The Mendoza Lab is a theory and computation group in the Department of Chemical Engineering and Materials Science at Michigan State University, with an affiliation in Physics and Astronomy. We work like renaissance scientists: problems are chosen for their importance, and we bring whatever physics, chemistry, mathematics or computer science they need. Since 2005 Prof. Mendoza-Cortés and his collaborators have published 69 peer-reviewed articles (37 of them since the lab moved to MSU in 2020), 7 preprints and 2 open textbooks, cited more than 10,000 times.
What we work on
Latest news
-
Paper
A ferrocene-functionalized framework that exceeds the ultimate hydrogen-storage targets
Marcus Djokic designed a covalent organic framework with ferrocene binding sites that, in multiscale simulations, stores 18 wt% and 72.6 g/L of hydrogen at room temperature and 700 bar, beyond the U.S. Department of Energy ultimate targets. npj Energy Materials; provisional patent filed.
-
Talk
Invited seminar at Central Michigan University
From next-generation AI-powered force fields to automated discovery and heavy-element chemistry: an overview of the group's newest methods.
-
Paper
How sodium reshapes layered birnessite
With Tec de Monterrey, we show from first principles how inserting sodium changes the structure and electronic states of layered potassium birnessite, a candidate electrode and catalyst. The Journal of Chemical Physics 165, 084703.
-
Grant
NSF funds our study of synthetic mimics of the oxygen-evolving complex
A collaborative NSF Chemistry award with Michael Zdilla (Temple University) will use theory and synthesis to understand how heterocubane clusters, topological mimics of the water-oxidizing center of Photosystem II, release oxygen. Award CHE-2555662, 2026 to 2029.
-
Talk
Invited talk at Los Alamos National Laboratory
Fantastic simulations and where to find them: next-generation AI-powered force fields, automated discovery and the secrets of heavy elements.
-
Milestone
Congratulations, Dr. Maldonado-Lopez and Dr. Rodriguez
Daniel Maldonado-Lopez (2D materials, ammonia production and porous materials) and Austin Rodriguez (machine-learned interatomic potentials for reactions) completed their Ph.D. degrees in Chemical Engineering and Materials Science.
Recent publications
-
Projected Hessian Learning: Fast Curvature Supervision for Accurate Machine-Learning Interatomic Potentials
Machine Learning: Science and Technology 2026, 7, 055037
Training machine-learning interatomic potentials on Hessians, the second derivatives that describe the curvature of a potential energy surface, improves accuracy, but full Hessians are costly to compute and store. We introduce Projected Hessian Learning, which supervises curvature using only Hessian-vector products along random probe directions, and benchmark it on reactants, products, transition states and reaction paths. It matches full-Hessian training accuracy for energies, forces and Hessians while training more than 24 times faster per epoch.
-
How Does Intercalation Reshape Layered Structures? A First-Principles Study of Sodium Insertion in Layered Potassium Birnessite
The Journal of Chemical Physics 2026, 165, 084703
Layered manganese oxides such as birnessite can host ions between their sheets, but how intercalation reshapes their structure and electronic properties is not well understood. Using hybrid density functional theory, we follow sodium insertion into potassium birnessite through formation energies, simulated Raman and X-ray diffraction patterns, diffusion barriers and binding energies. Intercalation shifts manganese oxidation states and distorts the lattice, and some intercalated structures behave as bipolar magnetic semiconductors, suggesting uses in spintronics.
-
Ferrocene-functionalized covalent organic framework exceeding the ultimate hydrogen storage targets: a first-principles multiscale computational study
npj Energy Materials 2026, 1, 18
Storing enough hydrogen in a compact, lightweight tank is a key hurdle for hydrogen-powered vehicles, for which the Department of Energy sets targets of 6.5 wt% and 50 g/L. Following our multi-binding-sites strategy for covalent organic frameworks (COFs), we design MSUCOF-4-FeCp by adding ferrocene groups to the IRCOF-102 framework and evaluate it with first-principles multiscale simulations. At room temperature and 700 bar it stores 18.0 wt% and 72.6 g/L, a lower-cost alternative to precious-metal COFs.
-
Leveraging Polymorphism in YbCuBi to Map Transport and Elastic Properties
Chemistry of Materials 2026, 38, 3, 1425–1434
Zintl compounds with honeycomb layers, such as YbCuBi, combine vacancy tolerance with low thermal conductivity, but how their structure controls heat transport is unclear. Below 410 K, YbCuBi changes from flat to corrugated copper–bismuth layers, and we exploit this polymorphism using inelastic neutron scattering, resonant ultrasound spectroscopy, thermal conductivity measurements and first-principles calculations. The results tie honeycomb corrugation directly to elastic and thermal transport properties, guiding the design of materials with tailored thermal behavior.
Join us. Ph.D. applicants can come through Chemical Engineering, Materials Science, Physics, Chemistry, CMSE, Mathematics and other MSU programs. Graduates of the group become academic descendants of Marie Curie and of Ralph Fowler, Paul Dirac's doctoral advisor, and anyone who publishes with us gets an Erdős number of at most 6. Read how to join.