Cross-Sectional Imaging of Boundary Lubrication Layer Formed by

Sep 29, 2017 - Influence of Ionic Strength on the Deposition of Metal–Phenolic Networks. Langmuir. Guo, Richardson, Besford, Christofferson, Dai, On...
1 downloads 8 Views 8MB Size
Article pubs.acs.org/Langmuir

Cross-Sectional Imaging of Boundary Lubrication Layer Formed by Fatty Acid by Means of Frequency-Modulation Atomic Force Microscopy Tomoko Hirayama,*,†,‡ Ryota Kawamura,§ Keita Fujino,§ Takashi Matsuoka,† Hiroshi Komiya,† and Hiroshi Onishi∥ †

Department of Mechanical Engineering and §Graduate School of Science and Engineering, Doshisha University 1-3 Miyakodani, Tatara, Kyotanabe, Kyoto 610-0394, Japan ‡ PRESTO, Japan Science and Technology Agency 7 Gobancho, Chiyoda-ku, Tokyo 102-0076, Japan ∥ Department of Chemistry, Kobe University 1-1 Rokkodai-cho, Nada-ku, Kobe, Hyogo 657-8501, Japan ABSTRACT: To observe in situ the adsorption of fatty acid onto metal surfaces, cross-sectional images of the adsorption layer were acquired by frequency-modulation atomic force microscopy (FM-AFM). Hexadecane and palmitic acid were used as the base oil and typical fatty acid, respectively. A Cu-coated silicon wafer was prepared as the target substrate. The solvation structure formed by hexadecane molecules at the interface between the Cu substrate and the hexadecane was observed, and the layer pitch was found to be about 0.6 nm, which corresponds to the height of hexadecane molecules. This demonstrates that hexadecane molecules physically adsorbed onto the surface due to van der Waals forces with lying orientation because hexadecane is a nonpolar hydrocarbon. When hexadecane with palmitic acid was put on the Cu substrate instead of pure hexadecane, an adsorption layer of palmitic acid was observed at the interface. The layer pitch was about 2.5−2.8 nm, which matches the chain length of palmitic acid molecules well. This indicates that the original adsorption layer was monolayer or single bilayer in the local area. In addition, a cross-sectional image captured 1 h after observation started to reveal that the adsorbed additive layer gradually grew up to be thicker than about 20 nm due to an external stimulus, such as cantilever oscillation. This is the first report of in situ observation of an adsorbed layer by FM-AFM in the tribology field and demonstrates that FM-AFM is useful for clarifying the actual boundary lubrication mechanism.

1. INTRODUCTION Understanding the state and behavior of a boundary lubrication layer formed by additive molecules and its role is an important goal because the formation of a boundary layer greatly affects the coefficient of friction under boundary lubricated conditions. A typical model of a boundary lubrication layer is Hardy’s monolayer model.1 In accordance with this model, we assert that additive molecules with a polar group diffuse in the base oil and spontaneously adsorb due to the polar group onto a metal surface. Also in contrast to Hardy’s model, Allen proposed a multilayer model in 1969,2 and the state of the adsorbed additive layer has been repeatedly discussed. Gellman called this situation a “monolayer-multilayer controversy”.3 Since the most important issue is the sustainability and durability of the boundary layer, the state, monolayer or multilayer, may not be important. However, obtaining physical and chemical information about the actual solid−liquid interface would be quite useful for clarifying the friction reduction mechanism. From a historic perspective on analytical study on boundary lubrication layers, ex situ analytical approaches began to be applied to tribological surfaces in the 1960s and have been used © XXXX American Chemical Society

with high frequency; for example, scanning electron microscope with energy dispersive X-ray microanalyzer (SEM-EDX), Auger electron spectroscopy (AES), and X-ray photoelectron spectroscopy (XPS) are still indispensable for chemical composition analysis of boundary lubrication film produced through chemical reactions between lubricant and metal substrate. Their usage revealed the friction reduction and antiwear mechanisms achieved by extreme pressure agents such as phosphoric acid ester4−6 and zinc dialkyldithiophosphate (ZDDP).7−9 In addition, the use of transmission electron microscopy (TEM) for samples cut by focused ion beam (FIB) is increasing for crosssectional observation of actual sliding surfaces after friction.10−12 However, the targets by such analyzers remain “hard” materials and are definitely not “soft” materials formed by adsorption of additive molecules. To observe the soft layers formed by additive adsorption, in situ analysis is necessary. The recent development of in situ Received: July 20, 2017 Revised: September 2, 2017

A

DOI: 10.1021/acs.langmuir.7b02528 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 1. Diagram of FM-AFM apparatus and tip scanning trace on the x−z plane. FM-AFM is quite sensitive to interaction force around the top of the cantilever; the solvation structure in solution can be imaged cross-sectionally as a density map of liquid molecules from the Δf map obtained in x−z plane scanning.

resolution of this improved FM-AFM was calculated to be 10 pN on the basis of the theory proposed by Sader et al.44 Images of boundary lubrication layers have been produced as simple diagrams and shared among tribology researchers since the 1920s. Now, the use of FM-AFM has the potential to produce not simply diagrams but “actual” cross-sectional images of boundary lubrication layers. The aim of this study was to acquire an actual image of a boundary layer formed by adsorbed fatty acid molecules dissolved in base oil and to clarify the orientation and behavior of the layer.

analyzers for solid−liquid interfaces is noteworthy and their application to tribological study is rapidly increasing in fact. Infrared (IR) spectroscopy had been commonly used since the early 1990s,13−16 and attenuated total reflection infrared spectroscopy (ATR-IR) in particular has been used to obtain chemical information about adsorbed additive layers on substrate surfaces soaked in lubricant.17−21 Infrared light can penetrate crystals made from silicon or germanium, so in situ analysis of the metal surface/lubricant interface is possible if thin metal film is first deposited onto the crystal surface. Another method that is increasingly being applied is sum frequency generation (SFG) spectroscopy, which also uses light, and its use has clarified the interfacial structure and molecular orientations of adsorbed additive layers.22−25 Neutron reflectometry (NR)26−29 and quartz crystal microbalance with dissipation monitoring (QCM-D)30−33 are also useful for in situ evaluation of the thickness and density of adsorbed additive layers, and the combined use of these methods has led to further understanding of the boundary lubrication mechanism. A more recently developed method is frequency-modulation atomic force microscopy (FM-AFM), which has opened a new door to research on solid−liquid interfaces. Atomic-level observation of a liquid environment by FM-AFM was achieved in 200534 due to the use of special technology for low-noise measurement.35 An improved low-noise FM-AFM was used to produce a cross-sectional density map of liquid at the interface, and such maps revealed that the density of liquid at the interface is not homogeneous, as predicted in the 1960s, not only in water36−39 but also in hydrocarbons.40−43 The minimum force

2. FREQUENCY-MODULATION ATOMIC FORCE MICROSCOPE AFM generally has two basic operation modes: “contact mode” and “dynamic mode”. Contact mode is suitable for measuring the static atomic force between a substrate and cantilever, so it is commonly used for estimating the three-dimensional surface topography of hard materials. Dynamic mode, or “tappingmode”, is suitable for measuring the surface topography of soft materials. The latter is better at preventing surface damage during cantilever scanning and has higher sensitivity due to the cancellation of electrostatic force. Dynamic mode is further divided into “amplitude-modulation (AM) mode” and “frequency-modulation (FM) mode”. There is widespread use of AM-AFM especially in soft matter physics field, while there is still little use of FM-AFM due to the greater difficulty in controlling its operation. However, FM-AFM is more sensitive in estimating the structure of soft materials covering liquid due to its higher signal-to-noise ratio.35 In the FM-AFM B

DOI: 10.1021/acs.langmuir.7b02528 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir apparatus, the cantilever is oscillated at the frequency of selfexcited vibration by using a rectangular piezoelectric element. When the head of the cantilever comes close to the surface, the interaction force between the substrate and the top of the cantilever shifts the self-excited frequency of the cantilever oscillation. The surface topography of materials is obtained by moving the cantilever along the z-axis to each (x,y) point, thereby keeping the shift, Δf, constant, as shown in Figure 1a. The most attractive feature of FM-AFM is its ability to draw a Δf map on the x−z plane, as shown in Figure 1b. Because FM-AFM is quite sensitive to the interaction force around the top of the cantilever (less than 10 pN in principle), the solvation structure in the solution can be imaged crosssectionally as a density map of liquid molecules from the Δf map. Though there is still discussion on the interpretation of the relationship between the Δf map and density distribution map of liquid molecules,45,46 FM-AFM can nevertheless produce a cross-sectional image of liquid molecules, which cannot be done with any other method.

3. EXPERIMENTAL SECTION 3.1. Substrate. A 0.2-mm-thick silicon wafer was cut to about 10 × 10 mm square with a diamond cutter and cleaned with ethanol in an ultrasonic bath. After cleaning, Cu film about 100 nm thick was deposited onto the wafer surface by DC sputtering (SVC-7000DCII, Sanyu Electron). We used Cu as the target substrate because Cu is one of the most common and stable metals and is suitable for a first trial because of the higher reactivity and adsorption property of general additive molecules to Cu (especially Cu2O). To investigate the composition ratio of Cu metal and oxidized Cu, XPS and AES spectra were obtained by XPS analyzer (Versa Probe II, Ulvac-Phi) and they are shown in Figure 2 (analyzed by a software “CasaXPS”). The upper XPS spectrum mainly consists of Cu and/or Cu2O peaks and a Cu(OH)2 peak, and the lower AES spectrum consists of a Cu peak and Cu2O and/or Cu(OH)2 peaks. Separting the peaks was not easy, but it was assumed that the deposited target substrate had natural oxidation and hydroxylation film consisting of Cu2O and Cu(OH)2 on the top surface with a Cu:Cu2O:Cu(OH)2 coverage ratio of about 0.33:0.37:0.30, respectively, as calculated from each peak area shown in Figure 2. 3.2. Lubricant. Hexadecane (C16H34, Sigma-Aldrich, >99%) and palmitic acid (C15H31COOH, Sigma-Aldrich, >99%) were used as the base oil and additive because hexadecane is a nonpolar and stable molecule like a commonly used industrial base oil such as poly-α-olefin and palmitic acid is a typical saturated fatty acid used for friction reduction. The palmitic acid was mixed and dissolved into hexadecane in a centrifuge tube in boiling water. The concentration of palmitic acid was 0.01, 0.03, or 0.3 mass% in accordance with the purpose of each trial. 3.3. Apparatus. A commercially available FM-AFM apparatus (SPM-8000FM, Shimadzu) and a cantilever (PPP-NCHAuD, Nanosensors) suitable for FM-AFM observation were used. The specifications of the cantilever are shown in Table 1. FM-AFM can scan not only along the x−z plane but also along the x−y plane, so a crosssectional image showing the density distribution map of liquid molecules on the substrate can be obtained. To obtain an image in solution, a special cantilever holder with a glass slit at the top was used, and the solution was sandwiched between the glass slit and substrate and kept in the gap by its surface tension, as shown in Figure 3, where surface tensions of water and hexadecane are 72.8 and 27.5 mN/m@ 20 °C, respectively. The AFM unit was put in a chamber preventing signal noise from air and the chamber was set on an actively controlled antivibration table. The room temperature and humidity were kept constant: 23 °C and 40%. 3.4. Pre-Evaluation of Apparatus Performance − Hydration Structure on Mica. As a pre-evaluation of the performance of the FM-AFM apparatus, KCl aqueous solution with a KCl concentration

Figure 2. XPS (upper) and AES (lower) spectra for Cu-sputtered substrate. Al Kα X-ray (1486.6 eV) source was used. Deposited target substrate had natural oxidation and hydroxylation film consisting of Cu2O and Cu(OH)2 on top surface with Cu:Cu2O:Cu(OH)2 coverage ratio of about 0.33:0.37:0.30, respectively.

Table 1. Specifications of Cantilever for FM-AFM Observation Length [μm] Spring constant [N/m] Resonant frequency [kHz] Tip material Tip radius [nm] Tip length [μm]

125 38−40 330 Silicon 7 10−15

of 0.05 mol/L and mica crystal were prepared for use as the liquid and substrate. The mica crystal was peeled to form a thin sheet immediately before the test. The surface topography obtained using x−y scanning mode with Δf constant is shown in Figure 4. The honeycomb structure of the mica cleavage plane is clearly evident in the atomic image even though it was in aqueous solution. A crosssectional image derived from the Δf map and the Δf profile along the line on the map are shown in Figure 5a and b, respectively. The crosssectional image shows a few layer structures. The white and black bands correspond to Δf during x−z scanning. The Δf was larger than average in the white bands, and smaller than average in the black bands. Though the precise relationship between the Δf map and density distribution of liquid molecules is still unclear, the simplest interpretation is that the white and black bands correspond to high and low density areas of liquid molecules. This interpretation is based on the equation-of-state model, which predicts that the interaction force at the top of a cantilever is higher if the molecule density is higher than average, and several researchers are endeavoring to clarify the interpretation of the relationship.45,46 On the basis of this simple interpretation, Figure 5a shows H2O molecules in a layer structure, a “hydration structure”, on a mica surface, as also shown in previous research.38,39 The pitch of the Δf wave shown in Figure 5b is about 0.23 nm, which matches the height of H2O molecules. The results of C

DOI: 10.1021/acs.langmuir.7b02528 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 3. Illustration of FM-AFM observation at solid−liquid interface. Solution was sandwiched and kept between upper glass slit and lower substrate by surface tension.

Figure 4. Atomic image of mica surface obtained by x−y plane scanning in KCl solution with FM-AFM. Image of honeycomb structure of mica cleavage plane agrees well with atomic model theoretically predicted. this pre-evaluation demonstrated that this FM-AFM apparatus has sufficiently high performance for directly observing the structure of liquid molecules even in solution.

4. CROSS-SECTIONAL IMAGING AT SOLID−LIQUID INTERFACE 4.1. Solvation Structure of Base Oil on Copper Substrate. The first step was to investigate the structure of hexadecane molecules on a Cu substrate. Figure 6a shows a cross-sectional image captured by x−z scanning at the interface of the Cu substrate and hexadecane. A black-and-white layered structure is evident on the substrate. The black and white areas respectively correspond to low and high density areas of hexadecane molecules, as is the case of the hydration structure described above. This image shows that hexadecane molecules formed a multilayered solvation structure on the Cu substrate. The Δf profile along the line in the image is shown in Figure 6b. It is a clear vibrational Δf profile with a pitch of about 0.6 nm. This result is quite similar to those for the cases of hexadecane on an alkanethiol monolayer40 and 1-dodecanol on graphite;41 vibrational Δf profiles with a pitch of about 0.6 nm at the interface were also clearly observed for these cases. We found that the hydrocarbon molecules orient with their molecular axes parallel to the surface because the layer pitch did not match the chain length of the hydrocarbon

Figure 5. Hydration structure due to water molecules on mica surface. Pitch of Δf wave was about 0.23 nm, which matched the height of H2O molecules.

molecules but instead matched their height. We also found that the height of the hexadecane molecules matched the pitch, demonstrating that the hexadecane molecules were lying and forming a few layers with parallel orientation on the Cu surface. The adsorption is categorized as physical adsorption due to van der Waals forces (the hexadecane molecules did not adhere to the substrate but instead formed a stacking structure) because hexadecane is a nonpolar molecule. 4.2. Interfacial Structure between Cu Substrate and Base Oil with Additive. Our second step was to change the liquid from hexadecane to hexadecane with palmitic acid with a D

DOI: 10.1021/acs.langmuir.7b02528 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 7. Adsorbed layer formed by palmitic acid on Cu surface. Additive concentration was 0.01 mass%. Red dotted lines in (a) indicate layer boundary. Layer pitch agreed well with length of palmitic acid, demonstrating that monolayer and bilayer of acid molecules were locally formed.

Figure 6. Solvation structure formed by hexadecane molecules on Cu surface. Layer pitch was constant and matched height of hexadecane molecules, demonstrating that hexadecane molecules formed a few stacked layers with parallel orientation on the surface.

concentration of 0.01 mass%. A cross-sectional image of the interface is shown in Figure 7a. A layered structure is clearly evident on the Cu substrate. The layer on the substrate apparently was the adsorbed additive layer formed by palmitic acid because the layer was much more visible compared with those in Figures 5a and 6a. The FM-AFM was able to visualize the additive adsorbed layer, a first in the area of tribological study. The amplitude of the cantilever oscillation was set larger for x−z scanning in the observation of the adsorbed additive layer compared with that for the observation of the solvation structure because the additive layer had high density, and therefore the cantilever did not enter the layer if the amplitude of the cantilever was set smaller. The amplitude for the observation shown in Figure 6a was set to about 6−8 Å while that for the observation shown in Figure 7a was set to about 15 Å, which was measured in hexadecane bulk liquid. The additive layer could be observed only when the amplitude of the cantilever oscillation increased, but then the layered solvation structure of hexadecane molecules could not be observed. This means that the amplitude of the cantilever should be tuned in accordance with the target material. The need for larger amplitude oscillation

Figure 8. Cross-sectional image of Si surface/lubricant interface. Additive concentration was 0.3 mass%. Bright area (=high Δf area) was not observed, demonstrating that adsorbed additive layer did not form on Si surface. E

DOI: 10.1021/acs.langmuir.7b02528 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Figure 9. Growth of adsorbed additive layer on Cu surface. Additive concentration was 0.03 mass%. Thickness of adsorbed additive layer gradually increased from (a) to (c), demonstrating that palmitic acid grew from monolayer to multilayer while maintaining layer structure due to cantilever oscillation.

A cross-sectional image of the interface of the Si substrate and hexadecane with palmitic acid with a concentration of 0.3 mass% is shown in Figure 8. The amplitude of the cantilever oscillation was then set larger, so the layered solvation layer formed by hexadecane molecules could not be seen. We can see from Figure 8 that only the surface of the Si substrate could be observed and that an adsorbed additive layer on the Si substrate could not be observed even though the additive concentration was much higher. This demonstrates that an adsorbed additive layer forms only on an oxidized metal surface, as expected. 4.3. Growth of Adsorbed Additive Layer. Our next step was to increase the concentration of palmitic acid in the hexadecane to 0.03 mass%. Cross-sectional images captured at the interface of the Cu substrate and liquid immediately after the observation started, 10 min later, and 1 h later are shown in Figure 9. The layer structure formed by palmitic acid was thin, less than 10 nm at maximum, immediately after the observation started and gradually thickened to over 20 nm 1 h later. However, a layer with a pitch of 2−3 nm is clearly evident even in

to capture the image in Figure 7a means that the adsorbed additive layer was harder with a higher density of molecules than the solvation layers formed by pure hexadecane. Figure 7a shows that there were at least two parallel layers on the Cu surface, as indicated by the red dotted lines. As is wellknown, palmitic acid chemically adsorbs on an oxidized Cu surface and becomes copper palmitate through dehydration condensation: Cu 2O + 2C15H31COOH → 2(C15H31COO)2 Cu + H 2O

The Δf profile along the white line in Figure 7a is shown in Figure 7b. It shows that the first and second layer pitches were about 2.8 and 2.5 nm, respectively. They were close to the chain length of palmitic acid molecules, about 2.4 nm. The color of the first layer was brighter than that of the second. This means that the monolayer and bilayer of the palmitic acid (or copper palmitate) were locally formed and arranged in a clear layered structure on the Cu surface, like those in Hardy’s and Allen’s diagrams. F

DOI: 10.1021/acs.langmuir.7b02528 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir Figure 9c. This demonstrates that the palmitic acid grew from a monolayer to a multilayer while maintaining the layer structure. Though the reason the additive layer tilted up from the surface of the Cu substrate is unclear, it is possible that part of the layer structure broke locally and peeled off from the surface during the probe scanning, and “new” additive molecules covered the surface of substrate. To investigate the effect of probe scanning, we also took a cross-sectional image 1 h after we put the liquid on the substrate in the apparatus and held with the remainder to stop the cantilever oscillation. Interestingly, growth of the layer structure was not observed, and the image captured 1 h after starting was almost the same as that in Figure 9a. This indicates that the growth of the layer structure was driven by an external stimulus, such as cantilever oscillation. Conversely, the layer easily grew even with a small stimulus, so the additive layer is more likely to be a multilayer, several 10 nm thick, under actual tribological conditions. 4.4. Desorption of Additive Layer by Rinsing with Hexane. After capturing the image shown in Figure 9c, we rinsed the substrate with pure hexane and captured a crosssectional image in hexadecane again. In the rinsing process, shown in Figure 10, the substrate was soaked in a beaker with a

Figure 11. Adsorbed additive layer in hexadecane after rinsing with hexane. Thin additive layer only about 5 nm thick remained on surface, indicating that chemically adsorbed additive layer with higher adhesion was left.

were in a tilting orientation taken to minimize potential.47 The image in Figure 7a shows a layer pitch of about 2.5 nm, which almost matches the molecular chain length. However, the additive layer captured as a multilayer “almost always” tilted up from the surface of the substrate, as shown in Figure 9. This was not due to equipment problems, such as heat drift and noise. A likely explanation is that the bottom-layer molecules in such a multilayer adsorbed onto the substrate surface with a tilting orientation, resulting in the multilayer being completely tilted. Several experimental studies using SFG,25 AFM,48 and XAS49 also indicated that the adsorbed molecules tilted although their experimental conditions were not the same as ours. The reported tilting angles ranged from 30° to 60° (depending on the situation), which matches that of the molecules in the images we captured. Regarding the growth of adsorbed additive layers into multilayers, several papers with macrotribological experiments accepted this as natural though the most of papers with molecular simulation show “monolayer” model.47,50 For example, although it is well-known that the formation of a “metal soap” layer between a substrate and fatty acid due to a chemical reaction greatly reduces friction, the layer is typically recognized as being thicker than a monolayer.51,52 The X-ray absorption spectroscopy analysis by Fischer et al. also revealed the growth of a stearic acid layer on a copper surface, especially under rubbing conditions, as judged from the increase in peak value.53 The measurement of oil film thickness under elastohydrodynamic lubricated (EHL) conditions by Ratoi et al. demonstrated that a thicker film with a thickness of around 8−10 nm forms even under low-speed EHL conditions when “wet” stearic acid is mixed into the base oil.54,55 Copper is naturally reactive, so the growth of an adsorbed additive layer as we presented may be natural and not surprising. “Monolayer or multilayer” − FM-AFM is a unique apparatus that enables actual cross-sectional imaging of the boundary lubrication layer through direct observation. The target substrate must be flat because for high resolution along z-axis, but are not limited for metal surface. In addition, all lubricants with various additives can be targeted. The use of FM-AFM is promising for further understanding of the actual boundary

Figure 10. Rinsing process to remove physically adsorbed additive layer. Substrate was subsequently set again in apparatus, and interface was observed again in pure hexadecane.

sufficient volume of pure hexane and shaken by hand. After being rinsed three times, the substrate was set again in the FM-AFM apparatus and pure hexadecane was dropped onto the surface before the observation. The captured cross-sectional image is shown in Figure 11. We can see that a thin layer, only about 5 nm thick, was remained on the substrate surface. Previous studies found that hexane-rinsing washed away the physically adsorbed layer, leaving the chemically adsorbed layer. The remaining layer was not clearly arranged in a layered structure, but the concept that the remaining layer was a chemically adsorbed one with higher adhesion to the substrate is reasonable. 4.5. Discussion. Recently, approaches based on molecular simulation are increasingly being used to clarify the boundary lubrication layer in lubricant on top of a substrate. Several papers have reported the orientation and behavior of a boundary lubrication layer formed by fatty acid, and most of them reported that the molecules that adsorbed onto a metal surface G

DOI: 10.1021/acs.langmuir.7b02528 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

(10) Ito, K.; Martin, J. M.; Minfray, C.; Kato, K. Low-friction tribofilm formed by the reaction of ZDDP on iron oxide. Tribol. Int. 2006, 39 (12), 1538−1544. (11) Kosarieh, S.; Morina, A.; Flemming, J.; Lainé, E.; Neville, A. Wear Mechanisms of Hydrogenated DLC in Oils Containing MoDTC. Tribol. Lett. 2016, 64, 4. (12) Erdemir, A.; Ramirez, G.; Eryilmaz, O. L.; Narayanan, B.; Liao, Y.; Kamath, G.; Sankaranarayanan, S. K. R. S. Carbon-based tribofilms from lubricating oils. Nature 2016, 536, 67−71. (13) Perez, J. M.; Ku, C. S.; Pei, P.; Hegemann, B. E.; Hsu, S. M. Characterization of Tricresylphosphate Lubricating Films by MicroFourier Transform Infrared Spectroscopy. Tribol. Trans. 1990, 33 (1), 131−139. (14) Hu, Z.-S.; Hsu, S. M.; Wang, P. S. Tribochemical and Thermochemical Reactions of Stearic Acid on Copper Surfaces Studied by Infrared Microspectroscopy. Tribol. Trans. 1992, 35 (1), 189−193. (15) Willermet, P. A.; Carter, R. O., III; Boulos, E. N. Lubricantderived tribochemical films − An infra-red spectroscopic study. Tribol. Int. 1992, 25 (6), 371−380. (16) Sun, J.-X.; Hu, Z.-S.; Hsu, S. M. The Effect of Concentration, Solvent, and Temperature on Aggregation of a Commercial Calcium Sulfonate Additive as Studied by FTIR and Light Scattering Techniques. Tribol. Trans. 1997, 40 (4), 633−638. (17) Cann, P. M.; Spikes, H. A. In-Contact IR Spectroscopy of Hydrocarbon Lubricants. Tribol. Lett. 2005, 19, 289−297. (18) Piras, F. M.; Rossi, A.; Spencer, N. D. Growth of Tribological Films: In Situ Characterization Based on Attenuated Total Reflection Infrared Spectroscopy. Langmuir 2002, 18 (17), 6606−6613. (19) Piras, F. M.; Rossi, A.; Spencer, N. D. Combined in situ (ATR FT-IR) and ex situ (XPS) Study of the ZnDTP-Iron Surface Interaction. Tribol. Lett. 2003, 15 (3), 181−191. (20) Mangolini, F.; Rossi, A.; Spencer, N. D. Chemical Reactivity of Triphenyl Phosphorothionate (TPPT) with Iron: An ATR/FT-IR and XPS Investigation. J. Phys. Chem. C 2011, 115 (4), 1339−1354. (21) Sasaki, K.; Inayoshi, N.; Tashiro, K. Development of New In Situ Observation System for Dynamic Study of Lubricant Molecules on Metal Friction Surfaces by Two-Dimensional Fast-Imaging FourierTransform Infrared-Attenuated Total Reflection Spectrometer. Rev. Sci. Instrum. 2008, 79, 123702. (22) Ward, R. N.; Duffy, D. C.; Davies, P. B.; Bain, C. D. SumFrequency Spectroscopy of Surfactants Adsorbed at a Flat Hydrophobic Surface. J. Phys. Chem. 1994, 98 (34), 8536−8542. (23) Chen, Z.; Ward, R.; Tian, Y.; Baldelli, S.; Opdahl, A.; Shen, Y.R.; Somorjai, G. A. Detection of Hydrophobic End Groups on Polymer Surfaces by Sum-Frequency Generation Vibrational Spectroscopy. J. Am. Chem. Soc. 2000, 122 (43), 10615−10620. (24) Miyake, K.; Kume, T.; Nakano, M.; Korenaga, A.; Takiwatari, K.; Tsuboi, R.; Sasaki, S. Effects of Surface Chemical Properties on the Frictional Properties of Self-Assembled Monolayers Lubricated with Oleic Acid. Tribol. Online 2012, 7 (4), 218−224. (25) Watanabe, S.; Nakano, M.; Miyake, K.; Sasaki, S. Analysis of the Interfacial Molecular Behavior of a Lubrication Film of n-Dodecane Containing Stearic Acid under Lubricating Conditions by Sum Frequency Generation Spectroscopy. Langmuir 2016, 32, 13649− 13656. (26) Hirayama, T.; Torii, T.; Konishi, Y.; Maeda, M.; Matsuoka, T.; Inoue, K.; Hino, M.; Yamazaki, D.; Takeda, M. Thickness and density of adsorbed additive layer on metal surface in lubricant by neutron reflectometry. Tribol. Int. 2012, 54, 100−105. (27) Kalin, M.; Simič, R.; Hirayama, T.; Geue, T.; Korelis, P. Neutron-reflectometry study of alcohol adsorption on various DLC coatings. Appl. Surf. Sci. 2014, 288 (1), 405−410. (28) Lin, B.; Zhu, H.; Tieu, A. K.; Hirayama, T.; Kosasih, B.; Novareza, O. Adsorbed Film Structure and Tribological Performance of Aqueous Copolymer Lubricants. Wear 2015, 332−333, 1262−1272. (29) Wood, M. H.; Welbourn, R. J. L.; Charlton, T.; Zarbakhsh, A.; Casford, M. T.; Clarke, S. M. Hexadecylamine Adsorption at the Iron Oxide-Oil Interface. Langmuir 2013, 29 (45), 13735−13742.

lubrication mechanism and will be rapidly widespread as a powerful method in tribology field in near future.

5. CONCLUSION This study focused on the adsorbed additive layer for boundary lubrication, and FM-AFM was used to capture cross-sectional images of the layer in lubricant. Formation of a solvation layer by base oil molecules was clearly observed using FM-AFM. The layer pitch was about 0.6 nm, which matches the molecular height. This demonstrates that the base oil molecules without polarized functional groups were physically adsorbed onto the metal surface with lying orientation. The in situ observation of a boundary lubrication layer formed by fatty acid by FM-AFM was successful. The crosssectional image showed that the fatty acid molecules clearly formed a layer structure with a layer pitch of about 2.5 nm, which matches the molecular length of acid. Continuous scanning using a cantilever revealed that the adsorbed additive layer gradually grew and ultimately reached a thickness of about 20 nm. This demonstrates that the adsorbed additive layer easily grew due to an external stimulus.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel. +81-774-65-6413. Fax. +81-774-65-6827. ORCID

Tomoko Hirayama: 0000-0001-7138-7407 Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This study was supported by JST PRESTO (Grant No. JPMJPR13C9) and the Canon Foundation. The authors are grateful to Dr. Ryohei Kokawa of Shimadzu Corporation for his helpful advice and guidance.



REFERENCES

(1) Hutchings, I. Tribology: Friction and Wear of Engineering Materials; Butterworth-Heinemann: London, 1992. (2) Allen, C. M.; Drauglis, E. Boundary layer lubrication: monolayer or multilayer. Wear 1969, 14 (5), 363−384. (3) Gellman, A. J.; Spencer, N. D. Surface Chemistry in Tribology. Proc. Inst. Mech. Eng., Part J 2002, 216 (6), 443−461. (4) Najman, M. N.; Kasrai, M.; Bancroft, G. M.; Miller, A. Study of the Chemistry of Films Generated from Phosphate Ester Additives on 52100 Steel Using X-ray Absorption Spectroscopy. Tribol. Lett. 2002, 13 (3), 209−218. (5) Najman, M. N.; Kasrai, M.; Bancroft, G. M. Chemistry of Antiwear Films from Ashless Thiophosphate Oil Additives. Tribol. Lett. 2004, 17 (2), 217−229. (6) Mistry, K. K.; Morina, A.; Erdemir, A.; Neville, A. Tribological Performance of EP Lubricants with Phosphorus-Based Additives. Tribol. Trans. 2013, 56 (4), 645−651. (7) Jahanmir, S. Wear reduction and surface layer formation by a ZDDP additive. J. Tribol. 1987, 109, 577−586. (8) Nicholls, M. A.; Do, T.; Norton, P. R.; Kasrai, M.; Bancroft, G. M. Review of the lubrication of metallic surfaces by zinc dialkyldithiophosphates. Tribol. Int. 2005, 38, 15−39. (9) Martin, J. M.; Onodera, T.; Minfray, C.; Dassenoy, F.; Miyamoto, A. The origin of anti-wear chemistry of ZDDP. Faraday Discuss. 2012, 156, 311−323. H

DOI: 10.1021/acs.langmuir.7b02528 Langmuir XXXX, XXX, XXX−XXX

Article

Langmuir

Modifiers Adsorbed on Iron Oxide Surfaces. Langmuir 2016, 32 (18), 4450−4463. (51) Bowden, F. P.; Tabor, D. Friction and Lubrication of Solids; Oxford University Press: Oxford, UK, 1950. (52) Stachowiak, G. W.; Batchelor, A. W. Engineering Tribology; Elsevier: Amsterdam, Netherlands, 1993. (53) Fischer, D. A.; Hu, Z. S.; Hsu, S. M. Molecular orientation and bonding of monolayer stearicacid on a copper surface prepared in air. Tribol. Lett. 1997, 3 (1), 41−45. (54) Spikes, H. A. Direct Observation of Boundary Layers. Langmuir 1996, 12 (19), 4567−4573. (55) Ratoi, M.; Anghel, V.; Bovington, C. H.; Spikes, H. A. Mechanisms of oiliness additives. Tribol. Int. 2000, 33, 241−247.

(30) Muller, M. T.; Yan, X.; Lee, S.; Perry, S. S.; Spencer, N. D. Lubrication Properties of a Brushlike Copolymer as a Function of the Amount of Solvent Absorbed within the Brush. Macromolecules 2005, 38 (13), 5706−5713. (31) Evans, K. O.; Biresaw, G. Quartz crystal microbalance investigation of the structure of adsorbed soybean oil and methyl oleate onto steel surface. Thin Solid Films 2010, 519, 900−905. (32) Stokes, J. R.; Macakova, L.; Chojnicka-Paszun, A.; de Kruif, C. G.; de Jongh, H. H. Lubrication, adsorption, and rheology of aqueous polysaccharide solutions. Langmuir 2011, 27 (7), 3474−3484. (33) Veeregowda, D. H.; Busscher, H. J.; Vissink, A.; Jager, D.-J.; Sharma, P. K.; van der Mei, H. C. Role of Structure and Glycosylation of Adsorbed Protein Films in Biolubrication. PLoS One 2012, 7 (8), e42600. (34) Fukuma, T.; Kobayashi, K.; Matsushige, K.; Yamada, H. True atomic resolution in liquid by frequency-modulation atomic force microscopy. Appl. Phys. Lett. 2005, 86, 193108. (35) Fukuma, T.; Kobayashi, K.; Matsushige, K.; Yamada, H. Development of low noise cantilever deflection sensor for multienvironment frequency-modulation atomic force microscopy. Rev. Sci. Instrum. 2005, 76, 053704. (36) Fukuma, T.; Jarvis, S. P. Development of liquid-environment frequency modulation atomic force microscope with low noise deflection sensor for cantilevers of various dimensions. Rev. Sci. Instrum. 2006, 77, 043701. (37) Higgins, M. J.; Polcik, M.; Fukuma, T.; Sader, J. E.; Nakayama, Y.; Jarvis, S. P. Structured Water Layers Adjacent to Biological Membranes. Biophys. J. 2006, 91 (7), 2532−2542. (38) Fukuma, T.; Higgins, M. J.; Jarvis, S. P. Direct Imaging of Individual Intrinsic Hydration Layers on Lipid Bilayers at Angstrom Resolution. Biophys. J. 2007, 92 (10), 3603−3609. (39) Hiasa, T.; Kimura, K.; Onishi, H. Hydration of hydrophilic thiolate monolayers visualized by atomic force microscopy. Phys. Chem. Chem. Phys. 2012, 14, 8419−8424. (40) Hiasa, T.; Kimura, K.; Onishi, H. Two-dimensional distribution of liquid hydrocarbons facing alkanethiol monolayers visualized by frequency modulation atomic force microscopy. Colloids Surf., A 2012, 396, 203−207. (41) Hiasa, T.; Kimura, K.; Onishi, H. Cross-Sectional Structure of Liquid 1-Decanol over Graphite. J. Phys. Chem. C 2012, 116, 26475− 26479. (42) Hiasa, T.; Kimura, K.; Onishi, H. Minitips in FrequencyModulation Atomic Force Microscopy at Liquid-Solid Interfaces. Jpn. J. Appl. Phys. 2012, 51, 025703. (43) Hiasa, T.; Kimura, K.; Onishi, H. Interfacial Structure of Primary and Tertiary Liquid Alcohols over Hydrophilic Thiolate Monolayers. J. Phys. Chem. C 2013, 117, 5730−5735. (44) Sader, J. E.; Jarvis, S. P. Accurate formulas for interaction force and energy in frequency modulation force spectroscopy. Appl. Phys. Lett. 2004, 84 (10), 1801−1803. (45) Amano, K.; Suzuki, K.; Fukuma, T.; Takahashi, O.; Onishi, H. The relationship between local liquid density and force applied on a tip of atomic force microscope: A theoretical analysis for simple liquids. J. Chem. Phys. 2013, 139, 224710. (46) Watkins, M.; Reischl, B. A simple approximation for forces exerted on an AFM tip in liquid. J. Chem. Phys. 2013, 138, 154703. (47) Doig, M.; Warrens, C. P.; Camp, P. J. Structure and Friction of Stearic Acid and Oleic Acid Films Adsorbed on Iron Oxide Surfaces in Squalane. Langmuir 2014, 30 (1), 186−195. (48) Campen, S.; Green, J. H.; Lamb, G. D.; Spikes, H. A. In Situ Study of Model Organic Friction Modifiers Using Liquid Cell AFM; Saturated and Mono-unsaturated Carboxylic Acids. Tribol. Lett. 2015, 57, 18. (49) Fischer, D. A.; Hu, Z. S.; Hsu, S. M. Molecular orientation and bonding of monolayer stearic acid on a copper surface prepared in air. Tribol. Lett. 1997, 3, 41. (50) Ewen, J. P.; Gattinoni, C.; Morgan, N.; Spikes, H. A.; Dini, D. Nonequilibrium Molecular Dynamics Simulations of Organic Friction I

DOI: 10.1021/acs.langmuir.7b02528 Langmuir XXXX, XXX, XXX−XXX