Hard Mask Strip

The hard-mask strip — ACL Strip or HM Removal depending on the maker — removes the amorphous carbon, poly, or nitride hard mask once it has finished protecting the real feature etch beneath it, closing the loop on the exact trade this masking scheme started with: a sacrificial layer survived the demanding etch so photoresist didn't have to, and now that job is finished, the sacrificial layer has to go too. Nothing about this step adds or defines geometry the way the etches before it did — its entire purpose is cleanly removing material that was never meant to be part of the finished chip, without disturbing the feature it was protecting.

Hard mask has finished protecting the real etch
active area, trench, or contact already formed
Strip method chosen for the material and the layer underneath
Dry ash
O2 plasma, for amorphous carbon
Wet strip
selective chemistry, for poly or nitride
Clean surface, feature fully revealed
ready for the next deposition or etch

Choosing dry ash versus wet strip is less about personal preference and more about what's directly underneath the hard mask at that point in the flow. An oxygen-plasma dry ash efficiently converts amorphous carbon to volatile byproducts and works well when the structures beneath it can tolerate plasma exposure, but a wet chemical strip is often preferred when the underlying layer is more sensitive to plasma damage or when complete removal matters more than throughput — DRAM process integration picks between the two specifically based on what sits one layer down, not based on which strip method is generically "better."

The Mask's Job Ends Here removing what was never meant to stay feature etched, mask still present hard mask residue on top strip mask removed, feature clean nothing left but the real feature the feature underneath has to come out exactly as the etch left it — the strip isn't allowed to change it

Residual hard-mask material left behind by an incomplete strip is a quiet, deferred problem rather than an immediately visible one. A trace of amorphous carbon or nitride that survives the strip doesn't announce itself at this step — it can interfere with the next deposition's adhesion, block a later contact from landing cleanly, or simply sit there as a particle risk for several process steps downstream before it ever shows up as a yield loss that has to be traced back through the flow to this exact cleanup step; this is why hard-mask strip gets inspected for completeness at nearly the same level of scrutiny as the feature etch it follows, even though it looks like a simple cleanup operation.

Because this step repeats after essentially every hard-mask-open-and-etch sequence in the flow, strip chemistry and the feature it's cleaning up are treated as a matched pair from the very beginning of process development. Ashing and wet-strip platforms from suppliers that also build the fab's etch tools are chosen specifically for selectivity against whatever sits beneath each particular hard mask at each particular point in the flow — the active area, the buried wordline, the storage node hole — meaning there isn't one universal hard-mask strip recipe used everywhere, there's a family of closely related strip steps each tuned to the layer it's protecting underneath.

Read the hard-mask strip through a *the-scaffold-comes-down* lens rather than a *just-cleaning-up* lens — every protective layer this session has described getting used to survive a demanding etch eventually has to be removed completely and without a trace, and how cleanly that removal happens is just as consequential to final yield as how well the mask protected the feature in the first place.

Take hard mask strip further

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