1 Introduction

With rapid improvement of modern industrial technology requirements, equipment parts are constantly developing toward the requirements of high speed, high load, high reliability and long service life. As a supporting component, the self-lubricating spherical plain bearing plays an important role in oil-free lubricating parts in engineering machinery, aerospace and other military equipment (Ref 1). It is a kind of spherical plain bearing without adding lubricant, which has been widely applied in helicopter and aircraft because of its simple structure, small size, impact resistance, corrosion resistance and long service life, etc. The spherical plain bearing is composed of an outer ring with an inner sphere and an inner ring with an outer sphere, and the self-lubricating composite solidified on the inner sphere of the outer ring determines the self-lubricating effect and service life of bearing. Direct friction of the inner and outer ring with bearing steel material is replaced by sliding friction between the composite and the outer spherical surface of inner ring, with the advantages of large bearing capacity, stable self-lubricating effect, simple configuration, impact resistance and convenient maintenance. However, if the self-lubricating composite material loses efficacy once, the bearing steel can be glued in a short time under high-frequency and high-load condition, resulting in failure of the whole component. For example, the failure of the self-lubricating pad in spherical plain bearing in the tail-rotor pitch control push rod can directly lead to the loss of control of the helicopter blade. The aerodynamic force at the negative angle of attack makes the blade swing down and sweep to the tail beam, which will inevitably cause aircraft accidents and casualties (Ref 2). Therefore, the self-lubricating composite determines performance of the spherical plain bearing (Ref 3, 4).

The metal matrix-reinforced self-lubricating composites (copper, copper alloy, etc.) meet the performance requirements of most low-speed spherical plain bearings, but have the disadvantages of high friction coefficient and high wear rate under high-speed condition (Ref 5). For this reason, the fiber-reinforced PTFE fabric composites have been widely studied and applied in spherical plain bearings due to their good stability and wear resistance (Ref 6, 7). Most attention has been paid on the glass fiber, aramid fiber, carbon fiber, polyester fiber, etc. (Ref 8,9,10,11,12,13,14). The modification of tribological properties of self-lubricating materials with different fibers has different characteristics. Gu studied the friction and wear properties of PTFE/Kevlar fabric composite under vacuum condition and found that Kevlar fiber could significantly improve the wear resistance of the material (Ref 8). Yang investigated the friction and wear properties of PTFE/Nomex fabric/phenolic composite at various ambient temperatures, indicating that the nomex fiber has high-temperature resistance (Ref 10). Li discussed the tribological properties of PTFE/Kevlar fabric composite with different fabric density (Ref 14). Aglan et al. systematically evaluated the fatigue fracture mechanism of glass fiber-reinforced PTFE (GF/PTFE) composites (Ref 15,16,17). Su et al. compared the sliding friction and wear performances of carbon fiber and glass fiber-reinforcced PTFE fabric composite, respectively, and noted that GF/PTFE composite combines the advantage of good wear resistance from glass fiber and the self-lubricating capacity from PTFE fiber (Ref 18, 19). Under different environment conditions, the friction and wear mechanisms of the self-lubricating composite have also been revealed (Ref 20,21,22). Tribological behaviors of Kevlar fiber, carbon fiber and glass fiber-reinforced polyimide were also compared under different serving conditions (Ref 23). Results show that the CF/PI and GF/PI composites have the lowest and the highest friction coefficient, respectively, and the Kevlar fiber shows excellent wear resistance. Zhang et al. studied the composition, preparation process and properties of different fiber-reinforced PTFE composites (Ref 24,25,26,27). He found that the Kevlar fiber and glass fiber-reinforced PTFE composite exhibits good friction and wear performance under high-speed condition. Recently, the Kevlar fiber has being widely used in equipment parts under harsh service conditions due to its excellent fatigue resistance.

The self-lubricating fiber-reinforced composites have attracted many researchers’ attention, mainly focused on the component of different reinforced fibers and their friction properties. Self-lubricating fiber-reinforced composites have good wear resistance and long life cycle, but there is still a lack of life cycle experiment on this material. For this, the high-frequency and high-load oscillatory tribotester was used to carry out tribological test on Kevalr/PTFE fiber-reinforced composite, investigating the self-lubricating and damage behaviors of the material during its life cycle and exploring the key factors to influence the tribological and life properties. Based on the test and analysis results, the preparation process of material was improved, and the corresponding results were expected to promote the preparation process and application development of high-performance self-lubricating composites.

2 Experimental

2.1 Materials

In this study, the Kevlar fiber-reinforced PTFE self-lubricating composite was prepared with satin weave mode by Lanzhou Institute of Chemical Physics, Chinese Academy of Science. Kevlar fiber fabric (with density of 260 g/m2) is woven from fiber Kevlar 49 provided by DuPont Company, and the physical properties are listed in Table 1. The adhesive resin is an organosilicon-modified phenolic acetal resin, provided by Shanghai Xinguang Resin Co., Ltd. The preparation process is as follows. Kevlar and PTFE fabrics were washed with petroleum ether and ethanol for more than 36 h to remove impurities. After drying, it is repeatedly impregnated and brushed in the bonding resin solution containing solid lubricant filler to make the mass fraction of fiber fabric between 75 and 80%.

Table 1 Physical properties of Kevlar 49
Full size table

The specimen was obtained with a thickness of 0.38 ± 0.02 mm. It was adhered on a base (with a surface roughness of 0.45 mm) with a radius of curvature of 50 ± 0.05 mm and a radian of 60 ± 0.5°. The phenolic acetal resin was solidified in the curing process at the curing temperature of 180 °C, pressure of 0.3 MPa and curing time of 2 h. The initial surface of material is shown in Fig. 1. Peel strength of material adhered was from 0.35 to 0.40 N/mm, and peel angle between the composite and base surface was 140 ± 40°. Fabric tightness was not more than 0.07 mm. The counterpart sliding against the specimen was 9Cr18Mo, processed with subzero treatment temperature minus 55 °C and with thermal insulation time no less than 1 h according to the requirement of high-carbon chromium bearing stainless steel technology. The counterpart had a diameter of 100 mm, with a surface roughness of 0.16 lm, degreased with acetone before test. The test was performed with maintaining temperature of 25 °C and relative humidity of 60%.

Fig. 1
figure 1

SEM images of the initial sample surface

Full size image

2.2 Experimental Methods

According to the serving conditions of composites, life cycle test was carried out at load of 20 MPa (40 kN), frequency of 8 Hz and oscillating angle of ± 3.2°. The structure of tribotester is shown in Fig. 2. The friction pair used shaft-pad structure. The surface-to-surface contact mode can not only simulate the friction form in the spherical plain bearing, but could also eliminate the interference of systematic factors such as the structure and assembly of the bearing on the wear performance of the material. Friction pair was driven by oscillating cylinder mounted on the main shaft, with the oscillating frequency from 0 to 30 Hz. Torque and torque angle were measured by torque sensor and photoelectric encoder, respectively. The loading was completed by a hydraulic cylinder under the friction pair through a closed-loop control, ranging from 10 kN (5 MPa) to 400kN (200 MPa). It could ensure a loading force accuracy of ± 1% during oscillating friction. Friction temperature was measured in real time by a contact thermocouple at the position with a distance of 0.5 mm from the surface. Noted that the friction temperature mentioned in the study is actually the temperature close to the friction surface, which is only used as the reference temperature of the contact surface. The wear depth and deformation of friction pair were measured by displacement sensor.

Fig. 2
figure 2

Structure of the tribotester

Full size image

At the initial stage, pre-grinding was performed for 30 min to ensure the material surface being in a good lubricating state and reaching an accuracy of measurement. During test, the friction coefficient, friction temperature and deformation with oscillating cycles were monitored and recorded in real time. Wear debris were collected at an interval of 25-50 h. The residual thickness of the composite would also be measured, taking points from 6 to 36 mm in the width direction and from 6 to 48 mm in the length direction, with an interval of 6 mm. Therefore, a total of 48 points were measured. When the friction coefficient exceeds the threshold value of 0.18 or the friction temperature monitored exceeds safety threshold of 160 °C, the tribotester will stop automatically. But it should be noted that the sample may have failed before that.

3 Results

3.1 Friction and Wear Performance

The self-lubricating property of the Kevlar/PTFE fiber-reinforced composites can be characterized by friction coefficient and friction temperature, as shown in Fig. 3. The friction coefficient is obviously higher at the initial stage (0 ~ 28 h) with a mean value of 0.113 and fluctuates greatly. During this stage, the friction coefficient has a notable periodic decline and the self-lubricating performance of the material is extremely unstable. Meanwhile, friction temperature increases gradually at the initial stage and changes with the fluctuation of friction coefficient. It is indicated that the friction coefficient can also synchronously reflect the self-lubricating properties of friction surface. In terms of action mechanism, the increase in friction surface temperature caused by friction heat gradually changes the matrix of self-lubricating material from amorphous glass state to high elastic state. The mechanical properties, bonding strength and structural stability of Kevlar fiber, PTFE and resin are greatly weakened, which is finally reflected in the changes of friction coefficient and wear rate. Mean value of the friction coefficient in the middle stage (100 ~ 129 h) decreases from 0.113 to 0.098, and the fluctuation amplitude decreases greatly. The friction temperature at this stage is maintained at 129 °C, and the self-lubricating performance of the material is apparently improved. The friction coefficient starts to increase after 200 h, and the friction temperature has exceeded 148 °C. This indicates that the self-lubricating effect of the material decreases and the wear is severe. At 236 ~ 242 h in the later life, the friction coefficient fluctuates greatly with the average value as high as 0.162, and the friction temperature reaches 158 °C and shows an upward trend. It indicates that the material has entered the failure stage and it is found that local wear has occurred through the surface of the sample after shutdown. Although the wear of the material is up to 242 h, the self-lubricating effect and reliability of the material have been in an unstable state in the later life. So the actual service life of this material must be less than 242 h.

Fig. 3
figure 3

Friction coefficient and friction temperature in life cycle

Full size image

At the initial stage (0-28 h, 100-129 h, 200-219 h), friction coefficient and friction temperature change with time similarly. Seen from Fig. 3, the time for the friction coefficient at each stage to enter the dynamic stability is faster than the friction temperature. This is because that the former is mainly up to the time of stable transfer film formed, whereas the latter mainly is decided by the time of the friction heat generated, accumulated, dissipated and reached to an equilibrium. The transfer film contains PTFE adhesive layer and resin adhesive layer. Initially, the transfer film on the counterpart has not been fully formed, most of which are resins with weak adhesion, and the content of PTFE with good adhesion is very small. Due to the large regional mobility of the attached resin, the performance of the formed transfer film is poor, so the lubrication effect at this stage is unstable. When the wear is accumulated to a certain extent, friction heat tends to be stable and the corresponding relation with the fluctuation of friction coefficient decreases (236-242 h).

In order to analyze the variation characteristics of friction coefficient and friction temperature in the whole life cycle, average value is taken every 25-50 h as the reference value of this period, as given in Fig. 4. The initial friction coefficient and friction temperature are higher and fluctuate greatly. This is the running-in stage. In the middle stable stage, the friction coefficient is low with a small fluctuation, and the friction temperature begins to increase. In the later failure stage, the friction coefficient and friction temperature increase synchronously and fluctuate greatly. Based on the friction coefficient, the self-lubricating performance of the material can be divided into three stages: running-in, stable and failure stage. According to the friction temperature, it only includes stable and failure stage. Seen from previous test results, when the friction coefficient exceeds 0.15, the material enters to the failure stage. When it reaches 0.16, it indicates that the material is serious damaged and the self-lubricating performance is invalid. At this time, it is determined that the material life is terminated. In addition, it is also judged that the material failure occurs when the friction temperature exceeds 150 °C.

Fig. 4
figure 4

Periodic mean change of friction temperature and friction coefficient

Full size image

The residual thickness of the specimen is measured at an interval of 25-50 h during wear process. The values of 6 points in the width direction are averaged, and the variation of wear residual thickness with time at 8 positions in the length direction can be obtained, as shown in Fig. 5. Initial thickness at different positions is 400 ~ 440 μm, but the difference of residual thickness at different positions begins to increase after running-in stage and the nonuniformity of surface wear is becoming more and more obvious. The wear rate in running-in stage and failure stage is obviously large, whereas it decreased slightly in the stable stage. The change of wear thickness verifies that the life cycle of the sample being consistent with the judgment of three stages. Therefore, the wear behavior and life stage can be accurately reflected by the residual thickness of the sample. In practical application, the effective and reliable service life of the material should be controlled within the stable stage. When material enters the failure stage, the spherical plain bearing may fail at any time due to fluctuation and sudden failure of self-lubricating performance, which could affect the safety of the whole machine.

Fig. 5
figure 5

Wear characteristics of specimen in whole life stage

Full size image

3.2 Analysis of Nonuniform Wear

Figure 5 shows that both wear depth and the wear areas have nonlinear characteristics and change irregularly with time. That is areas with severe or slight wear change with time. The residual thickness of different areas was measured to analyze the wear characteristics of the sample surface in different life stages, as shown in Fig. 6. The initial thickness distribution is nonuniform due to the error of sample during pasting and curing process, and this intensifies the wear obviously and reduces the wear resistance and self-lubricating effect of the material. This effect is more significant in the late life in Fig. 5. Seen from the residual thickness distribution of the sample after 53 and 96 h, it can be judged that the serious wear area has changed and the discreteness of wear loss in different areas is more obvious. The residual thickness discreteness at six time points in the life stage listed in Fig. 6 is as δ242h(39.7) > δ53h(34.4) > δ203h(32.8) > δ161h(31.8) > δ96h(29.2) > δinitial(27.7). It is indicated that the discreteness of the sample thickness is the largest at the initial stage (53 h) and after failure (242 h). When the wear reaches a stable stage, the surface tends to be flat and uniform under the reciprocating action of friction.

Fig. 6
figure 6

Residual thickness of the wear surface at different time points

Full size image

After wear failure, the residual thickness distribution of the sample shows that the wear degree of different areas is significantly different, and the reinforced fiber and matrix are worn out in the seriously worn areas. Figure 7 presents the sample and counterpart surfaces and fluorine (F) element distribution after wear failure. In Fig. 7(a), it shows an obvious worn out and spalling on the surface of the sample. The severe wear occurring in the local area is a direct reason for the final failure of the material, rather than the uniform wear failure of the whole material. Figure 7(c) presents the SEM image of the non-worn area of the sample surface, and there are a large number of fatigue spalling cracks. By comparing with the F element distribution on specimen in Fig. 7(e), it can be judged that there are still a large number of PTFE components and some exfoliated resin (mark in Fig. 7c) in the non-worn area of the sample surface. It can be seen from the surface of the counterpart ring in Fig. 7(b) that the dark areas where adhesion occurs are also unevenly distributed. Figure 7(d) shows that there is a transfer film formed from the surface of self-lubricating material to the surface of counterpart ring during friction. This could also be confirmed by the content of F element in EDS analysis listed in Table 2. In addition, from the energy-dispersive spectrometer (EDS) analysis in Table 2 and area 3 marked in Fig. 7(d), it can be found that metal particles appear on the surface of the counterpart ring, indicating that there is abrasive wear during the friction process.

Fig. 7
figure 7

The sample and counterpart surfaces and fluorine (F) element distribution after wear failure

Full size image
Table 2 EDS analysis result corresponding to the positions marked in Fig. 7 (after test) (wt.%)
Full size table

Wear debris were tested by micro-SEM analysis to reveal the wear difference of the sample in different life stages, as shown in Fig. 8. In the initial stage, wear debris is unevenly distributed and isolated, because the surface of friction pair has not been covered by a uniform and stable transfer film in the running-in stage. Some large flake debris is produced under the mechanical scraping of the oscillating motion, and this is also the reason for the severe wear and its large discreteness in the middle and late stage shown in Fig. 5. EDS analysis is done corresponding to 4 positions marked in Fig. 8, and results show that the wear products contains different amounts of PTFE (in Table 3). Because PTFE has strong adsorption and stable film-forming performance, it is main component of the transfer film. By comparing with the EDS data listed in Table 2, it can be judged that some PTFE components on the surface of self-lubricating material are transferred to the surface of the counterpart ring to form a transfer film during friction, and some are extruded out of the friction surface together with resin and reinforcing fiber as wear products. Therefore, the consumption of PTFE will inevitably affect the uniformity and stability of transfer film, resulting in the fluctuation of wear performance.

Fig. 8
figure 8

Wear debris at different time points

Full size image
Table 3 EDS analysis result corresponding to the positions marked in Fig. 8 (after test) (wt.%)
Full size table

When the uniform transfer film is formed on friction surfaces, the sample substrate is no longer in direct contact with the counterpart and is blocked by the transfer film. This effectively reduces the friction coefficient and wear. In Fig. 8, the difference of wear debris size at 100 and 150 h is obviously smaller. The wear debris has fine geometry, small size, large amount and almost no blocks, which shows that the self-lubricating and wear properties reach a stable state.

The wear debris after failure (242 h) of the material mainly contains large flaky products. Its shape is consistent with the worn out area in Fig. 7 and is also the same with the surface without worn out but having obvious cracks in Fig. 9(a). Although the large flaky debris occurs in the local area, the metal is directly worn against the reinforcing fiber due to the loss of self-lubricating material isolation between the specimen and its counterpart (Fig. 9b). Therefore, the sample will be worn out in a short time (Fig. 9c and d), and the peeling depth of the material is large at later wear stage. Figure 10 shows the 3D morphology and contour of the sample surface at the later wear stage. It can be seen that the penetration depth of the crack part causing material peeling is as high as 100 ~ 120 μm. Peeling in the later wear stage can lead to sudden failure when the residual thickness is high. So the actual service life of the self-lubricating materials should be limited to a stable period. Once entering the failure stage with large performance fluctuation (Fig. 5), even if the surface is complete, local wear failure may occur at any time, resulting in serious mechanical problems.

Fig. 9
figure 9

Characteristics of the specimen surface with local worn out

Full size image
Fig. 10
figure 10

The 3D morphology of worn surface after 242 h

Full size image

3.3 Effect of Thickness Uniformity on Wear Life

From above analysis, it can be judged that the uniformity of wear thickness of the sample has a decisive impact on the service life. During the life cycle, the wear area shows obvious nonuniformity. It is mainly due to the uneven curing and bonding of the sample, resulting in the concentration of friction contact stress as well as the inconsistent of mechanical damage. The state of the transfer film formed in different areas is different. In view of this, the curing temperature of phenolic acetal resin for the sample preparation was changed from 180 to 130 °C, curing time prolonged from 120 to 240 min, and the curing pressure maintained at 0.3 MPa. A high-precision curing mold was redesigned based on original process to match the radius of curvature of the sample holder. Life cycle test was carried out on the second batch of samples thus prepared, and the results are presented in Fig. 11 and 12.

Fig. 11
figure 11

Residual thickness distribution of worn surface at different life stages

Full size image
Fig. 12
figure 12

Wear characteristics of the improved sample in the whole life stage

Full size image

In Fig. 11(a), the surface uniformity of the sample is significantly improved after improving the curing process, and only the thickness error of the edge is still large. The overall thickness variance of different regions is decreased from 27.7 to 19.5. Figure 11(b) shows the surface of sample with residual thickness after wear failure. The thickness discreteness of worn surface is greater than that in the initial state. Because transfer film of the fiber-reinforced composites changes dynamically during wear process, the self-lubricating effects are varying in different areas on friction surfaces. Therefore, it is inevitable to cause inconsistent wear degree in different areas. However, the overall uniformity is still improved after failure, with the residual thickness variance decreased from 39.7 (in Fig. 6f) to 25.9. It shows that the consistency of the material surface is improved in the wear process.

Improving the initial uniformity of materials can effectively improve wear consistency of the surface. Figure 12 presents the change in residual thickness of the sample after re-curing in whole life stage. Serious wear occurred in the initial state, but the discreteness of wear loss at different areas is better during whole life cycle compared to that shown in Fig. 5. The reason is that the uniformity of the initial thickness of the sample is improved, which avoids the large stress concentration on the contact surface during the friction process. This could reduce wear caused by the mechanical force on the sample and could also avoid excessive temperature of the local surface, which make the wear of the overall friction surface tend to be uniform. Therefore, the stable stage can be divided into two parts, including slight nonuniform wear and serious nonuniform wear stage. In the previous stage, the self-lubricating performance of the material is extremely stable, and the wear in each contact area is basically the same. In the latter stage, the anisotropy of the composite is obvious with the increase in wear, which make the heterogeneity of the material obvious. The discreteness of wear degree in different areas increases gradually after 190 h. At this time, the friction surface changes dynamically and plays an important role during the self-lubricating process. At this time, the dynamic characteristics of the friction surface in the process of self-lubrication play a major role, resulting in the increase in the difference of wear degree in different areas. From the overall life, the wear time of the re-cured sample increased from 242 to 293 h, with an overall life increased by 51 h. The average wear rate has been decreased from 3.43 × 10−5 to 2.65 × 10−5 mm3 (Nm)−1. Therefore, the wear difference at different friction areas in the wear process could be reduced and the stable stage in the life cycle could be prolonged by improving the uniformity of the sample. It plays a significant role in improving the properties and service life of materials.

4 Discussion

From whole life wear test results, the surface of Kevlar/PTFE fiber-reinforced composites produces resin and PTFE wear debris under the mechanical force from the counterpart ring during the friction process. Under the action of van der Waals force and Coulomb force, some debris is transferred to the surface of the counterpart ring to form a transfer film, and the other happens local adhesive wear on the surface of the counterpart ring. These two parts of material are finally extruded out of the wear surface in the form of flake spalling, which is consistent with the description of transfer film in Ref 24 and 27. Obviously, the uneven surface of the material reduces the adhesion stability of the transfer film and adhesive material, resulting in the intensification of flake spalling and extrusion. In contrast, a uniform friction contact surface helps to improve the uniformity and stability of the transfer film and adhesive materials and makes the wear of materials dominated by the extrusion consumption of small wear debris, avoiding the occurrence of large spalling wear products. By optimizing the uniformity of the thickness of the self-lubricating material, the average wear rate is 2.65 × 10−5 mm3 (Nm)−1. Compared with the optimal wear rate (3.2 × 10−5 mm3 (Nm)−1) obtained by air-plasma treatment for Kevlar fabric in literature (Ref 27), the research of this study has a more significant effect on reducing the wear rate. Therefore, under the same preparation process conditions, the wear resistance of Kevlar/PTFE fiber-reinforced composites can be significantly improved by improving the uniformity of thickness of the Kevlar/PTFE fiber-reinforced composites.

5 Conclusions

The Kevlar/PTFE fiber-reinforced composites were studied under high oscillating frequency conditions, and the whole life test was performed to analyze the friction and wear properties of the material. It shows that the self-lubricating and wear performance have obvious stage characteristics, including running-in stage, stable stage and failure stage. The first two stages show the effective life of the material in practical application, accounting for about 85% of the total life.

The friction heat and friction coefficient can timely and accurately reflect the performance and life characteristics of Kevlar/PTFE fiber-reinforced composites. When the friction coefficient exceeds 0.15, the material enters the failure stage of the later life. When it reaches 0.16, the material life can be determined to be terminated. In addition, when the friction temperature exceeds 150 °C, the material failure can also be determined. Therefore, the real-time monitoring of friction temperature and friction coefficient is an effective method to judge the self-lubricating performance and life stage of materials.

The damage of Kevlar/PTFE fiber-reinforced composites has significant nonlinear dynamic characteristics. The wear degree changes with time, showing the characteristics of nonuniform wear and the dynamic characteristics of local serious wear. The former leads to the fluctuation of the self-lubricating performance of the friction pair, while the latter is easy to cause the local wear failure of the material an sudden accidents.

The self-lubricating performance and wear life of Kevlar/PTFE fiber-reinforced composites can be significantly improved by improving the curing process as well as the uniformity and bonding firmness of the matrix. In this way, the wear life of the material can be increased from 242 to 293 h, and the effective life can be extended by 41 h.