The metal-semiconductor field effect transistor (MESFET) is quite similar to the vastly more popular MOSFET. In the MOSFET structure, an insulating layer (usually silicon dioxide) lies between a semiconductor channel and a highly conductive gate that controls the conductivity in the channel. This insulating layer is absent in the MESFET structure; a metallic gate directly contacts the semiconductor substrate. However, a MOSFET without the insulating layer is not functional. It would look like this:
A nonoperative MESFET structure
| Metal |
| S | | D |
The junction between the metal gate and the semiconductor substrate is either an ohmic contact or a Schottky diode. In either case, the gate could not control the conductivity of the channel, because any carriers it induced in the channel would be swept away as gate current.
Therefore the MESFET requires an additional modification. A lightly-doped region of the same doping type as the source and drain is introduced under the metal gate. For convenience, I will assume the source and drain are n-type. The MESFET structure is the following:
The actual MESFET structure
Source | Metal | Drain
|n+ | n |n+ |
The metal is chosen such that it makes a Schottky diode with the semiconductor. Like a p-n junction, the Schottky junction contains a depletion region, a portion of the semiconductor near the junction that is depleted of current carriers (electrons and holes).
There are two types of MESFET's. If the depletion region does not extend all the way to the p-type substrate, the MESFET is a depletion-mode MESFET. A depletion-mode MESFET is conductive (on) when no gate-to-source voltage is applied and is turned off upon the application of a negative gate-to-source voltage, which increases the width of the depletion region such that it "pinches off" the channel. It turns out to be difficult to make digital circuits with depletion-mode transistors, since such circuits require two different power supply voltages.
In an enhancement-mode MESFET, the depletion region is wide enough to pinch off the channel without applied voltage. Therefore the enhancement-mode MESFET is naturally off. When a positive voltage is applied between the gate and source, the depletion region shrinks, and the channel becomes conductive. Unfortunately, a positive gate-to-source voltage puts the Schottky diode in forward bias, where a large current can flow. For this reason and because of fabrication difficulties, it is also difficult to make digital circuits from enhancement-mode MESFET's.
In other words, the MESFET is inferior to the MOSFET. So why even consider the MESFET? The answer is that some semiconductors with good electrical properties (e.g. gallium arsenide} do not grow high-quality insulating materials in the same way that silicon grows silicon dioxide. In the 80s and 90s there was considerable interest in gallium arsenide MESFETs, and Seymour Cray used them in his supercomputers. It now appears that silicon CMOS will remain the technology of choice for almost all integrated circuits.