ReaxFF Reactive Molecular Dynamics Simulations of Mechano-Chemical Decomposition of Perfluoropolyether Lubricants in Heat-Assisted Magnetic Recording

Abstract

A thorough understanding of the decomposition of perfluoropolyether (PFPE) lubricants is crucial to achieve heat-assisted magnetic recording (HAMR). In contrast to previous studies, which focused on thermal and catalytic decompositions, we gain insights into the mechano-chemical decomposition of PFPE films confined between the head and disk by performing reactive molecular dynamics simulations with our new ReaxFF force field. By quantifying the decomposition time constants under the operation conditions of HAMR, we infer that, within a heating time of ∼1 ns, pure thermal decomposition hardly occurs, whereas mechano-chemical decomposition is highly likely to occur. The decomposition rate constant of the PFPE films subjected to confined shear increases with normal pressure. The increase is well-fitted by a linear stress-activated Arrhenius curve at high normal pressures, whereas this is not the case at low normal pressures. We caution against extrapolating the linear stress-activated Arrhenius curve, which could cause significant overestimation of decomposition rate constants at low normal pressures. We find that the mechano-chemical decomposition of PFPEs is mainly attributed to the dissociation of C–OH and ether C–O bonds in the polar end groups, and in the main chain, the C–O bond is more likely to dissociate than the C–C bond.

Mechanochemistry of Phosphate Esters Confined between Sliding Iron Surfaces

“…Thus, the activation volume represents a ~7 % deformation of both molecules. This percentage deformation is comparable to that calculated from previous NEMD simulation results for α-pinene (~3 %) 33 and allyl alcohol (~7 %) 32 oligomerisation between sliding silica surfaces as well as the decomposition of PFPEs (~2 %) 36 between DLC surfaces.…”

“…55 Previous NEMD simulations with ReaxFF have shown that ΔV* is pressure-dependent for the decomposition of PFPEs between DLC surfaces. 36 One possible explanation for the increase in ΔV* with increasing temperature observed in the current simulations is that the elastic modulus of the α-Fe(110) surfaces is expected to decrease by around 10 % as the temperature is increased from 300 K to 500 K. 56 This is important since a recent DFT study has shown that ΔV* is inversely proportional to the contact stiffness. 57 To enable visualisation of the simultaneous effects of temperature and stress on reactivity, the decomposition rate data was also plotted as a 3D graph in Figure 4a and Figure 4b.…”

“…During the heating, compression, and sliding phase, there is an exponential decay in the number of intact TNBP and TSBP molecules with simulation time, which is indicative of a first-order reaction. 36 The NEMD simulation data (dashed lines) during the production phase are therefore fit to an exponential function (solid lines) to determine the reaction rate. In previous experiments for ZDDP 12,15,16 and DDP, 16 the rate of tribofilm formation has followed either zero-order or fractional-order kinetics.…”

“…28 For example, NEMD simulations have recently been used to study the tribopolymerization of aldehydes between alumina surfaces, 29 phosphoric acid, 30,31 allyl alcohols, 32 and terpenes 33 between sliding silica surfaces, and cyclopropane carboxylic acid between iron oxide surfaces. 34 They have also been used to study the mechanochemical decomposition of alkyl sulfides between sliding iron surfaces 35 and perfluoropolyethers (PFPEs) between DLC surfaces, 36 the vapour phase lubrication of trialkylphosphates 37 and trialkylphosphites 38,39 between iron surfaces, and the ultralow friction of organic friction modifier additives between DLC surfaces. 40 Most of these NEMD simulations employed many-body empirical force fields, [30][31][32][33][34][35][36] while others use first principles methods, 29,[37][38][39] or sometimes a combination of both of these techniques.…”

“…34 They have also been used to study the mechanochemical decomposition of alkyl sulfides between sliding iron surfaces 35 and perfluoropolyethers (PFPEs) between DLC surfaces, 36 the vapour phase lubrication of trialkylphosphates 37 and trialkylphosphites 38,39 between iron surfaces, and the ultralow friction of organic friction modifier additives between DLC surfaces. 40 Most of these NEMD simulations employed many-body empirical force fields, [30][31][32][33][34][35][36] while others use first principles methods, 29,[37][38][39] or sometimes a combination of both of these techniques. 40 First principles methods are extremely computationally expensive and are therefore limited to single molecules, sub-nanosecond timescales, and very severe conditions.…”