Search

The electrical properties of thick-film vanadate-glass switches, with both planar and sandwich geometries, have been investigated using a remote double-pulse technique. The effect on delay time of pulse separation, overvoltage and polarity were examined. No forming step was required for any of the devices. Photographic evidence of filament formation confirmed the thermal model of switching suggested by these experiments. Additional evidence, such as the effect of temperature on threshold voltage and electrical conductivity, established the importance of the semiconductor-metal transition in the VO 2 present in the material for the switching action. A computer simulation based on the discontinuous change in electrical conductivity, occurring at 68°C, and due to this transition, was shown to be in broad agreement with the experimental facts.

Methods are described for making reproducible and robust vanadate-glass threshold switches both by a bulk and a thick-film technique; the latter method lendsitself to the large-scale production of the devices. The electrical characteristics obtained using dc, sine-wave and square-wave inputs are described and the various electrical parameters such as: threshold voltage, holding current, delay time, etc. are given. The results of a statistical analysis involving 30 alumina substrates with 12 switches on each are described, the population being analysed in terms of the standard deviation and mean deviation of the dc threshold voltage and the resistance of the off-state. An attempt is made to correlate the variation in parameters from device to device with variables involved in the furnace-firing schedule as well as the printing.

Thick-film, niobate-glass threshold switches have been made in reducing gas atmospheres at high temperatures (1100–1400°C) using SiO2 and B2O3 as glass formers. Their temperature characteristics are far superior to those of similar vanadate switches previously reported [1,2], and they are stable at low ON currents ( The process consists in pre-firing the printed, powdered glass in air and then reducing in argon/hydrogen gas. While the reducing conditions are not critical and virtually full reduction to tetragonal NbO2 occurs, the pre-firing conditions in air are critical if high-resistance devices are to be obtained. Many devices can be made on a single alumina chip and the results of a statistical analysis are given. The electrical characteristics — measured, chiefly, using a double-pulse technique to examine the change in delay time with pulse separation, ON current, pulse width, etc. — are similar to those previously reported for the vanadate switches [2]. These measurements, together with the results of a microscope and electron-microprobe examination, indicate that a thermal mechanism of switching with filament or channel formation is the most likely. A steady-state computer simulation based on the thermal model indicates the temperatures to be expected during switching, and these are calculated to be less than the NbO2 semiconductor-metal transition temperature of roughly 800°C, for all the cases examined.

Weight gain/time curves have been obtained for samples of beryllium sheet fabricated by a variety of different routes, in dry carbon dioxide at 500–1000° C, in oxygen at 600–1000° C and in carbon dioxide containing water vapour at 700° C for times up to 10 000 h. Protective films were formed on most materials in dry carbon dioxide at 1 atm pressure and temperatures up to 850° C. At 1000° C breakaway was obtained after 1500 h. The presence of 3 vol. % water vapour in carbon dioxide at 1 atm resulted in breakaway at 700° C. The subsequent linear rates of attack were markedly dependent upon the method used to fabricate the metal and differed by as much as a factor of 1000 in certain instances. Most of the differences in the behaviour of materials could not be related to impurity content but were consistent with the view that protection is associated with an intergranular network of oxide in certain batches of metal. No significant protection from breakaway in wet gas was provided by oxide films previously grown on the fabricated samples in dry carbon dioxide. The behaviour of metal in dry oxygen also varied with the method of fabrication. In some cases it was very similar to that in dry carbon dioxide, whilst in others breakaway was obtained at 700° and 850° C.