The Effect of Precipitate State on the Creep Resistance of Mg-Sn Alloys

Transcription

1 The Effect of Precipitate State on the Creep Resistance of Mg-Sn Alloys 1 M. A. Gibson, 2 X. Fang, 3 C. J. Bettles* and 3 C. R. Hutchinson 1 CAST CRC, CSIRO Process Science & Engineering, Private Bag 33, Clayton South MDC, Clayton, 3169, Victoria, Australia 2 Monash Centre for Electron Microscopy (MCEM), Monash University, 3800 VIC, Australia 3 ARC Centre of Excellence for Design in Light Metals, Department of Materials Engineering, Monash University, Clayton, 3800, Victoria, Australia *corresponding author. Phone , Fax , colleen.bettles@monash.edu Abstract The tensile creep performance of Mg-Sn alloys, containing Zn and Zn+Na alloying additions, at temperatures of 177 C and 150 C have been measured. Under a load of 60MPa, the Mg- Sn-Zn-Na alloy shows secondary creep rates five orders of magnitude slower than the binary alloy and approximately three orders of magnitude slower than the ternary Mg-Sn-Zn alloy. Transmission electron microscopy shows that the origin of the excellent creep properties of the Mg-Sn-Zn-Na alloy is a remarkable thermal stability of the precipitate structure. Keywords: magnesium alloys; creep; precipitation; There is currently much interest in the potential use of magnesium alloys as structural materials in automobiles as a means of reducing mass by substitution for heavier aluminium and steel components. The increasing demands of fuel efficiency suggest that this interest will continue if magnesium alloys that realise some of the potential can be developed. An application receiving particular attention is the use of magnesium in powertrain applications. Applications in transmission housings and crankcases require high performance at elevated temperatures (>150 C) and under these conditions the stability of microstructure (and therefore properties) is of key importance. Those alloys that derive their strength from precipitation hardening are usually considered the most promising for such applications and the thermal stability of the precipitate distribution is an important component of the overall microstructural stability.

2 One alloy system that has recently received considerable interest is the Mg-Sn system. This is an age hardenable system with the potential to form ~10 vol. % of Mg 2 Sn. The Mg 2 Sn precipitate (FCC, a=0.676nm, point group m3 m ) [1] has a high melting temperature (~770 C) and alloys based on this system might be thought to show some promise for use in applications requiring elevated temperature creep resistance. Precipitation in the binary system was characterised many years ago: Mg 2 Sn forms from the supersaturared matrix with a plate/lath shaped morphology and does not appear to be preceded by any metastable phases at typical aging temperatures [2]. The age hardening response is only modest, despite the large volume fraction of precipitates that may be formed [3]. The reason for the poor precipitate strengthening response is that the precipitates form, in a relatively coarse distribution, with a lath shaped morphology parallel to the basal planes of the Mg matrix. This configuration is relatively inefficient at impeding dislocation motion [4]. Mendis et al. [4,5] have recently pursued two avenues for improving the room temperature precipitation hardening response of Mg-Sn based alloys. The first approach involved using microalloying additions to catalyse a greatly refined distribution of precipitates. The basic idea was to create heterogeneities in the matrix at which Mg 2 Sn may precipitate. A qualitative thermo-kinetic criteria for choosing microalloying additions was proposed and, using additions of Na, a two orders of magnitude increase in the number density of precipitates and an augmentation of the hardening increment of 270% were achieved [4]. In a second, more ambitious, approach, Mendis et al. used Zn additions to attempt to manipulate the shape of the Mg 2 Sn precipitates [5]. The premise was to manipulate the lattice misfit normal to the basal plane of the magnesium matrix in an effort to change the shape of the Mg 2 Sn precipitates. This was also shown to be successful and a substantial increase in age hardening was realized [5]. A similar result has also been reported by Sasaki et al. [6]. As a final step, the two approaches were used simultaneously (Na+Zn additions) realizing a further increase in the precipitation hardening response. In all cases the precipitate remained almost exclusively Mg 2 Sn. The age hardening responses of the binary alloy, the alloy containing Zn, and the alloy containing Zn+Na have been reported by Mendis et al. [5].

3 However, designing microstructures that provide high strength (room temperature or elevated temperature) is not the same as designing materials for creep resistance. In addition to static strength, creep resistance also requires stability of microstructure. Under conditions where the creep strain is accommodated mainly by dislocation motion, stability in the precipitate state is crucial for creep resistance. Approaches to strengthening that involve an increase in precipitate number density suffer from a potential disadvantage for applications requiring creep resistance in that, for a given volume fraction of precipitates, increases in the number density necessarily correspond to a decrease in the mean particle size (R). The driving force for coarsening processes is the excess energy stored in the particle/matrix interfaces and the coarsening rate, dr, of a particle distribution scales as 1 2, e.g. [7]. Thus, the increase in dt R strength of precipitate strengthened alloys achieved through refinements in scale may come at the cost of a decrease in the thermal stability of the precipitate distribution. The Mg-Sn-Zn alloys examined by Mendis et al. did not show a large change in precipitate number density (a factor of 2 increase only) [5]. The increase in strengthening was mainly due to changes in precipitate shape. As a result, the Zn additions are not expected to adversely affect the thermal stability and, therefore, the creep properties of such alloys might be expected to be interesting. However, the substantial increase in strength realized by the alloy containing Zn+Na was achieved by a two orders of magnitude increase in the precipitate number density. A surprising result was that the long term aging behaviour of the alloy containing Zn+Na did not result in significant softening of the alloy, even for times approaching 1000h at 200 C. This result could not have been predicted using the approaches outlined by Mendis et al. for selecting alloying additions [4,5]. Such stability in properties indicates stability in microstructure and suggests that this alloy may have very interesting creep properties. In this work, the tensile creep properties of a binary Mg-Sn alloy, a ternary Mg-Sn-Zn alloy and an alloy containing both Zn and Na additions are examined. Three alloys, with compositions shown in Table 1, were cast from high purity elements into a sand mould under an atmosphere of AM-cover (0.2wt.% tetrafluoroethane and 99.8wt.% nitrogen gas) to protect the melt. The Na was added to the quaternary alloy in the form of a Sn-Na master alloy. The alloys were heat treated at 345 C for 2h prior to heating to 500 C

4 and homogenisation for 6h followed by quenching. Each of the three alloys was then aged to peak strength at 200 C. Peak strength for the binary, ternary and quaternary alloys corresponded to ~1000h, ~200h and ~7h, respectively. The average grain size of the alloys was ~ m. Tensile creep samples were spark discharge machined from the ingots, with gauge dimensions of 25mm x 5mm x 3mm. Creep tests were performed in oil at 177 C with a load of 60MPa. Selected tests at 150 C or with loads of 35MPa were also performed. Table 1 The precipitate structures, before and after creep deformation, were observed using conventional transmission electron microscopy with a Phillips CM-20 operating at 200kV. The foils were prepared by low-angle ion milling using a Gatan Precision Ion Polishing System. An accelerating voltage of 5.0 kev and a gun current of 18µA were used for the first step with an operating angle of 4º to produce a tiny hole in the middle of the foil. The foils were then thinned using the acceleration voltage of 3.0 kev, gun current of 6µA and operating angle of 2º. The tensile creep strain as a function of loading time at 177 C under a load of 60MPa is shown in Fig. 1a for the three alloys. The creep performance of the binary alloy is quite poor and this is consistent with previous reports [8] and expectations based on the low static strength of this alloy. The performance of the ternary alloy containing Zn is much improved compared with the binary alloy but the performance of the Zn+Na containing quaternary alloy is rather striking. The creep performance is orders of magnitude better than both the binary and ternary alloys. The creep performance of the Zn+Na containing quaternary alloy is plotted in Fig. 1b with an extended time axis to illustrate the creep lifetime. Lifetimes in excess of 400h were obtained under tensile loading conditions at both 177 C and 150 C. The tests were stopped soon after 400h. Figure 1 The secondary creep rates are summarised in Table 2 and are compared with other common magnesium alloys reported in the literature. Table 2

5 The quaternary Zn+Na containing alloy has a creep rate approximately five orders of magnitude lower than the binary alloy under loads of 60MPa. The creep performance is two orders of magnitude better than AZ91 and one order of magnitude better than the creep resistant Mg-Sn-Al-Si (TAS831) alloy at 150 C. One of the objectives of this work is to develop reasonably low cost creep resistant magnesium alloys. A suitable benchmark with which to compare creep performance may be the AE42 alloy. As can be seen in Table 2, the Mg-Sn-Zn-Na alloy has secondary creep rates comparable to, or a little better than, the AE42 alloy, although it must be acknowledged that the AE42 alloy was tested in the die cast condition. The creep rates of the Mg-Sn-Zn-Na alloy are also shown to be comparable to a series of RE containing alloys that have recently been tested under the same conditions [11]. The peak aged microstructures of Mg-Sn, Mg-Sn-Zn and Mg-Sn-Zn-Na alloys have been characterized in the past [4,5]. The microstructures of the Mg-Sn alloys used in this study are not reproduced here. Transmission electron micrographs of the peak aged microstructures of the Mg-Sn-Zn and Mg-Sn-Zn-Na alloys used in this study, taken with the electron beam closely parallel to both the c r and a r axes of the Magnesium lattice, are shown in Figs. 2 and 3 (a and b), respectively. The microstructures are consistent with those previously reported [5]. The dominant effect of Zn additions is to change the shape of the Mg 2 Sn precipitates (from laths elongated along the basal planes of the Mg matrix to more equiaxed precipitates, cf. [4,5]) without a large change in the number density of precipitates. Microbeam electron diffraction (MBED) was used to confirm that the precipitate identity remains Mg 2 Sn. The addition of both Zn and Na leads to a large increase in the number density of precipitates that form, and a more equiaxed shape than those in the binary alloy. This is also consistent with previous reports [5]. Figure 2 Figure 3 Microstructures for the two alloys in the as-crept condition are shown in Figs. 2 and 3 (c and d) respectively. As can be seen in Fig. 2, there is a clear coarsening of the Mg-Sn-Zn microstructure. MBED confirms that the precipitate identity remains Mg 2 Sn. In contrast, the as-crept microstructure of the Mg-Sn-Zn-Na alloy (Fig. 3) shows very little change from the

6 peak aged structure, confirming the thermal stability of the precipitate structure that was suggested by the long term aging behaviour. The excellent creep resistance of the Mg-Sn-Zn- Na alloy is linked to the excellent thermal stability of this microstructure. The important question is: What is the origin of the thermal stability of the precipitates in the Mg-Sn-Zn-Na alloy? Since such thermal stability is not exhibited by the alloy containing Zn additions only, it is tempting to look to the Na for an explanation. As outlined by Mendis et al. [4], Na additions were chosen because, based on thermodynamic grounds, they were expected to cluster on the Mg lattice during aging. The rationale was that the heterogeneities that form will act as potent nucleation sites for the Mg 2 Sn precipitates. If the rate of formation of the Na clusters, and the scale with which they form, is faster and finer than the intrinsic kinetics and scale of the Mg 2 Sn precipitation in the binary alloy, then a refined distribution of Mg 2 Sn precipitates would be expected. Since the Na additions are expected to be intimately linked with the Mg 2 Sn precipitates, it may be expected that the Na is either incorporated into the Mg 2 Sn precipitate as it grows or may be located at the precipitate/matrix interface. In either case, it would not be surprising for such segregation to affect the lattice misfit between the precipitate and matrix, and therefore the interfacial energy which drives coarsening processes, or in the case of incorporation of the Na into the Mg 2 Sn precipitates, the intrinsic thermodynamic stability of the Mg 2 Sn precipitates. The observed beneficial effects on the thermal stability of the Mg 2 Sn precipitate distribution in the Mg-Sn-Zn-Na alloy could not have been predicted using the alloy element selection procedures outlined by Mendis et al. [4,5]. More sophisticated calculations would be required. The obvious next step is to experimentally identify the exact location of the Na additions in the microstructure, and any possible association with the Zn additions. This information would aid in the optimisation of the microalloying additions for the most beneficial impact on the mechanical properties. In addition, at present this alloy system cannot be grain refined by any commercial grain refining agents. It is expected that the yield strength would be increased if a finer as-cast grain size could be achieved, and this would lead to an increase in the maximum load that could be accommodated prior to the creep limit being exceeded. These issues are the subject of current work. This work illustrates that, through the judicious selection of alloying additions, it is possible to very substantially affect the size and shape of Mg 2 Sn precipitates in Mg-Sn based alloys.

7 These modifications are beneficial for the tensile creep performance of these alloys. In particular, The Mg-Sn-Zn-Na alloy exhibits excellent thermal stability of the precipitate structure leading to tensile creep properties that are of practical significance and have a five orders of magnitude improvement in secondary creep rate compared to the binary Mg-Sn alloy. ACKNOWLEDGEMENTS The CAST Co-operative Research Centre was established under, and is funded in part by, the Australian Government's Co-operative Research Centres Scheme. The Australian Research Council is gratefully acknowledged for funding the ARC Centre of Excellence for Design in Light Metals. The assistance of Mr. Daniel Graham in the sand casting of the alloy plates used in this work is gratefully acknowledged. REFERENCES [1] Villars, P., Calvert, L.D. in Pearson's Handbook of Crystallographic Data for Intermetallic Phases, Materials Park, ASM International, [2] Derge, G., Kommell, A.R., Mehl, R.F., Trans. AIME., 1937, vol. 124, p [3] van der Planken, J., J. Mat. Sci. Letts., 1969, vol. 4, p [4] Mendis, C.L., Bettles, C.J., Gibson, M.A., Gorsse, S., Hutchinson, C.R., Philosophical Magazine Letters, 86, p. 443, [5] Mendis, C.L., Bettles, C.J., Gibson, M.A., Hutchinson, C.R., Materials Science and Engineering A, , p. 163, [6] Sasaki, T.T., Oh-ishi, K., Ohkubo, T., Hono, K., Scripta Materialia, 55(3), p , [7] Martin, J.W., Doherty, R.D., Cantor, B., Stability of Microstructure in Metallic Systems. Cambridge, Cambridge University Press, [8] Wei, S., Chen, Y., Tang, Y., Liu, H., Xiao, S., Niu, G., Zhang, X., Zhao, Y., Materials Science and Engineering A., 492, p. 20, 2008.

8 [9] Kang, D.H., Park, S.S., Oh, Y.S., Kim, N.J., Materials Science and Engineering A, , p. 318, [10] Bai, J., Sun, Y.S., Xun, S., Xue, F., Zhu, T.B., Materials Science and Engineering A, 419, p. 181, [11] Zhu, S.M., Gibson, M.A., Easton, M.A., Nie, J.F., Scripta Materialia, 2010, in press.

9 FIGURE CAPTIONS Figure 1. Tensile creep curves for a) the Mg-Sn, Mg-Sn-Zn and Mg-Sn-Zn-Na alloys at 177 C under a load of 60MPa and b) the Mg-Sn-Zn-Na alloys at 177 C and 150 C under a load of 60MPa. These tests were stopped after ~ 400h. Figure 2. Bright field TEM micrographs of the Mg-Sn-Zn alloy before (a and b) and after (c and d) creep loading at 177 C with 60MPa for a duration of ~4h. (a and c) b [0001], b and d) b [1120]) Figure 3. Bright field TEM micrographs of the Mg-Sn-Zn-Na alloy before (a and b) and after (c and d) creep loading at 177 C with 60MPa for a duration of ~400h. (a and c) b [0001], b and d) b [1120])

10 Table 1. Compositions of the three Mg-Sn based alloys considered in this work Alloy Composition (wt. %) Sn Zn Na Mg-Sn <0.005 Mg-Sn-Zn <0.005 Mg-Sn-Zn-Na

11 Table 2. Experimentally measured secondary tensile creep rates and comparison with values reported from the literature Alloy Secondary creep rates (s -1 ) 150 C/ C/35 MPa 177 C/60 MPa MPa Mg-Sn 1.2 x x x 10-5 Mg-Sn-Zn 8.4 x x x 10-6 Mg-Sn-Zn-Na 5.6 x x x 10-9 AZ x 10-8 [9]* Mg-Sn-Al-Si 4 x 10-9 [9]* (TAS831) AE x 10-8 [10] # Mg-3.44wt.%La 2.1 x 10-9 [11] Mg-2.87wt.%Ce 1.2 x 10-9 [11] Mg-2.60wt.%Nd 3.5 x [11] These values are only upper limit estimates as the material was still in the primary creep stage when the tests were terminated. * The TAS831 and AZ91 alloys were tested in the die cast condition under a load of 50MPa. # The AE42 alloy was tested in the die cast condition at 175 C under a load of 70MPa.

12 Figure 1

13 Figure 2

14 Figure 3