Nonequilibrium dynamics provides an unprecedented point of view onto the interacting many-body systems. I am fascinated by that approach because it builds the understanding through surprising connections; via thermalization we can learn a lot about the non-ergodic systems, emergent time-scales can actually give information on the interactions between different subsystems, while some extraordinary properties can be triggered or accessed only by driving. Nonequilibrium as a research topic seems inexhaustible since almost all applications of correlated matter use or rely on nonequilibrium situations.

Broad areas of interest:
non-equilibrium dynamics of many-body systems,
open systems,
tensor networks,
machine learning.

Concrete past research topics:
Unsupervised learning of state complexity We used autoencoders to detect local complexity of time evolved states.
Reviving integrability in realistic materials. We showed that driving can reactivate conservation laws of integrable models in realistic setups (solid state materials, trapped ions) and can stabilize highly non-thermal steady states described by generalized Gibbs ensembles.
Critical properties of many-body localized (MBL) systems in open setups. While numerical studies of MBL in closed system are very limited, we looked at this problem through a fresh (and seemingly forbiden) perspective: by coupling a disordered system to baths.
Pump-probe dynamics of Mott insulators. We studied different stages of relaxation after a (photo) excitation of Mott insulators. Most importantly, we proposed mechanisms that explained recombination of holon-doublon excitons.
Dielectric breakdown of Mott insulator. We determined threshold field for excessive production of charged particles as a consequence of applying a constant electric field.

Past group members

---

Sourav Nandy --> Max Planck Institute for the Physics of Complex Systems, Dresden, Germany machine learning open systems Floquet dynamics

Madhumita Sarkar --> University of Exeter, UK Driven electronic systems