TABLE OF CONTENTS Over twenty years ago the first gallium arsenide transistors which used a diffused-gate structure were reported by Turner et al (1967) giving useful gains in the lower megahertz frequency bands. In 1969 Middel-hoek realized a silicon Metal Semiconductor Field Effect Transistor (MESFET) with a 1 ?m gate length by projection masking which had a maximum frequency of oscillation of 12 GHz (Middelhoek, 1970a, b). This was comparable to the maximum frequency of oscillation, f max, of the best bipolar transistors at that time.
In 1971 a significant step was made by Turner et al, when 1 ?m gate length FETs on GaAs were made having f max equal to 50 GHz and useful gains up to 18 GHz.
The substantial improvement in FET performance over the silicon bipolar transistor is due mainly to two reasons:-
In gallium arsenide the conduction electrons have a six times larger mobility and twice the peak drift velocity of those in silicon (Ruch et al, 1970).
The active layer is grown on a semi-insulating GaAs substrate with resistivity larger than 10 7 ? cm. This compares with a typical value of 30 ? cm for intrinsic silicon.
The first property results in lower parasitic resistances, larger transconductances and shorter electron transit times.
The second property results in lower gate-bonding pad parasitic capacitance when the gate pad is on the semi-insulating (SI) substrate.
By the early 1970's it was clear that the gallium arsenide MESFET could be used...
