Article pubs.acs.org/Langmuir
Nanotribological Properties of Hexadecanethiol Self-Assembled Monolayers on Au(111): Structure, Temperature, and Velocity Torben Marx, Ximeng Shen, Dirk Dietzel, and André Schirmeisen* Institute for Applied Physics, Justus-Liebig-Universität, 35392 Gießen, Germany ABSTRACT: Self-assembled monolayers (SAM) are promising building blocks for the optimization of a large variety of systems both on the nano- and on the microscale. Among other applications, SAM are often used as protective coating or friction modifiers. In this work, we have used hexadecanethiol SAM on Au(111) as a model system and studied the different mechanisms of energy dissipation during temperature and velocity dependent friction force microscopy (FFM). In a number of cases, the SAM remained stable during atomic force microscopy experiments and friction−velocity isotherms related dissipation to an activation energy. In other cases, friction experiments lead to an irreversible deterioration of the SAM. This can rather be associated with the general SAM structure that was analyzed by scanning tunneling microscopy and showed a large variety of potential breakdown points like, for example, grain boundaries, step edges, or substrate-related holes in the SAM.
■
INTRODUCTION The term self-assembly refers to the spontaneous formation of organized objects out of smaller units, which are typically either atoms or whole molecules. In many cases, self-assembly leads to the formation of monolayer coatings on surfaces. Using such coatings of self-assembled monolayers (SAMs)1−3 is nowadays a widely applied strategy to tune mechanical and chemical properties of surfaces. SAM coatings can be utilized for a wide range of applications2 including, for example, electrode modification in electrochemistry,4,5 lithography,6 cell culture applications,7 and research of biointerfaces.8 A particularly promising area of application is nanomechanics, where a more widespread use of nano- and microelectromechanical systems (NEMS and MEMS) is often impeded by problems related to friction and wear. Here, selfassembled monolayers could provide a means to protect these surfaces. At the same time, other essential properties of the samples, for example, electrical conductivity, chemical reactivity, thermal insulation or friction coefficient, can be modified by SAMs. In many cases, the behavior is significantly influenced by the choice of end groups, which, for example, was shown for tribological properties9 using friction force microscopy (FFM).10 Apart from the end groups, also the chain length of the molecules forming SAMs is one of the parameters determining mechanical and electrical properties, and recently it was shown how thermal transport parameters diminish with increasing chain length.11 However, in most applications, the mechanical stability of the SAM is of paramount importance. Therefore, it is crucial to understand and identify the different dissipation channels of frictional energy. In this work we used hexadecanethiol HS© 2017 American Chemical Society
[CH2]15-CH3 on Au(111) as a model system and friction force microscopy is applied to the self-assembled monolayers (Figure 1) in order to identify the characteristic activation energy EA of our system. More specifically, this is done by combined temperature and velocity-dependent friction measurements on the monolayers and data analysis based on the frequency− temperature superposition model. This approach was demon-
Figure 1. Schematic representation of our sample system in interaction with a moving AFM tip. The Au(111) substrate is covered by a self-assembled monolayer on top of which the AFM tip is sliding. The torsion of the AFM tip during sliding is proportional to the friction force. Received: April 4, 2017 Revised: May 24, 2017 Published: June 12, 2017 6005
DOI: 10.1021/acs.langmuir.7b01131 Langmuir 2017, 33, 6005−6010
Article
Langmuir strated by Sills et al.,12−14 who used the transformation of friction−velocity isotherms into a single master curve that revealed activation energies EA of polymer systems. In our case, application of the frequency−temperature superposition model yields a characteristic energy barrier, which opens up a new route to understand frictional dissipation in SAM structures. Apart from intrinsic properties of the SAM also dissipation channels due to defects or grain boundaries can become important. Therefore, also the structure of the SAM has been analyzed using high-resolution scanning tunneling microscopy (STM).15 In principle, Alkanethiol SAM on Au(111) are already well studied systems as long as short chains are concerned. Long chains, as used in this work, are much less explored with so far only one publication related to them.16 This lack of analysis is probably mostly due to low electronic transport properties of longer chains, which makes structural analysis by STM difficult. Nonetheless, we have chosen rather long chains because these can be expected to be more susceptible to mechanical deterioration. Interestingly, apart from features relevant with respect to energy dissipation mechanisms, our STM analysis also showed three characteristic SAM phases that had not yet been observed for this chain length.
■
Figure 2. Characterization of the Au substrate and SAM using atomic force microscopy. (a) Terrace of the flame-annealed Au(111) substrate. The image clearly shows the typical herringbone reconstruction. (b) Atomic force microscopy image of a SAM on Au(111). The most prominent features are the substrate related steps in the lower part of the image and the frequent typical gold vacancy islands. (c) High-resolution friction force image of a SAM on Au(111). The atomic level stick slip reveals the anticipated hexagonal symmetry of the SAM.
EXPERIMENTAL SECTION
Sample Preparation and Topography Characterization. To prepare the SAM we have first used sputter evaporation to prepare a thin layer of gold (thickness of approximately 50 μm) onto freshly cleaved MICA that was heated to 400 °C before evaporation to remove residual moisture. After the evaporation, the Au film was annealed at a temperature of about 650−750 °C to create a wellordered surface. To ensure a sufficient quality of the Au(111) substrate for the self-assembly process it can be analyzed by AFM and the lateral force signal revealed large terraces and the characteristic herringbone reconstruction17,18 of the Au(111)-plane (Figure 2a), where even atomic level stick slip resolution can be achieved (not shown). After the annealing process, the Au surface was put into alkanethiol solution for several hours, during which time the SAM was formed. As a last step, the sample was dried under a flow of hydrogen gas and directly transferred into the UHV System (Omicron VT-AFM/STM) for further analysis. Here, AFM imaging directly reveals the SAM formation (Figure 2). More specifically, Figure 2b shows a larger area of the sample covered by SAM with characteristic step edges related to terraces of the Au-substrate in the lower part of the image and a tear in the SAM in the left part of the image. Additionally most parts of the SAM appear to be covered by darker spots. The height difference of these spots equals the height of a Au monolayer. Therefore, these spots can be identified as typical gold vacancy islands, which appear as pits within the continuous SAM.19 Finally, Figure 2c shows a highresolution lateral force image of the SAM. A regular stick slip pattern with hexagonal symmetry is detected as the AFM tip explores the potential energy surface, which is primarily related to the end groups of the hexadecanethiol. Additionally, we have used scanning tunneling microscopy with currents of I ∼ 2−4 pA and voltages of U ∼ 2−3 V to characterize the SAM with even higher resolution. An overview over three typical sample areas is shown in Figure 3, which nicely illustrates that the SAMs usually consist of adjacent domains of limited size. In Figure 3a, we can identify six different domains (I−VI) with homogeneously orientated molecules, while the two other images (b,c) reveal a similar structure. It is well-known that alkanethiol SAMs on Au(111) form a (√3 × √3)R30° structure (α-phase) from which further substructures can be formed as c(4 × 2) superlattices (β−ζ structures).20 Several of these structures can be observed by zooming into areas from Figure 3. First, a magnification of domain III reveals the β-phase for the hexadecanethiol. Additionally, two more phases have been found
experimentally at other sample positions, namely the α-phase and the γ-phase as shown in Figure 3, the existence of all of which had so far not been confirmed for long chains like hexadecanethiol. Nonetheless, another aspect of the images shown in Figure 3 might be of even higher importance with respect to the velocity dependent friction measurements. The idea of preparing a Au(111) substrate with subsequent deposition of a self-assembled monolayer suggest a very flat and highly ordered sample. However, this is only the case for length scales of a few nm. On larger length scales, we can observe a variety of features such as grain boundaries, step edges, gold vacancy islands, or simply areas where SAM coverage seems to be incomplete or molecules remained flat on the substrate. The only way to improve the film quality might be using gold single crystals instead of flame annealed substrates. But even in this case, the size of terraces will still be limited and related defects can occur. In any case, one has to keep in mind that the defect rich structure of the SAM might impact friction force microscopy analysis, when the AFM tip exerts mechanical force to the SAMs. Analysis of Energy Dissipation by Combined Velocity and Temperature-Dependent Measurements. Typically, the thermally activated Prandtl Thomlison model as introduced by Gnecco et al.21 yields excellent results in friction force microscopy as long as hard surfaces with periodic surface potentials are considered. In principle, the stick slip image Figure 2c might suggest that this model might also be applicable to this system. But it is well-known that friction of SAMs is also governed by their viscoelastic properties, which are considered in the molecular spring model for compliant SAMs.22 In this case, the influence of sliding velocity becomes more complex and energy disspation peaks can arise, if the excitation rate is well adjusted to the relaxation time τ of the compliant system.23 For polymer systems, it was found that this relaxation time τ is reflected in the frequency response of dynamic mechanical analysis. Here, a maximum energy loss occurs if the inverse excitation frequency matches τ. In AFM experiments, the sliding velocity can be considered as equivalent to the excitation frequency23 and friction depends sensitively on the adjustment of v with respect to τ. 6006
DOI: 10.1021/acs.langmuir.7b01131 Langmuir 2017, 33, 6005−6010
Article
Langmuir
of the increasing and decreasing branch allows to assess, whether this sequence was recorded without lasting damage to the SAM. Indeed, this turned out to be a frequent effect of the measurements. Figure 4a shows an example of velocity-dependent friction where the increasing and decreasing branch do not match.
Figure 3. STM images of typical hexadecanethiol SAM structures on Au(111). (a) The topmost image shows an overview of the surface where different domains (I−VI) can be identified. A magnification of domain III reveals the β-phase, whereas magnifications of two further regions (b,c) additionally reveal the γ- and α-phase as potential SAM structures.
Figure 4. Normalized friction forces extracted from two sequences of lateral force images. (a) Typical example of friction changes induced by subsequent imaging (T = 100 K). The labels at each data point indicate the cyclic sequence of the images. Ramping the velocity up and down induces a gap between the two branches, resulting in friction values that are up to 40% higher for the second branch. (b) Example of subsequent imaging, where no friction changes are induced (T = 300 K). The same velocity ramping procedure was used as in (a).
In this context also the temperature plays an important role, since the relaxation time is typically temperature dependent with the time constant τ(T) = τ0eEA/kT and EA is the activation energy of the system. Therefore, measurements at different temperatures yield different peak velocities, or more generally, the velocity isotherms are shifted but can be transferred into a unique master curve by using a temperature dependent shift factor aT. This procedure is called frequency− temperature superposition24 and finally the activation energy EA can be extracted from an arrhenius plot of the shift factors EA.12−14,25 In our case, the SAMs needed to be cooled down for the time constants to match the AFM excitation. It is helpful if all friction isotherms show the dissipation peak but it is not necessary because a master curve can also be built if this is not the case.25 Consequently, to apply the frequency−temperature superposition technique to the SAM system, we have measured friction at a very low constant normal forces of about FN = 1 nN as a function of sliding velocity v for nine temperatures between 30 K and 300 K. At each temperature, several sequences of friction force images were automatically recorded by the AFM at different velocities in a maximum range from v = 150 nm/s to v = 15 μm/s and scan sizes of 30 nm × 30 nm. Each sequence started at low velocities and the velocity was first gradually increased before it was decreased again, once the maximum sliding velocity was reached. Thereby, two images were recorded at the same velocity for each sequence and comparison
We tried to improve the data acquisition process by, for example, using delay times for SAM rejuvenation between measurements at different velocities. However, such strategies had very little impact on the result, which suggest that indeed often irreversible changes have been induced to the SAM possibly originating from the different defects discussed previously. Therefore, doing a check by ramping the sliding velocity up and down is of crucial importance because otherwise unreliable data curves might enter the superposition process. Nonetheless, in many cases friction turned out to be well reproducible with matching friction levels for ramping the speed up and down, especially if the velocity range and thus the number of points is slightly reduced. This means that these data curves are not affected by major defects induced in the SAM. One example is shown in Figure 4b, where ramping v up and down yields almost identical results. Together with equivalent curves measured at eight further temperatures, such data can be used as foundation for the previously described frequency−temperature superposition. Please note that friction measurements do not allow us to distinguish the different SAM phases. Therefore, we have to assume that different phases have 6007
DOI: 10.1021/acs.langmuir.7b01131 Langmuir 2017, 33, 6005−6010
Article
Langmuir been introduced to the master curve, which thus effectively averages across the different phases.
expected to increase monotonously with decreasing temperature. Instead, this peak suggests the occurrence of a secondary effect, namely a resonance excitation of a SAM inherent dissipation channel. Therefore, the frequency−temperature superposition method was applied and the resulting master curve is shown in Figure 6a. Additionally, the corresponding shift factors aT are shown in Figure 6b as a function of the inverse temperature, while the vertical shift, that is, ΔFfriction(T) is shown in Figure 6c. Both curves of Figure 6a,b document that the approach of frequency−temperature superposition can well be applied to the SAM system. A universal mastercurve is found and the required shift factors show a good alignment in the Arrhenius type plot from which the slope of a linear fit yields an activation energy of EA = 58 ± 2 meV. Please note that the shift factors for the two lowest temperatures have not been included into the fit of Figure 6b. The reason for that is not related to the sample system but lies instead in the instruments increasing error for surface temperatures below 60 K. Also the vertical shift factors ΔFfriction of Figure 6c show a systematic behavior. For low temperatures, no vertical shift is needed, while friction has to be shifted up for higher temperatures. Interestingly, this is a behavior that is very similar to results by Sills et al. obtained on polystyrene.14 Here, the strength of the polymers α-peak was assumed to be “diluted” by lower energy dissipation associated with phenyl side chains at lower temperatures. In our case, a similar SAM inherent process might occur. In a more general picture, it seems plausible that higher temperatures should reduce friction effects unrelated to the anticipated bond breaking of methylene units. However, at this point no final assessment of the origin of the behavior of ΔFfriction can be given. Additionally, we have to keep in mind that the shift factors also compensate for slight variations of the surface found at the different sample positions
■
RESULTS AND DISCUSSION The following section will now describe the application and results of the frequency−temperature superposition. To see if this approach is indeed justified, we can first take a look at the general temperature dependence of friction, as it is shown in Figure 5. Here we find that friction is relatively stable in the
Figure 5. Temperature dependence of sliding friction for three different velocities measured on a self-assembled monolayer of hexadecanethiol. The gray line serves as a guide to the eye and points out the general trend of temperature dependent friction.
temperature range of 150 K < T < 300 K, before it increases by a factor of approximately 2 and reaches a maximum at 100 K, similar to measurements by Fujita et al., who observed a friction peak on a dodecanethiol SAM.26 Such a behavior cannot be explained by the typical temperature dependence of thermally activated friction27 alone. In this case, friction would be
Figure 6. Frequency−temperature superposition of the friction data recorded for the hexadecanethiol SAM on Au(111). (a) Superposition on data recorded for nine different temperatures. The measurement at 90 K has been used as starting point without shift. (b) lateral shift factor aT plotted versus the inverse temperature. The first seven data points have been used for the linear fit procedure resulting in the dashed black line. (c) Force shifts ΔFfriction applied to the different curves in (a). The red line illustrates the general trend. 6008
DOI: 10.1021/acs.langmuir.7b01131 Langmuir 2017, 33, 6005−6010
Article
Langmuir
(2) Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Self-Assembled Monolayers of Thiolates on Metals as a Form of Nanotechnology. Chem. Rev. 2005, 105, 1103−1170. (3) Vericat, C.; Vela, M. E.; Benitez, G.; Carro, P.; Salvarezza, R. C. Self-assembled monolayers of thiols and dithiols on gold: new challenges for a well-known system. Chem. Soc. Rev. 2010, 39, 1805. (4) Janek, R. P.; Fawcett, W. R.; Ulman, A. Impedance spectroscopy of self-assembled monolayers on Au (111): sodium ferrocyanide charge transfer at modified electrodes. Langmuir 1998, 14, 3011− 3018. (5) Felgenhauer, T.; Yan, C.; Geyer, W.; Rong, H.-T.; Goelzhaeuser, A.; Buck, M. Electrode modification by electron-induced patterning of aromatic self-assembled monolayers. Appl. Phys. Lett. 2001, 79, 3323− 3325. (6) Smith, R. K.; Lewis, P. A.; Weiss, P. S. Patterning self-assembled monolayers. Prog. Surf. Sci. 2004, 75, 1−68. (7) Hudalla, G. A.; Murphy, W. L. Chemically well-defined selfassembled monolayers for cell culture: toward mimicking the natural ECM. Soft Matter 2011, 7, 9561−9571. (8) Mrksich, M.; Whitesides, G. M. Using self-assembled monolayers to understand the interactions of man-made surfaces with proteins and cells. Annu. Rev. Biophys. Biomol. Struct. 1996, 25, 55−78. (9) Cong, P.; Kubo, T.; Nanao, H.; Minami, I.; Mori, S. Effect of selfassembled monolayers modified slider on head-disk tribology under volatile organic contamination. Tribol. Lett. 2007, 27, 137−143. (10) Mate, C. M.; Mcclelland, G. M.; Erlandsson, R.; Chiang, S. Atomic-Scale Friction of a Tungsten Tip on a Graphite Surface. Phys. Rev. Lett. 1987, 59, 1942−1946. (11) Meier, T.; Menges, F.; Nirmalraj, P.; Hoelscher, H.; Riel, H.; Gotsmann, B. Length-Dependent Thermal Transport along Molecular Chains. Phys. Rev. Lett. 2014, 113, 060801. (12) Sills, S.; Overney, R. M. Creeping Friction Dynamics and Molecular Dissipation Mechanisms in Glassy Polymers. Phys. Rev. Lett. 2003, 91, 095501. (13) Sills, S.; Overney, R. M. Applied Scanning Probe Methods II; Springer, 2005. (14) Sills, S.; Gray, T.; Overney, R. M. Molecular dissipation phenomena of nanoscopic friction in the heterogeneous relaxation regime of a glass former. J. Chem. Phys. 2005, 123, 134902. (15) Binnig, G.; Rohrer, H. Scanning tunneling microscopy. Helvetica Physica Acta 1982, 55, 726−735. (16) Mendoza, S. M.; Arfaoui, I.; Zanarini, S.; Paolucci, F.; Rudolf, P. Improvements in the Characterization of the Crystalline Structure of Acid-Terminated Alkanethiol Self-Assembled Monolayers on Au(111). Langmuir 2007, 23, 582−588. (17) Buergi, L.; Brune, H.; Kern, K. Imaging of Electron Potential Landscapes on Au(111). Phys. Rev. Lett. 2002, 89, 176801. (18) Hanke, F.; Bjoerk, J. Structure and local reactivity of the Au(111) surface reconstruction. Phys. Rev. B: Condens. Matter Mater. Phys. 2013, 87, 235422. (19) Poirier, G. E. Characterization of Organosulfur Molecular Monolayers on Au(111) using Scanning Tunneling Microscopy. Chem. Rev. 1997, 97, 1117−1128. (20) Luessem, B.; Mueller-Meskamp, L.; Karthaeuser, S.; Waser, R. A New Phase of the c(4 × 2) Superstructure of Alkanethiols Grown by Vapor Phase Deposition on Gold. Langmuir 2005, 21, 5256−5258. (21) Gnecco, E.; Bennewitz, R.; Gyalog, T.; Loppacher, C.; Bammerlin, M.; Meyer, E.; Guentherodt, H. J. Velocity dependence of atomic friction. Phys. Rev. Lett. 2000, 84, 1172−1175. (22) Tambe, N. S.; Bhushan, B. Friction model for the velocity dependence of nanoscale friction. Nanotechnology 2005, 16, 2309. (23) Hammerschmidt, J. A.; Gladfelter, W. L.; Haugstad, G. Probing Polymer Viscoelastic Relaxations with Temperature-Controlled Friction Force Microscopy. Macromolecules 1999, 32, 3360−3367. (24) Strobl, G. The Physics of Polymers; Springer, 1996. (25) Grosch, K. A. The Relation between the Friction and ViscoElastic Properties of Rubber. Proc. R. Soc. London, Ser. A 1963, 274, 21−39.
during the series of temperature-dependent measurements, for example, related to different phases. Especially this second effect limits the possibility for further interpretation of Figure 6c. At this point, we can only hypothesize about the energy dissipation process in our experiments. The AFM tip is probably plowing through the thiol chains, that is, these chains bend away from the tip due to the tip pressure but still stay bonded to the surface. The characteristic energy of this plowing process would then be related to the binding energy of the chains with each other. There are no literature values available for direct comparison, however, helium reflectivity measurements reported that the desorption energy of the same thiols as used in this study is around 65 meV, if lying flat on a gold surface.28 This may be viewed as an order magnitude estimate for the interthiol interaction for upright standing thiol films. On the other hand, one may assume that the AFM tip is exciting vibrational modes within the thiol chains. However, literature reports much higher energy values for typical thiol vibrational modes.29
■
CONCLUSION In this work, we analyzed frictional properties of hexadecanethiol as a function of temperature and velocity. As a first step, the SAM was characterized by high-resolution scanning tunneling microscopy and friction force microscopy. These measurements showed that hexadecanethiol on Au(111) forms monolayers consisting of of relatively small domain, where different structures (namely the α-, β-, and γ-phases) could be identified by STM while at the same time also atomic level stick slip of the SAM was measured. Care has to be taken to avoid irreversible changes of the SAM during friction measurements, but once this is achieved, the friction data can be analyzed following the strategy of frequency−temperature superposition, from which an activation energy of EA = 58 ± 2 meV related to SAM inherent dissipation could be extracted. One might envision this process as a partial unzipping of the organized molecule chains by the tip. However, after showing the general principle of frequency velocity superposition applied to hexadecanethiol SAM in this work, further experiments on SAMs of different chain lengths might be an adequate way to substantiate the conclusion of energy dissipation due to unzipping of SAMs. If this process indeed works as assumed, the energy barrier should not depend on the chain length. Financial support was provided by the German Research Foundation (Project DI917/5-1) and in part by COST Action MP1303 and the Laboratory for Material Science of the Justus Liebig University Giessen.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. ORCID
Dirk Dietzel: 0000-0001-6158-6971 Notes
The authors declare no competing financial interest.
■
REFERENCES
(1) Ulman, A. Formation and structure of self-assembled monolayers. Chem. Rev. 1996, 96, 1533−1554. 6009
DOI: 10.1021/acs.langmuir.7b01131 Langmuir 2017, 33, 6005−6010
Article
Langmuir (26) Fujita, M.; Fujihira, M. Effect of temperature on friction observed between a Si 3 N 4 tip and a dodecanethiol self-assembled monolayer on Au (111). Ultramicroscopy 2002, 91, 227−230. (27) Jansen, L.; Hoelscher, H.; Fuchs, H.; Schirmeisen, A. Temperature Dependence of Atomic-Scale Stick-Slip Friction. Phys. Rev. Lett. 2010, 104, 256101. (28) Wetterer, S. M.; Lavrich, D. J.; Cummings, T.; Bernasek, S. L.; Scoles, G. Energetics and Kinetics of the Physisorption of Hydrocarbons on Au(111). J. Phys. Chem. B 1998, 102, 9266−9275. (29) Kato, H. S.; Noh, J.; Hara, M.; Kawai, M. An HREELS Study of Alkanethiol Self-Assembled Monolayers on Au(111). J. Phys. Chem. B 2002, 106, 9655−9658.
6010
DOI: 10.1021/acs.langmuir.7b01131 Langmuir 2017, 33, 6005−6010
