- Manufacturing of microsystems such as lab-on-chip analytical sensors, flow control, and micropropulsion devices will take a giant leap forward if cost-effective predictive modeling is available in areas where experimental diagnostic is limited or impossible. Gas flows in microsystems are characterized by low Reynolds and large Knudsen numbers, and large surface-to-volume ratios. This leads to flow phenomena in microsystems that are drastically different from their macroscale counterparts. For example, thermal stresses generated by temperature gradients in devices manufactured from now-available low thermal conductivity materials, such as aerogels, can be exploited to built gas pumps and actuators. The work in this area is aimed at the development of methods and tools for deterministic solution of the Boltzmann equation, which governs microscale gas flows; coupling of fluid flow and thermal/structural material response and quantification of uncertainties of the simulations.
- Development of Engineering Molecular Mechanics — The two most widely used numerical methods suitable for molecular simulations are Molecular Statics (MS) and Molecular Dynamics (MD). While MD simulations can be used to simulate systems at any temperature, they involve enormous computational resources and, hence, pose a crippling limitation on the size and time scales of the systems that can be modeled. On the other hand, MS simulations involve significantly lesser computational resources but are valid only at 0° K (absolute zero). We are developing a greatly simplified equivalent static molecular simulation technique that offers the same results as MD simulations but with a fraction of the computing time. The essence of this approach lies in the selective use of some bulk properties such as coefficients of thermal expansion which are readily available to achieve the simplification, We call this approach "Engineering Molecular Mechanics" (EMM) because it utilizes bulk engineering properties and it can be easily applied to large scale engineering problems which need nanoscale designs at finite temperatures. We have done some preliminary studies to show the effectiveness and efficiency of the EMM technique. Preliminary results indicate savings of at least three orders of magnitudes in computation time using EMM when compared to MD simulations of the same system. We plan to expand the EMM capability to simulate systems involving multi-physics phenomenon like corrosion where chemical reactions play a significant role.
- Nonconventional Multiscale Continuum Models — Although molecular mechanics is capable of providing detailed information of materials at the atomistic or molecular level, the size of the simulation model is limited by the present time computing capability. We develop and use nonconventional continuum models to perform the conversion of the result obtained from molecular models into appropriate bulk properties for practical engineering analysis and design of large scale systems The issue of "hand shaking" between the EMM model and the nonconventional continuum model with micro/nano-structure has been addressed. By "hand shaking" we do not mean modeling a material with both EMM and continuum models simultaneously as practiced by many researchers in the computational mechanics community. The continuum models we have developed contain relevant micro/nano-structures that can accurately capture the selected behavior predicted by EMM and are truly multiscale continuum models.
- Manufacturing Nanolaminates by a Self-Assembly Method — One of the most promising ways of utilizing the unique material properties found at the nanoscale is to combine the nano particles with a matrix material such as a polymer to form a composite material (nanocomposite). Unfortunately, current technology for making nanocomposites is limited to mixing or standard self assembly methods, both of which have severe limitations. We have investigated a self assembly method to make nanolaminates consisting of hundreds or thousands of nanometer thick laminae. This is done using the Spin Assisted Layer by Layer Self Assembly (SALbL) by which layers of different nanomaterials are formed according to the design. While the nanolamina will be only nanometer thick, their lateral dimensions can be made rather large.