Dr. Vinay Arya
Research

Controlling transport under nanoscale confinement

When water, ions and polymers are squeezed into channels a few molecules wide, continuum theory breaks down: interfacial forces, entropy, electrostatics and molecular layering take over. I use molecular dynamics, physics-informed machine learning and experiments to understand these effects and use them in water and energy technologies. This work is done with Prof. Chirodeep Bakli in the Tensore Lab, IIT Kharagpur.

PhD thesisDesalinationPolymersOil–waterPhysics-informed AIPassive coolingBeyond the thesisMethodsCollaborations
PhD thesis · IIT Kharagpur · February 2026

Transport control under confinement: molecular mechanisms, soft matter effects, and multiscale integration

The thesis builds a physics-driven multiscale framework that connects molecular mechanisms to system-level performance. It establishes design principles for graphene desalination membranes, shows how grafted polymers tune slip and selectivity, uncovers passive oil–water fractionation pathways, and links atomistic descriptors to device-scale predictions with physics-informed machine learning. Finally, it applies this understanding to polyacrylamide–LiCl hydrogels for passive cooling of electronics and solar panels.

Supervisor: Prof. Chirodeep Bakli · School of Energy Science and Engineering · Supported by the Prime Minister's Research Fellowship
Water molecules flowing through a graphene nanochannel
Hydration shells of Na+ and Cl- ions in bulk water and at a graphene interface
Ions shed part of their hydration shell at the graphene interface.
Thesis chapter 3

Water and ion transport for desalination

Key resultChanging membrane wettability alone tunes how ions and water pass, and which ion is preferred (Cl⁻ over Na⁺), without new nanofabrication.

Pressure-driven flow through functionalised graphene nanopores. Water structuration acts as an invisible gate, hydrophilic pores pass more water, and ion selectivity comes from hydration asymmetry: dehydration energy barriers combined with van der Waals interactions decide which ions get through.

Molecular dynamics snapshot of polymer chains flowing in a graphene nanochannel
Polymer chains under nano-confinement between graphene-like walls.
Thesis chapter 4

Polymer-mediated nanofluidic transport

Key resultAn NPzT simulation route reaches densely grafted brushes longer than half the channel height, configurations conventional all-atom MD can't produce.

How wettability shapes polymer density layering, slip length and flow enhancement through competing enthalpy and entropy. Grafted polymer brushes reshape water at the molecular level: grafting density, chain length and channel size act as control knobs that tune flow resistance, slip and brush folding, captured in a unified descriptor.

Nanoscale journal cover featuring the oil spill fractionation work
Featured on the cover of Nanoscale, 2024.
Thesis chapter 5

Passive oil–water fractionation

Key resultDistinct separation zones that shift with temperature give a design map for separating oil spills across real environmental conditions.

Multiphase transport of oil–water mixtures in nanochannels. Composition-driven segregation and wall–fluid attraction determine which phase sticks to the surface and which slips, giving a design map for separating oil spills passively, without pumps or chemicals.

Design map showing where oil stays near or away from the graphene-like surface
Design map: oil near (yellow) or away from (purple) the surface, at 335 K and 370 K.
Physics-informed neural network architecture with PDE, boundary and data losses
PINN architecture combining data, boundary and governing-equation losses.
Thesis chapter 6

Physics-informed AI for multiscale integration

Key resultMatches molecular dynamics with R² > 0.97 for electroosmotic velocity while cutting computation from hours to minutes.

A physics-informed neural network that embeds atomistic descriptors into the governing transport equations, so a few hundred MD simulations are enough for fast, interpretable predictions of electroosmotic flow in graphene nanochannels and of system-level metrics such as specific water production.

Hydrogel synthesis, transparency and adhesion, and wind-tunnel thermoelectric set-up
PAM–LiCl hydrogel: chemistry, transparency, adhesion and the wind-tunnel thermoelectric test.
Thesis chapter 7

Hydrogels for thermal management & renewable energy

Key resultCools by evaporation during the day and recharges with atmospheric moisture at night: no power and no feed water.

Polyacrylamide–lithium chloride composite hydrogels that cool by evaporation during the day and recharge with atmospheric moisture at night. On solar panels they lower operating temperature and recover power output in rooftop tests; on high-heat-flux electronics, combined with thermoelectric modules and airflow, they cool and recover energy at the same time.

Beyond the thesis

Wetting, boiling & capillarity

Topography and wettability in boiling on graphene; capillary filling that defies Lucas–Washburn.

Electrolytes for energy storage

MD insight into electrolyte structure for high-voltage pseudocapacitors and Li–S batteries.

2D materials & composites

Twisted bilayer graphene, a graphene–MoSeS sandwich Ongoing, rubber–epoxy interfaces and 3D-printed anthill mechanics.

Methods & tools

Simulation
All-atom molecular dynamics (EMD / NEMD)CFD: ANSYS, COMSOLMultiscale transportEnergyPlusAnsys Workbench
Machine learning
Physics-informed neural networksScientific MLDeep learningSurrogate modelling & optimisation
Experiments
Hydrogel synthesisSolar simulator & I–V tracingInfrared thermographyWind-tunnel testingWetting characterisation
Computing
PythonMATLABData analytics

Collaborations

IIT Kharagpur, India

Prof. Suman Chakraborty

  • Nanofluidic flow enhancement (JCP 2023)
  • Capillary filling in CNTs (JCP 2024)
  • Boiling on graphene (Energy 2025)
  • Ion-specific interfaces (JCIS 2026)
Jain University, Bengaluru, India

Dr. Debasis Ghosh

  • Flexi-pseudocapacitor (Small 2024)
  • Li–S batteries (Chem. Eur. J. 2025)
  • MnO2 pseudocapacitor (JMCA 2026)
University of Ulsan, South Korea

Prof. BoHung Kim

  • Boiling over functionalized graphene (Energy 2025)
INST Mohali, India

Dr. Chandan Bera

  • Twisted bilayer graphene (JAP 2023)
IIT Kharagpur, India

Prof. Chandra Sekhar Tiwary

  • 3D-printed anthill mechanics (2022)

Interested in working together?I welcome collaborations that connect simulations and machine learning with experiments.

Get in touch