Begin with the residue’s history
Post-etch residue removal prepares an etched surface for the next interface. The unwanted material may come from altered resist, deliberately deposited sidewall passivation or etch products involving the exposed stack. These sources can coexist, but they are not interchangeable. “EKC” is a family of named cleaners; the label alone cannot identify the residue, prescribe its removal mechanism or prove compatibility.
The first question is therefore diagnostic: what remains, where is it, and what changed during etch or strip?
Process map
This step lives inside the 28nm Planar Flow course
Trace the real handoff from V1M2 etch through residue removal to barrier, seed and copper filling; distinguish surface cleaning from recovery of low-k damage. Full Step text requires 28nm course access; the step specifies no EKC formulation.
Account access is confirmed on the step page.
How residues arrive at different locations
Fluorocarbon dielectric etching can intentionally deposit sidewall polymer to help control the profile. Material that serves a useful role during etching can become unwanted before subsequent films are deposited. The cited study links removal of these deposits with the adhesion and coverage requirements of later materials.
Photoresist can also become chemically modified during plasma exposure. Residues may include an organic matrix with contributions from other materials encountered in the stack. The all-wet resist/BARC study found fluorine- and titanium-containing species in the resist crust after the particular hard-mask opening process it examined. This is evidence of a mixed residue in that experiment, not a universal composition for every etch.
Location changes the problem: a top-surface deposit, sidewall film and via-bottom material have different access and consequences. The node name alone cannot establish which residue exists or whether a wet clean is needed after every plasma operation.
Residue removal and damage repair are different
A remaining film is added or retained material. Plasma damage can instead be a change within the dielectric, including loss of carbon-containing groups and increased moisture affinity. Removing a visible coating does not necessarily restore the underlying material’s original dielectric properties. The low-k review treats damage and recovery as distinct issues.
Conversely, an apparently cleared cavity can result from etching away part of its supporting dielectric. That is material loss, not evidence of selective residue removal. A successful assessment must consider the retained profile and material state as well as the disappearance of the deposit.
Match the observation to the question
Cross-sectional microscopy can locate obvious deposits and dimensional changes. Surface-sensitive chemical measurements such as XPS can help distinguish fluorocarbon-rich material from the underlying surface, while vibrational spectroscopy can identify changes in relevant bonds. Each method has its own sampling depth and geometry; absence of a visible particle does not prove absence of a thin chemical residue.
Electrical or dielectric measurements answer another question: whether the structure still performs its intended function. The cited studies combine morphological, chemical and property measurements rather than treating one clean-looking image as a reliability result. Interpretation should compare the same stack and processing history.
The handoff from opening to surface preparation and fill
After a dielectric opening is formed, its sidewalls and bottom become interfaces for subsequent materials. Remaining deposits can interfere with adhesion or coverage. In a copper damascene sequence, barrier and seed formation precede electrochemical fill; those operations do not replace preparation of the receiving surface. CMP later removes excess field material rather than cleaning an already buried interface.
The current 28nm topology places Post ET Polymer Residue Removal after V1M2 etch and before M2 barrier-stack deposition, followed by copper seed and copper electrochemical plating. This locates a real preparation stage; it does not identify that stage as a particular EKC formulation. Cleaning behavior and acceptance still depend on the residue and stack.
What the existing teaching route shows
The public 28nm Flow overview locates dielectric opening and the transition to metal formation. Its dual-damascene topic compares V1M2 etch, M2 copper electrochemical plating and M2 copper CMP. This is a three-stop sequence view, not an uninterrupted manufacturing recipe: it omits the cleaning station and other intervening operations and displays no EKC formulation.
Entering the topic anonymously at its first stop opens two summary stops. An existing anonymous window remains tied to its original entry; changing the entry does not expand it. Free registration opens the full three-stop teaching, while detailed Step permissions remain independent. Anonymous overview reading may also be abbreviated; registration opens the full overview.
Choose a cleaning approach only after diagnosis
For the detailed surface-preparation explanation, open the 28nm course step. It follows V1M2 etch and leads to barrier and seed deposition, separating residue removal from low-k damage recovery. Full Step text requires 28nm course access; free registration alone does not unlock it. The step identifies no EKC formulation. The public overview and the existing three-stop topic remain sequence comparisons; the topic omits cleaning.
Once the residue and retained materials are understood, compare removal, metal compatibility, dielectric change and the rinse/dry handoff. The companion wet-cleaning article develops those selection questions. A mechanism demonstrated for one modified polymer or one chemical solution should remain bounded to its evidence, rather than becoming a universal EKC explanation.
References
Effect of UV Irradiation on Modification and Subsequent Wet Removal of Model and Post-Etch Fluorocarbon Residues
Q. Le, J. D. Marneffe, T. Conard, I. Vaesen, H. Struyf, G. Vereecke
Removal of Photoresist and BARC in Cu BEOL using an All-wet Process
Q. Le, A. Klipp, M. Lux, Yunlong Li, L. Zhao, G. Vereecke
Plasma processing of low-k dielectrics
M. Baklanov, J. D. Marneffe, D. Shamiryan, A. Urbanowicz, Hualiang Shi, T. Rakhimova et al.