|Research:||Kinetic Monte Carlo Simulations|
|Publications:||Publications by Mao in our database|
Dr. Zugang Mao
Materials Science and Engineering
2220 North Campus Drive
Evanston, IL 60208
Email: z-mao2 at northwestern.edu
Current Research Project
Performs computer simulations of the temporal evolution of the microstructures, on a subnanometer scale, of technologically significant metallic alloys utilizing the kinetic Monte Carlo algorithm. And determination of all the necessary energetic parameters necessary for describing those alloys and installing them in the kinetic Monte Carlo code.
The main research of this work involves studying the microstructure evolutions and grain boundaries of nickel based superalloy systems, especially the early stage ordering, phase separations and precipitations. Kinetic Monte Carlo (KMC) simulations are performed in parallel with the 3D atom probe results in our group. This method is using residence time algorithm in a rigid crystal lattice. There are several million different types atoms simulated in one three-dimensional system. The kinetic mechanism is through vacancy diffusion model. The diffusion process is via vacancy jumps toward nearest neighboring atoms and the jumps are thermally activated. At each KMC step, one transition is performed. It is relatively efficient in comparison with Metropolis Monte Carlo (MMC) simulations since MMC has high rejection rate at low temperatures. Also KMC time is defined in a physically meaningful way through each jumping step.
Grand Canonical Monte Carlo (GCMC) simulations and first principle calculations are performed to obtain the necessary thermodynamic parameters such as exchange energies and saddle point energies. The exchange energies are optimized to fitting to phase diagram of ternary system at certain temperature. Both pair potentials and BSF codes (based on Bozzolo-Smith-Ferrante method) are used for interaction energies in KMC simulations. First principle calculations will also involve in this issue.
Previous Research Project
Computational modeling of materials and physical chemistry, specifically molecular dynamics simulations to study the behavior of molecular fluids in carbon nanotubes and investigate the electronic structure of material interfaces and relate to the effect of the interface on mechanical properties and thermal and electrical transport across the interface. For fluids in carbon nanotubes, our computational methods are molecular dynamics simulations. The basic thought of this method is to solve Newton's classic equations of motion for a system via an interaction potential. For interface project, ab-initio total energy calculations are performed to study the properties of relaxed metal oxide low angle grain boundary structures and metal-Gas and metal-metal oxide heterogeneous interfaces. The calculations use density functional and pseudopotential theories in conjunction with plan wave expansions within the framework of a supercell approach, and are performed on high-end workstations.
1. Investigated the properties of dynamic molecular flow in nanotubes Three different types of molecules, CH4, C2H4, and C2H6 through single-walled carbon nanotubules with diameters in the range of 7.1 Angstroms to 36 Angstroms were considered. Form simulation data to data, molecular dynamic flow slowed in nanotubules with diameters below 36 Angstroms in agreement with experimental observations. In addition, high initial flow rates are predicted to lead to chemical reactions among some reactive molecules and between some molecules and the tubule walls. In simulation studies of dynamic flow, several factors were predicted to affect the flow velocity over time. The first is the size (diameter) of the tubules. As the diameter of carbon nanotubules decreased, the dynamic flow of the fluid molecules slowed more quickly. The density of the fluids was also shown to have an affect. At higher fluid densities, the velocity of the molecules slowed more quickly due to increased interactions between the fluid molecules. Next, the type of fluid molecule in the flow was considered. Fluids of ethane and ethylene slowed more quickly than methane since they have stronger interactions with the tubule walls. Finally, the rigidity of the tubule walls affected the dynamic molecular flow. Non-rigid tubules slowed fluids down more quickly than rigid tubules because dynamic tubules disturb the motion of the nearest flowing. This causes more collisions between the molecules and the tubule walls. The motion of the fluid molecules also tends to make the tubules expand and straighten.
2. Examined the diffusion behavior of single-type molecular fluids in carbon nanotubes at room temperature. Four-type organic molecules were used to in simulations: methane, ethane, ethylene, and butane. From the simulation data to date, in (10,0) tubes the diffusion behavior is from normal mode to single-file mode with the increase of the molecular size. The effect of different termination tubes was investigated in calculations. Hydrogen-termination tubes will make small size molecules more difficult to diffuse into the tubes but little effect on larger size molecules. The pore-pore effects are also tested in our simulations.
3. Modeled the separation mechanism of binary fluids. Three binary systems were investigated: methane/ethane, methane/n-butane, and methane/t-butane. Through simulations, we first try to understanding the diffusion mode of different type molecules in carbon nanotubes, model the separation mechanism of binary system in carbon nanotubes based on molecular diffusion, and determine the affecting components on molecular diffusion for binary fluids. From simulation data to data, we found that in small diameter tubes, small molecules diffuse and large molecules diffuse much slower or not diffuse in carbon nanotubes. When diameter of tubes becomes large, large molecules begin to diffuse. The fluid molecules have more difficulties to diffuse in hydrogen-ended tubes than carbon-ended tubes but the separation trend not change. In nanotube bundles, the diffusion behavior of binary molecules is a little different from single tube. The diffusion coefficient is 2/3 of single tube.
4. Investigated the electronic structure of material interfaces and relate to the effect of the interface on mechanical properties and thermal and electrical transport across the interface. Ab-initio total energy calculations are performed to study the properties of relaxed metal oxide low angle grain boundary structures and metal-gas and metal-metal oxide heterogeneous interfaces. The calculations use density functional and pseudopotential theories in conjunction with plan wave expansions within the framework of a supercell approach, and are performed on high-end workstations. We are studying the properties of the grain boundary structures in ZrO2 (cubic) and in TiO2 (rutile) and comparing with experimental results obtained using a scanning transmission electron microscope (STEM) at ORNL and a new transmission electron microscope (TEM).
5. Theoretical predicted of a spiral diffusion path for non-Spherical organic molecules in carbon nanotubes. The diffusive behavior of ethane and ethylene in single-walled carbon nanotubes was investigated using classical molecular dynamics simulations and density functional theory calculations. At low molecular densities, these non-spherical molecules are predicted to follow a spiral path inside nanotubes with diameters of 13-22 Å. Following this spiral path maximizes the interaction of molecular C-C bonds with the C-C bonds in the nanotubes. Spherical molecules, such as methane, are not predicted to follow a spiral diffusion path. This result quantifies the manner in which molecular shape and chemical bonding affects molecule-nanotube interactions and indicates the generality of spherical transport through nanotubes.
6. Write computer programs to test different algorithm for the molecular dynamics computational simulation approach.