WEAR IN Al-Si ALLOYS UNDER DRY SLIDING CONDITIONS K. MOHAMMED JASIM* and E. S. DWARAKADASA+ Scientific
(Received March 11, 1987; accepted April 8,1987)
Summary Al-Si alloys containing 3 - 22 wt.% silicon have been slid against a continuously rotating steel disc under dry sliding conditions in the ranges of bearing pressure and sliding speed of 6 - 195 kPa and 58 - 580 cm s-l respectively. The steady state wear rates increase linearly with the bearing pressures in two distinct regions marked by a sharp transition point. This behaviour is indicative of an oxidative mechanism at low bearing pressure and a combined oxidative-metallic wear above the transition point. At all pressures above the transition point gross failure of metal takes place at the interface, preceded by severe plastic deformation. Evidence for the above process has been obtained in terms of debris structure and examination using scanning electron microscopy of the worn surface. The wear behaviour and wear rates have been found to be a function of silicon content and not dependent on the initial structure or the distribution of the silicon phase. This is due to the formation of a subsurface deformation layer in which the silicon phase is finely fragmented into spheroidal particles. The size of the fragments is nearly constant at 3 - 5 pm and is independent of all the parameters involved in the process and explains the non-dependency of the wear rate on the internal structure. However, the depth to which the silicon phase is fragmented, i.e. the subsurface damage, is critically dependent on all the experimental parameters. The wear rate in Al-Si alloys shows a steady decrease up to the eutectic composition and thereafter it decreases, and is explained on the basis of the structure of the subsurface region which is homogeneous. In high silicon alloys the subsurface region does not show any signs of plastic deformation having occurred in the direction of sliding. The results obtained have been used to propose a mechanism of metal removal during wear in Al-Si alloys.
*Present address: Department of Metallurgy and Materials Science, Imperial College of Science and Technology, London SW7 2BP, U.K. ‘Present address: Department of Metallurgy, Indian Institute of Science, Bangalore 560 012, India. 0043-1648/87/$3.50
0 Elaevier Sequoia/Printed
in The Netherlands
1. Introduction Loss of material removal by wear is a universal phenomenon and has attracted the attention of many a scientist, engineer and technologist owing to its importance. In recent years, efforts have been directed towards the study of the actual mechanisms involved in the material removal due to rubbing action at the interfaces. The chief objective of research in this area is to minimize loss due to wear at all levels. One important outcome of several such investigations is the fact that aluminium-based alloys, especially Al-% alloys, make good substitutes for cast iron components in applications in which resistance to wear is desired. As a consequence, the wear behaviour of Al-S1 alloys has been studied in detail to reveal that the eutectic and hypereutectic compositions show a very low wear rate compared with other compositions as for many other engineering materials. One of the several methods of examining wear is to look at the subsurface effects, because it is known that in ductile materials [l] subsurface deformation influences the wear behaviour. In single-phase materials it was difficult to quantify the extent of deformation while the use of suitable two phase alloys enabled  measurement of the depth of damage from the structural changes that result. In this respect, the Al-Si eutectic alloy was shown [2, 31 to be an ideal system. However, because of the fact that in most wear experiments the tendency is to use a range of pressures to study its effect on wear rate, and usually these pressures are rather high, subsurface effects have not been clearly identified. Use of high pressures is clearly derogative in the examination of subsurface effects and results with the use of small loads have not been reported. In a recent study [4, 51, the topography of the worn surface has been found to display a characteristic appearance related to the wear process and reminiscent of the material removal mechanism. Also, it is well known that the shape, size and size distribution of the debris particles are closely related to the wear mechanism. In spite of the large amount of data available in the literature on the wear of materials under widely varying experimental conditions, no attempt has been made so far to correlate worn surface topography and subsurface damage with wear behaviour or debris characteristics. A detailed study was therefore planned of the wear behaviour of Al-Si alloys under unlubricated adhesive low pressure low speed conditions. The results obtained have been presented and discussed in this communication.
2. Materials and experimental procedures Binary Al-Si alloys were made starting from 2S grade commercial purity aluminium and an Al-22wt.%Si master alloy. The chief impurities in both were iron and titanium which were probably used for grain refinement. A desired quantity of aluminium was first melted in an alumina crucible in a
gas-fired furnace and a suitable quantity of master alloy, taken in the form of small pieces and well wrapped in thick aluminium foil, was added to the melt, After a good stirring to allow for homogenization the alloy melt was poured into metallic moulds to obtain chill cast ingots 15 mm in diameter and 150 mm long. All the rods were used in the as-cast condition. Alloys were made conforming to nominal silicon contents of 3, 8, 13, 16 and 22 wt.% Si. The actual chemical compositions of the alloys were checked by spectrographic analysis to determine that the silicon content was within 5% of the nominal value. In addition, the compositions were also checked by comparison with the microstructures of standard Al-Si alloys. The cast rods, which were machined down to about 15 mm in length, were prepared with flat end faces. Use of conical or spherical-tipped wear pins would give better control and reproducibility. However, the upright cylinder geometry was preferred as it lends itself to easier subsurface examination. Wear experiments were conducted on a standard pin-on-disc machine with a continuously rotating steel disc as the counterface. Two different steel discs were used having average hardness values of 30 HRC and 50 HRC. The wear pin, supported with its axis vertical at one end of a lever arm, was allowed to bear against the steel surface and the bearing pressure was varied by varying the bearing load by the addition of dead-weights in steps of 0.5 N. The volume wear rate was calculated by the weight loss measurements after 60 min con~uous run. The relative sliding speed at the interface was varied by varying the rotational speed of the disc while the distance of the pin from the disc axis was kept constant. During the wear runs it was observed that within a few minutes the disc surface was coated with fine debris particles. In order to retain actual application conditions, no attempt was made to clear or remove the debris but it was left, in fact, for the entire running time. The wear conditions reported in this paper, strictly, therefore, represent three-body interaction conditions. Thus, although the wear experiments were designed for the study of adhesive wear, they are not truly adhesive in nature. The debris particles, especially of the oxide type, cause an appreciable and observable amount of abrasive wear. At the end of the wear run the wear pin was transferred to the specimen chamber of a Cambridge Stereoscan model S410 scanning electron miicroscope to study the structure of the worn surface. The pin was then subjected to 5” oblique  sectioning and polishing to observe the subsurface structure. Examination of the subsurface damage was carried out for both the etched and the unetched conditions in the optical microscope as well as in the scanning electron microscope. The debris particles collected on the disc surface were, at the end of the wear run, collected in a vial and examined in the scanning electron microscope after a time lapse of 1 - 5 days. The vials were stored in a desiccator until examination and there is the possibility of some oxidation occurring before they were taken to the microscope.
Before starting a new wear experiment the disc surface was cleaned, ground with 600 grade Sic emery paper, cleaned and degreased so that the starting conditions for every wear run were exactly identical.
3. Results and discussion 3.1. Load-wear rate relationship The variation in the adhesive wear rates of the binary Al-Si alloys is plotted in Fig. 1 as a function of the bearing pressure in the range 6 - 195 kPa. The variation shown in Fig. 1 is typical of the behaviour of all the Al-Si alloys studied and only representative data for hypoeutectic, eutectic and hypereutectic alloys are plotted in Fig. 1. The wear rate increased linearly in two distinct regions in almost all cases. These two regions have been referred to as the “low pressure” region and the “high pressure” region which are separated by a sharp transition point. A similar behaviour had been observed in pure aluminium , Al-Ni eutectic alloy  and Al-22wt.%Si alloy [ 7, 81. The wear rates observed in the low pressure region did not show much variation with varying alloy composition or sliding speed. Wear in this region is interpreted as reflecting the fracture of the oxide layer at the wear interface, especially since the load levels involved are insufficient to cause deep penetration and deformation in the metal below the oxide. Also, as was shown by Razavizadeh and Eyre , the temperature levels reached at the interface are not high enough to cause the extent of oxidation necessary for the observed amount of debris. Thus, in the low pressure region,
Fig. 1. Variation of the adhesive wear rates of binary Al-Si alloys as a function of bearing pressure.
wear is primarily controlled by the fracture and removal of the oxide debris particles. As supporting evidence, in the low pressure region, deformation in the subsurface regions was not observed. The effect of increasing the pressure below the transition point is to accelerate the fracture of the oxide and thus cause increased wear. As the surface oxide is removed, the fresh metal exposed is further oxidized. The exact way in which wear continues, by a process of oxidation, fracture and removal of oxide and fresh oxidation, is not clearly known as yet. However, it is clear enough that, for a study of subsurface deformation during the wear process, it is necessary to use pressure values larger than that indicated by the transition point for the particular conditions of pressure, speed and alloy composition. It was also noted that the effect of increasing speed was to shift the transition point to higher loads. When the counterface hardness was increased, the transition point pressure was also shifted to higher values. However, irrespective of the value of the hardness of the steel disc, the disc material was found to be involved in the wear process, as was indicated by the presence of magnetic particles in the debris. Rolling of these particles in the interface was thought to contribute considerably to the surface deformation observed. Hence, in order to maintain comparable sliding conditions, the disc hardness was kept constant. It is significant to note that, under abrasive conditions of wear [lo], subsurface deformation was absent in all the Al-Si alloys. The slope of the linear plots of wear rate us. pressure in the high pressure region varied considerably with alloy composition and the other wear conditions. This region was characterized by the formation of metallic debris particles present along with the fine oxide debris. It can therefore be said that in this region wear is occurring both by oxide removal and metal failure at the interface. The laminar shapes of the metallic debris particles provide ample evidence for adhesive wear taking place as per the delamination theory proposed by Suh [l]. Large-sized debris particles rolling through the interface cause plastic deformation in the surface layers of the softer Al-Si wear pin. As the bearing pressure increases, the amount of metallic debris also increases, leading to increased depths of subsurface damage. At very high values of bearing pressure the deformation at the surface becomes large enough to obscure the details of the process as observed in terms of the surface topography (given in the next section). It therefore becomes clear from this study that, for the investigation of subsurface deformation effects, it is necessary to select proper values of load, depending on the wear system. A similar conclusion was reached for 60/40 brass slid against carbon steel discs [ll]. Use of very high pressures is probably one of the reasons it has not been possible to make a detailed study of subsurface deformation effects in earlier investigations. 3.2. Effect of silicon content The effect of silicon content on the wear rate of aluminium is shown in Fig. 2. As has been observed by other investigators [ 12 - 151, the adhesive wear rate in aluminium is reduced by the addition of silicon to reach a
30HRC Steel -_-o_-.--_x--
Fig. 2. Effect of silicon content speed, 125 cm s-l).
SOHRC Steel -o-
12 15 WeightPercentage Silicon
the adhesive wear rate of At-Si alloys (relative &ding
minimum value at about the eutectic composition. Further addition of silicon increases the wear rate. These results have been discussed in detail by others . Careful examination of the plots in Fig. 2 reveals the fact that the effect of silicon is more pronounced at intermediate pressures than at high pressures. Varying the sliding speed had little influence on the nature of the plots in Fig. 2. However, the effect of silicon on the wear rate of aluminium became less pronounced as the sliding speed increased. Increasing the counterface hardness also had a similar effect. 3.3. Topography of the worn surface Interest in the study of worn surface topography has registered a steep increase since the av~ab~ity of the scanning electron microscope in which the specimen could be examined with no further preparation and with a very large depth of field. There have been several attempts [2, 4, 141 at correlating the wear behaviour with the wear surface topography. A clear correlation has not yet emerged. In an earlier study 1171, it was shown that the use of a two phase material lends itself to easy and direct identification of the extent of subsurface damage. In this study, it was realized that use of the two phase alloys also gives the facility of studying the characteristics of the wear surface topo~aphy. Thus it becomes clear that to study the wear
Fig, 3. Secondary electron images of freshly worn surfaces of fa) pure commercial aluminium, (b) Al-Bwt.%Si alloy and fc) Al-lGwt.%Si alfoy. Bearing pressure, 52 kPa; sliding speed, 126 cm s-l. Arrows indicate sliding direction.
mechanism it is also necessary to choose proper values of pressure and speed, and suitable disc materials and the test materials. In ductile materials it is known that the surface layers undergo plastic deformation. The wear surface is expected to show typical wear track patterns. In the case of the metal, aiuminium, the surface presents a very confusing picture, as shown in Fig. 3(a). Addition of silicon changes the topography so that it can be more easily analysed, as can be seen from Figs. 3(b) and 3(c). All the surfaces shown in Fig. 3 are worn under similar conditions. From debris particle examination, the occurrence of a delamination process could be deduced but a surface topographic structure, ~m~iscent of a delamination process, could be clearly seen only in the Al-Si alloys under specific conditions of wear. The effect on the surface structure of increasing the bearing pressure is shown in Fig. 4. Although deformation in the surface layer is more obvious,
Fig. 4. Secondary electron image of freshly worn Al-8wt.%Si alloy. Bearing pressure, 155 kPa; sliding speed, 125 cm s-I. Arrow indicates sliding direction.
the number of cracks appearing on the surface and the number of fracture islands increase, cracks become shallower and the details of the wear process become less obvious. As shown by Bai et al. [ 161, this feature is also related to the extent of subsurface deformation. In the eutectic and hypereutectic Al-Si alloys the wear surface was characterized by regions of failure with highly granular inner surfaces as shown in Fig. 5. In these alloys the occurrence of surface plastic deformation was clear from the shapes of microcracks present in the regions in between the fractured islands. These cracks curve out in the direction of sliding showing the deformation. A magnified view of a region containing such microcracks, marked in Fig. 5, is shown in Fig. 6. The formation of the fractured islands, as well as the microcracks, have been found to bear a direct relation to the composition of the wear pin material 3.4. Structure of the subsurface region Microstructural examination carried out on oblique sections  revealed the nature and depth of subsurface deformation caused by the sliding process. In pure aluminium a clear zone of plastic deformation was observable (as shown in Fig. 7(a)) which is the microstructure of the subsurface region of a commercial aluminium sample. The good contrast observed is mainly because of the impurity particles present in the commercial material. Very close to the wear surface, the presence of compacted oxide is also apparent . Figures 5, 6 and 7(b) - 7(d) show the effect of the presence of silicon on the subsurface structure and deformation. That deformation has occurred is still obvious in the hypoeutectic AHwt.%Si alloy (Fig. 7(b)), while it is less obvious in the eutectic Al-13wt.%Si alloy (Fig. 7(c)) and in the hypereutectic Al-lGwt.%Si alloy it is not observed at all. It can also be seen that the depth to which the compacted oxide occurs continu-
Fig. 5. Granular appearance of islands of wear fracture in a freshly worn Al-lGwt.%Si alloy. Bearing pressure, 105 kPa; sliding speed, 133 cm s-l. Arrow indicates sliding direction. Fig. 6. High magnification secondary electron image of area shown in Fig. 5, showing cracks from the subsurface region opening out into the surface.
ously decreases as the silicon content is increased. This is again observed in the hypereutectic Al-Si alloy where the region of compacted oxide decreases in depth as the silicon content of the alloy is increased. The oblique sections shown in Fig. 7 also reveal the presence of deep “V” grooves at the intersection of the wear surface and the polishing plane (the line of intersection between the dark areas on the right and the bright areas on the left is the line of intersection in each case). The depth of these grooves is indicative of both the amount of deformation and the average size of the debris particles. As the amount of silicon is increased, the depth of the grooves decreases. Possibly, the presence of silicon causes fracture and removal of the debris much quicker than in aluminium. Deformation, both at the wear surface as well as in the subsurface region, decreases, as is obvious from an examination of the structures shown in Fig. 7. In all the Al-Si alloys, irrespective of the silicon content or the pressure or the speed of testing, the subsurface region was always punctuated by the presence of finely fragmented silicon particles. The average size of the silicon particles measured in nearly all the Al-Si alloys was nearly the same, in the range 3 - 5 pm. A similar observation had been made by Eyre  for hypereutectic Al-Si alloys and in Al-Al-Ni and other two phase aluminiumbased alloys by Rohatgi and Pai [ 151. Mohammed Jasim and Dwarakadasa had observed that in Al-13wt.%Si and AG22wt.%Si [6,8] alloys the silicon particles were not only fragmented but were also spheroidized. The effect of varying the bearing pressure and sliding speed of the experiment was to
affect the depth to which the fragmentation took place. Thus during steady state wear at the wear interface it is always a composite material of silicon particles distributed in an aluminium matrix that is presented to the rubbing action. This clearly explains why the wear rate for Al-Si alloys is independent of the initial silicon structure but is only dependent on the silicon content. However, at very low bearing pressure levels in hypereutectic Al-Si alloys, the primary silicon phases remain unfragmented until they reach the wear interface (as can be seen in Fig. 5). Care should therefore be exercised in choosing proper bearing pressures while using the hypereutectic alloys. The mechanism by which the silicon phase becomes fragmented during steady state wear was discussed earlier in terms of the wear in Al-22wt.%Si [6, 81 and Al-13wt.%Si [3, 131 alloys. Observations now made with other binary alloys confirm that the fragmentation takes place in the same way in all these alloys. Mechanical compressive forces in the surface layer of the pin, due to the normal component of the bearing pressure, crack the silicon needles while the tangential component of the pressure or, more precisely, the frictional drag assists in the redistribution of the cracked silicon uniformly in the deformed region. The fact that the fragmented silicon particles neither appear in the flow lines of plastic deformation nor possess the angular shapes of the mechanically cracked silicon phase clearly indicates that an additional thermally assisted process is in operation. This has been analysed in greater detail elsewhere [13, 181. The high temperatures reached at the interface lead to a solution of the silicon phase at the Si-Al interface and later, when the material cools, the silicon redeposits giving rise to rounded silicon particles. However, in the absence of proper measurements of the temperature levels reached at the rubbing interface, this rather oversimplified picture cannot be further refined. 3.5. Subsurface damage and wear behaviour From the observations made on the wear behaviour of Al-Si alloys under adhesive conditions, it is possible to build a consistent picture of the mechanism of material removal in these alloys. Under steady state conditions there occurs a zone consisting of a deformed layer immediately below the wear surface in which the silicon phase is finely fragmented. The size of the silicon particles is indicated by the level of compressive forces that can be transmitted by the deforming matrix. Failure by a delamination process is clearly indicated by the shapes of the debris particles. This means that, by the repeated action of the bearing pressure, cracks nucleate in the siliconmatrix interface. When the effective crack length for a number of particles, possibly lying in a plane parallel to the wear interface, reaches a critical value, catastrophic failure in the matrix between these particles results in the removal of the laminar debris. Cracking within the small-sized silicon particles is considered unlikely because, if this were possible, the silicon phase would be reduced to smaller-sized particles. The detached laminate would actually be removed from the surface of the pin material when, under the frictional drag, a number of tensile cracks have nucleated and opened out on
the surface. When these cracks and the subsurface plane of separation meet, the debris particle is detached. In a recent study  it has been observed that if the Al-Si alloys consist of spheroids of silicon of different sizes (as can be produced by a process of modification or by long-term homogenization of the alloy) the silicon phase does not become fragmented, even when the initial particle size is far larger than the average fragment size observed in the subsurface region of 3 - 5 lrn. This observation clearly indicates that the actual process in operation is more complicated than the simple picture derived above. Further observations would be required in order to refine the mechanism proposed and work is in progress in this direction. 4. Conclusions (1) The adhesive wear rates and wear characteristics of binary Al-Si alloys, studied as a function of bearing pressure, sliding speed, disc hardness and alloy composition, indicate that the wear in these alloys is related to a zone of subsurface damage in which the silicon-rich phase is fragmented into spheroids 3 - 5 pm in size. The non-dependency of wear rate on silicon structure is thus obvious. (2) The eutectic alloy shows the lower wear rate compared with all the alloys in the Al-Si system. (3) The subsurface region does not show any signs that plastic deformation has occurred in the hypereutectic alloys. (4) The spheroidal particles of the fragments show that combined mechanical and thermal processes take place in the subsurface layer. References 1 N. P. Suh, Wear, 44 (1977) 1. 2 E. S. Dwarakadasa, B. N. Pramiia Bai and S. K. Biswas, Metall. Mater. Technol., 10 (1981) 91. 3 B. N. Pramila Bai, E. S. Dwarakadasa and S. K. Biswas, Wear, 71 (1981) 381. 4 A. D. Sarkar and J. Clarke, Wear, 75 (1982) 71. 5 E. S. Dwarakadasa and Ft. S. Yaseen, Wear, 84 (1983) 375. 6 K. Mohammed Jasim, Wear, 98 (1984) 183. 7 K. Mohammed Jasim and E. S. Dwarakadasa, Wear, 82 (1982) 377. 8 K. Mohammed Jasim and E. S. Dwarakadasa, Mater. Sci. Lett., 1 (1982) 503. 9 K. Razavizadeh and T. S. Eyre, Wear, 82 (1982) 325. 10 R. I. AzaI, B. Abdul Reda, E. S. Dwarakadasa and A. Ismail, Report, 1983 (Department of Metallurgy, University of Technology, Baghdad, Iraq). 11 Y. B. Sheath, Thesis, Department of Mechanical Engineering, University of Technology, Baghdad, Iraq, 1983. 12 T. S. Eyre, Microstruct. Sci., 8 (1980) 141. 13 K. Mohammed Jasim and E. 5. Dwarakadasa, Adhesive wear characteristics in the dry sliding of as-cast ahrminium-silicon alloys, submitted to Wear. 14 R. Shivnath, P. K. Sen Gupta and T. S. Eyre, Br. Foundryman, 70 (1977) 333. 15 P. K. Rohatgi and B. C. Pai, Wear, 28 (1974) 353. 16 B. N. Pramila Bai, E. S. Dwarakadasa and S. K. Biswas, Wear, 76 (1982) 211. 17 R. S. Yaseen and E. S. Dwarakadasa, Wear, 85 (1983) 213. 18 K. Mohammed Jasim, Thesis, Department of Production Engineering and Metallurgy, University of Technology, Baghdad, Iraq, 1983.