1. Introduction

An informal number of papers dealing with the tribological behaviors of polymer materials have been published. That is why the polymers are extending over a great area used in sliding components like such as gears, cams, breaks, clutches, bearings, wheels and bushes. Adhesive wear includes galling, fretting, scuffing and surface fatigue. This refers to the damage produced when two mating surfaces move relative to each other under a normal load. Surface asperities interact and very high stresses, strain, and strain rates are generated in localized regions[1]. This type of wear occurs in bearings, piston rings, cylinders and in electrical contacts. In recent years attention has been focused on the sliding wear behavior of polymers and their composites due to their increasing use as bushings and seals in machinery.

The research by various authors[2-4] reported that the friction between polymers can be attributed by the two main mechanisms i.e, deformation and adhesion. The

From the point of view of the coefficient of friction, Brentnall and Lancaster[6] reported that the friction coefficient of polymers rubbing against metals decreases with the increase in load. Some researchers[7-9] reported that the coefficient off friction value increases with the increase in load. Finally, Byett and Allen[10] and Friedrich et al.[11] have reported that the coefficient off friction value increases with the increase in load. Many polymeric materials have an excellent strength-to-weight ratio, good corrosion resistance, wider choice of the materials, dimensional stability, high impact strength, light weight and ease to manufacturing. Some polymers also possess excellent tribological properties[12]. The polymers can be considered to be one of the competitive materials for tribological applications because of their low friction values against steel counterparts, good damping properties, and self lubricating abilities.

It has been observed that by assimilating filler particles in the polymer based composites, synergistic effects may be achieved in the form of higher modulus and reduction of the material cost[13-15]. The inclusion of such particles into polymers for commercial applications is focused at the cost reduction and stiffness improvement[16, 17]. Various kinds of polymers and polymer-matrix composites with different types of fillers such as Al₂O₃, SiC, TiO₂, fly ash etc. have been the subject of extensive research in recent years as found from the literature[18-21]. But the potential of particulate fillers CaCO₃ and CaSO₄ in vinyl ester matrix has not been reported so far. So, in this paper we are using these fillers (CaCO₃ and CaSO₄) with vinyl ester matrix and studying the effect of such fillers on mechanical and dry sliding wear behavior of the composites. Polymer composites containing different fillers and/or reinforcements that are frequently used for applications like automotive parts, gear assemblies, tub/ shower industries etc. in which friction and wear are critical issues. Calcium carbonate (CaCO₃) and calcium sulfate (CaSO₄) are the fillers, which are used in the automobile parts and tub/ shower industries respectively. However, study of the effect of such filler addition is necessary to ensure that the mechanical properties of the composites are not affected adversely by the addition of such fillers.

The importance of mechanical and tribological properties has convinced many researchers to study the friction and wear behaviour and to improve the wear resistance of polymer based composites. From the above mentioned literature it is understood that there is tremendous potentials of these fillers used for research on vinyl ester composites and studied under dry sliding conditions. Since, the purpose of this paper is to study the mechanical and dry sliding wear behaviour of fillers CaCO₃ and CaSO₄ filled vinyl ester composites. Taguchi method is used to optimize the process parameters of sliding wear in order to reduce the number of experiments without sacrificing the information.

2. Experimental Details

2.1. Specimen Preparation

The matrix used in this work is vinyl ester resin (density 1.28 gm/cc) was supplied by Northern Polymer Pvt. Ltd. New Delhi. Methyl ethyl ketone peroxide (MEKP- 1.5%), Cobalt Naphthenate (1.5%) was used as catalyst and accelerator respectively. Three different types of composites are prepared for the study. The Calcium Sulfate (CaSO₄) and Calcium Carbonate (CaCO₃) having particle size 1.813 µm and 1.620 µm respectively were used as a filler material collected from the Pioneer Chemical Corporation, Delhi.

The composites are made homogeneously. Firstly they are properly sterilized in a jar and then simply pour into the 10mm diameter test tubes. There are three different types of composites are made for the current study with 0, 10 and 20 wt.% of the filler content. The accelerator Cobalt Naphthenate 1.5% is mixed thoroughly in vinyl ester resin and then catalyst 1.5% Methyl ethyl ketone peroxide (MEKP) was mixed in the resins prior to reinforcement. Before pouring the composite solution in the test tube, the test tube is sprayed with a release agent (Silicon spray) to ensure that the part will not adhere to the test tube after the curing of the composites samples. The cast of each composite is cured for 24 hour at room temperature before it is removed from the test tube. The other composite samples with fillers CaCO₃ and CaSO₄ of fixed weights (10 wt.% and 20 wt.%) percentage were fabricated by the same technique. The fillers CaSO₄ and CaCO₃ were mixed thoroughly in the vinyl ester resin mechanically before pouring into the test tubes. The composites prepared for this study are designated as WCGV₁, WCGV₂, WCGV₃, WCGV₄, and WCGV₅ respectively. The composition and designation of the composites prepared for this study are listed in Table 1.

Table 1. Composition and designation of the composites Designation Detail composition WCGV1 Vinyl ester WCGV2 Vinyl ester +10 wt.% CaSO4 WCGV3 Vinyl ester +20 wt.% CaSO4 WCGV4 Vinyl ester +10 wt.% CaCO3 WCGV5 Vinyl ester +20 wt.% CaCO3

2.2. Friction and Wear Test Apparatus

Dry sliding wear tests were conducted on a pin-on-disc friction and wear monitoring test rig (DUCOM) as per ASTM G 99. The cylindrical pin specimens of 10mm diameter and 30mm length were tested against a disc made of hardened ground steel (EN-32, hardness 72HRC, surface roughness R_a= 0.07 µm). The specimen was held stationary and the disc was rotated while a normal force was applied through a lever mechanism. The schematic diagram of the pin-on-disc apparatus is shown in the Figure 1. During the test, frictional force was measured by the transducer mounted on the loading arm. The frictional force readings were taken as the average of 100 readings of every 40 seconds for the required time period. For this purpose a microprocessor controlled data acquisition system was used. The average mass loss was used to calculate to the specific wear rate (K_s). The tests were conducted with sliding velocity (1.57, 2.62, 3.67 m/sec.), normal load (20, 40, 60 N), filler content (0, 10, 20 %) for the sliding distance of 1000, 3000 and 5000 m. Sliding wear data reported here is the average of two runs. The initial weight before run and final weight after run is measured using a precision electronic balance with an accuracy of ± 0.01 mg. The specific wear rate (mm³/Nm) is then expressed on ‘volume loss’ basis.

Table 2. Parameters setting and levels for various control factors for wear test Control Factors Symbols Units Levels I II III Velocity A m/sec. 1.57 2.62 3.67 Normal Load B N 20 40 60 Filler Content C % 0 10 20 Sliding Distance D m 1000 3000 5000

Where, K_s is the specific wear rate (mm³/Nm), is the mass loss in the test duration in gm, is the density of the composite (gm/cm³), F_n is the applied normal load (N), L is the sliding distance (m). The parameters setting and levels for various control factors for wear test are shown in the Table 2.

Figure 1. Schematic diagram of pin-on disc apparatus

2.3. Scanning Electron Microscopy

A FEI quanta FEG450 was used to analyze the worn surfaces of the polymer composites. The composite samples are mounted on stubs with gold plating. To enhance the conductivity of the samples, thin films of platinum are vacuum evaporated onto them before the photomicrographs were taken.

2.4. Experimental Design

Taguchi design of experiment is a powerful analysis tool which is adopted for optimizing design parameters. Taguchi method provides the designer with a systematic and efficient approach for experimentation to determine near optimum settings of design parameters for performance, quality and cost[22-25]. The most important stage in the design of experiment lies in the selection of the control factors. In the present work, the impact of the four such factors are studied using L₂₇ (3¹³) orthogonal array which has 27 rows corresponding to the number of tests (20 degree of freedom) with 13 columns at three levels. The operating conditions under which sliding wear tests carried out are given in the Table 2.

In conventional full factorial experimental design, it would require 3^{4 =} 81 runs to study four factors each at three levels whereas, Taguchi’s factorial experiment approach reduces it to only 27 runs offering a great advantage in terms of experimental time and cost. The experimental observations are transformed into a signal-to-noise (S/N) ratio. There are three S/N ratios available depending upon the type of characteristics (smaller-the-better, larger-the-better, nominal-the better). The S/N ratio for minimum (friction and wear rate) coming under smaller is better characteristic, which can be calculated as logarithmic transformation of the loss function as shown below[26]

(1)

Where ‘n’ is the repeated number trial conditions and y₁, y₂………y_n are the response of the friction and sliding wear characteristics. “Lower is better” (LB) characteristic, with the above S/N ratio transformation is suitable for minimizations of coefficient of friction and specific wear rate. The standard linear graph is used to assign the factors and interactions to various columns of the orthogonal array (OA).

The plan of experiments is as follows: the first column is assigned to the velocity (A), the second column to normal load (B), the fifth column to filler content (C) and the ninth column to sliding distance (D) the third and fourth column are assigned to (A×B)₁ and (A×B)₂ respectively to estimate interaction between the velocity (A) and the normal load (B), the sixth and seventh column are to (B×C)₁ and (B×C)₂ respectively to estimate the interaction between the normal load (B) and the filler content (C), the eight and eleventh column are assigned to (A×C)₁ and (A×C)₂ respectively to estimate the interaction between the velocity (A) and the filler content (C) and the remaining columns are used to estimate the experimental error. The linear graph for L₂₇ array is shown in the Figure 2.

Figure 2. Linear graphs for L27 array

3. Results and Discussion

The characterization of the composites reveals that inclusion of any particulate filler has very strong influence not only on the mechanical properties of composites but also on their sliding wear behavior. By incorporating these particulate fillers into the vinyl ester matrix, synergistic effects, as expected were achieved in the form of modified mechanical properties and improved sliding wear resistance. A comparative study of modified behavior of the composites against the two different types of fillers is presented.

3.1. Density

The composite under this investigation consists of three components such as matrix, fiber and particulate filler. Hence, the density of the composite can be calculated using rule-of-mixture as shown in the following expression Agarwal and Broutman[27].

(2)

Where, W and ρ represents the weight fraction and density respectively. The suffix m, f, and p stand for the matrix, fiber and particulate filler respectively.

The actual or experimental density of the composite, however, can be determined by simple water immersion technique (Archimedes principle). The volume fraction of voids (V_v) in the composites is calculated using the following equation:

(3)

Table 3. Composite designations and their experimental and theoretical densities Composite Designation CompositeCompositions ExperimentalDensity(g/cm3) TheoreticalDensity(g/cm3) Void Fraction WCGV1 Vinyl ester 1.2738 1.2800 0.4844 WCGV2 Vinyl ester +10 wt.% CaSO4 1.3307 1.3570 1.9381 WCGV3 Vinyl ester +20 wt.% CaSO4 1.3891 1.4439 3.7953 WCGV4 Vinyl ester +10 wt.% CaCO3 1.3400 1.3513 0.8362 WCGV5 Vinyl ester +20 wt.% CaCO3 1.4022 1.4311 2.0194

It can be noticed from Table 3 that composite density values calculated from weight fractions using Eq. (2) are not in agreement with the experimentally determined values. The difference between the composite or theoretical density and the actual or experimental density is a measure of voids and pores present in the composites. It is clear from the Table 3 that volume fraction of voids is small in WCGV₁ due to absence of particulate fillers in it. As the filler content of CaSO₄ and CaCO₃ composites (WCGV₂-WCGV₅) increases from 10 wt.% to 20 wt.% the volume fractions of voids increases as shown in the Table 3. The voids are more in the CaSO₄ filler as comparison to the CaCO₃ filler. This may due to the particle size variations, because the particle size of the CaSO₄ (1.813 µm) and CaCO₃ (1.620 µm) is higher respectively. The voids significantly affect some of the mechanical properties and even the performance of composites. Higher void contents usually mean lower fatigue strength, greater susceptibility to water penetration and weathering[27]. The knowledge of the void content is usually important for the estimation of the quality of the composites.

3.2. Mechanical Properties

The experimental values of the properties of the particulate filled CaCO₃ and CaSO₄ composites under this investigation are presented in Table 4.

Table 4. Mechanical properties of the composites Composites Tensile strength (MPa) Tensile Modulus (GPa) Flexural strength (MPa) Compressive Strength (MPa) Hardness (HRB) WCGV1 39.04 0.583 103.15 9.51 17.5 WCGV2 21.65 0.797 69.45 8.81 21.3 WCGV3 18.58 0.808 59.0 6.27 10.2 WCGV4 24.81 0.698 101.75 8.05 22.15 WCGV5 20.77 0.744 107.6 10.52 19.6

Figure 3. Variation of tensile strength of composites with filler type and content

The tensile test is performed on the universal testing machine (UTM) Hounsfield H25KS as per ASTM standard D 3039-76[28]. It is seen that in all the samples irrespective of the filler material the tensile strength of the composite decreases with increase in filler content. The pure vinyl ester has strength of 39.04 MPa in tension and it may be seen from Table 4 that this value drops to 21.65 MPa and 18.58 MPa with addition of 10 wt.% and 20 wt.% of CaSO₄ filler respectively. The same phenomenon has been seen in CaCO₃ filler_, the values are drops to 24.81 MPa to 20.77 MPa with the addition of the 10 wt.% and 20 wt.% of CaCO₃ filler.

Among the two fillers taken in this study, the inclusion of CaSO₄ filler causes maximum reduction in the composite strength. It may occurs due to the interface bonding between the vinyl ester matrix and CaSO₄ filler is not good to transfer the tensile stress as that in CaCO₃ filler content. The one more reason is that the corner points of the irregular shaped particulates result in stress concentration in the vinyl ester matrix.

Now, it is interesting to note that the properties for tensile modulus is increasing with the addition of the both the filler CaSO₄ and CaCO₃ at 10 wt.% and 20 wt.% respectively as shown in Figure 4.

Figure 4. Variation of tensile modulus of composites with filler type and content

Figure 5. Variation of flexural strength of composites with filler type and content

The flexural test is conducted on the same UTM as per ASTM standard D 2344-84[29]. Figure 5 shows the comparison of flexural strengths of the composites obtained experimentally from the 3-point bend tests for composites (WCGV₁-WCGV₅). The flexural strength for the filler CaCO₃ is increasing with the addition of the 10 wt.% to 20 wt.% filler content. But, for the filler CaSO₄ it decreases with the addition of the filler content. Now, from the results it may now be suggested that CaCO₃ is potential candidate to be used as filler in making high flexural strength composites with the increase of the reinforcement of the filler in comparison to CaSO₄ filler. The flexural properties are of great importance for any structural element. Composite materials used in structures are prone to fail in bending and therefore the development of new composites with improved flexural characteristics is essential. From the results it may now be suggested that CaCO₃ is potential candidate to be used as filler in making high flexural strength composites with the increase of the reinforcement of the filler in comparison to CaSO₄ filler. CaSO₄ holds good flexural strength at 10 wt.% filler content also more than WCGV₁ and WCGV₄ composites as shown in Figure 5. There may be one reason for this that the voids in WCGV₂ (1.9381) is more in comparison to WCGV₄ (0.8362).

Figure 6. Variation of compression strength of composites with filler type and content

Figure 7. Variation of Rockwell hardness of composites with filler type and content

Figure 6 shows the compressive strengths of the composites obtained experimentally for the composites (WCGV₁-WCGV₅). The compressive strength for filler CaSO₄ decreases continuously with the addition of the 10 wt.% and 20 wt.% filler content and for filler CaCO₃, it firstly decreases at the addition of 10 wt.% filler content and then increases with the addition of 20 wt.% filler content as shown in the Figure 6.

Figure 7 shows the hardness of the composites (WCGV₁-WCGV₅) as obtained from the experiments. For composite WCGV₁, the hardness is recorded as 17.5 HRB while for WCGV₂ and WCGV₃ the values are recorded as 21.3 HRB and 10.2 HRB respectively. On the other hand the values for composites WCGV₄ and WCGV₅ are recorded as 22.15 HRB and 19.6 HRB respectively. From the analysis of the results we observed that the hardness of the CaCO₃ filler is more as comparison to CaSO₄.

3.3. Analysis of Experimental Results

On the other hand, from Table 8 the overall mean for the S/N ratio of the coefficient of friction and the specific wear rate for CaSO₄ are found to be 3.3670 db and 89.8949 db respectively. Here, we saw that the overall mean for the S/N ratio of the coefficient of friction and specific wear rate is more in CaCO₃ in comparison to CaSO₄ which means that the coefficient of friction and specific wear rate is less in CaCO₃. The analysis of the experimental data is carried using the software MINITAB 16 specially used for the design of experiment applications. Before analyzing the experimental data using this software for predicting the measure of performance, the possible interaction between control factors are considered. Thus factorial design incorporates a simple means of testing for the presence of the interaction effects.

Table 5. Experimental design for CaCO3 using L27 array Velocity(m/sec.) Normal Load(N) Filler Content(%) Sliding Distance(m) COF(µ) S/N Ratio(db) Wear(mm3/Nm) S/N Ratio(db) 1.57 20 0 1000 0.70 3.09804 0.0000771 82.259 1.57 20 10 3000 0.59 4.58296 0.0000462 86.707 1.57 20 20 5000 0.54 5.35212 0.0000384 88.313 1.57 40 0 3000 0.75 2.49877 0.0000853 81.381 1.57 40 10 5000 0.58 4.73144 0.0000386 88.268 1.57 40 20 1000 0.45 6.93575 0.0000273 91.277 1.57 60 0 5000 0.80 1.93820 0.0000932 80.612 1.57 60 10 1000 0.47 6.55804 0.0000312 90.117 1.57 60 20 3000 0.39 8.17871 0.0000262 91.634 2.62 20 0 3000 0.78 2.15811 0.0000825 81.671 2.62 20 10 5000 0.69 3.22302 0.0000474 86.484 2.62 20 20 1000 0.64 3.87640 0.0000333 89.551 2.62 40 0 5000 0.83 1.61844 0.0000952 80.427 2.62 40 10 1000 0.59 4.58296 0.00000822 101.703 2.62 40 20 3000 0.50 6.02060 0.00000493 106.143 2.62 60 0 1000 0.75 2.49877 0.0000858 81.330 2.62 60 10 3000 0.48 6.37518 0.00000792 102.025 2.62 60 20 5000 0.46 6.74484 0.00000423 107.473 3.67 20 0 5000 0.85 1.41162 0.0000982 80.158 3.67 20 10 1000 0.76 2.38373 0.00000863 101.280 3.67 20 20 3000 0.61 4.29340 0.00000561 105.021 3.67 40 0 1000 0.78 2.15811 0.0000832 81.598 3.67 40 10 3000 0.66 3.60912 0.00000606 104.351 3.67 40 20 5000 0.60 4.43697 0.00000451 106.916 3.67 60 0 3000 0.80 1.93820 0.0000942 80.519 3.67 60 10 5000 0.57 4.88250 0.00000411 107.723 3.67 60 20 1000 0.52 5.67993 0.00000212 113.473

Figure 8 and 9 for CaCO₃ shows graphically the effect of four control factors on coefficient of friction and specific wear rate of the composite specimens WCGV₁, WCGV₄ and WCGV₅. The analysis of the results gives the combination factors resulting in minimum coefficient of friction and specific wear rate of the composites. Analysis of these results leads to the conclusion that factors combination A₁, B₃, C₃ and D₂ gives minimum coefficient of friction as shown in the Figure 8. It is observed that the interaction B×C shows significant effect on the coefficient of friction. Similarly the combination factors A₃, B_3, C₃ and D₂ gives minimum specific wear rate as shown in the Figure 9. It is observed that interaction A×C has significant effect on the specific wear rate.

Figure 8. Effect of control factor on Coefficient of friction. (For CaCO3)

Figure 9. Effect of control factor on Specific wear rate. (For CaCO3)

For CaSO₄ Figure 10 and 11 shows graphically the effect of four control factors on coefficient of friction and specific wear rate of the composite specimens WCGV₁, WCGV₂ and WCGV₃. Analysis of these results leads to the conclusion that factors combination A₁, B_3, C₂ and D₁ gives minimum coefficient of friction as shown in the Figure 10. It is observed that the interaction B×C again shows significant effect on the coefficient of friction as in case of CaCO₃ filler. The combination factors A₃, B₃, C₂ and D₁ gives minimum specific wear rate as shown in the Figure 11. It is observed that interaction A×C has significant effect on the specific wear rate.

Figure 10. Effects of control factor for coefficient of friction. (For CaSO4)

Figure 11. Effect of control factor on Specific wear rate. (For CaSO4)

3.4. ANOVA and Effects of Factors

In order to understand the impact of various control factors like velocity (A), normal load (B), filler content (C) and sliding distance (D) and interaction on the response of experimental data it is desirable to develop the analysis of variance (ANOVA) to find the significant factors as well as interactions. ANOVA allows analyzing the influence of each variable on the total variance of the results. For CaCO₃, Table 6a shows the results of ANOVA for the specific wear rate and Table 7a shows the results of ANOVA for coefficient of friction and for CaSO₄, Table 9a shows the results of ANOVA for the specific wear rate and Table 10a shows the results of ANOVA for coefficient of friction. The analyses are performed with a level of significance 5% means at 95% level of confidence. In ANOVA table, the column shows the percentage contribution (P) of each variable in the total variation indicating the influence of specific wear rate and coefficient of friction.

For filler CaCO₃ it can be observed from the ANOVA Table 6a for specific wear rate that the filler content (P=57.981%), velocity (P=17.981%), normal load (P=5.523%) and the interactions A×C (P=10.218%), B×C (P=3.283) and A×B (P=3.097%) has significant influence on the specific wear rate. However, the control factor sliding distance (P=0.303%) does not have a significant effect (both physically and statistically) on specific wear rate as their values are quit smaller than error (P=1.613%) so they are neglected. From the analysis of ANOVA and response Table 6b of the S/N ratio of specific wear rate, it is observed that the control factor filler content (C) has major impact on the specific wear rate followed by velocity (A), normal load (B) and sliding distance (D). It means that with increasing the filler content, velocity and normal load the specific wear rate decreases i.e., increase the wear resistance as observed from the Figure 9.

Table 6a. ANOVA table for Specific wear rate. (For CaCO3) Source DF Seq SS Adj SS Adj MS F P (%) A 2 563.46 563.46 281.732 33.45 17.981 B 2 173.08 173.08 86.540 10.27 5.523 C 2 1816.92 1816.92 908.462 107.86 57.981 D 2 9.51 9.51 4.754 0.56 0.303 A*B 4 97.06 97.06 24.266 2.88 3.097 A*C 4 320.19 320.19 80.047 9.50 10.218 B*C 4 102.87 102.87 25.718 3.05 3.283 Residual Error 6 50.54 50.54 8.423 1.613 Total 26 3133.63 100.00

Table 6b. Response table for Specific wear rate. (For CaCO3) Level A B C D 1 86.73 89.05 81.11 92.51 2 92.98 93.56 96.52 93.27 3 97.89 94.99 99.98 91.82 Delta 11.16 5.94 18.87 1.45 Rank 2 3 1 4

Table 7a. ANOVA table for Coefficient of friction. (For CaCO3) Source DF Seq SS Adj SS Adj MS F P (%) A 2 9.5096 9.5096 4.7548 33.39 10.265 B 2 11.6173 11.6173 5.8086 40.79 12.540 C 2 59.8532 59.8532 29.9266 210.15 64.607 D 2 1.6144 1.6144 0.8072 5.67 1.743 A*B 4 0.6990 0.6990 0.1748 1.23 0.755 A*C 4 1.7495 1.7495 0.4374 3.07 1.889 B*C 4 6.7443 6.7443 1.6861 11.84 7.279 Residual Error 6 0.8544 0.8544 0.1424 0.922 Total 26 92.6416 100.00

Table 7b. Response table for coefficient of friction. (For CaCO3) Level A B C D 1 4.875 3.375 2.146 4.197 2 4.122 4.066 4.548 4.406 3 3.422 4.977 5.724 3.815 Delta 1.453 1.602 3.578 0.591 Rank 3 2 1 4

In the same way from the ANOVA Table 7a for coefficient of friction the filler content (P=64.607%), normal load (P=12.540%), velocity (P=10.265%), sliding distance (P=1.743%) and the interactions B×C (P=7.279%), A×C (P=1.889%) has significant effect on the coefficient of friction. But, interaction A×B (P=0.755%) does not have a significant effect (both physically and statistically) on coefficient of friction as its value is quit smaller than error (P=0.922%) so it may neglected. So, from the analysis of ANOVA and response Table 7b of the S/N ratio of coefficient of friction, it is observed that the filler content (C) has major influence followed by normal load (B), velocity (A) and sliding distance (D) as for the coefficient of friction.

For filler CaSO_4, it can be observed from the ANOVA Table 9a for specific wear rate that filler content (P=44.375%), velocity (P=27.788%) and the interactions A×C (P=20.423%) and B×C (P=2.474%) has significant effect on specific wear rate. However, the control factors normal load (P=1.326%), sliding distance (P=0.815%) and interaction A×B (P=0.859%) do not have a significant effect (both physically and statistically) on specific wear rate as their values are quit smaller than residual error (P=1.941%) so they are neglected. From the analysis of ANOVA and response Table 9b of the S/N ratio for specific wear rate, it is observed that the control factor filler content (C) has major impact on the specific wear rate followed by velocity (A), the normal load (B) and sliding distance (D). It means that for filler CaSO₄, with the increases of the filler content, velocity and normal load the specific wear rate decreases i.e., the wear resistance is good as observed from the Figure 11. But from the Figure 11 we also observed that for CaSO₄ the filler content plays adverse effect when filler content is increases from 10 wt.% to 20 wt.%. At 10 wt.% for CaSO₄ the specific wear rate is decreased and for 20 wt.%, it further increases.

Table 8. Experimental design for CaSO4 using L27 array Velocity(m/sec.) Normal Load (N) Filler Content (%) Sliding Distance (m) COF(µ) S/N Ratio(db) Wear(mm3/Nm) S/N Ratio(db) 1.57 20 0 1000 0.70 3.09804 0.0000771 82.259 1.57 20 10 3000 0.60 4.43697 0.0000432 87.290 1.57 20 20 5000 0.63 4.01319 0.0000483 86.321 1.57 40 0 3000 0.75 2.49877 0.0000853 81.381 1.57 40 10 5000 0.57 4.88250 0.0000373 88.566 1.57 40 20 1000 0.64 3.87640 0.0000498 86.055 1.57 60 0 5000 0.80 1.93820 0.0000932 80.612 1.57 60 10 1000 0.48 6.37518 0.0000325 89.762 1.57 60 20 3000 0.70 3.09804 0.0000535 85.433 2.62 20 0 3000 0.78 2.15811 0.0000825 81.671 2.62 20 10 5000 0.68 3.34982 0.0000458 86.783 2.62 20 20 1000 0.72 2.85335 0.0000588 84.612 2.62 40 0 5000 0.83 1.61844 0.0000952 80.427 2.62 40 10 1000 0.57 4.88250 0.00000812 101.809 2.62 40 20 3000 0.75 2.49877 0.0000685 83.286 2.62 60 0 1000 0.75 2.49877 0.0000858 81.330 2.62 60 10 3000 0.52 5.67993 0.00000783 102.125 2.62 60 20 5000 0.64 3.87640 0.0000786 82.092 3.67 20 0 5000 0.85 1.41162 0.0000982 80.158 3.67 20 10 1000 0.64 3.87640 0.00000652 103.715 3.67 20 20 3000 0.79 2.04746 0.00000785 102.103 3.67 40 0 1000 0.78 2.15811 0.0000832 81.598 3.67 40 10 3000 0.63 4.01319 0.00000558 105.067 3.67 40 20 5000 0.69 3.22302 0.00000582 104.702 3.67 60 0 3000 0.80 1.93820 0.0000942 80.519 3.67 60 10 5000 0.58 4.73144 0.00000333 109.551 3.67 60 20 1000 0.64 3.87640 0.00000401 107.937

Table 9a. ANOVA Table for Specific wear rate. (For CaSO4) Source DF Seq SS Adj SS Adj MS F P (%) A 2 747.54 747.54 373.769 42.95 27.788 B 2 35.66 35.66 17.831 2.05 1.326 C 2 1193.77 1193.77 596.883 68.58 44.375 D 2 21.93 21.93 10.967 1.26 0.815 A*B 4 23.11 23.11 5.777 0.66 0.859 A*C 4 549.40 549.40 137.351 15.78 20.423 B*C 4 66.55 66.55 16.636 1.91 2.474 Residual Error 6 52.22 52.22 8.703 1.941 Total 26 2690.17 100.00

Table 9b. Response table for Specific wear rate. (For CaSO4) Level A B C D 1 85.30 88.32 81.11 91.01 2 87.13 90.32 97.19 89.88 3 97.26 91.04 91.39 88.80 Delta 11.96 2.72 16.08 2.21 Rank 2 3 1 4

Table 10a. ANOVA table for Coefficient of friction. (For CaSO4) Source DF Seq SS Adj SS Adj MS F P (%) A 2 2.8080 2.8080 1.4040 6.47 6.655 B 2 2.6152 2.6152 1.3076 6.02 6.199 C 2 29.3058 29.3058 14.6529 67.47 69.460 D 2 1.7235 1.7235 0.8618 3.97 4.085 A*B 4 1.7612 1.7612 0.4403 2.03 4.174 A*C 4 0.1993 0.1993 0.0498 0.23 0.472 B*C 4 2.4749 2.4749 0.6187 2.85 5.866 Residual Error 6 1.3030 1.3030 0.2172 3.088 Total 26 42.1908 100

Table 10b. Response table for coefficient of friction. (For CaSO4) Level A B C D 1 3.802 3.027 2.146 3.722 2 3.268 3.295 4.692 3.152 3 3.031 3.779 3.263 3.227 Delta 0.771 0.752 2.546 0.570 Rank 2 3 1 4

In the same way from the ANOVA Table 10a for coefficient of friction the filler content (P=69.460%), velocity (P=6.655%), normal load (P=6.199%), sliding distance (P=4.085%) and the interactions B×C (P=5.866%) and A×B (P=4.174%) has significant effect on the coefficient of friction. However, the interaction A×C (P=0.472%) do not have significant effect on the coefficient of friction as their values are quite smaller than the residual error (P=3.088%), so it may neglected. From the analysis of the ANOVA and the response Table 10b for coefficient of friction it is observed that the filler content (C) has major influence followed by the velocity (A) normal load (B) and the sliding distance (D). From the Fig. 10 for CaSO_4, it is observed that the coefficient of friction is also increase with the increase of the filler content. It means that in 10 wt.% the coefficient of friction is less as comparison to 20 wt.% i.e., the reinforcement of the CaSO₄ filler at 10 wt.% is more wear resistance.

3.5. Surface Morphology

Figures 12a-c are the SEM pictures of composites for minimum, maximum and nominal wear test conditions for CaCO₃ filled vinyl ester composites. It has been found from the experimental analysis that the minimum wear occurs at 3.67 m/sec., 60 N, 20 wt. %, 1000 m test parameter conditions, the maximum wear occurs at 3.67 m/sec., 20 N, 0 wt. %, 5000 m test parameter conditions and the nominal wear occurs at 3.67 m/sec., 20 N, 10 wt. %, 1000 m test parameter conditions as shown in the Figures 12a-c.

The micrograph in Figure 12a shows the resinous and matrix region. The filler CaCO₃ covered the matrix region which results in less wear. Figure 12b shows the debris and wedge formation regions due to long sliding distance. Vinyl ester debris was adhered into the filler region and micro cracks were identified which increases the wear rate. Figure 12c showing the thin layer formation and debris which results the nominal wear.

Similarly, Figures 13a-c are the SEM pictures of composites for minimum, maximum and nominal wear test conditions for CaSO4 filled vinyl ester composites. It has been found from the experimental analysis that the minimum wear occurs at 3.67 m/sec., 60 N, 10 wt. %, 5000 m test parameter conditions, the maximum wear occurs at 3.67 m/sec., 20 N, 0 wt. %, 5000 m test parameter conditions and the nominal wear occurs at 3.67 m/sec., 20 N, 20 wt. %, 3000 m test parameter conditions as shown in the Figures 13a-c.

Figure 12. SEM pictures of composites at (a) 3.67 m/sec., 60N, 20 wt. %, 1000 m (b) 3.67 m/sec., 20N, 0 wt. %, 5000 m (c) 3.67 m/sec., 20N, 10 wt. %, 1000 m. (For CaCO3 filled vinyl ester composites)

Figure 13. SEM pictures of composites at (a) 3.67 m/sec., 60N, 10 wt. %, 5000 m (b) 3.67 m/sec., 20N, 0 wt. %, 5000 m (c) 3.67 m/sec., 20N, 20 wt. %, 3000 m. (For CaSO4 filled vinyl ester composites)

Figure 13a shows the resinous and filler region as in case of CaCO₃ which results in the lesser wear rate because the filler covers the maximum matrix region. Figure 13b found the maximum wear rate due to wedge formation and matrix debris adhered into the filler region as in the case of CaCO₃ which increases the wear rate. Figure 13c shows the nominal wear rate due to thin transfer layer formation.

3.6. Confirmation Experiments

The confirmation experiment is the final step in the design of experiments process. It predicts and verifies the improvements in the observed values through the use of optimal combination level of control factors. For filler CaCO_3, the confirmation experiment was performed by taking an arbitrary set of factor combination A₂ B₂ C₃ D₁ to predict the coefficient of friction and for specific wear rate factor setting is A₂ B₂ C₁ D₁. Now, the estimated S/N ratio for coefficient of friction can be calculated with the help of the following predictive equation.

(4)

Where, is the predicted average of CaCO₃ for coefficient of friction, is the overall experimental average and is the mean response for factors and interactions at designed levels. By combining all the terms Eq. (4) reduces to

(5)

A new combination of factor levels A₂ B₂ C₃ D₁ are used to predict the S/N ratio of coefficient of friction through predictive Eq. (5) and is found to be = 5.6781 db. For each of performance measures an experiment is conducted for different combination of factors and results are compared with those obtained from the predictive equation as shown in the Table 11. Similarly a predictive equation is developed for estimating S/N ratio of specific wear rate as shown in Eq. (6).

(6)

Where, is the predictive average of CaCO₃ for specific wear rate, is the overall experimental average and is the mean response for factors and interactions at designed levels. By combining all the terms Eq. (6) reduces to

(7)

A new combination of factor levels A₂ B₂ C₁ D₁ are used to predict the S/N ratio of specific wear rate through predictive Eq. (7) and is found to be = 81.1478 db. For each of performance measures an experiment is conducted for the prediction equation as shown in Table 11. The resulting equations seem to be capable of predicting the coefficient of friction and specific wear rate. An error of 6.49% for the S/N ratio of the coefficient of friction and 5.41% for the S/N ratio of the specific wear rate is observed.

Table 11. Results of the confirmation experiments for the coefficient of friction and specific wear rate. (For CaCO3) Optimal control parameters Error Prediction Experimental Level A2 B2 C3 D1 A2 B2 C3 D1 (%) S/N ratio for coefficient of friction (db) 5.6781 6.0727 6.49 Level A2 B2 C1 D1 A2 B2 C1 D1 (%) S/N ratio for specific wear rate (db) 81.1478 85.7860 5.41

For filler CaSO_4, the confirmation experiment was performed by taking an arbitrary set of factor combination A₂ B₂ C₃ D₁ to predict the coefficient of friction and for specific wear rate factor setting is A₂ B₂ C₁ D₁ as same for filler CaCO₃. Now, the estimated S/N ratio for coefficient of friction can be calculated with the help of the following predictive Eq. (8).

(8)

Where, is the predicted average of CaSO₄ for coefficient of friction, is the overall experimental average and is the mean response for factors and interactions at designed levels. By combining all the terms Eq. (8) reduces to

(9)

A new combination of factor levels A₂ B₂ C₃ D₁ are used to predict the S/N ratio of coefficient of friction through predictive Eq. (9) and is found to be = 3.2593 db. For each of performance measures an experiment is conducted for different combination of factors and results are compared with those obtained from the predictive equation as shown in the Table 12. Similarly a predictive equation is developed for estimating S/N ratio of specific wear rate as shown in Eq. (10).

(10)

Where, is the predictive average of CaSO₄ for specific wear rate, is the overall experimental average and is the mean response for factors and interactions at designed levels. By combining all the terms Eq. (10) reduces to

(11)

A new combination of factor levels A₂ B₂ C₁ D₁ are used to predict the S/N ratio of specific wear rate through predictive Eq. (11) and is found to be = 82.2855 db. For each of performance measures an experiment is conducted for the prediction equation as shown in Table 12. The resulting equations seem to be capable of predicting the coefficient of friction and specific wear rate. An error of 9.18% for the S/N ratio of the coefficient of friction and 8.04% for the S/N ratio of the specific wear rate is observed.

Table 12. Results of the confirmation experiments for the coefficient of friction and specific wear rate. (For CaSO4) Optimal control parameters Error Prediction Experimental Level A2 B2 C3 D1 A2 B2 C3 D1 (%) S/N ratio for coefficient of friction (db) 3.2593 3.5888 9.18 Level A2 B2 C1 D1 A2 B2 C1 D1 (%) S/N ratio for specific wear rate (db) 82.2855 89.4786 8.04

4. Conclusions

An experimental study has been carried out for friction and dry sliding wear of vinyl ester matrix with fillers CaCO₃ and CaSO₄ sliding against smooth stainless steel counterface using Taguchi experimental design. Taguchi’s design of experiment method can be used to analyze the coefficient of friction and the dry sliding wear of polymer matrix composites as presented in this research paper. The following conclusions can be drawn from the present study:-

(1) The tensile strength for both the fillers CaCO₃ and CaSO₄ decreases with the increase of the filler content. While the tensile modulus for both the fillers increases with the increase of filler content.

(2) For filler CaCO₃, flexural and compressive strength increases with the increase of filler content. While, hardness is increases at 10 wt.% and then decreases at 20 wt.%.

(3) For filler CaSO₄, flexural and compressive strength decreases with the increase of filler content. While, the hardness is increases at 10 wt.% and then decreases at 20 wt.% as same in case of CaCO₃ filler content.

(4) For pure vinyl ester the coefficient of friction and specific wear rate increases with the increase of normal load, sliding velocity and sliding distance.

(5) The coefficient of friction and specific wear rate for CaCO₃ filler decreases with the increase of filler content. But, for filler CaSO₄ the coefficient of friction and specific wear rate decreases at 10 wt.% and then increases at 20 wt.%.

(6) For both the fillers CaCO₃ and CaSO₄ it is observed that the control factor filler content (C) has major impact on the specific wear rate followed by velocity (A), normal load (B) and sliding distance (D).

(7) The predictive equations based on Taguchi approach is successfully used for the prediction of effect of four control factors and predicted results are consistent with the experimental observations.

(8) It is demonstrated that Taguchi approach based on ANOVA well reflect the effects of various factors on the friction and sliding wear loss.

The significance of this current research is to create the interest between the various young researchers towards such fillers which are good not only in the mechanical behaviour but also good in the friction and dry sliding wear behaviour.

ACKNOWLEDGEMENTS

Although the interpretations and opinions expressed in this paper are our own. They are based upon the experimental work and are carried out in National Institute of Technology, Hamirpur. We would like to thank NIT Hamirpur for their able support. We would also like to thank Vivek Ramdas Desale and Akant Kumar Singh for their valuable support in between the work.