Theoretical Roadmap Links Molecular Structure to Memristive Function for Neuromorphic Computing
Summarize this article with:
Neuromorphic computing seeks to revolutionise information processing by mimicking the human brain, and central to this effort are memristive materials which combine memory and computation within a single device. Salvador Cardona-Serra from the Universitat de València, and colleagues, now present a theoretical framework to guide the development of organic memristors, materials offering advantages in tunability, cost, and biocompatibility over existing inorganic options.
The team outlines a multiscale computational approach, integrating techniques such as molecular dynamics, to understand how molecular structure dictates memristive behaviour. This work addresses a critical gap in the field, paving the way for the rational design of chemically engineered synaptic materials and accelerating progress towards brain-inspired computing technologies. Molecular Devices, Memristors and Spintronic Effects This research encompasses a broad investigation into molecular materials for advanced electronic and spintronic devices, focusing on memristors, spintronics, and related nanoscale systems. Scientists are exploring organic memristors, manipulating electrode designs to control device characteristics, and investigating how molecules interact within these devices. The work extends to nanoscale materials, such as spin-crossover nanoparticles, integrated into complex structures to achieve tunable properties, and explores the potential of peptides as platforms for quantum computing. Researchers are designing molecular systems that can control electron spin, including single-molecule magnets and materials exhibiting chiral-induced spin selectivity, also investigating spin filtering in peptides and nanoparticles. This interdisciplinary effort combines materials science, chemistry, physics, and computational methods to design and understand novel molecular materials for advanced electronic and spintronic devices, demonstrating the importance of theoretical simulations in guiding materials discovery and optimization. The research employs a diverse range of computational techniques, from quantum mechanical calculations using programs like Gaussian, ORCA, and SIESTA, underpinned by Density Functional Theory, to molecular dynamics simulations with GROMACS, NAMD, cp2k, and Amber. Ring-Polymer Molecular Dynamics incorporates quantum effects into these simulations, while coarse-graining simplifies complex systems.
Kinetic Monte Carlo simulations model the evolution of systems, particularly useful for studying switching in memristors, utilizing custom platforms for this purpose. Global optimization is achieved through algorithms like the Artificial Bee Colony method, and software like ESPResSo simulates soft matter systems. Techniques like ONIOM combine high- and low-level quantum calculations, and GFN2-xTB provides fast and accurate quantum chemical results. Monte Carlo simulations are used for various calculations, and multigraining combines fine- and coarse-grained simulations.
Multiscale Simulations Design Advanced Organic Memristors Scientists have developed a comprehensive multiscale computational methodology to design next-generation organic memristors, devices that emulate biological synapses. This work integrates multiple computational approaches, starting with detailed quantum mechanical calculations to understand fundamental material properties. Molecular trajectories are obtained by numerically solving Newton’s equations of motion, enabling simulations involving up to tens of thousands of atoms, a significant increase in scale compared to traditional methods.
The team progressed to molecular dynamics (MD) simulations, which consider the temporal evolution of the system, and employed coarse-grained molecular dynamics (CGMD), allowing for the modeling of systems containing up to one million atoms by grouping atoms into ‘pseudo-atoms’. The methodology also incorporates combined Quantum Mechanics/Molecular Mechanics (QM/MM) methods, treating selected atoms with quantum mechanical calculations while applying simpler molecular mechanics to the rest of the system. Finally, the researchers implemented effective differential equations based on finite elements and finite differences, enabling the prediction of macroscopic electrical memristive behavior using locally obtained parameters. Monte Carlo methods were also used to simulate the device’s active region, employing mean field approximation and particle interactions consistent with previously determined parameters. This comprehensive approach allows for the assessment of critical material features, including the number of accessible conductance states, cycling endurance, energy consumption, and switching time, ultimately aiming to rationally design molecular memristors with enhanced properties through chemical design.
Multiscale Simulations Design Organic Memristors Scientists are pioneering a multiscale computational approach to design next-generation organic memristors, essential components for brain-inspired computing. This work addresses a critical gap in the field by exploring the potential of molecular and polymeric systems. The study employs a combination of quantum chemistry and molecular dynamics simulations to understand how molecular structure dictates memristive function, enabling the rational design of new synaptic materials. Researchers investigate three key mechanisms, ionic migration, redox-driven switching, and conduction interplay in chiral molecules, as representative pathways toward molecular neuromorphic hardware. The methodology begins with quantum chemistry calculations to determine the electronic structure and properties of candidate molecules, providing fundamental insights into their potential for memristive behavior. Subsequently, molecular dynamics simulations model the dynamic behavior of these molecules, capturing how they respond to applied voltages and currents over time, allowing scientists to observe ionic migration, conformational changes, and the formation of conductive filaments. The simulations accurately model the materials at the nanoscale, revealing how molecular arrangements influence resistance changes and hysteresis loops. By systematically varying molecular structures and simulating their behavior, scientists identify key design principles for optimizing device performance, promising enhanced miniaturization, energy efficiency, and performance, exceeding the limits of conventional transistor-based technologies, and contributing to the development of advanced neuromorphic systems with improved computational capabilities and energy efficiency. Multiscale Modelling of Molecular Memristor Switching This research presents a multiscale computational framework designed to accelerate the development of organic molecular memristors, key components in next-generation neuromorphic computing systems. Scientists successfully integrated diverse computational techniques, including molecular dynamics and quantum chemical methods, to investigate the fundamental mechanisms governing memristive switching at the molecular level. Three prominent switching mechanisms, ionic migration, redox-driven processes, and conduction in chiral molecules, were examined, providing insights into how molecular structure influences electronic behavior.
The team’s work demonstrates the potential of theoretically guided design to create chemically engineered synaptic materials with tailored properties. By combining these computational approaches, researchers can now predict and optimize molecular structures for enhanced memristive performance, potentially overcoming limitations of existing inorganic materials. The authors highlight the ongoing development of efficient algorithms and software packages to address the computational demands of these multiscale simulations, paving the way for more efficient and biocompatible neuromorphic devices. 👉 More information 🗞 Toward a Theoretical Roadmap for Organic Memristive Materials 🧠 ArXiv: https://arxiv.org/abs/2512.05617 Tags: