Book a demo

Cut patent&paper research from weeks to hours with PatSnap Eureka AI!

Try now

Atomic Layer Etching Sub-5nm — PatSnap Eureka

Atomic Layer Etching Sub-5nm — PatSnap Eureka
Sub-5nm Semiconductor Patterning

Atomic Layer Etching: Selectivity at Sub-5nm Patterning

Conventional RIE cannot achieve the surface binding energy discrimination or CD uniformity demanded at sub-5nm nodes. ALE's self-limiting sequential reactions deliver atomic-scale material removal precision — and the patent landscape reveals exactly how the industry is deploying it.

ALE Self-Limiting Cycle — Two-Step Process

Each ALE cycle converts only the outermost atomic layer, then volatilises it — leaving bulk material untouched.

ALE Self-Limiting Two-Step Cycle: Step 1 Surface Modification converts outermost monolayer; Step 2 Selective Removal volatilises only the modified layer; bulk material remains intact after each cycle. Schematic of the ALE sequential self-limiting half-reaction cycle as described in Lam Research patent filings. The self-limitation means only the chemically converted monolayer is removed per cycle, providing inherent cross-material selectivity based on surface binding energy differences. Bulk Material Outermost Layer STEP 1 Surface Modification STEP 2 Selective Removal Halogen / Oxidant Modified layer only Volatile by-product SELF-LIMITING — reaction terminates spontaneously
50+
Patents analysed across CN, US, KR, TW, JP, WO, SG
7
Jurisdictions covered in the ALE patent corpus
15nm
Feature width threshold for passivation-assisted ALE
1 ML
Minimum etch-stop layer thickness achievable with ALE
Fundamental Mechanisms

Self-Limiting Surface Chemistry — The Foundation of ALE Selectivity

ALE is distinguished from continuous etching processes by its reliance on sequential, self-limiting surface reactions. In each cycle, a surface modification step chemically converts only the outermost atomic layer(s) of the target material into a weakened intermediate — a "modified layer" — while leaving the bulk unchanged. A subsequent removal step then volatilises exclusively this modified layer, leaving a pristine, unreacted surface for the next cycle.

This self-limitation is the primary source of selectivity: only the modified species are susceptible to removal, and once the modified layer is consumed, the reaction terminates spontaneously. According to PatSnap's patent analytics, Lam Research's directionality-control architecture — filed 2017–2025 across US, CN, TW, KR, SG, and WO jurisdictions — establishes that a bias voltage applied during surface modification controls the conversion depth to at least one monolayer, while the removal step exploits a ligand-exchange reaction to volatilise the modified layer.

The energy framework underlying selectivity is most rigorously quantified in the "Designer ALE" family of patents. The method requires that process conditions for surface modification produce an energy less than the modification energy but greater than the desorption energy, and the removal step must produce an energy greater than the desorption energy but less than the surface binding energy of the material to prevent physical sputtering. By "calculating synergy" across these energy windows, engineers can maximise the process window for ALE and rationally design systems where one material is etched while an adjacent material — with a different surface binding energy — remains untouched. This atomic energy-based framework has been validated on semiconductors (Si, Ge, SiGe, GaN), metals (W, Co, Cu, Ta), dielectrics (SiO2, SiN), and hard masks. Learn more about advanced materials innovation intelligence from PatSnap.

Thermal ALE — the fully isotropic variant — achieves selectivity through sequential gas-phase reactions without plasma. As demonstrated by the University of Colorado, the method involves contacting a metal oxide substrate first with an oxidant to convert the surface, then with a fluorinating agent, with both steps being self-limiting at the monolayer level. Rate enhancement is possible through ligand-exchange reactions in which a metal precursor (e.g., Sn(acac)2 or trimethylaluminum) accepts fluorine from the metal fluoride surface and transfers its ligands to form volatile species such as AlF(acac)2 or AlF(CH3)2.

≥1 ML
Modification depth controlled per ALE cycle via bias voltage
3–5nm
Stop layer thickness RIE requires vs. ~1 monolayer for ALE
3 modes
Plasma (anisotropic), thermal (isotropic), hybrid wet/gas-phase ALE
6+
Material classes validated: Si, Ge, SiGe, GaN, metals, dielectrics
  • Surface modification converts only the outermost monolayer
  • Removal step volatilises modified layer exclusively
  • Reaction terminates spontaneously once modified layer consumed
  • Selectivity derived from surface binding energy differences
  • Plasma treatment eliminates residues after removal step
Selectivity Applications

How ALE Delivers Cross-Material Selectivity at Sub-5nm

Four critical selectivity mechanisms drawn from the patent corpus, each addressing a distinct failure mode of conventional RIE at advanced nodes.

EUV Integration

Si-Selective ALE with Metal-Oxide EUV Resist

Tokyo Electron's 2026 filing demonstrates ALE selectively etching Si relative to a metal-oxide photoresist layer formed by EUV lithography. The ALE cycle exposes Si to a halogen-containing gas (CF4 or NF3) without plasma to form a modified Si surface layer, then removes it using an Ar-containing plasma, while the metal-oxide resist — having distinct chemistry — resists attack. An alternative variant pre-treats Si with H-containing plasma to reduce native oxide before fluorination.

Angstrom-level pattern fidelity
Etch-Stop Control

Monolayer-Thin Etch-Stop Layers

Tokyo Electron's 2022 filing demonstrates that when using an ALE process, a stop layer as thin as approximately one monolayer is sufficient for the etch to self-terminate at the interface between the second film and the first film. This is impossible with RIE, where non-uniformities and micro-loading effects cause the etch front to penetrate stop layers of 3–5nm thickness. The mechanism exploits the chemistry-dependent nature of ALE modification: once the dissimilar stop material is encountered, the surface modification step simply does not produce a reactive intermediate.

Critical for GAA gate dielectric patterning
High-Aspect-Ratio Features

Passivation-Assisted ALE for Sub-15nm Features

For features with widths of 15nm or smaller, Lam Research's 2026 filing adds a passivation step in which the substrate is exposed to a passivating agent plasma to coat sidewall surfaces, preventing lateral modification and limiting active etching to the bottom surface only. This sidewall passivation selectively protects already-etched surfaces while the directional ion plasma continues to drive material removal at the feature floor, avoiding depth-loading entirely.

Eliminates depth loading in narrow features
Low-k Dielectric BEOL

RIE Lag Compensation via Quasi-ALE

Tokyo Electron's 2023 filing demonstrates a quasi-ALE (Q-ALE) process exhibiting reverse etch lag combined with a conventional RIE process exhibiting forward lag; the net combination provides a desired, controllable etch lag across all CD values. The Q-ALE process operates as a cyclic deposition-purge-etch sequence that, through its deposition sub-step, preferentially loads narrow features first, producing the reverse lag needed to compensate. This inter-process selectivity — controlling etch rate as a function of CD — cannot be replicated by continuous etch approaches.

Unique to self-limiting cyclic ALE
PatSnap Eureka

Search the full ALE patent landscape in seconds

Access 50+ ALE patents from Lam Research, Tokyo Electron, TSMC, SMIC, and more — with AI-powered analysis.

Analyse ALE IP on Eureka →
Patent Intelligence

ALE Innovation Landscape — Key Data from 50+ Patents

Visualising the assignee distribution, ALE variant coverage, and application domain spread across the analysed patent corpus.

ALE Patent Portfolio by Leading Assignee

Lam Research holds the dominant position; Tokyo Electron is the second most active assignee across the 50+ patent corpus filed 2000–2026.

ALE Patent Portfolio by Leading Assignee: Lam Research (dominant, largest portfolio), Tokyo Electron (second most active), University of Colorado (foundational chemistry), TSMC (device-manufacturer), SMIC (device-manufacturer), NAURA (domestic equipment), Jimai (III-V materials). Relative patent portfolio sizes for key ALE assignees derived from analysis of 50+ patents filed 2000–2026 across CN, US, KR, TW, JP, WO, and SG jurisdictions via PatSnap Eureka. Lam Research leads with the broadest technology coverage including directionality control, Designer ALE, passivation-assisted ALE, and metal ALE. Very High High Medium Low Minimal Dominant Lam Research 2nd Tokyo Electron Foundational Univ. Colorado Device TSMC Device SMIC Regional NAURA / Jimai

ALE Process Variant Coverage in Patent Corpus

Plasma (anisotropic) ALE leads coverage; thermal, quasi-ALE, and wet ALE variants address complementary selectivity challenges across the corpus.

ALE Process Variant Coverage: Plasma ALE (anisotropic, largest coverage), Thermal ALE (isotropic, fluorination and ligand-exchange), Quasi-ALE (reverse etch lag compensation), Wet ALE (solubility-limited, polycrystalline films), Integrated ALD/ALE (CD correction). Distribution of ALE process variant coverage across 50+ patent filings analysed via PatSnap Eureka. Plasma-based anisotropic ALE dominates, but thermal, quasi-, and wet ALE variants each address distinct selectivity challenges that plasma ALE cannot solve alone. 4 ALE modes Plasma ALE Anisotropic, bias-modulated Thermal ALE Isotropic, fluorination/ligand-exchange Quasi-ALE Reverse etch lag compensation Wet / ALD-ALE Hybrid and integrated CD correction

Want live ALE patent data filtered by assignee, jurisdiction, or application domain?

Run Your ALE Patent Search →
Engineering Implementations

Six Key ALE Application Domains at Sub-5nm

From EUV stochastic defect repair to metal interconnect patterning, the patent corpus reveals how ALE is being deployed across the full device stack.

🔗

Integrated ALD/ALE for Active CD Correction

Lam Research's 2025 filing describes etching a feature mask pattern to a width narrower than the desired final CD, then using conformal ALD to deposit a passivation layer that expands the mask width to the target value, followed by ALE of the underlying layer to transfer the corrected pattern at precisely the desired width. This ALD-then-ALE scheme decouples lithographic resolution limits from final device CD — all within a single vacuum environment.

EUV Stochastic Defect Correction

EUV lithography at sub-5nm nodes introduces stochastic defects — missing or merged features caused by photon shot noise and resist non-uniformity. Lam Research's cyclic process combines aspect-ratio-dependent deposition (via PECVD) with aspect-ratio-independent ALE. Because partially formed features present a different aspect ratio than fully opened features, the deposition step preferentially fills defective sites while the ALE step uniformly removes material at a constant per-cycle depth, systematically reducing defect density and line-edge roughness (LER).

⚙️

Contact Patterning in FinFET Processes

TSMC's 2018 filing discloses a method where two sequential ALE steps with different operating parameters open contact vias: the first ALE step removes the bulk of the interlayer dielectric, while a second ALE step with adjusted parameters exposes the source/drain features with minimal over-etch. The parameter-switching capability of ALE enables a progressive, controlled etch stop approach that is impossible with a single-chemistry continuous plasma etch — especially relevant for gate contacts at sub-7nm nodes where the ILD is only a few nanometers thick.

✂️

Sidewall Trimming for Sub-Lithographic CDs

SMIC's 2021 and 2023 filings demonstrate using ALE to trim the sidewalls of lithographically defined pattern layers. An organic layer formed on the sidewall reduces the bond energy of outermost atoms below that of the inner lattice atoms. During organic layer removal, the weakened outermost atoms are co-removed at a precisely controlled rate per ALE cycle. By repeating this sequence, sub-lithographic CDs can be achieved with angstrom-level dose control per cycle — directly enabling sub-5nm pitch patterning beyond optical resolution limits.

🔒
Unlock Metal ALE & Spatial Planarisation Insights
Explore the full engineering implementations — including subtractive metal interconnect patterning and monolayer-level planarisation — in the complete patent dataset on PatSnap Eureka.
Metal ALE (Lam Research KR) Spatial ALE planarisation + full patent texts
Access Full ALE Dataset →
Competitive Landscape

Key Assignees and Their ALE Innovation Focus Areas

The patent corpus spans seven jurisdictions. Here is how the leading assignees divide the ALE innovation space, based on PatSnap patent analytics.

Assignee Primary ALE Focus Areas Key Jurisdictions Filing Period Role
Lam Research Corporation Directionality control, Designer ALE, ALD/ALE integration, passivation-assisted ALE, metal ALE, EUV defect correction US, CN, TW, KR, SG, WO 2017–2025 Dominant
Tokyo Electron Limited (TEL) Quasi-ALE for RIE lag, Si-selective ALE with EUV resist, ultrathin ESL, wet/gas-phase hybrid ALE, fast ALE via ion mass segregation CN, US, TW 2016–2025 2nd Most Active
University of Colorado Foundational thermal ALE chemistry: sequential oxidation/fluorination, ligand-exchange rate enhancement US, WO 2016–2020 Foundational
TSMC ALE-enabled contact patterning in advanced FinFET / gate-last CMOS; two-step parameter-switching ALE for ILD opening US 2018 Device Mfr.
SMIC ALE sidewall trimming for sub-lithographic CD control; angstrom-level dose control per cycle US 2021–2023 Device Mfr.
Beijing NAURA Apparatus patents for process uniformity via dual-chamber electrode design CN 2023 Regional Equip.
Zhejiang Jimai ALE applied to III-V compound semiconductor patterning CN 2024 Regional Materials

Track ALE competitor filings in real time

Monitor Lam Research, Tokyo Electron, TSMC, and SMIC patent activity as it happens — with PatSnap Eureka alerts.

Set Up Patent Monitoring →
Application Domain Analysis

Where ALE Selectivity Is Being Applied Across the Device Stack

From front-end gate patterning to BEOL interconnects, ALE addresses distinct selectivity challenges at each level of the device stack.

ALE Application Domains — Patent Coverage by Stack Level

Gate/channel patterning and EUV integration lead patent coverage; metal interconnects and planarisation represent growing areas.

ALE Application Domains by Device Stack Level: Gate/Channel Patterning (highest coverage, sub-5nm FinFET and GAA), EUV Stochastic Defect Repair (high, LER and missing features), Contact Patterning (medium-high, ILD opening), Sidewall Trimming (medium, sub-lithographic CD), Metal Interconnects (medium, subtractive Cu/W), Low-k Dielectric BEOL (medium, RIE lag compensation), Planarisation (growing, spatial ALE). Coverage of ALE application domains across the semiconductor device stack, derived from 50+ patent filings analysed via PatSnap Eureka. Gate and channel patterning at sub-5nm FinFET and GAA nodes leads, with EUV defect repair integration as the second most prominent application area. 25% 50% 75% 100% Gate / Channel High EUV Defect Repair High Contact Patterning Med-High Sidewall Trimming Medium Metal Interconnects Medium Low-k BEOL Medium Planarisation Growing

ALE Patent Filing Jurisdictions in the Corpus

CN and US dominate filing activity; KR, TW, WO, JP, and SG reflect the global strategic breadth of ALE IP protection.

ALE Patent Filing Jurisdictions: CN (China, largest volume), US (United States, second largest), KR (South Korea), TW (Taiwan), WO (PCT international), JP (Japan), SG (Singapore). Corpus spans all 7 jurisdictions with dates from early 2000s to 2026. Distribution of ALE patent filings across 7 jurisdictions (CN, US, KR, TW, WO, JP, SG) from the 50+ patent corpus analysed via PatSnap Eureka. CN and US account for the majority of filings, reflecting both equipment manufacturer and device manufacturer strategic IP positions. High Med-H Med Low CN CN US US KR KR TW TW WO WO JP JP SG SG Source: PatSnap Eureka · 50+ ALE patents · Early 2000s–2026

Explore the full ALE patent dataset — filtered by jurisdiction, assignee, or technology area.

Search ALE Patents on Eureka →
Frequently asked questions

Atomic Layer Etching at Sub-5nm — Key Questions Answered

Still have questions about ALE selectivity mechanisms or patent strategy? Let PatSnap Eureka answer them for you.

Ask PatSnap Eureka Your ALE Question →
PatSnap Eureka

Map the Full ALE Patent Landscape — From Mechanisms to Assignees

Join 18,000+ innovators already using PatSnap Eureka to accelerate their R&D. Search 50+ ALE patents across 7 jurisdictions, track competitor filings, and identify white-space opportunities at sub-5nm.

References

  1. Control of directionality in atomic layer etching — Lam Research Corporation, 2019 (US)
  2. Control of directionality in atomic layer etching — Lam Research Corporation, 2018 (WO)
  3. Designer Atomic Layer Etching (设计者原子层蚀刻) — Lam Research Corporation, 2020 (CN)
  4. Designer ALE and Tantalum ALE (设计者原子层蚀刻和钽的原子层蚀刻) — Lam Research Corporation, 2024 (CN)
  5. Atomic layer etching processes using sequential, self-limiting thermal reactions comprising oxidation and fluorination — University of Colorado, 2019 (US)
  6. Atomic layer etching processes using sequential, self-limiting thermal reactions comprising oxidation and fluorination — University of Colorado, 2018 (US)
  7. Enhancement of thermal atomic layer etching — University of Colorado, 2020 (US)
  8. Novel methods of atomic layer etching (ALE) using sequential, self-limiting thermal reactions — University of Colorado, 2016 (WO)
  9. Selective ALE of Si-based materials (Si基材料的选择性原子层蚀刻) — Tokyo Electron, 2026 (CN)
  10. Method using ultrathin etch-stop layer in selective ALE (在选择性原子层蚀刻中使用超薄蚀刻停止层的方法) — Tokyo Electron, 2022 (CN)
  11. Atomic layer etching using passivation and directional plasma (使用钝化和定向性等离子体的原子层蚀刻) — Lam Research Corporation, 2026 (CN)
  12. Atomic layer etching using passivation and directional plasma — Lam Research Corporation, 2025 (TW)
  13. Method for reducing RIE lag in low-k dielectric etching (用于在低K电介质蚀刻中减少反应离子蚀刻滞后的方法) — Tokyo Electron, 2023 (CN)
  14. ALD and ALE in a single plasma chamber for CD control (针对关键尺寸控制在单一等离子体室中的原子层沉积和蚀刻) — Lam Research Corporation, 2025 (CN)
  15. Integrated atomic-scale processes: ALD and ALE (集成原子级工艺:ALD和ALE) — Lam Research Corporation, 2021 (CN)
  16. Eliminating yield impact of stochastic events in lithography (消除光刻中随机数的收率影响) — Lam Research Corporation, 2020 (CN)
  17. Semiconductor device and fabrication method therefor — TSMC, 2018 (US)
  18. Method for forming semiconductor structure — SMIC, 2021 (US)
  19. Atomic layer etching for subtractive metal etching — Lam Research Corporation, 2021 (KR)
  20. Systems and Methods for Improving Planarity using Selective Atomic Layer Etching (ALE) — Tokyo Electron, 2022 (US)
  21. Semiconductor Industry Association (SIA) — semiconductor technology roadmap and advanced node challenges
  22. IEEE — advanced semiconductor process control and atomic-scale patterning publications
  23. WIPO — global patent database and international filing statistics

All data and statistics on this page are sourced from the references above and from PatSnap's proprietary innovation intelligence platform. Patent analysis conducted via PatSnap Eureka.

Ask PatSnap Eureka
Ask PatSnap Eureka
AI innovation intelligence · always on
Ask anything about atomic layer etching at sub-5nm.
PatSnap Eureka searches patents and research to answer instantly.
Try asking
Powered by PatSnap Eureka