Quantum Simulation of Molecules

From electron spins to enzymes, and from classical to quantum computers.

We develop and apply quantum-chemical and quantum-dynamical methods to understand how electronic structure, spin, and nuclear motion work together to determine the properties of molecules. Most of our work involves open-shell systems, where several electronic states lie close in energy and coupling between them, through spin–orbit and vibronic interactions, controls magnetism, spectroscopy, and photophysics. The same principles also guide our studies of enzyme reactions and our work on algorithms for quantum computers.

Methods and software used in the group →

1. Spin-state chemistry and molecular magnetism

Transition-metal complexes with nearly degenerate spin states can switch between them with temperature, light, or a change in coordination. The same near-degeneracy gives rise to large magnetic anisotropy, which is the basis of single-molecule magnets. We use multireference ab initio methods to compute spin-state energetics, zero-field splitting, and g-tensors, and relate them to the ligand field and molecular geometry.

Questions we ask

Methods: CASSCF/NEVPT2, QD-NEVPT2, spin–orbit state interaction, ab initio ligand-field theory, spin Hamiltonians, MCTDH.

Selected papers

Back to top

2. Molecular spin qubits

Paramagnetic molecules with an isolated S = 1/2 ground state, such as Cu(II), V(IV), and Cr(V) complexes, are candidate qubits for quantum information. Their usefulness is limited by spin relaxation, which happens mainly through coupling of the spin to molecular vibrations and lattice phonons. We compute this spin–vibrational coupling from first principles, identify which vibrational modes cause relaxation, and propose ligand designs that suppress it.

Questions we ask

Methods: multireference and DFT spin-Hamiltonian parameters, normal-mode spin–phonon coupling, relaxation-rate models.

Selected papers

Back to top

3. Vibronic and spin–orbit coupling in spectroscopy and photophysics

When electronic states are close in energy, the Born–Oppenheimer picture breaks down: vibrations mix the states (Jahn–Teller and Renner–Teller effects), and spin–orbit coupling mixes states of different spin. We build model vibronic Hamiltonians from ab initio data and solve the nuclear dynamics quantum mechanically. This lets us simulate and assign photoelectron and photodetachment spectra, and explain processes such as fluorescence quenching and luminescence in transition-metal and organic chromophores.

Questions we ask

Methods: diabatic vibronic-coupling Hamiltonians, MCTDH/ML-MCTDH wavepacket dynamics, Dyson orbitals, multireference electronic structure.

Selected papers

Back to top

4. Quantum computing for chemistry

Quantum computers may one day solve electronic-structure and dynamics problems that are out of reach for classical computers. Current hardware is small and noisy, so algorithms must use as few qubits and gates as possible. We develop variational algorithms for ground states and quantum dynamics, reduce qubit Hamiltonians, and test the methods on simulators and real quantum hardware. This work is carried out with the group of Prof. Sonjoy Majumder (Physics, IIT Kharagpur).

Questions we ask

Methods: VQE and its variants (SHARC-VQE), projective variational quantum dynamics (BLOP-VQD), Jordan–Wigner and Bravyi–Kitaev mappings; Qiskit, PySCF.

Selected papers

Back to top

5. Enzyme mechanisms and biomolecular simulation

We study how enzymes catalyze reactions, how mutations and metal substitution change their activity, and how conformational changes control substrate binding. Hybrid QM/MM simulations describe bond breaking and formation at the active site, and classical molecular dynamics follows the larger-scale protein motion. The main systems we study are bacterial enzymes of the lysine biosynthesis pathway (DapE, DapF) as antibiotic targets, lipoxygenases, and flavin-dependent dehalogenases, including iodotyrosine deiodinase.

Questions we ask

Methods: QM/MM molecular dynamics, free-energy simulations, classical MD and network analysis, docking and virtual screening.

Selected papers

Back to top

6. Electronic structure for functional materials

Together with experimental groups, we use electronic-structure calculations to explain and predict the properties of organic semiconductors, redox-active polymers for energy storage, battery redox mediators, photocatalysts, and fluorescent sensors. Our part is usually to link molecular structure to charge transport, redox potentials, excited states, and reaction pathways.

Methods: DFT and TD-DFT, charge-transfer and mobility models, periodic calculations (CP2K), reaction-path analysis.

Selected papers

Back to top

7. Methods and software

Multireference electronic structureCASSCF, NEVPT2, QD-NEVPT2, spin–orbit coupling, ab initio ligand-field theory
Single-reference methodsDFT, TD-DFT, DLPNO-CCSD(T)
Quantum nuclear dynamicsVibronic-coupling Hamiltonians, MCTDH, ML-MCTDH
Biomolecular simulationQM/MM molecular dynamics, classical MD, free-energy methods
Quantum computingVariational algorithms on simulators and IBM quantum hardware
SoftwareORCA, Gaussian, CP2K, PySCF, Qiskit, QuTiP, ASE, TorchANI

Back to top