Etch Chamber Match Budget

# Etch Chamber Match Budget

An etch chamber match budget allocates the allowed difference in wafer results between a reference chamber and a candidate chamber. It defines the comparison metric, matching limit and evidence needed to establish that difference. Copying a recipe is not the same as matching its result.

The general plasma-chamber-matching topic already exists. This entry focuses on a measured comparison and its acceptance budget, rather than duplicating the full equipment discussion.

## Compare the same process response

Choose a specific result such as depth at defined sites, a defined CD or a profile parameter. Identify the reference, sample plan, wafer condition and measurement method. A depth comparison does not establish a CD or residue match.

Etch Chamber Match Budgetthe same recipe settings do not prove the same wafer resultComparable incoming wafersChamber ASame nominal recipe100 nmChamber BSame nominal recipe103 nmIllustrative measured sample mean depthTwo sample means are not a complete matching qualification · control wafer and measurement variation.

Chambers with the same nominal settings can differ in their physical state and response. Hardware condition, process history, maintenance and measurement effects are possible contributors. A result difference does not identify which contributor caused it; establish the cause before selecting a correction.

Incoming wafer differences can also affect the comparison. Use suitable comparable material and patterning, and design the run sequence to distinguish chamber effects from time, lot or wafer variation. A candidate processed much later than the reference may be confounded with drift.

## Separate product limits from matching limits

Use a calculated teaching example: the reference sample mean depth is 100 nm and the candidate sample mean is 103 nm. Their difference, defined here as candidate minus reference, is +3 nm. Both means lie inside an assumed product interval of 95-105 nm.

Product Pass Is Not a Match Passexample means pass 95-105 nm product limits, but differ by 3 nmIllustrative matching limit: |B - A| at most 2 nmB - A = +3 nm-4-20+2+3+4Difference between sampled mean depths (nm)Calculated example · uncertainty, sample size and spatial matching still require evaluation.

If the assumed matching requirement is an absolute mean-depth difference no greater than 2 nm, the +3 nm comparison fails that nominal requirement. Passing product limits and passing chamber-matching limits are different questions. Neither set of limits should be invented from the other after inspecting the data.

These sample means are illustrative. Means inside product limits do not prove that individual sites or wafers pass. The graph also omits measurement uncertainty and process variation, so it is not a statistical equivalence test or a complete chamber qualification.

## Define what the budget covers

State whether the matching requirement applies to a mean difference, a bounded site difference, a spatial map or a statistical population quantity. A single average can hide a center-to-edge mismatch if positive and negative differences cancel.

Measurement repeatability, tool-to-tool measurement differences, sampling and within-chamber variation affect the comparison. If two readings use the same measurement system, some errors may be correlated; if different systems are used, a tool offset can be mistaken for a chamber offset. Use an explicit measurement and variation model rather than adding every observed spread as an independent term.

A deterministic allowance and a standard-deviation-based allocation are not interchangeable. For a statistical claim, define the confidence, equivalence bounds and experimental design appropriate to the question. A small dataset cannot establish stable long-term matching merely because its two means are close.

## Check more than one scalar

Compare depth maps, CD, profile, mask loss and protected-layer integrity when they are relevant requirements. Include representative patterns and wafer positions. Specify chamber state, such as the relevant period after maintenance or conditioning, because matching may change over the chamber's process history.

A time adjustment might shift one mean while changing another response. Do not assume a percentage correction to time solves a multivariable mismatch. Any proposed adjustment needs verification on the actual module and its full acceptance criteria.

## Maintain the comparison over time

Keep a qualified reference and monitor the metrics that support interchangeability. Repeat the required checks after changes that can affect the process. A reference chamber can itself drift; labeling it the reference does not establish that it remains on target.

The useful match is a defined agreement in wafer results, maintained across the intended operating conditions. Keep the product window, chamber-comparison limit and measurement evidence separate and review them together.

## Sources and scope

  • [Lam Research — Virtual Sensor Innovation for Higher Productivity](https://newsroom.lamresearch.com/virtual-sensor-innovation-for-higher-productivity?blog=true): chamber matching and process-distribution context.
  • [Lam Research — Controlling Variability and Cost at 3 nm and Beyond](https://newsroom.lamresearch.com/Controlling-Variability-and-Cost-at-3-nm-and-Beyond?blog=true): equipment variability and chamber-matching challenges.
  • [NIST — Accuracy Considerations for Critical Dimension Semiconductor Metrology](https://www.nist.gov/publications/accuracy-considerations-critical-dimension-semiconductor-metrology): reference measurements and accuracy considerations.

Original diagrams and assumed depth limits. Values illustrate the distinction between product acceptance and chamber matching, not a qualified production matching plan.

Take etch chamber match budget further

Ask the copilot about this term, or have our engineers assess it against your process.

Ask CFSGPT about this → Request a technical assessment → Open the CFS app