub y]Al[sub 1-y] As structure (x> 0.53, y < 0.52) with a strain-balanced In[sub x]Ga[sub 1-x]As/In[sub y]Al[sub 1-y]As SL. Most of the results were verified using a simplified 1D model, but some of them were verified by more complicated 2D-model calculations. [ABSTRACT FROM AUTHOR]

Two groups of current carriers naturally coexist in ballistic and quasiballistic diodes and field-effect transistors (FETs): (1) traversing ballistic current carriers emitted by a source and absorbed by a drain, and (2) nontraversing (nonparticipating in a current flow) carriers that are in equilibrium with the drain carrier reservoir. Therefore, the convective two-stream instability develops in such diodes and FETs with appropriate physical and geometrical parameters. It can result in oscillatory regimes. In this article, we consider development of the two-stream instability in n[sup +]nn[sup +] diodes with a doped bulk n-base, n[sup +]nn[sup +] diodes with a modulation doped n-channel base, and also in diodes with a gated n-channel base (that is in ballistic FETs) where a gate potential controls electron concentration in the channel. Since oscillatory regimes in such devices are restricted by pair electron-electron interaction between electrons belonging to the two different streams and participating in the instability process, we suggest a new type FET with two parallel n channels. Current-carrying electrons from the primary channel interact with slow electrons from the parallel additional channel that simultaneously serves as a controlling gate for the primary channel. In this design, electron streams participating in the two-stream instability are spatially separated, and their pair interaction is suppressed. Along with analytic estimates, we present results of numerical simulations for the ballistic FETs that prove the existence of oscillatory regimes in terahertz frequency ranges. © 2000 American Institute of Physics. [ABSTRACT FROM AUTHOR]

Presents information on a study which examined the negative-effective-mass ballistic field-effect transistors. Ballistic carrier transport in diode structures; Ballistic carrier transport in gated structures; Numerical simulation procedure; Simulation results; Conclusions.

Presents a study which investigated the stationary distribution of ballistic current carriers in a thin doped base. Assessment of the negative effective mass part in the carrier dispersion law; Characteristics of current densities; Calculation of the carrier ballistic current in the quasiclassical approximation.

Examines the contributions of nonlocal mechanisms to nonlinear transport in semiconductors with emphasis on hot-electron emission at heterojunctions and its variations. Relationship of the formation of negative differential resistance region (N-NDR) branches of the current-voltage characteristics with hot-electron effects; Methods used to derive N-NDR; Steps in describing N-NDR by real-space transfer (RST); Types of devices based on RST.

There exist two regimes of the negative-effective-mass (NEM) instability in ballistic p[sup +]pp[sup +]-diodes with two-dimensional hole gas in the p-base: the instability of homogeneous NEM-hole distribution in a quasineutral plasma region, and the instability of a thin accumulation layer, which forms inside a depletion region and contains NEM holes. Both instabilities lead to the development of terahertz oscillatory regimes. The regimes' simulation in the inductance-loaded diodes with base lengths 0.05-0.15 μm demonstrates that such loads substantially enlarge the voltage range of the second regime and give rise to oscillatory regimes, which do not appear in unloaded diodes at all. Efficiencies of different oscillatory regimes are estimated. © 2000 American Institute of Physics. [ABSTRACT FROM AUTHOR]