Alexander Urban
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Alexander Urban has a total of 129 co-authors. Here are the top co-authors by number of shared publications:
- Nongnuch Artrith
Utrecht University
10 shared publications - Alan C. West
Columbia University
3 shared publications In Won Yeu
Department of Chemical Engineering, Columbia University, New York, USA
3 shared publicationsJames M. Stevenson
Schrödinger, New York, USA
3 shared publicationsQian Wang
Department of Chemical Engineering, Columbia University, New York, New York 10027, United States
3 shared publicationsXinhao Li
Department of Chemical Engineering, Columbia University, New York, New York 10027, United States
3 shared publications
Latest Publications
- Open Access
- 14 October 2026
- Published by Springer Nature in Nature
- Vol. 658 (8136) , 636-642
Evidence for multi-Li+ clustering in lithium-ion battery electrolytes
- 14 October 2026
- Published by Springer Nature in Nature
- Vol. 658 (8136) , 636-642
Open AccessEvidence for multi-Li+ clustering in lithium-ion battery electrolytes
- 5 November 2025
- Published by American Chemical Society (ACS) in ACS Sustainable Resource Management
- Vol. 2 (11) , 2283-2291
Fast and Selective Leaching of Pyrrhotite in the Presence of Pentlandite
- 5 November 2025
- Published by American Chemical Society (ACS) in ACS Sustainable Resource Management
- Vol. 2 (11) , 2283-2291
Fast and Selective Leaching of Pyrrhotite in the Presence of Pentlandite
- Open Access
- 22 September 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Evidence for significant multi-Li+ clustering in common lithium-ion battery electrolytes
Lithium-ion battery (LIB) electrolytes have been empirically optimized to maximize battery lifetime and charge rate, but a detailed understanding of how their atomic-scale structure controls their electrochemical behaviors remains lacking. Here, we combine machine-learning-enabled atomistic simulations, nuclear magnetic resonance measurements, electrochemical data, and quantum-chemical calculations to reveal previously unstudied multi-Li+ complexes without anions that are uniquely facile for reduction. This surprising insight explains the success of empirically-chosen solvent mixtures and provides a foundation for the rational design of improved electrolytes. Furthermore, our machine learning force field (MLFF) based LIB electrolyte simulations accurately reproduce a wide range of condensed phase experiments, despite being parameterized only to quantum chemical gas phase cluster calculations, and hence provide a compelling demonstration that MLFF models developed from first principles can robustly represent a complex chemical system to atomic resolution.
- 22 September 2025
- Published by American Chemical Society (ACS) in ChemRxiv
Open AccessEvidence for significant multi-Li+ clustering in common lithium-ion battery electrolytes
Lithium-ion battery (LIB) electrolytes have been empirically optimized to maximize battery lifetime and charge rate, but a detailed understanding of how their atomic-scale structure controls their electrochemical behaviors remains lacking. Here, we combine machine-learning-enabled atomistic simulations, nuclear magnetic resonance measurements, electrochemical data, and quantum-chemical calculations to reveal previously unstudied multi-Li+ complexes without anions that are uniquely facile for reduction. This surprising insight explains the success of empirically-chosen solvent mixtures and provides a foundation for the rational design of improved electrolytes. Furthermore, our machine learning force field (MLFF) based LIB electrolyte simulations accurately reproduce a wide range of condensed phase experiments, despite being parameterized only to quantum chemical gas phase cluster calculations, and hence provide a compelling demonstration that MLFF models developed from first principles can robustly represent a complex chemical system to atomic resolution.
- 16 July 2025
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 163 (3)
r2SCAN+rVV10+U parameterization for 3d transition metal sulfides for thermochemistry
Transition metal sulfides play a critical role in energy storage, catalysis, and metallurgy. Here, we assess the predictive accuracy of density-functional theory with the r2SCAN functional for thermochemical properties of transition-metal sulfides. We find that a correction of the method’s intrinsic self-interaction error can further improve the accuracy of phase diagrams compared with experimental observation. A systematic Hubbard-U parameterization across 22 transition metal sulfide systems resulted in a ∼70% reduction in the root mean square error of formation energies. r2SCAN+U predicts binary and ternary phase diagrams as well as Pourbaix diagrams in excellent agreement with experimental references and captures phase stability trends that are incorrectly described by uncorrected r2SCAN. These advancements establish a reliable and computationally efficient framework for modeling transition-metal sulfides with broad applicability to the design of energy storage materials, electrocatalysts, and sustainable metallurgical processes.
- 16 July 2025
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 163 (3)
r2SCAN+rVV10+U parameterization for 3d transition metal sulfides for thermochemistry
Transition metal sulfides play a critical role in energy storage, catalysis, and metallurgy. Here, we assess the predictive accuracy of density-functional theory with the r2SCAN functional for thermochemical properties of transition-metal sulfides. We find that a correction of the method’s intrinsic self-interaction error can further improve the accuracy of phase diagrams compared with experimental observation. A systematic Hubbard-U parameterization across 22 transition metal sulfide systems resulted in a ∼70% reduction in the root mean square error of formation energies. r2SCAN+U predicts binary and ternary phase diagrams as well as Pourbaix diagrams in excellent agreement with experimental references and captures phase stability trends that are incorrectly described by uncorrected r2SCAN. These advancements establish a reliable and computationally efficient framework for modeling transition-metal sulfides with broad applicability to the design of energy storage materials, electrocatalysts, and sustainable metallurgical processes.
- Open Access
- 28 May 2025
- Published by Springer Nature in npj Computational Materials
- Vol. 11 (1) , 1-13
Scalable training of neural network potentials for complex interfaces through data augmentation
- 28 May 2025
- Published by Springer Nature in npj Computational Materials
- Vol. 11 (1) , 1-13
Open AccessScalable training of neural network potentials for complex interfaces through data augmentation
- Open Access
Scalable Training of Neural Network Potentials for Complex Interfaces Through Data Augmentation
Artificial neural network (ANN) potentials enable highly accurate atomistic simulations of complex materials at unprecedented scales. Despite their promise, training ANN potentials to represent intricate potential energy surfaces (PES) with transferability to diverse chemical environments remains computationally intensive, especially when atomic force data are incorporated to improve PES gradients. Here, we present an efficient ANN potential training methodology that uses Gaussian process regression (GPR) to incorporate atomic forces into ANN training, leading to accurate PES models with fewer additional first-principles calculations and a reduced computational effort for training. Our GPR-ANN approach generates synthetic energy data from force information in the reference dataset, thus augmenting the training datasets and bypassing direct force training. Benchmark tests on hybrid density-functional theory data for ethylene carbonate (EC) molecules and Li metal-EC interfaces, relevant for lithium metal battery applications, demonstrate that GPR-ANN potentials achieve accuracies comparable to fully force-trained ANNs with a significantly reduced computational overhead. Detailed comparisons show that the method improves both data efficiency and scalability for complex interfaces and heterogeneous environments. This work establishes the GPR-ANN method as a powerful and scalable framework for constructing high-fidelity machine learning interatomic potentials, offering the computational and memory efficiency critical for the large-scale simulations needed for the simulation of materials interfaces.
Open AccessScalable Training of Neural Network Potentials for Complex Interfaces Through Data Augmentation
Artificial neural network (ANN) potentials enable highly accurate atomistic simulations of complex materials at unprecedented scales. Despite their promise, training ANN potentials to represent intricate potential energy surfaces (PES) with transferability to diverse chemical environments remains computationally intensive, especially when atomic force data are incorporated to improve PES gradients. Here, we present an efficient ANN potential training methodology that uses Gaussian process regression (GPR) to incorporate atomic forces into ANN training, leading to accurate PES models with fewer additional first-principles calculations and a reduced computational effort for training. Our GPR-ANN approach generates synthetic energy data from force information in the reference dataset, thus augmenting the training datasets and bypassing direct force training. Benchmark tests on hybrid density-functional theory data for ethylene carbonate (EC) molecules and Li metal-EC interfaces, relevant for lithium metal battery applications, demonstrate that GPR-ANN potentials achieve accuracies comparable to fully force-trained ANNs with a significantly reduced computational overhead. Detailed comparisons show that the method improves both data efficiency and scalability for complex interfaces and heterogeneous environments. This work establishes the GPR-ANN method as a powerful and scalable framework for constructing high-fidelity machine learning interatomic potentials, offering the computational and memory efficiency critical for the large-scale simulations needed for the simulation of materials interfaces.
- 18 November 2024
- Published by American Chemical Society (ACS) in ACS Applied Energy Materials
- Vol. 7 (23) , 10784-10794
Mechanism of the Layered-to-Spinel Phase Transformation in Li0.5NiO2
- 18 November 2024
- Published by American Chemical Society (ACS) in ACS Applied Energy Materials
- Vol. 7 (23) , 10784-10794
Mechanism of the Layered-to-Spinel Phase Transformation in Li0.5NiO2
- Open Access
- 28 May 2024
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 160 (20)
Discovering melting temperature prediction models of inorganic solids by combining supervised and unsupervised learning
The melting temperature is important for materials design because of its relationship with thermal stability, synthesis, and processing conditions. Current empirical and computational melting point estimation techniques are limited in scope, computational feasibility, or interpretability. We report the development of a machine learning methodology for predicting melting temperatures of binary ionic solid materials. We evaluated different machine-learning models trained on a dataset of the melting points of 476 non-metallic crystalline binary compounds using materials embeddings constructed from elemental properties and density-functional theory calculations as model inputs. A direct supervised-learning approach yields a mean absolute error of around 180 K but suffers from low interpretability. We find that the fidelity of predictions can further be improved by introducing an additional unsupervised-learning step that first classifies the materials before the melting-point regression. Not only does this two-step model exhibit improved accuracy, but the approach also provides a level of interpretability with insights into feature importance and different types of melting that depend on the specific atomic bonding inside a material. Motivated by this finding, we used a symbolic learning approach to find interpretable physical models for the melting temperature, which recovered the best-performing features from both prior models and provided additional interpretability.
- 28 May 2024
- Published by AIP Publishing in The Journal of Chemical Physics
- Vol. 160 (20)
Open AccessDiscovering melting temperature prediction models of inorganic solids by combining supervised and unsupervised learning
The melting temperature is important for materials design because of its relationship with thermal stability, synthesis, and processing conditions. Current empirical and computational melting point estimation techniques are limited in scope, computational feasibility, or interpretability. We report the development of a machine learning methodology for predicting melting temperatures of binary ionic solid materials. We evaluated different machine-learning models trained on a dataset of the melting points of 476 non-metallic crystalline binary compounds using materials embeddings constructed from elemental properties and density-functional theory calculations as model inputs. A direct supervised-learning approach yields a mean absolute error of around 180 K but suffers from low interpretability. We find that the fidelity of predictions can further be improved by introducing an additional unsupervised-learning step that first classifies the materials before the melting-point regression. Not only does this two-step model exhibit improved accuracy, but the approach also provides a level of interpretability with insights into feature importance and different types of melting that depend on the specific atomic bonding inside a material. Motivated by this finding, we used a symbolic learning approach to find interpretable physical models for the melting temperature, which recovered the best-performing features from both prior models and provided additional interpretability.
- Open Access
Discovering Melting Temperature Prediction Models of Inorganic Solids by Combining Supervised and Unsupervised Learning
The melting temperature is important for materials design because of its relationship with thermal stability, synthesis, and processing conditions. Current empirical and computational melting point estimation techniques are limited in scope, computational feasibility, or interpretability. We report the development of a machine learning methodology for predicting melting temperatures of binary ionic solid materials. We evaluated different machine-learning models trained on a data set of the melting points of 476 non-metallic crystalline binary compounds, using materials embeddings constructed from elemental properties and density-functional theory calculations as model inputs. A direct supervised-learning approach yields a mean absolute error of around 180~K but suffers from low interpretability. We find that the fidelity of predictions can further be improved by introducing an additional unsupervised-learning step that first classifies the materials before the melting-point regression. Not only does this two-step model exhibit improved accuracy, but the approach also provides a level of interpretability with insights into feature importance and different types of melting that depend on the specific atomic bonding inside a material. Motivated by this finding, we used a symbolic learning approach to find interpretable physical models for the melting temperature, which recovered the best-performing features from both prior models and provided additional interpretability.
Open AccessDiscovering Melting Temperature Prediction Models of Inorganic Solids by Combining Supervised and Unsupervised Learning
The melting temperature is important for materials design because of its relationship with thermal stability, synthesis, and processing conditions. Current empirical and computational melting point estimation techniques are limited in scope, computational feasibility, or interpretability. We report the development of a machine learning methodology for predicting melting temperatures of binary ionic solid materials. We evaluated different machine-learning models trained on a data set of the melting points of 476 non-metallic crystalline binary compounds, using materials embeddings constructed from elemental properties and density-functional theory calculations as model inputs. A direct supervised-learning approach yields a mean absolute error of around 180~K but suffers from low interpretability. We find that the fidelity of predictions can further be improved by introducing an additional unsupervised-learning step that first classifies the materials before the melting-point regression. Not only does this two-step model exhibit improved accuracy, but the approach also provides a level of interpretability with insights into feature importance and different types of melting that depend on the specific atomic bonding inside a material. Motivated by this finding, we used a symbolic learning approach to find interpretable physical models for the melting temperature, which recovered the best-performing features from both prior models and provided additional interpretability.
- Open Access
Constructing and Compressing Global Moment Descriptors from Local Atomic Environments
Local atomic environment descriptors (LAEDs) are used in the materials science and chemistry communities, for example, for the development of machine learning interatomic potentials. Despite the fact that LAEDs have been extensively studied and benchmarked for various applications, global structure descriptors (GSDs), i.e., descriptors for entire molecules or crystal structures, have been mostly developed independently based on other approaches. Here, we propose a systematically improvable methodology for constructing a space of representations of GSDs from LAEDs by incorporating statistical information and information about chemical elements. We apply the method to construct GSDs of varying complexity for lithium thiophosphate structures that are of interest as solid electrolytes and use an information-theoretic approach to obtain an optimally compressed GSD. Finally, we report the performance of the compressed GSD for energy prediction tasks.
Open AccessConstructing and Compressing Global Moment Descriptors from Local Atomic Environments
Local atomic environment descriptors (LAEDs) are used in the materials science and chemistry communities, for example, for the development of machine learning interatomic potentials. Despite the fact that LAEDs have been extensively studied and benchmarked for various applications, global structure descriptors (GSDs), i.e., descriptors for entire molecules or crystal structures, have been mostly developed independently based on other approaches. Here, we propose a systematically improvable methodology for constructing a space of representations of GSDs from LAEDs by incorporating statistical information and information about chemical elements. We apply the method to construct GSDs of varying complexity for lithium thiophosphate structures that are of interest as solid electrolytes and use an information-theoretic approach to obtain an optimally compressed GSD. Finally, we report the performance of the compressed GSD for energy prediction tasks.