arXiv subjects

Jan Michael Y. Carrillo

Publications and source records attributed to Jan Michael Y. Carrillo.

3 recordsLinked to original sources

arXiv2026

Symbolic Ensemble Learning Enables Discovery of Fast Accurate Physics-Based Interatomic Potentials

Machine learning has transformed materials simulation by delivering force fields with ab initio accuracy, yet bridging the gap between high-dimensional regression and physical interpretability remains a grand challenge. Conventional analytical potentials offer transparency but often fail to capture the complexity of far-from-ground state regimes. Here, we introduce a hybrid symbolic-neural framework that unifies the interpretability of the Embedded Atom Method (EAM) with the adaptability of data-driven learning. Using Equation Learner Neural Networks (EqNNs) trained on density functional theory (DFT) data, we obtain interpretable models for aluminum through three distinct training protocols: random initialization trained via Monte Carlo Tree Search (MCTS) and gradient descent, and two transfer-learning strategies initialized from a copper potential - one employing MCTS followed by gradient descent, and the other using gradient descent only. We find that while all three resulting symbolic models achieve sub-10 meV/atom accuracy, they occupy distinct local minima in the functional landscape, exhibiting complementary trade-offs across phonon dispersion, surface energetics, and elastic response. By integrating these diverse functional forms through a weighted symbolic ensemble, we derive a composite potential that surpasses the fidelity of its constituent models. The resulting ensemble effectively mitigates individual biases, delivering superior consistency with DFT benchmarks across equation-of-state curvature, phonon spectra, and melting dynamics. This approach demonstrates that combining transfer learning with ensemble symbolic regression yields compact, transparent potentials capable of robust prediction across equilibrium and non-equilibrium states.

cond-mat.mtrl-sci↗

arXiv2024

Enhanced Capacitance in Surfactant Mediated Ion Complexes

Ion complexes hold the key for various energy transfer and communication systems in nature and industries. Achieving controllable mechanical and electrical properties in these complex systems is highly desirable but remains challenging. In this work, we propose the use of amphiphilic molecules to mediate salt crystallization. The resultant ionic interfaces can be tunable through reduced surface tensions of crystal facets, an additional intermolecular modifier, and the diffusion-limited crystallization during controlled solvent evaporation. Specifically, we utilized the ionic surfactant Sodium Lauryl sulfoacetate (SLSA) to mediate the crystallization of sodium chloride. Citric acid (CA) was adopted as a hydrogen-bond modifier. Systems mediated by SLSA exhibited the highest capacitance, and a significant enhancement of capacitance was also observed in systems with both SLSA and CA. The substantially increased capacitance of these ionic complexes can be attributed to changes in interfacial and crystalline grain structures. Transmission electron microscopy (TEM), optical microscopy (OM), and Finite Element Analysis (FEA) were used to study the effects of surfactant molecules in these ionic complexes. Understanding the role of ionic complexation in defining the thermodynamic and kinetic process of the crystallization process will contribute to better optimization in nucleation and scalability of organic and inorganic crystal production.

cond-mat.mtrl-sci↗

arXiv2023

Accelerated Design of Block Copolymers: An Unbiased Exploration Strategy via Fusion of Molecular Dynamics Simulations and Machine Learning

Star block copolymers (s-BCPs) have potential applications as novel surfactants or amphiphiles for emulsification, compatbilization, chemical transformations and separations. s-BCPs are star-shaped macromolecules comprised of linear chains of different chemical blocks (e.g., solvophilic and solvophobic blocks) that are covalently joined at one junction point. Various parameters of these macromolecules can be tuned to obtain desired surface properties, including the number of arms, composition of the arms, and the degree-of-polymerization of the blocks (or the length of the arm). This makes identification of the optimal s-BCP design highly non-trivial as the total number of plausible s-BCPs architectures is experimentally or computationally intractable. In this work, we use molecular dynamics (MD) simulations coupled with reinforcement learning based Monte Carlo tree search (MCTS) to identify s-BCPs designs that minimize the interfacial tension between polar and non-polar solvents. We first validate the MCTS approach for design of small- and medium-sized s-BCPs, and then use it to efficiently identify sequences of copolymer blocks for large-sized s-BCPs. The structural origins of interfacial tension in these systems are also identified using the configurations obtained from MD simulations. Chemical insights on the arrangement of copolymer blocks that promote lower interfacial tension were mined using machine learning (ML) techniques. Overall, this work provides an efficient approach to solve design problems via fusion of simulations and ML and provide important groundwork for future experimental investigation of s-BCPs sequences for various applications.

cond-mat.soft↗