Etch Mask Erosion Budget

# Etch Mask Erosion Budget

An etch mask erosion budget checks whether a patterned mask survives with enough thickness and the required lateral shape throughout pattern transfer. Vertical thinning, edge retreat and corner faceting are different changes. A mask that is still present can nevertheless transfer an unacceptable feature.

The earlier selectivity budget estimates thickness consumption. This entry adds the geometry of a retained mask line and its edges; it does not assume that a blanket-film selectivity ratio predicts the complete mask shape.

## Separate top loss from edge movement

Vertical erosion reduces the mask thickness at a defined location. Lateral retreat changes the position of its edge at a defined height. Faceting changes the corner shape, so an edge position measured near the top can differ from one measured near the bottom.

Etch Mask Erosion Budgeta mask can retain thickness while its edges changeOriginal mask outlineRemaining maskUnderlying materialThinningat the topEdge retreat measured at a defined heightCorner facetSchematic retained mask line · geometry exaggerated · not a target-profile prediction.

The dashed outline is the original mask and the solid shape is the remaining mask. The underlying material is shown only to locate the interface. The drawing does not predict a final target linewidth or an etched sidewall angle.

For a retained mask line with equal inward movement at both measured edges, its width decreases by twice the per-edge retreat. For a mask opening, retreat of the surrounding material can instead enlarge the opening. Always state whether the measured CD describes a material line or an empty opening.

## A two-dimensional survival example

Assume a retained mask line starts 50 nm thick and 120 nm wide at the specified measurement locations. Suppose the complete sequence consumes 25 nm of thickness and moves each measured edge inward by 3 nm. The remaining thickness is 50 - 25 = 25 nm, and the remaining line width is 120 - 2 × 3 = 114 nm.

Check Two Different Limitsremaining thickness does not prove that the mask width is acceptableDimensionStartingAfter etchMinimumMask thickness50 nm25 nm15 nmMask-line width120 nm114 nm110 nm25 nm thinning · 3 nm retreat at each edgeBoth checks pass only within this simplified example.Assumed limits and constant losses · corners, uncertainty and target CD need separate checks.

If the example requires at least 15 nm of thickness and at least 110 nm of mask width, those two checks pass under the assumptions. The unallocated differences are 10 nm in thickness and 4 nm in total width. They are different dimensions with different acceptance criteria; do not add them together as a single margin.

The numbers are arithmetic teaching values, not representative removal rates or a production specification. Faceting and roughness can make the thinnest point or narrowest relevant mask section more limiting than the two reported measurements.

## Use the full exposure history

Include the actual coating opening, breakthrough, main etch, landing and overetch steps during which the mask is exposed. A step that removes little thickness can still change the edge shape. Conversely, a central thickness measurement can miss loss concentrated near a corner.

If a constant-rate model is appropriate, thickness loss is the sum of the vertical loss in each step. A lateral-loss calculation needs separately qualified edge behavior. Do not substitute the vertical mask removal rate for a lateral rate without evidence. Material, angular exposure, surface modification and process conditions can affect the shape evolution.

A thicker starting mask can improve one thickness allowance, but it is not automatically a solution to edge fidelity. Its patterning, aspect ratio, transfer profile and eventual strip compatibility also matter.

## Relate mask geometry to the target

The final mask width is not automatically the final target width. The target responds to the evolving mask, its own surface chemistry, profile control and exposure history. Measure the target CD and profile at the intended stage rather than applying a final-mask correction as if it captured the entire transfer.

Treat asymmetric edge motion separately from total linewidth loss. A line can retain its width while shifting position if its two edges move differently. Edge placement, linewidth, roughness and thickness can therefore require separate checks.

## Verify the limiting shape

Compare defined pre-etch and post-etch mask measurements, then inspect relevant intermediate or limiting conditions where practical. Include dense and isolated patterns and appropriate wafer positions. Verify the target shape, residue and protected-material integrity alongside mask survival.

A useful mask budget protects the pattern as well as the remaining film. Passing a thickness check is one part of that evidence, not the complete result.

## Sources and scope

  • [Development of hard masks for reactive ion beam angled etching of diamond](https://arxiv.org/abs/2201.06416): primary research on mask choice and erosion in a specified angled-etch application.
  • [Oxford Instruments — Deep Reactive Ion Etching](https://plasma.oxinst.com/products/dsie/): process control and selectivity to masking materials.
  • [Lam Research — Etch Essentials](https://newsroom.lamresearch.com/etch-essentials-semiconductor-manufacturing): mask selectivity and profile-control context.

Original diagrams and calculated geometry example. Material-specific mask behavior and acceptance limits require qualification for the intended process.

Take etch mask erosion 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