Recent advances in machine learning guided mechanical properties prediction and design of two-dimensional materials

Thin-Walled Structures, Год журнала: 2025, Номер unknown, С. 113261 - 113261

Опубликована: Март 1, 2025

Язык: Английский

---

Machine Learning Interatomic Potential for Modeling the Mechanical and Thermal Properties of Naphthyl-Based Nanotubes

Journal of Chemical Theory and Computation, Год журнала: 2025, Номер unknown

Опубликована: Янв. 28, 2025

Two-dimensional (2D) nanomaterials are at the forefront of potential technological advancements. Carbon-based materials have been extensively studied since synthesizing graphene, which revealed properties great interest for novel applications across diverse scientific and domains. New carbon allotropes continue to be explored theoretically, with several successful synthesis processes carbon-based recently achieved. In this context, study investigates mechanical thermal DHQ-based monolayers nanotubes, a allotrope characterized by 4-, 6-, 10-membered rings, route using naphthalene as molecular precursor. A machine-learned interatomic (MLIP) was developed explore nanomaterial's behavior larger scales than those accessible through first-principles calculations. The MLIP trained on data derived from DFT/PBE (density functional theory/Perdew–Burke–Ernzerhof) level ab initio dynamics (AIMD). Classical (CMD) simulations, employing MLIP, that Young's modulus nanotubes ranges 127 243 N/m, depending chirality diameter, fracture occurring strains between 13.6 17.4% initial length. Regarding response, critical temperature 2200 K identified, marking onset transition an amorphous phase higher temperatures.

Язык: Английский

---

Transferability of machine-learning interatomic potential to α-Fe nanocrystalline deformation

International Journal of Mechanical Sciences, Год журнала: 2025, Номер unknown, С. 110132 - 110132

Опубликована: Март 1, 2025

Язык: Английский

---

Navigating the Evolution of Carbon Nitride Research: Integrating Machine Learning into Conventional Approaches

Physical Chemistry Chemical Physics, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

Carbon nitride research has reached a promising point in today's endeavours with diverse applications including photocatalysis, energy storage, and sensing due to their unique electronic structural properties. Recent advances machine learning (ML) have opened new avenues for exploring optimizing the potential of these materials. This study presents comprehensive review integration ML techniques carbon an introduction CN classifications recent advancements. We discuss methodologies employed, such as supervised learning, unsupervised reinforcement predicting material properties, synthesis conditions, enhancing performance metrics. Key findings indicate that algorithms can significantly reduce experimental trial-and-error, accelerate discovery processes, provide deeper insights into structure-property relationships nitride. The synergistic effect combining traditional approaches is highlighted, showcasing studies where driven models successfully predicted novel compositions enhanced functional Future directions this field are also proposed, emphasizing need high-quality datasets, advanced models, interdisciplinary collaborations fully realize materials next-generation technologies.

Язык: Английский

---

The evolution of machine learning potentials for molecules, reactions and materials

Chemical Society Reviews, Год журнала: 2025, Номер unknown

Опубликована: Янв. 1, 2025

This review offers a comprehensive overview of the development machine learning potentials for molecules, reactions, and materials over past two decades, evolving from traditional models to state-of-the-art.

Язык: Английский

---

A Hybrid Machine Learning Framework for Predicting Hydrogen Storage Capacities in Metal Hydrides: Unsupervised Feature Learning with Deep Neural Networks

ACS Applied Materials & Interfaces, Год журнала: 2025, Номер unknown

Опубликована: Май 12, 2025

In this study, we present a sophisticated hybrid machine-learning framework that significantly improves the accuracy of predicting hydrogen storage capacities in metal hydrides. This is critical challenge due to scarcity experimental data and complexity high-dimensional feature spaces. Our approach employs power unsupervised learning through use state-of-the-art autoencoder. autoencoder trained on elemental descriptors obtained from Mendeleev software, enabling extraction meaningful lower-dimensional latent space input data. representation serves as basis for our deep multilayer perceptron (MLP) model, which consists five layers shows good precision capacities. Furthermore, results show very agreement with density functional theory (DFT). addition addressing limitations caused by limited unevenly distributed field materials, also focus discovering new materials promising opportunities storage. These were identified using both feature-based approaches predictions generated large language model (LLM). A significant highlight work integration decoder-only LLM based GPT-2, fine-tuned generation Using such an approach, have discovered selected subset subsequently validated (DFT) calculations.

Язык: Английский

---

A machine learning-assisted exploration of the structural stability, electronic, optical, heat conduction and mechanical properties of C3N4 graphitic carbon nitride monolayers

Deleted Journal, Год журнала: 2024, Номер unknown, С. 100024 - 100024

Опубликована: Дек. 1, 2024

Язык: Английский

---

A Multi-Method Approach to Analyzing MOFs for Chemical Warfare Simulant Capture: Molecular Simulation, Machine Learning, and Molecular Fingerprints

Nanomaterials, Год журнала: 2025, Номер 15(3), С. 183 - 183

Опубликована: Янв. 24, 2025

Mustard gas (HD) is a well-known chemical warfare agent, recognized for its extreme toxicity and severe hazards. Metal–organic frameworks (MOFs), with their unique structural properties, show significant potential HD adsorption applications. Due to the hazards of HD, most experimental studies focus on simulants, but molecular simulation research these simulants remains limited. Simulation analyses can uncover structure–performance relationships enable validation, optimizing methods, improving material design performance predictions. This study integrates simulations, machine learning (ML), fingerprinting (MFs) identify MOFs high simulant diethyl sulfide (DES), followed by in-depth analysis comparison. First, are categorized into Top, Middle, Bottom materials based efficiency. Univariate analysis, learning, then used compare distinguishing features fingerprints each category. helps optimal ranges Top materials, providing reference initial screening. Machine feature importance combined SHAP identifies key that significantly influence model predictions across categories, offering valuable insights future design. Molecular fingerprint reveals critical combinations, showing optimized when such as metal oxides, nitrogen-containing heterocycles, six-membered rings, C=C double bonds co-exist. The integrated using HTCS, ML, MFs provides new perspectives designing high-performance demonstrates developing CWAs simulants.

Язык: Английский

---

Hierarchical multiscale fracture modeling of carbon-nitride nanosheet reinforced composites by combining cohesive phase-field and molecular dynamics

Composite Structures, Год журнала: 2025, Номер unknown, С. 118942 - 118942

Опубликована: Фев. 1, 2025

Язык: Английский

---

Multiscale simulation and machine learning facilitated design of two-dimensional nanomaterials-based tunnel field-effect transistors: A review

APL Machine Learning, Год журнала: 2025, Номер 3(1)

Опубликована: Март 1, 2025

Traditional transistors based on complementary metal–oxide–semiconductor and field-effect are facing significant limitations as device scaling reaches the limits of Moore’s law. These include increased leakage currents, pronounced short-channel effects, quantum tunneling through gate oxide, leading to higher power consumption deviations from ideal behavior. Tunnel Field-Effect Transistors (TFETs) can overcome these challenges by utilizing charge carriers switch between off states achieve a subthreshold swing below 60 mV/decade. This allows for lower consumption, continued scaling, improved performance in low-power applications. review focuses design operation TFETs, emphasizing optimization material selection advanced simulation techniques. The discussion will specifically address use two-dimensional materials TFET explore methods ranging multi-scale approaches machine learning-driven optimization.

Язык: Английский

---