Article pubs.acs.org/JPCC
Tribochemical Degradation of Polytetrafluoroethylene Catalyzed by Copper and Aluminum Surfaces Tasuku Onodera,*,† Takayuki Nakakawaji,† Koshi Adachi,‡ Kazue Kurihara,§,∥ and Momoji Kubo⊥ †
Center for Technology Innovation − Materials, Research, and Development Group, Hitachi, Ltd. 7-1-1 Omika, Hitachi 319-1292, Japan ‡ Department of Nanomechanics, Graduate School of Engineering, Tohoku University 6-6-1 Aoba, Aramaki, Aoba-ku, Sendai 980-8579, Japan § WPI Advanced Institute for Materials Research, ∥Institute of Multidisciplinary Research for Advanced Materials, and ⊥Institute for Materials Research, Tohoku University 2-1-1 Katahira, Aoba-ku, Sendai 980-8577, Japan ABSTRACT: The tribochemical reaction of polytetrafluoroethylene (PTFE) resin sliding against a metallic surface was investigated for a purpose of improving its tribological performance. The tribochemical reaction between PTFE and aluminum surface has been investigated experimentally and theoretically [Onodera et al., J. Phys. Chem. C 2015, 119, 15954−15962]. One of the important results is that the wear and self-lubrication properties can be controlled by the catalytic activity of the oxidized aluminum surfaces with different crystal structures (α- and γ-alumina). In particular, the amount of wear was higher on the γ-alumina surface because it easily activates tribochemical degradation of PTFE and suppresses the formation of transfer film (a necessary phenomenon for reducing wear). Accordingly, for controlling catalytic activity a copper surface was tested as a model surface exhibiting a weaker reactivity than an aluminum surface. A thermogravimetric analysis proved that the copper surface showed a weaker catalytic effect in regard to PTFE degradation in comparison with an aluminum surface. Metallic fluoride, causing less transfer film to form, appeared on the aluminum surface but hardly any was observed on the copper surface. An experiment and density functional theory (DFT) calculations showed lower catalytic activity on copper surfaces. A friction test was also carried out to determine the relationship between the catalytic activity of metallic surfaces on PTFE and the resultant wear properties. A rubbed copper surface was fully covered with flakelike PTFE transfer film, and no metallic fluoride was detected. Transfer film was less abundant on an aluminum surface because aluminum fluoride tribochemically formed. Reflecting these chemical changes, the wear amount of PTFE on the copper surface was less than that on the aluminum surface. The results of this investigation indicate that suppressing the catalytic effect during friction process of PTFE is a promising way to improve its tribological performance. The use of a copper-based material will be effective in industrial uses of PTFE, because it suppresses catalytic tribochemical degradation. contact without any other external fields. In this case, the oxidized surface of aluminum interacts as a “tribocatalyst” and a defluorination reaction takes place. The tribocatalytic effect of aluminum surfaces changes the friction and wear properties of PTFE.10 Aluminum fluoride, which is a product of the tribocatalytic reaction, suppresses the formation of PTFE transfer film on aluminum surfaces. Figure 1 shows the results of a molecular dynamics (MD) simulation for passivated and fluorinated surfaces of aluminum.9 The PTFE transfer film firmly forms on the passivated surface (a) by contributing many hydrogen bonds on the interface (blue dotted line), while it hardly sticks to the fluorinated surface (b). The difficulty of transfer film formation on fluorinated surfaces is caused by the electrostatic repulsive interaction between the
1. INTRODUCTION Metallic materials act as powerful catalysts for several organic compounds. The related research includes methanol synthesis using copper and zinc,1 ammonia synthesis using cobalt,2 and phenol to cyclohexane conversion using molybdenum.3 Synthesis and degradation of hydrocarbon polymer compounds have also been studied. For example, polyethylene can be synthesized through the use of a vanadium catalyst4 and can be degraded with a platinum catalyst.5 A silica−alumina complex acts as a catalyst for depolymerization of polystyrene.6 On the other hand, fluoropolymers such as polytetrafluoroethylene (PTFE) hardly decompose even on metallic materials because of the strong covalent bonds between the carbon and fluorine atoms in its molecular structure. One of the few ways to decompose PTFE is application of an electric field with an alkaline metal. 7 On the other hand, in our previous investigation8−10 we found that a metallic aluminum surface, surprisingly, easily reacts with PTFE when it is in frictional © 2016 American Chemical Society
Received: January 25, 2016 Revised: May 6, 2016 Published: May 6, 2016 10857
DOI: 10.1021/acs.jpcc.6b00799 J. Phys. Chem. C 2016, 120, 10857−10865
Article
The Journal of Physical Chemistry C
6200 thermogravimetric/differential thermal analyzer (TG/ DTA) was used. Two samples of mixed-powder, copper with PTFE and aluminum with PTFE, were prepared to demonstrate the situation of contact between oxidized metallic surfaces and PTFE. The chemical composition and crystal structure of the metal oxide surface are important considerations in this analysis because they affect the resultant catalytic activity. Figure 2 shows X-ray diffraction (XRD)
Figure 1. Results of molecular dynamics simulation for (a) passivated and (b) fluorinated surfaces of α-alumina (0001).9
fluorinated surface and PTFE. The formation of transfer film is one of the most important phenomena to reduce friction and wear of PTFE because this lubricious film is able to inhibit direct contact of PTFE with metallic surface despite its nanometric-scale thickness,11 that is, 10 nm. Also, the coverage of the transfer film is important to describe the self-lubrication mechanism of PTFE.12 However, formation of PTFE transfer film is difficult on a surface covered with aluminum fluoride due to the electrostatic repulsive interaction between the two materials; thus, the tribocatalytic reaction worsens the friction and wear performance of PTFE on aluminum surfaces. Furthermore, fluoride forms more easily on metallic surfaces under low-humidity conditions because the catalytic center is always exposed during the friction process.9 These findings suggest us that prevention of fluoride formation, especially under low-humidity conditions, must be the key to improving the friction and wear performances of PTFE sliding against a metallic surface. The tribochemical effect of the metallic surface should be investigated in order to address the issue of fluoride formation. To precoat the metallic surface with a covalent material is one of the ways to prevent the tribocatalytic effect. Mori et al. studied the tribocatalytic reaction of gaseous fluorocarbons with aluminum. Their results showed that the aluminum oxide acts as a stronger catalyst than the aluminum nitride which exhibits greater covalency.13−15 Another way is to use a metallic material that has lower chemical affinity with fluorine than that of aluminum. This method is more feasible than the former idea, which uses hard material (nitride) and thus may cause mechanical wear. Accordingly, in this study the tribocatalytic effect of copper surfaces (copper is a representative material showing low chemical affinity with fluorine) sliding against PTFE was investigated in a thermal analysis and friction test. A density functional theory (DFT) calculation was also conducted to support the results of these experiments. The resultant tribological performances of PTFE and the chemical nature of the rubbed copper surfaces were compared with the case of aluminum,10 which is catalytically active for PTFE decomposition.
Figure 2. XRD patterns for surfaces of a pure copper specimen used in the friction test. The diffraction intensity is plotted on a log scale.
patterns from the surface of a pure copper specimen used in the friction test (see Section 2.2). A Rigaku Smart Lab 90TF was used in this analysis, and Cu K-α was chosen as the X-ray source. The diffraction peaks in the figure can be assigned to Cu and Cu2O crystal structures. This is a reasonable result because many researchers have reported that Cu2O usually forms on copper surfaces.16−20 On the basis of this analysis, metallic Cu powder was used to consider the Cu2O surface. Our previous report10 on aluminum oxide compared the catalytic activities of α-Al2O3 and γ-Al2O3 surfaces and showed that the catalytic effect of γ-Al2O3 is stronger for PTFE degradation than that of α-Al2O3. In this paper, γ-Al2O3 was chosen as the model compound of the aluminum surface. The Cu and γ-Al2O3 powders (produced by Kojundo Chemical) used in this study had average particle sizes of about 1 and 3 μm, respectively. An average particle size of the PTFE powder (produced by SynQuest) was about 0.5 μm and average molecular weight of 5 × 104. All samples were prepared with the same mass ratio of 50% PTFE and 50% metal. The sample chamber was purged by pure nitrogen gas with flow rate of 200 cm3/min. The samples were heated to 773 K for 360 min duration to observe the distinct effects of the catalytic reaction. After the sample chamber was cooled to room temperature, the heated residue was sampled. Subsequently, the chemical products in the heated powders was analyzed by X-ray photoelectron spectroscopy (XPS), using an Ulvac-Phi PHI5600. In XPS analysis, a focused monochromatic Al K-α X-ray (14 keV) beam was scanned over a 800 μm-diameter spot. The photoelectron takeoff angle relative to the sample normal was 45°. 2.2. Friction Test. Evaluating the tribochemical effect on the friction and wear performances of PTFE with different metals was next studied by conducting friction test using a pinon-disk tribometer (Bruker UMT-3 with a rotary drive). Because the main objective of this experiment was to measure
2. METHODS 2.1. Thermal Analysis. Comparing the catalytic activities of two metals (copper and aluminum) in degrading PTFE polymer should be the first step in our investigation. For this purpose, a thermal analysis was chosen. A Seiko Instruments 10858
DOI: 10.1021/acs.jpcc.6b00799 J. Phys. Chem. C 2016, 120, 10857−10865
Article
The Journal of Physical Chemistry C
a temperature of 773 K. At time zero, the weight 100% equals the mass of PTFE with metal. The TG/DTA results for a pure PTFE powder sample without any metal oxides are shown for reference. These thermograms trace the thermal degradation of PTFE, including formation of gaseous molecules such as small fragments of fluorocarbon and hydrogen fluoride. The Cu2O and γ-Al2O3 mixed samples showed catalytic effects causing thermal degradation of PTFE because these metal oxides accelerated the decrease in the material weight. The degradation rates, however, were different; it can be clearly seen that weight loss is more significant for the sample with γAl2O3 than for the sample with Cu2O. In the Cu2O case, the PTFE degradation is not so different from the case without any metal. Therefore, the oxidized surface of copper probably acts as a mild catalyst for PTFE degradation. Detecting the chemical products in the thermochemical reaction between PTFE and metal oxides is the next step in evaluating the catalytic activity. For this purpose, XPS was performed on heated residue obtained in the TG/DTA tests. Figure 4 shows the fluorine-1s XPS spectra for the TG/DTA-
the wear amount of PTFE, a 1/4 in. PTFE ball was put in contact with copper (99.99% purity) and aluminum disks (99% purity). The diameter and thickness of each disk were 30 and 4 mm, respectively. The arithmetic average roughness (Ra) of both disk surfaces was 0.05 μm. The micro-Vickers hardness was measured using a Shimadzu HMV-G under 1 N loading. A value of 40 HV was obtained for the aluminum specimen and 105 HV for the copper specimen. In the friction test, a normal load of 1 N was applied while the rotation speed was set to 100 rpm. The equivalent sliding speed was 0.0523 m/s. The test was conducted for 720 min (temperature of 293−298 K and relative humidity over 50%). After the friction test, the sliding surface of the metal disks was observed by scanning electron microscopy (SEM) (Hitachi SU1510) and energy-dispersive X-ray spectroscopy (EDX) (Horiba EMAX). An XPS analysis was also carried out to determine the chemical bonds. 2.3. Density Functional Theory. The DFT calculation was conducted in order to confirm the results of the above two experiments. The DMol3 program implemented in Accelrys Materials Studio Version 4.3 was used. Double numerical basis sets with polarization and generalized gradient approximation in terms of Perdew−Burke−Ernzerhof (PBE) exchangecorrelation functionals21,22 were employed. Transition states (TS) were searched with the complete LST/QST method, where the linear synchronous transit (LST) maximization was performed for the coordinates interpolated between a reactant and a product, followed by repeated conjugated gradient minimizations and quadratic synchronous transit (QST) maximizations until a TS was located.23,24 This search method has been successful to find the TS structure for a catalyst,25,26 metallic surface,27 and organic systems.28,29 The calculation model are described in Section 3.1. Accelrys Materials Studio Visualizer was used to visualize all atomic models.
3. RESULTS AND DISCUSSION 3.1. Catalytic Activity in PTFE Degradation. The difference in the catalytic effect between copper and aluminum was studied in the TG/DTA analysis. Figure 3 shows the weight loss of PTFE sample powder mixed with two different metal oxides (Cu with a Cu2O oxidized surface and γ-Al2O3) at
Figure 4. Fluorine-1s XPS spectrum for the residue of PTFE with and without γ-Al2O3 or Cu (Cu2O oxidized surface) heated at 773 K.
obtained residue samples. The XPS spectrum for a nonheated PTFE sample without any metal is shown for reference. There is only one peak for the nonheated sample, which is of the carbon and fluorine bond in PTFE (binding energy of 689.3 eV). Similar spectra can be seen in the case of the heated residue of PTFE mixed with Cu (Cu2O surface) and no other chemical bonds are evident. Thus, only unreacted PTFE remained in this sample. On the other hand, the peak of the carbon and fluorine bond slightly shifted to a lower binding energy in the case of the γ-Al2O3-including sample. This shift was caused by a formation of intermediate including a carbon double bond in the PTFE backbone (FCCF). Note that a carbon double bond formed in the earlier experiment on PTFE degradation.7 Moreover, another peak can be seen; it is assigned to metallic fluoride (aluminum fluoride in this case). These results imply that the PTFE was catalytically decomposed by γ-Al2O3 and released fluorine, forming aluminum fluoride instead. These experimental analyses suggest that the catalytic activity (for PTFE degradation) of Cu2O on the Cu surface is significantly lower than that of γ-Al2O3. Because metallic
Figure 3. Thermograms obtained by TG/DTA at 773 K for PTFE powder mixed with and without γ-Al2O3 or Cu (Cu2O oxidized surface). 10859
DOI: 10.1021/acs.jpcc.6b00799 J. Phys. Chem. C 2016, 120, 10857−10865
Article
The Journal of Physical Chemistry C
Figure 5. DFT-optimized models of reactant, transition state, and product for C5F12 molecule adsorbed on (a) Cu2O (111) and (b) γ-Al2O3 (100) surfaces. Atoms on the topmost surface are expressed as large balls. The size of the surface area is (a) a = 1.046 nm and b = 1.208 nm. (b) a = 1.117 nm and b = 0.841 nm.
fluoride, which inhibits transfer film formation, hardly formed on the Cu2O surface, the tribological properties of PTFE were probably better than in the case of aluminum. The next section describes the tribological test using copper and aluminum specimens to test this assumption. Next, a DFT calculation was performed to confirm the experimental analyses described above and quantitatively estimate the catalytic effect of the two metallic materials. The calculation model included a fluorocarbon molecule (C5F12 as a model compound of PTFE) and a metal oxide surface. Cu2O (111) and γ-Al2O3 (100) oxide surfaces were modeled; these are well-known stable surfaces for Cu2O30 and γ-Al2O3.31 The model surfaces included 120 and 80 atoms, respectively. The surface was not passivated by several hydroxyl groups (chemisorbed water molecules in the atmosphere) because they usually detach to reform water in heating32 or sliding33 conditions. One C5F12 molecule was placed on each surface and its geometry was optimized by DFT to obtain the most stable structure of molecular adsorption. Figure 5a,b shows the optimized adsorption structures as the reactants of the chemical reaction in which two fluorine atoms dissociate from the C5F12 molecule and two bonds form between the metal and fluorine atoms (see Figure 7 in ref 10). The chemical products include fluorinated metal surface and C5F10 molecule with a carbon double bond. The geometries of the product were also optimized by DFT. The complete LST/QST method was then used to obtain the TS structure. The DFT-obtained structure for the TS and product are also shown in Figure 5a,b. In the TS structure (a), the carbon and fluorine bond is extended to 0.259 nm (initial bond distance was 0.135 nm) on the Cu2O (111) surface, meaning that two fluorine atoms (radicals) are released. On the other hand, fluorine radicals do not appear on the γ-Al2O3 (100) surface, while one bond between aluminum and fluorine atoms already formed (bond distance was 0.170 nm). Figure 6 shows the energy diagram for these defluorination reactions on each metallic surface. In the figure, the energy values are relative to
Figure 6. Energy diagram for the chemical reaction of C5F12 fluorocarbon molecule adsorbed on Cu2O (111) and γ-Al2O3 (100) surfaces. The products include a C5F10 molecule and a partially fluorinated surface. The energy of each reactant was set to zero.
each reactant. Reflecting the formation of fluorine radicals on the Cu2O (111) surface, the TS energy involved in the C5F12 molecule reaction is significantly higher than that on the γAl2O3 (100) surface (over 200 kJ/mol). The defluorination reaction in the presence of metallic fluoride is thus more difficult on the copper surface than on the aluminum surface. These results show the same trend as the experimental analyses described above. Thus, the chemical stability of the TS, viz., formation of fluorine radicals, is the key to the formation of metallic fluoride which suppresses transfer film formation during a friction process. The DFT calculation revealed the molecular mechanism behind the reaction of PTFE degradation. 3.2. Effect of Catalytic Reaction on Tribological Behavior. The results of the thermal analysis and supporting 10860
DOI: 10.1021/acs.jpcc.6b00799 J. Phys. Chem. C 2016, 120, 10857−10865
Article
The Journal of Physical Chemistry C
To summarize the two wear periods, specific wear rates were calculated under the friction conditions we used. PTFE rubbed against copper had 6.32 × 10−5 mm3/(N·m), while PTFE rubbed against aluminum had 1.39 × 10−4 mm3/(N·m). The amount of wear on copper was almost half that on aluminum and this potentially points to a lower catalytic activity. These results indicate that the wear of PTFE is significantly influenced by the catalytic activity of the metallic counter surfaces. Even though the micro-Vickers hardness of the copper disk (105 HV) was 2.5 times larger than that of the aluminum disk (40 HV), the copper surface caused less wear. Also, by referring to a nanoindentaiton test, a native oxide layer of copper is 7.2−12.3 GPa, which depend on heat treatment temperature,36 and 9.5 GPa for aluminum.37 Thus, the hardness of two native oxide layers (Cu2O and Al2O3) were nearly the same or Cu2O layer is harder than Al2O3 layer. These facts also explain that in the case of PTFE friction the amount of wear is determined by not only the mechanical properties of the metal but also the catalytic effect. In order to examine the relationship between the formation of transfer film and wear performance, the surfaces of the metals and PTFE balls were observed after the 720 min friction test. Figure 8 shows optical microscope images, SEM images, EDX-obtained fluorine mapping images, and fluorine-1s XPS spectra for copper and aluminum surfaces rubbed with PTFE. Similarly, the optical microscope image and fluorine-1s XPS spectra for the PTFE ball sliding with the copper and aluminum disks are shown in Figure 9. In Figure 8a,b, it can be observed that a flakelike matter appeared on the rubbed copper surface while distinct wear scars occur on the rubbed aluminum surface. To further investigate this difference, a representative geometry for the cross-section of wear track on the copper and aluminum surfaces, which were measured by KEYENCE KS1100 laser displacement meter, is also shown in Figure 10. On the copper surface, there is only a small roughness and no clear wear can be observed. In contrast, reflecting wear scar observed in the optical microscope and SEM images, a scratched part on the aluminum surface is clearly found (wear depth is approximately 6 μm). On the other hand, in Figure 9a any clear differences are not observed in the wear scar on two ball surface. Several wear debris are presented on only the aluminum surface. By referring to the EDX mapping and XPS spectrum in Figure 8c,d, the flakelike matter on the copper surface could be definitely assigned to the PTFE transfer film. Furthermore, no metal fluoride was detected. As described in Section 3.1, the lower catalytic activity of the copper surface (Cu2O native oxide) should lead to lower wear as a result of forming a firmly attached transfer film. On the other hand, on the rubbed aluminum surface fluorine is less abundant (see EDX image) and aluminum fluoride also forms (see XPS spectrum). Less transfer film and metal fluoride, as induced by the higher catalytic activity of the aluminum surface, would be the cause of the PTFE wear in the friction test. In the XPS spectrum for PTFE ball (Figure 9b), metal fluoride was not detected in both copper and aluminum cases. This result indicates that the metal and its fluoride form were not transferred to PTFE surface during friction. The mechanism of transfer film formation on copper surface is quite different from that on aluminum surface because of less chemical affinity with fluorine. Above XPS results strongly support this difference. From our theoretical and experimental results on the aluminum case, it can be inferred that the
DFT calculation suggested that the copper surface is less active in the defluorination reaction than the aluminum surface is. A fluorinated metal surface, which is the product of a chemical reaction between the metal and PTFE, decreases the amount of PTFE transfer film formed by the electrostatic repulsive interaction acting on the interface.9 The tribological performance of PTFE is probably better when it is rubbed on a copper surface than when it is rubbed on an aluminum surface. The second investigation of this study, accordingly, assessed the influence of the catalytic reaction of the metallic surface on the tribological behavior of PTFE. Friction tests were performed on copper and aluminum disks in contact with a pure PTFE ball. As described in Section 2.2, copper (99.99% purity) and aluminum disks (99% purity) were used, and native oxide Cu2O formed on the copper surface (see Figure 2). Here, the friction tests were performed three times for each disk to check reproducibility. Each test shows a similar friction coefficient and the steady friction coefficient on average was 0.142 for the copper surface and 0.134 for the aluminum surface. The range of friction coefficient obtained is in good agreement with that found in several papers that also discussed the friction between PTFE and metal.12,34,35 Although the friction was a little higher on the copper surface, there were no significant differences in the frictional behaviors on these surfaces as far as the applied conditions (contact load, sliding speed, and environment) went. Figure 7 shows the wear of the PTFE ball over the time in which the friction was applied. A downward displacement of
Figure 7. Wear of PTFE ball sliding against copper and aluminum disks.
the PTFE ball (the vertical axis of the graph) was employed as a representative amount of wear. Both wear curves are divided into two parts: a running-in period (denoted as T1) and a steady-state period (denoted as T2). The rate of wear is significantly higher during the running-in period than in the steady-state period because the gradient of the graph is steeper. The time for the transition from running-in to steady-state differs between the copper and the aluminum surfaces (37 and 89 min, respectively). This means that PTFE reaches a stable wear stated earlier on the copper surface than on the aluminum surface. In the steady-state period, PTFE shows a gradual increase in wear on both surfaces; however, the wear rate again differs between the two surfaces: 0.0353 μm/min for the copper surface and 0.0468 μm/min for the aluminum surface. The rate of wear on the copper surface is slower in the steadystate period. 10861
DOI: 10.1021/acs.jpcc.6b00799 J. Phys. Chem. C 2016, 120, 10857−10865
Article
The Journal of Physical Chemistry C
Figure 8. (a) Optical microscope, (b) SEM, and (c) EDX images of wear scars on copper and aluminum disks sliding against a PTFE ball. (d) XPS fluorine-1s spectrum for the same two disk samples. In the EDX images, green means the existence of fluorine atoms.
formation of transfer film is controlled by the aluminum fluoride. In the copper case, the fluoride was not detected on the surface rubbed with PTFE. Probably, the main driving force of the transfer film formation on copper surface is not a chemical interaction. A mechanical effect, such as anchoring effect by surface roughness, should be one of the possible mechanism for transfer film formation. The wear mode of PTFE is also different between the copper and aluminum cases because a distinct wear scar was observed only on the aluminum surface. The metal fluoride chemically formed may cause the wear of aluminum surface. This material is inherently harder than the original aluminum surface and acts as an abrasive particle (wear debris). In fact, the surface of aluminum was scratched as shown in Figures 8a and 10. On the copper surface, no abrasive wear has been observed because the metal fluoride was not formed. In order to prove this hypothesis on wear induced by abrasive particle of metal fluoride, XPS analysis for the wear debris is required. In our next paper, detailed wear mechanism will be discussed by referring to the XPS analysis for wear debris. Additionally, no clear difference in the friction coefficient was found between the copper and aluminum cases (0.142 for the
copper and 0.134 for the aluminum) while the PTFE wear was influenced by the catalytic activity of metal surface. One of the possible explanations for this fact is that the friction is determined by the topmost surface of the metal. By referring to above XPS results in Figure 8d, the topmost surface of both metals is covered by the PTFE transfer film although there is a difference in the amount of transfer film between two metal surfaces. It is speculated that presence of PTFE transfer film controls friction and its amount or thickness rarely contributes to determining friction. To provide additional quantitative support for the above experimental results, a DFT calculation was conducted comparing the adhesive energies of PTFE transfer films on oxidized and fluorinated metallic surfaces. Figure 11 shows the models and the resultant adhesive energy. Note that the adhesive energy, EA, was calculated as follows EA = EM/CF − (EM + ECF)
(1)
where EM is the energy of the oxide surface (Cu2O (111) or γAl2O3 (100)), ECF is the energy for the C5F12 molecule (a model compound of PTFE), and EM/CF is the energy for the C5F12 molecule adsorbed on the metal oxide surface. A larger 10862
DOI: 10.1021/acs.jpcc.6b00799 J. Phys. Chem. C 2016, 120, 10857−10865
Article
The Journal of Physical Chemistry C
Figure 9. (a) Optical microscope image and (b) XPS fluorine-1s spectrum for wear scars on the PTFE ball sliding against copper and aluminum.
Figure 10. A cross-section of wear track on copper and aluminum surfaces. A center of wear track is set to zero and its vicinity is only shown.
Figure 11. DFT-obtained adhesive energy of C5F12 molecule on oxidized and fluorinated surfaces of copper and aluminum.
negative value means strong adhesion of the C5F12 molecule. The figure clearly indicates that the fluorocarbon more easily attaches to the oxidized surface than to the fluorinated one for both copper and aluminum. The reason for the lower adhesion on the fluorinated surface is the electrostatic repulsive interaction with the fluorine atoms in the C5F12 molecule. Although the fluorinated surface of copper decreases the adhesive energy of the fluorocarbon, this material is harder to form during a friction process in comparison with the
aluminum case (see the description of the simulation in Section 3.1). Furthermore, in nature oxidized copper surfaces show better adhesion for fluorocarbons (physical adsorption). Besides a mechanical anchoring effect described above, this is also one of the advantages of using copper as a surface. It can be inferred that these merits of copper were exhibited in the friction experiment as a regular formation of transfer film with a flake-like morphology. 10863
DOI: 10.1021/acs.jpcc.6b00799 J. Phys. Chem. C 2016, 120, 10857−10865
Article
The Journal of Physical Chemistry C
(3) Itthibenchapong, V.; Ratanatawanate, C.; Oura, M.; Faungnawakij, K. A Facile and Low-Cost Synthesis of MoS2 for Hydrodeoxygenation of Phenol. Catal. Commun. 2015, 68, 31−35. (4) Ochedzan-Siodlak, W.; Bihun, A. Direct Synthesis of Fibrous High Molecular Weight Polyethylene Using Vanadium Catalysts Supported on an SiO2 Ionic Liquid System. Polym. Int. 2015, 64, 1600−1606. (5) Roozbehani, B.; Sakaki, S. A.; Shishesaz, M.; Abdollahkhani, N.; Hamedifar, S. Taguchi Method Approach on Catalytic Degradation of Polyethylene and Polypropylene into Gasoline. Clean Technol. Environ. Policy 2015, 17, 1873−1882. (6) Moqadam, S. I.; Mirdrikvand, M.; Roozbehani, B.; Kharaghani, A.; Shishehsaz, M. R. Polystyrene Pyrolysis Using Silica-Alumina Catalyst in Fluidized Bed Reactor. Clean Technol. Environ. Policy 2015, 17, 1847−1860. (7) Kavan, L.; Dousek, F. P.; Janda, P.; Weber, J. Carbonization of Highly Oriented Poly(tetrafluoroethylene). Chem. Mater. 1999, 11, 329−335. (8) Onodera, T.; Kawasaki, K.; Nakakawaji, T.; Higuchi, Y.; Ozawa, N.; Kurihara, K.; Kubo, M. Chemical Reaction Mechanism of Polytetrafluoroethylene on Aluminum Surface under Friction Condition. J. Phys. Chem. C 2014, 118, 5390−5396. (9) Onodera, T.; Kawasaki, K.; Nakakawaji, T.; Higuchi, Y.; Ozawa, N.; Kurihara, K.; Kubo, M. Effect of Tribochemical Reaction on Transfer-film Formation by Poly(tetrafluoroethylene). J. Phys. Chem. C 2014, 118, 11820−11826. (10) Onodera, T.; Kawasaki, K.; Nakakawaji, T.; Higuchi, Y.; Ozawa, N.; Kurihara, K.; Kubo, M. Tribocatalytic Reaction of Polytetrafluoroethylene Sliding on an Aluminum Surface. J. Phys. Chem. C 2015, 119, 15954−15962. (11) Bahadur, S. The Development of Transfer Layers and Their Role in Polymer Tribology. Wear 2000, 245, 92−99. (12) Ye, J.; Khare, H. S.; Burris, D. L. Transfer Film Evolution and Its Role in Promoting Ultra-Low Wear of a PTFE Nanocomposite. Wear 2013, 297, 1095−1102. (13) Wu, X.; Cong, P.; Mori, S. Adsorption and reactions of HFC134a Gas on the Nascent Surface of Alumina. Appl. Surf. Sci. 2002, 201, 115−122. (14) Wu, X.; Cong, P.; Nanao, H.; Kobayashi, K.; Mori, S. Chemisorption and Tribochemical Reaction Mechanisms of HFC134a on Nascent Ceramic Surface. Langmuir 2002, 18, 10122−10127. (15) Cong, P.; Wu, X.; Nanao, H.; Mori, S. Tribological Characteristics and Tribochemical Reactions of Various Ceramics Lubricated with HFC-134a Gas. Tribol. Lett. 2003, 15, 65−72. (16) Rhodin, T. N. Low Temperature Oxidation of Copper. I. Physical Mechanism. J. Am. Chem. Soc. 1950, 72, 5102−5106. (17) Allen, J. A. Oxide Films on Electrolytically Polished Copper Surfaces. Trans. Faraday Soc. 1952, 48, 273−279. (18) Wilhelm, S. M.; Tanizawa, Y.; Liu, C. Y.; Hackerman, N. A Photo-Electrochemical Investigation of Semiconducting Oxide Films on Copper. Corros. Sci. 1982, 22, 791−805. (19) Chawla, S. K.; Rickett, B. I.; Sankarraman, N.; Payer, J. H. An Xray Photo-Electron Spectroscopic Investigation of the Air-Formed Film on Copper. Corros. Sci. 1992, 33, 1617−1631. (20) Suzuki, S.; Ishikawa, Y.; Isshiki, M.; Waseda, Y. Native Oxide Layers Formed on the Surface of Ultra High-Purity Iron and Copper Investigated by Angle Resolved XPS. Mater. Trans., JIM 1997, 38, 1004−1009. (21) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (22) Perdew, J. P.; Burke, K.; Ernzerhof, M. Errata: Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1997, 78, 1396. (23) Halgren, T. A.; Lipscomb, W. N. The Synchronous-transit Method for Determining Reaction Pathways and Locating Molecular Transition States. Chem. Phys. Lett. 1977, 49, 225−232. (24) Bell, S.; Crighton, J. S. Locating Transition States. J. Chem. Phys. 1984, 80, 2464−2475. (25) Sasaki, T.; Tada, M.; Iwasawa, Y. Density Functional Theory Study on the Re Cluster/HZSM-5 Catalysis for Direct Phenol
It can be concluded from this study that the catalytic effect of the surface controls the wear of PTFE. The stronger catalytic effect of aluminum causes more wear because of metallic fluoride forming through a tribochemical reaction. Lower activity of the metallic surface is an important factor to improving the tribological properties. The results of this study show that the use of copper-based materials is one way to reduce PTFE wear.
4. CONCLUSIONS The catalytic effects of two metallic surfaces, copper and aluminum, were experimentally and theoretically examined in regard to the tribological performance of PTFE. To investigate the chemical reactivity of copper and aluminum, a TG/DTA experimental analysis for powder samples of PTFE with metals was performed. The results of the thermal analysis showed that the copper surface potentially exhibited a weaker catalytic effect on PTFE degradation than the aluminum surface. Metallic fluoride formed in the aluminum sample, while none formed in the copper sample. The DFT calculations accounted for this difference. In the case of copper, unstable structures with fluorine radicals increased the activation energy of PTFE degradation. By using a pin-on-disk tribometer, the effect of the catalytic reaction on tribological performance was studied. A PTFE ball without any additives was put in contact with pure copper and aluminum disks. The PTFE wear on the copper surface was less than that on the aluminum surface. EDX and XPS chemical analyses indicated that the PTFE transfer film formed on the copper surface whereas not so much formed on the aluminum surface. The aluminum surface instead had aluminum fluoride on it. The antiwear performance of PTFE is thereby worsened by the aluminum fluoride. Accordingly, preventing the catalytic effect of a metal surface would be a way to improve tribological performance. A copper-based material is effective for industrial uses of PTFE because it suppresses tribochemical degradation in shearing. The next step of our investigation will be to seek additives for PTFE-based composite materials that will improve the tribological performance of the metal/resin sliding interface.
■
AUTHOR INFORMATION
Corresponding Author
*E-mail:
[email protected]. Tel: +81-294-525111. Fax: +81-294-52-7622. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS This work was supported by “Tohoku Innovative Materials Technology Initiatives for Reconstruction (TIMT)” funded by the Ministry of Education, Culture, Sports, Science and Technology (MEXT) and Reconstruction Agency, Japan.
■
REFERENCES
(1) Lei, H.; Hou, Z.; Xie, J. Hydrogenation of CO2 to CH3OH over CuO/ZnO/Al2O3 Catalysts Prepared via a Solvent-Free Routine. Fuel 2016, 164, 191−198. (2) Tarka, A.; Zybert, M.; Truszkiewicz, E.; Mierzwa, B.; Kepinski, L.; Moszynski, D.; Rarug-Pilecka, R. Effect of a Barium Promoter on the Stability and Activity of Carbon-Supported Cobalt Catalysts for Ammonia Synthesis. ChemCatChem 2015, 7, 2836−2839. 10864
DOI: 10.1021/acs.jpcc.6b00799 J. Phys. Chem. C 2016, 120, 10857−10865
Article
The Journal of Physical Chemistry C Synthesis from Benzene and Molecular Oxygen: Active Re Structure and Reaction Mechanism. Top. Catal. 2009, 52, 880−887. (26) Ao, Z. M.; Peeters, F. M. Electric Field: A Catalyst for Hydrogenation of Graphene. Appl. Phys. Lett. 2010, 96, 253106. (27) Zhang, W.; Wu, P.; Li, Z.; Yang, J. First-Principles Thermodynamics of Graphene Growth on Cu Surfaces. J. Phys. Chem. C 2011, 115, 17782−17787. (28) Tian, L.; Wang, J.; Shen, B.; Liu, J. Building a Kinetic Model for Steam Cracking by the Method of Structure-Oriented Lumping. Energy Fuels 2010, 24, 4380−4386. (29) Ao, Z. M.; Hernández-Nieves, A. D.; Peeters, F. M.; Li, S. The Electric Field as a Novel Switch for Uptake/Release of Hydrogen for Storage in Nitrogen Doped Graphene. Phys. Chem. Chem. Phys. 2012, 14, 1463−1467. (30) Bendavid, L. I.; Carter, E. A. First-Principles Predictions of the Structure, Stability, and Photocatalytic Potential of Cu2O Surfaces. J. Phys. Chem. B 2013, 117, 15750−15760. (31) Digne, M.; Sautet, P.; Raybaud, P.; Euzen, P.; Toulhoat, H. Use of DFT to Achieve a Rational Understanding of Acid-Basic Properties of γ-Alumina Surfaces. J. Catal. 2004, 226, 54−68. (32) Bhattacharya, I. N.; Das, S. C.; Mukherjee, P. S.; Paul, S.; Mitra, P. K. Thermal Decomposition of Precipitated Fine Aluminum Trihydroxide. Scand. J. Metall. 2004, 33, 211−219. (33) Danchevskaya, M. N.; Ivakin, Y. D.; Martynova, L. F.; Zuy, A. I.; Muravieva, G. P.; Lazarev, V. B. Investigation of Thermal Transformations in Aluminum Hydroxides Subjected to Mechanical Treatment. J. Therm. Anal. 1996, 46, 1215−1222. (34) Krick, B. A.; Ewin, J. J.; Blackman, G. S.; Junk, C. P.; Sawyer, W. G. Environmental Dependence of Ultra-low Wear Behavior of Polytetrafluoroethylene (PTFE) and Alumina Composites Suggests Tribochemical Mechanisms. Tribol. Int. 2012, 51, 42−46. (35) Pitenis, A. A.; Ewin, J. J.; Harris, K. L.; Sawyer, W. G.; Krick, B. A. In Vacuo Tribological Behavior of Polytetrafluoroethylene (PTFE) and Alumina Nanocomposites: The Importance of Water for Ultralow Wear. Tribol. Lett. 2014, 53, 189−197. (36) Jian, S. R.; Chen, G. J.; Hsu, W. M. Mechanical Properties of Cu2O Thin Films by Nanoindentation. Materials 2013, 6, 4505−4513. (37) Jõgiaas, T.; Zabels, R.; Tamm, A.; Merisalu, M.; Hussainova, I.; Heikkilä, M.; Mändar, H.; Kukli, K.; Ritala, M.; Leskelä, M. Mechanical Properties of Aluminum, Zirconium, Hafnium and Tantalum Oxides and Their Nanolaminates Grown by Atomic Layer Deposition. Surf. Coat. Technol. 2015, 282, 36−42.
10865
DOI: 10.1021/acs.jpcc.6b00799 J. Phys. Chem. C 2016, 120, 10857−10865