arXiv subjects

Marc J. Cawkwell

Publications and source records attributed to Marc J. Cawkwell.

8 recordsLinked to original sources

arXiv2026

Differential Learning for Robust Prediction of Thermal Stability with Application to Energetic Materials

Predicting thermal stability during handling and storage is essential for the design of safe and reliable energetic materials. However, experimental measurements vary significantly across laboratories due to differences in protocols and analysis methods, making it difficult to train reliable predictive models. We address this challenge through differential learning. Rather than predicting absolute decomposition temperatures, we instead train message passing neural networks to predict relative differences between pairs of molecules. This approach reduces sensitivity to systematic experimental errors and achieves >85% accuracy in ranking compounds by thermal stability, outperforming conventional regression methods on the same heterogeneous dataset. To understand what drives these predictions, we compare neural network models with interpretable alternatives built from descriptors derived from ab initio calculations and cheminformatics software. This analysis identifies bond dissociation enthalpy as a key determinant of thermal stability rankings, providing further insight into the complex chemistry of thermal decomposition. The differential learning framework generalizes across model architectures, from graph neural networks to classical descriptor-based approaches. Our results demonstrate that learning relative properties rather than absolute values offers a practical solution for modeling noisy experimental data, with direct applications in materials design where thermal stability predictions inform safety protocols.

physics.chem-ph↗

arXiv2026

Generative Chemical Language Models for Energetic Materials Discovery

The discovery of new energetic materials remains a pressing challenge hindered by limited availability of high-quality data. To address this, we have developed generative molecular language models that have been pretrained on extensive chemical data and then fine-tuned with curated energetic materials datasets. This transfer-learning strategy extends the chemical language model capabilities beyond the pharmacological space in which they have been predominantly developed, offering a framework applicable to other data-spare discovery problems. Furthermore, we discuss the benefits of fragment-based molecular encodings for chemical language models, in particular in constructing synthetically accessible structures. Together, these advances provide a foundation for accelerating the design of next-generation energetic materials with demanding performance requirements.

physics.chem-ph↗

arXiv2023

Extended Lagrangian Born-Oppenheimer Molecular Dynamics with DFT+U

Extended Lagrangian Born-Oppenheimer molecular dynamics (XL-BOMD) [Phys. Rev. Lett. vol. 100, 123004 (2008)] is combined with Kohn-Sham density functional theory (DFT) using a DFT+U correction based on the Hubbard model. This combined XL-BOMD and DFT+U approach allows efficient Born-Oppenheimer molecular dynamics simulations with orbital-dependent corrections beyond regular Kohn-Sham density functional theory. The extended Lagrangian formulation eliminates the need for the iterative self-consistent-field optimization of the electronic ground state prior to the force evaluations, which is required in regular direct Born-Oppenheimer molecular dynamics simulations. This method provides accurate and stable molecular trajectories, while reducing the computational cost per time step. The combined XL-BOMD and DFT+U approach is demonstrated with molecular dynamics simulations of a nitromethane molecular liquid and a system of solid nuclear fuel, UO$_2$, using self-consistent-charge density functional based tight-binding theory.

physics.chem-ph↗

arXiv2023

Spin-Polarized Extended Lagrangian Born-Oppenheimer Molecular Dynamics

We present a generalization of Extended Lagrangian Born-Oppenheimer molecular dynamics [Phys. Rev. Lett. vol. 100, 123004 (2008); Eur. Phys. J. B vol. 94, 164 (2021)] that also includes the electronic spin-degrees of freedom as extended dynamical variables. To integrate the combined spin and charge degrees of freedom, we use a preconditioned low-rank Krylov subspace approximation. Our approach is demonstrated for quantum-mechanical molecular dynamics simulations of iron, using spin-polarized self-consistent charge density functional tight-binding theory. We also show how the low-rank Krylov subspace approximation can be used to accelerate the self-consistent field convergence.

physics.chem-ph↗

arXiv2016

Graph-based linear scaling electronic structure theory

We show how graph theory can be combined with quantum theory to calculate the electronic structure of large complex systems. The graph formalism is general and applicable to a broad range of electronic structure methods and materials, including challenging systems such as biomolecules. The methodology combines well-controlled accuracy, low computational cost, and natural low-communication parallelism. This combination addresses substantial shortcomings of linear scaling electronic structure theory, in particular with respect to quantum-based molecular dynamics simulations.

physics.comp-ph↗

arXiv2015

Canonical density matrix perturbation theory

Density matrix perturbation theory [Niklasson and Challacombe, Phys. Rev. Lett. 92, 193001 (2004)] is generalized to canonical (NVT) free energy ensembles in tight-binding, Hartree-Fock or Kohn-Sham density functional theory. The canonical density matrix perturbation theory can be used to calculate temperature dependent response properties from the coupled perturbed self-consistent field equations as in density functional perturbation theory. The method is well suited to take advantage of sparse matrix algebra to achieve linear scaling complexity in the computational cost as a function of system size for sufficiently large non-metallic materials and metals at high temperatures.

physics.chem-ph↗

arXiv2014

Generalized Extended Lagrangian Born-Oppenheimer Molecular Dynamics

Extended Lagrangian Born-Oppenheimer molecular dynamics based on Kohn-Sham density functional theory is generalized in the limit of vanishing self-consistent field optimization prior to the force evaluations. The equations of motion are derived directly from the extended Lagrangian under the condition of an adiabatic separation between the nuclear and the electronic degrees of freedom. We show how this separation is automatically fulfilled and system independent. The generalized equations of motion require only one diagonalization per time step and are applicable to a broader range of materials with improved accuracy and stability compared to previous formulations.

physics.chem-ph↗

arXiv2012

Fast method for quantum mechanical molecular dynamics

With the continuous growth of processing power for scientific computing, first principles Born-Oppenheimer molecular dynamics (MD) simulations are becoming increasingly popular for the study of a wide range of problems in materials science, chemistry and biology. Nevertheless, the computational cost still remains prohibitively large in many cases, particularly in comparison to classical MD simulations using empirical force fields. Here we show how to circumvent the major computational bottleneck in Born-Oppenheimer MD simulations arising from the self-consistent-charge optimization. The optimization-free quantum mechanical MD method is demonstrated for density functional tight-binding theory. The molecular trajectories are almost indistinguishable from an "exact" microcanonical Born-Oppenheimer MD simulation even when linear scaling sparse matrix algebra is used. Our findings drastically reduce the computational gap between classical and quantum mechanical MD simulations.

cond-mat.mtrl-sci↗