July 7, 2014, IMEC Technology Forum, San Francisco—Wilfred Vandervorst described the challenges and science of measurements, and noted the great differences between lab and fab use modes and requirements. The latest tools are enabling new insights into fundamental physics and in ramping yields on the next-generations processes.
The labs tend to work on the nano-scale properties of materials whereas the fabs are looking for cost effective manufacturing tools that can provide data on process control and in-die performance monitors. The appropriate scale and metrology can enable yield, so the focal points and time scales of the equipment vary between lab and fab.
The ROI of metrology depends upon the use case. R & D areas want better and more accurate data, while the fab wants sufficient data to resolve yield issues. The R & D tools are capable of atomic-scale resolution and are limited by the tool suppliers at a single atom resolution.
The change in measurements are apparent in FinFETs; 1-D, 2-D, 2+-D and full 3-D require tools and techniques to image doping characteristics and compare differences in performance. The researchers are looking at atomic profiles to identify defects in lattice structures and doping profiles. This information correlates to differences in leakage, as more defects means more leakage. By labeling individual atoms, the images can show lattice dislocations and areas for stress buildup.
Creating partnerships with tool suppliers allows full 3-D imaging and tomography. A scalpel SSRM can scrape layers to measure conductivity on a per-layer basis. This data can then be used to modify the processing in manufacturing. One example of this level of work is the imaging of filaments in RRAM. A 2nm contact probe shows the possibility for a 2nm RRAM. The tools enable insight into fundamental functions in heterogeneous materials.
Manufacturing is looking for physics limits; volume, time to data, statistical relevance and the tools tend to be device limited. The fab uses an array of features to tune the processes. They still need 3-D imaging to track defects and other detailed phenomena for their work, but are working at a gross scale compared to the R & D work. A self-focused SIM depends on physics and the sample to define resolution. Raman imaging uses polarized light to identify features.
IMEC works with the tool suppliers, universities, and manufacturers to define the requirements of both groups and enable appropriate tools for each. They also collaborate with others in the ecosystem to develop standards and proof of concept test vehicles.
