Wear 265 (2008) 1902–1908
Contents lists available at ScienceDirect
Wear journal homepage: www.elsevier.com/locate/wear
Sliding wear behaviour of thixoformed AlSiCuFe alloys Yucel Birol a,∗ , Feriha Birol b a b
Materials Institute, Marmara Research Center, TUBI˙ TAK, Kocaeli, Turkey R&D Center, ARC¸ELI˙ K, Tuzla, I˙ stanbul, Turkey
a r t i c l e
i n f o
Article history: Received 29 December 2007 Received in revised form 30 April 2008 Accepted 2 May 2008 Available online 17 June 2008 Keywords: Thixoforming Aluminium alloys Sliding wear
a b s t r a c t The effect of Si content on the wear properties of thixoformed AlSiCuFe alloys was investigated in the present work. Wear rates, which decrease with increasing Si content until 12 wt%, start increasing at hypereutectic Si levels. The inferior wear resistance of thixoformed hypereutectic alloys is linked with the response of the coarse Si particles to severe wear conditions. The primary Si particles fracture and eventually drop out during the wear test leaving large cavities on the surface. The T6 heat treatment provided considerable hardening yet failed to improve wear properties. It is inferred from the wear rates of thixoformed and heat-treated samples that the wear properties of the thixoformed AlSiCuFe alloys is governed largely by the size and distribution of coarse Si particles and that the impact of hardness is only secondary. © 2008 Elsevier B.V. All rights reserved.
1. Introduction Compressors which use low-viscosity lubricants enjoy superior performance but suffer heavier wear problems and thus require a connecting rod material more resistant to wear than the current high pressure die cast (HPDC) AlSiCuFe alloy . Hypereutectic Al–Si-based alloys, which offer high wear resistance, high strength, high hardness and low thermal expansion [2,3] appear to be the material of choice. These attributes, together with excellent castability and reduced density, make these alloys very competitive in heavy wear applications . However, the use of conventional cast grades has been restricted owing to their high latent heat and consequent long solidiﬁcation time which results in die wear, segregation and excessive growth of primary silicon particles and unfavourable shrinkage behaviour [5,6]. Thixoformed parts may be attractive for this demanding application since the casting temperature and heat content in thixoforming are relatively lower, the primary silicon is ﬁner and uniformly distributed and the shrinkage is much less than that of a molten alloy . Furthermore, the die ﬁlling process can be controlled to eliminate porosity, thanks to the high viscosity of semisolid alloys. Recently, wear properties of several AlSiCuFe alloys produced by high pressure die casting and thixoforming were investigated by the authors in response to the need to replace the HPDC AlSiCuFe alloy currently used in connecting rods in compressors . While the wear properties of the thixoformed AlSiCuFe alloys were encour-
∗ Corresponding author. Tel.: +90 262 6773084; fax: +90 262 6412309. E-mail address: [email protected]
ov.tr (Y. Birol). 0043-1648/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.wear.2008.05.001
aging, a sound analysis of their performances with respect to Si content was not possible. The present work was undertaken to explore the impact of Si content on the wear properties of thixoformed AlSiCuFe alloys tested under conditions faced by connecting rods in compressors. 2. Experimental A series of experimental AlSiCuFe alloys were obtained by adding 4–12 wt% elemental Si to AlSi8Cu3Fe alloy the HPDC grade of which is used in connecting rods in compressors (Table 1). Ingots of these alloys were melted and the melts thus obtained were cooled to within 5–15 ◦ C of their liquidus points before they were poured into a permanent mould in order to produce the non-dendritic ingots for thixoforming. Details of thixoforming feedstock preparation are given in . The slugs machined from these ingots were thixoformed into preforms after they were heated in situ in the semisolid range, between 568 and 573 ◦ C, for 5 min in a laboratory press. The thixoformed preforms were then machined precisely into wear ring test samples (Fig. 1). A second set of thixoformed samples were heat-treated to the T6 temper, by solutionizing at 500 ◦ C for 1 h, followed by forced air cooling before ageing at 175 ◦ C for 8 h. A modiﬁed Block on Ring test unit (Fig. 2) was employed to identify the wear properties of the thixoformed AlSiCuFe alloys. The wear tests were performed under exactly the same conditions connecting rods face in service (Table 2). The counterface block material was a heat-treated DIN 100Cr6 steel, the common wrist pin material in compressors. The test chamber was continuously purged with a R600a refrigerant gas. Experiments were conducted
Y. Birol, F. Birol / Wear 265 (2008) 1902–1908
Fig. 1. Preform produced by thixoforming and the wear test ring samples machined from the thixoformed part. Table 1 Chemical compositions of the alloys used in the present work Alloy
1 2 3 4
8.52 12.65 16.32 19.73
1.099 1.109 1.056 1.044
2.940 2.600 2.934 2.623
0.1712 0.1813 0.1708 0.1635
0.1532 0.1193 0.1395 0.1049
0.018 0.019 0.018 0.019
0.046 0.040 0.047 0.044
0.975 0.907 0.953 0.893
Table 2 Wear test parameters Block (counterface) material
Load (N) Contact pressure (MPa) Sliding distance (m) Rotation speed (rpm) Frequency (Hz) Time (min) Pendulum motion angle (◦ ) Lubrication Lubricant Temperature (◦ C) Continuous purging of R600a (refrigerant gas) into the chamber
350 277 370 100 1.6 210 60 Semi-submerged Mineral oil 75
with a contact pressure of 277 MPa under semi-submerged conditions in a mineral oil lubricant (0.007 Pa s viscosity at 40 ◦ C) at 75 ◦ C. Each test lasted 210 min with an oscillation frequency of 1.6 Hz. The extent of wear in the test rings was estimated from weight loss measurements. The samples were cleaned ultrasonically in trichloroethylene before and after each test. An analytical balance with an accuracy of 0.1 mg was used to measure the weight of the samples before and after each test. The worn surfaces of the tested
samples were examined with a scanning electron microscope ﬁtted with an energy dispersive X-ray spectroscopy (EDS). Samples sectioned from the as-cast ingots, thixoformed and heat-treated parts were prepared with standard metallographic practices. These samples were etched with a 0.5% HF solution before they were examined with optical and scanning electron microscopes. X-ray diffraction (XRD) was employed for the identiﬁcation of intermetallic particles with a Shimadzu XRD 6000 Diffractometer equipped with Cu K␣ radiation. The hardness of the thixoformed and heat-treated samples was measured in Brinel (HB) units with a load of 31.25 kg and a 2.5 mm diameter indenter.
3. Results and discussion As-cast microstructures of the parent AlSi8Cu3Fe alloy and the three experimental alloys with 4–12 wt% additional Si are illustrated in Fig. 3. The parent alloy (alloy 1) exhibits a coarse dendritic structure of the ␣-Al solid solution with an interdendritic network of Al–Si eutectic phase, conﬁrming its hypoeutectic composition. Intermetallic particles, found by XRD and metallographic analysis to be of the ␤-Al5 FeSi and CuAl2 variety, were also noted at interdendritic sites. Alloy 2 is dominated by a uniform distribution of Si plates and needles in an aluminium solid solution matrix (Fig. 3b). Several small primary Si particles place this alloy very near, yet slightly above the eutectic point. Increasing number of coarse blocky Si particles are noted in the next two hypereutectic alloys (Fig. 3c and d). These particles are larger in size, occupy a greater volume fraction and have thus become the predominant feature in alloy 4 which has as much as 20 wt% Si (Fig. 3d). The scale of the dendritic structure is much smaller in the hypereutectic alloys (alloys 3 and 4) than in the hypoeutectic alloy (alloy 1). Marked changes are noted in all alloys after thixoforming. The eutectic Si plates and needles in the cast ingots are largely modiﬁed into compact blocky particles upon semisolid soaking while the aluminium solid solution matrix is rearranged into a more or less uniform distribution of ␣-Al globules. Hence, a variety of features in the cast alloys are replaced by a simple mixture of ␣-Al globules and predominantly compact blocky Si particles, regardless of the Si level (Fig. 4a). The latter are bigger and more frequent in the hypereutectic alloys while ␣-Al globules become smaller
Fig. 2. Modiﬁed block-on-ring wear test equipment used in the present work.
Y. Birol, F. Birol / Wear 265 (2008) 1902–1908 Table 3 Wear test results for the thixoformed (TF) and heat-treated (T6) alloys
Fig. 3. As-cast microstructures of the AlSiCuFe alloys: (a) alloy 1, (b) alloy 2, (c) alloy 3 and (d) alloy 4.
Weight loss (mg)
Wear rate, k × 10−5 (mm3 /Nm)
4.2 ± 0.9 3.7 ± 0.9
1.20 ± 0.24 1.06 ± 0.24
82.0 ± 0.2 121.7 ± 0.6
1.8 ± 0.4 3.0 ± 0.5
0.52 ± 0.10 0.86 ± 0.14
84.8 ± 1.0 121.7 ± 0.6
3.5 ± 0.5 3.2 ± 0.4
1.00 ± 0.14 0.92 ± 0.10
82.0 ± 0.2 127.0 ± 1.0
4.4 ± 0.4 4.7 ± 0.4
1.26 ± 0.10 1.34 ± 0.10
84.8 ± 1.0 130.7 ± 1.2
with increasing Si content. With ␣-Al globules invariably smaller than 100 m and with no evidence of intraglobular liquid, the thixoformed microstructures are just as good as those previously reported for similar alloys [4,5,10]. High temperatures and long soaking times involved in the solution heat treatment apparently promoted rounding of the Si particles, the interglobular ones in particular (Fig. 5). Further modiﬁcation of the Si plates and needles, which have survived the semisolid heating cycle, at the relatively lower solutionizing temperature, is believed to be due to the much longer duration of the solution heat treatment. The primary Si particles and Fe/Cu-based intermetallics, on the other hand, resisted spheroidization during the solution treatment. The very ﬁne intraglobular particles which are noted after the T6 treatment are Mg2 Si particles which have precipitated during quenching from the solution heat treatment. The hardness of the thixoformed parts which ranged between 82 and 96 HB have increased to 121–131 HB after the T6 heat treatment, increasing with increasing Si content both in the thixoformed state and after the T6 treatment (Fig. 6). The wear test results are summarized in Table 3, in units of weight loss and speciﬁc wear rate (worn volume per unit sliding distance per unit load). The wear rates are substantial suggesting that a state of severe wear prevailed under the test conditions employed. Severe wear is evidenced also by the SEM micrographs which show that the original machining marks are largely missing across the worn surfaces of ring samples (Fig. 7). Extreme conditions of stress close to the worn surface have led to extensive plastic deformation as evidenced by the degree of particle fragmentation particularly in hypereutectic alloys (Fig. 8). The detailed investigation of the contact surfaces of the counter-surface and the ring samples has shown substantial material transfer from the latter to the counter-surface suggesting that adhesive wear was occurring. The transferred layer was shown by EDS analysis to be aluminium-rich. The transfer of aluminium together with the deformation of the surface layers has apparently changed the topological features of the contact surfaces. It is clear from Fig. 9 that the wear rates of the experimental AlSiCuFe alloys decrease with increasing Si content until 12 wt%. The wear resistance of the parent alloy has more than doubled with a mere addition of 4 wt% Si. This trend is reversed, however, once the Si content increases above 12 wt% in spite of an ever increasing hardness in this composition range. Hence, of the four alloys tested in the present work, the near-eutectic alloy is the most wear-resistant. The favorable effect of Si and of hardness on wear resistance of hypereutectic Al–Si alloys is thus not at all evident. This contrasts with some earlier work which reports that the wear resistance improves with increasing Si and hardness [11–13]. The present authors have also reported such a trend in similar thixoformed alloys . However, the Si content of the thixoformed alloys investigated in  ranged between 8 and 14 wt% and did not
Y. Birol, F. Birol / Wear 265 (2008) 1902–1908
Fig. 5. Microstructures of thixoformed alloys after T6 heat treatment: (a) alloy 1, (b) alloy 2, (c) alloy 3 and (d) alloy 4.
Fig. 4. Microstructures of thixoformed alloys: (a) alloy 1, (b) alloy 2, (c) alloy 3 and (d) alloy 4.
Y. Birol, F. Birol / Wear 265 (2008) 1902–1908
Fig. 6. Change in hardness with Si content in thixoformed and heat-treated alloys.
Fig. 7. Features of the worn surface of the ring test samples and the neighbouring original machined surface.
extend to higher levels as it did in the present work. Besides, the compositions of the alloys investigated in  did not allow a sound analysis of the wear properties in terms of Si content which has been a subject of great controversy over the years [11–28]. There is considerable amount of work which claims Si content to be insignif-
Fig. 8. Particle fragmentation on the worn surface of alloy 4.
Fig. 9. Change in speciﬁc wear rates with Si content for test samples in the asthixoformed and heat-treated states.
icant regarding wear resistance [27,28] while others have reported a minimum in wear rate either at the eutectic composition [23,29,30] or at higher Si levels [14–16]. It is suspected that the role of Si particles may be different in thixoformed Al–Si-based alloys from that in conventional cast grades. After all, processing of Al–Si alloys is well established to impact their wear properties through its effect on Si particle characteristics [7,12,19,31]. Ward et al.  reported a decrease in wear resistance with increasing Si in thixoformed alloys and blamed the difﬁculty of thixoforming high Si alloys for their inferior wear properties. It should be noted, however, that the HPDC grade of the near-eutectic alloy (alloy 2) wears nearly 3 times more under the same conditions , implying that thixoforming is a viable alternative as the manufacturing route. The features of the worn surfaces are fully consistent with the speciﬁc wear rate measurements and with the microstructures of the respective alloys (Fig. 10). The grooves depicted on the worn surfaces of the hypoeutectic alloy (alloy 1) are shallow and uniform (Fig. 10a). The situation is much improved in the case of the near-eutectic alloy (alloy 2) which shows a more or less homogeneous surface with limited localized wear damage (Fig. 10b). The hypereutectic alloy (alloy 4), on the other hand, shows, in addition to prominent wear marks, a heterogeneous distribution of cavities some of which are very large and deep (Fig. 10c). The vast majority of the coarser particles which were once perfectly embedded in the matrix are either sitting inside these cavities fragmented (Fig. 8) or entirely missing. Some Si depletion from the worn surfaces is evidenced by EDS analysis which has shown the fragments to be invariably of the Si variety. The intermetallic particles, on the other hand, were almost always retained in the matrix. Whether in eutectic or primary form, Si phase in Al–Si-based alloys exists in the microstructure as discrete particles due to a very limited solid solubility in aluminium. The interfacial regions between the Si particles and the matrix are thus quite prone to cracking and the tendency for cracking increases with the size of the particles . The complex stress state involved in sliding wear, leads to fragmentation of the coarse primary silicon particles which, as such, can no longer be accommodated by the aluminium matrix. Coarse primary Si particles eventually drop out from the matrix leaving behind large surface cavities and a bare matrix in contact with the counter-surface during the wear process (Fig. 11). Eutectic Si particles, on the other hand, are much ﬁner and more uniformly
Y. Birol, F. Birol / Wear 265 (2008) 1902–1908
Fig. 10. Features of the worn surfaces of the ring test samples in the as-thixoformed state: (a) alloy 1, (b) alloy 2 and (c) alloy 4.
distributed in the matrix than the primary Si phase. The cracking tendency of eutectic particles is substantially reduced which in turn allows the Si phase to carry load more effectively. A neareutectic alloy, such as alloy 2 in the present work, enjoys the largest population of such favourable particles and a much improved wear resistance. It is inferred from the foregoing that the hypereutectic alloys are clearly at a disadvantage. Their inferior wear properties may be
attributed to debonding of large primary Si particles at the matrix/Si interface and to the predominating cracking tendency of these particles. This situation is probably aggravated in thixoformed alloys due to the oxidation of the mushy alloy in the semisolid temperature range producing even weaker interfaces between the Si particles and the aluminium matrix. The fragmented Si particles not only contribute to the weight loss directly by dropping out from the surface but also act as abrasives and lead to additional abrasive wear (Fig. 12). It is thus fair to claim that the coarse primary Si particles hurt, rather than improve, the wear performance of the thixoformed Al–Si-based alloys under severe wear conditions. The T6 heat treatment provided considerable hardening in all alloys yet failed to improve their wear resistance. The ranking of the alloys in terms of speciﬁc wear rates did not change either as conﬁrmed by the analysis of the worn surfaces of the respective alloys (Fig. 13). The near-eutectic alloy (alloy 2), once again, is the most wear-resistant in the T6 temper. This is not surprising considering that the heat treatment did not change the microstructural features of the thixoformed alloys in an appreciable fashion. It is evident from the performance of the thixoformed and heat-treated alloys that the wear performance of the thixoformed Al–Si-based alloys is governed largely by the size and distribution of coarse Si particles, and that the impact of hardness is only secondary. While heat treatment is reported to affect the wear resistance of Al–Si alloys, often in a favorable way [13,19,32,33], this does not necessarily hold true for thixoformed alloys. No systematic effect of heat treatment was found on the wear properties of similar thixoformed alloys in recent studies [7,8].
Fig. 11. Wear mechanism in hypereutectic AlSiCuFe alloys.
Fig. 12. Evidence of abrasive wear produced by loose Si particles.
Y. Birol, F. Birol / Wear 265 (2008) 1902–1908
Fig. 13. Features of the worn surfaces of the ring test samples in the heat-treated state: (a) alloy 1, (b) alloy 2 and (c) alloy 4.
4. Conclusions There appears to be a clear distinction in the wear properties of thixoformed AlSiCuFe alloys with respect to the eutectic composition under severe wear conditions. Wear rates which decrease until 12 wt% Si start increasing at hypereutectic Si levels. Hence, of the four alloys tested in the present work, the near-eutectic alloy is the most wear-resistant. The inferior wear properties of thixoformed hypereutectic alloys are accounted for by the response of the coarse Si particles to severe wear conditions. The majority of the primary Si particles fracture, and as such, can no longer be accommodated by the aluminium matrix. They eventually drop out during the wear test leaving large cavities on the surface. Having failed to change the microstructural features in an appreciable fashion, the T6 heat treatment did not improve the wear resistance in spite of increased hardness. It is inferred from the wear rates of thixoformed and heat-treated samples that the wear properties of the thixoformed AlSiCuFe alloys is governed largely by the size and distribution of coarse Si particles and that the impact of hardness is only secondary. Acknowledgements It is a pleasure to thank O. C¸akır and F. Alageyik for their help in the experimental part of this work. The ﬁnancial support of the State Planning Organization of Turkey is gratefully acknowledged. References  Y. Birol, F. Birol, Wear properties of thixoformed and high pressure die cast aluminium alloys for connecting rod applications in compressors, in: E. Cueto, F. Chinesta (Eds.), 10th ESAFORM Conference on Material Forming, Zaragoza, Spain, 2007, pp. 1167–1172.
   
                           
G.C. Pratt, Int. Mater. Rev. 18 (1973) 1. P.K. Rohatgi, R. Asthana, S. Das, Int. Met. Rev. 131 (1986) 115. L. Lasa, J.M. Rodriguez-Ibabe, Scripta Mater. 46 (2002) 477–481. P. Kapranos, D.H. Kirkwood, H.V. Atkinson, J.T. Rheinlander, J.J. Bentzen, P.T. Toft, C.P. Debel, G. Laslaz, L. Maenner, S. Blais, J.M. Rodriguez-Ibabe, L. Lasa, P. Giordano, G. Chiarmetta, A. Giese, J. Mater. Process. Technol. 135 (2003) 271–277. S. Midson, J. Keist, J. Svare, SAE 2002 World Congress, Detroit, Michigan, March 4–7, 2002, 2002-01-394. P.J. Ward, H.V. Atkinson, P.R.G. Anderson, L.G. Elias, B. Garcia, L. Kahlen, J.-M. Rodriguez-ibabe, Acta Mater. 44 (1996) 1717–1727. Y. Birol, F. Birol, Wear 265 (2008) 590–597. Y. Birol, J. Mater. Sci. 43 (2008) 3577–3581. C.M. Chen, C.C. Yang, C.G. Chao, Mater. Sci. Eng. A 397 (2005) 178–189. D.K. Dwivedi, Mater. Des. 27 (2006) 610–616. S.C. Lim, M. Gupta, Y.F. Leng, E.J. Lavernia, J. Mater. Process. Technol. 63 (1997) 865–870. M. Harun, I.A. Talib, A.R. Daud, Wear 194 (1996) 54–59. V.C. Srivastava, S.N. Ojha, Mater. Sci. Technol. 20 (2004) 1632–1638. B.K. Prasad, K. Venkateswaralu, O.P. Modi, A.K. Jha, S. Das, R. Dasgupta, A.H. Yegneswaran, Metall. Mater. Trans. 29A (1998) 2747–2752. F.A. Davis, T.S. Eyre, Tribol. Int. 27 (1994) 171–181. A.K. Prasada Rao, B.S. Murty, M. Chakraborty, Mater. Sci. Eng. A 395 (2005) 323–326. F. Wang, H. Liu, Y. Ma, Y. Jin, Mater. Des. 25 (2004) 163–166. L. Lasa, J.M. Rodriguez-Ibabe, Mater. Sci. Eng. A363 (2003) 193–202. G. Timmermans, L. Froyen, Wear 230 (1999) 105–117. A.S. Reddy, B.N.P. Bai, K.S.S. Murthy, S.K. Biswas, Wear 171 (1994) 115–127. H. Torabian, J.P. Pathak, S.N. Tiwari, Wear 172 (1994) 49– 58. K.M. Jassim, E.S. Dwarakadasa, J. Mater. Sci. Lett. 11 (1992) 421–423. V.K. Kanth, B.N.P. Bai, S.K. Biswas, Scripta Metall. Mater. 24 (1990) 267– 272. R. Antoniou, D.W. Borland, Mater. Sci. Eng. 93 (1987) 57. A.D. Sarkar, J. Clarke, Wear 61 (1980) 157–167. B.N.P. Bai, H.S. Shamasunder, S.K. Biswas, Mater. Sci. Eng. A104 (1988) L11–L14. B.N.P. Bai, S.K. Biswas, Wear 120 (1987) 61–74. J.D. Clarke, A.D. Sarkar, Wear 54 (1979) 7–16. R.M. Farrel, T.S. Eyre, Wear 15 (1970) 359–372. F. Wang, Y. Ma, Z. Zhang, X. Cui, Y. Jin, Wear 256 (2004) 342–345. R.D. Ott, C.A. Blue, M.L. Santella, P.J. Blau, Wear 251 (2001) 868–874. B.N.P. Bai, S.K. Biswas, Acta Metall. Mater. 39 (1991) 833.