Autonomous heterogeneous catalyst discovery with a self-evolving multi-agent digital twin

• URL: http://arxiv.org/abs/2606.05050v1
• Date: Wed, 03 Jun 2026 16:10:58 GMT
• Title: Autonomous heterogeneous catalyst discovery with a self-evolving multi-agent digital twin
• Abstract summary: CatDT (Catalysis Digital Twin) is a self-evolving multi-agent system that builds an autonomous digital twin of a working catalyst.<n>Eight specialized agents and 27 scientific tools predict stable facets, reconstruct working surfaces, enumerate and rank reaction pathways, locate transition states, and compute kinetics in 5-30 min on a single GPU.
• Score: 1.2947828742859666
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Theoretical heterogeneous catalysis promises rapid catalyst discovery, yet computational and machine-learning predictions often deviate from experiment and stay confined to narrow material families, for want of a faithful, condition-aware catalytic simulator. We present CatDT (Catalysis Digital Twin), a self-evolving multi-agent system that builds an autonomous digital twin of a working catalyst, unifying gas-solid and liquid-solid modeling. From only a bulk crystal and a natural-language reaction description, eight specialized agents and 27 scientific tools predict stable facets, reconstruct working surfaces, enumerate and rank reaction pathways, locate transition states, and compute kinetics in 5-30 min on a single GPU. Two innovations address the hardest steps: UniMech finds dominant pathways for novel materials at over $10^3\times$ lower cost than exhaustive enumeration by fusing agent-guided proposals with energy-cached graph search, and a memory-augmented reinforcement loop raises barrier-calculation success from 41\% to 84\% across 600 catalytic surfaces. Across seven gas-solid benchmarks -- stepped metals, single-atom catalysts, ordered intermetallics, vacancy-rich 2D sulfides and carbides, and a strong-metal--support-interaction (SMSI) interface -- every CatDT prediction lies within 0.5-2 times experiment over four orders of magnitude. For propane dehydrogenation, CatDT independently discovers non-precious candidates rivaling the Pt-based industrial benchmark, with a proposed Ni@ZrO$_2$ SMSI overlayer reaching a simulated TOF of $1.63~\text{s}^{-1}$ at $\sim$100\% selectivity. More broadly, the decisive factor for a faithful catalyst digital twin -- or any multi-stage scientific simulator -- is not raw LLM capability but the engineered harness around it: deterministic tools, persistent memory, and verified self-improvement that compound across models, tools, and runs.

Related papers

• Catalyst Diffusion Transformer: Generative Inverse Design of Heterogeneous Catalysts [0.18665975431697432]
  Catalyst Diffusion Transformer (CatDiT) is a unified framework for inverse catalyst design.<n>CatDiT generates valid and novel structures ranging from intermetallic alloys to oxide surfaces.
  arXiv  Detail & Related papers  (2026-07-27T11:11:33Z)
• Science Earth: Towards A Planet-Scale Operating System for AI-Native Scientific Discovery [51.70607714288616]
  Science Earth is a planet-scale scientific runtime in which any capability can connect to any other.<n>Its underlying EACN protocol lets capabilities discover one another, negotiate task ownership, and adjudicate across incompatible evidentiary standards.<n>Two runs validate this under structurally distinct conditions.
  arXiv  Detail & Related papers  (2026-05-31T16:05:41Z)
• Benchmark Dataset for Catalysis on 2D MXenes [20.3137851025657]
  We focus on two-dimensional (2D) Ti$VCT$_y$ MXenes, whose versatile surface chemistry makes them compelling candidates for study.<n>To address this challenge, we generate a comprehensive dataset of 50,000 DFT calculations for training and 10,000 for testing.<n>We train and validate widely used machine learning interatomic potential (MLIP) models, including EquiformerV2, MACE, MatRIS, and UPET, that accurately predict atomic forces and formation energies.
  arXiv  Detail & Related papers  (2026-05-30T16:18:17Z)
• Agentic Fusion of Large Atomic and Language Models to Accelerate Superconductor Discovery [85.54953285509167]
  Scaling to 2.4 million equilibrium crystals, ElementsClaw identifies 68,000 high-confidence candidates in just 28 GPU hours.<n>We experimentally synthesize and verify four novel superconductors.
  arXiv  Detail & Related papers  (2026-04-26T15:14:38Z)
• LLM-Driven Discovery of High-Entropy Catalysts via Retrieval-Augmented Generation [2.2716047672389306]
  This paper demonstrates how large language models can assist the catalyst discovery process.<n>We introduce a retrieval-augmented generation framework that enables GPT-4 to navigate chemical space.<n>Our approach generated over 250 catalyst candidates with an 82% thermodynamic stability rate.
  arXiv  Detail & Related papers  (2026-03-16T14:10:52Z)
• SimulCost: A Cost-Aware Benchmark and Toolkit for Automating Physics Simulations with LLMs [56.07550353240028]
  We introduce SimulCost, the first benchmark targeting cost-sensitive parameter tuning in physics simulations.<n>SimulCost compares LLM tuning cost-sensitive parameters against traditional scanning approach in both accuracy and computational cost.<n>Each simulator's cost is analytically defined and platform-independent.
  arXiv  Detail & Related papers  (2026-03-11T05:00:48Z)
• Catalyst-Agent: Autonomous heterogeneous catalyst screening and optimization with an LLM Agent [10.596902977676807]
  We introduce Catalyst-Agent, a server-based, LLM-powered AI agent.<n>It can explore vast material databases using the OPTIMADE API.<n> Catalyst-Agent achieves a success rate of 23-34 percent among all the materials it chooses and evaluates.
  arXiv  Detail & Related papers  (2026-03-01T22:44:56Z)
• Quantum algorithm for simulating non-adiabatic dynamics at metallic surfaces [0.034998703934432676]
  Non-adiabatic dynamics at molecule-metal interfaces govern diverse technologically important phenomena.<n>We develop a highly optimized quantum algorithm for simulating realistic molecule-metal interfaces.
  arXiv  Detail & Related papers  (2026-01-22T19:00:07Z)
• Hierarchical Multi-agent Large Language Model Reasoning for Autonomous Functional Materials Discovery [0.0]
  We introduce Materials Agents for Simulation and Theory in Electronic-structure Reasoning (Master)<n>In MASTER, a multimodal system translates natural language into density functional theory, while higher-level reasoning agents guide discovery through a hierarchy of strategies.<n> Reasoning trajectories reveal chemically grounded decisions that cannot be explained by sampling or semantic bias.
  arXiv  Detail & Related papers  (2025-12-15T22:08:18Z)
• Zero-shot Autonomous Microscopy for Scalable and Intelligent Characterization of 2D Materials [41.856704526703595]
  characterization of atomic-scale materials traditionally requires human experts with months to years of specialized training.<n>This bottleneck drives demand for fully autonomous experimentation systems capable of comprehending research objectives without requiring large training datasets.<n>We present ATOMIC, an end-to-end framework that integrates foundation models to enable fully autonomous, zero-shot characterization of 2D materials.
  arXiv  Detail & Related papers  (2025-04-14T14:49:45Z)
• Diff-PIC: Revolutionizing Particle-In-Cell Nuclear Fusion Simulation with Diffusion Models [38.46100610494588]
  Nuclear fusion, generally seen as an ultimate solution, has been the focus of intensive research for nearly a century. Recent advancements in Inertial Confinement Fusion have drawn significant attention to fusion research. Laser-Plasma Interaction is critical for ensuring fusion stability and efficiency.
  arXiv  Detail & Related papers  (2024-08-03T19:42:31Z)
• PhAST: Physics-Aware, Scalable, and Task-specific GNNs for Accelerated Catalyst Design [102.9593507372373]
  Catalyst materials play a crucial role in the electrochemical reactions involved in industrial processes. Machine learning holds the potential to efficiently model materials properties from large amounts of data. We propose task-specific innovations applicable to most architectures, enhancing both computational efficiency and accuracy.
  arXiv  Detail & Related papers  (2022-11-22T05:24:30Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.