- Email: [email protected]
Enhancement of properties in cast Mg–Y–Zn rod processed by severe plastic deformation
Materials Science & Engineering A 615 (2014) 198–207
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Enhancement of properties in cast Mg–Y–Zn rod processed by severe plastic deformation Rimma Lapovok a,n, Xiang Gao a, Jian-Feng Nie a, Yuri Estrin a, Suveen N. Mathaudhu b a b
Materials Engineering Department, Monash University, Clayton, Vic 3800, Australia Materials Science Division, U.S. Army Research Office, Research Triangle Park, NC 27709, USA
art ic l e i nf o
a b s t r a c t
Article history: Received 10 June 2014 Received in revised form 21 July 2014 Accepted 21 July 2014 Available online 30 July 2014
Ternary Mg–Y–Zn alloys have attracted considerable attention from researchers due to their excellent mechanical properties and unique microstructures, particularly from the presence of long-period stacking-order (LPSO) phases. Microstructural variations and the resulting mechanical properties can be affected by various processing routes, particularly those involving severe plastic deformation of a cast billet. The approach used in this work was based on subjecting cast Mg92Y4Zn4 (composition in wt%) billet to severe plastic deformation by three different routes, namely equal channel angular pressing (ECAP), high pressure torsion (HPT) and ECAP followed by HPT, with the aim of refining the microstructure and improving mechanical properties. Samples processed by ECAP were annealed by post-processing and tested in compression and tension. The effect of the processing route and the process parameters on the microstructure and the hardness of the Mg–Y–Zn alloy is reported. An overall positive effect of annealing treatment on the mechanical properties of ECAP-processed alloy is demonstrated. & 2014 Elsevier B.V. All rights reserved.
Keywords: Mg–Y–Zn alloys LPSO phases Severe plastic deformation Equal channel angular pressing High pressure torsion
1. Introduction Addition of rare-earth (RE) elements is known to increase the strength and to improve the creep properties of magnesium alloys at high temperatures due to the formation of intermetallic phases [1–4]. Most RE-containing magnesium alloys show an excellent precipitation hardening response [1–4]; however, Mg–Y–Zn ternary alloys are not precipitation hardenable [5]. Superior mechanical properties in these alloys are achieved by other techniques, for example, warm extrusion of rapidly solidified powder (strength of up to 600 MPa and 5% tensile ductility) [6,7], or working of cast billet [8–10], which results in strengthening by grain refinement. The microstructure after such processing usually contains a mixture of nano-scale magnesium grains (100–200 nm) and intermetallic particles that exhibit a range of long-period stackingorder (LPSO) structures, among which 18R and 14H are the most common [11]. The notation used to designate the LPSO intermetallic phases is defined in Ref. [11]. A unique feature of such LPSO structures is their lamellar morphology that comprises alternating layers of magnesium and intermetallic lamellae. The accumulated experimental evidence indicates that mechanical properties of the Mg–Y–Zn alloys can be tailored by proper control of the volume n
Corresponding author. Tel.: +61 399059294; fax: +61 399054940. E-mail address: [email protected] (R. Lapovok).
http://dx.doi.org/10.1016/j.msea.2014.07.068 0921-5093/& 2014 Elsevier B.V. All rights reserved.
fraction and size of the lamellae of the LPSO structures. It is expected that an ultrahigh strength alloy may be obtained if a nano-scale distribution of LPSO structures is achieved. It is not possible to achieve a redistribution of LPSO structures by deformation with conventional metal forming processes, as they have a geometrical limitation on the amount of strain imparted to the material. However, by severe plastic deformation (SPD) methods [12] a billet can be subjected to virtually unlimited strain without a change in its shape and dimensions. The most well-known SPD methods are equal channel angular pressing (ECAP) [13] and high pressure torsion (HPT) [14]. The effects of SPD on the mechanical properties of Mg–Y–Zn alloys have not been explored to date. It can be expected, however, that severe plastic deformation may lead to much finer grains and potential dispersion of the LPSO structures, and therefore much higher yield strengths. A comparative study of microstructure and properties of Mg92Y4Zn4 cast billets severely deformed by different SPD techniques has been performed in this work.
2. Materials and experimental methods Mg92Y4Zn4 (composition in wt%) billet was prepared by permanent mould casting followed by homogenisation for 24 h at 450 1C in a muffle furnace. After homogenisation the samples were
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
quenched in water to 70 1C. Cast billets were subjected to severe plastic deformation by three different routes, namely ECAP, HPT and ECAP followed by HPT. Samples 10 mm in diameter and 37 mm in length were machined from the homogenised billet. Four passes of ECAP (route BC in which the specimen is rotated by 901 about the pressing direction between the passes) were performed using a 901 die at 350, 450 and 500 1C under a back-pressure of 185 MPa (Fig. 1a). ECAP with back-pressure was performed using a specially designed unit equipped with a heated die to permit isothermal processing within the temperature range of 20–500 1C. Backpressure was controlled through a horizontal hydraulic cylinder to maintain a pre-set pressure level, while the unit was placed in a
199
1000 kN press to exercise the forward pressure. Each pass of ECAP introduced a strain of ε ¼1.15 in the material. In a separate set of experiments, disk shaped samples of 1 mm thickness and diameter of 8 mm were cut from as-cast and ECAP processed samples and subjected to 10 revolutions in an HPT die (Fig. 1b) under a pressure of 1.9 GPa at room temperature. Each revolution introduced a strain around ε ¼ 14 near the rim of the specimen. The schedule of experiments is shown in Table 1. Annealing after ECAP was carried out in a range of temperatures from 200 to 500 1C and for different times from 10 to 30 min. After all deformation routes the microstructure was examined using an Olympus GX51 optical microscope (OM) and a Philips
Fig. 1. Severe plastic deformation processes used in this research: (a) ECAP with back-pressure and (b) HPT. Table 1 Experimental schedules. ECAP temperature, 1C ECAP HPT ECAP þHPT ECAP þAnn
350 – 350 –
HPT temperature, 1C 450 – 450 450
500 – – 500
Fig. 2. Initial microstructure of cast billet before deformation: (a) as-cast and (b) homogenised.
– 20 20 –
200
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
A
<111>
Precipitates 01-1 110 101
Mg-matrix
<001>
B
-200 020
0002 α
20 nm
50 nm
-1101 α -4221 14 H
0002 14 H
14H
Mg-matrix
Fig. 3. TEM images of microstructure after homogenisation: (a) precipitates and (b) LPSO particles (BF images and SAED patterns).
Fig. 4. Microstructure of deformed billet after ECAP at 350 1C: (a) one pass and (b) four passes.
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
201
Fig. 5. TEM bright field (BF) images of deformed billet after 4 passes of ECAP at 350 1C: (a) magnesium matrix; (b) BF image and SAED pattern recorded from a 14H particle; (c) BF image and MBED pattern recorded from a Mg3Y2Zn3 (fcc: a¼ 0.6848 nm) particle.
Fig. 6. OM of samples deformed by room temperature HPT with initial conditions: (a) as-cast (arrows indicate shear bands) and (b) ECAP processed at 350 1C.
202
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
Fig. 7. TEM images of samples deformed by 4 passes ECAP at 350 1C followed by room temperature HPT: (a) magnesium matrix; (b) and (c) BF images and SAED pattern recorded from a 14H particle; (d) BF images and MBED patterns recorded from Mg3Y2Zn3 (fcc: a¼ 0.6848 nm) particles.
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
CM20 transmission electron microscope (TEM) operated at 200 kV. Vickers hardness (ASTM E384-10e2) was measured using a Struers Duramin A-300 hardness tester with a load of 0.1 kg (HV0.1). Seven indents were produced per sample. For XRD analysis, a Philips X-ray diffractometer equipped with a Cu anode operated at 40 kV and 25 mA was used.
3. Results and discussion 3.1. Microstructure The microstructure of the as-cast alloy contains a large volume fraction of relatively fine-scale lamella (Fig. 2a) which are reported to have high concentrations of yttrium and zinc [5]. This microstructure is quite different from that normally observed in as-cast Mg-alloys which typically consist of magnesium grains and secondary phases in grain boundaries. This lamellar structure represents a mixture of 18R and α-Mg phases [5,11,15–18]. The eutectic formed between the lamellar structures, seen as the dark regions in Fig. 2a, is expected to be comprised of the intermetallic W-phase and the α-Mg phase [19,20]. After homogenisation the lamellar platelets are still the dominating feature in the microstructure, but then they have become much coarser. The eutectic is still observed
203
in regions isolated by the lamellar platelets (Fig. 2b). The 18R phase formed in the as-cast microstructure generally transforms to 14H after prolonged heat treatments at high temperatures [5,11,24]; however, the microstructure of the studied alloy after homogenisation contained α-Mg grains (Fig. 3a) and both 18R and 14H LPSO particles (Fig. 3b). The diffraction pattern obtained suggested that it has a BCC crystal structure. During severe plastic deformation by ECAP at 350 1C, the strain is distributed non-uniformly between the skeleton LPSO phase and magnesium matrix, with the strain mainly supported by the LPSO phase. After the first ECAP pass, a strain of about 1.15 is introduced into the magnesium matrix, and this level of strain appears to be large enough to deform the LPSO structure, which becomes wavy, as can be seen in Fig. 4a. As the number of passes increased to four, a significant amount of strain (ε ¼4.6) was introduced into both the matrix and LPSO phase (Fig. 4b). This gives rise to extreme grain refinement (grain size around 200–300 nm) in magnesium matrix, which can be seen in TEM images (Fig.5a). The microstructure produced by four ECAP passes also contains 14H particles and coarse Mg3Y2Zn3 (face-centred cubic: a ¼0.6848 nm) particles (Fig. 5b and c). A closer inspection reveals that there exist several domains or kink bands [21–24] within the 14H particles. Each 14H domain or kink band is rotated at a small angle with respect to the surrounding domains, as
Fig. 8. TEM images of samples deformed by 4 passes ECAP at 450 1C: (a) magnesium matrix; (b and c) BF images of 14H particles; (d) SAED pattern recorded from a 14H particle.
204
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
Fig. 9. TEM images of samples deformed by 4 passes ECAP at 450 1C followed by room temperature HPT: (a) magnesium matrix; (b and c) BF images of 14H particles; (d) SAED pattern recorded from a 14H particle.
revealed by a small rotation of lattice fringes in the bright field (BF) image and some splitting of diffraction spots in the selected area electron diffraction (SAED) pattern, both shown by arrows in Fig. 5b. The micro beam electron diffraction (MBED) patterns recorded from a Mg3Y2Zn3 particle are shown in Fig 5c. More strain can be introduced if samples are subjected to additional HPT deformation. During HPT enormously high strain is produced and the adiabatic temperature increase is sufficient for dynamic recrystallisation (DRX) to occur. The microstructures of HPT samples with as-cast or after ECAP initial conditions are shown for comparison in Fig. 6. It can be seen that if HPT is applied to the as-cast sample, the plastic flow is unstable and shear bands are formed (Fig. 6a). Samples pre-strained by ECAP have LPSO platelets with dynamically recrystallised small grains in the regions between these platelets (Fig. 6b). The TEM study of samples processed by four ECAP passes at 350 1C followed by room temperature HPT also reveals fine α-Mg grains, 14H particles and coarse Mg3Y2Zn3 particles (Fig. 7a, b, d). As expected, far more severe deformation in the 14H phases is observed in this case than after one ECAP pass alone. It appears that there are kinks formed within 14H particles (Fig. 7c). The rotation angles of individual 14H domains with respect to their surrounding domains appear to be larger, as evident from the tilt of the lattice fringes in the
BF image and the split of the diffraction spots in the SAED pattern (Fig. 7b). The angle between domains D1 and D2 (as well as D3 and D4) is approximately 11 while the angle between domains D1 and D3 (as well as D2 and D4) is approximately 51. Samples deformed by four ECAP passes at 450 1C contained mainly fine α-Mg grains ( 1 μm) and 14H particles (Fig. 8). An inspection of the lattice fringe image of the 14H phase shows that there are some defects within the 14H particles (see arrows) (Fig. 8b). In addition, there appear several domains within an individual 14H particle. Similarly, the microstructure of samples after four passes ECAP at 450 1C followed by HPT also contains mainly fine α-Mg grains and 14H particles (Fig. 9). The microstructure of samples deformed by four ECAP passes at 500 1C also comprised of α-Mg grains and 14H particles, Fig. 10. A significantly smaller number of defects and few, if any, domains are detectable within a 14H particle. This may be attributed to the higher processing temperature. 3.2. Mechanical properties Mechanical properties were evaluated by microhardness measurements and, in the case of ECAP processed material, by uniaxial tension and compression testing. (Unfortunately, due to the small
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
205
Fig. 10. TEM images of samples deformed by 4 passes ECAP at 500 1C: (a) magnesium matrix; (b and c) BF image and SAED pattern recorded from a 14H particle.
Table 2 Vickers hardness of Mg92Y4Zn4 with different processing histories. Sample
LPSO Phase
Mg-grains
Hardness, HV (0.1) Homogenised ECAP 1 pass ECAP 4 pass HPT ECAP þ HPT
907 4 1277 7 1367 6 1537 5 1617 4
108 75 120 75 144 77 109 75
size of HPT and ECAP þHPT processed samples, no tensile or compression tests were possible.) Measurements on samples with the different processing histories studied showed an increase in microhardness by 46 HV due to severe plastic deformation in the LPSO phase after four passes of ECAP and a further increment of 25 HV as a result of subsequent HPT deformation (Table 2). The microhardness of the magnesium grains increased less significantly (18–30 HV), possibly due to dynamic recrystallisation or recovery within the magnesium matrix caused by heat release during HPT deformation (Fig. 7b). The unexpectedly large microhardness values in magnesium
grains in samples deformed by HPT only may be attributed to difficulties in isolating magnesium grains from the LPSO phase within the narrow bands of each. We note that the LPSO phases are the majority components of the microstructure which controls the load bearing capability of the alloy, hence the strengthening of these phases together with grain refinement of alpha-Mg matrix and dispersion of LPSO phase by the SPD processing significantly contribute to the overall strength of the material. It can thus be concluded from Table 2 that with the larger number of ECAP passes, greater strengthening is achieved. An additional HPT step raises the strength of the alloy even further. Comparing HPT with ECAP taken as two individual processes, we notice that HPT imparts a larger strengthening to the material than ECAP does. The tensile and compression tests conducted on ECAP processed samples revealed the following picture. While all ECAP processed samples exhibited strength well in excess of that of the homogenised as-cast material, their ductility was reduced. For example, after four ECAP passes at 500 1C, the tensile strength was 337 MPa and tensile ductility was 3.4%, while compression strength and ductility were 509 MPa and 3.7%, respectively. To improve ductility, several annealing treatments were performed on the samples that were subjected to four ECAP passes at 450 and
206
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
Fig. 11. True stress vs. true strain (compression test) for samples subjected to 4 ECAP passes at 450 1C (left) and 500 1C (right) followed by annealing treatment.
Fig. 12. True stress vs. true strain (tensile test) for samples subjected to 4 ECAP passes at 450 1C (left) and 500 1C (right) followed by annealing treatment at 400 1C.
500 1C. The true stress vs. true strain curves obtained in the compression tests are shown in Fig. 11. It can be seen that both a short high temperature annealing (500 1C/10 min) and a longer annealing at a lower temperature (300 1C/30 min) improve the mechanical properties of the SPD processed samples. To further investigate the influence of the annealing time on the mechanical properties, tensile samples were annealed at 400 1C for different times (Fig. 12). We note that after annealing at 400 1C the maximum strength in compression, around 550 MPa, was much higher than the ultimate tensile strength in tension (350–400 MPa). The ductility in compression was also very high: around 25–35%, as compared to that in tension (12–18%). It can be seen that SPD processing followed by short annealing leads to levels of ductility close to the ductility of homogenised cast material in compression ( 30%) and a multiple of that in tension (12–18% vs. 3%). For both types of loading, the strength of the SPD processed alloy with a short post-processing annealing was much higher (by 130– 250 MPa) than that of the homogenised cast alloy. A remarkable observation is that the sign of the tension vs. compression anisotropy, as given by the difference between the yield strength in tension and that in compression, was reversed by ECAP followed by annealing. While the magnitude of this asymmetry did not change much in terms of the absolute value, the relative
quantity (i.e., the magnitude of this difference divided by the higher of the two yield stress values) dropped quite appreciably, from about 0.6 to about 0.4.
4. Conclusions Cast Mg92Y4Zn4 (composition in wt%) billet was subjected to severe plastic deformation by three different routes, namely ECAP, HPT and ECAP followed by HPT. The LPSO phases were shown to deform by bending and rearrangement within a soft magnesium matrix, where a substantial grain refinement down to 200– 300 nm takes place. Further increases in plastic strain and the associated temperature rise appear to initiate dynamic recrystallisation and/or dynamic recovery of the magnesium matrix leading to a degree of softening therein. However, the overall microhardness of the material remains high, as most of the applied load is carried by the harder intermetallic LPSO particles. It can be concluded that high hardness and – by induction – high strength of Mg92Y4Zn4 rod can be achieved by severe plastic deformation processing. This conclusion was confirmed by uniaxial tension and compression tests of the material that underwent different levels of ECAP processing. It was also found that the increase of strength and hardness by SPD processing was achieved at the cost of
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
ductility. However, short-time annealing was demonstrated to increase ductility by about 30% in compression and 12–18% in tension (as compared to the as-homogenised alloy), while maintaining the tensile strength at a very high level of 550 MPa. The overall conclusion that can be made on the basis of the results presented is that the processing of the as-cast alloy Mg92Y4Zn4 by four passes of Route Bc equal channel angular pressing at 500 1C, followed by annealing for 12 min at 400 1C produces the most favourable combination of mechanical properties in the alloy. In view of a relatively small incremental change of microhardness (and presumably strength) of the material due to an extra HPT step subsequent to ECAP, we do not recommend the inclusion of this step in the processing route. Acknowledgement This project was supported by the U.S. Army International Technology Center – Pacific and the US Army Research Laboratory (FA5209-11-T-0043), under programme managers Dr. Christopher Drew and Dr. Vincent Hammond, respectively. References [1] C. Antion, P. Donnadieu, F. Perrard, A. Deschamps, C. Tassin, A. Pisch, Acta Mater. 51 (18) (2003) 5335–5348. [2] J.F. Nie, B.C. Muddle, Acta Mater. 48 (4) (2000) 1691–1703. [3] X. Gao, S.M. He, X.Q. Zeng, L.M. Peng, W.J. Ding, J.F. Nie, Mater. Sci. Eng. A 431 (1–2) (2006) 322–327. [4] P.J. Apps, H. Karimzadeh, J.F. King, G.W. Lorimer, Scr. Mater. 48 (8) (2003) 1023–1028.
207
[5] Y.M. Zhu, A.J. Morton, J.F. Nie, Mater. Sci. Forum 561–565 (Part 1) (2007) 151–154. [6] T. Inoue, Y. Kawamura, M. Matsushita, K. Hayashi, J. Koike, J. Mater. Res. 16 (7) (2001) 1894–1900. [7] Y. Kawamura, K. Hayashi, A. Inoue, T. Masumoto, Mater. Trans. 42 (7) (2001) 1172–1176. [8] X. Zheng, J. Dong, D. Yin, W. Liu, F. Wang, L. Jin, W. Ding, Mater. Sci. Eng. A 527 (16–17) (2010) 3690–3694. [9] K.U. Kainer, J. Wendt, K. Hantzsche, J. Bohlen, S. Yi, D. Letzig, Mater. Sci. Forum 654–656 (2010) 580–585. [10] J. Bohlen, S. Yi, D. Letzig, K.U. Kainer, Mater. Sci. Eng. A 527 (26) (2010) 7092–7098. [11] Y.M. Zhu, A.J. Morton, M. Weyland, J.F. Nie, Acta Mater. 58 (2) (2010) 464–475. [12] R.Z. Valiev, Y. Estrin, Z. Horita, T.G. Langdon, M.J. Zehetbauer, Y.T. Zhu, J. Met. 58 (4) (2006) 33–39. [13] R. Lapovok, The role of back-pressure in equal channel angular extrusion, J. Mater. Sci. 40 (2) (2005) 341–346. [14] J. Cizek, I. Prochazka, B. Smola, I. Stulikova, R. Kuzel, Z. Matej, V. Cherkaska, R. K. Islamgaliev, O. Kulyasova, Mater. Sci. Forum 503–504 (2006) 149–154. [15] K. Milicka, P. Pérez, F. Dobeš, G. Garcés, P. Adeva, Mater. Sci. Eng. A 510–511 (2009) 377–381. [16] D.H. Ping, K. Hono, Y. Kawamura, A. Inoue, Philos. Mag. Lett. 82 (2002) 543–551. [17] K. Hagihara, Y. Sugino, Y. Fukusumi, Y. Umakoshi, T. Nakano, Mater. Trans. 52 (2011) 1096–1103. [18] K. Hagihara, A. Kinoshita, Y. Sugino, M. Yamasaki, Y. Kawamura, H.Y. Yasuda, Y. Umakoshi, Intermetallics 18 (2010) 1079–1085. [19] G. Shao, V. Varsani, Z. Fan, Comput. Coupling Ph. Diagr. Thermochem. 30 (3) (2006) 286–295. [20] J. Grobner, A. Kozlov, X.Y. Fang, J. Geng, J.F. Nie, R. Schmid-Fetzer, Acta Mater. 60 (2012) 5948–5962. [21] K. Hagihara, Y. Sugino, Y. Fukusumi, Y. Umakoshi, T. Nakano, Mater. Trans. 52 (6) (2011) 1096–1103. [22] K. Hagihara, N. Yokotani, Y. Umakoshi, Intermetallics 18 (2010) 267–276. [23] X.H. Shao, Z.Q. Yang, X.L. Ma, Acta Mater. 58 (2010) 4760–4771. [24] J.F. Nie, Y.M. Zhu, A.J. Morton, Metall. Mater. Trans. A 45 (8) (2014) 3338–3348.
Contents lists available at ScienceDirect
Materials Science & Engineering A journal homepage: www.elsevier.com/locate/msea
Enhancement of properties in cast Mg–Y–Zn rod processed by severe plastic deformation Rimma Lapovok a,n, Xiang Gao a, Jian-Feng Nie a, Yuri Estrin a, Suveen N. Mathaudhu b a b
Materials Engineering Department, Monash University, Clayton, Vic 3800, Australia Materials Science Division, U.S. Army Research Office, Research Triangle Park, NC 27709, USA
art ic l e i nf o
a b s t r a c t
Article history: Received 10 June 2014 Received in revised form 21 July 2014 Accepted 21 July 2014 Available online 30 July 2014
Ternary Mg–Y–Zn alloys have attracted considerable attention from researchers due to their excellent mechanical properties and unique microstructures, particularly from the presence of long-period stacking-order (LPSO) phases. Microstructural variations and the resulting mechanical properties can be affected by various processing routes, particularly those involving severe plastic deformation of a cast billet. The approach used in this work was based on subjecting cast Mg92Y4Zn4 (composition in wt%) billet to severe plastic deformation by three different routes, namely equal channel angular pressing (ECAP), high pressure torsion (HPT) and ECAP followed by HPT, with the aim of refining the microstructure and improving mechanical properties. Samples processed by ECAP were annealed by post-processing and tested in compression and tension. The effect of the processing route and the process parameters on the microstructure and the hardness of the Mg–Y–Zn alloy is reported. An overall positive effect of annealing treatment on the mechanical properties of ECAP-processed alloy is demonstrated. & 2014 Elsevier B.V. All rights reserved.
Keywords: Mg–Y–Zn alloys LPSO phases Severe plastic deformation Equal channel angular pressing High pressure torsion
1. Introduction Addition of rare-earth (RE) elements is known to increase the strength and to improve the creep properties of magnesium alloys at high temperatures due to the formation of intermetallic phases [1–4]. Most RE-containing magnesium alloys show an excellent precipitation hardening response [1–4]; however, Mg–Y–Zn ternary alloys are not precipitation hardenable [5]. Superior mechanical properties in these alloys are achieved by other techniques, for example, warm extrusion of rapidly solidified powder (strength of up to 600 MPa and 5% tensile ductility) [6,7], or working of cast billet [8–10], which results in strengthening by grain refinement. The microstructure after such processing usually contains a mixture of nano-scale magnesium grains (100–200 nm) and intermetallic particles that exhibit a range of long-period stackingorder (LPSO) structures, among which 18R and 14H are the most common [11]. The notation used to designate the LPSO intermetallic phases is defined in Ref. [11]. A unique feature of such LPSO structures is their lamellar morphology that comprises alternating layers of magnesium and intermetallic lamellae. The accumulated experimental evidence indicates that mechanical properties of the Mg–Y–Zn alloys can be tailored by proper control of the volume n
Corresponding author. Tel.: +61 399059294; fax: +61 399054940. E-mail address: [email protected] (R. Lapovok).
http://dx.doi.org/10.1016/j.msea.2014.07.068 0921-5093/& 2014 Elsevier B.V. All rights reserved.
fraction and size of the lamellae of the LPSO structures. It is expected that an ultrahigh strength alloy may be obtained if a nano-scale distribution of LPSO structures is achieved. It is not possible to achieve a redistribution of LPSO structures by deformation with conventional metal forming processes, as they have a geometrical limitation on the amount of strain imparted to the material. However, by severe plastic deformation (SPD) methods [12] a billet can be subjected to virtually unlimited strain without a change in its shape and dimensions. The most well-known SPD methods are equal channel angular pressing (ECAP) [13] and high pressure torsion (HPT) [14]. The effects of SPD on the mechanical properties of Mg–Y–Zn alloys have not been explored to date. It can be expected, however, that severe plastic deformation may lead to much finer grains and potential dispersion of the LPSO structures, and therefore much higher yield strengths. A comparative study of microstructure and properties of Mg92Y4Zn4 cast billets severely deformed by different SPD techniques has been performed in this work.
2. Materials and experimental methods Mg92Y4Zn4 (composition in wt%) billet was prepared by permanent mould casting followed by homogenisation for 24 h at 450 1C in a muffle furnace. After homogenisation the samples were
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
quenched in water to 70 1C. Cast billets were subjected to severe plastic deformation by three different routes, namely ECAP, HPT and ECAP followed by HPT. Samples 10 mm in diameter and 37 mm in length were machined from the homogenised billet. Four passes of ECAP (route BC in which the specimen is rotated by 901 about the pressing direction between the passes) were performed using a 901 die at 350, 450 and 500 1C under a back-pressure of 185 MPa (Fig. 1a). ECAP with back-pressure was performed using a specially designed unit equipped with a heated die to permit isothermal processing within the temperature range of 20–500 1C. Backpressure was controlled through a horizontal hydraulic cylinder to maintain a pre-set pressure level, while the unit was placed in a
199
1000 kN press to exercise the forward pressure. Each pass of ECAP introduced a strain of ε ¼1.15 in the material. In a separate set of experiments, disk shaped samples of 1 mm thickness and diameter of 8 mm were cut from as-cast and ECAP processed samples and subjected to 10 revolutions in an HPT die (Fig. 1b) under a pressure of 1.9 GPa at room temperature. Each revolution introduced a strain around ε ¼ 14 near the rim of the specimen. The schedule of experiments is shown in Table 1. Annealing after ECAP was carried out in a range of temperatures from 200 to 500 1C and for different times from 10 to 30 min. After all deformation routes the microstructure was examined using an Olympus GX51 optical microscope (OM) and a Philips
Fig. 1. Severe plastic deformation processes used in this research: (a) ECAP with back-pressure and (b) HPT. Table 1 Experimental schedules. ECAP temperature, 1C ECAP HPT ECAP þHPT ECAP þAnn
350 – 350 –
HPT temperature, 1C 450 – 450 450
500 – – 500
Fig. 2. Initial microstructure of cast billet before deformation: (a) as-cast and (b) homogenised.
– 20 20 –
200
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
A
<111>
Precipitates 01-1 110 101
Mg-matrix
<001>
B
-200 020
0002 α
20 nm
50 nm
-1101 α -4221 14 H
0002 14 H
14H
Mg-matrix
Fig. 3. TEM images of microstructure after homogenisation: (a) precipitates and (b) LPSO particles (BF images and SAED patterns).
Fig. 4. Microstructure of deformed billet after ECAP at 350 1C: (a) one pass and (b) four passes.
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
201
Fig. 5. TEM bright field (BF) images of deformed billet after 4 passes of ECAP at 350 1C: (a) magnesium matrix; (b) BF image and SAED pattern recorded from a 14H particle; (c) BF image and MBED pattern recorded from a Mg3Y2Zn3 (fcc: a¼ 0.6848 nm) particle.
Fig. 6. OM of samples deformed by room temperature HPT with initial conditions: (a) as-cast (arrows indicate shear bands) and (b) ECAP processed at 350 1C.
202
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
Fig. 7. TEM images of samples deformed by 4 passes ECAP at 350 1C followed by room temperature HPT: (a) magnesium matrix; (b) and (c) BF images and SAED pattern recorded from a 14H particle; (d) BF images and MBED patterns recorded from Mg3Y2Zn3 (fcc: a¼ 0.6848 nm) particles.
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
CM20 transmission electron microscope (TEM) operated at 200 kV. Vickers hardness (ASTM E384-10e2) was measured using a Struers Duramin A-300 hardness tester with a load of 0.1 kg (HV0.1). Seven indents were produced per sample. For XRD analysis, a Philips X-ray diffractometer equipped with a Cu anode operated at 40 kV and 25 mA was used.
3. Results and discussion 3.1. Microstructure The microstructure of the as-cast alloy contains a large volume fraction of relatively fine-scale lamella (Fig. 2a) which are reported to have high concentrations of yttrium and zinc [5]. This microstructure is quite different from that normally observed in as-cast Mg-alloys which typically consist of magnesium grains and secondary phases in grain boundaries. This lamellar structure represents a mixture of 18R and α-Mg phases [5,11,15–18]. The eutectic formed between the lamellar structures, seen as the dark regions in Fig. 2a, is expected to be comprised of the intermetallic W-phase and the α-Mg phase [19,20]. After homogenisation the lamellar platelets are still the dominating feature in the microstructure, but then they have become much coarser. The eutectic is still observed
203
in regions isolated by the lamellar platelets (Fig. 2b). The 18R phase formed in the as-cast microstructure generally transforms to 14H after prolonged heat treatments at high temperatures [5,11,24]; however, the microstructure of the studied alloy after homogenisation contained α-Mg grains (Fig. 3a) and both 18R and 14H LPSO particles (Fig. 3b). The diffraction pattern obtained suggested that it has a BCC crystal structure. During severe plastic deformation by ECAP at 350 1C, the strain is distributed non-uniformly between the skeleton LPSO phase and magnesium matrix, with the strain mainly supported by the LPSO phase. After the first ECAP pass, a strain of about 1.15 is introduced into the magnesium matrix, and this level of strain appears to be large enough to deform the LPSO structure, which becomes wavy, as can be seen in Fig. 4a. As the number of passes increased to four, a significant amount of strain (ε ¼4.6) was introduced into both the matrix and LPSO phase (Fig. 4b). This gives rise to extreme grain refinement (grain size around 200–300 nm) in magnesium matrix, which can be seen in TEM images (Fig.5a). The microstructure produced by four ECAP passes also contains 14H particles and coarse Mg3Y2Zn3 (face-centred cubic: a ¼0.6848 nm) particles (Fig. 5b and c). A closer inspection reveals that there exist several domains or kink bands [21–24] within the 14H particles. Each 14H domain or kink band is rotated at a small angle with respect to the surrounding domains, as
Fig. 8. TEM images of samples deformed by 4 passes ECAP at 450 1C: (a) magnesium matrix; (b and c) BF images of 14H particles; (d) SAED pattern recorded from a 14H particle.
204
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
Fig. 9. TEM images of samples deformed by 4 passes ECAP at 450 1C followed by room temperature HPT: (a) magnesium matrix; (b and c) BF images of 14H particles; (d) SAED pattern recorded from a 14H particle.
revealed by a small rotation of lattice fringes in the bright field (BF) image and some splitting of diffraction spots in the selected area electron diffraction (SAED) pattern, both shown by arrows in Fig. 5b. The micro beam electron diffraction (MBED) patterns recorded from a Mg3Y2Zn3 particle are shown in Fig 5c. More strain can be introduced if samples are subjected to additional HPT deformation. During HPT enormously high strain is produced and the adiabatic temperature increase is sufficient for dynamic recrystallisation (DRX) to occur. The microstructures of HPT samples with as-cast or after ECAP initial conditions are shown for comparison in Fig. 6. It can be seen that if HPT is applied to the as-cast sample, the plastic flow is unstable and shear bands are formed (Fig. 6a). Samples pre-strained by ECAP have LPSO platelets with dynamically recrystallised small grains in the regions between these platelets (Fig. 6b). The TEM study of samples processed by four ECAP passes at 350 1C followed by room temperature HPT also reveals fine α-Mg grains, 14H particles and coarse Mg3Y2Zn3 particles (Fig. 7a, b, d). As expected, far more severe deformation in the 14H phases is observed in this case than after one ECAP pass alone. It appears that there are kinks formed within 14H particles (Fig. 7c). The rotation angles of individual 14H domains with respect to their surrounding domains appear to be larger, as evident from the tilt of the lattice fringes in the
BF image and the split of the diffraction spots in the SAED pattern (Fig. 7b). The angle between domains D1 and D2 (as well as D3 and D4) is approximately 11 while the angle between domains D1 and D3 (as well as D2 and D4) is approximately 51. Samples deformed by four ECAP passes at 450 1C contained mainly fine α-Mg grains ( 1 μm) and 14H particles (Fig. 8). An inspection of the lattice fringe image of the 14H phase shows that there are some defects within the 14H particles (see arrows) (Fig. 8b). In addition, there appear several domains within an individual 14H particle. Similarly, the microstructure of samples after four passes ECAP at 450 1C followed by HPT also contains mainly fine α-Mg grains and 14H particles (Fig. 9). The microstructure of samples deformed by four ECAP passes at 500 1C also comprised of α-Mg grains and 14H particles, Fig. 10. A significantly smaller number of defects and few, if any, domains are detectable within a 14H particle. This may be attributed to the higher processing temperature. 3.2. Mechanical properties Mechanical properties were evaluated by microhardness measurements and, in the case of ECAP processed material, by uniaxial tension and compression testing. (Unfortunately, due to the small
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
205
Fig. 10. TEM images of samples deformed by 4 passes ECAP at 500 1C: (a) magnesium matrix; (b and c) BF image and SAED pattern recorded from a 14H particle.
Table 2 Vickers hardness of Mg92Y4Zn4 with different processing histories. Sample
LPSO Phase
Mg-grains
Hardness, HV (0.1) Homogenised ECAP 1 pass ECAP 4 pass HPT ECAP þ HPT
907 4 1277 7 1367 6 1537 5 1617 4
108 75 120 75 144 77 109 75
size of HPT and ECAP þHPT processed samples, no tensile or compression tests were possible.) Measurements on samples with the different processing histories studied showed an increase in microhardness by 46 HV due to severe plastic deformation in the LPSO phase after four passes of ECAP and a further increment of 25 HV as a result of subsequent HPT deformation (Table 2). The microhardness of the magnesium grains increased less significantly (18–30 HV), possibly due to dynamic recrystallisation or recovery within the magnesium matrix caused by heat release during HPT deformation (Fig. 7b). The unexpectedly large microhardness values in magnesium
grains in samples deformed by HPT only may be attributed to difficulties in isolating magnesium grains from the LPSO phase within the narrow bands of each. We note that the LPSO phases are the majority components of the microstructure which controls the load bearing capability of the alloy, hence the strengthening of these phases together with grain refinement of alpha-Mg matrix and dispersion of LPSO phase by the SPD processing significantly contribute to the overall strength of the material. It can thus be concluded from Table 2 that with the larger number of ECAP passes, greater strengthening is achieved. An additional HPT step raises the strength of the alloy even further. Comparing HPT with ECAP taken as two individual processes, we notice that HPT imparts a larger strengthening to the material than ECAP does. The tensile and compression tests conducted on ECAP processed samples revealed the following picture. While all ECAP processed samples exhibited strength well in excess of that of the homogenised as-cast material, their ductility was reduced. For example, after four ECAP passes at 500 1C, the tensile strength was 337 MPa and tensile ductility was 3.4%, while compression strength and ductility were 509 MPa and 3.7%, respectively. To improve ductility, several annealing treatments were performed on the samples that were subjected to four ECAP passes at 450 and
206
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
Fig. 11. True stress vs. true strain (compression test) for samples subjected to 4 ECAP passes at 450 1C (left) and 500 1C (right) followed by annealing treatment.
Fig. 12. True stress vs. true strain (tensile test) for samples subjected to 4 ECAP passes at 450 1C (left) and 500 1C (right) followed by annealing treatment at 400 1C.
500 1C. The true stress vs. true strain curves obtained in the compression tests are shown in Fig. 11. It can be seen that both a short high temperature annealing (500 1C/10 min) and a longer annealing at a lower temperature (300 1C/30 min) improve the mechanical properties of the SPD processed samples. To further investigate the influence of the annealing time on the mechanical properties, tensile samples were annealed at 400 1C for different times (Fig. 12). We note that after annealing at 400 1C the maximum strength in compression, around 550 MPa, was much higher than the ultimate tensile strength in tension (350–400 MPa). The ductility in compression was also very high: around 25–35%, as compared to that in tension (12–18%). It can be seen that SPD processing followed by short annealing leads to levels of ductility close to the ductility of homogenised cast material in compression ( 30%) and a multiple of that in tension (12–18% vs. 3%). For both types of loading, the strength of the SPD processed alloy with a short post-processing annealing was much higher (by 130– 250 MPa) than that of the homogenised cast alloy. A remarkable observation is that the sign of the tension vs. compression anisotropy, as given by the difference between the yield strength in tension and that in compression, was reversed by ECAP followed by annealing. While the magnitude of this asymmetry did not change much in terms of the absolute value, the relative
quantity (i.e., the magnitude of this difference divided by the higher of the two yield stress values) dropped quite appreciably, from about 0.6 to about 0.4.
4. Conclusions Cast Mg92Y4Zn4 (composition in wt%) billet was subjected to severe plastic deformation by three different routes, namely ECAP, HPT and ECAP followed by HPT. The LPSO phases were shown to deform by bending and rearrangement within a soft magnesium matrix, where a substantial grain refinement down to 200– 300 nm takes place. Further increases in plastic strain and the associated temperature rise appear to initiate dynamic recrystallisation and/or dynamic recovery of the magnesium matrix leading to a degree of softening therein. However, the overall microhardness of the material remains high, as most of the applied load is carried by the harder intermetallic LPSO particles. It can be concluded that high hardness and – by induction – high strength of Mg92Y4Zn4 rod can be achieved by severe plastic deformation processing. This conclusion was confirmed by uniaxial tension and compression tests of the material that underwent different levels of ECAP processing. It was also found that the increase of strength and hardness by SPD processing was achieved at the cost of
R. Lapovok et al. / Materials Science & Engineering A 615 (2014) 198–207
ductility. However, short-time annealing was demonstrated to increase ductility by about 30% in compression and 12–18% in tension (as compared to the as-homogenised alloy), while maintaining the tensile strength at a very high level of 550 MPa. The overall conclusion that can be made on the basis of the results presented is that the processing of the as-cast alloy Mg92Y4Zn4 by four passes of Route Bc equal channel angular pressing at 500 1C, followed by annealing for 12 min at 400 1C produces the most favourable combination of mechanical properties in the alloy. In view of a relatively small incremental change of microhardness (and presumably strength) of the material due to an extra HPT step subsequent to ECAP, we do not recommend the inclusion of this step in the processing route. Acknowledgement This project was supported by the U.S. Army International Technology Center – Pacific and the US Army Research Laboratory (FA5209-11-T-0043), under programme managers Dr. Christopher Drew and Dr. Vincent Hammond, respectively. References [1] C. Antion, P. Donnadieu, F. Perrard, A. Deschamps, C. Tassin, A. Pisch, Acta Mater. 51 (18) (2003) 5335–5348. [2] J.F. Nie, B.C. Muddle, Acta Mater. 48 (4) (2000) 1691–1703. [3] X. Gao, S.M. He, X.Q. Zeng, L.M. Peng, W.J. Ding, J.F. Nie, Mater. Sci. Eng. A 431 (1–2) (2006) 322–327. [4] P.J. Apps, H. Karimzadeh, J.F. King, G.W. Lorimer, Scr. Mater. 48 (8) (2003) 1023–1028.
207
[5] Y.M. Zhu, A.J. Morton, J.F. Nie, Mater. Sci. Forum 561–565 (Part 1) (2007) 151–154. [6] T. Inoue, Y. Kawamura, M. Matsushita, K. Hayashi, J. Koike, J. Mater. Res. 16 (7) (2001) 1894–1900. [7] Y. Kawamura, K. Hayashi, A. Inoue, T. Masumoto, Mater. Trans. 42 (7) (2001) 1172–1176. [8] X. Zheng, J. Dong, D. Yin, W. Liu, F. Wang, L. Jin, W. Ding, Mater. Sci. Eng. A 527 (16–17) (2010) 3690–3694. [9] K.U. Kainer, J. Wendt, K. Hantzsche, J. Bohlen, S. Yi, D. Letzig, Mater. Sci. Forum 654–656 (2010) 580–585. [10] J. Bohlen, S. Yi, D. Letzig, K.U. Kainer, Mater. Sci. Eng. A 527 (26) (2010) 7092–7098. [11] Y.M. Zhu, A.J. Morton, M. Weyland, J.F. Nie, Acta Mater. 58 (2) (2010) 464–475. [12] R.Z. Valiev, Y. Estrin, Z. Horita, T.G. Langdon, M.J. Zehetbauer, Y.T. Zhu, J. Met. 58 (4) (2006) 33–39. [13] R. Lapovok, The role of back-pressure in equal channel angular extrusion, J. Mater. Sci. 40 (2) (2005) 341–346. [14] J. Cizek, I. Prochazka, B. Smola, I. Stulikova, R. Kuzel, Z. Matej, V. Cherkaska, R. K. Islamgaliev, O. Kulyasova, Mater. Sci. Forum 503–504 (2006) 149–154. [15] K. Milicka, P. Pérez, F. Dobeš, G. Garcés, P. Adeva, Mater. Sci. Eng. A 510–511 (2009) 377–381. [16] D.H. Ping, K. Hono, Y. Kawamura, A. Inoue, Philos. Mag. Lett. 82 (2002) 543–551. [17] K. Hagihara, Y. Sugino, Y. Fukusumi, Y. Umakoshi, T. Nakano, Mater. Trans. 52 (2011) 1096–1103. [18] K. Hagihara, A. Kinoshita, Y. Sugino, M. Yamasaki, Y. Kawamura, H.Y. Yasuda, Y. Umakoshi, Intermetallics 18 (2010) 1079–1085. [19] G. Shao, V. Varsani, Z. Fan, Comput. Coupling Ph. Diagr. Thermochem. 30 (3) (2006) 286–295. [20] J. Grobner, A. Kozlov, X.Y. Fang, J. Geng, J.F. Nie, R. Schmid-Fetzer, Acta Mater. 60 (2012) 5948–5962. [21] K. Hagihara, Y. Sugino, Y. Fukusumi, Y. Umakoshi, T. Nakano, Mater. Trans. 52 (6) (2011) 1096–1103. [22] K. Hagihara, N. Yokotani, Y. Umakoshi, Intermetallics 18 (2010) 267–276. [23] X.H. Shao, Z.Q. Yang, X.L. Ma, Acta Mater. 58 (2010) 4760–4771. [24] J.F. Nie, Y.M. Zhu, A.J. Morton, Metall. Mater. Trans. A 45 (8) (2014) 3338–3348.