next up previous contents
Next: 2.1 Limitations of the Channel Material InGaAs Up: Dissertation Helmut Brech Previous: 1 Introduction

2 The Principles of a HEMT
HEMTs are field effect transistors where the current flow between two ohmic contacts, source and drain, is controlled by a third contact, the gate. Most often the gate is a Schottky contact. In contrast to ion implanted MESFETs HEMTs are based on epitaxially grown layers with different band gaps Eg. A schematic cross section of a HEMT with a T­shaped gate is shown in  Figure 2.1.

Figure 2.1 Schematic cross section of a High Electron Mobility Transistor (HEMT). Depending on the use of GaAs or AlGaAs for the buffer layer the HEMT is called single heterojunction HEMT (SH­HEMT) or double heterojunction HEMT (DH­HEMT) respectively.

In the vicinity of a semiconductor heterojunction electrons are transferred from the material with the higher conduction band energy EC to the material with the lower EC where they can occupy a lower energy state. This can be a large number of electrons especially if the semiconductor with the high EC barrier is doped. Near the interface a two dimensional electron gas (2DEG), the channel, is created. This way it is possible to separate the electrons in the channel from their donor atoms which reduces Coulomb scattering and hence increases the mobility of the conducting electrons. If the channel is built only by a single heterojunction the electrons are penetrating into the buffer under the channel very easily where their mobility is usually lower and the control of the gate is poor. To keep the electrons in the channel a second energy barrier below the channel can be introduced by a material with a higher EC than the channel material.

In  Figure 2.2 the band gaps of the most important III-V semiconductors and the available substrates are shown. AlGaAs/InGaAs on GaAs substrate nowadays is the most widely used material system and will be investigated in this work. The principal advantages and disadvantages of HEMTs based on InP substrates will be addressed in the appropriate sections.

Figure 2.2 Lattice constant versus band gap of the most important semiconductors. The bold line represents the AlGaAs/InGaAs system. The lattice constant of GaAs and AlAs are very similar whereas the lattice constant of InGaAs is significantly larger for all In contents.

If two semiconductors with different band gap energies are joined together the difference is divided up into a band gap offset in the valence band DEV and a band gap offset in the conduction band DEC. One of the most common made assumptions for the AlGaAs/InGaAs material system is 40 % valence band offset and 60 % conduction band offset. This is only valid for Al contents below about 45 %. For higher Al contents the bandgap of AlGaAs changes from direct to indirect.

In  Figure 2.1 such an AlGaAs/InGaAs HEMT with a delta doped upper barrier layer is shown. The conduction band energy under the gate along the cutting line A­A' is shown in  Figure2.3. The conduction band of the channel relative to the Fermi level EF is determined by DEC, the doping level ND, the barrier height of the Schottky contact qFB, the gate to channel separation dGC, and the applied voltage on the gate VGS. To obtain high drain currents ID and high transconductance gm it is favorable to maximize qFB, ND, DEC, and to minimize dGC. If a homogeneously doped upper barrier layer is used qFB, ND, and dGC are directly related to each other. A decrease in dGC reduces the total doping in the barrier layer which shifts the threshold voltage (VT) to more positive values and thus reduces ID max. If ND is increased FB of the Schottky contact is reduced.

Figure 2.3 Conduction band diagram of a delta doped HEMT. The Fermi level EF and the quantum energy level Ee of the electrons in the channel are indicated by the dashed line.

This direct trade-off can be overcome if a delta doping is used. A delta doping in an (Al)GaAs layer can be realized by growing pure silicon for a short period of time within the growth of an undoped AlGaAs layer. This way ND is not reduced by a reduction of dGC. The sheet doping concentration can be adjusted by the amount of silicon incorporated in the crystal and the activation of the dopands. The activation depends on various parameters of MBE growth such as substrate temperature. The upper limit of activated sheet doping concentration is in the order of .

FB is about 650 meV and decreases if the separation between the delta doping and the gate gets below 10 nm. The achievable minimum dGC highly depends on the applied technology. Tight process control can yield .

The aim of the channel is to provide a high current density. The electron concentration is mainly determined by DEC as well as the doping concentration in the barrier layers and its distance to the channel. To increase DEC it is favorable both to reduce Eg of the channel material and to increase Eg of the barrier layers.

next up previous contents
Next: 2.1 Limitations of the Channel Material InGaAs Up: Dissertation Helmut Brech Previous: 1 Introduction

Helmut Brech