Tribo-Chemistry of Phosphonium-Derived Ionic Liquids

The structure, name, and abbreviation (code) of the ionic liquid samples used in this experiment are shown in Table 1. BMIM-TFSA was selected as the reference sample. Methoxyethyl-substituted ammonium salt (Et₂MeMeOEN-TFSA) was examined as a nitrogen cation with a structure analogous to methoxyethyl-substituted phosphonium salt (Et₃MeOEP-TFSA). All the samples employed in the tribo-tests were of reagent grade and used as received.

The friction-reducing properties of the samples were determined using a T-type pendulum friction tester [26]. This test is conducted on a small quantity (approximately 1 cm³) of a sample in order to determine its friction coefficient under boundary conditions with good repeatability. Thirty tests were run for each sample and the average of friction coefficient for all test runs was reported. A ball-on-flat-type tribo-tester in a reciprocating motion was employed to evaluate the anti-wear properties of sample. Two tests were run for each set of test parameters and the average of two test runs was reported if differences in the results were <15%. After the test, the wear scar diameter of the ball and wear width on flat specimen were measured using an optical microscope. The details of the test conditions are shown in Tables 2 and 3.

After the ball-on-flat-type test was conducted, the resultant flat was ultrasonically cleaned in hexane and the rubbed surfaces were studied by means of Auger electron spectroscopy (AES) and X-ray photoelectron spectroscopy (XPS). The conditions of the surface analyses are shown in Tables 4 and 5.

The preliminary results of the pendulum friction test were obtained in order to determine the friction-reducing properties of a lubricant under boundary conditions with good repeatability. The friction coefficients of all the phosphonium salts were observed to be lower than that of BIMI-TFSA, as shown in Fig. 1. As we have reported previously, the pendulum friction test indicated that the friction coefficient of 1-alkyl-3-methylimidazolium salts decreased with an increase in the chain length of the alkyl group [12, 14]. On the other hand, as shown in Fig. 2, the alkyl group appeared to have a slight effect on the friction coefficient of phosphonium salts. The addition of a benzyl substituent resulted in a slight increase in the friction coefficient. Et₃MeOEP-TFSA exhibited a considerably low friction coefficient as compared to that of Et₂MeMeOEN-TFSA. When the friction coefficients of phosphate salts such as Bu3MeP-MePhos or Bu4P-EtTPhos were compared to those of TFSA salts such as Bu3PMe-TFSA, it was observed that the phosphate anion (in the case of Bu3MeP-MePhos) afforded a slightly high friction coefficient to the phosphate salt whereas the thiophosphate anion (in the case of Bu4P-EtTPhos) afforded a low friction coefficient to the phosphate salt. These results indicate the effect of the anion on the tribological properties of phosphonium salts.

The results of the ball-on-flat-type test also showed that the friction coefficient of phosphonium salts was lower than that of BMIM-TFSA; the only phosphonium salt that was an exception to this was the Et₃MeOEP-TFSA (Fig. 3). The wear width on the ball and flat were also measured after the tribo-test was conducted (Fig. 4). The wear by (thio)phosphate salts was lower than that by TFSA salts.

We are interested in the surface chemistry involved in the tribological process. Therefore, we selected Bu₃PMe-TFSA, Bu₃MeP-MePhos, and Bu₄P-EtTPhos for further investigations. The tribological properties of the selected salts were evaluated as functions of the applied load using a ball-on-flat-type tester. We determined the relationship between the chemical structure and the friction-reducing properties (Fig. 5). For a load of 40 N, all the phosphonium salts exhibited a friction coefficient of approximately 0.08 and the 1,3-dialkylimidazolium salt exhibited a higher friction coefficient of 0.095. For a load of 60 N, only the (thio)phosphate salts exhibited a low friction coefficient (approximately 0.08). The TFSA salts, i.e., both the BMIM salt and the R₄P salt, exhibited a high friction coefficient of 0.09. These results indicate that cationic moiety is responsible for the friction coefficient for a load of 40 N, while anionic moieties are responsible for the friction coefficient for a load of 60 N. For a load of 10 and 20 N, both the anionic and the cationic moieties influence the friction coefficient. Consequently, in the case of cationic moieties, the friction coefficient decreases in the following order: 1,3-dialkylimidazolium > tetraalkylphosphonium. Further, in the case of anionic moieties, it decreases in the following order: TFSA > phosphate > thiophosphate.

The anionic moieties clearly have an effect on the wear properties of the salts. The wear caused by the (thio)phosphate salts was considerably lower than that caused by the TFSA salts (Fig. 6). The presence of the phosphonium cation contributed to the reduction in the wear caused by the TFSA salt. For a load of 10–40 N, the width of the wear caused by Bu₃PMe-TFSA was lower than that caused by BMIM-TFSA. However, for a load of 60 N, the phosphonium cation appeared to have no effect on wear reduction. These results indicate that the anionic moieties are greatly responsible for the wear. The cationic moieties affect the wear reducing properties under low-load conditions.

Figure 7 shows the elemental mapping on the flat specimen including inside and outside the wear track. We found that the concentration of oxygen and phosphorus inside the wear track was higher than that outside it. The presence of phosphate inside the wear track was confirmed by the XPS spectrum, which showed the binding energy of P 2p at 133 eV (Fig. 8). These results indicate that the tribochemical reaction of Bu₃PMe-MePhos yielded a boundary film composed of iron phosphate. It is well recognized that the phosphate boundary film is instrumental in improving the tribological properties [27]. Sulfoxides, sulphides, and phosphate were detected on the surface rubbed with Bu₄P-EtTPhos (Fig. 9). Sulphides are also known as key compounds that improve tribological properties [28]. The (thio)phosphate salts, Bu₃PMe-MePhos and Bu₄P-EtTPhos, are analogous in chemical structure to conventional antiwear additives such as tricresyl phosphates and zinc dialkyldithiophospates. Therefore, the tribological properties and surface chemistry of (thio)phosphate salts are closely related to those of conventional lubricants.

In contrast to the observations of the elemental mapping on the flat specimen rubbed with Bu₃PMe-MePhos, no detectable phosphorus was found inside the wear track in the case of the specimen rubbed with the TFSA salt (Bu₃PMe-TFSA) (Fig. 10). These results indicate that anionic moieties are major reactants in the tribochemical reactions under these conditions. The tribochemical reaction of TFSA salts has been reported to yield metal fluorides on rubbed surfaces [29]. In fact, we detected both organic fluoride (C–F at the binding energy of 688 eV) and inorganic fluoride (Fe–F at the binding energy of 684 eV) on the surface rubbed with Bu₃PMe-TFSA (Fig. 11). The former is the TFSA adsorbed on the surface and the latter is the TFSA reacted with iron on the surface. The ratio of the adsorbed fluoride to the reacted fluoride changes with the change in the load applied in the tribo-test. In the case of a 20 N load, a greater proportion of adsorbed TFSA was observed on the rubbed surface. On the other hand, in the case of a 60 N load, a greater proportion of reacted TFSA was observed on the rubbed surface. It is well understood that activation energy sufficient for the tribochemical reaction of TFSA was provided at a higher load.

Detectable amounts of phosphate as well as sulfoxides were observed on the surface rubbed with Bu₃MeP-TFSA (Fig. 12). The presence of the phosphate indicates that the tribochemical reaction of the cationic moiety occurred during the tribo-test [22]. Both adsorbed TFSA and reacted TFSA contained sulfoxide moieties. Therefore, the presence of both these species is possible.

When the specimen was rubbed with the reference sample (BMIM-TFSA), metal fluorides, sulfoxides, and sulphides were detected on the rubbed surfaces (Fig. 13). A careful study of the fluorine XPS spectra shown in Figs. 11 and 13 indicates the differences between the tribochemical reactions of Bu₃MeP-TFSA and BMIM-TFSA. In the case of a 20 N load, a greater proportion of adsorbed TFSA was observed in the case of Bu₃MeP-TFSA, whereas a greater proportion of reacted TFSA was observed in the case of BMIM-TFSA. The sulphur XPS spectra shown in Figs. 12 and 13 also suggest that the tribochemical reaction of the anions in BMIM-TFSA. A considerable amount of sulphide was detected on the surface rubbed with BMIM-TFSA. However, no sulphides—if any exist—were detected on the surface rubbed with Bu₃MeP-TFSA. From these results, it can be concluded that for a 20 N load, the phosphonium cation reduced the tribochemical reaction of the TFSA anion to some extent. However, the cation did not reduce the reaction of the anion for a load of 60 N. As shown in Table 6, no phosphorus was detected on the rubbed surface. These results indicate that these conditions were more favourable to the reaction of the anion than the reaction of the cation.

The results of the ball-on-flat-type test reveal the effects of the anionic moieties on friction and wear. The tribological properties of (thio)phosphate salts were observed to be better than those of TFSA salts (Figs. 5, 6). As mentioned above, the reaction of the (thio)phosphate salts yielded a phosphate-containing boundary film and the reaction of the TFSA salts yielded a fluoride-containing boundary film. From these results, it can be concluded that phosphate films exhibit better tribological properties than fluoride films. For a load of 20 N, the tribological properties of Bu₃MeP-TFSA, which contains a phosphorus cation and a fluorine anion, were observed to be intermediate between those of BMIM-TFSA and Bu₃MeP-MePhos. The reaction of Bu₃MeP-TFSA yielded a boundary film that comprised both phosphate and fluoride and contained an adsorbed TFSA on its uppermost surface. These results indicate that both phosphate and fluoride ions are responsible for the tribological properties of the salt. For a load of 60 N, Bu₃MeP-TFSA yielded a fluoride film and no detectable phosphate film. Under these conditions, the tribological properties of Bu₃MeP-TFSA and BIMI-TFSA were quite similar. We wish to propose a model of the boundary film on the basis of the similarities in the tribological properties and surface chemistry of these salts (Fig. 14).

It has been stated that the tribological properties of iron fluorides are better than those of naked metal surfaces [30, 31]. Despite the fact that the iron fluorides possess physical properties that are conducive to better tribological properties, they have certain disadvantages from a chemical perspective. The hydrolysis of iron fluorides yields iron hydroxides and iron oxides and hence results in the rusting of iron. Iron fluorides are one of the Lewis acids that catalyze the degradation of lubricants [29, 32]. Therefore, a decrease in the tribochemical reaction of the TFSA anion appears to be instrumental in improving the tribological properties of RT-ILs.

The contact stress applied in the pendulum test (1.1 GPa) was lower than that applied in the ball-on-flat-type test for a load of 20 N (1.3 GPa). Therefore, it was assumed that the tribochemical reaction of TFSA was inhibited during the pendulum friction test, and hence, the effect of the phosphate boundary film on friction was possibly greater in the pendulum friction test. This is probably the reason for the low friction coefficient of phosphonium salts as compared to that of 1,3-dialkylimidazolium salts. However, the surface analyses of the surface rubbed during the pendulum friction test were difficult to conduct because no measurable wear scars were obtained in this tribo-test.