International Colloquium Tribology
ict
expert verlag Tübingen
125
2022
231
Molecular Dynamics Study of the Adsorption of Organic Friction Modifiers on Iron Oxide Surfaces
125
2022
Pablo Navarro Acero
Stephan Mohr
Marco Bernabei
Carlos Férnandez
Beatriz Dominguez
James Ewen
ict2310441
23rd International Colloquium Tribology - January 2022 441 Molecular Dynamics Study of the Adsorption of Organic Friction Modifiers on Iron Oxide Surfaces Pablo Navarro Acero Nextmol (Bytelab Solutions SL), Carrer de Roc Boronat 117, 08018 Barcelona, Spain Barcelona Supercomputing Center (BSC-CNS), Carrer Jordi Girona 29, 08034 Barcelona, Spain Stephan Mohr Nextmol (Bytelab Solutions SL), Carrer de Roc Boronat 117, 08018 Barcelona, Spain Barcelona Supercomputing Center (BSC-CNS), Carrer Jordi Girona 29, 08034 Barcelona, Spain Corresponding author: stephan.mohr@nextmol.com Marco Bernabei Repsol Technology Lab, DC Technology & Corporate Venturing, Agustín de Betancourt s/ n, 28935 Mostoles, Madrid, Spain Carlos Fernández Repsol Technology Lab, DC Technology & Corporate Venturing, Agustín de Betancourt s/ n, 28935 Mostoles, Madrid, Spain Beatriz Dominguez Repsol Technology Lab, DC Technology & Corporate Venturing, Agustín de Betancourt s/ n, 28935 Mostoles, Madrid, Spain James Ewen Department of Mechanical Engineering, Imperial College London, South Kensington Campus, London, SW7 2AZ, U.K 1. Introduction The adsorption and self-assembly of surfactants at solid-liquid interfaces plays a key role in a wide range of technological processes. In boundary lubrication conditions, lubricant additives that reduce friction are vital to increase the energy efficiency of moving engine components. However, due to concerns with respect to detrimental side effects caused by certain chemical elements found in conventional friction modifier additives, there is an increasing interest in so-called organic friction modifiers (OFM) that only contain C, H, O and N atoms. The performance of OFMs, as well as other additives, depends critically on their ability to adsorb onto the metal surface of the engine components and to form self-assembled layers that then result in the desired friction reduction. Therefore, understanding the relation between the initial concentration of the additive in the lubricant and the resulting surface coverage is extremely important for lubricant formulations and other surfactant applications in general. 2. Computational Modelling Approach In recent years, molecular simulations have provided additional insights into the nanoscale behavior of a wide range of lubricant additives, including OFMs. These insights at the atomic scale are essential to develop new additives with improved performance. Thanks to considerable advances in both theory and algorithmic development seen in recent years, combined with the steadily increasing power of modern supercomputers, it is now possible to simulate complex systems and phenomena with high accuracy and within reasonable time. In the current study, we used Molecular Dynamics (MD) to study the adsorption isotherm of three different OFMs (stearic acid, glycerol monooleate and glycerol monoostearate) onto a hematite surface with hydrogenated 1-decene trimer (main constituent of PAO 4 base oil) acting as bulk solvent. 2.1 Adsorption PMF First, we calculated the potential of mean force (PMF) of the adsorption process using the adaptive biasing force (ABF) importance sampling method [1]. We found that the adsorption strength of GMO is 442 23rd International Colloquium Tribology - January 2022 Molecular Dynamics Study of the Adsorption of Organic Friction Modifiers on Iron Oxide Surfaces considerably higher compared to GMS, which in turn is comparable to SA. Moreover, the PMF profiles are interpreted in terms of headgroup-surface and tail-solvent interactions, in order to explain the differences in the adsorption energies (Figure 1 shows as illustrative example the case of GMS). Additionally, we show that the PMF calculation using ABF is more efficient than the analogous calculation with Umbrella Sampling. Figure 1: Adsorption PMF for GMS. 2.2 Hard-disk area Second, we estimated the area occupied by each molecule on the surface in the high coverage limit. To this end, we carried out simulated annealing simulations for films of various densities of OFMs pre-adsorbed onto the surface, with the objective of determining the maximum coverage and thus the area occupied by each molecule. The results are shown in Figure 2. We obtained a similar (hard-disk) surface area for GMS and GMO (0.23 nm 2 ), whereas the result is smaller for SA (0.16 nm 2 ). 2.3 Adsorption isotherms from MTT Finally, we combined the adsorption energy and the surface area of each OFM to determine the respective adsorption isotherms using the Molecular Thermodynamic Theory (MTT) model developed by Blankschtein et al. [2]. The results are shown in Figure 3. Out of the three OFMs that have been studied, SA yields the highest maximum coverage (3.6 nm -2 ) due to its smaller headgroup size compared to GMO and GMS. Since both GMO and GMS have the same headgroup, they yield a very similar maximum coverage at high concentrations, where the surface coverage is mostly determined by the packing efficiency of the molecules. GMO yields the highest coverage of the three OFMs at low concentration, which is due to its larger adsorption energy. GMS has a coverage between GMO and SA at low concentration because of its intermediate adsorption energy. Overall, the results are in good agreement with reported experiments, even though there are some slight deviations, in particular concerning the difference between the experimental results for saturated and unsaturated surfactants and our computational results for GMO and GMS. We propose that these deviations are mainly due to differences in the kinetic barriers to the formation of high coverage monolayers. These kinetic barriers can be understood in terms of the steric hindrance for molecules adsorbing into partially formed monolayers, whereas lateral interactions between the tail groups and differences in aggregation (inverse micelle formation) play probably a secondary role. We discuss possible ways of taking these factors into account in future studies. Figure 2: Inital and final surface coverage of the OFMs following the annealing simulations. Figure 3: Adsorption isotherms for the studied OFMs calculated with the MTT model. 23rd International Colloquium Tribology - January 2022 443 Molecular Dynamics Study of the Adsorption of Organic Friction Modifiers on Iron Oxide Surfaces 3. Experimental validation We validated the calculated adsorption energies through high frequency reciprocating rig (HFRR) friction measurements. First, we measured the friction coefficient in boundary lubrication conditions as a function of the molar fraction of OFM added to the base oil. Subsequently, we applied the Jahanmir and Beltzer model [3] to relate the friction coefficient in boundary lubrication conditions with surface coverage. In this way, we obtain the surface coverage as a function of the OFM concentration. Finally, fitting these results with the Temkin adsorption isotherm yields an experimental estimate of the adsorption energy. The obtained values are in very good agreement with the predictions made by the simulations for SA and GMS, whereas for GMO they are slightly worse (Table 1). We attribute this discrepancy once again to the steric hindrance for molecules adsorbing into partially formed monolayers, which is larger for GMO due to the kink in the unsaturated tail group. SA GMS GMO Simulation -27.8 -34.2 -41.8 Experiment -26.4 *) -33.2 -24.9 Table 1: Adsorption energies (in kJ/ mol) calculated from simulation and experiment. *) taken from [4]. 4. Conclusion We have used both computational and experimental methods to assess the adsorption of three OFM molecules onto an iron oxide surface. The overall good agreement between simulation and experiments demonstrates that MD is an accurate and efficient tool to study the adsorption of OFMs at interfaces. Moreover, the possibility to evaluate OFM molecules in silico allows to develop automated high-throughput workflows to perform a virtual screening of many molecules, out of which only the best candidates will eventually be synthetized and tested experimentally. This approach can accelerate the development of new additives and chemicals, also beyond the area of tribology. References [1] Darve, E.; Rodríguez-Gómez, D.; Pohorille, A. Adaptive biasing force method for scalar and vector free energy calculations. Journal of Chemical Physics 2008, 128. [2] Nikas, Y. J.; Puvvada, S.; Blankschtein, D. Surface Tensions of Aqueous Nonionic Surfactant Mixtures. Langmuir 1992, 8, 2680. [3] Jahanmir, S.; Beltzer, M. An adsorption model for friction in boundary lubrication. ASLE Transactions 1986, 29, 423. [4] Jaishankar, A.; Jusufi, A.; Vreeland, J. L.; Deighton, S.; Pellettiere, J.; Schilowitz, A. M. Adsorption of Stearic Acid at the Iron Oxide/ Oil Interface: Theory, Experiments, and Modeling. Langmuir 2019, 35, 2033.
