Mechanical Alloying of Mg-Al Alloy with Addition of Metal Silicides

Materials Transactions, Vol. 45, No. 7 (2004) pp. 2410 to 2416 #2004 The Japan Institute of Metals Mechanical Alloying of Mg-Al Alloy with Addition of Metal Silicides Akihiro Yamazaki*, Junichi Kaneko and Makoto Sugamata Department of Mechanical Engineering, College of Industrial Technology, Nihon University, Narashino 275-8575, Japan Mg-Al alloy powder was mechanically alloyed (MA) by addition of metal silicides (TiSi 2, CrSi 2 and WSi 2 ), using a planetary ball mill under an Ar atmosphere. The MA powders were consolidated via vacuum hot pressing and hot extrusion. Solid state reactions in the MA powders and extruded materials were examined by XRD and TEM. Mechanical properties were evaluated by hardness and compression tests. In all alloy systems, in situ formation of Mg 2 Si and metal silicides with lower Si content than the added silicides occurred by solid state reaction during MA and subsequent heating processes. All of the as-extruded materials with fine dispersoids and fine matrix grain structures exhibited high hardness and specific proof strength. The as-extruded material of the Mg-Al-TiSi 2 system showed the highest hardness of 141 HV, 0.2% compressive proof stress of 657 MPa and specific proof stress of 337 MPa/(Mg/m 3 ) at room temperature. (Received December 24, 2003; Accepted May 28, 2004) Keywords: powder metallurgy, mechanical alloying, solid state reaction, magnesium-aluminum alloy, metal silicide 1. Introduction The metal silicides, with their high melting points and narrow compositional ranges, are difficult to produce via conventional melting and casting. However, it can be produced by mechanical alloying (MA) because formation of metal silicides by solid state reaction can be promoted. It has been reported that metal silicides such as NbSi 2, 1) MoSi 2, 1,2) WSi 2, 3) Mg 2 Si, 4) etc. can be synthesized by the MA process. In the MA process, mixing can be achieved at the nano level, even for a system in which alloying is difficult to perform by melting and casting. It has been reported that when metal oxides thermodynamically less stable than MgO are added to pure Mg or Mg- Al alloy, the added oxides may decompose and MgO is formed during MA and subsequent heating processes. 5 8) Magnesium compounds can be dispersed in the Mg matrix at the nano-size level by applying such oxygen displacement reactions. In addition, the presence of solute Al promotes the decomposition of oxides in the situation where reduced metals can form aluminide compounds. Metals such as Ti, Cr, W, etc., which differ markedly from magnesium in density and melting point, form no binary compound under equilibrium state with Mg. 9) For this reason, alloying performed by melting and casting leads to significant segregation due to gravity, making it difficult to obtain materials with uniform structures. As shown in the standard free energy-temperature diagrams of metal silicides (Fig. 1 10) ), TiSi 2, CrSi 2, and WSi 2 are thermodynamically less stable than Mg 2 Si. Thus, these silicides should decompose and Mg 2 Si should be formed in the matrix of the Mg solid solution. It has been shown that when TiSi 2, CrSi 2, and WSi 2 are added to magnesium, MA and subsequent heat treatment were able to produce Mgbased materials with a dispersion of nano-sized Mg 2 Si particles. 11) In this paper, various metal silicides were added to a Mg- Al alloy. Solid state reactions during MA and subsequent heat treatment were studied and mechanical properties of the P/M materials were evaluated. *Research Fellow, Nihon University Fig. 1 The standard free energies of formation of silicides as a function of temperature. Table 1 Chemical composition of AZ91 alloy used in this work (mass%). Alloy Al Zn Mn Si Cu Ni Fe Ca Mg AZ91 9.2 0.7 0.27 0.02 0.001 0.000 0.005 0.002 bal. 2. Experimental Procedures 2.1 Test materials The Mg-Al alloy was sourced from machined chips of AZ91 Mg alloy. The chemical composition of the AZ91 alloy used in the experiment is shown in Table 1. Although this alloy contains 0.7 mass% of Zn, Zn was not involved in any solid state reactions in the present work. Powders of TiSi 2, CrSi 2, and WSi 2 (purity 99% or more; particle size 2 5 mm) were added by an amount of 5 at%si. The starting compositions of the test materials are shown in Table 2. As the MA process control agent, stearic acid was added by an amount of 3.85 mass% with respect to the total powder charge. 2.2 MA process and consolidation MA was carried out using a planetary ball mill (Fritsch; P- 5) with a stainless steel container of 500 ml and balls of 10 mm in diameter. The mixed powder charge was 25 g, and the ball to powder weight ratio was set to about 16:1. The

Mechanical Alloying of Mg-Al Alloy with Addition of Metal Silicides 2411 Table 2 Nominal compositions of the tested materials. Matrix Additive Material (mass%) (mol%) (mass%) 7.5%TiSi 2 Mg-8.0%Al-10%TiSi 2 Mg-9.2%Al-0.7%Zn 7.5%CrSi 2 Mg-8.0%Al-11%CrSi 2 7.5%WSi 2 Mg-7.3%Al-18%WSi 2 mixed powder was sealed in the container together with the balls and stearic acid under an argon atmosphere. By rotating the container at 300 min 1 (autorotation: 651 min 1 ), MA was conducted for 108 ks. The resulting MA powder was taken out of the container inside a globe box under Ar atmosphere. Subsequently, MA powder was sealed in an AZ31 alloy container, and hot-pressed (HP) at 673 K for 3.6 ks at a pressure of 100 MPa under a vacuum of 1:3 10 3 Pa. The HP materials were then hot extruded at a reduction ratio of 25:1 at 623 K. Extruded rods with a diameter of 7 mm were produced. 2.3 Evaluation of test materials For the MA powder and the extruded material, changes of constituent phases during MA process and subsequent heating were evaluated using a Shimadzu XD-610 X-ray diffractometer (XRD). XRD was performed at a scanning rate of 16:7 10 3 deg/s using Cu K radiation at 40 kv and 60 mv. HP and extruded materials were prepared with polished surfaces for XRD. Furthermore, crystal size of the MA powders, was estimated by applying Hall s method 12) using eq. (1). For each of the diffraction lines, the relationship between ( cos =) and (sin =) was obtained, and crystal size was estimated by the least square method. The half line width ( sample ) was obtained by eq. (2), i.e. the measured half line width ( exp ) was compensated by the half line width ( inst )of annealed AZ91 alloy. cos sample ¼ sin þ 1 ð1þ " 2 sample ¼ 2 exp 2 inst ð2þ where " represents crystalline size, is the wavelength of incident X-ray, is half line width, is the effective nonuniform strain and is the Bragg angle. The microstructure of the extruded material was examined using a JEOL JEM-2010 transmission electron microscope (TEM). TEM specimens were prepared by electrolytic polishing using an ethanol 5 vol% nitric acid solution at about 245 K. Following polishing, ion milling (Gatan 600) was performed for 43.2 ks at an irradiation angle of 15, gun voltage of 2 kv and gun current of 0.5 ma. The Vickers hardness and compressive yield strength (0.2% proof stress) were measured. To determine the hardness, loads of 98 mn and 9.8 N were applied to the MA powder and extruded material, respectively. Average hardness values were obtained from 10 and 20 measurements on the MA powders and extruded materials, respectively. The hardness of the as-extruded material was measured after heating for 7.2 ks at each temperature in air. Specimens of 5 mm in diameter and 5 to 7.5 mm in height were prepared for compression testing of the as-extruded materials. Compression tests were performed at an initial strain rate of 0.001 s 1 by using three specimens for each material. The density of the as-extruded materials was obtained using the Archimedes method. 3. Experimental Results and Discussion 3.1 Solid state reaction and microstructures Solid state reaction during MA and subsequent heating processes was studied for the MA powders and extruded materials by XRD. The XRD results for the Mg-Al-TiSi 2 system are shown in Fig. 2. Weak diffraction lines from Mg 2 Si were detected for the MA powder. By heat treating the extruded material at 773 K, the intensity of the Mg 2 Si diffraction lines increased, while those from TiSi 2 remained unchanged. This reveals that solid state reaction occurred as the result of MA process and subsequent heating, and a displacement reaction involving Si took place leading to formation of Mg 2 Si. However, the presence of Ti due to decomposition of TiSi 2 could not be confirmed by XRD. TEM microstructures of the as-extruded material of Mg- Al-TiSi 2 system are shown in Fig. 3. A bright-field image is shown in Fig. 3(a), and a selected-area electron diffraction image obtained from this visual field is shown in Fig. 3(d). A dark-field image of Mg(101) obtained from the spot marked as (B) in Fig. 3(d) is shown in Fig. 3(b), which shows Mg-Al alloy grains of 100 200 nm in size. In contrast, the crystal size of the Mg-Al alloy matrix estimated from XRD diffraction lines for the MA powder was 32 nm. This suggests that grain growth occurred during consolidation. However, Fig. 2 X-ray diffraction patterns of MA powders and extruded materials of Mg-Al-TiSi 2 system; (a) AZ91, (b) TiSi 2, (c) MA powder for 108 ks, (d) as-extruded material and (e) extruded material heated at 773 K for 7.2 ks.

2412 A. Yamazaki, J. Kaneko and M. Sugamata Fig. 3 Transmission electron micrographs of as-extruded material of Mg- Al-TiSi 2 system; (a) bright field image, (b) dark field image of Mg (101) spot, (c) dark field image of Mg 2 Si (220) spot and (d) selected area diffraction pattern from the observed area. the grain boundaries of the as-extruded material were not straight but wavy as shown in Fig. 3(b). Thus, grain growth was suppressed due to pinning by the dispersed particles. A dark-field image obtained from the spot of Mg 2 Si(220) marked as (C) in Fig. 3(d) is shown in Fig. 3(c). Thus, it was confirmed that the dispersed Mg 2 Si particles are as fine as approximately 50 nm in size. Figure 4(b) shows an EDS spectrum obtained from a dispersed particle marked in a TEM micrograph of the asextruded material of the Mg-Al-TiSi 2 system shown in Fig. 4(a). It was shown that the dispersed particle contains both Ti and Si. Quantitative analysis was performed. The atomic ratio of Si:Ti was determined to be 50:24 for the coarse dispersoid in the Mg-Al alloy matrix not shown in Fig. 4(a). This ratio is approximately equal to 2:1. The result of quantitative analysis of a smaller particle shown by an arrow in Fig. 4(a) under the same condition showed that the atomic composition ratio of Si:Ti was 11:8 as shown in Fig. 4(b). It is apparent that the Ti content is higher than that of TiSi 2.This suggests that some Si was separated from TiSi 2 in the Mg-Al alloy matrix, leading to formation of Mg 2 Si and silicide particles with an increased Ti content. This explains why Ti aluminide phase was not observed by XRD although Mg 2 Si was formed after decomposition of TiSi 2. Thus, solute Al did not influence the decomposition of TiSi 2. Figure 5 shows the results of XRD for the MA powder and as-extruded material of the Mg-Al-CrSi 2 system. MA after 108 ks led to the formation of Mg 2 Si, revealing that CrSi 2 Fig. 4 Transmission electron micrograph of as-extruded material of Mg- Al-TiSi 2 system; (a) bright field image and (b) EDS spectrum of the precipitate of the bright field image. was partly decomposed. After heat treating of the as-extruded material at 773 K for 7.2 ks, the diffraction lines for CrSi 2 disappeared completely, while those of Cr 3 Si became detectable. An aluminide compound of Cr could not be confirmed. Observed diffraction lines of AlFe indicate occurrence of Fe contamination from the milling steel balls and vial. Figure 6 shows TEM microstructures of the as-extruded material of the Mg-Al-CrSi 2 system. Fig. 6(d) is a selectedarea electron diffraction pattern obtained from a bright field image shown in Fig. 6(a). A dark-field image from the Mg 2 Si(220) diffraction spots marked with (B) in Fig. 6(d) is shown in Fig. 6(b), and that from the spot of Mg(101) marked with (C) is shown in Fig. 6(c). This suggests that the Mg 2 Si formed by solid state reaction is dispersed as particles with sizes less than 30 nm in the Mg-Al alloy matrix. The matrix was observed to have nano-crystal grains of 100 200 nm in size. The matrix grain size of the MA powder was estimated to be 28 nm by Hall s method, which is smaller than that of the as-extruded material. Grain growth during hot-extrusion occurred only to a limited extent due to grain boundary pinning by the fine dispersoids. Figure 7 summarizes the results of XRD on the Mg-Al- WSi 2 system. Even after heating of the MA powder,

Mechanical Alloying of Mg-Al Alloy with Addition of Metal Silicides 2413 Fig. 5 X-ray diffraction patterns of MA powders and extruded materials of Mg-Al-CrSi 2 system; (a) AZ91, (b) CrSi 2, (c) MA powder for 108 ks, (d) as-extruded material and (e) extruded material heated at 773 K for 7.2 ks. Fig. 7 X-ray diffraction patterns of MA powders and extruded materials of Mg-Al-WSi 2 system; (a) AZ91, (b) WSi 2, (c) MA powder for 108 ks, (d) as-extruded material and (e) extruded material heated at 773 K for 7.2 ks. Fig. 8 Transmission electron micrograph of as-extruded material of Mg-Al-WSi 2 system. Fig. 6 Transmission electron micrographs of as-extruded material of Mg- Al-CrSi 2 system; (a) bright field image, (b) dark field image of Mg 2 Si (220) spots, (c) dark field image of Mg (101) spot and (d) selected area diffraction pattern from the observed area. decomposition of WSi 2 was not confirmed. The diffraction lines from Al 12 Mg 17 disappeared in the MA powder, revealing that solutionizing of Al into Mg occurred during the MA process. Figure 8 shows TEM microstructure of the as-extruded material for the Mg-Al-WSi 2 system, in which large dispersed particles over 100 nm as well as fine dispersoids of about 10 nm were found. The crystal size of the Mg-Al alloy matrix in the as-extruded material was about 200 nm, whereas that of the MA powder was 23 nm, estimated using Hall s method. Although grain growth took place during hot extrusion, nano-order crystalline size was still maintained due to the pinning effect of the fine dispersed particles as in the case of the other systems.

2414 A. Yamazaki, J. Kaneko and M. Sugamata Fig. 9 Transmission electron micrographs of as-extruded material of Mg- Al-WSi 2 system; (a) bright field image, (b) dark field image of Mg 2 Si (220) spots, (c) selected area diffraction pattern from the observed area and (d) EDS spectrum of the precipitate of the bright field image. Figure 9 shows TEM microstructures of the as-extruded material of the Mg-Al-WSi 2 system. Fig. 9(c) shows a selected-area electron diffraction pattern obtained from a bright-field image shown in Fig. 9(a). From a dark-field image (Fig. 9(b)) obtained from the Mg 2 Si(220) diffraction spots marked with an arrow in Fig. 9(c), Mg 2 Si particles of less than 50 nm in size were formed, although this could not be confirmed by XRD. Thus, partial decomposition of WSi 2 and formation of Mg 2 Si occurred. An EDS spectrum of a dispersed particle of about 100 nm marked with an arrow in Fig. 9(a) is shown in Fig. 9(d), revealing the presence of a compound containing W and Si. When quantitative analysis was performed on the EDS spectrum, the atomic ratio of Si to W was estimated to be 50:45. This suggests that a silicide with a higher W content was formed by the decomposition of WSi 2 in the Mg-Al alloy matrix in this system, in a similar manner to the cases of the other systems. This explains why a W-aluminide compound was not detected in this system. These results reveal that metal silicides with a lower Si content such as Mg 2 Si and Cr 3 Si are generated when metal silicides (TiSi 2, CrSi 2, and WSi 2 ) are mechanically alloyed with Mg-Al and subsequently heat treated. The standard free energy of the silicide per mol of Si is summarized in Fig. 10. 10) Fig. 10(a), Fig. 10(b) and Fig. 10(c) show the relationships between the standard free energy and temperature for Ti silicide, Cr silicide and W silicide, respectively. In these figures, the standard free energy for Mg 2 Si is also shown. From these figures, it is evident that Ti 5 Si 3,Cr 5 Si 3 and W 5 Si 3 with lower Si content ratio are thermodynamically more stable than the added silicides of Fig. 10 The standard free energies of formation of silicides and Mg 2 Si as a function of temperature, (a) Ti silicides, (b) Cr silicides, (c) W silicides. TiSi 2, CrSi 2 and WSi 2. Above all, Ti 5 Si 3 and Cr 5 Si 3 are thermo-dynamically more stable than Mg 2 Si. Cr 3 Si detected by XRD in the Mg-Al-CrSi 2 system is thermodynamically more stable than Cr 5 Si 3 with its higher Si content ratio. Table 3 10) summarizes the chemical reactions when the added metal silicides are decomposed in Mg. Table 3 shows the change in standard free energy, G at 298 K in the relevant reaction and the G values converted to per mol of Mg 2 Si. It is suggested that a solid state reaction take place due to the

Mechanical Alloying of Mg-Al Alloy with Addition of Metal Silicides 2415 Table 3 The driving force of displacement reactions of Si in Mg. Reaction G at 298 K kj kjmol 1 Mg 2 Si 14Mg + 5TiSi 2! 7Mg 2 Si + Ti 5 Si 3 401 57:3 14Mg + 5WSi 2! 7Mg 2 Si + W 5 Si 3 357 51:1 14Mg + 5CrSi 2! 7Mg 2 Si + Cr 5 Si 3 400 57:2 10Mg + 3CrSi 2! 5Mg 2 Si + Cr 3 Si 373 74:6 driving force to form both Mg 2 Si and silicide of lower Si content such as Cr 3 Si. In the Mg-Al-CrSi 2 system, Cr 3 Si, which is thermodynamically the most stable, was formed and further reaction did not occur in Mg since Cr 3 Si is more stable than Mg 2 Si. Therefore, Cr or Ti was not detected by XRD as a constituent phase. In the Mg-Al-WSi 2 system, W 5 Si 3 is thermo-dynamically more stable than WSi 2 but less stable than Mg 2 Si, and hence, WSi 2 was partially decomposed and Mg 2 Si and W 5 Si 3 were formed during heating at 773 K. From the values of G per 1 mol of Mg 2 Si in Table 3, the driving force of the decomposition is highest for CrSi 2, and lowest for WSi 2, which explains the XRD results. It is also suggested that there was no involvement of the solute Al in the solid state reaction, and thus formation of aluminide compound (except FeAl) was not detected. 3.2 Mechanical properties The hardness of the MA powders of the Mg-Al-TiSi 2, Mg- Al-CrSi 2 and Mg-Al-WSi 2 were 218, 196 and 206 HV, respectively. These high hardness values may be attributable to the fine dispersion and to nano-crystal grain structure of the matrix. Figure 11 shows the hardness values of the extruded Mg- Al-TiSi 2, Mg-Al-CrSi 2, and Mg-Al-WSi 2 systems after heating for 7.2 ks at various temperatures. The hardness values of the as-extruded materials were 141 HV, 134 HV and 129 HV for the Mg-Al-TiSi 2, Mg-Al-CrSi 2 and Mg-Al- WSi 2 systems, respectively. The hardness of the as-extruded materials was appreciably lower than that of the MA powders. It appears that the materials were softened by hot extrusion. The hardness of all the as-extruded materials remained unchanged after heat treating up to 573 K. No softening was observed for the extruded materials of Mg-Al- TiSi 2 and Mg-Al-CrSi 2 after heating at 673 K. The hardness of all as-extruded materials decreased after heating at 773 K. The more pronounced softening observed after heating of Mg-Al-WSi 2 at 673 K and 773 K may be caused by a decrease in amounts of the in situ formed silicides with lower Si content such as W 5 Si 3 and Mg 2 Si due to the higher stability of WSi 2 in the Mg-Al alloy matrix. The compressive proof stress by compression test of the extruded materials is shown in Fig. 12, in which values observed for the AZ91 extrude material via ingot metallurgy (I/M) are also included. The density, hardness, compressive proof stress and specific proof strength of the as-extruded materials are summarized in Table 4. In the compressive load-displacement curves, the plastic deformation was clearly recognized. It is clear that these P/M materials show remarkably high specific proof strength due to fine dispersoids and nano-scale grain size of the Mg-Al matrix. 4. Conclusions MA was performed with addition of transition metal silicides (TiSi 2, CrSi 2 and WSi 2 ) to a Mg-Al alloy. The MA powders were consolidated to the P/M materials via vacuum hot pressing and hot extrusion. Solid state reactions were studied using XRD and TEM during MA and after sub- Fig. 12 Compressive proof strength at room temperature for as-extruded materials of Mg-Al-metal silicide systems and of AZ91 I/M material. Table 4 The results of hardness and compression tests of as-extruded materials of Mg-Al-metal silicide systems. Fig. 11 Changes in hardness of extruded materials of Mg-Al-metal silicide systems on isochronal heating for 7.2 ks at various temperatures. System Mg-Al-TiSi 2 Mg-Al-CrSi 2 Mg-Al-WSi 2 AZ91 I/M Hardness (HV) 141 134 129 58.3 Density (Mgm 3 ) 1.95 2.03 2.40 1.83 0.2% Compressive Proof Stress 657 590 535 274 (MPa) Specific Proof Strength 337 291 223 150 (MPaMg 1 m 3 )

2416 A. Yamazaki, J. Kaneko and M. Sugamata sequent heat treatment. The mechanical properties were examined by hardness and compression tests. The obtained results are summarized as follows: (1) During MA and subsequent heat treatment of Mg-Almetal silicide systems, Mg 2 Si and silicides of lower Si content ratio were formed. Formation of aluminide compounds of transition metals was not observed. (2) The MA powders showed nano-size matrix grain structures in the order of 10 nm. In the as-extruded material, fine grains of 100 200 nm were formed in the Mg-Al matrix with dispersion of Mg 2 Si particles smaller than 50 nm. (3) Hardness values of the as-extruded materials for the Mg-Al-TiSi 2, Mg-Al-CrSi 2 and Mg-Al-WSi 2 systems were 141, 134 and 129 HV, respectively. In the Mg-Al- TiSi 2 and Mg-Al-CrSi 2 systems, hardness values were unchanged after heating at 673 K for 7.2 ks. (4) In the as-extruded material of the Mg-Al-TiSi 2 system, the 0.2% compressive proof stress and specific proof stress were 657 MPa and 337 MPa/(Mg/m 3 ), respectively. In addition, the material showed some ductility. Acknowledgements The authors would like to thank Masahiro Kubota, Yuichi Asano and Fumihiko Takano for their assistance with the experimental work. REFERENCES 1) L. Liu, F. Padella, W. Guo and M. Magini: Acta Metall. Mater. 43 (1995) 3755 3761. 2) R. B. Schwarz, S. R. Srinivasan, J. J. Petrovic and C. J. Maggiore: Mater. Sci. Eng. A155 (1992) 75 83. 3) E. Ivanov: Proceedings of the 2nd International Conference on Structural Applications of Mechanical Alloying, (ASM International, 1993) 415 419. 4) L. Lu, M. O. Lai and M. L. Hoe: Nanostructured Materials 10 (1998) 551 563. 5) A. Yamazaki, J. Kaneko and M. Sugamata: J. Japan Soc. Powder and Powder Metall. 48 (2001) 61 66. 6) A. Yamazaki, J. Kaneko and M. Sugamata: J. Japan Soc. Powder and Powder Metall. 48 (2001) 397 403. 7) A. Yamazaki, J. Kaneko and M. Sugamata: J. Japan Soc. Powder and Powder Metall. 48 (2001) 935 942. 8) A. Yamazaki, J. Kaneko and M. Sugamata: J. JILM 52 (2002) 421 425. 9) Phase Diagrams of Binary Magnesium Alloys, ed. A. A. Nayeb- Hashemi and J. B. Clark, (ASM International, 1988). 10) I. Barin: Thermochemical Data of Pure Substances, (VCH, New York, 1989). 11) A. Yamazaki, J. Kaneko and M. Sugamata: J. Japan Soc. Powder and Powder Metall. 48 (2001) 404 411. 12) W. H. Hall: J. Inst. Met. 75 (1948 1949) 1127.