The development of miniature circuitry, and particularly of the hybrid film circuit, has demanded more and more capacitance in a smaller volume of capacitor. There are three possible ways in which this can be achieved. One is to produce capacitors without wires or encapsulation or any of the other appendages previously required by equipment designers. Another is to use thinner layers of dielectric, but this is limited by the voltage that the capacitor must withstand. The third is the use of higher dielectric constant materials, but unfortunately it is generally found that the higher the dielectric constant of a material the poorer is its stability.

This last point is illustrated by comparing the change in capacitance with temperature of two typical materials: one having a nominal dielectric constant of 1200 (known as K1200) will change its value of K by as much as 20 per cent when its temperature is raised from −55°C (the lower limit stipulated by international specification) to room temperature, while a low dielectric constant ceramic having a K of, say, 30 – generally used for capacitors known as NPO for negative-positive-zero – will vary only by at most 0.25 per cent with the same change of temperature.

The monolithic multilayer ceramic capacitor is produced by screen printing on to strips of “green” ceramic of the barium titanate type an ink or paste incorporating finely divided platinum or a mixture of platinum group metals. A stack of these strips is then assembled and fired at about 1300°C to mature the substrate and fuse the layers into a block or chip. Metallising the ends of the chip with silver-palladium or gold-palladium then completes the contact between alternate electrodes. By this technique very thin dielectrics, which individually would be hopelessly fragile, may be used, while the fused structure is robust, hermetically sealed, and can be immediately incorporated into a hybrid circuit.

Ceramic substrates of many compositions can be used in monolithic chips, so that very high dielectric constants are available, but the stability of the capacitor clearly cannot be greater than that of the ceramic itself. The desire to use high K materials in the interests of miniaturisation can thus lead to difficulties because of the relatively poor stability of these materials. When subjected to high temperature processes such as soldering, for example, high K capacitors can show a drift in capacitance beyond acceptable limits.

Some recent work in the Electronics Laboratory of Matthey Printed Products was therefore designed to evaluate the change in capacitance a monolithic chip capacitor is likely to undergo during the assembly and subsequent processing of a hybrid microcircuit, and to provide the user with information that will allow him to anticipate this behaviour and make allowance for it. The detailed results of this work, reported at the recent Inter-Nepcon 70 Conference at Brighton, showed that, while the variations are predictable and repeatable and in some circumstances reversible, because of the diversity of techniques employed in the industry it would be unrealistic for a manufacturer of ceramic chips to specify a time by which a chip will recover its original capacitance after a soldering operation or any other heating cycle.

It was found, however, that the application of high voltages can accelerate the ageing and so stabilise the capacitance.

Further development along these lines should result in components having a limited temperature range in which they will operate satisfactorily, but having at the same time an unrivalled combination of stability and capacitance/volume ratio.