Research paperComputational MLIPComputed RamanComputational DFTLarge Scale Raman Spectrum Calculations in Defective 2D Materials using Deep LearningOlivier Malenfant-Thuot, Dounia Shaaban Kabakibo, Simon Blackburn, Bruno Rousseau et al.arXiv·2025·10.1088/1361-648x/ada106·arXiv:2410.20417AbstractWe introduce a machine learning prediction workflow to study the impact of defects on the Raman response of 2D materials. By combining machine-learned interatomic potentials, the Raman-active Γ-weighted density of states method, and splitting configurations into independent patches, we reach simulation sizes in the tens of thousands of atoms, with diagonalization becoming the main bottleneck. The method is applied to isotopic graphene and defective hexagonal boron nitride, and the predicted Raman response is compared with experiment with good agreement.Read more
Pristine and isotopically modified graphene configurations used as DFT-labeled training data and Raman workflow inputs.1 characterizationSimulated Supercell DftCStudied MaterialExpand
Pristine and vacancy-defective hBN configurations used as DFT-labeled training data and Raman workflow inputs.1 characterizationSimulated Supercell Dfth-BNStudied MaterialExpand
Research paperComputational MLIPComputed RamanComputational DFTLarge Scale Raman Spectrum Calculations in Defective 2D Materials using Deep LearningOlivier Malenfant-Thuot, Dounia Shaaban Kabakibo, Simon Blackburn, Bruno Rousseau et al.arXiv·2025·10.1088/1361-648x/ada106·arXiv:2410.20417AbstractWe introduce a machine learning prediction workflow to study the impact of defects on the Raman response of 2D materials. By combining machine-learned interatomic potentials, the Raman-active Γ-weighted density of states method, and splitting configurations into independent patches, we reach simulation sizes in the tens of thousands of atoms, with diagonalization becoming the main bottleneck. The method is applied to isotopic graphene and defective hexagonal boron nitride, and the predicted Raman response is compared with experiment with good agreement.Read more
Pristine and isotopically modified graphene configurations used as DFT-labeled training data and Raman workflow inputs.1 characterizationSimulated Supercell DftCStudied MaterialExpand
Pristine and vacancy-defective hBN configurations used as DFT-labeled training data and Raman workflow inputs.1 characterizationSimulated Supercell Dfth-BNStudied MaterialExpand
Research paperComputational MLIPComputed RamanComputational DFTLarge Scale Raman Spectrum Calculations in Defective 2D Materials using Deep LearningOlivier Malenfant-Thuot, Dounia Shaaban Kabakibo, Simon Blackburn, Bruno Rousseau et al.arXiv·2025·10.1088/1361-648x/ada106·arXiv:2410.20417AbstractWe introduce a machine learning prediction workflow to study the impact of defects on the Raman response of 2D materials. By combining machine-learned interatomic potentials, the Raman-active Γ-weighted density of states method, and splitting configurations into independent patches, we reach simulation sizes in the tens of thousands of atoms, with diagonalization becoming the main bottleneck. The method is applied to isotopic graphene and defective hexagonal boron nitride, and the predicted Raman response is compared with experiment with good agreement.Read more
Pristine and isotopically modified graphene configurations used as DFT-labeled training data and Raman workflow inputs.1 characterizationSimulated Supercell DftCStudied MaterialExpand
Pristine and vacancy-defective hBN configurations used as DFT-labeled training data and Raman workflow inputs.1 characterizationSimulated Supercell Dfth-BNStudied MaterialExpand
Research paperComputational MLIPComputed RamanComputational DFTLarge Scale Raman Spectrum Calculations in Defective 2D Materials using Deep LearningOlivier Malenfant-Thuot, Dounia Shaaban Kabakibo, Simon Blackburn, Bruno Rousseau et al.arXiv·2025·10.1088/1361-648x/ada106·arXiv:2410.20417AbstractWe introduce a machine learning prediction workflow to study the impact of defects on the Raman response of 2D materials. By combining machine-learned interatomic potentials, the Raman-active Γ-weighted density of states method, and splitting configurations into independent patches, we reach simulation sizes in the tens of thousands of atoms, with diagonalization becoming the main bottleneck. The method is applied to isotopic graphene and defective hexagonal boron nitride, and the predicted Raman response is compared with experiment with good agreement.Read more
Pristine and isotopically modified graphene configurations used as DFT-labeled training data and Raman workflow inputs.1 characterizationSimulated Supercell DftCStudied MaterialExpand
Pristine and vacancy-defective hBN configurations used as DFT-labeled training data and Raman workflow inputs.1 characterizationSimulated Supercell Dfth-BNStudied MaterialExpand