+44 (0)24 7671 8970
More publications     •     Advertise with us     •     Contact us
Technical Insight

Magazine Feature
This article was originally featured in the edition:
Volume 28 Issue 7

Advancing Ga2O3 doping


Nitrous oxide, ammonia, and several carbon-containing molecules are strong contenders for doping gallium oxide. But which one’s the best?


There’s a lot to like about Ga2O3. Thanks to a bandgap of around 4.8 eV, this semiconductor promises to provide more efficient switching than today’s rising duo, SiC and GaN. What’s more, crystals of Ga2O3 can be formed from the melt, indicating the potential for low-cost production of ultra-wide bandgap devices.

However, the picture is not all rosy. In addition to concerns over a limited thermal conductivity for Ga2O3, there are significant issues associated with doping [1].

Today it seems that the chances of effective p-type doping in Ga2O3 are rather bleak, due to the small energy dispersion of the valance band and the large effective masses in valance band states. These inherent weaknesses are limiting Ga2O3 device architectures to those that are unipolar. Although researchers have investigated whether impurities such as magnesium, iron, and nitrogen could unlock the door to p-type conductivity, all attempts resulted in deep acceptors.

Such efforts are still valuable, though, because deep acceptors are beneficial for engineering Ga2O3 power devices. When Ga2O3 substrates are not intentionally doped, they still exhibit unintentional n-type conductivity, associated with significant levels of silicon, a background impurity. One can compensate for this by adding acceptor impurities, such as magnesium or iron, to the Ga2O3 melt – this enables the production of semi-insulating or highly resistive β-Ga2O3 substrates. The addition of deep acceptor impurities causes the equilibrium Fermi level to move away from the conduction band edge and closer to acceptor dopant states. This shift in the Fermi level allows a junction to form with adjacent n-type material, thus opening the door to the realization of potential barriers for voltage blocking. Ultimately, this could enable enhanced-mode Ga2O3 MOSFETs with an accepter-doped channel.