Friction of Mixed and Single-Component Aromatic Monolayers in

Publication Date (Web): January 13, 2006 ... Contact Mechanics of Nanometer-Scale Molecular Contacts: Correlation between Adhesion, Friction, and Hydr...
0 downloads 0 Views 292KB Size
J. Phys. Chem. B 2006, 110, 2209-2218

2209

Friction of Mixed and Single-Component Aromatic Monolayers in Contacts of Different Adhesive Strength M. Ruths* Department of Physical Chemistry, Åbo Akademi UniVersity, Porthansgatan 3-5, FI-20500 Åbo, Finland ReceiVed: August 24, 2004; In Final Form: October 17, 2005

Friction force microscopy has been used to study single-component and mixed self-assembled monolayers of aminothiophenol and thiophenol on gold. The friction forces and transition pressures of mixed monolayers were intermediate to the ones of single-component monolayers, and varied systematically with composition. The strength of the adhesion was altered by working in dry N2 gas or in ethanol. In all systems studied, low adhesion (in ethanol) resulted in a linear dependence of the friction on load already at low loads, whereas high adhesion (in dry N2) gave an apparent area-dependence. However, for a given monolayer composition, similar transition pressures were observed in dry N2 and in ethanol, suggesting that the overall monolayer structure was not strongly altered by the presence of ethanol. Similar observations were made for very closepacked monolayers of octadecanethiol.

Introduction Compared to the wealth of information available on the formation1-4 and friction5-18 of self-assembled monolayers of alkanethiols and alkylsilanes, relatively few studies have concentrated on aromatic monolayers. Aromatic molecules are less flexible than linear and branched alkanes, and they have more complex intermolecular interactions in solution and at interfaces. This makes their friction-modifying properties interesting from a fundamental physical-chemical perspective, since one current focus of the research on boundary lubrication is on correlating the structure of confined monolayers and thin films with their friction response. Aromatic systems are also important from a practical point of view. Certain sulfurcontaining aromatic compounds such as thiophenes and thiols occur naturally as impurities in diesel fuel19 and in mineraloil-based lubricants, where some of them are known to act as antioxidants and antiwear agents.20-24 Similar molecules are also used as friction modifiers in aluminum-on-steel sliding.21 It is well-known that the removal of sulfur-containing compounds and aromatics significantly reduces the lubricity of fuels,19,25 and aromatic, heteroaromatic, and polyaromatic molecules with polar substituents are added to refined fuels and lubricant formulations to improve their performance.19,20,25,26 Despite their obvious importance, the lubricating function at the molecular level of such additives and of the sulfur-containing aromatic molecules naturally present in the oil is not well understood. Our current knowledge of the formation and structure of selfassembled aromatic monolayers comes from molecular optics and electrochemical applications, and a small number of computer simulations.27-52 Aromatic monolayer surfaces that expose reactive groups are used, for example, to attach conducting polymer layers to electrodes,34,45 or for linking nanoparticles to a surface.46 The orientation and packing density of aromatic molecules in a monolayer are strongly affected by lateral interactions, which can be altered by introducing polar groups * Present address: Department of Chemistry, University of Massachusetts Lowell, Lowell, MA 01854. E-mail: [email protected]. Phone: (978) 934 3692. Fax: (978) 934 3013.

or increasing the rotational freedom through flexible linkers between the anchoring group and the aromatic part. Increased lateral interactions typically give a more upright orientation and higher packing density,36,41,45,50 which is also reflected in the friction response of the monolayers.53,54 The substrate and the adsorption conditions also affect the packing. This investigation is concerned with how friction can be controlled by changing the composition and surrounding conditions of model systems of simple aromatic thiols. The monolayers used in this work were formed on template-stripped gold by adsorption from ethanol solution. The equilibrium structure formed by pure thiophenol (TP) adsorbed on gold (111) from ethanol is a loose-packed monolayer31 where no structure can be seen when imaged with an atomic force microscope (AFM).33,36 4-Aminothiophenol (ATP) forms a more closepacked monolayer with a more upright molecular orientation than TP.29,32,42,46 Mixed monolayers of ATP and TP can be formed on gold by a 20 min adsorption from ethanol solution,34 and it has been suggested that after this short adsorption time, the surface composition is equal to the mole fraction of the two thiols in solution.34 The friction of these monolayer systems has been studied in a single-asperity contact using friction force microscopy, a technique55-57 based on atomic force microscopy (AFM). Several experimental studies on solid surfaces and monolayer systems have shown that the friction force F in an adhesive, single-asperity contact increases in proportion to the contact area at low loads (adhesion-controlled friction),8,17,58-61 and there is a finite friction force even at zero applied external load (L ) 0). In contrast, experiments in systems with low adhesion show a linear increase in friction force with load (load-controlled friction) already at low loads, and F ) 0 at L ) 0.6,13,15,53,54,59,62 Furthermore, in systems that are adhesion-controlled at low loads, the friction force often becomes linear at high loads. These experimental observations are generally expressed as59

F ) ScA + µL

(1)

where Sc is the “critical shear stress” (a constant) and µ is the

10.1021/jp0461706 CCC: $33.50 © 2006 American Chemical Society Published on Web 01/13/2006

2210 J. Phys. Chem. B, Vol. 110, No. 5, 2006 “friction coefficient”. Alternatively, this empirical equation can be written as “shear strength”63 by dividing it by the contact area to obtain S ) Sc + µp, where p is the pressure L/A. Several models have been proposed to explain this type of friction response. One is the “cobblestone model”, which proposes that work has to be done normal to the surfaces to separate them and enable them to slide laterally past each other.59,62 This work is assumed to consist of two separate components that in this simple model are additive: (i) the work against adhesion and (ii) the work against load (normal force). The work against load is independent of the contact area. In this model, only Sc depends on the adhesion (the interfacial energy, γ), which implies that the area-dependence of the friction force would be strongly reduced and even removed if the adhesion were reduced. As a result, only the term µL in eq 1 would remain, which has been observed experimentally in systems where the surfaces repel each other,59,62 and also in systems with low adhesion, γ ≈ 1-4 mJ/m2.53,54 Other models assume that the friction is always proportional to the contact area.64,65 Since the area of an elastically deforming single-asperity contact typically does not increase linearly with load,66-71 observations of a linear dependence of the friction force on load may be ascribed to a “pressure-dependent shear stress”. While a pressure dependence of the shear stress certainly is possible, it is unlikely that this is the full explanation for the linear load-dependence seen in many systems (lubricated and unlubricated, see, for example, refs 13, 54, 59, 72, and 73) over large load regimes. It would require the pressure dependence of the shear stress in all these very different systems to exactly compensate for the incremental increase in contact area with increasing load. To further complicate the interpretation, such systems can at only slightly altered conditions also show a friction force proportional to the contact area, i.e., a constant shear stress (with no pressure dependence). The effects of contact area and adhesion on friction,73,74 and the occurrence of dilatency during sliding,75,76 are also being investigated in computer simulations. Some of the simple models above have recently been discussed in connection with a molecular dynamics simulation that suggests that in many cases, a dependence on the contact area is not the best way to describe friction.73 The aim of this work is to evaluate the effects of molecular structure and monolayer composition on the friction of different aromatic monolayers in contacts of different adhesive strength. Since the strength of the adhesion is expected to affect the magnitude and functional form of the friction force as a function of load, some experiments were done in dry N2, where the adhesion is high, and some were done in ethanol, where the adhesion is low. Although small solvent molecules can interact with the outermost regions of adsorbed monolayers,77 and penetrate into very loose-packed ones, they are less likely to affect the monolayer structure in the case of close-packed monolayers. Comparisons of our measured adhesion forces in ethanol with calculated van der Waals forces suggest that a full layer of ethanol molecules is not present between the monolayer and the sliding tip, and the experiments in ethanol therefore also show the effects of monolayer structure on friction. Materials and Methods Self-Assembled Monolayers. 4-Aminothiophenol (ATP, Fluka, 96.2%) and thiophenol (TP, Aldrich, 99.8%) were used as received. Details on the sample preparation are given in ref 53. Some of the single-component and mixed aromatic monolayers used for friction measurements were formed on templatestripped, polystyrene-supported gold53,78 by adsorption for 20

Ruths min from 2 mM solution in ethanol (99.5%, Primalco, Finland). It has been proposed that this procedure results in monolayers with the same composition as the xATP (mole fraction of ATP of the total thiol, ATP and TP) in solution.34 Mixed monolayers with longer adsorption times were also prepared for contact angle measurements, and the friction of one mixed monolayer system with xATP ) 0.42 was measured after an adsorption time of 23 h. Single-component monolayers of ATP, TP, and octadecanethiol (C18, Fluka, 99.4%, used as received) were also formed on template-stripped gold and on gold-covered AFM tips by adsorption from 2 mM solution for 47 h. The rms roughness of the template-stripped gold was 0.2-0.3 nm, measured over an area of 1 µm2. At adsorption equilibrium, TP has a reported molecular area of 0.38 nm2.36,44 Compared to the area of a vertically oriented benzene molecule (based on covalent and van der Waals radii) of 0.21 nm2,36,41,47 the TP molecules are thus loose-packed and also strongly tilted. Different experiments and simulations indicate a tilt from the surface normal of 42°;41 49° (monolayer thickness 6 Å);49 76°;39 and 80° on Au(111),43,44 and different tilt angles on other gold crystal planes.43,44 Recent work suggests that the tilt angles would depend on the coverage.51,52 ATP has a reported molecular area of 0.17 nm2,42 0.19 nm2,29,46 0.22 nm2,32 or 0.23 nm2,45 and an orientation near-parallel to the surface normal.34,45 One report shows an AFM image of a regular structure.29 Adsorption of TP or ATP from aqueous solution31,40,47 or on other substrates gives orientations and molecular areas27,28,44,45 different from the ones above. Experiments using a quartz crystal microbalance have shown that the adsorption of ATP on polycrystalline gold from 1 mM ethanol solution takes 70 min to reach equilibrium.46 An adsorption time of more than 1 h is also indicated in ref 42. The adsorption of ATP from a more dilute solution, 0.1 mM, takes about 2 h to reach equilibrium.32 The desorption of ATP from a well-formed monolayer immersed in ethanol has been shown to be very slow.38 Contact Angle Measurements. The advancing and receding contact angles of water were measured from recorded images of sessile drops. The syringe needle was kept in the drop so that the drop volume could be increased and subsequently decreased. The images were obtained with a CCD camera mounted on an Elma/Kru¨ss G1 contact angle meter with a video monitoring system. The drops were symmetrical and typically had a base diameter of 1-2 mm. Each of the points used to construct the plot in Figure 1 is the average of 7-10 measurements, with a resulting error in θadv - θrec of (1°. Friction Force Microscopy. The friction force F was measured with an atomic force microscope (Multimode AFM, NanoScope III controller, Digital Instruments). Friction data were obtained with unfunctionalized Si tips (exposing a native silicon oxide surface), and with monolayer-functionalized goldcovered tips (MikroMasch, Estonia), as indicated in the figure legends. The normal and torsional spring constants, kN and kl, of the rectangular cantilevers were determined by measuring their dimensions with scanning electron microscopy. The calculations of the spring constants79,80 and the calibration of the signal from the AFM are described in detail in refs 53 and 54. The spring constants and the radii of the tips, R, which were obtained by reverse imaging of a calibration sample (model TGT01, MikroMasch, Estonia), are given in the figure legends. The radii did not change after the friction measurements, indicating that there was no significant tip wear. No damage to the monolayers could be detected within the investigated range of loads.

Friction Forces and Transition Pressures of Monolayers Measurements were done either in dry N2 gas (Woikoski, Finland, purity 99.7%) in a chamber enclosing the AFM (the relative humidity was typically