NIST logo
*
Bookmark and Share

Computational Methods for Materials Science

Summary:

The design of innovative materials is an essential task to advance fields such as medicine, optics, and mechatronics.  

Nano-materials are extremely valuable to enhance fundamental material properties such as thermal and electrical conductivity, viscosity, and various mechanical properties.

In order to understand the behavior of nano-materials in solutions and other complex environments, existing experimental techniques such as normal mode resonance (NMR) and paramagnetic relaxation enhancement (PRE) need to be complemented by computational models and measurements.

Description:

To better understand and quantify nano-material dynamics, binding mechanics/affinity, and aggregation pathways in solutions at the atomistic level (e.g., carbon nanotubes-surfactant), the following research topics are being addressed.

  • We are designing Monte Carlo techniques for the detection of first-encounter binding modes based on electrostatic and hydrophobic forces and on molecular surface complementarity. We are using configuration bias Monte Carlo techniques for ab initio molecular binding prediction. This approach also enables the detection of "weak" binding modes, which may be essential in understanding certain molecular mechanisms but are hard to detect with traditional experimental techniques.

First-encounter binding modes Electrostatic (red-blue) and hydrophobic (white) patches on molecular surface Electrostatic (red-blue) and hydrophobic (white) patches on molecular surface

Computationally obtained binding modesExperimentally obtained binding modesProbability distribution of computationally obtained binding modes

  • We are designing molecular dynamics protocols and trajectory analysis tools based on explicit solvent representation, which enables a more accurate investigation of the induced fit that follows first-encounter binding. 

Simulation in pure water box (Ph~7)PCA of molecular trajectory: ribbons show collective motions

  • We are designing adaptive, multigrid-based techniques for the efficient numerical solution of PDE systems and non-linear equations that enable the use of more accurate implicit solvent models, especially in relation to interfacial effects. 

  • We are implementing parallelization schemes by multi-threading on CPU and by porting to GPU for enhanced efficiency of the above techniques.  

  • ZENO is a Monte-Carlo simulation code that uses Random-Walks to compute the intrinsic viscosity of molecules and nano-materials. We are developing enhanced spatial data structures and algorithms that will accelerate the simulation by several orders of magnitude and, as such, can potentially transform the way material researchers use such simulations.

ZENO

We will parallelize the enhanced data structures and algorithms to take advantage of multicore architectures to yield an additional order of magnitude acceleration.

Potential applications: The above computational techniques and resulting software can be used to elucidate complex molecular mechanisms related to nano-materials.  Potential applications include the following.

  • Novel CNT-based material design procedures through a better understanding of CNT-surfactant interaction and of chirality-based CNT separation dynamics in solution.

Novel CNT-based material design

  • The design of polymeric materials with enhanced properties such as shear viscosity and electric conductivity through a better understanding of CNT aggregation mechanisms.
  • Improved drug delivery vectors through a better understanding of nanomaterial-based matter manipulation mechanisms (e.g., molecular transport properties, intrinsic molecule viscosity)

Lead Organizational Unit:

itl

Staff:

Walid Keyrouz, Information Systems Group