Skip to content
SASemiAgoraSemiconductor knowledgeResearch services
Menu
Research services한국어 안내LearnTraceCompareReproducePlanSearch
ExploreKnowledge AtlasDetailed mapSimulationsDomain mapGuided journeysLabs
WorkspacesScenario ComparisonReproducibility RegistryExperiment PlannerResearch Project BuilderClaim-Evidence WorkbenchData Contract
TrustEvidence Knowledge GraphMetric OntologyCitation ProvenanceSource WorkspaceEvidence BrowserPublic Release LedgerSource PolicyCorrections
MoreSearchPracticeCalculatorsResourcesResearch atlasRelease provenanceReading RoomResearch servicesAbout SemiAgora
SASemiAgoraSemiconductor knowledge & engineeringOpen run card

PROCESS REPLAY LAB / SA-INT-ETCH-DEPTH-043

Etch depth is a family of evidence questions, not one recipe knob.
Process / Equipment / Metrology / Yield

HAR transport and profile control, ALE cycle completeness, selective-etch differentiation, DRIE passivation balance, strip and ash residue removal, and wet-etch mass transfer remain separate while profile, selectivity, mask, residue, metrology, and role evidence are compared on one bounded atlas.

Open simulationStudy lesson ->
packetSA-INT-ETCH-DEPTH-043runSimulation ObjectlearnEvidence SequencemethodMethodslimitsBoundaries

EVIDENCE SEQUENCE

What the learner should inspect first.

This surface connects the daily-scan source signal to an original SemiAgora simulation object without copying source text, code, figures, PDK files, or process recipes.

01

Choose HAR, ALE, selective etch, DRIE, strip and ash, or wet etch before loading any row data; the initial atlas carries compact metadata only.

Open the run card to inspect the saved rows, methods, limits, and export files behind this step.

Inspect run card ->
02

Hold each family-specific mechanism and state axis fixed, then move through eight ordered teaching states without reading the order as time, depth, or recipe progression.

Open the run card to inspect the saved rows, methods, limits, and export files behind this step.

Inspect run card ->
03

Compare profile, selectivity, transport, mask, residue, surface, and metrology proxies within one family; do not compare unlike proxy names as a universal performance score.

Open the run card to inspect the saved rows, methods, limits, and export files behind this step.

Inspect run card ->
04

Read operator, process-engineer, integration, and yield gates as separate ownership questions while confirming that three exact integer ledgers close at zero residual.

Open the run card to inspect the saved rows, methods, limits, and export files behind this step.

Inspect run card ->
05

End with the owned measurements, materials, geometry, uncertainty, validation, genealogy, and accountable review needed to replace every categorical handoff.

Open the run card to inspect the saved rows, methods, limits, and export files behind this step.

Inspect run card ->

METHODS

Precomputed model contract.

The goal is useful study scaffolding and paper-planning discipline, not calibrated production evidence.

method

Six packet-complete deterministic Cartesian sweeps contribute 41,472 rows, 20,736 physical configurations, and paired teaching-assumption and replacement-slot evidence postures.

The simulation object keeps the same statement beside downloadable JSON and CSV evidence.

method

The browser receives compact metadata first and loads exactly one 11-12 MB family CSV after explicit selection; all six row-bearing JSON artifacts remain canonical evidence only.

The simulation object keeps the same statement beside downloadable JSON and CSV evidence.

method

Each family keeps its own normalized proxy equations and three exact integer teaching ledgers; the atlas aligns evidence dimensions and role questions without inventing a cross-family physical scale.

The simulation object keeps the same statement beside downloadable JSON and CSV evidence.

method

One hundred twenty link-only source records and ninety-six bounded claim nodes anchor terminology, rights, applicability, and replacement-evidence questions without supplying generated numbers.

The simulation object keeps the same statement beside downloadable JSON and CSV evidence.

LIMITS

Useful, but deliberately bounded.

No account, upload, payment, comment, newsletter, live solver, facility workflow, or product-release surface is enabled.

non-claim

No named chemistry, material stack, concentration, pressure, flow, power, bias, temperature, time, rate, depth, CD, angle, selectivity ratio, roughness, residue, damage, collapse, or endpoint value.

Keep this boundary attached when reusing the card for lesson writing, figure planning, or research notes.

non-claim

No equipment setpoint, gas or chemical delivery command, recipe, maintenance or safety instruction, process optimization, supplier ranking, wafer disposition, qualification, signoff, or production claim.

Keep this boundary attached when reusing the card for lesson writing, figure planning, or research notes.

non-claim

The atlas compares original dimensionless teaching constructs only; it does not solve plasma, surface, transport, feature-profile, liquid, or equipment physics and cannot establish causal or transferable process guidance.

Keep this boundary attached when reusing the card for lesson writing, figure planning, or research notes.

SOURCE ANCHOR

Link-only public source spine.

SemiAgora uses the source as a signal and citation anchor, then writes original summaries and original toy data with explicit limitations.

Official, primary, and DOI link-only metadata spines for six etch evidence families

https://science.osti.gov/-/media/fes/pdf/2023/DOE_FES_PlasmaScience_Semiconductors_Final.pdf

Open source ->