Silicon dioxide (SiO2) is commonly used as the gate insulator in field effect transistors, however, it is now approaching its limits with regards to certain properties and, as a result, the semiconductor industry is in search of a replacement. These new materials require a higher dielectric constant (high-k) than SiO2 and must be compatible with the conducting silicon channels of the devices. However, many known high-k compounds are rigid ionic crystals with electronically active defects which preclude their use as gate materials.

Fig. 1: Transmission electron microscopy (TEM) images (20x20 nm) of the self-formed epitaxial apatite thin film on Si taken from two perpendicular directions. The TEM images show a, strong epitaxy and b, weak epitaxy directions.

Now, Dmitry Kukuruznyak and colleagues1 at the Advanced Electronic Materials Centre, Japan, report that rare earth aluminium-silicon apatites grown on a silicon substrate form a new class of high-k dielectric materials which have similar atomic-scale adaptability to native SiO2 and hence could be used as gate materials (Fig. 1).

The success of these high-k oxides is because of their ‘pliancy’ and ability to self-reconstruct the structural arrangement of the silicon substrate; essentially emulating the behaviour of SiO2.

The precursors to the rare earth aluminium-silicon apatites are solid oxides such as CeO2, Al2O3 and HfO2. Kukuruznyak and colleagues report on their discovery of the range of precursor mixtures that results in the desired properties. “We can mix the solid oxides in all possible proportions and make complex oxide combinatorial libraries of thin films with many thousands of different compositions on a single silicon wafer,” says Kukuruznyak. These films are then annealed at high temperatures, the precursors react with the silicon substrate and areas of epitaxial growth are located using X-ray diffraction. In these areas of epitaxial growth, the oxide and silicon have matching atomic arrangements in the interfacial region.

These ionic compounds which emulate the self-reconstructive properties of silicon dioxide show great promise for use in semiconductor devices, however, before they are incorporated into devices, optimization of the chemical composition, structure, processing conditions and electrical performance are necessary.

According to Kukuruznyak, the conditions in which these high-k oxides form are similar to those currently used in the semiconductor industry and hence manufacturers would not need to modify their equipment to make these next generation chips.