Tribological Properties of Alkylsilane Self-Assembled Monolayers

Because of high hydrophobicity, low surface energies, and compact packing ..... a Lockheed Martin Company, for the United States Department of Energy'...
0 downloads 0 Views 279KB Size
Download PDF
11744

Langmuir 2005, 21, 11744-11748

Tribological Properties of Alkylsilane Self-Assembled Monolayers Christian D. Lorenz,* Michael Chandross, Gary S. Grest, Mark J. Stevens, and Edmund B. Webb III Sandia National Laboratories, Albuquerque, New Mexico 87185 Received June 28, 2005. In Final Form: September 23, 2005

In this study, we perform molecular dynamics simulations of adhesive contact and friction between alkylsilane Si(OH)3(CX2)10CX3 and alkoxylsilane Si(OH)2(CX2)10CX3 (where X ) H or F) self-assembled monolayers (SAMs) on an amorphous silica substrate. The alkylsilane SAMs are primarily hydrogenbonded or physisorbed to the surface. The alkoxylsilane SAMs are covalently bonded or chemisorbed to the surface. Previously, we studied the chemisorbed systems. In this work, we study the physisorbed systems and compare the tribological properties with the chemisorbed systems. Furthermore, we examine how water at the interface of the SAMs and substrate affects the tribological properties of the physisorbed systems. When less than a third of a monolayer is present, very little difference in the microscopic friction coefficient µ or shear stresses is observed. For increasing amounts of water, the values of µ and the shear stresses decrease; this effect is somewhat more pronounced for fluorocarbon alkylsilane SAMs than for the hydrocarbon SAMs. The observed decrease in friction is a consequence of a slip plane that occurs in the water as the amount of water is increased. We studied the frictional behavior using relative shear velocities ranging from v ) 2 cm/s to 2 m/s. Similar to previously reported results for alkoxylsilane SAMs, the values of the measured stress and µ for the alkylsilane SAM systems decrease monotonically with v.

1. Introduction Microelectromechanical systems (MEMS) and emerging nanoelectromechanical systems (NEMS) have experienced a rapid increase in technological interest in the past decade.1-3 As the dimensions of the systems decrease, the surface area/volume ratio increases. The large surface area/volume ratio in MEMS/NEMS causes surface forces (i.e., adhesion and friction) to dominate over inertial forces.4 Therefore, tribological limitations, such as stiction, friction, and wear, are major problems that limit the efficiency, power output, steady-state motion, and reliability of MEMS/NEMS devices.5-7 Nanometer-thick films of organic self-assembled monolayers (SAMs), which have long been used to modify the chemical nature of surfaces, have attracted attention as lubricants for MEMS/NEMS devices. Because of high hydrophobicity, low surface energies, and compact packing structures, SAMs usually have low adhesion and friction, which results in minimal energy losses. SAMs also show stability within a wide range of environmental conditions. The self-assembly characteristic allows SAMs to coat MEMS surfaces well, including the undersides. Among the various types of SAM coatings, the best candidates for MEMS lubrication are alkylsilane and alkoxyl8 molecules. Alkylsilane SAMs are stabilized by hydrogen bonds to the substrate, hydrogen bonds within the SAM, and some covalent bonds to the substrate.9-11 * To whom correspondence should be addressed. E-mail: [email protected]. (1) Zhao, Y.-P.; Wang, L. S.; Yu, T. X. J. Adhes. Sci. Technol. 2003, 17, 519. (2) Komvopoulos, K. J. Adhes. Sci. Technol. 2003, 17, 477. (3) Maboudian, R.; Carraro, C. Annu. Rev. Phys. Chem. 2004, 55, 35. (4) Spearing, S. M. Acta Mater. 2000, 48, 179. (5) Maboudian, R. Surf. Sci. Rep. 1998, 30, 207. (6) Maboudian, R.; Ashurst, W. R.; Carraro, C. Sens. Actuators, A 2000, 82, 219. (7) Bhushan, B. J. Vac. Sci. Technol., B 2003, 21, 2262. (8) Major, R. C.; Kim, H. I.; Houston, J. E.; Zhu, X. Y. Tribol. Lett. 2003, 14, 237.

Alkoxyl SAMs were developed using a deposition and chemistry technique8 that produces SAMs that are primarily covalently bonded to the SiO2 substrate. These SAMs should be more stable to wear and aging. Our previous work treated alkoxylsilane SAMs because of their greater simplicity. Here, we present the results of extensive molecular dynamics (MD) simulations to better understand the tribological properties of physisorbed alkylsilane SAMs on amorphous silicon oxide (the typical surface in MEMS) for both hydrocarbon and fluorocarbon SAMs. We also study the effect of differing amounts of water, in the monolayer range, at the substrate-SAMs interface. Water can help self-assembly by providing a more flexible surface to which hydrogen bonds can form; on the other hand, too much water can lead to degradation.12 Our results for alkylsilane SAMs in the absence of water are compared to prior results for alkoxylsilane SAMs, and little difference in frictional behavior is found. In addition, our results show that the friction for the hydrocarbon alkylsilane SAMs is approximately the same as for fluorocarbon SAMs, which agrees with the experimental observations that the friction of perfluorodecyltrichlorosilane (C8F17C2H4SiCl3, FTS) coated MEMS µ ) 0.08-0.1613,14 and OTS-coated MEMS µ ) 0.07-0.1214,15 are within the same range. While the addition of a small amount of water (1/3 monolayer) has (9) Tripp, C.; Hair, M. Langmuir 1995, 11, 149. (10) Parikh, A. N.; Liedberg, B.; Atre, S. V.; Ho, M.; Allara, D. J. Phys. Chem. 1995, 99, 9996. (11) Stevens, M. J. Langmuir 1999, 15, 2773. (12) Bunker, B. C.; Carpick, R. W.; Assink, R. A.; Thomas, M. L.; Hankins, M. G.; Voigt, J. A.; Sipola, D.; de Boer, M. P.; Gulley, G. L. Langmuir 2000, 16, 7742. (13) DePalma, V.; Tillman, N. Langmuir 1989, 5, 868. (14) Srinivasan, U.; Foster, J. D.; Habib, U.; Howe, R. T.; Maboudian, R.; Senft, D. C.; Dugger, M. T. Solid-State Sensor and Actuator Workshop: Technical Digest 1998, 156. (15) Cle´chet, P.; Martelet, C.; Belin, M.; Zarrad, H.; Jaffrezic-Renault, N.; Fayeulle, S. Sens. Actuators, A 1994, 44, 77.

10.1021/la051741m CCC: $30.25 © 2005 American Chemical Society Published on Web 11/04/2005

Tribological Properties of Alkylsilane SAMs

Figure 1. Snapshots from the simulations of alkylsilane SAMs on amorphous SiO2 for hydrocarbon SAMs (a) with no water, (b) a monolayer of water, (c) fluorocarbon SAMs with no water, and (d) a monolayer of water. The applied normal pressure is 600 MPa, and the relative shear velocity is 2 m/s. Colors: SAMs/ substrate, carbon (light blue), fluorine (green), hydrogen (black), oxygen (red), and silicon (yellow); water, oxygen (gray) and hydrogen (blue).

little effect on friction, we observe a significant decrease in friction for 1 monolayer of water for both fluorocarbon and hydrocarbon alkylsilane SAMs. This decrease correlates with the formation of a slip plane in the layer of water at the interface between the substrate and the SAM. Decreasing shear velocity decreases friction, and this effect is more pronounced in the presence of water. 2. Computational Procedure The alkylsilane systems studied here consist of a pair of apposing SAMs physisorbed to an amorphous SiO2 substrate, as shown in Figure 1. The alkylsilane molecule has a silicon headgroup that is attached to three hydroxyl groups with either a hydrocarbon or fluorocarbon backbone [i.e., Si(OH)3(CX2)10CX3, where X ) H or F]. The alkoxylsilane molecule differs from the alkylsilane molecule in that one of the hydroxyl groups in the silicon headgroup has been replaced with a chemical bond between the silicon and the substrate [i.e., Si(OH)2(CX2)10CX3]. There is a small difference between the alkoxylsilane SAMs that we treated previously and the experimentally studied alkoxyl SAMs.8 Because our early work followed the alkylsilane research, we treated alkoxylsilane SAMs [Si(OH)2(CH2)xCH3] with a silane headgroup that is attached to the substrate, while the recent experiments use alkoxyl SAMs [CH2(CH2)xCH3] with a ethylene headgroup. Once bonded to the silica surface, the only difference is the side groups bonded to the Si atom. We believe that the tribological properties of the alkoxylsilane SAMs are representative of what would be seen for the alkoxyl SAMs, because the difference in the two molecules is on the end attached to the substrate. The alkoxylsilane SAMs are generated in the same manner as discussed in ref 11. The alkylsilane SAM chains are placed randomly within 3 Å of the surface. Initially, the alkylsilane chains are aligned perpendicularly to the amorphous substrate. Upon equilibration, the chains quickly move to bring their

Langmuir, Vol. 21, No. 25, 2005 11745 headgroups closer to the hydroxyls on the substrate. The SAMs are placed at coverages consistent with experimentally observed coverages (32.0 Å2/chain for perfluorinated SAMs16,17 and 25.0 Å2/chain for hydrocarbon SAMs.18,19) Our previous work20 on hydrocarbon alkoxylsilane SAMs showed that the friction coefficient is only weakly dependent upon the chain length and coverage; therefore, in this work, we have studied only one chain length and coverage for each type of SAM. We used the previously reported procedure20,21 to generate an amorphous silica substrate with dimensions 51 × 49 Å and a coverage of hydroxyl groups of 4.0 OH/nm2, which is in the range of experimentally observed values. In the cases where water is present, the water is placed between the SAMs and the silica substrate as is shown in parts b and d of Figure 1. In this study, a monolayer is defined as 300 molecules, which was found from separate MD simulations of water on the amorphous silica substrate. When the alkylsilane SAMs are placed over the water on the substrate, a small amount of water (J10%) moves between the chains, particularly when the systems are compressed. We use the same force field employed in our previous study of fluorinated alkoxylsilane SAMs on amorphous silica,22-24 which combined interaction terms from the OPLS force field22-24 and the COMPASS force field.25,26 The force field parameters for the hydrocarbon SAMs that were not reported previously were taken directly from ref 22. Water was modeled with the TIP3P force field.27 Adhesion simulations are conducted by bringing the two apposing SAMs together at a constant velocity of 0.8 m/s, during which we monitor the change in structure and energetics as a function of the separation distance. For the shear simulations, we apply a constant normal pressure P⊥ (ranging from 50 MPa to 1 GPa) perpendicular to the silica surfaces, while the two SAMs are sheared in opposite directions in the x direction at constant velocities - v/2 and v/2, which correspond to a shear velocity v. We have studied v in the range from 2 cm/s to 2 m/s. To explore the friction in the attractive region (P⊥ < 0), we carried out simulations at constant separation. The range of shear velocities simulated overlaps the range for MEMS devices; however, velocities in the model are above those realized in AFM experiments.28 Simulations discussed in this paper were performed with the LAMMPS MD code.29 In all simulations, the temperature is fixed at 298 K and controlled with a Langevin thermostat, with damping applied only in the direction perpendicular to sliding and compression with a time constant of 0.01 fs-1. Integration is performed using rRESPA,30 with time steps of 0.3 fs for the bond forces, 0.6 fs for the other intramolecular forces, and 1.2 fs for nonbond interactions for the systems that did not have any water present. In the systems with water, the SHAKE algorithm31 was used to constrain the bond lengths of the water molecule and a 1 fs time step was used. The van der Waals (vdW) interaction is cutoff at 10 Å. A slab version of the particle-particle particlemesh algorithm32 is used to compute the long-range Coulombic (16) Schonherr, H.; Vancso, G. J. Langmuir 1997, 13, 3769. (17) Schonherr, H.; Vancso, G. J. Mater. Sci. Eng. 1999, 8-9, 243. (18) Tidswell, I. M.; Ocko, B. M.; Pershan, P. S.; Wasserman, S. R.; Whitesides, G. M.; Axe, J. D. Phys. Rev. B 1990, 41, 1111. (19) Kojio, K.; Takahara, A.; Kajiyama, T. Langmuir 1998, 14, 971. (20) Chandross, M.; Webb, E. B., III; Stevens, M. J.; Grest, G. S.; Garofalini, S. H. Phys. Rev. Lett. 2004, 93, 166103. (21) Lorenz, C. D.; Webb, E. B., III; Stevens, M. J.; Chandross, M.; Grest, G. S. Tribol. Lett. 2005, 19, 93. (22) Jorgensen, W. L.; Maxwell, D. S.; Tirado-Rives, J. J. Am. Chem. Soc. 1996, 118, 11225. (23) Watkins, E. K.; Jorgensen, W. L. J. Phys. Chem. A 2001, 105, 4118. (24) Jorgensen, W. L. Private communication, 2003. (25) Sun, H. Macromolecules 1995, 28, 701. (26) Sun, H.; Rigby, D. Spectrochim. Acta, Part A 1997, 53, 1301. (27) Jorgensen, W. L.; Chandrasekhar, J.; Madura, J. D.; Impey, R. W.; Klein, M. L. J. Chem. Phys. 1983, 79, 926. (28) Zhang, Q.; Archer, L. A. J. Phys. Chem. B 2003, 107, 13123. (29) Plimpton, S. J. J. Comput. Phys. 1995, 117, 1. (30) Plimpton, S. J.; Pollock, R.; Stevens, M. J. Proceedings of the Eighth SIAM Conference on Parallel Processing for Scientific Computing; Heath, M., Ed.; SIAM: Philadelphia, 1997; pp 8-21. (31) Ryckaert, J.-P.; Ciccotti, G.; Berendsen, H. J. C. J. Comput. Phys. 1977, 23, 327. (32) Crozier, P. S.; Rowley, R. L.; Henderson, D. J. Chem. Phys. 2001, 114, 7513.

11746

Langmuir, Vol. 21, No. 25, 2005

Figure 2. (a) Plot of the normal pressure P⊥ as a function of the SAM monolayer separation during compression at a rate of 0.8 m/s for the fluorinated alkylsilane SAMs on an amorphous SiO2 substrate with no water (O) and with a monolayer of water (0). Results for the alkoxylsilane SAMs attached to the SiO2 substrate (]) are shown for comparison. (b) Similar plot for the hydrocarbon SAMs. The inset plots show an enlarged region near the attractive minimums of the alkylsilane SAM systems. interactions. This version models an isolated slab that is infinitely periodic in the x and y dimensions and finite in the z dimension.

3. Results A. Adhesion. Figure 2 shows the normal pressure P⊥ as a function of separation for both the alkylsilane and alkoxylsilane SAMs. In the initial configuration, the terminal groups on the SAMs are separated by 15 Å, and the chains are perpendicular to the substrate. The gap between SAMs is reduced by advancing the surfaces toward one another. The total displacement is 20 Å, which corresponds to a 5 Å overlap between the two monolayers with respect to the starting structure or a separation of -5 Å. At a separation of 15 Å, P⊥ is negligible. As the separation is decreased, P⊥ first becomes slightly negative because of the attractive vdW interactions and then becomes positive because repulsive interactions dominate at short separations. At the initial separation, each monolayer relaxes and tilts. The tilt angle for the alkylsilane SAMs (∼10° for X ) F and ∼30° for X ) H) is larger than the tilt angle of the alkoxylsilane (∼0°) SAMs that are chemically bonded to the substrate. This tilt for the alkylsilane SAMs results in a thinner film and therefore an increased average separation of the terminal groups compared to the alkoxylsilane SAM films. The alkylsilane chains are free to move and thereby more readily relieve stress, unlike the alkoxylsilane chains that are restrained by their chemical bonds to the substrate. Thus, the rise in P⊥ because of repulsive interactions occurs at smaller separations for the alkylsilane SAMs (∼5.0 Å for fluorinated SAMs and ∼4.0 Å for hydrocarbon SAMs) than for the alkoxylsilane SAMs (∼7.5 Å for fluorinated SAMs and ∼6.0 Å for hydrocarbon SAMs). This difference

Lorenz et al.

between X ) F and H may be due to the higher density of the hydrocarbon alkylsilane SAMs (4.0 molecules/nm2) than for the fluorocarbon alkylsilane SAMs (3.1 molecules/ nm2). The maximum adhesions of both the fluorinated and hydrocarbon alkylsilane SAMs (as shown in the inset of Figure 2) are very small and within uncertainty are identical to that found for the alkoxylsilane SAMs (∼25 MPa for fluorinated SAMs and ∼50 MPa for hydrocarbon SAMs). However, because of the noise in the data, it is difficult to make any comments on the relative surface free energies of the hydrocarbon and fluorocarbon SAMs. For the fluorinated SAMs, water at the interface results in a less steep rise in P⊥ once the SAMs overlap. The chains are pressing into the water layer as the monolayers are brought together instead of the chains contacting the stiff substrate. Water molecules penetrate between chains, resulting in a less steep rise in the repulsive force with a decreasing separation. In the hydrocarbon alkylsilane SAMs case, this behavior is not observed, which may be due to the fact that it is more difficult to force water between the more densely packed hydrocarbon chains than the less densely packed fluorocarbon chains. Thus, the mechanism of stress relief provided by the water film for the fluorocarbon systems is not observed for hydrocarbon chains. The presence of water does not significantly affect the maximum adhesion of either the hydrocarbon or fluorocarbon alkylsilane SAMs as seen in the inset of parts a and b of Figure 2, because the adhesion is determined by the interaction of the tails of the chains (opposite from the headgroups and water film). Note that the effect of solvation is not addressed in these simulations because we are using a fixed amount of water and also are using classical force fields that do not account for the reactions that would occur during solvation. B. Friction. Figure 3 shows the measured shear stress τ as a function of the applied pressure P⊥ for the fluorinated and hydrocarbon SAMs. The shear stress values are determined by averaging the measured τ for 6 ns of steadystate behavior; this was preceded by 2.5 ns of simulation to achieve the steady state. From these data, we can calculate a microscopic friction coefficient µ by considering that the friction depends upon both the load and area, i.e., F ) RA + µL, where F is the friction force, R is a constant, A is the area, and L is the applied load. Because our area is constant, the slope of the data in Figures 3 and 4 gives µ. The typically measured or “macroscopic” friction coefficient encompasses a number of issues (i.e. microscopic roughness and asperity contact) that we do not treat in our simulations. When we calculate the microscopic µ, we provide a necessary input to calculate the macroscopic µ. Table 1 shows µ for the various fluorinated and hydrocarbon SAM systems. Generally, the values of µ are the same within error for the alkylsilane and alkoxylsilane SAMs for both the fluorinated and hydrocarbon systems. This agrees with experiments that have shown that the friction coefficients for an alkoxylsilane (OTS, µ ) 0.070.12)14,15 and an alkylsilane (OTE, µ ∼ 0.16)33 are very similar. Figure 3 also shows that the measured τ values at different P⊥ are nearly identical for the alkylsilane and the alkoxylsilane SAM systems. Without water, the frictional behavior of the hydrocarbon and fluorocarbon SAM systems is similar; the values of µ at v ) 2 m/s and 20 cm/s are within error of each other. This is in agreement with values of µ reported for hydrocarbon (OTS, µ ) 0.07-0.12)14,15 and perfluorinated SAMs (FTS, µ ) 0.08-0.16)13,14 on MEMS, which (33) Qian, L.; Tian, F.; Xiao, X. Tribol. Lett. 2003, 15, 169.

Tribological Properties of Alkylsilane SAMs

Figure 3. (a) Plot of shear stress τ as a function of the applied pressure P⊥ during shear at a relative velocity of 2.0 m/s for the fluorocarbon alkylsilane SAMs on amorphous SiO2 with no water (0), with one-third of a monolayer of water (b), and with a monolayer of water (]). Results for alkoxylsilane SAMs attached to amorphous SiO2 with no water (O) are also shown. For a comparison, data for amorphous silica with only a monolayer of water and no SAMs are shown (4). (b) Similar plot for the hydrocarbon SAMs on amorphous SiO2. The inset plots show an enlarged region near the attractive minimums of the alkylsilane SAM systems. Error bars were omitted where the measured error is smaller than the representative symbol.

generally operate in the mm/s to cm/s range. Also, the measured τ values at each P⊥ for the hydrocarbon and fluorocarbon SAMs are very similar. The friction coefficient of the alkylsilane SAM systems with water present decreases for both the fluorocarbon and hydrocarbon alkylsilane systems. In the case of the fluorinated alkylsilane SAMs, we observe that a small amount of water has a weak effect on the value of µ (1/3 monolayer, µ ) 0.163) as compared to when no water is present (µ ) 0.187). However, when the amount of water is increased, we observe that the value of µ decreases significantly [µ(monolayer) ) 0.072]. For the hydrocarbon alkylsilane SAMs, we also observe that when there is a large amount of water present the value of µ decreases [µ(monolayer) ) 0.132 < µ(no water) ) 0.190]. For comparison, we calculated µ for the amorphous silica substrate with water but no SAMs. The value of µ that we obtain for a monolayer of water between two silica substrates is 0.035 ( 0.001, which compares well to the value that Qian et al.33 found for the friction between silica at high humidity and an AFM tip made from Si3N4 (µ ∼ 0.03). The presence of water also affects the individual values of τ at different P⊥. The calculated τ values decrease significantly with an increasing amount of water. The inset plots of Figure 3 show the frictional behavior of the alkoxylsilane and alkylsilane SAM systems in the attractive region, which previously has not been measured in simulations. These results were obtained by shearing at constant separation at a relative shear velocity of 2.0 m/s. A natural consequence of the shape of the attractive region is that the same P⊥ (3/4 of a monolayer) of water, there is very little motion at the SAM/SAM interface. Therefore, the friction measured in these cases is mainly the friction between the water and the SAM molecules. The velocity difference between the water molecules and the substrate also increases as the amount of water increases. This is a result of water molecules that have left the distinct water layer and occupy space within the SAM. In Figure 5, we do not distinguish between water molecules in the water layer and those in the SAM. Because the latter move at the same rate as the SAM, the velocity in Figure 5 is somewhat less than the velocity of just the distinct water layer. This only decreases slightly the calculated slip between the water and the substrate. Results for the hydrocarbon alkylsilane SAMs are nearly identical to the fluorocarbon alkylsilane SAMs. In both SAM systems, a slip plane develops between the water layer and the SAM and it is the existence of this slip that accounts for the reduction in friction observed when sufficient water is present. 4. Conclusion This work presents the first simulations of alkylsilane (physisorbed) SAMs on amorphous silica. These are also the first simulations to study the effect of water between SAMs and the substrate. We observe that the repulsive region of the alkylsilane SAMs begins at smaller separations than for the alkoxylsilane SAMs and that the presence of water causes the repulsive region to be softer than when there is no water present. The frictional coefficients of the alkylsilane SAMs are nearly identical to that of the alkoxylsilane SAMs, which is in good agreement with experiment.14,15,33 We also investigated the effect of the relative shear velocity on the frictional behavior and found that both the calculated τ and µ for the alkylsilane SAM systems decrease monotonically with the relative shear velocity, which is consistent with the previous simulation results for hydrocarbon alkoxylsilane SAMs.20 We studied the effect of water on the frictional behavior and found that when a small amount of water (e1/3 monolayer) is present there is little difference in µ or τ, while when sufficient water (∼1 monolayer) is present there are large decreases in µ and τ. The friction reduction is shown to result from a slip plane that develops between the water layer and SAMs. We are conducting simulations that model the AFM experimental setup with tip-shaped silica substrates in contact with flat substrates that will address whether the slip plane causes the wear properties to be poor as well as provide further insight into the physics of the nanotribological properties of such a system. Acknowledgment. Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000. LA051741M