Effects of the shapes and dimensions of mullite whisker on the friction and wear behaviors of resin-based friction materials

Effects of the shapes and dimensions of mullite whisker on the friction and wear behaviors of resin-based friction materials

Author’s Accepted Manuscript Effects of the shapes and dimensions of mullite whisker on the friction and wear behaviors of resinbased friction materia...

1MB Sizes 0 Downloads 30 Views

Author’s Accepted Manuscript Effects of the shapes and dimensions of mullite whisker on the friction and wear behaviors of resinbased friction materials Zhengjia Ji, Wanyue Luo, Keke Zhou, Shuen Hou, Qifeng Zhang, Jiangyu Li, Hongyun Jin www.elsevier.com/locate/wear

PII: DOI: Reference:

S0043-1648(17)31578-8 https://doi.org/10.1016/j.wear.2018.03.018 WEA102388

To appear in: Wear Received date: 31 October 2017 Revised date: 26 March 2018 Accepted date: 26 March 2018 Cite this article as: Zhengjia Ji, Wanyue Luo, Keke Zhou, Shuen Hou, Qifeng Zhang, Jiangyu Li and Hongyun Jin, Effects of the shapes and dimensions of mullite whisker on the friction and wear behaviors of resin-based friction materials, Wear, https://doi.org/10.1016/j.wear.2018.03.018 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Effects of the shapes and dimensions of mullite whisker on the friction and wear behaviors of resin-based friction materials

Zhengjia Jia, Wanyue Luoa, Keke Zhoua, Shuen Houa, Qifeng Zhangb, Jiangyu Lic, Hongyun Jina*


Faculty of Materials Science and Chemistry, China University of Geosciences, Wuhan 430074,

PR China b

Department of Electrical and Computer Engineering, North Dakota State University, Fargo, USA


Department of Mechanical Engineering, University of Washington, Seattle, WA 98195-2600 USA


Corresponding author: Hongyun Jin Tel.: +86 13476118841. [email protected]

Abstract In this study, mullite whiskers were prepared by the molten salt method and added into the friction composites. It was of interest to examine the effects of shapes and dimensions of mullite whisker addition on the sliding friction and wear characteristics of resin-based friction materials over a range of temperatures. The whiskers with higher mullite phase contents enhanced the friction stability. In addition, the wear rate increased with increasing aspect ratio of the mullite whisker. Among all the specimens, the whiskers with an average aspect ratio of 14 demonstrated the best improvement in the tribological performance. The friction coefficient variance of the specimen was small, 7.9e-4 and 6.5e-4 during the fade and recovery process, respectively. The wear rate 4 wt%. Moreover, the worn surfaces of the composites were analyzed, revealing that the 1

whiskers with higher aspect ratios are easy to peel off from the resin matrix and abrasive wear was found to be the main wear mechanism.

Keywords: mullite whisker; aspect ratio; phase composition; friction and wear

1. Introduction In recent years, hybrid polymer-based composite friction materials are generally applied in the friction brake systems of automobiles [1]. These materials are required to possess good friction stability and moderation, noise inexistence or vibration, and an excellent wear resistance along with the minimum wear to counterparts during wear [2–4]. As a consequence, multiphase composites have been developed for an automotive brake friction material and normally consisted of an excess of 10 ingredients [5]. The friction material performance is usually affected by the type and relative amount of ingredients, which are classified into four types: binders, fillers, reinforcing fibers, and friction modifiers [6]. Among these four components, the strong effects of reinforcing fibers on the strength, stiffness, and thermal stability along with friction and wear property preservation are well recognized [7]. Metallic fibers [8, 9], ceramic fibers [10, 11], organic fibers [12], and natural fibers [13] are commonly utilized in friction materials. Therefore, effects of the type, relative amount, and length of fibers on the friction behavior of composites have been extensively investigated by different research groups. Ozturk et al. [12] studied the effect of various types of fibers such as rock-wools, ceramics, E-glasses, and steel wool fibers on the tribological property. Both the highest friction coefficient and specific wear rate were obtained with the E-glass and the steel wool fiber-reinforced 2

composites, respectively. Zhou et al. [13] synthesized carbon fiber-reinforced PA6/PPS composites (PA6/PPS-CF) and discovered that the average friction coefficient value was lower than that of free-CF PA6/PPS blend with increasing carbon fiber contents from 5% to 15%, and the wear rate of the composites tended to increase. Zhang et al. [14] reported that epoxy composites with longer carbon fibers (nominal 400 μm long) exhibited a better wear resistance and a slightly decreased friction coefficient compared to that with shorter carbon fibers (nominal 90-μm long). Among these various fibers, the mullite whiskers demonstrated a low heat conduction coefficient, a low thermal expansion coefficient, a high creep resistance, and a thermal shock resistance. Cai et al. [15] utilized mullite to improve the tribological performance of friction materials and discovered that the materials filling with mullite could reduce the sensitivity of the friction materials in both load and speed. Wang et al. [16] discovered that the ceramic-matrix friction materials mixed with both steel and mullite fibers improved the friction stability, whereas the friction coefficients increased as the mullite fiber content increased. Cho et al. [17] examined two different specimens containing 10% and 30% of mullite reinforced composites, demonstrating the similar tendency for both the sliding friction and wear tests; however, they exhibited lower wear rate in comparison to the glass fiber reinforced composites. Moreover, research demonstrated that the friction coefficients of the PTFE-based composites were strongly dependent on the crystal structure of the whiskers. In addition, for the PTFE anti-wear property improvement, the whisker-like filler improved compared to the particle-like filler [18]. A large number of studies emphasized the interaction between the dosages of mullite whiskers along with the macroscopic feature of fillers and friction, and wear behaviors, whereas 3

only a few studies focused on the effect of the whisker geometry such as the aspect ratio and the phase composition on the friction and wear properties. In this study, the aspect ratio and phase composition were controlled by changing the condition of synthesis, and the mechanism of the whisker morphology and phase composition alteration effect on the friction and wear behavior was investigated. For this purpose, the friction material specimens with various mullite whiskers were prepared, and the friction tests were performed. The effect of mullite aspect ratio and phase compositions on the friction stability and the wear resistance of the friction materials were evaluated and discussed. 2. Experimental 2.1 Mullite Synthesis Mullite whiskers were fabricated by the molten-salt growth method in the presence of silicon dioxide (SiO2, analytically pure) as the silica resource, aluminum sulfate hydrate (Al2(SO4)3·18H2O), analytically pure) as the alumina precursor, and sodium sulfate (Na2SO4, analytically pure) as the molten system. First, Al2(SO4)3·18H2O was calcined at 500 C for 3 h to remove the crystal water. Then, Al2(SO4)3 and SiO2 (with a molar ratio of 3:2) along with the molten Na2SO4 salt were taken in ethyl alcohol and mixed into a ball mill for 4 h. Then, the mixture was dried completely in an oven at a constant temperature. The samples were finally heated in a muffle furnace at 900, 950, and 1000 C for 1, 3, and 7 h, respectively. Table 1 presents the calcination temperature and duration for the synthesis of the mullite whisker samples. The obtained products were washed for removing sulfate ions, which was confirmed with barium nitrate (0.01 mol/L). The phase structure of the mullite was determined using a D8-FOCUS X–ray powder diffraction analyzer (XRD) (German Brooke AXS co., LTD). At last, the morphology of 4

the mullite whisker and worn surface were characterized by ultra high-resolution field emission scanning electron microscopy (HR-SEM, SU8010).

Table 1 Synthesis conditions of mullite whiskers Mullite whisker Calcination temperature (C) Holding time (h)
















2.2 Specimen preparation The detailed ingredients are listed in Table 2. In this formulation, both the phenolic resin and the NBR were selected as the binders, the mullite whiskers and the aramid pulp as the reinforcing constituents, graphite as the friction modifier, and the barites as the fillers. Table 2 Composition of friction materials Raw material Compositions in wt%

Aramid pulp 20





Phenolic resin 15

Graphite 5

Mullite whisker 15

The friction materials were prepared by the hot press molding. All the raw ingredients were weighed with a precision of 0.01 g and mixed in a high speed blender for 3-4 times, until a uniform dispersion was obtained. The mixtures were molded for 6 min under 180 C and 10 MPa using an XL10025T hot pressing forming hydraulic machine (Wuhan Xianglong friction material equipment co., LTD). During the hot pressing, a double short release was carried out for the release of moisture and gaseous by-products because of the phenolic resin cross-linking reactions, and the composite swelling was prevented. All the specimens sustained a post-cured process in an oven at 180 C for 4 h to release the residual stress and were consequently polished to form 5

uniform and maximum contact areas for the friction performance tests. Three pairs of specimens were prepared for each whisker formulation. The mullite whiskers were labeled as N9003, N9501, N9503, N9507, and N10003, and the friction materials were labeled as N9003, N9501, N9503, N9507, and N10003. The first number after the “N” in the code is the calcination temperature in degrees (C), and the last digit in the code refers to the number of hours of holding time. For comparison, reference friction material without adding mullite whisker was also prepared (Ref). 2.3 Testing procedure The friction tests were performed using an XD-MSM constant speed tester (Xiangyang Xinyi friction sealing equipment co., LTD) in accordance with the China national standard GB 5763-2008. China national standard GB 5763-2008 is a new improved version of GB 5763/98 standard mentioned in literature [19]. The standard developed by General Administration of Quality Supervision,Inspection and Quarantine of the People's Republic of China aim to establish a friction test program of brake linings for automobiles. The testing machine schematic is presented in Fig. 1. The disk with a 185-210 HB brinell hardness maintains a constant speed. The test specimens had the common dimensions of 25×25×6 mm3. Two specimens were press-fitted to the surface of a gray cast iron rotor disk with a friction radius of 0.15 m. The applied pressure on the specimens was retained at 0.98 MPa. The surface temperature of the specimens was determined by measuring the temperature of the brake disc. The distance between the thermocouple and specimen was in the range 0.05–0.1 m. Before each test, the specimen and the disk were polished with abrasive papers to reach a surface roughness of 0.30 μm and then cleaned with acetone. In the fade testing procedure, the friction disk was rotated for 5000 revolutions at the constant testing temperatures of 100, 150, 200, 250, 300 and 350 C. The mass 6

of the test specimens was noted subsequently at each testing temperature for the wear rate calculation of each period. In the recovery testing procedure, the friction disk was rotated for 1500 revolutions at the constant testing temperatures of 300, 250, 200, 150, and 100 C. The temperature rise mainly depends on the friction heat and auxiliary heating of the heating pipe. The temperature drop mainly depends on the cooling water. Each temperature is controlled with computer as temperature feedback. The average frictional force at each testing temperature was measured to calculate the friction coefficient.

Fig. 1 XD-MSM constant-speed friction testing machine The fade ratio (at 350 C) and recovery ratio (at 100 C) were defined as follows: Rfade= μf-350/μf-max×100%


Rrecovery= μr-100/μr-max×100%


where μf-350 and μf-max are the μ at 350 C and the μ at the highest friction coefficient during the fade test procedure, respectively, and μr-100 and μr-max are the μ at 100 C and the μ at the highest friction coefficient during the recovery testing procedure. The fade fluctuations and recovery fluctuations were defined as follows: Ffade = ∑ni=1(μf − μi )2⁄n


Frecovery = ∑nj=1(μr − μj )2⁄n

(4) 7

where n is the number of temperature conditions for each test, μf and μr are the average friction coefficients during the fade and recovery processes, respectively, and μi and μj are the friction coefficient at each specified temperature during the fade process and recovery process, respectively. The wear values of the specimens were measured by the corresponding weight measurements prior to and following each temperature test using an analytical balance (Model: AL 104, Mettler Toledo, China) with a precision of 10-4. The wear rate W was calculated from the mass loss using the following equation: Wear rate (wt%) = [(W1- W2)/W1]×100%


where W1 and W2 are the W values prior to and following the test at each specified temperature during the fade testing procedure. In the tests, at least three specimens were tested for each mullite whisker composite for objective and accurate results. 3. Results and Discussion 3.1 Phase and morphology of mullite whiskers


Fig. 2 XRD spectrum of mullite whiskers under various synthesis conditions Fig. 2 shows the XRD patterns of five mullite whisker samples. The XRD pattern of samples N9503 and N9507 presented a pure mullite phase; however, that of sample N9003 exhibited impure phases resulting from silica and NaAl3(SO4)2OH6. As the temperature increased to 950 C, the peak intensity of the mullite phases increased gradually, whereas an amount of silica phase for N9501 sample remained. At 950 C, with sufficient heat preservation, the mullite was the only crystalline phase existing in samples N9503 and N9507. With further increase in calcination temperature, the mullite decomposed and an impurity phase appeared again in sample N10003.

Fig. 3 Mullite whiskers at various synthesis conditions: a. sample N9003, b. sample N9501, c. 9

sample N9503, d. sample N9507, e. sample N10003 The SEM images of the five mullite whisker samples are shown in Fig. 3. The corresponding size and aspect ratio are listed in Table 3. Fig. 3(a) shows the growth of tinny whiskers on sample N9003, presenting a cluster shape. As observed in Fig. 3(b), the whiskers with an average aspect ratio of 6.0 encircled the high-sized mass in N9501 sample. As the reaction time increased, tinny whiskers from sample N9503 gradually dispersed and eventually turned to short rod-like whiskers with the uniform size distribution as shown in Fig. 3(c), revealing a length of 1.17–3.24 μm, a diameter in the range 220–340 nm, and an aspect ratio of 14.0. The aspect ratio of sample N9507 was 23.0. Needle-like whiskers along with nano-sized graininess being scattered reduced the sample homogeneity. Regarding sample N10003 in Fig. 3(e), the overreaction resulted in the secondary growth of whiskers, leading to the formation of flakes and particles. Through comparison, the aspect ratio increased from 6.0 to 23.0 with increasing dwelling time for samples N9501, N9503, and N9507. The optimal calcination condition for the whisker synthesis was found to be 950 C for 3 h. Table 3 Mullite whisker dimensions at various synthesis conditions Specimens



900 C /3 h






950 C /1

950 C /3

950 C /7

1000 C /3






















Average Aspect ratio

3.2 Friction performance


Fig. 4 Friction coefficient variation of composites versus temperature: a. Fade process, b. Recovery process The friction tests were performed at constant sliding conditions. Three pairs of specimens were tested for each composition, and the average of three tests was reported, exhibiting error bars. The purpose of the tests was to investigate the effect of both phase composite and aspect ratio effects of the mullite whiskers on the friction coefficient alteration. Fig. 4 demonstrates the friction coefficient variation of five specimens at each temperature. Both the adhesion and the deformation resistance of the materials are known to change as a function of temperature, and the friction coefficient normally changes with the sliding temperature [8]. The friction coefficients 11

were 0.45, 039, 0.43 0.38, and 0.42 for specimens N9003, N10003, N9503, N9501, and N9507 at the beginning of the tests, respectively. At the end of the tests, these values became approximately 0.43, 0.41, 0.47, 0.41, and 0.44. As shown in Fig. 4(a), during the fade testing, the friction coefficient of specimen N9503 increased as the temperature raised from 100 C to 300 C and decreased when the temperature increased further to 350 C. The increase in the friction coefficient could be attributed to the glass transition of the friction material resin binder [20]. The decrease in the friction coefficient resulted from the heat fade effect caused by the friction force decrease at an elevated temperature and is closely related to the binder resin thermal decomposition [21]. In contrast, the curve diagram of N9003 specimen dropped vigorously at 250 C, demonstrating a noticeable fade at lower temperatures. This resulted from non-uniformity of the whiskers, increased the stress concentration, thereby deteriorating the mechanical property of composite. The scenarios of N9507 and N10003 specimens are similar, because mullites are not easily dispersed in the matrix. N9503 specimen exhibited higher fade resistance. A possible reason is that the high-purity mullite phase and the well-dispersed whisker morphology changed the glass-to-rubber transition temperature. Regarding the recovery process in Fig. 4(b), the friction coefficient of all the specimens initially increased and then decreased as the temperature decreased further, except specimens N9503 and N10003. The recovery performance was affected by both the surface layer morphology and wear debris alteration [22]. At a higher surface temperature, the increase in the friction coefficient was primarily correlated to the wear debris formation, resulting in the counterpart disc scraping. The decreased friction coefficient with decreasing temperature resulted from the change in the rheology between the surface layer and the wear debris. 12

Fig. 5 Effects of mullite whisker on the friction coefficient, fade ratio, and recovery ratio variance: a. Variance of μ at the fade process and fade ratio, b. Variance of μ at the recovery process and recovery ratio In order to better understand the friction coefficient changes, examining the friction coefficient, fade ratio, and recovery ratio variance is necessary. The results are presented in Fig. 5. As shown in Fig. 5(a), the order of the friction coefficient stability at the fade process was: N9507> N9503> N9003> N10003> N9501. The μ variance was 7.9e-4 for N9503 specimen and did not appear to be significantly different compared to the most stable N9507 specimen. In addition, the order of the fade ratio of the specimens is as follows: N10003> N9503> N9507> N9501> N9003, whereas only very small distinction between N10003 specimen (92%) and N9503 specimen (93%) existed. It was discovered that the friction stability of N9503 specimen improved more than those 13

of the other four specimens. As presented in Fig. 5(b), the order of stability of friction coefficient at the recovery process is as follows: N9503> N10003> N9003> N9507> N9501. Furthermore, the recovery ratio is as follows: N9503> N10003> N9507> N9003> N9501. The μ variance and recovery ratio were 6.5e-4 and 88% for N9503 specimen, respectively, demonstrating the best friction property at the recovery stage. The ratios in Table 3 show that the friction coefficient stability increases in the beginning and decreases subsequently with increasing aspect ratio. N9503 specimen demonstrated the improved recovery properties and friction stability over the other specimens during the recovery testing. The dispersed histological structure and the non-existent agglomeration of N9503 sample mullite whiskers, presented in Fig. 3(c), could be responsible for the friction coefficient and high stability. Homo-dispersed structure is believed to be beneficial to the bond between the whiskers and the resin matrix, and thus strengthens the toughness. N9501 specimen presented the maximum fluctuation throughout the testing stage, which might be attributed to the presence of impure silica graininess in the mullite whiskers and the irregular lumps in the whisker structure, weakening the whisker reinforcement effect (Fig. 3(b)). The mullite and silica thermo-mechanical data indicated that both the breaking tenacity and breaking strength of the mullite toughening ceramic composite were higher than the corresponding properties of silica reinforced composite [23], apart from the property of NaAl3(SO4)2OH6. These impure phases weakened the friction stability, caused by the poor combination with the resin and the lower compressive property compared to the mullite whiskers. Compared to the other samples [15], the friction material with poor compressive properties might be quite more sensitive to the heat increase, causing an apparent friction coefficient fluctuation. This phenomenon could result from the added-in whiskers, with various aspect ratios and 14

phase impurities. Fig. 3(a) shows that the inadequate reaction temperature of 900 C caused the cluster structure, whereas the whisker anisotropic growth also displayed a counteraction on the friction property. N9501 sample whiskers presented a cluster structure mostly, combined with the high size of the block materials, known as the silica impure phase (Fig. 2). As the reaction heat increased, the lengthening and widening of the whiskers became apparent. The homogeneously dispersed structure of N9503 sample with an aspect ratio of 14 and the best phase purity absolutely presented the best reinforcement in the friction property. In contrast, as the aspect ratio (sample N9507) increased and the mullite composition results in the appealed impurity phase (sample N10003), the specimen friction coefficients became significantly unstable. 3.3 Wear performance The weight loss measurements were conducted following 5000 revolutions of each sample at each temperature. Fig. 6 presents the effects of the mullite whiskers on the wear rate of the friction material composition. Fig. 6(a) clearly shows that the wear rate increased with increasing temperature in the beginning as a result of temperature increase on the sliding interface. The trend of downward fluctuations occurred in the range 200–300 C and could be attributed to both the broken whiskers and the detached resin particles, which were compacted forming the frictional contact area with a good wear resistance. Subsequently, the friction material was protected from wear [24]. As the heat increased further, the matrix decomposed and invalidated, consequently the severe softening and detachment increased. Fig. 6(b) shows that N9003 specimen gives the lowest sum wear rate and N10003 specimen achieves the highest. A possible reason might be due to the nanoparticles encircling the whiskers as shown in Fig. 3(e), accelerating the wear during the friction. Regarding specimens N9501, 15

N9503, and N9507, the sum wear rate increased with the elevated aspect ratio in the range 6.0– 23.0. The possible reason is that the impact strength of longer whiskers weakened in comparison to the shorter whiskers, with the same volume content, whereas the shorter whiskers had an increased amount of whisker ends, indicating a higher number of primary plateaus, which were both mechanically stable and wear resistant [25]. For specimen N10003, the whisker morphology reveals a high number of flake particles, giving rise to the poor interfacial adhesion between the fibers and the matrix. Moreover, the hard asperities and the broken fibers presented on the counter surface film increased the friction thrust, absolutely amplifying the separation and fracture of the reinforcing fibers. This phenomenon could increase the specimen wear rate because of the counterface film damage.


Fig. 6. Variation in the composite wear rates with elevated temperature. (a) Wear rate at each temperature regime, (b) wear rate sum of each sample during the fade process 3.4 Worn surface analyses The friction and wear behavior of the friction materials were closely related to the corresponding worn surfaces [16]. In order to understand the mullite whisker effects on the wear mechanism, the friction worn surfaces were observed, as displayed in Fig. 7. Figs. 7(a) and (b) are the worn surface micrographs of specimen N9003. Compared to the other composites, specimen N9003 had a low quantity of fine wear debris with smooth high-area secondary plateaus on the surface, indicating a high friction coefficient value with an excellent stability. The grooving presented as a surface damage phenomenon associated to the surface fatigue. This might be due to the formation of hard particles from the resin wear debris trapped on the surface. Specimen N9501 composite displayed the fine debris covered on the worn surfaces with a low number of primary plateaus, as shown in Figs. 7(c) and (d). Besides, resin definitely deformed to a certain extent, due to the deterioration in the whisker/matrix bonding, as presented in Fig. 7(d), related to the poor cohesive action between the agglomerated whiskers and the resin. Certain grooves have also been observed in the micrographs of samples N9501, N9503, and 17

N9507, indicating abrasive wear as the main wear mechanism. Figs. 7(e) and (f) demonstrate the specimen N9503 worn surface, with a significant amount of crazing and secondary plateaus. The presence of crazing was thought to arise from the form of localized surface fatigue that occurred in brittle polymers, characterized by low-sized crazes [26]. Moreover, the debris piled up and sustained a compaction by the shear forces and the normal pressure and the friction heat, forming secondary plateaus [24, 27]. Regarding N9507 sample in Figs. 7(g) and (h), it could be clearly observed that the surface fatigue and a high amount of crazing existed encircling the contact surfaces, having been caused by low-cracks cracks where the two edges bridged by the nanometric fibrils. The latter could lead to the mass removal from the surface, which was considered as a particular wear mechanism in polymers. In addition, the fibrils have been reported to be an important surface damage phenomenon in relation to the plastic deformation for brittle polymers. As a candidate of the highest wear rate, specimen N10003 severely damaged as shown in the SEM images in Figs. 7(i) and (j), displaying a large area surface fatigues and resin decompositions, whereas the surface fatigue presence in this type of tests might be because of compressive and tensile stress suffered by the material surface [28]. Plastic deformation probably occurred due to the counter-body asperities pressure. The pressure directly increased the roughness of counterface, elevating the specific wear rate [29]. The main wear mechanism was the fatigue wear. Both the particles and the block structure of the whisker morphology might also be the possible causes for the fatigue mechanism.


Fig. 7 SEM micrographs of composite worn surfaces: specimen N9003 (a, b); Specimen N9501 (c, d); Specimen N9503 (e, f); Specimen N9507 (g, h); Specimen N10003 (i; j) 19

Furthermore, as revealed by the morphology of all the specimens, the area of the actual contact was confined within the contact plateaus. The friction increase was related to the formation of primary plateaus, consisting of broken whiskers and hard particles embedded onto the surface of the composites [25], whereas the smooth protruding patches on the surfaces represented the secondary plateaus. When the composite rough surfaces were worn, the primary plateaus formed as a result the actual possible contact area between the composites and the counterfaces increased [30]. An increased area of actual contact was believed to be a result of increased friction coefficient. The nature of the typical wear debris consisted of the sheared-deformed polymer matrix, containing low-sized broken and pulled-out whisker elements along with the metallic counterpart wear powder. The particles could either be removed from the contact zone immediately, breaking the composite surface or remain localized as the layers transferred. The composites demonstrating lower wear resistance (specimens N9507 and N10003) produced quite extensive fine spherical and certain plate-shaped wear particles, along with the loose and uncompacted contact plateaus of the worn surface, which could be easily peeled off from the worn surface.

Conclusions In this study, various mullite whiskers were synthesized and the friction and wear behaviors of mullite whiskers reinforced resin friction materials were studied. Based on the aforementioned results, the following conclusions are summarized: 1. The whiskers with higher mullite phase contents enhanced the friction stability of composite. The existence of impurities such as silica and NaAl3(SO4)2OH6 was detrimental for the 20

friction stability, whereas the toughening and reinforcing effects on the resin were weaker than the corresponding effects on mullite, reducing the resin decomposition and surface fatigue quite easily. 2. As the whisker aspect ratio increased from 6.0 to23.0, the wear resistance of resin friction materials decreased. The shorter whiskers had a higher number of whisker ends and therefore a higher number of primary plateaus, which were mechanically stable and wear resistant. The whiskers with excessive aspect ratios were easy to peel off from the resin matrix, making abrasive wear as the main wear mechanism. 3. For the friction and wear performances of resin friction material with mullite whisker, the optimal average aspect ratio was 14.0. The material formulation with high aspect ratio and favorable dispersibility in the resin matrix (treated for 3 h at 950 C) had the best frictional stability and wear resistance. 4. Both the grooving and fine wear debris appeared on the wear surface, indicating that the main wear mechanism was the abrasive wear, whereas for the friction material with mullite flakes and particles, the main wear mechanism was the fatigue wear because of the plastic deformation and surface fatigue. Overall, it was concluded that shapes and dimensions of mullite whisker in friction materials affected the friction and wear performance significantly, however, the physical and mechanical properties of brake friction materials was outside the scope of this study and will be reported in the near future.

Acknowledgments 21

The authors would like to acknowledge the financial support from the National Key Research and Development Program of China (2016YFA0201001) and the National Natural Science Foundation of China (11627801).

References [1] J. Bijwe, Composites as friction materials: Recent developments in non-asbestos fiber reinforced friction materials—a review, Polymer Composites, 18 (1997) 378–396. [2] B. Satapathy, J. Bijwe, Fade and recovery behavior of non-asbestos organic (NAO) composite friction materials based on combinations of rock fibers and organic fibers, Journal of Reinforced Plastics and Composites, 24 (2005) 563–577. [3] R. Hinrichs, M.R. Soares, R. Lamb, M.R. Soares, M. Vasconcellos, Phase characterization of debris generated in brake pad coefficient of friction tests, Wear, 270 (2011) 515–519. [4] S.J. Kim, J.Y. Lee, H. Jang, Effect of humidity on friction characteristics of automotive friction materials, KSTLE International Journal, 2 (2001) 150–153. [5] S.J. Kim, H. Jang, Friction and wear of friction materials containing two different phenolic resins reinforced with aramid pulp, Tribology International, 33 (2000) 477–484. [6] A. Shojaei, M. Fahimian, B. Derakhshandeh, Thermally conductive rubber-based composite friction materials for railroad brakes – Thermal conduction characteristics, Composites Science & Technology, 67 (2007) 2665–2674. [7] P. Gopal, L.R. Dharani, F.D. Blum, Fade and wear characteristics of a glass-fiber-reinforced phenolic friction material, Wear, 174 (1994) 119–127. [8] H. Jang, K. Ko, S. Kim, R. Basch, J. Fash, The effect of metal fibers on the friction 22

performance of automotive brake friction materials, Wear, 256 (2004) 406–414. [9] Z.J.Z.Y.D. Xu, L.G.N. Huawei, Effect of Steel Fiber and Iron Oxide Powder Content on Frictional and Wear Performance of Semi-metallic Friction Materials [J], Lubrication Engineering, 1 (2011) 019. [10] F. Wang, Y. Liu, Effects of Mullite Fiber Content on Friction and Wear Properties of Ceramic-based Friction Material, Cailiao Gongcheng/journal of Materials Engineering, (2012) 61–65. [11] F.H. Wang, Y. Liu, Friction and Wear Properties of Ceramic-Based Composites with a Mixture of Steel Fiber and Aluminium-Silicate Fiber, Materials for Mechanical Engineering, (2013). [12] B. Öztürk, F. Arslan, S. Öztürk, Effects of different kinds of fibers on mechanical and tribological properties of brake friction materials, Tribology Transactions, 56 (2013) 536–545. [13] S. Zhou, Q. Zhang, C. Wu, J. Huang, Effect of carbon fiber reinforcement on the mechanical and tribological properties of polyamide6/polyphenylene sulfide composites, Materials & Design, 44 (2013) 493–499. [14] H. Zhang, Z. Zhang, K. Friedrich, Effect of fiber length on the wear resistance of short carbon fiber reinforced epoxy composites, Composites Science and Technology, 67 (2007) 222–230. [15] P. Cai, Y. Wang, T. Wang, Q. Wang, Improving tribological behaviors of friction material by mullite, Tribology International, 93 (2016) 282–288. [16] F. Wang, Y. Liu, Mechanical and tribological properties of ceramic-matrix friction materials with steel fiber and mullite fiber, Materials & Design, 57 (2014) 449–455. [17] B.R. Cho, H.Y. Lee, Characterization of friction materials reinforced by porous mullite, 23

Materials science forum, Trans Tech Publications, 2006, pp. 650–653. [18] Y.J. Shi, X. Feng, H.Y. Wang, C. Liu, X.H. Lu, Effects of filler crystal structure and shape on the tribological properties of PTFE composites, Tribology International, 40 (2007) 1195–1203. [19] Y. Liu, Z. Fan, H. Ma, Y. Tan, J. Qiao, Application of nano powdered rubber in friction materials, Wear, 261 (2006) 225–229. [20] Y. Wu, M. Zeng, Q. Xu, S. Hou, H. Jin, L. Fan, Effects of glass-to-rubber transition of thermosetting resin matrix on the friction and wear properties of friction materials, Tribology International, 54 (2012) 51–57. [21] Z. Ji, H. Jin, W. Luo, F. Cheng, Y. Chen, Y. Ren, Y. Wu, S. Hou, The effect of crystallinity of potassium titanate whisker on the tribological behavior of NAO friction materials, Tribology International, 107 (2017) 213–220. [22] M. Kumar, J. Bijwe, NAO friction materials with various metal powders: Tribological evaluation on full-scale inertia dynamometer, Wear, 269 (2010) 826–837. [23] H. Schneider, J. Schreuer, B. Hildmann, Structure and properties of mullite—A review, Journal of the European Ceramic Society, 28 (2008) 329–344. [24] E.J. Lee, H.J. Hwang, W.G. Lee, K.H. Cho, H. Jang, Morphology and Toughness of Abrasive Particles and Their Effects on the Friction and Wear of Friction Materials: A Case Study with Zircon and Quartz, Tribology Letters, 37 (2009) 637–644. [25] B. Öztürk, S. Öztürk, Effects of resin type and fiber length on the mechanical and tribological properties of brake friction materials, Tribology Letters, 42 (2011) 339–350. [26] C.E. Correa, S. Betancourt, A. Vázquez, P. Gañan, Wear resistance and friction behavior of thermoset matrix reinforced with Musaceae fiber bundles, Tribology International, 87 (2015) 24

57-64. [27] B.K. Satapathy, J. Bijwe, Composite friction materials based on organic fibres: Sensitivity of friction and wear to operating variables, Composites Part A Applied Science & Manufacturing, 37 (2006) 1557–1567. [28] M. Terheci, Microscopic investigation on the origin of wear by surface fatigue in dry sliding, Materials characterization, 45 (2000) 1–15. [29] B. Vishwanath, A.P. Verma, C.V.S.K. Rao, Friction and wear of a glass woven roving/modified phenolic composite, Composites, 21 (1990) 531–536. [30] J. Bijwe, N. Mazumdar, Influence of amount and modification of resin on fade and recovery behavior of non-asbestos organic (NAO) friction materials, Tribology Letters, 23 (2006) 215–222.


The mullite phase contents of whisker were studied on the influence of friction stability.

The aspect ratio of whisker was studied on the influence of wear resistance.

The pure mullite whisker with high aspect ratio showed improved friction stability and wear resistance.