As I have indicated in the previous post, after deciding on hafnia (HfO2) during the first stage of exploration (1st generation high-k gate dielectrics represented by transition metal oxides) the search for yet better performing gate dielectrics for MOSFETs continues. Among the rare earth metal oxides, which represent 2nd generation high-k gate dielectrics, gadolinium oxide,Gd2O3, also known as gadolinia, attracts more attention than other oxides in this class.
In terms of basic properties, gadolinia adequately meets requirements of an MOS gate dielectric, i.e. it displays k-value in the range 10-16, energy gap of 5.4 eV, a high conduction band offset with silicon, low leakage current, and a relatively high dielectric strength. What distinguishes gadolinium oxide among rare earth metal oxides is a set of other properties which in combination make it particularly promising as a gets dielectric on silicon.
Specifically, the cubic Gd2O3 has very close lattice-match to silicon; its lattice mismatch with Si(100) surface is within 0.5%. Hence, a cube-on-cube epitaxy between a Gd2O3 (100)-oriented crystal and the Si(100) surface, assuring very low interface state density is feasible. Furthermore, Gd2O3 is thermodynamically stable in contact with silicon at temperature exceeding 800 °C which means that formation of an interfacial oxide SiOx between Si and is Gd2O3 , which ruins EOT in the case of less thermally stable oxides on Si, can be prevented. As a result, in spite of the relatively modest dielectric constant, a Gd2O3 film thick enough to prevent excessive direct tunneling would still behave like a sub-1 nm SiO2.
All of the above is enough to put Gd2O3 in the spotlight. And, by the way, as some reports indicate, it also shows good promise in combination with GaAs.
Conclusion from these comments is that we are digging deeper and deeper into the periodic table to meet the needs of high-end semiconductor technology. As I said earlier (see post of January 12, 2010), this is an era of materials…..
