(Keynote) Metrology for Nanoscale Complex Semiconductor Systems

Abstract

Pushing the limits in IC-technology towards the nanometer scale, led to the development of complex systems and 3D-devices (like Finfets, TFET, nanowires) whereby novel materials and in particular interfacial interactions play a crucial role. These increasing technological challenges have boosted the importance of metrology and created a demand for concepts suited to probe very small volumes and enable atomic scale observations. One obvious solution is then to abandon 1D-metrology completely and to focus on metrology such as Atom probe tomography (APT) which is an extremely powerful method providing composition analysis within very small volumes (a few nm3) with high sensitivity and accuracy. Due to its excellent spatial and depth resolution (in many cases with the ability to resolve lattice planes and individual atoms) alloy composition analysis in small trenches, unexpected in diffusion in such volumes, dopant distribution and dopants decorating defects can in principle be identified. Nevertheless the presence of many materials with different evaporation fields does induce severe artefacts and trajectory aberrations which cannot always be corrected, causing severe measurement inaccuracies. Complementary to the resolving power of APT, is the application of scanning probes (spm, ssrm, c-afm) which enables to grasp the electrical activity of dopants or identify conduction paths within such volumes. As SPM is inherently a 2D-method, concepts for expanding into the depth dimension are explored (cfr Scalpel SPM, ion beam sputtering icw SPM,..). Applications in logic device engineering, failure analysis and memory cell development (RRAM, CBRAM, NAND) demonstrate the power of this approach. Moreover “standard” solutions like STM to image crystal structure and defects at the surface may need to be modified in view of the very close proximity of insulating regions.Despite their unique 3D-resolving power, APT and SPM suffer from a poor productivity and a lack of statistical averaging over large areas as required in more production oriented metrology. A solution can be found through ensemble measurements whereby spatial resolution is provided by the device under investigation and not by the probing beam. We will illustrate this concept through applications of “self focusing SIMS” which allows to determine the composition from trenches (SiGe, InGaAs,..) as small as 20 nm without having an ion beam with nm-resolution. Similarly crystallinity of III-V growth in narrow trenches (< 50 nm) can be obtained through channeling RBS whereby again we use a large beam but nevertheless probe the information from an array of very fine features. In all these cases, the averaging over a large array provides excellent statistics and in some cases dramatically improved productivity through the enhanced signal versus the case of a very focused probe beam. The latter is ultimately exemplified in Raman experiments on narrow SiGe-trenches where we demonstrate that the signal from very narrow features (20 nm) is significantly enhanced (50-100x) as compared to it blanket counterpart enabling to probe composition and structural properties from a small volume.