Negative differential conductance

Fig. 1.21. Tunneling spectroscopy in classic tunneling junctions, (a) If both electrodes are metallic, the HV curve is linear, (b) If one electrode has an energy gap, an edge occurs in the HV curve, (c) If both electrodes have energy gaps, two edges occur. A "negative differential conductance" exists. (After Giaever and Megerle, 1961).

Fig. 2. Measurement of G(V, B) for a 2 pm junction. Light shows positive and dark negative differential conductance. A smoothed background has been subtracted to emphasize the spectral peaks and the finite-size oscillations. The solid black lines are the expected dispersions of noninteracting electrons at the same electron densities as the lowest ID bands of the wires, ui) and li). The white lines are generated in a similar way but after rescaling the GaAs band-structure mass, and correspondingly the low-voltage slopes, by a factor of 0.7. Only the fines labeled by a, b, c, and d in the plot are found to trace out the visible peaks in G(V,B), with the fine d following the measured peak only at V > —10 mV.

Enormous progress has been achieved in the experimental realization of such nano-devices, we only mention the development of controllable single-molecule junctions [8]-[22] and scanning tunneling microscopy based techniques [23]— [44]. With their help, a plethora of interesting phenomena like rectification [18], negative differential conductance [9,35], Coulomb blockade [10,11,15,16,21, 23], Kondo effect [11,12], vibrational effects [10,13,14,16,21,25,31-33,35,36], and nanoscale memory effects [34,39,40,42,44], among others, have been demonstrated. 

Van Lien and Shklovskii, 1981), which are known to give rise to negative differential conductivity (Schmidlin, 1980 Van Lien and Shklovskii, 1981 Monroe, 1991 Gartstein et al., 1995). A key result of the work of Gartstein and... 

In the case of negative differential conductance, the current I decreases with increasing voltage U, and vice versa, which normally corresponds to an unstable situation. The actual electric response depends upon the attached circuit which in general contains - even in the absence of external load resistors - unavoidable resistive and reactive components like lead resistances, lead inductances, package inductances, and package capacitances. These... 

If a semiconductor element with negative differential conductance is operated in a reactive circuit, oscillatory instabilities may be induced by these reactive components, even if the relaxation time of the semiconductor is much smaller than that of the external circuit so that the semiconductor can be described by its stationary I U) characteristic and simply acts as a nonlinear resistor. Self-sustained semiconductor oscillations, where the semiconductor itself introduces an internal unstable temporal degree of freedom, must be distinguished from those circuit-induced oscillations. The self-sustained oscillations under time-independent external bias will be discussed in the following. Examples for internal degrees of freedom are the charge carrier density, or the electron temperature, or a junction capacitance within the device. Eq.(5.3) is then supplemented by a dynamic equation for this internal variable. It should be noted that the same class of models is also applicable to describe neural dynamics in the framework of the Hodgkin-Huxley equations [16]. 

Two important cases of negative differential conductivity (NDC) are described by an iV-shaped or an -shaped j (F) characteristic, and denoted by NNDC and SNDC, respectively. However, more complicated forms like Z-shaped, loop-shaped, or disconnected characteristics are also possible [15]. NNDC and SNDC are associated with voltage- or current-controlled instabilities, respectively. In the NNDC case the current density is a singlevalued function of the field, but the field is multivalued the F j) relation has three branches in a certain range of j. The SNDC case is complementary in the sense that F and j are interchanged. In case of NNDC, the NDC branch is often but not always - depending upon external circuit and boundary conditions - unstable against the formation of nonuniform field... 

Fig. 5.3-11 Left differential conductance as a function of the applied voltage for a 49-period superlattice at 20 K. Right schematic model to explain the 48 negative peaks in the differential conductance, (a) Zero bias (b) ground-state resonanttunneling conduction (c) first field localization, where resonant tunneling between the ground state and an adjacent excited state takes place (d) expansion of the high-field region by one additional quantum well. (After [3.62])...

In order to obtain devices with negative differential conductance, Kunze and Kowalsky (1988) have used ytterbium as a film gate. The low work function of Yb compared to the electron affinity of InAs increases the surface field. And, the low resistive tuimel junction serves as ohmic contact to the inversion layer. 

Among the two-terminal devices that can be imagined for UE [capacitors, inductors, rectifiers, negative differential resistance (NDR) devices], the simplest is a molecular wire, that is, a molecule capable of conducting electricity a nanoconductor or, equivalently, a nanoresistor. Even the most conductive of molecular wires has a minimum resistance. 

---