Etch Iso-Dense CD Bias Budget

# Etch Iso-Dense CD Bias Budget
An etch iso-dense CD bias budget defines how far an isolated feature's post-etch width may differ from a dense feature's width drawn at the same size, and the evidence needed to measure and bound that difference. It treats the layout around a feature as part of what sets its final critical dimension.
This budget does not restate the single-feature CD budget or the clearing-time loading budget, which are covered separately. It concentrates on one layout-dependent quantity: the CD gap between an isolated and a densely packed feature of the same drawn dimension after etch.
## See the layout in the result

Two openings drawn at the same width can finish the etch at different widths when one stands alone and the other sits in a dense array. The surrounding pattern changes the local supply of etchant and the removal of byproduct, so the same recipe transfers the two drawn widths differently.
The sketch places an isolated feature beside a dense group. It is not to scale and does not claim which side ends up wider. A published trench study notes that an isolated pattern can carry an equal or larger incoming width than a dense pattern and may enter the etch with a tapered profile, so the starting point already differs before the etch adds its own contribution.
The difference between the two post-etch CDs is the iso-dense bias. It is a property of the pair and the process together, not of either feature alone.
## Define the quantity precisely
| Element | What to fix | Why it matters |
|---|---|---|
| Drawn width | Same value for both features | Otherwise the comparison mixes design and process |
| Reference feature | Isolated or dense, stated explicitly | The sign of the bias depends on the choice |
| Checkpoint | After etch, after strip or after clean | Later steps can change the gap |
| Metrology | Same method and location for both | Technique differences can masquerade as bias |
A bias number is only meaningful once these are stated. A value quoted without its reference and checkpoint cannot be compared with another lab's number or with a later measurement.
## Build the budget around a matched pair

Measure a matched pair rather than a single feature. The isolated and dense CDs must come from the same drawn width, the same wafer where possible and the same checkpoint, so that the remaining difference is attributable to layout rather than to design or metrology.
Carry depth and profile alongside CD. The same study reports that iso-dense loading spans CD, depth and profile together, so a bias budget written only on top-down width can miss a depth or profile difference that matters to the next step.
Sample the pitches, densities and wafer positions the product uses. A gap measured between one isolated line and one dense array may not represent intermediate densities or edge-of-wafer positions.
## Act on the cause and check the limit against the next step
A reported trench investigation found that iso-to-dense CD bias loading could be reduced by tuning identified process knobs, which shows the gap is adjustable rather than fixed. It also shows that the fix is specific to that stack and recipe, found through designed experiments, not a universal setting.
The tolerable bias comes from the integration. A later liner, fill or contact may accept a range of CDs, or may require the isolated and dense features to match closely. There is no universal iso-dense bias that suits every layout, and a change that narrows the gap can affect rate, profile, selectivity and uniformity elsewhere.
## Close the budget with traceable evidence
Specify the drawn width, the reference feature, the sign, the measured pair, the depth and profile where relevant, and the checkpoint. Retain layout, recipe revision and chamber history so a later bias can be compared with the earlier one.
An iso-dense CD bias budget measures a matched pair and bounds the layout-driven gap between them. Measurement, reference definition and acceptance are separate questions, and each needs its own evidence.
## Sources and scope
- [A Study of Minimum Iso-Dense Etching CD Bias Loading Performance on Advanced Logic Trench Process (Lu et al., Applied Materials)](https://www.semiconchina.org/Semicon_China_Manager/upload/kindeditor/file/20250604/20250604115546_303.pdf): reports that iso-dense loading spans CD, depth and profile and that iso-to-dense CD bias loading can be tuned through identified process knobs.
- [Wikipedia — Etching (microfabrication)](https://en.wikipedia.org/wiki/Etching_(microfabrication)): defines etch bias and the difference between isotropic and anisotropic transfer that underlies layout-dependent width changes.
Original diagrams and explanatory synthesis. No universal iso-dense bias, direction or qualified recipe is implied.