Atomic layer etching of SiO$_2$ using sequential SF$_6$ gas and Ar plasma

Basic Information

• Paper ID: 2412.20653
• Title: Atomic layer etching of SiO$_2$ using sequential SF$_6$ gas and Ar plasma
• Authors: Jun Peng, Rakshith Venugopal, Robert Blick, Robert Zierold
• Classification: cond-mat.mtrl-sci (Condensed Matter Physics - Materials Science)
• Institutions: Center for Hybrid Nanostructures, University of Hamburg; Deutsches Elektronen-Synchrotron (DESY), Germany
• Paper Link: https://arxiv.org/abs/2412.20653

Abstract

With the continuous advancement of semiconductor technology, the demand for atomic layer processes has driven innovative techniques that play a crucial role in ongoing device miniaturization. Among these, atomic layer etching (ALE) has attracted increasing attention due to its precise control of material removal at the atomic scale. Although some thermal ALE processes have achieved sub-nanometer etching controllability, current practical ALE processes involving plasma steps suffer from excessively high etching rates due to the lack of highly synergistic ALE half-reactions. To overcome this limitation, this study developed an ALE process for etching silicon dioxide (SiO$_2$) on silicon wafers using sequential sulfur hexafluoride (SF$_6$) gas exposure and argon (Ar) plasma near room temperature, achieving a stable and consistent etching rate of approximately 1.4 Å/cycle. In this process, neither half-cycle reaction alone produces etching; etching occurs only when the two are repeated sequentially, indicating 100% synergistic effect.

Research Background and Motivation

Problem Definition

1. Core Issue: Existing plasma-assisted atomic layer etching (ALE) processes suffer from excessively high etching rates and insufficient synergistic effects, making it difficult to achieve atomic-level precision control.
2. Significance:
   • As Moore's Law approaches its limits, semiconductor manufacturing requires more precise atomic-level processing techniques
   • ALE has already been applied in logic devices at the 10 nm technology node
   • Emerging nanoelectronic devices such as quantum devices have increasingly stringent requirements for precise etching
3. Limitations of Existing Methods:
   • Thermal ALE can achieve sub-nanometer control but is limited by isotropic etching characteristics
   • Plasma ALE offers good directionality but exhibits weak synergistic effects (~80%) and excessive etching rates
   • Lack of highly synergistic ALE half-reactions
4. Research Motivation: Develop an SiO$_2$ etching process that combines the precision of thermal ALE with the directionality of plasma ALE, achieving 100% synergistic effect and atomic-level control.

Core Contributions

1. Development of a Novel ALE Process: Proposed an atomic layer etching method for SiO$_2$ using sequential SF$_6$ gas and Ar plasma, achieving a stable etching rate of 1.4 Å/cycle
2. Achievement of 100% Synergistic Effect: Demonstrated that individual half-reactions produce no etching, with effectiveness only when sequentially combined, achieving perfect synergy
3. Identification of Process Windows: Identified temperature window (room temperature ~40°C) and plasma power window (50-100W), providing guidance for process optimization
4. Verification of Directional Etching: Through morphological characterization of micro-nanopillars and hole structures, confirmed the excellent directional etching characteristics of this process
5. Provision of Scalable Solution: Using commercial RIE equipment and commonly available gases, demonstrating good scalability and practical applicability

Methodology Details

Process Definition

Input: SiO$_2$/Si wafer Output: Atomically precise etched SiO$_2$ surface Constraints: Room temperature operation, sub-nanometer etching control

Process Architecture

Four-Step Cyclic Process

1. Surface Modification Step: SF$_6$ molecules adsorb on exposed substrate surface in a self-limiting manner
2. Purge Step: Remove excess molecules, leaving a thin SF$_6$ layer on the SiO$_2$ surface
3. Removal Step: Ar plasma activation, generating Ar$^+$ ions and free electrons
4. Purge Step: Purge the reaction chamber again, exposing a fresh SiO$_2$ surface

Reaction Mechanism

• SF$_6$ plasma generates reactive species: SF$_5^+$, SF$_4^{2+}$, F radicals
• F radicals react with SiO$_2$ to produce volatile byproduct SiF$_4$
• Self-limiting adsorption of SF$_6$ ensures only a single surface layer is etched

Technical Innovations

1. Unique Gas Combination: First use of pure SF$_6$ gas combined with Ar plasma, avoiding the complexity of traditional fluorocarbon compounds
2. Perfect Synergy Design:
   • α (SF$_6$ individual contribution) = 0
   • β (Ar plasma individual contribution) = 0
   • Synergistic effect S = 100%
3. Room Temperature Operation: Unlike other ALE processes requiring elevated temperatures, this method operates effectively near room temperature
4. Self-Limiting Characteristics: SF$_6$ dosage reaches saturation at 25 sccm·s, confirming self-limiting adsorption characteristics

Experimental Setup

Sample Preparation

• Substrate: 4-inch SiO$_2$ (300 nm)/Si wafer, cut into 1×1 cm samples
• Cleaning: Acetone, isopropanol, and deionized water cleaning
• Equipment: Commercial reactive ion etching system (SenTech SI 500)

Standard Process Parameters

• Temperature: 23°C constant temperature
• Pressure: 1 Pa working pressure
• Gas Flow: 100 sccm continuous Ar flow, 20 sccm SF$_6$ pulse for 5 seconds
• Plasma: 100W ICP power, 60 seconds
• Purge Time: 30 seconds

Characterization Methods

• Thickness Measurement: Ellipsometer (SenTech), using Cauchy model
• Morphology Characterization: Scanning electron microscopy (Zeiss Crossbeam 550)
• Roughness Measurement: Atomic force microscopy (AFM, Dimension)
• Pattern Fabrication: Electron beam lithography system (Raith)

Evaluation Metrics

• Etch Per Cycle (EPC): Å/cycle
• Synergistic Effect: S = (EPC-(α+β))/EPC × 100%
• Surface Roughness: Ra value
• Uniformity: Standard deviation across wafer

Experimental Results

Main Results

Etching Performance

• EPC: 1.4 Å/cycle, linear fit R² ≈ 0.999
• Synergistic Effect: S = 100% (α = 0, β = 0)
• Uniformity: Standard deviation ~0.5 nm across 4×4 cm area
• Surface Quality: Ra ≈ 0.7 nm, maintaining low roughness

Comparison with Existing Technology

This process's EPC (1.4 Å/cycle) is significantly superior to plasma ALE methods reported over the past decade:

• C$_4$F$_8$/Ar plasma: 1.9-20 Å/cycle
• CHF$_3$/Ar plasma: 4.0-15 Å/cycle
• This process precision approaches thermal ALE levels (0.027-0.52 Å/cycle)

Process Window Characterization

Temperature Window

• Stable Range: Room temperature ~40°C, EPC remains stable
• High Temperature Decay: >40°C, EPC gradually decreases, possibly due to SF$_6$ molecular thermal desorption

Power Window

• Effective Range: 50-100W ICP power
• Low Power: <50W, insufficient energy, EPC decreases
• High Power: >100W, anomalous EPC decrease, possibly due to F radical concentration dilution and elastic scattering

Directional Etching Verification

Micro-nanopillar Experiment

• Pillar Diameter: 600 nm, height ~91 nm
• Etching Result: After 450 cycles, pillar height remains unchanged (89.6±1.00 nm), diameter unchanged
• Vertical Etching: Total etching thickness 62 nm, anisotropy ratio >27:1

Hole Structure Experiment

• Hole Diameter: 0.6 μm and 1.2 μm
• Result: Hole diameter remains constant during etching, confirming directional characteristics

Ablation Experiments

Single-Step Testing

1. SF$_6$ Exposure Only: EPC ≈ 0, no etching effect
2. Ar Plasma Only: EPC ≈ 0, no physical sputtering
3. Combined Process: EPC = 1.4 Å/cycle

Dose Dependency

• SF$_6$ Dosage: Saturation at 25 sccm·s, confirming self-limiting characteristics
• Plasma Time: 60 seconds as optimal parameter

Related Work

ALE Technology Development

1. History: ALE concept first proposed in 1988 for diamond etching
2. Thermal ALE: First Al$_2$O$_3$ thermal ALE reported by Lee and George in 2015
3. Plasma ALE: Widely applied but with suboptimal synergistic effects

Current Status of SiO$_2$ ALE Research

1. Thermal ALE Methods: Using trimethylaluminum as precursor, EPC <1 Å/cycle
2. Plasma Methods: Using fluorocarbon compounds to modify surface, EPC 2-20 Å/cycle
3. Infrared Thermal Etching: ALE methods combining thermal effects

Advantages of This Work

1. Higher Precision: Achieving thermal ALE-level precision in plasma ALE
2. Better Synergy: 100% vs. conventional ~80%
3. Simpler Process: Avoiding complex fluorocarbon chemistry
4. Room Temperature Operation: Reducing process complexity

Conclusions and Discussion

Main Conclusions

1. Successfully developed SF$_6$/Ar plasma SiO$_2$ ALE process, achieving stable etching of 1.4 Å/cycle
2. Achieved 100% synergistic effect, demonstrating process purity
3. Identified temperature window near room temperature and power window of 50-100W
4. Verified excellent directional etching characteristics and surface quality

Limitations

1. Material Limitations: Currently verified only on SiO$_2$, further research needed for other materials
2. Equipment Dependence: Requires precisely controlled RIE system
3. Etching Rate: Compared to conventional RIE, ALE rate is lower, affecting production efficiency
4. Cost Considerations: Precision control increases process cost

Future Directions

1. Material Extension: Explore application of this method on other dielectric materials
2. Process Optimization: Further improve etching rate and selectivity
3. Mechanism Research: Deeper understanding of SF$_6$/Ar plasma reaction mechanisms
4. Industrialization: Develop process parameters suitable for large-scale production

In-Depth Evaluation

Strengths

1. Strong Technical Innovation: First realization of 100% synergistic SF$_6$/Ar plasma ALE
2. Rigorous Experimental Design: Systematic verification of synergistic effects and process window characterization
3. Convincing Results: Linear relationship R²≈0.999, good reproducibility
4. High Practical Value: Using commercial equipment and common gases, easy to promote

Weaknesses

1. Insufficient Mechanistic Explanation: Explanation of EPC decrease at high power is somewhat speculative
2. Missing Selectivity Research: Does not address selective etching of different materials
3. Long-term Stability: Lacks long-term cycling stability data
4. Limited Temperature Range: Relatively narrow process window

Impact

1. Academic Contribution: Provides new process route for ALE field
2. Industrial Value: Provides precise etching solution for advanced semiconductor manufacturing
3. Reproducibility: Detailed experimental parameters facilitate reproduction by other research groups

Applicable Scenarios

1. Quantum Device Manufacturing: Quantum structures requiring atomic-level precision etching
2. Advanced Logic Devices: Precision processing at 10 nm and below technology nodes
3. MEMS Devices: Microelectromechanical systems requiring high-precision surface treatment
4. Optoelectronic Devices: Optical components with extremely high surface quality requirements

References

This paper cites 37 relevant references covering the development history of ALE technology, SiO$_2$ etching methods, and plasma chemistry, providing a solid theoretical foundation for the research.

Overall Assessment: This is a high-quality materials science research paper that makes important contributions to the atomic layer etching field. This work not only achieves technological breakthroughs but also demonstrates excellent practical value and industrialization prospects.