Research

Coupled physics, computed and measured

We combine physics-based simulation, reduced-order and data-driven modeling, and laboratory measurements, from the sheath of a plasma to the flow around an aircraft.

Low-temperature plasmas for materials processing

CCPICPharmonic balanceSOMAFOAM

Capacitively and inductively coupled plasmas (CCP, ICP) are the workhorses of semiconductor etching and deposition. We build and validate fluid models of these discharges in the open-source finite-volume solver SOMAFOAM, and develop the harmonic-balance method, which computes the periodic steady state of an RF plasma directly instead of following thousands of RF periods. Collaborations with Lam Research and Applied Materials connect the methods to process-scale problems.

Periodic steady state of an argon CCP over position and RF phase (electron density, temperature, potential), computed with the harmonic-balance method.
Periodic steady state of an argon CCP over position and RF phase (electron density, temperature, potential), computed with the harmonic-balance method.

Reduced-order and data-driven modeling

reduced-order modelsDMDmachine learninguncertainty

Design and control need models that run in milliseconds. We build reduced-order models that learn the full spatiotemporal state of complex systems from simulation data, using proper orthogonal decomposition, autoencoders, dynamic mode decomposition and Gaussian-process regression, and we quantify how much data a given accuracy requires. Our CCP model predicts the periodic state of a discharge in a fraction of a second and can be tried online.

Periodic state of a CCP from the harmonic-balance solver and predicted by the reduced-order model.
Periodic state of a CCP from the harmonic-balance solver and predicted by the reduced-order model.

Aerodynamics, rarefied and high-speed flows

unsteady aerodynamicsflight dynamicsDSMCshock waves

We study unsteady aerodynamics at low Reynolds number, such as the force regimes of plunging wings, where leading-edge vortices can more than double the mean lift and the fluctuating forces split into added-mass and circulatory parts, and flight dynamics problems such as formation flight in aerial refueling. Our work on rarefied and non-equilibrium gas flows uses direct simulation Monte Carlo and Boltzmann model equations for shock-tube and supersonic jet flows, gas-surface interactions and intermolecular potentials.

Vorticity around an ellipse plunging at Reynolds number 100, at four phases of the plunge period.
Vorticity around an ellipse plunging at Reynolds number 100, at four phases of the plunge period.

Cardiovascular and biological flows

biofluidspulsatile flowmachine learning

We model pulsatile blood flow, including the shear-thinning rheology of blood, and ask how much a few measurements can reveal about the flow. Current work combines verified simulations of pulsatile flow through stenosed arteries with machine learning to infer the geometry of a stenosis from sparse wall shear stress and pressure signals, with quantified uncertainty.

Verification and validation of the pulsatile blood-flow solver: shear-thinning pipe flow, Womersley flow, and flow through a 75% stenosis compared with direct numerical simulation.
Verification and validation of the pulsatile blood-flow solver: shear-thinning pipe flow, Womersley flow, and flow through a 75% stenosis compared with direct numerical simulation.

Microplasmas, gas breakdown and plasma applications

microplasmasgas breakdownexperimentsagriculture

Measurements and models of microscale gas discharges cover field-emission-assisted breakdown, modified Paschen curves and microwave microplasmas for tunable electronic devices. In the laboratory, atmospheric-pressure plasmas are applied to agriculture and water: irrigation water enrichment, plant growth, bipolar ionization in indoor farms and plasma-surface modification of biopolymers.

Field-emission-assisted microwave microplasma (Alamatsaz and Venkattraman, J. Appl. Phys. 2016).
Field-emission-assisted microwave microplasma (Alamatsaz and Venkattraman, J. Appl. Phys. 2016).