It is demonstrated that a significant improvement in the breakdown voltage of an SOI (silicon-on-insulator) MOSFET can be achieved by stress-induced damage to the source/body junction. The damage serves to degrade the injection efficiency of this junction and thus suppresses the parasitic lateral bipolar associated with source/body/drain. The experiment indicates the usefulness of source engineering to the latch problem

Fowler-Nordheim tunneling of electrons dominates the electron conduction mechanism in low temperature plasma grown silicon dioxide. Avalanche injection studies yield typical trap capture cross-sections of 10-15 and 10-17 cm2 following anodisation. If a high temperature post-oxidation anneal is used then the only capture cross-section observed has a value 10-18 cm2.

The effects of low temperature (≤ 600°C) annealing on the interface state density of plasma grown silicon dioxide are investigated. It has been found that the interface state density at midgap, Ditm, of plasma grown oxides can be an order of magnitude higher than that of thermally grown oxides after post-oxidation annealing (POA) treatment with SiO2/Si structure in forming gas. The differences in Ditm between the plasma grown oxide and the thermally grown oxide can be reduced by performing a post-metallization annealing (PMA) process. Thus, it would appear that there exist certain states at the plasma grown Si/SiO2 interface, which can be annealed out by atomic hydrogen, but not by molecular hydrogen, and these states are absent at the thermally grown oxide interface. The annealing rates of the PMA process decrease with the ambient pressure. A PMA treatment of the oxide in a RF hydrogen plasma can improve annealing efficiency.

Oxides have been grown on silicon using an oxygen/chlorine plasma at temperatures of 400°C and below. The electrical breakdown strength of the oxides together with the interface and bulk oxide properties after low-temperature annealing are considered. The electrical properties are shown to be more than adequate for device applications.

The noise behaviour of two types of electrometer is described. Values of the power spectral densities of the noise sources are derived from measurements, using a suitable theoretical model. For an MOS electrometer the input voltage noise density was found to have a value close to those reported elsewhere, although the dominant source of electrometer noise is shown to be a current noise source at the gate of the MOS transistor, which has not been extensively studied previously. In the case of the vibrating-capacitor electrometer a current noise associated with the vibrating capacitor was observed, the source of which is not clear. The amplitude distribution of the noise was also studied experimentally and found to be significantly non-Gaussian.

A multichannel system has been built to sample and store a single or repetitive waveform. In the repetitive mode the effects of noise on the signal can be reduced by averaging. The instrument dissects a 400 ns long signal of nominally 1 V peak amplitude through 21 sampling channels. In each channel the sample value is stored as an 8 bit digital word. The sampling can be repeated and added into store for signal averaging purposes up to 256 times. Readout is via either a CRO display or an X-Y recorder.

This paper describes the development of a closed loop servo-system designed to control gas throughput so as to maintain a constant pressure in a dynamic vacuum system. It is designed to work with pressure gauges having an electrical output signal and gas inlet needle valves adaptable for drive by a small servo motor. Traditional principles of operation are adopted, the gauge output signal after amplification being compared with a reference voltage. The difference between the two, controls the drive to the motor. The motor drive is in such a direction as to tend to maintain the pressure gauge output voltage equal to the reference. It is important that the system provides a rapid and stable closed loop response for widely varying types of operation with gauges having a fast response (e.g. ionization types) or slow response (e.g. Pirani type) and also with both long and short vacuum time constants.