Power conversion modules have been made of silicon for decades. So far, more and more power has been extracted, especially from power MOSFETS. However, the further increase in the efficiency of silicon is only possible with great economic effort. The transistor power loss can only be reduced with other materials. At the start is gallium nitride (GaN), which is used in LEDs. It will find its use in power electronics as a replacement for silicon.
The new material will be used in GaN HEMTs (Gallium Nitride High Electron Mobility Transistor). They are significantly better in terms of switching capacity and bandwidth. Although GaN-HEMTs have been known for some time, only since the silicon technology has been exhausted has research and implementation been invested.
Although GaN transistors have a relatively low dielectric strength, a relatively high frequency response and only a 3% better efficiency than silicon transistors. But one is only at the beginning of their development. One can therefore assume that there are still some things going on here in the future.
The best thing about GaN transistors is their similarity to power MOSFETs. Developers do not have to rethink. In addition, the threshold voltage is largely independent of the temperature, which greatly simplifies the development of circuits.
Advantage of Using
The advantages of gallium nitride reside in a very large 3.42 eV bandgap in the material combined with excellent transport properties in the channel of the field-effect transistor acting as a switch with motilities such as silicon. Another basic feature is the ability to grow heterostructures in addition to GaN using aluminum, resulting in lateral AlGaN / GaN switch devices, which are unusual for power electronics. Figure 1 shows the structure of a lateral power transistor as well as lateral, derived hetero-structure GaN Schottky diodes.
An alleged disadvantage for GaN arises initially from the fact that there are no volume crystals of sufficient diameter larger than 2 inches, on which epitaxy and power electronics technologies can be done. The growth of GaN on GaN substrates has thus far been reduced to experimental research.
Interesting for commercial applications, however, is the ability to grow GaN on silicon in the 111 orientation. This is not the orientation of CMOS technology, but it allows the application of Si-based technology approaches. The mismatch of lattice constants of AlGaN and GaN to silicon is considerable and requires strong efforts for high-quality growth. The ability to grow AlGaN or GaN heterostructures using the element aluminum allows particularly good control of a field-effect transistor channel in which the current is conducted.
The growth of the GaN layers on the silicon support must first be made possible by suitable growth conditions. A buffer layer compensates for the lattice constants of (Al) GaN on Si to compensate for the massive mismatch of 18%.
As noted, unlike conventional vertical transistor technologies such as Power MOSFETs or IGBTs, GaN transistors rely on a lateral HEMT (High Electron Mobility Transistor) technology. This means that the forward current does not flow through the bulk material, as it does with vertical devices, but laterally along with the AlGaN or GaN interface near the chip surface. This means that all connections (drain, gate and source) are on the chip top side and can be wired via conductor tracks.