A thickness number needs a fit and residuals.
Ellipsometry is useful only when the optical-stack assumption, fit score, and residual shape travel with the reported film thickness.
27.8 nm / 0.08 deg residual span / open simulation ->LESSON PACKET / SA-PROC-003
Process metrology.
Evidence before recipe.
Ellipsometry, sheet resistance, TLM, CD-SEM, overlay, SPC, and DOE belong together because a process claim becomes useful only when measurement method and uncertainty stay attached.
METROLOGY CHAIN
Turn measurements into reusable claims.
The MVP stays public-safe and recipe-free. It teaches which evidence belongs beside each process result: measurement method, fit residual, spatial map, extraction rule, control statistic, and DOE response surface.
A map beats a single electrical number.
Four-point probe data connects film, implant, anneal, and nonuniformity evidence. The mean is not enough unless center, edge, and spread are visible.
50.4 ohm/sq / +/-3.97% / open simulation ->The intercept tells a contact story.
TLM separates spacing-dependent sheet resistance from contact resistance. It turns raw I-V measurements into a geometry-aware extraction.
Rc 4 ohm/pad / Rs 80 ohm/sq / open simulation ->Linewidth and registration fail differently.
CD-SEM and overlay vectors keep pattern width, spatial variation, and layer-to-layer placement from collapsing into one vague lithography claim.
CD 100.61 nm / overlay p95 3.08 nm / open simulation ->Capability is spread plus centering.
Cp and Cpk separate potential from actual centered performance, while out-of-spec counts catch obvious excursions.
Cp 1.44 / Cpk 1.34 / open simulation ->A process window is a measured surface.
DOE keeps factor settings, center points, response tradeoffs, and curvature warnings attached before anyone calls a condition optimal.
64 nm/min / -110 MPa stress case / open simulation ->SIMULATION BUNDLE
Paper writers can start at the evidence card.
Each object keeps Explore, Run Card, Export, Methods, Limits, and Evidence in the same order, so future process lectures can scale without losing citation discipline.
Ellipsometry thickness fit
Replay a thin-film thickness fit and inspect residuals so a plotted film thickness is not separated from its optical-model assumptions.
Open Ellipsometry Fit ->Four-point probe sheet resistance
Replay a wafer-level sheet-resistance map and compute nonuniformity before treating a process split as electrically stable.
Open 4PP Sheet R ->TLM contact resistance extraction
Fit resistance versus contact spacing to separate sheet resistance, contact resistance, and geometry assumptions.
Open TLM Contact R ->CD-SEM and overlay control
Replay CD site statistics and overlay vectors so lithography evidence carries linewidth and registration together.
Open CD / Overlay ->SPC and process capability
Replay a CD sample set and compute Cp/Cpk so process spread, centering, and out-of-spec count stay distinct.
Open SPC / Cpk ->DOE response surface
Replay a two-factor process grid so deposition rate, film stress, center points, and curvature warnings travel together.
Open DOE / RSM ->SOURCE SPINE
Public references anchor the measurement vocabulary.
These sources support the teaching vocabulary. They do not authorize recipe copying, restricted manual redistribution, production qualification, or facility-operation claims.
NIST Thin Film RGTMs
https://www.nist.gov/programs-projects/thin-film-rgtms
Open source ->NIST dimensional metrology for nanoscale patterns
https://www.nist.gov/programs-projects/dimensional-metrology-nanoscale-patterns
Open source ->NIST 280DI sheet-resistance mapping system
https://www.nist.gov/laboratories/tools-instruments/nanofab-tool-four-dimensions-280di-sheet-resistance-mapping-system
Open source ->BYU Cleanroom contact resistance and TLM
https://cleanroom.byu.edu/contact_resistance
Open source ->NIST/SEMATECH process monitoring handbook
https://www.nist.gov/publications/nistsematech-engineering-statistics-handbook-chapter-6-process-or-product-monitoring
Open source ->NIST process capability
https://www.itl.nist.gov/div898/handbook/pmc/section1/pmc16.htm
Open source ->NIST/SEMATECH experimental design selection
https://www.itl.nist.gov/div898/handbook/pri/section3/pri33.htm
Open source ->DATA TO COLLECT NEXT
The next upgrade is measured, rights-clean metrology.
Useful process education does not need to expose confidential recipes. It needs trustworthy measurements, uncertainty, context, and clear permission to reuse figures and tables.
Ellipsometry or XRR thickness maps with calibration state, optical model, and uncertainty
Store the data with measurement method, uncertainty, source rights, process boundary, and non-claims.
Four-point probe maps with raw site coordinates, tool correction notes, and material context
Store the data with measurement method, uncertainty, source rights, process boundary, and non-claims.
TLM I-V sweeps with layout dimensions, fit residuals, and contact-stack boundary
Store the data with measurement method, uncertainty, source rights, process boundary, and non-claims.
CD-SEM sampling plans with CD distributions, overlay vectors, and measurement-system notes
Store the data with measurement method, uncertainty, source rights, process boundary, and non-claims.
SPC time-order datasets with subgroups, control limits, gauge studies, and excursion policy
Store the data with measurement method, uncertainty, source rights, process boundary, and non-claims.
DOE run matrices with randomization, replicates, response uncertainty, and residual analysis
Store the data with measurement method, uncertainty, source rights, process boundary, and non-claims.
Public boundary
No live tool operation, recipe upload, account workflow, payment, comment, newsletter signup, job/event submission, facility credential collection, or wafer release workflow is active in this MVP.