Atomistic Modeling and Analysis of Fluids in Bulk and at Interfaces for Lubricant Design

Ahmed, Jannat

doi:https://doi.org/10.21985/n2-trt0-g660

Work

Atomistic Modeling and Analysis of Fluids in Bulk and at Interfaces for Lubricant Design

Public

Download PDF

The functions of lubricants in moving interfaces are extensive; lubricants minimize friction, transfer heat, clean the rubbing surfaces, extract debris, and diminish wear. Rather than using separate lubricants for each of the functions, it is more desirable to make one lubricant serve multiple purposes, or a muti-functional lubricant. The goals of this thesis are to develop methods for computational characterization of the contributing aspects of fluids towards multi-functional lubricants, viz., 1) fluidity, 2) heat transfer, 3) lubricity in connection with tribochemistry, and to conceive formulas relating these properties with the atomistic features of the fluids to facilitate lubricant design. Computational methods using molecular dynamics were developed and simulations were carried out for numerical evaluations of lubricant bulk properties. Relationships between the molecular structure and bulk properties were identified, leading to the development of correlation equations. Tribochemistry at lubricated interfaces formed with metal-oxide surfaces was studied; processes of catalytic fragmentation and tribopolymerization of lubricants were identified, and the factors affecting those processes were captured.The bulk properties of lubricants analyzed include pour point and thermal conductivity. A molecular dynamics-based approach was developed to explore and identify the pour points of a number of traction fluids. Diffusion behaviors of several fluids were investigated by time-averaged mean squared displacement (MSD), and a clear change in the MSD-temperature trend was observed at a particular temperature for each fluid. The temperature at which the trend changes corresponded well with the experimentally observed pour point. Further, a correlation formula was obtained to link the pour point and the ratio of the solvent accessible surface area (SASA, defined as the locus surface of the center of a solvent probe when it rolls over the surface of the fluid molecule) to the total number of carbon atoms, as well as several other molecular structural features. The pour points obtained from the molecular dynamics (MD) simulation and formula prediction were found to be in agreement with existing experimental results of the fluids studied. Non-equilibrium molecular dynamics simulations were performed using two different force fields to study thermal conductivity of hydrocarbon base oils. Argon was analyzed as a reference for method evaluation, and the results revealed that the calculated conductivity strongly depends on the size of the computational domain. However, for hydrocarbon base oils, the dependence on the computation domain size was less prominent as the domain size was increased. The method of direct calculation in a sufficiently large computation domain, and that of reciprocal extrapolation with data calculated in much smaller domains were both applicable; both computation methods have distinct advantages in oil conductivity calculations. The calculated conductivities showed certain overpredictions when compared with experimentally measured results. The overprediction factor was related to the number of carbon atoms of the fluid molecules. The results also revealed that the thermal conductivity of a single-chain hydrocarbon fluid was linearly proportional to the number of carbon atoms. While each additional branch increased thermal conductivity slightly, the presence of multiple branches reduced it from the otherwise ideal linear relationship. A set of equations was formulated to correlate hydrocarbon fluids’ thermal conductivity with molecular characteristics in terms of the number of carbon atoms and that of branches. The in-situ formation of lubricious tribofilms was studied for a system involving chromium oxide surfaces and dodecane as the lubricant. Reactive molecular dynamics simulations were performed to study the fragmentation and tribopolymerization of dodecane. The effect of environmental oxygen on the rate of fragmentation was also studied. Thorough visualization of the simulation outputs revealed that chromium oxide acts as a catalyst to fragment dodecane molecules through the synergistic effects of Cr3+ (Lewis acid) sites and oxygen vacancies, thus escalating the tribopolymerization process. The investigation was further extended by simulating the same tribochemical system, but this time with the presence of CPCa precursor additive. The results showed that CPCa takes part in the tribopolymerization process and increases lubrication efficiency by attaching the tribopolymers to the surface. The developed methods and equations for the calculations and prediction of bulk properties can be used to predict lubricant properties prior to synthesis, making the design process more time-efficient and cost-effective. The atomistic analyses of the properties also provide a deeper understanding of the contribution of different features of molecular structures to bulk properties. The knowledge corresponding to the in-situ tribopolymerization of base oils may guide surface and lubricant designs for better friction and wear management. Furthermore, the thesis suggests multiple avenues of future research.

Creator