Influence of Roughness Parameters on Friction and Transfer Layer Formation

Surface texture of the harder mating surfaces is one of the key factors that control the coefficient of friction during sliding. In the present investigation, various kinds of surface textures were produced on the 080 M40 steel plates. For a given kind of surface texture, roughness was varied using various grits of emery papers or polishing powders. The surface textures were characterized in terms of roughness parameters using an optical profilometer. Sliding experiments were conducted to study the effect of surface texture on coefficient of friction using inclined pin-on-plate sliding tester at a sliding velocity of 2mm/s against the prepared hard plate using a soft AlMg alloy pin under both dry and lubricated conditions in ambient environment. Normal loads were varied from 0 to 120 N during the tests. Using scanning electron microscope (SEM) the surfaces of both the plate and pin materials were examined specifically to study the transfer layer formation in the former and damage in the latter. It was observed that the coefficient of friction and transfer layer formation are controlled by the surface texture of the harder mating surfaces under both dry and lubricated conditions. In addition, it was observed that among the surface roughness parameters, the average or the mean slope of the profile was found to explain the variations best. It was concluded that the coefficient of friction and transfer layer formation are strongly dependent on the mean slope of the profile regardless of surface textures under both dry and lubricated conditions. INTRODUCTION Friction is the resistance to motion during sliding that is experienced when one solid body moves tangentially over another with which it is in contact. The resistive tangential force, which acts in a direction directly opposite to the direction of motion, is called the friction force. Friction is commonly represented by the coefficient of friction, defined as the ratio of tangential force to the normal force. It is influenced by various factors such as surface texture, sliding speed, normal load, temperature, lubricants, and material properties. Considerable work has been done by many researchers to study the influence of these parameters on coefficient of friction [1-9]. It was reported earlier that surface texture indeed has an important role on coefficient of friction values during sliding [10-17]. Menezes et al. [10-12] studied the effect of surface texture on coefficient of friction and transfer layer formation under both dry and lubricated conditions for Al-Mg alloy [10], pure copper [11] and super purity aluminium [12] using inclined scratch test. Various kinds of surface textures – namely, unidirectional grinding marks, 8-ground, and random were prepared using simple metallographic techniques. Roughness, represented by Ra, of surfaces was varied over a range as they were prepared using different grit emery papers and abrasive powders. It was found that surface texture that promotes plane strain conditions near the interface causes higher plowing component and thus the higher coefficient of friction. On the other hand, surface texture that promotes plane stress conditions at the interface results in lower value for plowing component of friction. It was found that sliding perpendicular to the unidirectional grinding marks gives maximum friction force contributed by higher plowing component, and at the other extreme random texture results in lower friction values. It was observed that the roughness as given by Ra within the test range does not significantly affect the friction values. Further, surface textures were characterized in terms of roughness parameters and in the literature many roughness parameters [18] are available. The surface roughness parameter like Ra is used in general, to describe a surface. However, such a single roughness parameter, which is the universally recognized and most used parameter of surface roughness, is not sufficient to describe a functional characteristic like friction [19] and it is possible that two surface textures can have the same Ra, but their frictional characteristics could be different [8-12]. Considerable amount of work has also been done to study the effect of various roughness parameters on friction [19-22]. Lundberg [19] studied the influence of surface roughness parameters on normal sliding lubrication and reported that the Rmax and Rt to be the most significant parameters. Myers [20] conducted experiments using an inclined plane sliding tester to study the coefficient of friction between a test slider and sample disks. Twelve samples surfaces were fabricated from cold-rolled steel disks. Five of these had lapped finishes while the others had ground finishes. The author [20] studied the correlation coefficient between coefficient of friction with the three r.m.s. values corresponding to (i) surface profiles, (ii) first derivative of surface profiles and (iii) second derivative of surface profiles and concluded that the second one, namely, the r.m.s. of first derivative was most useful in predicting friction. Koura [21] studied the effect of surface texture on friction mechanism using universal testing machine. Steel specimens were prepared to various degrees of roughness by grinding, lapping and polishing. The results showed that the behavior of surfaces and thus friction during sliding depends on the degree of roughness. In the literature many individual or hybrid surface roughness parameters are in vogue. These include amplitude, spatial and hybrid parameters. However, it was noticed that like friction, the correlation coefficient between coefficient of friction and roughness parameters, was system dependent. A systematic study on the classification of surface textures was done recently by Stout and Blunt [23]. The authors [23] introduced the concept of engineered and non engineered surfaces and have sub-divided these into random, systematic, unstructured and structured derivatives. Most of the surface textures generated in the present study belong to one of these categories and defined in the experimental section. The aim of the present study is to characterize the surface textures in terms of roughness parameters using optical profilometer and to come out with a single roughness parameter which correlates with coefficient of friction regardless of surface textures. Experiments were done on hard counter surfaces by sliding soft pins using inclined pin-on-plate sliding tester. Scanning Electron Microscope (SEM) was used to reveal the morphology of the transfer layer formed on the plate surface as well as the damage on the pin surface. EXPERIMENTAL DETAILS Four types of surface textures were produced on 080 M40 steel plates. Type I, namely, structured directional surfaces, were produced on the steel plates with varying roughness by dry grinding the steel plates against dry emery papers of 220, 400, 600, 800 or 1000 grit size. For the directional surface texture, care was taken so that the grinding marks were unidirectional in nature. Type II, namely, structured non-directional surface texture, was generated on steel plates with varying roughness by moving the steel plate on dry emery papers of 220, 400, 600, 800 or 1000 grit size along a path with the shape of an “8” for about 500 times. Type III, namely, structured directional surface textures, similar to Type I was produced. Here the grinding marks direction was perpendicular to that of Type I. Type IV, namely, random texture, with varying roughness was generated under wet grinding conditions using a polishing wheel with any one of the three abrasive media such as SiC powder (600 and 1000 grit), Al2O3 powder (0.017 μm), and diamond paste (1-3 μm). Figures 1 (a), (b), (c) and (d) show the profiles of steel plate surfaces along with its 3D roughness parameter, Ra, generated by Types I, II, III, and IV respectively. In figure 1, the surface textures, namely, Type I, II and III were generated using 1000 grit emery papers while the Type IV was produced using 1000 grit SiC powder. Figure 1: Profiles of Types (a) I, (b) II, (c) III, and (d) IV surface textures. Experiments were conducted using an inclined pin-on-plate sliding tester, details of which were explained in earlier paper [24]. In the present study, soft material made of an Al-4Mg alloy was used as pins and hard material made of 080 M40 steel plates were used as counter part. The pins were 10 mm long, 3 mm in diameter with a tip radius of 1.5 mm. The dimensions of the 080 M40 steel plates were 28 mm x 20 mm x 10 mm (thickness). The pins were first machined, and then electro-polished to remove any work-hardened layer that might have formed during the machining. Hardness measurements of pin and plate were made at room temperature using a Vickers micro hardness tester with 100 gm load and 10-second dwell time. Average hardness numbers, obtained from 5 indentations, was found to be 105 and 208 for the pin and plate respectively. Before each experiment, the pins and steel plates were thoroughly cleaned first in an aqueous soap solution and then with acetone in an ultrasonic cleaner. The steel plate was fixed horizontally in the vice of the pin-on-plate sliding tester and then the vice-setup was tilted so that surface of the plate makes an angle of 1 o ± 0.05 o with respect to horizontal base. Then pins were slid at a sliding speed of 2 mm/s against the prepared steel plates starting from lower end to the higher end of the inclined surface for a track length of 10 mm. Normal load was varied from 0 to 100 N during the test. The advantage of 1 o inclination of the steel plate was that from a single experiment, the effect of normal load (up to the test limit of 100 N) on the coefficient of friction could be studied [24]. Dry tests were performed first to obtain five parallel wear tracks on the same steel plate. Each wear track was produced by a single sliding event. It was observed that the initial sphere-on-plate contact essentially became a flat-on-plate type contact even before the end of the first wear track. At the same time, it was observed that wear track width varied considerably in first three tests. For this reason, all the results presented were of the fourth wear track. The fifth wear track was made to confirm the consistency in results. It was observed that the coefficient of friction did not vary much for all these five wear tracks. After the dry tests, the pin was removed and a new pin was mounted on the vertical slide to perform lubricated tests. For the lubricated tests, a drop (i.e., 0.05 ml) of commercially available engine oil lubricant (‘Shell’ make 2-stroke oil) was applied on the surface of the same steel plate and the tests were performed to obtain another five parallel wear tracks on the steel plate similar to dry tests. The viscosity of lubricant oil was found to be 40 cSt at 40 o C and had the extreme pressure additive ZDDP (Zinc Dialkyl Dithiophosphate). The presence of ZDDP was confirmed using Fourier Transform Infrared spectroscopy technique. Both the dry and lubricated tests were done on the same steel plate so that the results of the dry and lubricated experiments will exclude variations during preparation of the steel plates. The dry tests were conducted first followed by the lubricated ones, to avoid any additional cleaning of the steel plates. After the tests, the profiles and surface roughness parameters of the steel plates were measured in the direction of the sliding on the bare surface away from the wear tracks using an optical profilometer. Later, the pins and steel plates were observed using a scanning electron microscope (SEM) to study the surface morphology. In the following sections micrographs of the central regions of both the pins and steel plates are presented. Finally, the effect of roughness parameters on coefficient of friction was investigated. RESULTS & DISCUSSION Figure 2 shows the variations of PSD (power spectral density) with spatial frequency for different types of surface profiles shown in figure 1 in the horizontal and vertical directions. PSD is used for characterizing both the asperity amplitude and spacing. It can be calculated by the Fourier decomposition of the measured surface into its component spatial frequencies. The PSD can be calculated using the expression: