NAVIS: A LAMMPS-Python framework for efficient computation of nanochannel velocity and thermal interfacial slip

TL;DR

The paper addresses the challenge of quantifying intrinsic hydrodynamic and thermal resistance at solid–fluid interfaces in nanoscale channels. It introduces NAVIS, a Python-LAMMPS toolkit that leverages equilibrium molecular dynamics to compute the Navier friction coefficient and Kapitza resistance from interfacial fluctuations, using robust time-correlation analyses and Laplace-domain transforms. The method is demonstrated on planar water–graphene (hydrodynamic slip) and cylindrical water–CNT (thermal slip) systems, with multiple post-processing approaches showing that Method-3 provides the most stable friction estimates and Kapitza resistance aligns with NEMD results while avoiding divergences associated with Green-Kubo in confined geometries. Overall, NAVIS offers a flexible, accurate, and efficient framework for extracting interfacial slip properties directly from equilibrium data, enabling better design and understanding of nanofluidic devices.

Abstract

We present NAVIS (NAnochannel Velocity and thermal Interfacial Slip), a LAMMPS-Python scripted toolkit for computing the Navier (hydrodynamic) friction coefficient and Kapitza (thermal) resistance at arbitrary solid-fluid interfaces. NAVIS is based on equilibrium molecular dynamics (EMD) methods for calculating the linear response friction and thermal resistance at the interface, as well as the corresponding velocity and temperature slips. The methodology is based on our previous studies (Hansen, et al., Phys. Rev. E 84, 016313 (2011); Varghese et al., J. Chem. Phys. 154, 184707 (2021); Alosious, et al., J. Chem. Phys. 151, 194502 (2019); Alosious, et al., Langmuir 37, 2355-2361 (2021)), and in this work we provide a pedagogical framework for the implementation of this toolkit on two systems: (i) a water-graphene system (for hydrodynamic slip) and (ii) a water-CNT system (for thermal slip). We provide detailed instructions for performing the EMD simulations using the LAMMPS package and processing the simulation outputs using Python modules to obtain the desired quantities of interest. We expect the toolkit to be useful for computational researchers studying interfacial friction and thermal transport, key factors for efficient and practical applications of nanofluidic systems.

Figures (6)

• Figure 1: A 2-D schematic diagram of the fluid molecules (depicted as solid circles) confined by planar walls. The actual system is three-dimensional and periodic in the $x$ and $y$ directions. $\Delta$ is the slab width.
• Figure 2: Flowchart outlining the steps involved in the computation of the Navier friction coefficient or Kapitza resistance.
• Figure 3: For a water-graphene system: (a) Normalized time correlation functions. Red filled squares represent $C_{u_{x}u_{x}} (t)$, and red filled circles represent $C_{u_{x}F_{x}} (t)$. The inset plot represents the corresponding normalized Laplace transform. (b) Variation of friction coefficient with correlation time intervals. The correlation lag times studied are $t$ = 5, 10, 20, 25, 40, 50, 80, and 100 ps. The error bars correspond to one standard error obtained from 50 independent simulations. (c) Histogram distribution of friction coefficient values and $\sigma$ represents the standard deviation of the distribution. (d) Comparison of $\xi_0$ calculated using different methods. Red filled circles represent $\xi_0$ calculated using Method-3 at different correlation lag intervals. The dashed line shows the $\xi_0$ calculated using the methodology proposed in Ref bocquet1994 at the correlation lag time t = 5 ps, and the dotted lines correspond to the upper and lower limits of the friction coefficient. The green shaded region represents the NEMD friction coefficient range at the external field $\sim 1.25 \times 10^{11}\ \textrm{m}\textrm{s}^{-2}$. Reprinted with permission from Varghese et al.varghese2021. Copyright (2021) American Institute of Physics.
• Figure 4: (a) Comparison of the Kapitza resistance obtained from the NEMD and the present EMD methods. (b) Comparison of the Kapitza resistance as a function of correlation time calculated using the present method and the Green-Kubo-like method of Barrat and Chiaruttini puech1986barrat2003kapitza. Reprinted with permission from Alosious, et al.alosious2021. Copyright (2021) American Chemical Society.
• Figure 5: (a) Schematic representation of the planar confined system. (b) Top view of the graphene channel. Reprinted with permission from Varghese et al.varghese2021. Copyright (2021) American Institute of Physics.