A hardmask retains a pattern so that material is removed through selected openings. An etch-stop layer uses a difference in removal rates to protect an interface or provide margin for changing to another etch stage. Both may use silicon nitride, but a shared material does not imply a shared job. Locate the layer relative to the material being removed, then ask which boundary it must preserve.
Follow the opening to identify the hardmask
A fine resist image must survive transfer into a deeper dielectric. If its carrier is consumed too early, the opening can widen or lose edge fidelity. A hardmask provides a more durable pattern carrier, allowing imaging and deeper material removal to be optimized separately.
At 14nm step 165, SiN is deposited as a hardmask. The blanket layer has not yet been opened and is not a finished contact. Subsequent patterning records where etching is allowed. The visible job at this stage is preparing a pattern-transfer surface, not demonstrating endpoint detection.
Blanket SiN hardmask after deposition; contact patterning has not yet opened it.
Process checkpoint
Understand PECVD SiN Hardmask Deposition in context
Locate the unopened hardmask, then follow the transfer of an opening into the dielectric.
Process context for “Hardmask versus etch-stop layer: pattern placement and depth control”: 14nm FinFET · CONTACT_SC · Step 165
Follow the depth to identify the stop
A correctly placed opening still needs a controlled depth. When the target dielectric has been removed to an intended interface, a more slowly removed material can protect what lies below or provide margin for switching etch stages. A stop layer is not completely inert: prolonged exposure can consume it, and inadequate selectivity reduces its protective margin.
The step description for step 173 calls for pattern transfer, depth control, and protection of gate caps and spacers. Those are distinct boundaries within the same integration problem. Its current cross-section shows the start of etching in the upper dielectric; it does not demonstrate arrival at, or successful stopping on, a lower stop layer.
Etching begins in the upper dielectric; this image does not establish arrival at a lower stop layer.
Why not call every SiN layer an etch stop?
SiN can serve in a hardmask or in a protective device structure. Selectivity is relative to the material being removed and the chemistry used. Patterned SiN above the target can carry an opening, while retained material below can protect an interface. Material choice alone does not identify the role.
A CMP stop layer is also not evidence of plasma-etch stopping. Polishing and plasma etching remove materials through different processes, even when their descriptions both use the word “stop.”
Two questions for any cross-section
Ask which layer retains lateral placement, then ask where removal must stop or change stages. An incorrect opening width points toward pattern-transfer and mask-loss questions. Removal beyond the intended interface points toward depth and selectivity questions. The failures can interact, but they should not be treated as interchangeable.
Sources
References
Study of the SiO2-to-Si3N4 etch selectivity mechanism in inductively coupled fluorocarbon plasmas and a comparison with the SiO2-to-Si mechanism
M. Schaepkens, T. Standaert, N. R. Rueger, Pgm Patrick Sebel, G. Oehrlein, J. M. Cook