AbstractEnglish Translation
Chemical mechanical polishing (CMP) and post-CMP cleaning are critical steps in semiconductor manufacturing, requiring atomic-scale flatness and complete removal of contaminants. This review examines the use of molecular dynamics (MD) simulations to elucidate atomic-scale mechanisms underlying these processes, focusing on four major MD methodologies: classical MD, reactive force field MD (ReaxFF), tight-binding quantum chemical MD (TB-QC MD), and ab initio MD (AIMD). Classical MD provides a foundation for simulating large-scale systems but lacks accuracy for modeling chemical reactions. ReaxFF allows real-time bond breaking and formation simulations during CMP. TB-QC MD combines quantum accuracy with classical efficiency, enabling exploration of chemical reaction effects on friction and material removal. AIMD directly calculates atomic interactions for precise depictions of chemical processes, albeit with high computational cost. MD simulations act as a 'computational microscope', enhancing CMP and postcleaning processes by quantifying interactions, material removal pathways, and contaminant desorption. Future research should address multiscale modeling challenges, improve AIMD efficiency, and develop accurate potential functions to propel semiconductor manufacturing toward greater precision and efficiency.
Executive Findings & Technical Breakthroughs
- •• • Classical MD enables large-scale simulations (up to millions of atoms) but cannot model chemical reactions, limiting its use in tribochemical processes; ReaxFF fills this gap by simulating bond breaking/formation in real time, critical for understanding material removal pathways. • • TB-QC MD integrates quantum accuracy with classical efficiency, allowing exploration of chemical reaction effects on friction and material removal at a fraction of AIMD's computational cost, making it suitable for systems where both reactivity and size matter. • • AIMD provides the most precise depiction of chemical processes by directly calculating atomic interactions, but its high computational expense restricts simulations to small systems and short timescales, necessitating efficiency improvements for industrial relevance. • • The review identifies force field parameterization for complex chemical environments and the need for experimental validation as key bottlenecks; future work must prioritize multiscale modeling and development of accurate potential functions to optimize CMP and post-CMP cleaning for sub-3 nm nodes.
Abstract
Chemical mechanical polishing (CMP) and post-CMP cleaning are critical steps in semiconductor manufacturing, requiring atomic-scale flatness and complete removal of contaminants. This review examines the use of molecular dynamics (MD) simulations to elucidate atomic-scale mechanisms underlying these processes, focusing on four major MD methodologies: classical MD, reactive force field MD (ReaxFF), tight-binding quantum chemical MD (TB-QC MD), and ab initio MD (AIMD). Classical MD provides a foundation for simulating large-scale systems but lacks accuracy for modeling chemical reactions. ReaxFF allows real-time bond breaking and formation simulations during CMP. TB-QC MD combines quantum accuracy with classical efficiency, enabling exploration of chemical reaction effects on friction and material removal. AIMD directly calculates atomic interactions for precise depictions of chemical processes, albeit with high computational cost. MD simulations act as a 'computational microscope', enhancing CMP and postcleaning processes by quantifying interactions, material removal pathways, and contaminant desorption. Future research should address multiscale modeling challenges, improve AIMD efficiency, and develop accurate potential functions to propel semiconductor manufacturing toward greater precision and efficiency.
1. Introduction
As integrated circuit (IC) manufacturing advances below 3 nm nodes, chemical mechanical polishing (CMP) and post-CMP cleaning face unprecedented demands for atomic-scale planarization and contaminant removal. The exponential increase in abrasive particles and reaction byproducts during CMP creates strongly bonded contaminants on nascent surfaces, threatening device yield and reliability. Conventional experimental approaches struggle to observe and control these atomic-scale phenomena, leaving a critical gap in understanding the fundamental mechanisms of material removal and contaminant desorption.
Molecular dynamics (MD) simulations have emerged as a powerful 'computational microscope' to bridge this gap. This review systematically evaluates four MD methodologies—classical MD, ReaxFF, TB-QC MD, and AIMD—each offering distinct trade-offs between system size, chemical accuracy, and computational cost. By quantifying interactions, material removal pathways, and contaminant dynamics, MD simulations provide actionable insights for optimizing CMP processes. However, challenges in force field parameterization and computational efficiency must be addressed to fully leverage these tools for next-generation semiconductor manufacturing.
Loading authentic research manuscript (Pages 1–5)...
Lifei Zhang, Ming Ji, Xinchun Lu (2026). Molecular Dynamics Simulations Addressing Atomic-Scale Core Issues in Chemical Mechanical Polishing and Post-CMP Cleaning: A Concise Review. SinoTechIntel Verified Research. https://doi.org/10.26599/FRICT.2025.9441197
Research & Educational Purpose Only:The translations, structured abstracts, analytical annotations, and data reports provided by SinoTechIntel are intended exclusively for academic research, corporate R&D benchmarking, and educational evaluation under international fair use principles.
Copyright Ownership: Source copyright remains with original Chinese academic publishers and authors. SinoTechIntel claims editorial rights over its original English translations and structural index enhancements.
Frequently Asked Questions
What are the key limitations of classical MD in simulating CMP processes, and how does ReaxFF overcome them?
Classical MD cannot model chemical reactions due to fixed bond connectivity, limiting its use to physical interactions. ReaxFF, however, allows bond breaking and formation in real time, enabling simulation of tribochemical reactions during polishing, which is essential for understanding material removal mechanisms.
How does TB-QC MD balance quantum accuracy and computational efficiency compared to AIMD?
TB-QC MD uses tight-binding approximations to reduce computational cost while retaining quantum mechanical accuracy, making it suitable for systems where chemical reactions and friction are coupled. AIMD offers higher accuracy but is computationally prohibitive for large systems, limiting its practical application.
What are the main challenges in adopting MD simulations for industrial CMP optimization?
Key challenges include accurate force field parameterization for complex chemical environments, the high computational cost of AIMD, and the need for experimental validation to confirm predictive accuracy. Multiscale modeling approaches are required to bridge atomic-scale insights to process-level outcomes.
What specific future research directions are recommended to advance MD simulations in CMP?
Future research should focus on improving AIMD efficiency, developing precise potential functions for complex chemical interactions, and integrating multiscale modeling to connect atomic-scale phenomena with macroscopic process parameters. These efforts will enable more accurate predictions and optimization of CMP and post-CMP cleaning.
Related Literature in This Field
Study on the effect of diketone lubricant on the tribological properties of angular contact ball bearings with skidding behavior
Skidding in angular contact ball bearings significantly increases friction, wear, and temperature, adversely affecting bearing performance and service life. Despite its critical impact, systematic investigations of lubrication behavior under skidding conditions remain scarce, with conventional lubricants often failing to provide stable low-friction operation. To address this gap, this study first calculated critical skidding parameters using a quasi-static model. Subsequently, experimental parameters for bearings with and without skidding were selected to evaluate tribological behaviors under three lubricants: base oil, commercial lubricant, and a diketone-based lubricant (PAO = 14 (20%)). Results demonstrate that under skidding conditions, the diketone lubricant achieved the lowest coefficient of friction (COF) of 0.0008 and temperature rise of 2.8 °C. Furthermore, diketone-lubricated bearings exhibited excellent anti-wear performance and an extremely short running-in period. The superior tribological performance is attributed to the synergistic effect of diketone molecular adsorption and chelation with iron atoms, which reduces friction and temperature rise. These findings highlight the potential of diketone lubricants to enhance bearing performance and durability under extreme operating conditions.
Steady Shear Rheological Response of Ferrofluids Containing Hydrophilic Fumed Silica under Magnetic Fields
This study investigates the steady shear rheological behavior of water-based ferrofluids composited with hydrophilic fumed silica under different magnetic field strengths, with particular attention paid to avoiding gelation that reduces fluidity. Seven composite ferrofluid samples were prepared and characterized. By adjusting the silica particle size and volume fraction, their effects on viscosity and yield stress were explored. As a result, pronounced shear-thinning behavior is observed in this dispersion, with their flow curves under different magnetic field strengths effectively scaled by the Mason number. A higher silica concentration or larger particle size increases the critical Mason number, showing that field-induced structures become more stable. In contrast, only high silica concentrations significantly enhance shear thinning, as reflected by a larger flow index, whereas particle size has little influence. Yield stress analysis further shows that macroscopic models capture normalized Bingham yield stress, while microscopic models better predict normalized static yield stress. Overall, this work demonstrates that hydrophilic fumed silica offers a simple and effective route for tuning the magnetorheology of water-based ferrofluids without inducing gelation, ensuring controllable rheology and good fluidity.
Slippery Liquid-Infused Porous Surface with Layered Double Hydroxides for Enhanced Corrosion and Wear Resistance of TC4 Alloys
Titanium alloys, particularly TC4 (Ti–6Al–4V), suffer from poor tribological performance and susceptibility to pitting corrosion, limiting their application in marine and biomedical fields. Layered double hydroxide (LDH) coatings offer potential protection but are hindered by the dense oxide layer on titanium alloys. In this study, a ZnAl LDH coating was fabricated on TC4 via in situ growth, followed by molybdate anion intercalation through ion exchange. A biomimetic slippery liquid-infused porous surface (SLIPS) was then created by UV-grafting polydimethylsiloxane (PDMS) onto the nanoporous LDH structure. The resulting surface exhibited excellent hydrophobicity, corrosion resistance, and wear resistance. Electrochemical tests (Tafel polarization and electrochemical impedance spectroscopy) demonstrated superior corrosion protection, with a low corrosion current density of 2.34×10−7 A/cm2. The infused silicone oil and ZnAl LDH nanosheets synergistically improved wear performance compared to bare TC4. This work provides insights into controllable in situ fabrication of LDH coatings and offers a novel strategy for broadening TC4 alloy applications in metal protection.
Manipulating the Conversion of Nanoscale Wear Debris into Tribofilm for Wear Reduction of Steel
Wear debris particles play a crucial role in frictional interfaces. Conventional understanding holds that debris accumulation causes severe wear. Interestingly, the debris from metal friction pairs includes anti-wear metal oxides generated by tribochemical reactions, which can form a protective oxidation film to resist wear. However, minimizing the abrasive damage caused by accumulated debris and using the anti-wear property of the metal oxides can be mutually exclusive. Here, a rational design of a coupling surface that manipulates nanoscale wear debris to resist further wear is reported. It consists of surface textures used to capture and temporarily store excess nanoscale wear debris, a deposited self-cleaning coating that subsequently helps transfer part of the captured debris into the sliding-contact interface, where it converts into a protective oxidation film. The coexistence of the two elements with contrasting properties in manipulating nanoscale wear debris considerably reduces wear under conditions of water lubrication, oil lubrication, and macroscale superlubricity. Our strategy achieves the manipulation and utilization of wear debris for anti-wear purposes. This work holds the potential to promote further investigation into the role of nanoscale wear debris and its utilization approaches.