Atomic layer epitaxy of rare earth oxide films on GaAs(111)A and their device properties

Abstract

The aggressive scaling of MOSFETs has created interest in using high-mobility III-V channel materials to replace traditional strained Si. However, it has been challenging to form high- \(\kappa\) dielectrics that can passivate III-V surfaces with a low interface state density (Dit). We deposited \(LaLuO_3\) high- dielectric layer by ALD on sulfur-passivated GaAs substrates. The precursors lanthanum tris(N,N'-diisopropylformamidinate), and lutetium tris(N,N'-diethylformamidinate) reacted with water vapor at \(350^oC\). The compositional ratio of La:Lu was about 1:1 by using one cycle of \(La_2O_3\) followed by one cycle of \(Lu_2O_3\) in one complete cycle of \(LaLuO_3\). Both high-resolution XRD analysis and TEM showed that ALD \(LaLuO_3\) formed epitaxially on GaAs(111)A substrates, as shown in Figures 1 and 2, respectively. The epitaxial layer exhibited a cubic structure with a lattice constant smaller than GaAs by 3.8%. The \(LaLuO_3\) film had a high degree of crystalline perfection and was relaxed and not strained. Electrical characterizations showed the measured dielectric constant of around 30, which is close to its bulk crystalline value. The interface had a low interface state density \((D_{it})\) of \(\sim 7×10^{11} cm^{-2}eV^{-1}\). The amount of lattice mismatch can be engineered by choosing various rare-earth oxides. ALD La2O3 formed cube-on-cube epitaxy on GaAs(111)A with a lattice constant just +0.9% larger than that of the substrate. The mismatch can be reduced to zero by adding some \(Y_2O_3\) to the \(La_2O_3\), using yttrium tris(N,N'-diisopropylactamidinate)/\(H_2O\) cycles. Perfect zero-mismatched epitaxy was achieved on GaAs(111)A by depositing \(La_{1.7}Y_{0.3}O_3\), as shown in Figure 3. The effects of mismatch on the electrical properties of eip-\(LaYO_3\) on GaAs(111)A were studied. These results suggest that atomic layer epitaxy of rare-earth oxides/GaAs(111)A is a promising structure for future generations of high-power/high-frequency analog devices or high-speed logic devices.