3166
Langmuir 2004, 20, 3166-3172
Autophobic Dewetting of Z-Tetraol Perfluoropolyether Lubricant Films on the Amorphous Nitrogenated Carbon Surface R. J. Waltman Hitachi Global Storage Technology, 5600 Cottle Road, San Jose, California 95193 Received April 21, 2003. In Final Form: January 23, 2004 The thermodynamic stability of thin films of the perfluoropolyether (PFPE) Z-Tetraol, as a function of molecular weight, on amorphous nitrogenated carbon, CNx, is investigated. An optical surface analyzer is used to image the autophobic dewetting of the Z-Tetraol films. Film dewetting results when the PFPE film thickness applied to the CNx surface exceeds a critical value. This critical dewetting thickness is identified as the monolayer thickness of the adsorbed PFPE film via measurements of the changes in the surface energy as a function of lubricant film thickness. The observed dewetting coincides with the film thickness at which the disjoining pressure goes to zero. The critical dewetting thickness is dependent on the PFPE molecular weight.
Introduction Molecularly thin liquid films in contact with solid surfaces have a large surface area to volume ratio; consequently, they can exhibit properties that significantly depart from those of their bulk (three-dimensional) analogues. For example, molecularly thin organic films on surfaces exhibit thickness-dependent diffusion coefficients and “terraced” flow that is a result of molecular ordering.1-3 Confinement effects may additionally affect boundary layer mobility, giving rise to nonclassical timedependent kinetics.4 Since these properties originate from the differences in the interaction energies between the liquid film/surface and the liquid/liquid systems, the physical descriptions of molecularly thin films on surfaces are often not scalable to their bulk counterparts. The dewetting of molecularly thin liquid films on surfaces is frequently observed and can be undesirable in technological applications. As an illustrative example, we consider the molecularly thin perfluoropolyether liquids that are employed as boundary lubricants on rigid magnetic recording disks for the reduction of wear between the intermittently contacting read-write slider (or “head”) and the disk surface. The chemical structures of some typical perfluoropolyether liquids used in this application are shown below.
The lubricant films are topically applied to the surface of the rigid disk. The film thicknesses employed are indeed molecularly thin, approximately 10-20 Å, which are (1) Zheng, X.; Rafailovich, M. H.; Sokolov, J.; Strzhemechny, Y.; Schwarz, S. A.; Sauer, B. B.; Rubenstein, M. Phys. Rev. Lett. 1997, 79, 241. (2) Cazabat, A. M.; Frayasse, N.; Heslot, F.; Carles, P. J. Phys. Chem. 1990, 94, 7581. (3) O’Connor, T. M.; Jhon, M. S.; Bauer, C. L.; Min, B. G.; Yoon, D. A.; Karis, T. E. Tribol. Lett. 1995, 1, 219. (4) Waltman, R. J.; Tyndall, G. W.; Pacansky, J.; Berry, R. J. Tribol. Lett. 1999, 7, 91.
comparable to their end-to-end distances when taken as ∼2Rg, where Rg is the radius of gyration.5 The polymer main chain is composed of a copolymer of the nonpolar perfluoromethylene oxide and perfluoroethylene oxide monomer units that are truncated by polar hydroxyl groups on both ends. Since the hydroxyl end groups preferentially interact with the polar sites of the underlying surface, exposing the nonpolar perfluoropolyether backbone to the air/film interface, an ordering of the first monolayer takes place perpendicular to the disk surface which leads to a minimum in the surface free energy as a function of surface coverage.6 When the monolayer film thickness is exceeded, the inability of the next liquid layer to spread on its own monolayer leads to autophobic dewetting.7-12 The origin for autophobic dewetting can therefore be attributed to differences in the interaction energies between the liquid/surface and the liquid/liquid systems. Analyses of the spreading profiles for Zdol films on surfaces indicate that the resulting oscillations in the polar surface energy as a function of film thickness induce oscillations in the resultant disjoining pressure that determine the stability of the Zdol films.13-15 This can be quantified by the negative derivative of the surface free energy with respect to film thickness, that is, the disjoining pressure Π(h) ) -dγ/dh, where γ is the surface free energy and h is the film thickness.13,16 When the total free energy of the surface decreases with increasing liquid film (5) Cotts, P. Macromolecules 1994, 27, 6487. (6) Tyndall, G. W.; Waltman, R. J. Proc. Mater. Res. Soc. 1998, 517, 408. (7) Hare, E. F.; Zisman, W. A. J. Phys. Chem. 1955, 59, 335. (8) Reiter, G.; Khanna, R. Langmuir 2000, 16, 6351. (9) Ferreira, P. G.; Ajdari, A.; Leibler, L. Macromolecules 1998, 31, 3994. (10) Derajaguin, D. V.; Voropayeva, T. N. J. Colloid Sci. 1964, 19, 113. (11) Derajaguin, D. V.; Churaev, N. V. J. Colloid Interface Sci. 1974, 49, 249. (12) Kim, H. I.; Mate, C. M.; Hannibal, K. A.; Perry, S. S. Phys. Rev. Lett. 1999, 82, 3496. (13) Tyndall, G. W.; Karis, T. E.; Jhon, M. S. Tribol. Trans. 1999, 42, 463. (14) Ma, X.; Gui, J.; Smoliar, L.; Grannen, K.; Marchon, B.; Jhon, M. S.; Bauer, C. L. J. Chem. Phys. 1999, 110, 3219. (15) Ma, X.; Gui, J.; Smoliar, L.; Grannen, K.; Marchon, B.; Bauer, C. L.; Jhon, M. S. Phys. Rev. E 1999, 59, 722. (16) Cazabat, A. M.; Fraysse, N.; Heslot, F.; Carles, P. J. Phys. Chem. 1990, 94, 7581.
10.1021/la0301700 CCC: $27.50 © 2004 American Chemical Society Published on Web 03/04/2004
Autophobic Dewetting of Lubricant Films
Langmuir, Vol. 20, No. 8, 2004 3167
Table 1. Characterization of the Z-Tetraol Lubricants Used in This Studya lubricant
Mn
Tetraol 1000 Tetraol 2200 #1 Tetraol 2200 #2 Tetraol 4000 Tetraol 1500 Tetraol 2210 Tetraol 3390 Tetraol 3820
1290 2210 2240 4200 1498 2212 3394 3820
a
Mw/Mn
1.29 1.34
C1/C2
% Tetraol
% Zdol
1.09 1.05 0.94 1.22 1.02 1.03 1.05 1.10
82.0 87.0 91.5 86.5 44.1 61.9 83.7 93.5
18.0 11.5 6.5 13.5 36.1 30.3 13.3 1.2
% -CF3
% -Cl
% bis adduct
8.0 2.0 0.6 0.5
6.9 5.0 2.1 4.6
1.5 2.0 4.9 0.8 0.3 0.3
Tetraol 1000, 2200, and 4000 are polydisperse, while Tetraol 1500, 2210, 3390, and 3820 are fractionated.
thickness, or equivalently when Π > 0, complete wetting of the surface by the liquid is thermodynamically favorable. When Π < 0, the surface is only partially wet by the liquid and droplets with a finite contact angle may be formed. The dewetting of molecularly thin perfluoropolyether Zdol films on rigid magnetic media has been recently shown to impact the dynamic spacing of the low-flying slider and the disk surface, causing spacing losses of as much as ∼4 nm due to slider interference with the liquid droplets on the disk surface.17 Since today’s hard disk drives operate at nominal flying heights of ∼10 nm, such spacing losses represent a significant fraction of the available clearance between the slider and the disk surface and will cause unwanted slider-disk interference and consequently a reduction in the tribological reliability of the head-disk interface. The critical dewetting thickness of Zdol films on disk surfaces has been found to depend on the molecular weight.17 Since the critical dewetting thickness for these films is equivalent to the monolayer thickness, it can be quantified via surface energy measurements (Π ) 0), by “titrating” the carbon surface with Zdol, or by (ellipsometric) imaging of the disk surface for lubricant droplets, all of which provide equivalent data.17 In this report, we investigate the film stability of Z-Tetraol as a function of molecular weight. The additional hydroxyl end groups in Z-Tetraol increase the possible number of inter- and intramolecular hydrogen bonds, which increases the level of adhesion to the underlying surface and reduces the radial creep of the lubricant film induced by the spinning disk.18 This could allow the use of a comparatively lower molecular weight (MW) perfluoropolyether lubricant to satisfy the requirements of both dynamic slider-disk clearance and adhesion to the disk surface in low-flying, high-rpm disk drives. However, the lower molecular weight could limit the maximum thickness that can be applied to the disk surface without risk of dewetting.
composition of the CNx surfaces has been previously quantified using X-ray photoelectron spectroscopy (XPS), Auger electron spectroscopy (AES), and Rutherford backscattering (RBS).19,20 The XPS measurements were made using a Phi Quantum 2000 ESCA system at a 75° takeoff angle. The Auger measurements were made using a PHI Auger 660 (Perkin-Elmer) operated at 3 kV electron gun voltage and 200 nA beam current. The CNx surface contained both oxygen and nitrogen. The AES results indicated a nitrogen level of 10.0 at. % (atomic %), while the XPS measurements provided a somewhat higher value of 13.8 at. %. Thus the nitrogen content in CNx was taken to be 10.0 < x < 13.8 at. %. The oxygen content was determined to be 7 at. % (XPS), and no hydrogen was detected in these films ( 0 for all Z-Tetraol film thicknesses, the van der Waals interactions between the perfluoropolyether backbone and CNx will favor the complete wetting of the surface by the Z-Tetraol lubricant. In contrast, the polar disjoining pressure (Πp) for Z-Tetraol 2200 on CNx changes sign from positive to negative at the film thickness corresponding to the minimum in γps (nominally 18.5 ( 1.5 Å). On the basis of the total disjoining pressure (Π ) Πp + Πd) presented in Figure 3, Z-Tetraol 2200 films of thickness less than 18.5 Å should be completely wetting on CNx surfaces. In contrast, film thicknesses greater than the critical thickness of 18.5 Å (where Π < 0) will be unstable and prone to undergo autophobic dewetting. The instability of Z-Tetraol films on CNx with thicknesses in excess of a monolayer was verified experimentally by imaging the CNx disks lubricated with a series of Z-Tetraol film thicknesses both below and above the monolayer film thickness. Representative images for
Figure 3. The disjoining pressure as a function of film thickness for Z-Tetraol 2200 #1 on CNx.
Figure 4. OSA images of Z-Tetraol 2200 #1 (actual Mn ) 2210) on CNx illustrating the onset of film instability and dewetting between 17.2 Å < h < 20.6 Å. The dark spots in the bottom image correspond to lubricant droplets. A subsequent immersion of the disk into lubricant solvent results in the removal of the droplets. Both images are taken after a head has flown once over the disk surfaces.
Z-Tetraol 2200 (#1) below (17 Å) and above (20 Å) the monolayer thickness are presented in Figure 4. When Z-Tetraol 2200 (#1) is applied to the disk surface at a film thickness of nominally 0, complete wetting of the surface by the liquid is thermodynamically stable. Conversely, when Π < 0, the surface is only partially wet by the surface and droplets with a finite contact angle can be formed. The critical dewetting thickness was shown to be a function of the lubricant MW, decreasing with decreasing MW. The results presented in Figure 5 further indicate that as the MW of the lubricant is decreased, the margin between some chosen nominal film thickness of the lubricant (say 15 Å for the purposes of the discussion) and the critical dewetting thickness decreases. Therefore, for a nominal lubricant film thickness that is only several angstroms below the critical dewetting thickness, the control of the film thickness uniformity on the disk becomes important. However, this could be further exacerbated in low-flying hard disk drives where the dynamics of the drive environment can additionally induce lubricant nonuniformity by head-disk interactions causing the formation of, for example, lubricant “moguls”.30 For example, Figure 6 shows the image of a disk surface that was coated with nominally 18.6 Å of Z-Tetraol 2200 (#1), which corresponds to the monolayer film thickness. The image of this disk surface clearly shows the occurrence of dewetting in local regions on the disk surface where small variations in the inherent film thickness put the total (30) Pit, R.; Marchon, B.; Meeks, S. W.; Velidandea, V. Tribol. Lett. 2001, 10, 133.
Figure 7. A log-log plot of the monolayer thickness for Z-Tetraol (9) and Zol (3) as a function of molecular weight (Mn) on CNx. The monolayer film thickness was determined from the critical dewetting thickness as observed by OSA images for each molecular weight. The least-squares fits of the data points for Z-Tetraol (slope ) 0.57) and Zdol (slope ) 1.01) are represented by the solid and dotted lines, respectively.
film thickness above that of its monolayer. As shown in Figure 6, the regions of the disk surface having locally slightly thicker lubricant could become susceptible to dewetting when a slider is flown over the disk surface. A trace of the lubricant profile accompanying the rightmost dewetting region shown in Figure 6 indicates that the lubricant is thicker in this region by only ∼0.8 Å, which is enough to cause the monolayer thickness to be exceeded and therefore dewetting to occur. The impact of lubricant droplets on slider-disk clearance has been previously analyzed.17 Its presence can severely decrease the dynamic slider-disk clearance. As a final commentary on the results of these experiments, we compare the MW dependence of the critical dewetting thicknesses, hc, for Z-Tetraol and Zdol on CNx. Figure 7 shows that hc scales linearly (slope ) 1) with the Zdol MW and as (MW)0.6 (slope ) 0.6) for Z-Tetraol. Since both lubricants share an identical main chain composed of a copolymer of perfluoromethylene oxide and perfluoroethylene oxide units, the difference in the slopes could be attributed to changes in the adsorbed film structure due to the different end groups. However, other possible influencing factors must first be considered, such as the compositional variations in bulk Z-Tetraol which could affect the adsorbed film structure and consequently the critical dewetting thickness. We note in Table 1 that the bulk Z-Tetraol lubricant samples used in these studies, both fractionated and polydisperse, are inherently mix-
Autophobic Dewetting of Lubricant Films
Figure 8. The critical dewetting thickness, hc, for mixtures of Z-Tetraol 2200 #1 and Zdol 4000 on CNx as a function of the weight fraction of Z-Tetraol in Zdol solutions of HFE-7100 (C5H3F9O) and Vertrel-XF (C5H2F10). The top of the vertical bar accompanying each symbol represents the thickness at which dewetting is observed, while the bottom of the vertical bar represents the thickness at which no dewetting is observed.
tures that are populated with significant fractions of end groups other than the tetraol adduct. These include the hydroxyl end groups on unreacted Zdol, a “bis-adduct” end group (-OCF2CH2-O-CH2-CH(OH)-CH2OCH2CH(OH)-CH2-OH), -CF3, and -Cl. Similar structural analyses on the Zdols used in these studies indicate >99% of hydroxyl end groups and therefore their chemical purity are not in question. For Z-Tetraol, the major “contaminant” appears to be unreacted Zdol and its presence, often in significant proportion to the tetraol adduct (Table 1), could cause the actual value of hc for Z-Tetraol to shift or change from an otherwise ideal tetraol adduct. Therefore, we conducted studies on the dependence of the critical dewetting thickness hc for mixtures of Z-Tetraol and Zdol on the CNx disk as a function of solvent used to apply the lubricant to the disk. The goal of these studies was to utilize the observed dependence of hc on lubricant MW to provide a methodology by which Zdol and Z-Tetraol on the CNx surface could be distinguished and therefore the impact of the Zdol contaminant on the Z-Tetraol hc be assessed. Since the hc values for Z-Tetraol 2200 #1 (actual Mn ) 2210) and Zdol 4000 (actual Mn ) 3560) on CNx are well-separated at nominally 19 and 28 Å, respectively, mixtures of the two lubricants are expected to exhibit a critical dewetting thickness between 19 Å < hc < 28 Å that could scale simply as a function of their concentration. However, the lubricant mixtures are deposited from solvent which could influence the adsorbed film structure depending upon the relative affinity of the lubricant for the solvent, surface, and/or other lubricant molecules. For a “good” solvent for both Z-Tetraol and Zdol, lubricantsolvent interactions may be similar enough such that there is no strong preference for adsorbing one or the other lubricant onto the CNx surface in which case hc could possibly scale as a function of the concentration of Z-Tetraol in Zdol. A “poorer” solvent for one or the other of the lubricants could instead allow its preferential adsorption onto the CNx surface due to the weaker interaction with the solvent. We note that a better or poorer solvent is simply a technical description of a solvent that tends to either increase or decrease the coil dimensions of the polymer in solution. The dependence of the critical dewetting thickness on mixtures of Z-Tetraol 2200 #1 and Zdol 4000 is presented in Figure 8. An approximately linear dependence of hc on solution concentration is observed for lubricant films
Langmuir, Vol. 20, No. 8, 2004 3171
deposited from Vertrel-XF (C5H2F10), while a decidedly nonlinear dependence of hc on solution concentration is obtained for lubricant films deposited from HFE-7100 (C5H3F9O). A more detailed study of the interactions between perfluoropolyether lubricants and solvents is forthcoming and remains beyond the scope of the current studies.31 The salient feature in Figure 8 is that the solvent HFE-7100 causes the preferential adsorption onto CNx of the tetraol adduct compared to the Zdol adduct. This is indicated by the similar hc for Z-Tetraol concentrations in Zdol of greater than ∼40%. The hc of the mixture begins to increase (toward Zdol) only when the concentration of Z-Tetraol in Zdol 4000 becomes less than ∼40%. This is to be contrasted to changes in hc observed for Z-Tetraol/ Zdol mixtures deposited from Vertrel-XF which instead show a roughly linear correlation as a function of concentration. Therefore, we are able to conclude that the hc values obtained for the Z-Tetraol lubricants shown in Figures 4-7 are more likely to be representative of (but not necessarily exclusive of) the tetraol adduct than of the Zdol adduct since the solvent HFE-7100 is used in all cases to generate the data. The conclusions presented above for the preferential adsorption of the tetraol adduct compared to the Zdol adduct on CNx are attributed to the solvent power of HFE7100. The solubility parameter is defined as the square root of the cohesive energy density:32
δi ) (∆Eiv/Vi)1/2 (cal/cm3)1/2
(7)
where ∆Eiv is the energy of vaporization and Vi is the molar volume of component i. The solubility parameter δ therefore describes the attractive strength between molecules and is used as a criterion for the solvent power. As the difference in δ between the polymer and solvent increases, the solvent becomes poorer. Equation 7 is related to the heat of mixing per unit volume as33
∆Hm/V ) (δ1 - δ2)2φ1φ2
(8)
where φi is the volume fraction of component i. Since the free energy of mixing is related to the enthalpy of mixing via ∆Gm ) ∆Hm - T∆Sm, the difference in the polymersolvent solubility parameter (δ1 - δ2)2 must be small for the components to be miscible. The solubility parameters for Zdol, Z-Tetraol, HFE-7100, and Vertrel-XF are summarized in Table 3 and indicate a larger (δ1 - δ2)2 for Z-Tetraol/solvent compared to Zdol/solvent and for the Z-Tetraol/HFE-7100 pair compared to the Z-Tetraol/ Vertrel-XF pair. Thus, there is preferential adsorption of Z-Tetraol onto the CNx surface compared to Z-Tetraol/ Vertrel-XF and Zdol/solvent. Assuming that the dependence of the Z-Tetraol critical dewetting thickness hc on MW presented in Figure 7 is primarily attributable to the tetraol end group, as suggested by the solvent deposition studies (Figure 8), we believe that the adsorbed film structure of Z-Tetraol on CNx could be substantially different from that of Zdol on CNx. The (MW)0.6 dependence of hc on the Z-Tetraol MW is consistent with the coil dimensions of the polymeric molecules near the surface being somewhat distorted from an unperturbed random coil (i.e., a radius of gyration or (MW)0.5 for the unperturbed coil). This picture is consid(31) Waltman, R. J.; Tyndall, G. W.; Wang, G. J.; Deng, H. Tribol. Lett. 2004, 16, 215. (32) Grulke, E. A. Polymer Handbook, 3rd ed.; Brandrup, J., Immergut, E. H., Eds.; John Wiley & Sons: New York, 1989; Vol. 7, p 519. (33) Fedors, R. F. Polym. Sci. Eng. 1974, 14, 147.
3172
Langmuir, Vol. 20, No. 8, 2004
Waltman
Table 3. Computed Solubility Parametersa for the Structures Shown Below chemical
chemical structure
δ (cal/cm3)1/2
HFE-7100 Vertrel-XF Zdol 4000 Z-Tetraol 2000
CF3-[CF2]3-O-CH3 CF3-CHF-CHF-CF2-CF3 HO-CH2-CF2-[OCF2]20-[OCF2CF2]20-O-CF2-CH2-OH HO-CH2-CH(OH)-CH2-CF2-[OCF2]10-[OCF2CF2]10-O-CF2-CH2-CH(OH)-CH2-OH
6 6.2 8.1 9.3
a
Reference 33.
erably different from the more loosely packed Zdol adsorbed film structure on CNx whose linear dependence on MW suggests a relatively large loop-to-train ratio.19 Concluding Remarks We have demonstrated that a maximum stable film thickness exists for Z-Tetraol on the CNx carbon surface. We have shown that this maximum stable thickness is molecular weight dependent and scales as (MW)0.6. ZTetraol applied in excess of the monolayer thickness can
be induced to form droplets or dewet simply by flying a slider over the disk surface. Acknowledgment. The author gratefully acknowledges J. Burns of the Materials Lab at this company for providing the NMR analyses of the polydisperse Z-Tetraol samples and G. W. Tyndall of the IBM Almaden Research Center for many fruitful discussions on surface energies. LA0301700
