Lubricant characterization by molecular simulation

Abstract

Lubrication is a phenomenon of immense practical importance and fundamental scientific interest, and the automobile engines of the future are envisioned by the Partnership for a New Generation Vehicle will require the development of improved lubricants that perform well at higher operating temperatures and higher engine speeds. The rheological properties of liquid alkanes of intermediate molecular sizes (C20H42-C40H82) are among the most important properties in lubricant performance. Though realistic study of these systems by molecular simulation has previously been limited by both high computational costs and the lack of potential models accurate over a wide range of physical conditions, the advent of massively parallel supercomputers has now made such studies possible. As an illustration of the ability of molecular simulation as a useful tool for lubricant development, we present the first molecular-simulation-based calculation of the kinematic viscosity index of an alkane liquid, viz. 2,6,10,15, 19,23-hexamethytetracosane, commonly called squalane. In the mass range of interest, squalane is one of the few commercially available isoparaffins and has been the subject of previous studies by molecular simulation (Mondello and Grest 1995;Mundy et al., 1997).Though numerous properties of a lubricant are important to end-use applications, the viscosity is considered most significant. Mundy et al. (1996) investigated the variation of decane's viscosity with pressure, calculating its pressureviscosity coefficient by equilibrium molecular dynamics simulations. The kinematic viscosity index (VI) is another widelyused industrial characterization of automotive lubricants. It was proposed by Dean and Davis (1929) as an indication of an oil's viscosity-temperature characteristics in terms of its Saybolt viscosities at 311 K (100°F) and 372 K (210°F). Two series of reference lubricating-oil fractions ( H and L ) were used for comparison. Series H exhibited little change of viscosity with temperature while the viscosities of series L oils exhibited large variation with temperature. Series H and L represented, respectively, the best and worst oils available in 1929. Series H oils were assigned a VI of 100, series L a Correspondence concerning this article should bc addressed to P. T. Cummings. value of 0. The VI of an oil under test ( T ) was calculated from the equation L -Hwhere U is the kinematic viscosity at 311 K of the oil in question, L and H , respectively, are the kinematic viscosities at 311 K of the series L and H having the same kinematic viscosity at 372 K as the oil T. Thus, the higher the VI the less the viscosity of an oil is affected by temperature and, therefore, the better the oil. Subsequently, Hardiman and Nissan (1945) proposed revision to the VI system because it had reached the end of usefulness in its previous form. It had become necessary to give VI values to oils with better viscosity-temperature behavior than those represented by 100 on the VI scale. Hardiman and Nissan found that the VI system faile...

Molecular Simulation of Poly-α-olefin Synthetic Lubricants: Impact of Molecular Architecture on Performance Properties

“…32 As far as viscositytemperature characteristics are concerned, we are aware of only one previous molecular simulation study in which the VI of the squalane molecule was calculated. 29 We should also mention the interesting work done by Mondello, Grest, Silbernagel, and Garcia, 20 in which they discuss the effect of branching density on the activation energy for diffusion. It is difficult to draw many firm conclusions from these studies, given that each had different objectives and different force fields were used to represent the fluid.…”

A molecular dynamics study of a short-chain polyethylene melt.

“…This effect is presented in Fig. 8 in terms of the normalized shear stress growth coefficient versus total strain for melts of C 100 , C 30 , and C 10 . The state points correspond to the experimental density (interpolated for both C 100 and C 30 ) at a temperature roughly between 30 and 60 K above each particular system's (experimental) melting point.…”

“…At the shear strain rates accessible to simulation, the liquid alkanes have shown certain rheological behaviors characteristic of polymer systems, including shear thinning (e.g. [9][10][11]) and non-zero normal stress differences (e.g. [12,13]).…”

New Experimental Data and Reference Models for the Viscosity and Density of Squalane

“…[43][44][45][46][47][48][49][50][51][52] In cases where an existing experimental data set was repeated (or it appeared that the same point was not re-measured) in subsequent articles, only the original data set was included in the regression data set: 64 , where there was a very limited data set available for the development of a reference model, the abundance of experimental data for squalane made it unnecessary to include molecular simulations in this work. 37,[65][66][67][68][69][70][71][72][73] Two points from Whitmore et al (1966) 8 were not included due to the inability of converting kinematic viscosities of sub-cooled squalane (219 and 233) K.…”