A. B. Fowler

Hot Electron Effects and Saturation Velocities in Silicon Inversion Layers

Carrier mobility variations were measured for electrons in silicon inversion layers. Near liquid‐helium temperatures and with electron densities less than about 2×1012 cm−2, the mobility increased with field and temperature. Correlation of the electron temperature with field indicates that ΔT varies with the field approximately as F3/2. Temperature dependence of the amplitude of the oscillatory magnetoconductance gave similar consistent results. At very high fields, the mobility decreased at all temperatures and electron densities. The electron drift velocity was measured for different surface orientations, substrate dopings, and ambient temperatures from 4.2° to 300°K. The drift velocity was found to saturate for fields greater than a few times 104 V/cm, depending upon the low field mobility, and was found to be lower than the reported bulk values. For (100), (111), and (110) surfaces, the limiting velocities at 300°K are (6.5±0.5) × 106, (5.5±0.5) × 106, and (4.0±0.5) × 106 cm/sec, respectively.

Temperature dependence of scattering in the inversion layer

Abstract Electron mobility has been measured in the silicon inversion layers of devices whose gate oxides were deliberately contaminated with Na+ ions. The mobility was measured as a function of inversion layer carrier density (1.0 × 1012 ⩽ NINV ⩽ 1.2 × 1013 cm−2), interfacial Na+ ion density (8.3 × 1010 ⩽ NOX ⩽ 2.02 × 1012 cm−2), substrate bias voltage (−15 ⩽ Vsub ⩽ +0.5 V) and temperature (4.2 ⩽ T ⩽ 80 K). From the data it is possible to separate the contributions to the mobility arising from ionized impurity scattering, surface roughness scattering and phonon scattering. The surface roughness scattering was found to be independent of temperature. The ionized impurity scattering was found to have components due to both single and multiple scattering processes, each of which were temperature dependent. Below 32 K there was no measureable phonon scattering.

Impurity bands in inversion layers

Abstract Impurity bands have been created and observed in inversion layers on silicon by drifting Na+ ions close to the SiSiO2 interface. The resulting hydrogen-like states can be made dense enough so that a peak in the conductivity can be observed at gate fields below the threshold for conduction in the conduction band. Three modes of conduction were observed: activation to the mobility edge, nearest neighbor hopping, and variable range hopping. From these measurements and by varying the density of ions information about binding energy, impurity band width, and the extent of the bound wave functions as a function of the electron density, ion concentration and substrate bias may be inferred.

One-dimensional conductance in silicon mosfet's

Abstract A review is given of experiments on the conductance of 1D MOSFETs. The types of samples studied, the phenomena observed and our theoretical understanding of these phenomena are discussed. Particular attention is given to the strong localization regime and the structure in the conductance as a function of gate voltage.

Thin-film circuit technology part III — Active thin-film devices

The insulated-gate thin-film transistor is the active device under the most intensive investigation today. Solutions to its problems should simplify the future development of other types of active thin-film devices The first article in this series1 discussed thin-film circuits and the considerable success that has been achieved in the fabrication of inactive circuit elements by evaporative techniques. However, the thin-film technology will not be complete until active devices can be made compatibly with the inactive ones. At present the active elements — usually transistors — are soldered on the circuit boards. The usefulness of evaporated circuitry would be immeasurably increased if the active devices themselves were also evaporated.

Fabrication of quantum devices in metals and semiconductors

Advances in nanolithography using electron‐beam techniques have allowed a great variety of quantum devices to be fabricated and tested. The critical dimensions governing the design of these devices will be discussed. Fabrication and experimental results of several such devices for studies will be reported. These include structures aiming at the observation of the electrostatic Aharonov–Bohm effect and nonlocal oscillations, superconducting weak links, and devices for the investigation of quantum scattering effects.

Magnetic field effects in strongly localized quasi-1D MOSFETs

Abstract We have measured the conductance as a function of gate voltage magnetic field, field orientation, and temperature of a quasi-1D silicon MOSFET in the strongly localized regime. At low temperatures and gate voltages, small changes in gate voltage produce large conductance fluctuations. The pattern of the fluctuations is changed dramatically as the magnetic field is increased, with some peaks vanishing and others growing out of the background. Much of the field dependence of the peak positions for different orientations and temperatures are due to Zeeman energy shifts.

On the effective mass and collision time of (100) Si surface electrons

Abstract We have measured the apparent mass of (100) Si surface electrons by thermal broadening of SdH oscillations of inversion and accumulation layers for a wide range of substrate doping level, geometry, interface charge density, magnetic field strength and substrate bias. It is shown that the apparent mass values determined in this way are very sensitive to these parameters and others which are not yet understood. Collision broadening of the SdH oscillations has been studied in terms of scattering time τ and compared with zero field mobility as well as average magnetoconductance relaxation times. Fair agreements were obtained between these quantities.

Magnetic field dependence of 2D sodium impurity band conduction in activated regions

We have measured the temperature dependence of conductance of electrons in 2D sodium induced impurity bands in silicon inversion layers from 4.2 to 80 K observing both activation of electrons to a mobility edge. E1, and nearest neighbor hopping, E3 in magnetic fields up to 25 T. The results are that E1 decreases slightly between 0 and about 20 T and then increases slightly. The prefactor decreases by about a factor of four. E3 increases monotonically. None of these results was expected.

Pressure Dependence of Barrier Heights in Ge–GaAs n—n Heterojunctions

The pressure dependence of barrier heights in Ge–GaAs n—n heterojunctions has been determined, for hydrostatic pressures up to 28 kbar, by means of electrical measurements. The results, expressed in terms of the pressure coefficient of the conduction‐band discontinuity ΔEc, are d(ΔEc)/dP = (6.7±1.0) meV/kbar for junctions having both (100) and (110) orientations. This result is seen to be in approximate agreement with predictions based upon considerations of bulk band energies only. It is concluded that the pressure dependence of any dipole‐layer contribution to ΔEc can be ignored. The results for d(ΔEc)/dP can then be used to obtain a value of ‐ (6.8±1.0) eV for the dilatational deformation potential for the conduction band of GaAs.