Patterning Challenges and Opportunities: Etch and Film Ying Zhang, Shahid Rauf, Ajay Ahatnagar, David Chu, Amulya Athayde, and Terry Y. Lee Applied Materials, Inc. SEMICON, Taiwan 2016 Sept. 07-09, 2016, Taipei, Taiwan
2 Outline Advanced nodes pose challenges for patterning These challenges demand new film and etch/removal capabilities Atomic Level Deposition Atomic Level Etch and Removal Low electron temperature plasma etch Highly selective radical based removal Closing remarks
Advanced nodes pose challenges for patterning
4 Patterning Technology Trend Lithography Technology Materials Engineering Lithography Technology 248nm 193nm 193i Litho multiple exposure EUV Complementary Litho e.g., 193i + EUV Key challenge: Overlay EPE Materials Engineering Etch Film ALD Gapfill Selective removal ALE Selective deposition/growth Key advantage: Enable self-align schemes Atomic Level Controllability
5 SAxP Flows In SAxP pitch splitting flows 1 litho step + many non-litho steps (film, etch, etc.) e.g.: SAQP: Litho Etch ALD Etch ALD Etch
6 CD/CDU/LER/LWR dominated by Litho, Etch and ALD In SAQP, there are 8 edges : Direct edge: = f (Litho CD/CDU/LER/LWR) S1 edge: = f (Litho and 1 st spacer CD/CDU/LER/LWR) S2 edge: = f (1 st and 2 nd spacer CD/CDU/LER/LWR) S1/S2 edge: = f (1 st, 1 st spacer and 2 nd spacer CD/CDU/LER/LWR) Source: Schenker, Intel SPIE 2016 To systematically reduce EPE: CD/CDU/LER/LWR of all edges at all steps need to be measured to trace down root causes Litho the key source of LER Etch/ALD the key for pitch walking
These challenges demand new film and etch/removal capabilities - ALD
8 Conventional ALD vs. Olympia TM Reconfigures ALD What is ALD? Conventional ALD Olympia TM ALD Divides CVD into two half-reactions Precursor Precursor Is self-limiting, producing uniform, conformal deposition A B A B Wafer is stationary A B On Off On Off A B Alternating chemistries Purge separates chemistries Primary technology used today Wafer travels continuously Spatially separated chemistries Chemistry-free zones isolate individual chemistries
Modular Design for Atomic-Level Engineering Precursor Precursor Treatment X ALD Mode Process Sequence Conventional ALD Thermal Plasma Enhanced A A B B p 20nm Titanium Nitride 20n m Silicon Nitride 20nm Titanium Oxide 20n m Silicon Oxide 100nm Aluminum Oxide Olympia TM ALD Atomic- Layer Treatment Versatility Broadens Spectrum of Achievable ALD Materials without Compromising Productivity A B X Source: Applied Materials, Inc. 9
These challenges demand new film and etch/removal capabilities - Etch
11 RIE Plasma etching patterning trend Mainstream plasma technologies Variety of CCP Variety of ICP ECR DSP/RP Add-on s Variety of RF pulsing technologies Mainstream plasma technologies Variety of CCP Variety of ICP ECR DSP/RP Add-on s Variety of RF pulsing technologies Thin Layer Etching (TLE) Atomic Layer Etching (ALE) Complex pulsing technologies Advanced radical etching Low T e plasmas Neutral beam
12 Basic Mechanisms of Reactive Ion Etching Ion-neutral reaction synergism One of the most important concepts of plasma-surface chemistry is the synergism of ion and neutral reactions Three key aspects of ion bombardment: Stimulate surface reactions Stimulate desorption or clear the surface of etch-inhibiting, nonvolatile residues Anisotropic or directional etching Ion Bombardment effects in Reactive Ion Etching Coburn and Winters, J. of App. Phys. 50. 3189-3196, 1979
SICLthick (A) 13 Low electron temperature, T e, plasmas Intuitively, lower T e lower V p lower ion energy lower damage ALE(?) How to control low ion energy, e.g., from <4eV to ~25eV? 50 40 30 5eV 10eV 25eV 50eV 100eV Radical Cl + Cl + Radical Cl + 25eV Cl + ~4 layers 20 10 Radical Cl + 5eV Cl + ~1-2 layers 0 0 1 2 3 4 5 Cl+ Fluence (ML) From Oliver Joubert, CNRS-LTM
14 Low T e Plasma Etch System A low T e plasma is produced in the processing chamber using energetic beam electrons in the 0.5 2.5 kev energy range. A separate inductively coupled plasma (ICP) based radical source is used in our system to provide accurate control over relative concentrations of radicals and ions Another important element in this plasma processing system is low frequency RF bias capability which allows control of ion energy in the 2 50 ev range Radical source e-beam source x Bias (wafer voltage)
15 Cross-section (Å 2 ) f e (au) Ion / Radical Composition: RF and Low T e Plasmas In an RF plasma (with T e = 4.0 ev), significantly more electrons can dissociate than ionize due to lower threshold for dissociation. In a low T e plasma produced using energetic electrons, radical / ion fraction is much lower. 6 Cl 2 1.2 f e @ T e = 4.0 ev Ebeam 1.0 4 s ion 0.8 f e @ T e = 0.2 ev 0.6 2 0.4 s diss 0.2 0 1 10 100 1000 Energy (ev) 0.0
Low T e Plasma can etch Si layer-layer with minimal damage The top surface can be more quantitatively analyzed using electron energy loss spectroscopy (EELS). The thickness of the amorphous layer at the top is similar for the unprocessed sample and the sample which has been etched in the low T e plasma only. When RF bias is applied to increase E i, the amorphous layer thickness increases. The sample that was etched in the inductively coupled plasma without bias shows similar damage to the 0.8 W etch case.
These challenges demand new film and etch/removal capabilities Selective Removal
18 What is Extreme Selectivity? Multiple Material Layers are Formed in a Structure Extreme Selectivity Enables Removal of Only One Material No Damage or Residues Remaining Selectra TM Removes Target Material without Damage to Others Critical for Patterning and 3D Architectures
19 Collapse Percentage (%) New Etch Methods Required to Continue Scaling Traditional Wet Etch Collapse of high aspect ratio structures Inability to penetrate small dimensions Traditional Dry Etch Lacks extreme selectivity Insufficient lateral etch control 100 80 Overetch at Top 60 40 20 Internal Image Incomplete Removal 0 Graph Courtesy of imec 10 15 20 25 30 Aspect Ratio Pattern Collapse Internal Image Tight Features Internal Image Lateral Control Traditional Etch Technologies Unable to Advance Moore s Law Insufficie nt at Bottom
20 How Does Selectra TM Achieve Extreme Selectivity? Plasma creates etchant chemistry Ions are blocked, chemistry passes through Damage-free, extreme selectivity etch without polymers The Selectra TM System Creates Tailored Chemistry for Extreme Selectivity
21 Extreme Selectivity Enables 10nm Multi-Patterning Pre- Selectra TM 9.3n m Si Ox SiN Ox Internal Image Post- Selectra TM Ox SiN 9.3n m No change in spacer width Ox Internal Image
Etch Amount (Å) Atomic-Level Precision Enables 10nm FinFET Pre-Selectra TM Post-Selectra TM Pre-Selectra TM Post-Selectra TM Ox Ox α-si 10 8 Si Internal Image Si Internal Image TiN Ox Ox TiN 6 0 Applied Materials Internal Structures 22 4 2 Silicon etch of two atomic layers Selectra TM Enables Fin Scaling and Penetration of Atomic-Level Structures Si Internal Image Si Internal Image Can access spaces <5 silicon atoms across
23 Lateral Etch Uniformity Enables 3D NAND Pre Etch Traditional Etch Selectra TM Etch Selectra TM Etch Creates Consistent Contact Resistance
24 Closing Remark Advanced nodes pose challenges for patterning Patterning trend: Litho dominating Litho/Materials engineering dominating Recent EUV emerging will help Litho, e.g., complementary litho, but not likely change this trend These challenges demand new film and etch/removal capabilities CD/CDU/LER/LWR play increasingly critical role in scaling Etch/Removal and Film play increasingly critical role in EPE reduction More opportunities for Film and Etch/Removal but key challenges are to have atomic level precision Atomic Layer Deposition Atomic Layer Etch and Removal Low electron temperature plasma etch Highly selective radical based removal