Frictional materials used in automobile brake linings are multifarious composite materials, and they are well-known as: i) High-steel (semi-metallic) brake pads containing 30% - 65% metal, ii) Low-steel (low-metallic) brake pads containing 10% - 30% metal, iii) Organic brake pads (also known as “non-asbestos organic” (NAO)) and iv) Hybrid brake pads, being a compromise between materials from group’s ii and iii. High-steel and low-steel friction materials possess several disadvantages such as tendency to corrosion, low thermal stability, uneven wear of brake disk, etc., have restricted their braking applications. Modern friction materials familiarly known as non-asbestos organic (NAO) made of thermoset composites have been utilized extensively, starting from bicycles, light commercial vehicles, heavy vehicles, airplanes, etc.    . These friction materials are a mixture of several ingredients, which includes many fillers, lubricants, friction modifiers and reinforcing fibers bonded together by a thermosetting resin     . Of several kinds of reinforcing fibers as support in polymer matrix composites, fibers made of glass, carbon, aramid and so on are broadly used. They are categorized by their aspect ratio. Polymers are further can be strengthened with different fillers that are accessible normally or synthesized in many forms such as, flakes, platelets, particles and so on to enhance their processability, mechanical, tribological and other performance, and in addition to reduce material cost. Filler particles of nano size with optimum loading percent have yielded the outstanding and synergistic performance in several characterization process. Many researchers have carried out the research on friction materials and the details of their research works has been discussed in Table 1.
Also, many investigations are made on tribological characterization aiming to evaluate fade and recovery behaviour of friction materials. However, abrasive wear from loose debris which is formed due to high pressure application need to be studied considering influential parameters like load, abrading distance and abrasive particle size. Hence, in view of cited literature above, the objective of the present work is to study the abrasive wear behaviour of phenolic friction materials. In particular, the three-body abrasive wear (3-BAW) behaviour of short carbon fiber reinforced phenolic friction composites with varying wt% of micro and nano-fillers, using Taguchi design of experiments and ANOVA to understand the control factors and their contributions affecting the wear characteristics.
2. Experimental Details
The source of the materials used to prepare the micro and nano fillers filled phenolic polymer composites in the present investigation are presented in Table 2.
2.2. Fabrication Method
The details of the fabrication method followed in the present work are discussed elsewhere  . The details of the hybrid composites prepared are listed in Table 3.
2.3. Three-Body Abrasive Wear Test
The phenolic friction composite in automotive application as brake pads encounters the metallic surface of the drum during braking. This results in the generation of debris leading to the peel out of the fillers from the composite brake pads. These fillers act as the third body at the interface of the brake pad and the metallic drum surface. Hence the study of three-body abrasive (3-BAW) behaviour of phenolic friction composites is worth to discover. These tests were carried out using Magnum make Rubber wheel abrasion tester (RWAT) in accordance with ASTM G-65-16  . Figure 1 shows the photograph of RWAT
Table 1. Literature review on polymeric friction materials.
Table 2. Materials used in present investigation
Table 3. Designations and constituents of thecomposites for Set-I and Set-II
apparatus used for the 3-BAW tests. The size of the test samples are maintained to 75 mm × 25 mm. The test procedure followed in the present work is as discussed elsewhere   , further the specific wear rate (Ks) are determined as mentioned in the reference  . The parameters for conducting the 3-BAW tests are listed in the Table 4.
2.4. Statistical Tool for Wear Characterization
The 3-BAW routine experiments were conducted for the test parameters listed in the Table 4. However, the Ks were found to be significant with applied normal load of 15 N and abrasive particle size of 300 µm. Hence they are considered as the constant factors in the present study. Further to understand the effect and contribution of abrading distance and filler content on Ks, they were considered for statistical analysis. Significant Ks was found at the abrading distances of 280, 570 and 1140 m, hence abrading distance at these levels were considered for statistical analysis. The control factors and levels listed in Table 5, are used for statistical analysis of wear. An orthogonal array (OA) L9 was chosen and the factors considered affecting the wear process are abrading distance (A) and filler
Figure 1. Dry sand rubber wheel abrasion test rig (RWAT).
Table 4. Test parameters considered for routine 3-BAW.
Table 5. Abrasive wear control factors and levels.
content (B). The experimental data obtained are transformed into signal-to-noise (SN) ratio by considering the minimization of Ks. Analysis of variance (ANOVA) is used to reveal the level of significance of factors influencing wear. The percentage contribution by each of the process parameter in the total sum of the squared deviations can be used to evaluate the importance of the process parameter change on the performance characteristic. If the P-value (probability of significance) for a factor in the table is less than 0.05 (95% confidence level) then it can be considered that, the effect of the factors is significant on the response.
3. Results and Discussion
3.1. Three Body Abrasion Wear Study
The abrasive wear behaviour of phenolic friction composites are determined on Rubber wheel abrasion tester (RWAT). Figure 2 demonstrates the 3D surface plot with filler content in X-axis, abrading distance in Y-axis and Ks in Z-axis, revealing the combined effect of filler content and abrading distance on wear behaviour of the phenolic friction composites.
The Ks decreases with the increase in abrading distance and with the inclusion of MWCNT along with molykote (Set-I series) and nano clay along with graphite (Set-II series), maintaining almost the same trend with marginal difference in Ks. However, the Ks of Set-I series is comparatively lesser than that of the Set-II series. The role of filler content can be observed in the plot, which results in the significant reduction of Ks. The Ks was found to be high without the nano fillers and inclusion of the same upto 0.5 wt%, has resulted in the decrease of Ks. However, routine experiments were conducted with the phenolic composites loaded with higher concentration of (i.e. 0.5 wt% to 1.0 wt%) nano fillers. It was
Figure 2. Surface plot revealing the combined effect of filler content and abrading distance on Ks of Set-I composites.
observed that the Ks increased beyond 0.5 wt% loading of nano fillers. The similar findings were observed by other researchers   . Hence in the present analysis, the composite with 0.5 wt% nano filler loading was considered.
3.2. Worn Surface Morphology
Figure 3 presents the SEM image of worn surface of polymeric friction composites (PFC) filled with 11.5 wt% of molykote + 0.5 wt% of MWCNT subjected to 15 N applied normal load, abraded to a distance 1140 m with abrasive particle size of 300 µm. This composite has demonstrated high resistance to abrasion wear in the study group. Very few filler pull-out (indicated as FP), fiber pull-out (indicated as SP) and fiber rupture (indicated as SR) can be seen in the Figure 3. This has resulted in low wear volume of the composite in the study group. Good bonding of matrix with the fiber is evident from the image resulting in improved abrasive wear resistance.
Figure 4 presents the SEM image of worn surface of PFC filled with 12 wt% graphite subjected to 15 N applied normal load, abraded to a distance 1140 m
Figure 3. SEM image of worn surface of PFC filled with 11.5 wt% molykote + 0.5 wt% MWCNT.
Figure 4. SEM image of worn surface of PFC filled with 12 wt% graphite.
with abrasive particle size of 300 µm. Inclusion of graphite has deteriorated the hardness of the composite resulting in low resistance to the abrasion wear. The matrix material (indicated as M) has been abraded and the short carbon fiber have been exposed to abrasive particles resulting in fiber rupture (indicated as SR), fiber pull-out and filler pull-out. This has resulted intense abrasion wear loss.
3.3. Statistical Analysis of Three-Body Abrasive Wear Data
Wear parameters that is abrading distance and filler content are considered as controlling factors at three different levels. Table 5 shows the control factors and their levels considered for the experimentation. The L9 orthogonal array (OA) of experiments along with the experimental wear responses are shown in Table 6. The responses was analyzed to obtain signal to noise ratio, using the MINITAB 17 software, specifically used for the design of experiments (DOE) applications. The mean of signal to noise ratio was found to be 33.8783 dB for Set-I and 32.766 dB for Set-II composites respectively.
Figure 5 shows the graphs denoting the effect of the control factors on the Ks of Set-I and Set-II composites respectively. Process parameter settings with the highest SN ratio always gave the optimum quality with minimum variance. The graphs show the change of the SN ratio when the setting of the control factor was changed from one level to the other. The best Ks were at the higher SN ratio values in the response graphs. From the graph, it is clear that control factor combination of A3 and B3 gives minimum Ks. Thus, minimum Ks for the developed composite materials are obtained when the abrading distance (A) and filler content (B) are at the highest level. The SN ratio response of Set-I and Set-II composites is presented in Table 7. The SN ratio delta values of A and B for Set-I composites are 2.65 and 8.35: and for Set-II composites are 2.65 and 9.21 respectively. The strongest influence on the Ks was shown by factor B, followed by factor A, in both material groups under study.
Table 6. OA of experiments, responses and corresponding SN ratios.
Figure 5. Plot demonstrating the main effects for SN ratio, (a) Set-I and (b) Set-II.
Table 7. Response table for SN ratio of Set-I and Set-II composites.
3.4. Analysis of Variance and the Effects of Control Factors
The ANOVA with Ks results are listed in Table 8 for Set-I and Set-II composites respectively. This analysis was undertaken for a level of significance of 5%, that is, for level of confidence 95%. The last column of the table indicates the order of significance among control factors. It could be observed from Table 8 that the control factor B (P value = 0.000) has greater static influence of 86.806% and A (P value = 0.013) has an influence of 11.6687% on Ks of the material system under study. Also for Set-II from Table 8 the control factor B (P value = 0.000) has greater static influence of 87.8719% and A (P value = 0.016) has an influence of 10.5838% on Ks of the material system under study. The present analysis shows that 3-BAW test parameters that are, abrading distance and filler content have both statistical and physical significance.
3.5. Confirmation Test
The confirmation test is the final step in the DOE process. The purpose of the confirmation test is to validate the conclusions drawn during the analysis phase. The estimated SN ratio for Ks using the optimum level of parameters are calculated as discussed elsewhere   .
The results of experimental confirmation were carried out by comparing the predicted Ks with the actual Ks using the optimal wear parameters are shown in Table 9. The change in the predicted to experimental results is about 1.41% (Set-I) and 1.19% (Set-II), which is well within the statistical confidence level. Therefore the Ks of the friction composite material under study can be predicted with an allowable limit of 1.41% and 1.19% respectively.
Table 8. Analysis of Variance for Ks.
DF: Degree of Freedom, Seq SS: sequential sum of squares, Adj MS: adjusted mean squares, F: variance P value: test statistics, PC (%): percentage of contribution.
Table 9. Confirmation test for the tested friction composite.
・ Inclusion of 11.5 wt% of MK and 0.5 wt% of MWCNT in PF composites exhibited highest abrasion resistance among the composites under study.
・ Nano fillers had beneficial effect on the abrasion behaviour of PF composites under study.
・ Filler concentration played a vital role with a contribution of around 87% in 3-BAW behaviour of the PF composites.
・ It is observed that minute agglomerates facilitated in forming a network which helped in improving the abrasion wear resistance.