Ellipsometry of very thin Al2O3-HfO2 stacks for nano-memory applications

Abstract

Nowadays, the way towards manufacturing of advanced flash memory devices with conformal layers in stacks or laminates depends on their physical characterisation, i.e. the layer and interface thicknesses, density and composition. The transition dimension between thin and very thin stacks is in the range of few nanometres. The assumption that the oxide layer properties in the very thin stacks are constant in depth or are the same as for the bulk material values is not correct. We have applied ellipsometry (Variable Angle Spectral Ellipsometry, and Multiple Angle Incident Ellipsometry) with appropriate algorithms for data interpretation to investigate very thin Al2O3/HfO2 stacks (5 bioxide blocks) grown by ALD with a total thickness below 20 nm. Depth profiling was used in simulations as a complimentary tool. A quantitative determination of HfO2 volume fraction in the stack depth was obtained. An independent determination of the thicknesses and the composition of blocks in the stacks were achieved. Lorenz-Lorentz mixing model was applied for the estimation of dielectric properties of the different layers in the stack depth.