Aniruddha Seal

Graduate Research

To design catalysts that perform under industrial conditions, we must efficiently simulate their dynamical behaviour with an accurate quantum-mechanical description. My graduate research addresses this by integrating multireference electronic-structure theory with machine-learned potentials and enhanced-sampling techniques to improve both accuracy and efficiency in molecular simulations, and by developing machine-learning–based algorithms for high-dimensional manifolds that capture multi-pathway chemical transitions with explicit many-body interactions.

Getting Multi-reference Electronic Structure ML ready

Machine learning interatomic potentials (MLPs) trained on multireference electronic-structure data offer a promising path for efficient simulation of strongly correlated systems. However, the development of these MLPs has been limited by inconsistent active space definitions (in multireference electronic-structure) across molecular geometries. I developed the weighted active space protocol (WASP), an systematic method for assigning consistent active spaces that allows training of multireference MLPs. WASP works by assigning the active space as a weighted sum of that of it's neigbours. This approach enables for the first time transition metal-based catalytic dynamics with multi-reference accuracy.

Seal, A., Perego, S., Hennefarth, M.R., Raucci, U., Bonati, L., Ferguson, A.L., Parrinello, M., Gagliardi, L. Weighted Active Space Protocol for Multireference Machine-Learned Potentials Proc. Natl. Acad. Sci. USA, Accepted Manuscript (2025)
Code Developed for this Project: github.com/GagliardiGroup/wasp

Enhanced Sampling Multi-Reference Reaction Dynamics

Accurately capturing chemical reactions requires multi-reference electronic structure methods to describe the topological features of potential energy surfaces. While most computational effort is usually spent overcoming the reaction barrier, the most intriguing dynamics often occur after crossing it. To address this, I introduce enhanced sampling with multi-reference electronic structure that accelerate barrier crossing by up to 10³–10⁴ times without corrupting the transition dynamics. This allows computational resources to focus on post-barrier dynamics, enabling faster estimation of rate constants with dynamical effects.

Seal, A., Gagliardi, L., Ferguson, A.L. Computing Reaction Kinetics with MC-PDFT-OPESf: Combining Multireference Electronic Structure and Enhanced Sampling Submitted (2025)
Code Developed for this Project: github.com/Ferg-Lab/PDFT_OPESf

Masters Research

Capturing Intra-molecular Vibrational Energy Flow

Vibrational energy flow dictates the reactivity of a molecule. Despite great progress in this area, most available methods operate on model Hamiltonians, restricting the pathways for energy transfer. Working with Upakarsamy Lourderaj (NISER), I developed a computational algorithm based on wavelet transform, that can track vibrational energy flow from full-dimensional on-the-fly trajectory simulations. We applied it to understand non-statistical dynamics of guanidine isomerization and thermal deazetization reactions.

Rashmi, R., Yadav, P. K., Seal, A., Paranjothy, M., and Lourderaj, U. E-Z Isomerization in Guanidine: Second-order Saddle Dynamics, Non-statisticality, and Time-frequency Analysis ChemPhysChem, 24.2, e202200640 (2023)
Yadav, K., Seal, A. and Lourderaj, U. Mechanisms and Dynamics of the Thermal Deazetization of 2,3-diazabicyclo[2.2.1]hept-2-ene (under preparation)
Code Developed for this Project: github.com/aniruddha-seal/pca-wavelet

Undergraduate Research

Chemistry-aware Modelling of Electrical Double Layers

Electrical double layers (EDLs) arise when an electrolyte is in contact with a charged surface, and can be found in batteries and many more devices. Over the last century, the development of Poisson-Boltzmann (PB) models have provided significant physical insight into the EDL structure and dynamics. However, a prominent knowledge gap has been the exclusion of van der Waals (vdW) interactions, which brings in the chemical specificity. With Ananth Govind Rajan (IISc Bengaluru), I developed an analytical extension to Gupta-Stone PB theory with the vdW interactions treated using Lennard-Jones potential.

Seal, A., Tiwari, U., Gupta, A., and Govind Rajan, A. Incorporating Ion-Specific van der Waals and Soft Repulsive Interactions in the Poisson-Boltzmann Theory of Electrical Double Layers Phys. Chem. Chem. Phys., Accepted Manuscript (2023)
Code Developed for this Project: github.com/agrgroup/PB-LJ

ADF-STEM images from Warner Group, UT Austin

Atomistic Modelling of Defects on 2D Materials

Defects are ubiquitous in nanomaterials, yet their role in modulating fluid flow is not well understood and often ignored in computational modelling. In this project, we developed a self-consistent protocol to model vacancy defects on 2D materials that involve transferring plane-wave density to partial charges. Combining this with classical MD simulations, the effect of defect composition and concentration on the structure and dynamics of interfacial fluid was studied at the hexagonal boron nitride(hBN)-water interface. We proposed design principles for defects based on our simulations to control fluid flow on 2D hBN layers.

Seal, A. and Govind Rajan, A. Modulating Water Slip Using Atomic-Scale Defects: Friction on Realistic Hexagonal Boron Nitride Surfaces Nano Lett., 21.19, 8008-8016 (2021).

Chemical Basis of Membrane Organization

Biological membrane is a complex self-assembly of lipids, sterols and proteins organized as a fluid bilayer of two closely stacked lipid leaflets. Differential molecular interactions among its constituents give rise to heterogeneities in the membrane lateral organization. Under certain conditions, heterogeneities in the two leaflets can be spatially synchronised and exist as registered domains across the bilayer. With Anand Srivastava (IISc Bengaluru), we developed an analytical theory to elucidate the driving forces that create and maintain domain registry across leaflets. I developed a computational workflow to parametrize artificial lipid molecules for all-atom simulations.

Sharma, A., Seal, A., Iyer, S. S. and Srivastava, A. Enthalpic and Entropic Contributions to Interleaflet Coupling Drive Domain Registration and Anti-registration in Biological Membrane Phys. Rev. E, 105.4, 044408 (2022)
Code Developed for this Project: github.com/codesrivastavalab/Membrane-Nanodomain-Registration/

Nat. Rev. Mol. Cell Biol, 18, 361–374 (2017)

Role of Water in Rhodopsin Ion Conduction

With Suman Chakrabarty (SNBNCBS), I studied the the localisation, structure, dynamics and energetics of the water molecules along the channel of a sodium-pumping rhodopsin ion channel. We identified the movement of the trapped water molecules to be gated by protein side-chains which controls ion-pumping, suggesting a water-mediated pathway.

Santra, M., Seal, A., Bhattacharjee, K. and Chakrabarty, S. Structural and dynamical heterogeneity of water trapped inside Na+-pumping KR2 rhodopsin in the dark state J. Chem. Phys., 154.21, 215101 (2021)

Page updated

Google Sites

Report abuse