Multiscale models for structure and dynamics of confined fluids

Sanghi, Tarun

Permalink

https://hdl.handle.net/2142/92711 Copy

Description

Title

Multiscale models for structure and dynamics of confined fluids

Author(s)

Sanghi, Tarun

Issue Date

2016-06-08

Director of Research (if dissertation) or Advisor (if thesis)

Aluru, Narayana R.

Doctoral Committee Chair(s)

Aluru, Narayana R.

Committee Member(s)

• Schulten, Klaus J.
• Schweizer, Kenneth S.
• Ferguson, Andrew L.

Department of Study

Mechanical Sci & Engineering

Discipline

Mechanical Engineering

Degree Granting Institution

University of Illinois at Urbana-Champaign

Degree Name

Ph.D.

Degree Level

Dissertation

Keyword(s)

• Multiscale modes
• Structure and Dynamics of Confined fluids
• Generalized Langevin Equation
• Thermal noise
• Empirical potential based Quasi-continuum Theory (EQT)

Abstract

In this dissertation, using systematic coarse-graining, we develop multiscale models to study structural and dynamical properties of confined fluids. With the advent of nanofluidics and nanobiotechnology, fluids confined inside nanometer scale geometries have become a subject of both fundamental investigation and applied research. An understanding of the structural and dynamical properties of fluids at nanoscale is essential for designing novel engineering applications such as nanofiltration, carbon-dioxide sequestration, single-file transport, nanomedicine and many others. Our structural model is based on an empirical potential based quasi-continuum theory (EQT). EQT is a multiscale theory that seamlessly integrates the interatomic potentials describing various atomic interactions into a continuum framework to obtain the equilibrium density and potential profiles of confined fluids in a self-consistent manner. The density and potential profiles obtained from it are comparable in accuracy with those obtained from particle-based methods such as molecular dynamics (MD) simulations. Also, being a continuum approach, EQT is very simple to implement and is computationally several orders of magnitude faster than MD simulations. The central task in EQT is the development of quasi-continuum potential models that accurately describe the wall-fluid and fluid-fluid interactions in confined fluids. Using systematic coarse-graining, we discuss the development of coarse-grained single-site (CGSS) pair-potentials and quasi-continuum potential models for poly-atomic fluids. Proposed potential models systematically incorporate the effect of size, geometric shape and orientation of poly-atomic fluids to predict the correct microstructure in confined environments. We take carbon-dioxide as an example fluid and demonstrate the applicability of the potentials models in EQT as well as coarse-grained MD (CG-MD) simulations to predict the center-of-mass (COM) density and potential profiles of carbon-dioxide inside slit-shape graphite nanochannels at several high and low pressure confinements. The results obtained from EQT and CG-MD simulations are found in good agreement with those obtained from all-atom MD (AA-MD) simulations. To develop dynamical models, one fundamental question is to understand the role of thermal noise in nanofluidic dynamics and transport. We discuss a combined memory function equation (MFE) and generalized Langevin equation (GLE) based approach (referred to as MFE/GLE formulation) to characterize thermal noise in molecular fluids. Using MFE/GLE formulation in conjunction with MD simulation, we extract and analyze the statistical properties of thermal noise in confined fluids. We find that the thermal noise correlation time of the confined fluid does not vary significantly across the confinement and is quite similar to that of the corresponding reference bulk state fluid. We show that it is the cross-correlation of the mean force with the molecular velocity that gives rise to the spatial anisotropy in the velocity-autocorrelation function of the confined fluids. Further, we demonstrate that using the noise characteristics of reference bulk state fluid, and the structural information obtained from EQT, GLE can be used to simulate the single- particle dynamical properties of confined fluids. As an application, we use the GLE formulation to compute the interfacial friction coefficient at solid-liquid interface. Interfacial friction coefficient is an important macroscopic modeling parameter that provides the atomistic to continuum bridge by incorporating the effect of the wall-lattice structure and the nature of wall-fluid interactions on the fluid transport. The attractive feature of the GLE approach is that all the inputs to the GLE are obtained from EQT and simulation data of the reference bulk state fluid, thereby eliminating the need to perform computationally expensive equilibrium molecular dynamics (EMD) simulation of the confined system to estimate the interfacial friction. We also use the GLE formulation to understand the memory effects in the dynamics and transport of nanoparticles such as fullerenes immersed in host fluid environment. Finally, we discuss a GLE based approach to simulate the dynamics of interacting-particles.

Graduation Semester

2016-08

Type of Resource

text

Permalink

http://hdl.handle.net/2142/92711

Copyright and License Information

Copyright 2016 Tarun Sanghi

Owning Collections