The high electron mobility transistor (HEMT)

The High Electron Mobility Transistor or HEMT is a form of field effect transistor (FET) that is used to provide very high levels of performance at microwave frequencies. It offers a combination of low noise figure combined with the ability to operate at microwave frequencies.

The development of the HEMT took many years and it was not until many years after the basic FET was established that the HEMT appeared on the market. The specific mode of carrier transport used in HEMTs was first investigated in 1969, but it was not until 1980 that the first experimental devices were available. During the 1980s they started to be used, but in view of their initial very high cost their use was considerably limited. Now with their cost somewhat less, they are more widely used, even finding uses in the mobile handset markets.

Construction
The key element within a HEMT is thespecialised pn junction that it uses. It is known as a hetero-junction and consists of a junction that uses different materials either side fot he junction. The most common materials used aluminium gallium arsenide (AlGaAs) and gallium arsenide (GaAs). Gallium arsenide is generally used because it provides a high level of basic electron mobility and this is crucial to the operation of the device. Silicon has a much lower level of electron mobility and as a result it is never used in a HEMT.

There is a variety of different structures that can be used within a HEMT, but all use basically the same manufacturing processes.

First an intrinsic layer of gallium arsenide is set down on the semi-insulating gallium arsenide layer. This is only about one micron thick. about one micron thick is set down. Next a very thin layer between 30 and 60 Angstroms of intrinsic aluminium gallium arsenide is set down on top of this. Its purpose is to ensure the separation of the hetero-junction interface from the doped aluminium gallium arsenide region. This is critical if the high electron mobility is to be achieved. The doped layer of aluminium gallium arsenide about 500 Angstroms thick is set down above this as shown in the diagrams. Precise control of the thickness of this layer is required and special techniques are required for the control of this.

There are two main structures that are used. These are the self aligned ion implanted structure and the recess gate structure. In the case of the self aligned ion implanted structure the gate, drain and source are set down and are generally metallic contacts, although source and drain contacts may sometimes be made from germanium. The gate is generally made from titanium, and it forms a minute reverse biased junction similar to that of the GaAsFET.,/p>

For the recess gate structure another layer of n-type gallium arsenide is set down to enable the drain and source contacts to be made. Areas are etched as shown in the diagram. The thickness under the gate is also very critical since the threshold voltage of the FET is determined by this. The size of the gate, and hence the channel is very small. Typically the gate is only 0.25 microns or less, enabling the device to have a very good high frequency performance.

Operation
The operation of the HEMT is somewhat different to other types of FET and as a result it is able to give a very much improved performance over the standard junction or MOS FETs.

Electrons from the n-type region move through the crystal lattice and many remain close to the hetero-junction. These electrons for a layer that is only one layer thick forming what is known as a two dimensional electron gas. Within this region the electrons are able to move freely because there are no other donor electrons or other items with which electrons will collide and the mobility of the electrons in the gas is very high.

A bias applied to the gate formed as a schottky barrier diode is used to modulate the number of electrons in the channel formed from the 2 D electron gas and in turn this controls the conductivity of the device. This can be compared to the more traditional types of FET where the width of the channel is changed by the gate bias.

Applications
The HEMT was originally developed for high speed applications. It was only when the first devices were fabricated that it was discovered they exhibited a very low noise figure. This is related to the nature of the two dimensional electron gas and the fact that there are less electron collisions.

As a result of their noise performance they are widely used in low noise small signal amplifiers, power amplifiers, oscillators and mixers operating at frequencies up to 60 GHz and more and it is anticipated that ultimately devices will be widely available for frequencies up to about 100 GHz.