Thermal decomposition and desorption of PFPE Zdol on a DLC substrate using quartic bond interaction potential

Abstract

In heat assisted magnetic recording (HAMR) system, heating of the hard disk magnetic layer is carried out by applying laser rays during the movement of the read/write head over the carbon overcoat for the purpose of reading and writing on its magnetic layer. Depletion of PFPE Zdol occurs because of thermal decomposition and desorption on a DLC substrate due to laser heating and this model is developed using the coarse-grained bead spring based on quartic and van der Waals interaction potential. The effects of temperature on the bond breaking phenomenon of PFPE Zdol due to thermal decomposition and thermal desorption were studied. To support the reliability of the present simulation results by a quartic potential, the end bead density and total bead density on a DLC substrate obtained by the finitely extensible non-linear elastic (FENE) and quartic potential are shown in a comparative manner.

---

1. Introduction

In MEMS/NEMS, PFPE films are used as a permanent lubricant to protect these structures from erosion.^1,2 The hard disk industry uses thin PFPEs (with a thickness of 1–2 nm) deposited onto an amorphous carbon overcoat to protect the underlying hard disks from mechanical damage during intermittent contacts with the read–write heads as the disk rotates. Conventional reading–writing technique to increase the memory of a hard disk has a limitation to reach above 1 TB per inch². The only way to surpass this value is through heat-assisted magnetic recording technology, which is an alternative way to reach memory density of up to 10 TB per inch². The main difficulty is the stability of the lubricants on the hard disk, which is a great concern to the scientific community. Ganesh et al.³ showed the successful fabrication of a thin, transparent, and homogenous coating of perfluoropolyether (PFPE) on a smooth glass surface by an electro spraying technique. The wetting behavior of three nanometer thick perfluoropolyether (PFPE) polymers with the same backbone but different end-groups was studied by contact angle tests, and the simultaneous governing mechanisms such as oleophobicity/hydrophilicity were investigated.⁴ Li and Wong⁵ carried out molecular dynamics simulations to study the thermal stability of thin lubricant films. Lubricant film thickness, bonding ratio, molecular weight and laser heating rate play important roles in the depletion of lubricant films.^6–9 Wu⁸ described a model for lubricant flow dynamics under a scanning laser beam and studied the mechanisms of the lubricant on both glass and aluminum disks. The main causes of the depletion of the lubricant film on a hard disk surface are thermal decomposition and desorption; however, the mechanisms of thermal decomposition and desorption are not understood either by the theoretical or experimental way to date.⁹ The magnetic layer of a hard disk is heated above its Curie temperature in heat-assisted magnetic recording systems.¹⁰ Pacansky and Waltman¹¹ reported that PFPEs degrade due to main chain scission when subjected to electron irradiation. Heller et al.¹² investigated the effect of localized laser heating on molecularly thin PFPE lubricants for short times. Zhu et al.¹³ used mass spectroscopy to study the thermal chemistry of laser-irradiated PFPEs on disk media. It was found that there is a stronger likelihood of laser-induced scission and rearrangement of the PFPE backbone chain than that of its end group with the laser power density in the range of 60–140 mJ cm⁻². Li et al.¹⁴ employed a molecular dynamics (MD) simulation to investigate the evolution and degradation characteristics of ultrathin PFPEs under laser irradiation. They observed that the lubricant diffused radially away from the heating spot and a raised ridge was formed around the depleted zone due to the non-uniform distribution of surface tension and thermo capillary stress. Waltman and Tyndall^15,16 measured the evaporation rates of both polydisperse and monodisperse PFPEs from a CH_x surface at various temperatures. Lim and Gellman⁹ investigated the kinetics of laser induced desorption and decomposition of Fomblin Zdol on carbon overcoats. They observed that under HAMR conditions, lubricant films with molecular weight of less than 3000 g mol⁻¹ desorb more than that of decomposition on the disk surface during rapid heating, whereas decomposition occurs at temperatures (<400 °C) expected in HAMR for higher molecular weight lubricants. The decomposition behavior of Fomblin Zdol and Fomblin Z on an amorphous carbon overcoat using thermal desorption spectroscopy,¹⁷ the decomposition mechanisms of PFPE Zdol at the head disk interface under sliding conditions,¹⁸ the desorption mechanisms of Fomblin Zdol on a graphite surface,¹⁹ the thermal stability of Fomblin Z on an Al₂O₃ surface,²⁰ the thermal decomposition of Fomblin Zdol on graphite and amorphous carbon surfaces,²¹ the thermal and electron-induced decomposition of Fomblin Zdol on a carbon overcoat,²² and the desorption and decomposition of polyperfluorinated ether using transmission infrared spectroscopy²³ were studied experimentally and reported. Lei et al.¹⁷ observed that the decomposition of Fomblin Zdol on an amorphous carbon surface occurs in the temperature range of 600–700 K and they suggest that desorption is the result of the decomposition of Fomblin Zdol. Kasai et al.²⁰ observed during the experiment that Fomblin Z on an Al₂O₃ surface degrades at 473 K in two stages and the vigorous second stage leads to complete loss of fluid. Vurens and Mate²¹ concluded from their study on Fombin Zol that the two desorption features of Fomblin Zdol on graphite surface happens at temperatures 620 and 770 K; on the other hand, the main desorption features of Fomblin Zdol on an amorphous carbon surface are at temperature 510 and 650 K.

Tagawa et al.^24,25 carried out a series of experiments to investigate the lubricant depletion characteristics of HAMR systems. They found that the lubricant depletion due to laser heating is largely affected by film thickness, bonding ratio, molecular weight and end groups of a lubricant. Ma et al.²⁶ studied lubricant depletion on a self-developed HAMR tester and found that the lubricant depletion depth is proportional to the logarithm of laser heating duration. Moreover, they suggested that almost all the lubricant on the disk would be depleted over the lifetime of the hard disk drive. Bei et al.³⁹ employed a molecular dynamics simulation coupled with a course-grained bead-spring polymer model to study the depletion of PFPE lubricants. Their study shows that the desorption of PFPE lubricants is favored over decomposition at high temperatures. The photo-thermal stability and tribological properties of Zdol lubricant, modified by substituting the OH groups with benzophenone, were investigated using laser beam exposure by Gauvin et al.²⁷ and they concluded that their modified Zdol lubricant is suitable for heat assistant magnetic recording systems.

From the above mentioned study, it is observed that the desorption and decomposition rates of functional lubricant films on a thin DLC film using a bond-breaking potential are not available. In the present study, we study desorption and decomposition of PFPE Zdol on a thin DLC film considering interaction of all carbon atoms with the lubricant film. A new potential quartic is used instead of the FENE potential, which allows for bond-breaking. To satisfy the current results obtained by the quartic potential, at low temperature, end bead density and total bead density with the film thickness were computed using quartic and FENE potential and are compared to each other. To study the validity of the present simulation results, the number density as a function of the film thickness and the decomposition and desorption rate with the film thickness are compared with the published results. The effects of temperature on the desorption and decomposition rates of PFPE Zdol were investigated in detail. With the graphical results, some snapshots during the depletion of lubricants are described to understand the physics of lubricant on a DLC substrate during heating.

2. Simulation details

Fig. 1 shows the interaction between the beads of lubricant molecules and a DLC substrate. The DLC substrate consists of carbon atoms and is prepared by subsequent heating and quenching of BCC or FCC diamond crystals by molecular dynamics simulations following the procedures^28–31 using the Tersoff potential.^32,33 The DLC substrate obtained from atomistic simulation was compressed to half of its original dimension so that no lubricant molecules could diffuse inside the DLC substrate during simulation.^34,35 Each lubricant molecule of PFPE Zdol was constructed considering the bead spring model. In this model, each bead consisted of several PFPE Zdol atoms. Each lubricant molecule consisted of 10 beads and among them, the end beads were considered as polar or functional beads and the rest of the beads were considered as backbone beads. The neighboring beads in each chain were connected by bonds, which were modeled using the quartic potential and this bond can break any time if the bond exceeds its allowable extension limit. In particular, the DLC substrate consists of many dangling bonds and for this reason the functional beads of lubricant molecules are attracted more than the rest of the lubricant molecules.^36,37 Moreover, the interaction between the functional beads is more than that of non-functional beads or that of functional and non-functional beads. Therefore, a special treatment was done between the interaction of functional beads to the DLC substrate and among functional beads. The detailed description of the potentials and their interaction are described in the theoretical section using Fig. 1.

Fig. 1 Coarse-grained bead spring model of functional lubricant PFPE Zdol on a DLC thin film.

900 lubricant molecules were considered in the present study. 900 lubricant molecules were placed on a DLC substrate and simulated for long time by the subsequent isothermal heating to obtain the initial configuration of the PFPE Zdol film, which is shown in Fig. 2.^34,35 After getting an initial configuration of the PFPE Zdol film on a DLC substrate, a three dimensional simulation was developed that considers periodic boundary conditions in the x and y directions and in the z direction shrink-wrapped boundary conditions³⁸ were used because during the simulation, the simulation box dimension altered in the z direction. The time step utilized during the simulation was 0.005τ. During the simulation, the positions of the carbon atoms in a DLC substrate were kept fixed. Isothermal heating was carried out on the molecules of PFPE Zdol lubricant molecules on a DLC substrate at different temperature. During simulations at high temperature, bond breaking occurred and new short chains were generated, which enhanced the thermal decomposition and desorption. The simulations were carried out separately at different isothermal heating treatment for 2 000 000 steps, for example T* = 1 to 6 using the NVT (N = same number of atoms, V = volume, T = temperature) canonical ensemble, and we calculated how many bonds break and how many lubricant molecules were washed away by desorption phenomenon during the simulation at a specific temperature at last time step. Gradual heating effects were also considered by heating the lubricant molecules from T* = 1 to 6 at a slow rate at a time step of 2 000 000 using the NVT canonical ensemble in the same simulation. The simulation was rescaled at every 20 steps. Thermal decomposition and thermal desorption of PFPE Zdol on a DLC substrate was calculated for each case at a time step of 2 000 000. To verify the soundness of the present simulation results, PFPE Zdol lubricant molecules on a DLC substrate were simulated using FENE and quartic potentials for bonding between beads of PFPE Zdol molecules separately keeping the other potential parameter the same at a very low temperature T* = 1.3.

Fig. 2 Front view of initial configuration of PFPE Zdol film on a DLC substrate.

3. Theories of functional lubricant on a thin DLC film

To describe the bond interaction between the neighboring beads in each lubricant molecule, FENE potential^39,40 was used in the coarse-grained bead spring model. As the FENE potential is unable to break bonds during the simulation, the quartic potential^41,42 was used instead of the FENE potential.^39,40where r is the distance between two neighboring beads, y = r − r_b shifts the quartic center from the origin and r_b is the cutoff length, wherein the potential goes smoothly to a local maximum. When the bond breaks, the quartic term is deactivated to impede the bond from reforming. This potential mimics the FENE potential by setting the following parameters as follows:^41,42

k₄ = 1434.3ε/σ⁴, b₁ = −0.7589σ, b₂ = 0.0, r_b = 1.5σ, and U₀ = 67.223ε.

The expression of non-bonded potential LJ 12/6 is as follows:

where σ is the L-J diameter of the beads and ε is the potential depth.

Eqn (2) is used for the interaction between non-functional beads and nonfunctional bead to functional bead of PFPE Zdol. As dangling bonds in the DLC film are available, polar beads attract toward the DLC surface to form bonds with hydrogen from their OH group and form an ionic bond with the DLC surface. Therefore, an extra attractive potential was added with the van der Waals interactions to satisfy the conditions.⁴⁴ For the extra interaction due to the functional effects of the end beads of functional lubricant PFPE Zdol,^{34,35,39,40,44} the expression used for the extra interaction between them by an exponential potential is as follows:

For the extra interaction due to the functional effects of the end beads of functional lubricant PFPE Zdol and DLC C atom,^{34,35,39,40,44} the expression used for the extra interaction between them by an exponential potential is as follows:

Combining the non-bonded potential LJ 12/6 and the exponential potential for the interaction between the end beads of functional lubricant PFPE Zdol, the expression of the resultant energy between polar end bead to polar end bead^{34,35,39,40,44} is as follows:

Combining the non-bonded potential LJ 12/6 and exponential potential for the interaction between the end beads to DLC C atoms, the expression of the resultant energy between polar end bead and DLC C atom^{34,35,39,40,44} is as follows:

The detailed descriptions of the parameters in the above mentioned equations are given in previous literature.^{34,35,39,40,43–45}

4. Results and discussion

From the experimental studies, it is observed that the depletion of PFPE Zdol lubricant on a DLC substrate occurs at high temperature. To date, the reason for the depletion of PFPE Zdol on a DLC substrate is not clear. Therefore, the theoretical analysis of the depletion of PFPE Zdol on a DLC substrate is necessary to understand the actual physics and chemistry of PFPE Zdol on a DLC substrate at low to high temperature. Two main reasons for the depletion of PFPE Zdol on a DLC substrate are assumed to be thermal decomposition and thermal desorption. In general, the static, kinetic and rheological properties of PFPE Zdol on a DLC substrate can be investigated using a coarse-grained bead spring model based on the FENE potential but it cannot be applied at higher temperatures because at these temperatures lubricant bonds break due to the resulting increased bond length. Therefore, the coarse-grained bead spring model based on the quartic potential allows for bond breaking during the simulation at higher temperatures. At a low temperature, T* = 1.3, PFPE Zdol on a DLC substrate was studied using two different potentials, the quartic and FENE, keeping the same van der Waals interaction and extra attractive potential to verify the reliability of the quartic potential on the coarse-grained bead spring model. Fig. 3(a) illustrates the end bead density of PFPE Zdol on a DLC substrate as a function of film thickness using two different potentials, quartic and FENE, based on coarse-grained bead spring models. Two layers of PFPE Zdol on a DLC films are formed in which the first peak of the end bead density of PFPE Zdol is largely higher than that of second peak for the case of both potentials. In the 1st layer, the end bead density of PFPE Zdol using a quartic potential is slightly smaller than that of the FENE potential except for the first layer, the end bead density of PFPE Zdol for the cases of both potentials coincide with each other. Fig. 3(b) illustrates the total bead density of PFPE Zdol on a DLC substrate using two different potentials, quartic and FENE, in a comparative manner. The first peak of the total bead density of PFPE Zdol using the quartic potential is higher than that of the first peak of the FENE potential. Except for the first peak, the total bead density of PFPE Zdol on a DLC substrate for both of the potentials is almost the same considering film thickness. From Fig. 3(a) and (b), it is clearly observed that the phase transition of PFPE Zdol originates between the first and second layers. Fig. 4 shows a comparative study of the number density of PFPE Zdol on a DLC substrate as a function of the film thickness using the quartic potential based coarse-grained bead spring model with that of PFPE on a single wall obtained by Bei et al., 2013. The obtained number density of PFPE Zdol on a DLC substrate qualitatively agrees well with that of Bei et al., 2013, and some of the discrepancies of our results with that of Bei et al., 2013 occurs because Bei et al., 2013 used non-functional PFPE lubricant instead of functional PFPF Zdol and single wall instead of a DLC substrate in their simulations. Fig. 5 illustrates the total bead density of PFPE Zdol on a DLC substrate at a higher temperature such as T* = 3.1, 3.4, 3.5 and 3.9. From the this figure, a clear phase transition zone of PFPE Zdol is observed wherein the liquid phase of PFPE Zdol transforms into a gaseous phase. From this figure, it is clear that with the increase of temperature, the total bead density not only decreases with the film thickness but also the film thickness of PFPE Zdol on a DLC substrate decreases with the increase of temperature. From Fig. 5, it is evident that heating of PFPE Zdol on a DLC substrate at a higher temperature is responsible for higher lubrication depletion than the heating of PFPE Zdol at a lower temperature. Fig. 6(a) shows the initial configuration of PFPE Zdol on a DLC film at a temperature T* = 3 at time step = 0. Fig. 6(b) shows the final configuration of PFPE Zdol on a DLC substrate during isothermal heating at time = 10 000τ. After time 10 000τ due to isothermal heating, the film of PFPE Zdol on a DLC substrate depletes which is clearly observed if compared to the snapshots of PFPE Zdol on a DLC substrate as observed in Fig. 6(a) and (b). At time 10 000τ, it is observed that in a few spots on the DLC substrate, PFPE Zdol lubricant is removed due to isothermal heating at temperature T* = 4.0.

Fig. 3 Effect of two different potentials such as quartic and FENE: (a) on the film thickness as a function of end bead density; (b) film thickness as a function of total bead density of PFPE Zdol lubricant molecules on a DLC film at a temperature, T* = 1.3.

Fig. 4 Comparitive study of the number density of beads of PFPE Zdol on a DLC substrate with that of non-functional PFPE on a single wall (Bei et al., 2013) at a reduced temperature T* = 1.3.

Fig. 5 Effect of temperature on the total bead density of PFPE Zdol lubricant as a function of film thickness at temperature T* = 3.1, 3.4, 3.5, 3.9.

Fig. 6 Snapshots of PFPE Zdol lubricant molecules on a DLC substrate at constant temperature, T* = 3: (a) at time, τ = 0, (b) a τ = 10 000 during isothermal heating.

Fig. 7 illustrates the comparative study of the bond percentage of PFPE Zdol on a DLC substrate with that of non-functional PFPE on a single wall obtained by Bei et al., 2013. Up to temperature T* = 4, the bond percentage of PFPE Zdol on a DLC substrate by us quantitatively agrees well with that of non-functional PFPE on a single wall obtained by Bei et al., 2013, but the deviation between the bond percentage of PFPE Zdol on a DLC substrate and non-functional lubricant PFPE on a single wall obtained by Bei et al., 2013 increases after temperature T* = 4 with the increase of temperature because the functionality of PFPE Zdol lubricants and the DLC substrate play an important role with such types of deviation from that of Bei et al., 2013. In the case of PFPE Zdol lubricants, functional beads form strong bonds with the DLC substrate. As the non-functional PFPE has no functional beads, they cannot form strong bonds with the carbon overcoat. Therefore, at high temperatures, the decomposition rate of PFPE Zdol should be lower than that of non-functional PFPE lubricants and this effect is strongly observed in the comparative study, as shown in Fig. 7. Fig. 8(a) illustrates the bond percentage of PFPE Zdol on a DLC substrate for the temperature range T* = 1.5 to 4. Fig. 8(b) illustrates the bond percentage of PFPE Zdol on a DLC substrate as a function of elapsed time during the simulation at the temperature range T* = 4 to 6. From Fig. 8(a), at temperature T* = 1.5, the bond percentage of PFPE Zdol is nearly 100%, which means that bond of PFPE Zdol on a DLC substrate does not break and ensures that at T* = 1.5, no thermal decomposition of PFPE Zdol on a DLC substrate happens at this temperature during the simulation. From Fig. 8(a), it is observed that at temperature T* = 2, the total bond percentage of PFPE Zdol on a DLC substrate is very high as compared to temperature T* = 3, i.e. at temperature T* = 2, the rate of the bond breaking phenomenon is extremely low as compared to at higher temperature T* = 3 or other temperatures. From Fig. 8(a) and (b), it is evident that with the increase of temperature, the bond percentage of PFPE Zdol on a DLC substrate decreases, which means that increased temperature enhances the bond breaking phenomenon of PFPE Zdol on a DLC substrate, increasing the thermal decomposition of PFPE Zdol on a DLC substrate. After the elapsed time 1000τ, the rate of bond percentage with time during the simulation gradually decreases as before because after the elapsed time 1000τ, many beads of PFPE Zdol lubricant exceed their force cut off distance and many short chains form after 1000τ, the percentage of long chain decreases for which the increasing rate of bond breaking decreases in the range of elapsed time from 1000τ to 10 000τ.

Fig. 7 Comparitive study of bond change of PFPE Zdol on a DLC substrate with that of non-functional PFPE on a single wall (Bei et al., 2013) with reduced temperature.

Fig. 8 Effect temperature on the bond percentage of PFPE Zdol lubricant molecules on a DLC substrate: (a) at temperature range, T* = 1.5–4; (b) at temperature range, T* = 4–6.

To validate the result of thermal desorption of PFPE Zdol on a DLC substrate obtained by us is compared with that of non-functional PFPE lubricant on a single wall obtained by Bei et al. 2013 at two different temperatures such as T* = 3 and 3.3, as shown in Fig. 9. For the case of temperature T* = 3, it is observed that weight percentage of PFPE Zdol on a DLC substrate obtained by us agrees well qualitatively with that of non-functional PFPE on a single wall obtained by Bei et al., 2013 but at T* = 3.3, the deviation of the weight percentage of PFPE Zdol on a DLC substrate obtained by us and the weight percentage of non-functional PFPE on a single wall obtained by Bei et al., 2013 is slightly larger as compared to temperature T* = 3 due to the functionality of the PFPE Zdol and DLC substrate. Fig. 10(a) illustrates the weight percentage of PFPE Zdol on a DLC substrate as a function of elapsed time for the temperature range T* = 1.5 to 4. Fig. 10(b) illustrates the weight percentage of PFPE Zdol on a DLC substrate as a function of elapsed time for the temperature range T* = 4 to 6. At temperatures T* = 1.5 and 2, the bond percentage of PFPE Zdol on a DLC substrate as a function of elapsed time decreases at a constant rate. At temperature T* = 1.5, the thermal desorption of PFPE Zdol is very low, which can be negligible because at elapsed time, the weight percentage of PFPE Zdol on a DLC substrate is close to 100%. With the increase of temperature, the weight percentage of PFPE Zdol on a DLC substrate decreases. The desorption rate of PFPE Zdol is several times than that of thermal decomposition during heating PFPE Zdol on a DLC substrate. Up to elapsed time 1000τ, the desorption rate of PFPE Zdol on a DLC substrate is constant and higher than that of the rest of the elapsed time. Thermal decomposition enhances the thermal desorption of PFPE Zdol because during heating many small lubricant chain forms due to bond breaking, which increases the thermal desorption. Fig. 11(a) illustrates the initial configuration of PFPE Zdol on a DLC substrate, which is heated at varying temperatures T* = 1–6 during elapsed time 10 000τ, as observed in Fig. 11. At the final stage, most of the PFPE Zdol molecules are washed away from the DLC substrate due to having higher temperature T* = 6 when it reaches at final stage of heating time 10 000τ, as shown in Fig. 11(b). Some lubricant molecules are attached on the DLC substrate and form strong bond on the DLC substrate. PFPE Zdol lubricant molecules on the DLC substrate form a strong bond and are strongly bonded with the DLC substrate and the lubricant molecules on the top surface of lubricant film are non-bonded. During heating, the kinetic energy of each bead increases; therefore, the bond length increases with the increase of temperature during heating. When the bond length exceeds its allowable limit, it breaks and new short chains appear on the film. During heating at high temperature, non-bonded lubricants are removed rapidly by desorption and the PFPE Zdol thickness decreases. As the lubricant on the DLC substrate forms strong bonds with the DLC substrate, all lubricant molecules are not washed away from the DLC substrate, as shown in Fig. 11(b). The bond percentage of PFPE Zdol on a DLC substrate is also studied in two different cases, as shown in Fig. 12(a) and (b). In the first case, the bond percentage of PFPE Zdol is studied as a function of elapsed time during heating temperature T* = 1–6, as shown in Fig. 12(a). In the second case, the bond percentage of PFPE Zdol is studied as a function of temperature T* during heating up to elapsed time 10 000τ, as shown in Fig. 12(b). During heating of PFPE Zdol on a DLC substrate, the bond percentage of PFPE Zdol decreases at a very low constant rate up to temperature T* = 1 to 3 or up to the elapsed time 4000τ during the simulation. After reaching temperature T* = 3 or the elapsed time 4000τ, the decreasing rate of bond percentage of PFPE Zdol on a DLC substrate increases up to T* = 5 or the elapsed time 7000τ and after reaching temperature T* = 5 or the elapsed time 7000τ, the bond percentage of PFPE Zdol on a DLC substrate increases at a constant rate to reach the final stage at time 10 000τ or temperature T* = 6. Thermal decomposition of PFPE Zdol is dominating in the temperature range T* = 4 to 6 which agrees qualitatively well with that of the experimental study of the thermal decomposition of Fomblin Zdol on a carbon surface by Lei et al.¹⁷ In the theoretical study of non-functional PFPE lubricants³⁹ and the experimental study of Fomblin Zdol on a carbon surface, desorption is favored over decomposition. Similar effects are also observed in our study and from the present study of thermal decomposition and desorption of PFPE Zdol on a DLC substrate, desorption of PFPE Zdol is favored over decomposition.

Fig. 9 Comparative study of mass change PFPE Zdol on a DLC substrate with that of non-functional PFPE on a single wall (Bei et al., 2013) with time τ.

Fig. 10 Effect of temperature on the weight percentage of PFPE Zdol lubricant molecules on a DLC substrate: (a) at temperature range, T* = 1.5–4; (b) at temperature range, T* = 4–6.

Fig. 11 Snapshots of PFPE Zdol lubricant molecules on a DLC substrate at varying temperature, T* = 1–6: (a) at time, τ = 0; (b) at τ = 10 000.

Fig. 12 Bond percentage of PFPE Zdol lubricant molecules on a DLC substrate: (a) as a function of elapsed time; (b) as a function temperature during heating at increased temperature T* = 1 to 6.

5. Conclusion

From the comparative study of the end bead density and total bead density of PFPE Zdol on a DLC substrate at a temperature T* = 1.3 obtained by quartic potential and FENE potential agree well each other. However, at higher temperature, the bond breaking phenomenon occurs, which cannot be solved using FENE potential but the quartic potential allows for bond breaking phenomenon during heating at high temperature. From the total bead density of PFPE Zdol using the quartic potential, a clear phase transition zone is observed in which liquid PFPE Zdol enters into the gaseous phase. Thermal decomposition and thermal desorption of PFPE Zdol on a DLC substrate increase with increased temperature. Thermal desorption is the principle cause of lub desorption. However, due to thermal decomposition, many long lubricant chain turn into short chains, which enhances the thermal desorption of PFPE Zdol lubricant molecules on a DLC substrate.

References

1. B. Marchon, IEEE Trans. Magn., 2009, 45, 872–876.
2. Z. Tao and B. Bhusan, Wear, 2005, 259, 1352–1361.
3. V. A. Ganesh, S. S. Dinachali, S. Jayaraman, R. Sridhar, H. K. Raut, A. Góra, A. Baji, A. S. Nair and S. Ramakrishna, RSC Adv., 2014, 4, 55263–55270.
4. Y. Wang, J. Knapp, A. Legere, J. Raney and L. Li, RSC Adv., 2015, 5, 30570–30576.
5. B. Li and C. H. Wong, Comput. Mater. Sci., 2014, 93, 11–14.
6. N. Tagawa, H. Andoh and H. Tani, Tribol. Lett., 2010, 37, 411–418.
7. N. Tagawa, T. Miki and H. Tani, Tribol. Lett., 2012, 47, 123–129.
8. L. Wu, Nanotechnology, 2007, 18, 215702.
9. M. S. Lim and A. J. Gellman, Tribol. Int., 2005, 38, 554–561.
10. W. A. Challener, C. Peng, A. V. Itagi, D. Karns, W. Peng, Y. Peng, X. Yang, X. Zhu, N. J. Gokemeijer, Y.-T. Hsia, G. Ju, R. E. Rottmayer, M. A. Seigler and E. C. Gage, Nat. Photonics, 2009, 3, 220–224.
11. J. Pacansky and R. J. Waltman, Chem. Mater., 1993, 5, 486–494.
12. J. Heller, C. M. Mate, C. C. Poon and A. C. Tan, Langmuir, 1999, 15, 8282–8287.
13. L. Zhu, T. Liew and T. C. Chong, Appl. Phys. A: Mater. Sci. Process., 2002, 75, 633–636.
14. Y. Li, C. H. Wong, B. Li, S. K. Yu, W. Hua and W. Zhou, Soft Matter, 2012, 8, 5649–5657.
15. R. J. Waltman, G. W. Tyndall, J. Pacansky and R. J. Berry, Tribol. Lett., 1999, 7, 91–102.
16. G. W. Tyndall and R. J. Waltman, J. Phys. Chem. B, 2000, 104, 7085–7095.
17. R. Z. Lei, A. J. Gellman and P. Jones, Tribol. Lett., 2001, 11, 1–5.
18. J. Wei, W. Fong, D. B. Bogy and C. S. Bhatia, Tribol. Lett., 1998, 5, 203–209.
19. K. R. Paserba and A. J. Gellman, J. Phys. Chem. B, 2001, 105, 12105–12110.
20. P. H. Kasai, W. T. Tang and P. Wheeler, Appl. Surf. Sci., 1991, 51, 201–211.
21. G. H. Vurens and C. M. Mate, Appl. Surf. Sci., 1992, 59, 281–287.
22. J.-L. Lin, C. S. Bhatia and J. T. Yates Jr, J. Vac. Sci. Technol., A, 1995, 13, 163–168.
23. L. M. Ng, E. Lyth, M. V. Zeller and D. L. Boyd, Langmuir, 1995, 11, 127–135.
24. N. Tagawa, R. Kakitani, H. Tani, N. Iketani and I. Nakano, IEEE Trans. Magn., 2009, 45, 877–882.
25. N. Tagawa and H. Tani, IEEE Trans. Magn., 2011, 47, 105–110.
26. Y. S. Ma, X. Y. Chen and B. Liu, Tribol. Lett., 2012, 47, 175–182.
27. M. Gauvin, H. Zhang, B. Suen, J. Lee, H. J. Kang and F. E. Talke, IEEE Trans. Magn., 2011, 47, 1849–1854.
28. H. Lan, T. Kato and C. Liu, Tribol. Int., 2011, 44, 1329–1332.
29. Z. D. Sha, P. S. Branicio, V. Sorkin, Q. X. Pei and Y. W. Zhang, Diamond Relat. Mater., 2011, 20, 1303–1309.
30. J. F. Ziegler, M. D. Ziegler and J. P. Biersack, Nucl. Instrum. Methods Phys. Res., Sect. B, 2010, 268, 1818–1823.
31. I. N. Remediakis, G. Kopidakis and P. C. Kelires, Acta Mater., 2008, 56, 5340–5344.
32. J. Tersoff, Phys. Rev. B: Condens. Matter Mater. Phys., 1988, 39, 5566–5568.
33. J. Tersoff, Phys. Rev. Lett., 1988, 61, 2879–2882.
34. S. K. Deb Nath, C. H. Wong, V. Sorkin, Z. D. Sha, Y. W. Zhang and S.-G. Kim, Tribol. Trans., 2013, 56, 255–267.
35. S. K. Deb Nath, Appl. Phys. A: Mater. Sci. Process., 2014, 117, 857–870.
36. P. H. Kasai, A. Wass and B. K. Yen, Journal of Information Storage and Processing Systems, 1999, 1, 245–258.
37. P. H. Kansai and A. M. Spool, IEEE Trans. Magn., 2001, 37, 929–933.
38. S. J. Plimpton, J. Comput. Phys., 1995, 117, 1–19.
39. L. Bei, C. H. Wong and Q. Chen, Soft Matter, 2013, 9, 700–708.
40. Q. GuO, S. Izumisawa, D. M. Phillips and M. S. Jhon, J. Appl. Phys., 2003, 93, 8707–8709.
41. M. Tsige and M. J. Stevens, Macromolecules, 2004, 37, 630–637.
42. M. J. Stevens, Macromolecules, 2001, 34, 1411–1415.
43. S. Izumisawa and M. S. Jhon, J. Appl. Phys., 2002, 91, 7583–7585.
44. P. S. Chung, H. Park and M. S. Jhon, IEEE Trans. Magn., 2009, 45, 3644–3647.
45. X. Li, Y. Hu, L. Jiang and J. Zhang, J. Chem. Phys., 2008, 128, 194904.

---