- Email: [email protected]
Quantitative fractographic analysis of variability in the tensile ductility of high-pressure die-cast AE44 Mg-alloy
Materials Science and Engineering A 427 (2006) 255–262
Quantitative fractographic analysis of variability in the tensile ductility of high-pressure die-cast AE44 Mg-alloy S.G. Lee a , G.R. Patel a , A.M. Gokhale a,∗ , A. Sreeranganathan a , M.F. Horstemeyer b a
School of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0245, USA b Department of Mechanical Engineering, Mississippi State University, Mississippi State, MS 39762, USA Received 4 August 2005; received in revised form 14 April 2006; accepted 24 April 2006
Abstract Cast magnesium alloys often exhibit large variability in fracture related mechanical properties such as ductility and strength. In this contribution, the variability in the tensile ductility of individually cast tensile test specimens of high-pressure die-cast AE44 Mg-alloy is examined at room temperature and at 394 K. Significant specimen-to-specimen variations in the ductility are observed at both temperatures. The variability in the ductility does not quantitatively correlate to the average volume fraction of porosity (or any other microstructural parameters) in the bulk threedimensional microstructure. The area fraction of porosity measured in the fracture surfaces of the tensile test specimens is much larger than the average volume fraction of the porosity in the corresponding bulk microstructure. Therefore, the fracture path preferentially goes through the regions of highly localized clusters of gas and shrinkage pores. Interestingly, at both test temperatures, the percent tensile ductility e shows a quantitative correlation with the area fraction of the porosity f in the corresponding fracture surfaces, which can be represented by the following simple equation e = e0 [1 − f]m , where e0 and m are empirical constants. © 2006 Elsevier B.V. All rights reserved. Keywords: AE44; High-pressure die-cast magnesium alloys; Quantitative fractography
1. Introduction There has been an increasing thrust lately on the development of lightweight cast magnesium alloy components for structural automotive and other applications. Majority of the Mg-alloy castings are made from Mg–Al base (AM series) and Mg–Al–Zn base (AZ series) alloys. AM and AZ series alloys exhibit superior die-castability and a good combination of strength and ductility [1]. These cast Mg-alloys are suitable for structural automotive components such as steering wheels, instruments panels, seat frames, and doorframes that do not experience elevated temperatures in service [2,3]. As AM and AZ series alloys do not have adequate high temperature strength and creep resistance above 400 K [4,5], they are not suitable for automotive components such as gearbox housing, oil pump housing, oil pans, and intake manifolds that operate at elevated temperatures. Consequently, cast Mg-alloys that have improved creep resistance and bolt load retention properties are required for such appli-
∗
Corresponding author. Tel.: +1 404 894 2887; fax: +1 404 894 9140. E-mail address: [email protected] (A.M. Gokhale).
0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2006.04.108
cations. Addition of rare earth elements (RE) such as Ce and La are known to improve creep resistance and corrosion resistance of Mg–Al base alloys [6]. As mischmetal (MM)1 is an economical way of adding such beneficial alloying elements, AE series of Mg–Al–RE base casting alloys such as AE42 were developed in the 1990s [7]. Recently, Hydro Magnesium developed a new high-pressure die-casting alloy, AE44, which has attractive high temperature mechanical properties, as well as diecastability, and corrosion resistance [8]. The alloy contains 4% Al and 4% MM. AE44 also has good fracture sensitive mechanical properties such as ductility and strength. High-pressure die-casting (HPDC) is the preferred manufacturing process for cast Mg-alloy components. Consequently, HPDC AE44 alloy is being considered for structural components such as automotive front engine cradle [9]. However, the HPDC Mg-alloys contain considerable amount of micro-porosity [10–12]. Microporosity and other casting defects appear to adversely affect the mechanical properties of the HPDC Mg-alloys and may lead
1 A typical MM is a standard grade cerium-based alloy (>50% cerium) with lanthanum (20–35%), neodymium (10–20%), and proesodymium (4–10%).
256
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
to significant variability in their fracture sensitive mechanical properties such ductility, fatigue life, and toughness. The variability in the mechanical properties is an important concern for the HPDC Mg-alloy components used for structural applications. Therefore, a systematic study of the correlations between the variability in the microstructure and the corresponding variability in the relevant fracture related mechanical properties of HPDC AE44 Mg-alloy is of significant interest. Although amount of micro-porosity appears to adversely affect the tensile ductility of HPDC Mg-alloys, attempts to establish quantitative correlations between the average amount of porosity in the bulk microstructure and fracture sensitive properties such as strength or ductility have often failed [13–15]. On the other hand, the amount of porosity observed in the fracture surfaces of tensile test specimens shows strong quantitative correlation with the fracture sensitive mechanical properties of HPDC AM50, AM60, and AZ91 Mg-alloys [14–16], which is similar to that observed in numerous cast Al-alloys [17–20]. Consequently, in this contribution, quantitative fractographic techniques have been used to explore the correlations between the variability in the tensile ductility of AE44 HPDC Mg-alloy and the geometric attributes of the micro-porosity present in the corresponding tensile fracture surfaces. For this purpose, the tensile tests have been performed on HPDC AE44 alloy specimens at room temperature and at 394 K. The next section of the paper gives description of the experimental work. The results of microscopy and quantitative fractography studies are presented in the subsequent sections. 2. Experimental 2.1. Materials and mechanical tests The experiments were performed on AE44 alloy, which is an Mg–Al–RE base alloy; its nominal composition is given in Table 1. Round tensile test specimens of 6 mm diameter were high-pressure die-cast using a metallic mold. HPDC Mgalloys have fine-grained microstructure (so called, “skin”) at the surface layers that is different from the interior bulk microstructure, and consequently, the constitutive behavior of the skin microstructure differs from that of the interior microstructure [21]. The HPDC components containing the skin have better mechanical properties. As a result, in the industrial practice, the skin is not removed from the high-pressure die-cast Mg-alloy components. Accordingly, in the present study, the skin was not Table 1 Nominal chemical composition of AE44 alloy Element
Composition (wt.%)
Al Mn Zn Si Cu RE Mg
3.5–4.4 0.1 (min) 0.22 (max) 0.1 (max) 0.01 (max) 4 Balance
machined off from the gage section of the specimens, but the grip sections were machined to ensure proper dimensions and alignment during the tensile tests. All tensile tests were performed at the engineering strain rate of 8.3 × 10−3 percent per second. The tests were performed at room temperature and at 394 K. The tensile test data at 394 K are important in the context of structural automotive applications of the HPDC AE44 alloy components, because the local temperature in some regions of an automobile can reach up to 394 K. At both temperatures, the tests were performed on at least 10 specimens. To ensure uniform specimen temperature, each specimen was soaked at the test temperature for 30 min before conducting the tensile test. The soaking treatment did not lead to any observable changes in the microstructure. The tests were performed as per ASTM E21-92 (1998) standard. 2.2. Metallography The metallography was performed on the samples drawn from the gage sections of the tensile test specimens. The metallographic planes containing the loading direction were utilized for metallography. These samples were mounted in standard metallographic mounts, and then they were ground using 320–1000 grit abrasive SiC papers. Immediately after the last grinding step, the specimens were washed with water, rinsed with methanol, and dried. The fine polishing was conducted using 6 and 1 m oil based diamond compounds. The final steps in polishing involved the use of 0.3 m alumina and 0.05 m colloidal silica. Glycerol and oil base lubricants were used during these polishing steps to avoid the formation of a surface film. The specimens were immediately rinsed with methanol and dried. The specimens were examined in the as-polished condition as well as in the etched condition. As-polished metallographic sections were used to measure the volume fraction and other attributes of porosity in the bulk microstructure using standard stereological techniques [22]. Digital image analysis was used to perform these measurements. The specimens were subsequently etched at room temperature using acetic glycol etching agent (20 ml acetic acid, 1 ml nitric acid, 60 ml ethylene glycol, and 19 ml water) to reveal the grain structure, eutectic constituents, and intermetallic compounds at the grain boundaries. 2.3. Quantitative fractography Fracture surfaces of the tensile test specimens were examined using a Hitachi S-4100 scanning electron microscope equipped with an energy dispersive X-ray analyzer. The fracture surfaces contain elongated shrinkage porosity as well as equiaxed gas (air) porosity. The gas porosity is due to the entrapped air in the mold cavity as well as due to hydrogen and other gases dissolved in the liquid alloy. The gas content in the pores cannot be determined from metallographic sections. Figs. 1 and 2 depict typical SEM fractographs of tensile fracture surfaces. In these fracture surfaces, the pre-existing shrinkage porosity can be detected due to presence of unfractured dendrites below the pores. Thus, the regions of the fracture surface that contain intact dendrites are essentially the areas occupied by porosity. Some
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
257
contain fine defects such as oxides. However, these defects are present in very small amounts compared to the porosity. 3. Results and discussion 3.1. Mechanical properties
Fig. 1. SEM fractograph of the fracture surface of a tensile test specimen depicting shrinkage porosity.
Fig. 3a shows typical true-stress (σ) versus true-strain (ε) curves of the high-pressure die-cast AE44 alloy at room temperature and at 394 K. Fig. 3b depicts the ln(σ) versus ln(ε) curves for the plastic part of the flow curves. The non-linearity of the plots in Fig. 3b reveals that in the HPDC AE44 alloy, the flow curve does not follow a simple power law equation often assumed in numerous theoretical models. It is well known that both slip and twinning contribute to plastic strain in Mg-alloys. Further, prism and pyramidal secondary slip systems become active at moderately elevated temperatures and/or at higher stresses. The relative contributions of the slip on primary basal slip system, secondary slip systems, and the twinning to the total global plastic strain varies with plastic strain and temperature, and therefore, work hardening rate varies with strain in a complex manner. Consequently, the flow behavior cannot be described by a simple power law type equation. As will be seen later, this has a bearing on
such regions are outlined in Fig. 1. On the other hand, some large gas pores (Fig. 2) can be easily identified due to their morphology. In the present work, the total area fraction of porosity (both gas and shrinkage pores) was measured in each tensile fracture surface using well-known quantitative fractographic techniques [14–16,23]. Entire fracture surfaces were scanned for these measurements. In addition to porosity, the fracture surfaces also
Fig. 2. SEM fractograph of the fracture surface of a tensile test specimen depicting gas (air) porosity.
Fig. 3. (a) True stress–strain curves of the alloy at room and at 394 K; (b) ln(σ) vs. ln(ε) curves for the plastic part of the flow curves.
258
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
Table 2 Tensile properties of AE44 specimens alloy at room temperature and at 394 K Sample ID
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Room temperature
394 K
YS (MPa)
UTS (MPa)
%El
YS (MPa)
UTS (MPa)
%El
135.1 128.3 131.0 136.5 135.1 138.6 131.0 130.3 143.4 135.1 137.2 135.8 132.4 135.1 135.1
237.9 233.7 242.7 252.4 242.7 253.7 244.8 251.7 242.7 239.3 243.4 228.9 255.8 237.9 248.9
7.1 6.1 8.2 13.1 10.2 11.1 7.1 10.1 7.1 8.2 5.1 6.1 12.1 9.2 9.1
111.7 108.9 122.0 111.7 111.0 111.0 103.4 110.3 109.6 122.7 108.9 112.4 109.6 106.2 123.4
161.4 162.7 149.6 162.0 158.6 160.0 157.2 159.3 171.0 149.6 165.5 161.4 157.2 171.0 146.9
38.8 34.7 41.9 30.6 24.2 35.4 37.4 24.2 30.3 41.9 32.3 37.8 30.3 22.5 39.8
the interpretation of the correlation between tensile ductility and fractographic attributes. Table 2 gives the data on the variability in the ductility, UTS, and yield stress of the tensile test specimens. Note that these specimens have the same average chemical composition and they were high-pressure die-cast using the same process parameters, and yet there is a significant specimen-to-specimen variability
in the tensile ductility. At room temperature, the tensile ductility varied from 5.1% to 13.1%, whereas at 394 K, it varied from 22.5% to 41.9%. There is some variability in the UTS as well, but the yield stress is about the same for the specimens tested at the same temperature. The decrease in the variability of ductility at higher temperature is likely to be due to the higher plasticity of the alloy at the higher temperature.
Fig. 4. (a) Digitally compressed seamless montage of 70 contiguous microstructural fields covering the complete thickness (y-axis) of the lowest tensile ductility sample. (b) Enlarged one field of view of the small window in (a).
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
259
3.2. Microstructural observations Fig. 4 depicts unetched microstructure of the tensile test specimens having the lowest tensile ductility at room temperature. The microstructure pertains to a metallographic plane containing the applied stress and it is in the gage length region of the specimen. Shrinkage and gas porosity is observed in the microstructure. The gas porosity is primarily due to the trapped air in the die cavity during the high-pressure die-casting process and dissolved gases in the liquid alloy, whereas shrinkage porosity is due to lack of liquid metal feeding the volume changes due to solidification. The shrinkage pores have crack-like morphologies, whereas the gas pores have equiaxed morphologies and they are larger in size. Table 3 gives the volume % of porosity in some of the tensile test specimens and their tensile ductility data. There appears to be no quantitative correlation between the average amount of porosity in the bulk microstructure and the tensile ductility in this HPDC AE44 alloy. Recently, Weiler and coworkers [24] used X-ray computed tomography to reconstruct the three-dimensional pore microstructure in 3 mm thick specimens of HPDC AM60B Mg-alloy, and have attempted to correlate amount of porosity with the tensile fracture strain. However, there are serious deficiencies in their experimental technique: (1) due to low resolution (pores smaller than 80 m could not be detected) they were not able to detect the shrinkage porosity at all because the thickness of crack-like shrinkage pores is on the order of few microns, and (2) gas pores smaller than 80 m also could not be detected. Consequently, Weiler and coworkers have seriously underestimated the total amount of porosity, as well as amount of gas porosity and average size of gas pores. It must be emphasized that crack-like shrinkage pores as well as spatially clustered small pores exist in significant amounts in the HPDC Mg-alloys, and they are also expected to deleteriously affect the fracture sensitive properties such as tensile fracture strain. Recently, Lee and coworkers [25] have reconstructed high-resolution (∼1 m) three-dimensional porosity distributions in HPDC AM50 alloy using montage serial sectioning technique [12] that clearly show the presence of significant amount of crack-like shrinkage pores, gas pores smaller than 80 m, and shrinkage pores topologically linked to the gas pores. As these features were not included in Weiler and coworkers [10] porosity data, their correlations between the porosity attributes and tensile fracture strain, and the associated conclusions need to be reexamined. Fig. 5 shows a typical etched microstructure of present alloy at a location near a free surface of a tensile test specimen. Observe that there is a “skin” region near the surface of the Table 3 Tensile ductility and bulk porosity in some of room temperature tensile test specimens Sample ID
Tensile ductility (%)
Total porosity (%)
11 10 6 4
5.1 8.2 11.1 13.1
0.6 0.6 0.9 0.4
Fig. 5. Etched optical microstructure showing a skin region near the surface of the sample.
specimen that has significantly finer grain structure as compared to that in the interior regions. In an earlier study, it has been shown that such differences in the microstructures of the skin and the interior regions in HPDC Mg-alloys can lead to differences in the local constitutive behavior and micro-hardness [21]. The fine-grain structure of the skin regions is beneficial for the room temperature mechanical properties, but may be deleterious to the creep resistance. Fig. 6 depicts high magnification views of the etched microstructure in the interior regions. The microstructure contains cored fine ␣-Mg-rich dendrites that are surrounded grain boundary region containing eutectic phases consisting of Al–Re intermetallic compounds (see Fig. 6b). In the present HPDC alloy both Al11 RE3 and Al2 RE type intermetallic compounds are present, which have also been reported in the earlier studies in AE42 alloy [26]. In the present HPDC alloy, the intermetallic compounds mostly contain Ce and/or La rare earth elements. 3.3. Relationship between variability in the tensile ductility and fractographic parameters Figs. 1 and 2 depict typical SEM fractographs of tensile fracture surfaces. In these fracture surfaces, the pre-existing porosity can be detected due to presence of unfractured dendrites below the pores. Thus, the regions of the fracture surface that contain intact dendrites are essentially the areas occupied by either shrinkage pore or gas pores. Some such regions are outlined in Fig. 1. On the other hand, some large gas pores (Fig. 2) can be easily identified due to their morphology. Fig. 7 shows a plot of ductility versus amount of porosity in the corresponding fracture surface for the specimens tested at room temperature, which demonstrates a quantitative correlation between these quantities. Similar correlation is also observed at 394 K. Fig. 8 shows the variation of tensile ductility (as represented by % plastic elongation to failure), e, versus the area fraction of porosity f observed in the fracture surface of the corresponding specimen, plotted in the format of ln(e) versus {−ln(1 − f)}. Interestingly, at both test temperatures, the data show quantitative correlation
260
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
Fig. 7. Tensile ductility vs. amount of porosity in the fracture surface for the specimens tested at room and 394 K temperatures.
Fig. 8. Variation of tensile ductility at room and 394 K temperatures vs. area fraction of porosity in the fracture surface plotted as [ln(e)] vs. [−ln(1 − f)].
Fig. 6. (a) Etched optical microstructure in the interior region showing α dendrites and eutectic phases. (b) SEM microstructure showing surrounded grain boundary region containing eutectic phases consisting of Al–Re intermetallic compounds.
between the tensile ductility e and the area fraction of porosity f in the corresponding fracture surface. The trends in Fig. 8 can be represented by the following simple equation: e = e0 [1 − f ]m
(1)
In Eq. (1), e0 and exponent m are empirical constants. Table 4 gives the values of e0 and m at the two temperatures. Interestingly, e0 increases significantly with temperature. It is tempting to interpret e0 as the ductility of a pore-free alloy having the same remaining microstructure, but it may not be true. This is because at very low porosity levels, some other competing damage accumulation mechanism may become dominant (for example, cracking of Al2 RE and/or Al11 RE3 eutectic constituents and
void growth around cracked particles), and consequently, Eq. (1) may not be applicable. Table 4 also reveals that the exponent m decreases significantly as the tensile test temperature is increased from room temperature to 394 K. This is probably due to higher plasticity of the alloys at 394 K due to activation of the secondary slip systems. Quantitative correlation between tensile ductility and amount of porosity in the fractures surfaces similar to Eq. (1) has been observed earlier in cast Al-alloys [17–20] as well as in high-pressure die-cast AM50 and AZ91 Mg-alloys [14–16]. Caceres and Selling [27] have given an equation similar to Eq. (1) that expresses quantitative correlation between the tensile ductility and amount of porosity in the fracture surfaces observed in some Al-alloys (which is based on earlier contributions of Bourcier et al. [28] and Marciniak and Kuczinski Table 4 Values of parameters e0 and m at room temperature and at 394 K Alloy
Temperatures (K)
e0
m
AE44
Room 394
15.8 45.9
34.7 14.2
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
[29]). However, Caceres and Selling’s model [27] assumes that the tensile flow curve of the alloy can be represented by a simple power law equation, which is certainly not the case for the present HPDC AE44 alloy as documented in Fig. 3b. Therefore, more research is needed to establish the theoretical basis for the correlation represented by Eq. (1) in the HPDC Mg-alloys where the tensile flow curves cannot be represented by simple power law equation. The present data reveal that the amount of porosity present in the fracture surfaces of the tensile test specimens is significantly higher than the average volume fraction of porosity in the bulk three-dimensional microstructure, and this is true for the specimens tested at both temperatures. Therefore, it is reasonable to conclude that the fracture path preferentially goes through the regions of the specimens containing large amount of localized (clustered) shrinkage and gas porosity. The quantitative correlation between the tensile ductility and the area fraction of the porosity in the fracture surfaces strongly supports this hypothesis. Therefore, the ductility of high-pressure die-cast AE44 alloy can be increased by decreasing the regions of localized clustered pores in the microstructure, which may not necessarily require decreasing the global average volume fraction of the pores in the three-dimensional microstructure. These observations also show that as the localized clustered pores constitute the extrema in the microstructure, and the fracture is governed by these extrema in the microstructural geometry rather than the average attributes such as average porosity volume fraction, average pore size. On the other hand, most of the current models and simulations of fracture in materials containing pores [30–33] utilize global average microstructural parameters (average volume fraction of pores, average pore size, etc.), and therefore, do not capture the variability in the fracture sensitive properties such as ductility in a set of specimens having the same average microstructure. Thus, there is a need to develop a new class of damage evolution and fracture models that incorporate the spatial clustering of shrinkage and gas pores in order to predict the variability in the fracture related properties of the HPDC Mg-alloys.
4. Summary and conclusions High-pressure die-cast AE44 Mg-alloy exhibits variability in the tensile ductility at room temperature as well as at 394 K. The extent of variability in the ductility is lower at 394 K. The variability in the tensile ductility does not quantitatively correlate to the average volume fraction of porosity in the bulk threedimensional microstructure. There is a quantitative correlation between the tensile ductility and the area fraction of the porosity present in the corresponding fracture surfaces, which can be represented by a simple power law type equation. The fracture path preferentially goes through the regions containing localized and clustered pores, and therefore, in these cast microstructures, the fracture related properties such as ductility are more sensitive such local extrema in the microstructure rather than global averages of microstructural attributes such as porosity volume fraction.
261
Acknowledgements The authors thank Richard Osborne and Don Penrod for discussions, encouragement, and support, and Westmoreland Mechanical Testing and Research Laboratories for conducting the tensile tests. The authors thank Magnesium Competence Centre, Hydro Aluminum Research Centre, Porsgrunn, Norway for high-pressure die-casting of the AE44 alloy tensile test specimens. The research was supported by research grants from Structural Cast Magnesium Development (SCMD) program of USCAR Project, the U.S. National Science Foundation (Grant no. DMR-0404668), and American Foundry Society. The financial support is gratefully acknowledged. Any opinions, findings, and conclusions or recommendations expressed in the paper are those of the authors and do not necessarily reflect the views of the funding agencies. References [1] J.F. King, in: K.U. Kainer (Ed.), Magnesium Alloys and Their Applications, Wiley, New York, 2000, p. 15. [2] B.L. Mordike, T. Elbert, Mater. Sci. Eng. A 302 (2001) 37–45. [3] B.R. Powell, L.J. Ouiment, J.A. Hynes, R.S. Beals, L. Kopka, P.P. Reid, Proceedings of the Symposium Magnesium Technology, TMS, Warrendale, 2004, pp. 3–10. [4] K.S. Nair, M.C. Mittal, Mater. Sci. For. 30 (1998) 89. [5] G. Pettersen, H. Westengen, R. Hoier, O. Lohne, Mater. Sci. Eng. A 207 (1996) 115. [6] G.S. Foerster, Proceedings of the Eighth SDCE International Casting Exposition and Congress, Paper no. G-T75-112, 1975. [7] T. Aune, H. Westengen, T. Ruden, SAE Technical Paper No. 94077, 1994. [8] P. Bakke, H. Westengen, in: N.R. Neelamegham, H.I. Kaplan, B.R. Powell (Eds.), Proceedings of the Magnesium Technology, TMS, Warrendale, 2005, pp. 291–296. [9] N. Li, R. Osborne, B. Cox, D. Penrod, in: N.R. Neelamegham, H.I. Kaplan, B.R. Powell (Eds.), Proceedings of the Magnesium Technology, TMS, Warrendale, 2005, p. 535. [10] A. Balasundaram, A.M. Gokhale, Mater. Characteriz. 46 (2001) 419–426. [11] A.K. Dahl, S. Sannes, D.H. St. John, H. Westengen, J. Light Met. 1 (2001) 99–103. [12] S.G. Lee, G.R. Patel, A.M. Gokhale, M. Evans, in: N.R. Neelamegham, H.I. Kaplan, B.R. Powell (Eds.), Proceedings of the Magnesium Technology, TMS, Warrendale, 2005, pp. 371–376. [13] A.L. Bowls, J.R. Grifiths, C.J. Davidson, in: J. Hryn (Ed.), Proceedings of the Magnesium Technology, TMS, Warrendale, PA, 2001, pp. 161–168. [14] A.M. Gokhale, G.R. Patel, in: J. Hryn (Ed.), Proceedings of the Symposium Magnesium Technology, TMS, Warrendale, 2001, pp. 195–199. [15] A.M. Gokhale, G.R. Patel, in: M. Skillinberg, S. Das (Eds.), Aluminum 2002: Proceedings of the TMS Symposium on Automotive Alloys, TMS, Warrendale, 2002, pp. 65–73. [16] S.G. Lee, G.R. Patel, A.M. Gokhale, A. Sreerangathan, M.F. Horstemeyer, Scripta Mater. 53 (2005) 851–856. [17] M.K. Surappa, E.W. Blank, J.C. Jaquet, Scripta Metall. 20 (1986) 1281–1286. [18] A.M. Gokhale, G.R. Patel, Mater. Sci. Eng. 392 (2005) 184–190. [19] A.M. Gokhale, G.R. Patel, Mater. Character. 54 (2005) 13–20. [20] A.M. Gokhale, G.R. Patel, Scripta Mater. 52 (2004) 237–241. [21] Z. Shan, A.M. Gokhale, Mater. Sci. Eng. A 361 (2003) 267–274. [22] A.M. Gokhale, ASM Metals Handbook, vol. 9, Metallography and Microstructures, ASM International, Materials Park, Ohio, 2004, pp. 428–438.
262
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
[23] A.M. Gokhale, in: R.J. Shipley, W.T. Becker (Eds.), ASM Metals Handbook, vol. 11, Failure Analysis and Prevention-2002, ASM International, Materials Park, Ohio, 2004, pp. 538–556. [24] J.P. Weiler, J.T. Wood, R.J. Klassen, E. Maire, R. Berkmortel, G. Wang, Mater. Sci. Eng. A 395 (2005) 315–320. [25] S.G. Lee, A.M. Gokhale, A. Sreeranganathan, Mater. Sci. Eng., in press. [26] M.O. Perkguleyuz, A.A. Kaya, in: A. Luo (Ed.), Proceedings of the Symposium Magnesium Technology, TMS, Warrendale, 2004, pp. 281–287.
[27] C.H. Caceres, B.I. Selling, Mater. Sci. Eng. 220 (1996) 109–116. [28] R.J. Bourcier, D.A. Koss, R.E. Smelser, O. Richmond, Acta Metall. 34 (1986) 2443–2453. [29] Z. Marciniak, K. Kuczinski, Int. J. Mech. Sci. 9 (1967) 609–620. [30] A.L. Gurson, J. Eng. Mater. Mech. 99 (1977) 2–12. [31] V. Tvergaard, Int. J. Fract. 33 (1985) 1079–1087. [32] F.A. McClintock, J. Appl. Mech. 35 (1968) 363. [33] L.M. Brown, J.D. Embury, Proceedings of the Third International Conference on the Microstructure and Design of Alloys, vol. 1, Institute of Metals, London, UK, 1973, pp. 164–173.
Quantitative fractographic analysis of variability in the tensile ductility of high-pressure die-cast AE44 Mg-alloy S.G. Lee a , G.R. Patel a , A.M. Gokhale a,∗ , A. Sreeranganathan a , M.F. Horstemeyer b a
School of Materials Science and Engineering, Georgia Institute of Technology, Atlanta, GA 30332-0245, USA b Department of Mechanical Engineering, Mississippi State University, Mississippi State, MS 39762, USA Received 4 August 2005; received in revised form 14 April 2006; accepted 24 April 2006
Abstract Cast magnesium alloys often exhibit large variability in fracture related mechanical properties such as ductility and strength. In this contribution, the variability in the tensile ductility of individually cast tensile test specimens of high-pressure die-cast AE44 Mg-alloy is examined at room temperature and at 394 K. Significant specimen-to-specimen variations in the ductility are observed at both temperatures. The variability in the ductility does not quantitatively correlate to the average volume fraction of porosity (or any other microstructural parameters) in the bulk threedimensional microstructure. The area fraction of porosity measured in the fracture surfaces of the tensile test specimens is much larger than the average volume fraction of the porosity in the corresponding bulk microstructure. Therefore, the fracture path preferentially goes through the regions of highly localized clusters of gas and shrinkage pores. Interestingly, at both test temperatures, the percent tensile ductility e shows a quantitative correlation with the area fraction of the porosity f in the corresponding fracture surfaces, which can be represented by the following simple equation e = e0 [1 − f]m , where e0 and m are empirical constants. © 2006 Elsevier B.V. All rights reserved. Keywords: AE44; High-pressure die-cast magnesium alloys; Quantitative fractography
1. Introduction There has been an increasing thrust lately on the development of lightweight cast magnesium alloy components for structural automotive and other applications. Majority of the Mg-alloy castings are made from Mg–Al base (AM series) and Mg–Al–Zn base (AZ series) alloys. AM and AZ series alloys exhibit superior die-castability and a good combination of strength and ductility [1]. These cast Mg-alloys are suitable for structural automotive components such as steering wheels, instruments panels, seat frames, and doorframes that do not experience elevated temperatures in service [2,3]. As AM and AZ series alloys do not have adequate high temperature strength and creep resistance above 400 K [4,5], they are not suitable for automotive components such as gearbox housing, oil pump housing, oil pans, and intake manifolds that operate at elevated temperatures. Consequently, cast Mg-alloys that have improved creep resistance and bolt load retention properties are required for such appli-
∗
Corresponding author. Tel.: +1 404 894 2887; fax: +1 404 894 9140. E-mail address: [email protected] (A.M. Gokhale).
0921-5093/$ – see front matter © 2006 Elsevier B.V. All rights reserved. doi:10.1016/j.msea.2006.04.108
cations. Addition of rare earth elements (RE) such as Ce and La are known to improve creep resistance and corrosion resistance of Mg–Al base alloys [6]. As mischmetal (MM)1 is an economical way of adding such beneficial alloying elements, AE series of Mg–Al–RE base casting alloys such as AE42 were developed in the 1990s [7]. Recently, Hydro Magnesium developed a new high-pressure die-casting alloy, AE44, which has attractive high temperature mechanical properties, as well as diecastability, and corrosion resistance [8]. The alloy contains 4% Al and 4% MM. AE44 also has good fracture sensitive mechanical properties such as ductility and strength. High-pressure die-casting (HPDC) is the preferred manufacturing process for cast Mg-alloy components. Consequently, HPDC AE44 alloy is being considered for structural components such as automotive front engine cradle [9]. However, the HPDC Mg-alloys contain considerable amount of micro-porosity [10–12]. Microporosity and other casting defects appear to adversely affect the mechanical properties of the HPDC Mg-alloys and may lead
1 A typical MM is a standard grade cerium-based alloy (>50% cerium) with lanthanum (20–35%), neodymium (10–20%), and proesodymium (4–10%).
256
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
to significant variability in their fracture sensitive mechanical properties such ductility, fatigue life, and toughness. The variability in the mechanical properties is an important concern for the HPDC Mg-alloy components used for structural applications. Therefore, a systematic study of the correlations between the variability in the microstructure and the corresponding variability in the relevant fracture related mechanical properties of HPDC AE44 Mg-alloy is of significant interest. Although amount of micro-porosity appears to adversely affect the tensile ductility of HPDC Mg-alloys, attempts to establish quantitative correlations between the average amount of porosity in the bulk microstructure and fracture sensitive properties such as strength or ductility have often failed [13–15]. On the other hand, the amount of porosity observed in the fracture surfaces of tensile test specimens shows strong quantitative correlation with the fracture sensitive mechanical properties of HPDC AM50, AM60, and AZ91 Mg-alloys [14–16], which is similar to that observed in numerous cast Al-alloys [17–20]. Consequently, in this contribution, quantitative fractographic techniques have been used to explore the correlations between the variability in the tensile ductility of AE44 HPDC Mg-alloy and the geometric attributes of the micro-porosity present in the corresponding tensile fracture surfaces. For this purpose, the tensile tests have been performed on HPDC AE44 alloy specimens at room temperature and at 394 K. The next section of the paper gives description of the experimental work. The results of microscopy and quantitative fractography studies are presented in the subsequent sections. 2. Experimental 2.1. Materials and mechanical tests The experiments were performed on AE44 alloy, which is an Mg–Al–RE base alloy; its nominal composition is given in Table 1. Round tensile test specimens of 6 mm diameter were high-pressure die-cast using a metallic mold. HPDC Mgalloys have fine-grained microstructure (so called, “skin”) at the surface layers that is different from the interior bulk microstructure, and consequently, the constitutive behavior of the skin microstructure differs from that of the interior microstructure [21]. The HPDC components containing the skin have better mechanical properties. As a result, in the industrial practice, the skin is not removed from the high-pressure die-cast Mg-alloy components. Accordingly, in the present study, the skin was not Table 1 Nominal chemical composition of AE44 alloy Element
Composition (wt.%)
Al Mn Zn Si Cu RE Mg
3.5–4.4 0.1 (min) 0.22 (max) 0.1 (max) 0.01 (max) 4 Balance
machined off from the gage section of the specimens, but the grip sections were machined to ensure proper dimensions and alignment during the tensile tests. All tensile tests were performed at the engineering strain rate of 8.3 × 10−3 percent per second. The tests were performed at room temperature and at 394 K. The tensile test data at 394 K are important in the context of structural automotive applications of the HPDC AE44 alloy components, because the local temperature in some regions of an automobile can reach up to 394 K. At both temperatures, the tests were performed on at least 10 specimens. To ensure uniform specimen temperature, each specimen was soaked at the test temperature for 30 min before conducting the tensile test. The soaking treatment did not lead to any observable changes in the microstructure. The tests were performed as per ASTM E21-92 (1998) standard. 2.2. Metallography The metallography was performed on the samples drawn from the gage sections of the tensile test specimens. The metallographic planes containing the loading direction were utilized for metallography. These samples were mounted in standard metallographic mounts, and then they were ground using 320–1000 grit abrasive SiC papers. Immediately after the last grinding step, the specimens were washed with water, rinsed with methanol, and dried. The fine polishing was conducted using 6 and 1 m oil based diamond compounds. The final steps in polishing involved the use of 0.3 m alumina and 0.05 m colloidal silica. Glycerol and oil base lubricants were used during these polishing steps to avoid the formation of a surface film. The specimens were immediately rinsed with methanol and dried. The specimens were examined in the as-polished condition as well as in the etched condition. As-polished metallographic sections were used to measure the volume fraction and other attributes of porosity in the bulk microstructure using standard stereological techniques [22]. Digital image analysis was used to perform these measurements. The specimens were subsequently etched at room temperature using acetic glycol etching agent (20 ml acetic acid, 1 ml nitric acid, 60 ml ethylene glycol, and 19 ml water) to reveal the grain structure, eutectic constituents, and intermetallic compounds at the grain boundaries. 2.3. Quantitative fractography Fracture surfaces of the tensile test specimens were examined using a Hitachi S-4100 scanning electron microscope equipped with an energy dispersive X-ray analyzer. The fracture surfaces contain elongated shrinkage porosity as well as equiaxed gas (air) porosity. The gas porosity is due to the entrapped air in the mold cavity as well as due to hydrogen and other gases dissolved in the liquid alloy. The gas content in the pores cannot be determined from metallographic sections. Figs. 1 and 2 depict typical SEM fractographs of tensile fracture surfaces. In these fracture surfaces, the pre-existing shrinkage porosity can be detected due to presence of unfractured dendrites below the pores. Thus, the regions of the fracture surface that contain intact dendrites are essentially the areas occupied by porosity. Some
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
257
contain fine defects such as oxides. However, these defects are present in very small amounts compared to the porosity. 3. Results and discussion 3.1. Mechanical properties
Fig. 1. SEM fractograph of the fracture surface of a tensile test specimen depicting shrinkage porosity.
Fig. 3a shows typical true-stress (σ) versus true-strain (ε) curves of the high-pressure die-cast AE44 alloy at room temperature and at 394 K. Fig. 3b depicts the ln(σ) versus ln(ε) curves for the plastic part of the flow curves. The non-linearity of the plots in Fig. 3b reveals that in the HPDC AE44 alloy, the flow curve does not follow a simple power law equation often assumed in numerous theoretical models. It is well known that both slip and twinning contribute to plastic strain in Mg-alloys. Further, prism and pyramidal secondary slip systems become active at moderately elevated temperatures and/or at higher stresses. The relative contributions of the slip on primary basal slip system, secondary slip systems, and the twinning to the total global plastic strain varies with plastic strain and temperature, and therefore, work hardening rate varies with strain in a complex manner. Consequently, the flow behavior cannot be described by a simple power law type equation. As will be seen later, this has a bearing on
such regions are outlined in Fig. 1. On the other hand, some large gas pores (Fig. 2) can be easily identified due to their morphology. In the present work, the total area fraction of porosity (both gas and shrinkage pores) was measured in each tensile fracture surface using well-known quantitative fractographic techniques [14–16,23]. Entire fracture surfaces were scanned for these measurements. In addition to porosity, the fracture surfaces also
Fig. 2. SEM fractograph of the fracture surface of a tensile test specimen depicting gas (air) porosity.
Fig. 3. (a) True stress–strain curves of the alloy at room and at 394 K; (b) ln(σ) vs. ln(ε) curves for the plastic part of the flow curves.
258
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
Table 2 Tensile properties of AE44 specimens alloy at room temperature and at 394 K Sample ID
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
Room temperature
394 K
YS (MPa)
UTS (MPa)
%El
YS (MPa)
UTS (MPa)
%El
135.1 128.3 131.0 136.5 135.1 138.6 131.0 130.3 143.4 135.1 137.2 135.8 132.4 135.1 135.1
237.9 233.7 242.7 252.4 242.7 253.7 244.8 251.7 242.7 239.3 243.4 228.9 255.8 237.9 248.9
7.1 6.1 8.2 13.1 10.2 11.1 7.1 10.1 7.1 8.2 5.1 6.1 12.1 9.2 9.1
111.7 108.9 122.0 111.7 111.0 111.0 103.4 110.3 109.6 122.7 108.9 112.4 109.6 106.2 123.4
161.4 162.7 149.6 162.0 158.6 160.0 157.2 159.3 171.0 149.6 165.5 161.4 157.2 171.0 146.9
38.8 34.7 41.9 30.6 24.2 35.4 37.4 24.2 30.3 41.9 32.3 37.8 30.3 22.5 39.8
the interpretation of the correlation between tensile ductility and fractographic attributes. Table 2 gives the data on the variability in the ductility, UTS, and yield stress of the tensile test specimens. Note that these specimens have the same average chemical composition and they were high-pressure die-cast using the same process parameters, and yet there is a significant specimen-to-specimen variability
in the tensile ductility. At room temperature, the tensile ductility varied from 5.1% to 13.1%, whereas at 394 K, it varied from 22.5% to 41.9%. There is some variability in the UTS as well, but the yield stress is about the same for the specimens tested at the same temperature. The decrease in the variability of ductility at higher temperature is likely to be due to the higher plasticity of the alloy at the higher temperature.
Fig. 4. (a) Digitally compressed seamless montage of 70 contiguous microstructural fields covering the complete thickness (y-axis) of the lowest tensile ductility sample. (b) Enlarged one field of view of the small window in (a).
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
259
3.2. Microstructural observations Fig. 4 depicts unetched microstructure of the tensile test specimens having the lowest tensile ductility at room temperature. The microstructure pertains to a metallographic plane containing the applied stress and it is in the gage length region of the specimen. Shrinkage and gas porosity is observed in the microstructure. The gas porosity is primarily due to the trapped air in the die cavity during the high-pressure die-casting process and dissolved gases in the liquid alloy, whereas shrinkage porosity is due to lack of liquid metal feeding the volume changes due to solidification. The shrinkage pores have crack-like morphologies, whereas the gas pores have equiaxed morphologies and they are larger in size. Table 3 gives the volume % of porosity in some of the tensile test specimens and their tensile ductility data. There appears to be no quantitative correlation between the average amount of porosity in the bulk microstructure and the tensile ductility in this HPDC AE44 alloy. Recently, Weiler and coworkers [24] used X-ray computed tomography to reconstruct the three-dimensional pore microstructure in 3 mm thick specimens of HPDC AM60B Mg-alloy, and have attempted to correlate amount of porosity with the tensile fracture strain. However, there are serious deficiencies in their experimental technique: (1) due to low resolution (pores smaller than 80 m could not be detected) they were not able to detect the shrinkage porosity at all because the thickness of crack-like shrinkage pores is on the order of few microns, and (2) gas pores smaller than 80 m also could not be detected. Consequently, Weiler and coworkers have seriously underestimated the total amount of porosity, as well as amount of gas porosity and average size of gas pores. It must be emphasized that crack-like shrinkage pores as well as spatially clustered small pores exist in significant amounts in the HPDC Mg-alloys, and they are also expected to deleteriously affect the fracture sensitive properties such as tensile fracture strain. Recently, Lee and coworkers [25] have reconstructed high-resolution (∼1 m) three-dimensional porosity distributions in HPDC AM50 alloy using montage serial sectioning technique [12] that clearly show the presence of significant amount of crack-like shrinkage pores, gas pores smaller than 80 m, and shrinkage pores topologically linked to the gas pores. As these features were not included in Weiler and coworkers [10] porosity data, their correlations between the porosity attributes and tensile fracture strain, and the associated conclusions need to be reexamined. Fig. 5 shows a typical etched microstructure of present alloy at a location near a free surface of a tensile test specimen. Observe that there is a “skin” region near the surface of the Table 3 Tensile ductility and bulk porosity in some of room temperature tensile test specimens Sample ID
Tensile ductility (%)
Total porosity (%)
11 10 6 4
5.1 8.2 11.1 13.1
0.6 0.6 0.9 0.4
Fig. 5. Etched optical microstructure showing a skin region near the surface of the sample.
specimen that has significantly finer grain structure as compared to that in the interior regions. In an earlier study, it has been shown that such differences in the microstructures of the skin and the interior regions in HPDC Mg-alloys can lead to differences in the local constitutive behavior and micro-hardness [21]. The fine-grain structure of the skin regions is beneficial for the room temperature mechanical properties, but may be deleterious to the creep resistance. Fig. 6 depicts high magnification views of the etched microstructure in the interior regions. The microstructure contains cored fine ␣-Mg-rich dendrites that are surrounded grain boundary region containing eutectic phases consisting of Al–Re intermetallic compounds (see Fig. 6b). In the present HPDC alloy both Al11 RE3 and Al2 RE type intermetallic compounds are present, which have also been reported in the earlier studies in AE42 alloy [26]. In the present HPDC alloy, the intermetallic compounds mostly contain Ce and/or La rare earth elements. 3.3. Relationship between variability in the tensile ductility and fractographic parameters Figs. 1 and 2 depict typical SEM fractographs of tensile fracture surfaces. In these fracture surfaces, the pre-existing porosity can be detected due to presence of unfractured dendrites below the pores. Thus, the regions of the fracture surface that contain intact dendrites are essentially the areas occupied by either shrinkage pore or gas pores. Some such regions are outlined in Fig. 1. On the other hand, some large gas pores (Fig. 2) can be easily identified due to their morphology. Fig. 7 shows a plot of ductility versus amount of porosity in the corresponding fracture surface for the specimens tested at room temperature, which demonstrates a quantitative correlation between these quantities. Similar correlation is also observed at 394 K. Fig. 8 shows the variation of tensile ductility (as represented by % plastic elongation to failure), e, versus the area fraction of porosity f observed in the fracture surface of the corresponding specimen, plotted in the format of ln(e) versus {−ln(1 − f)}. Interestingly, at both test temperatures, the data show quantitative correlation
260
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
Fig. 7. Tensile ductility vs. amount of porosity in the fracture surface for the specimens tested at room and 394 K temperatures.
Fig. 8. Variation of tensile ductility at room and 394 K temperatures vs. area fraction of porosity in the fracture surface plotted as [ln(e)] vs. [−ln(1 − f)].
Fig. 6. (a) Etched optical microstructure in the interior region showing α dendrites and eutectic phases. (b) SEM microstructure showing surrounded grain boundary region containing eutectic phases consisting of Al–Re intermetallic compounds.
between the tensile ductility e and the area fraction of porosity f in the corresponding fracture surface. The trends in Fig. 8 can be represented by the following simple equation: e = e0 [1 − f ]m
(1)
In Eq. (1), e0 and exponent m are empirical constants. Table 4 gives the values of e0 and m at the two temperatures. Interestingly, e0 increases significantly with temperature. It is tempting to interpret e0 as the ductility of a pore-free alloy having the same remaining microstructure, but it may not be true. This is because at very low porosity levels, some other competing damage accumulation mechanism may become dominant (for example, cracking of Al2 RE and/or Al11 RE3 eutectic constituents and
void growth around cracked particles), and consequently, Eq. (1) may not be applicable. Table 4 also reveals that the exponent m decreases significantly as the tensile test temperature is increased from room temperature to 394 K. This is probably due to higher plasticity of the alloys at 394 K due to activation of the secondary slip systems. Quantitative correlation between tensile ductility and amount of porosity in the fractures surfaces similar to Eq. (1) has been observed earlier in cast Al-alloys [17–20] as well as in high-pressure die-cast AM50 and AZ91 Mg-alloys [14–16]. Caceres and Selling [27] have given an equation similar to Eq. (1) that expresses quantitative correlation between the tensile ductility and amount of porosity in the fracture surfaces observed in some Al-alloys (which is based on earlier contributions of Bourcier et al. [28] and Marciniak and Kuczinski Table 4 Values of parameters e0 and m at room temperature and at 394 K Alloy
Temperatures (K)
e0
m
AE44
Room 394
15.8 45.9
34.7 14.2
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
[29]). However, Caceres and Selling’s model [27] assumes that the tensile flow curve of the alloy can be represented by a simple power law equation, which is certainly not the case for the present HPDC AE44 alloy as documented in Fig. 3b. Therefore, more research is needed to establish the theoretical basis for the correlation represented by Eq. (1) in the HPDC Mg-alloys where the tensile flow curves cannot be represented by simple power law equation. The present data reveal that the amount of porosity present in the fracture surfaces of the tensile test specimens is significantly higher than the average volume fraction of porosity in the bulk three-dimensional microstructure, and this is true for the specimens tested at both temperatures. Therefore, it is reasonable to conclude that the fracture path preferentially goes through the regions of the specimens containing large amount of localized (clustered) shrinkage and gas porosity. The quantitative correlation between the tensile ductility and the area fraction of the porosity in the fracture surfaces strongly supports this hypothesis. Therefore, the ductility of high-pressure die-cast AE44 alloy can be increased by decreasing the regions of localized clustered pores in the microstructure, which may not necessarily require decreasing the global average volume fraction of the pores in the three-dimensional microstructure. These observations also show that as the localized clustered pores constitute the extrema in the microstructure, and the fracture is governed by these extrema in the microstructural geometry rather than the average attributes such as average porosity volume fraction, average pore size. On the other hand, most of the current models and simulations of fracture in materials containing pores [30–33] utilize global average microstructural parameters (average volume fraction of pores, average pore size, etc.), and therefore, do not capture the variability in the fracture sensitive properties such as ductility in a set of specimens having the same average microstructure. Thus, there is a need to develop a new class of damage evolution and fracture models that incorporate the spatial clustering of shrinkage and gas pores in order to predict the variability in the fracture related properties of the HPDC Mg-alloys.
4. Summary and conclusions High-pressure die-cast AE44 Mg-alloy exhibits variability in the tensile ductility at room temperature as well as at 394 K. The extent of variability in the ductility is lower at 394 K. The variability in the tensile ductility does not quantitatively correlate to the average volume fraction of porosity in the bulk threedimensional microstructure. There is a quantitative correlation between the tensile ductility and the area fraction of the porosity present in the corresponding fracture surfaces, which can be represented by a simple power law type equation. The fracture path preferentially goes through the regions containing localized and clustered pores, and therefore, in these cast microstructures, the fracture related properties such as ductility are more sensitive such local extrema in the microstructure rather than global averages of microstructural attributes such as porosity volume fraction.
261
Acknowledgements The authors thank Richard Osborne and Don Penrod for discussions, encouragement, and support, and Westmoreland Mechanical Testing and Research Laboratories for conducting the tensile tests. The authors thank Magnesium Competence Centre, Hydro Aluminum Research Centre, Porsgrunn, Norway for high-pressure die-casting of the AE44 alloy tensile test specimens. The research was supported by research grants from Structural Cast Magnesium Development (SCMD) program of USCAR Project, the U.S. National Science Foundation (Grant no. DMR-0404668), and American Foundry Society. The financial support is gratefully acknowledged. Any opinions, findings, and conclusions or recommendations expressed in the paper are those of the authors and do not necessarily reflect the views of the funding agencies. References [1] J.F. King, in: K.U. Kainer (Ed.), Magnesium Alloys and Their Applications, Wiley, New York, 2000, p. 15. [2] B.L. Mordike, T. Elbert, Mater. Sci. Eng. A 302 (2001) 37–45. [3] B.R. Powell, L.J. Ouiment, J.A. Hynes, R.S. Beals, L. Kopka, P.P. Reid, Proceedings of the Symposium Magnesium Technology, TMS, Warrendale, 2004, pp. 3–10. [4] K.S. Nair, M.C. Mittal, Mater. Sci. For. 30 (1998) 89. [5] G. Pettersen, H. Westengen, R. Hoier, O. Lohne, Mater. Sci. Eng. A 207 (1996) 115. [6] G.S. Foerster, Proceedings of the Eighth SDCE International Casting Exposition and Congress, Paper no. G-T75-112, 1975. [7] T. Aune, H. Westengen, T. Ruden, SAE Technical Paper No. 94077, 1994. [8] P. Bakke, H. Westengen, in: N.R. Neelamegham, H.I. Kaplan, B.R. Powell (Eds.), Proceedings of the Magnesium Technology, TMS, Warrendale, 2005, pp. 291–296. [9] N. Li, R. Osborne, B. Cox, D. Penrod, in: N.R. Neelamegham, H.I. Kaplan, B.R. Powell (Eds.), Proceedings of the Magnesium Technology, TMS, Warrendale, 2005, p. 535. [10] A. Balasundaram, A.M. Gokhale, Mater. Characteriz. 46 (2001) 419–426. [11] A.K. Dahl, S. Sannes, D.H. St. John, H. Westengen, J. Light Met. 1 (2001) 99–103. [12] S.G. Lee, G.R. Patel, A.M. Gokhale, M. Evans, in: N.R. Neelamegham, H.I. Kaplan, B.R. Powell (Eds.), Proceedings of the Magnesium Technology, TMS, Warrendale, 2005, pp. 371–376. [13] A.L. Bowls, J.R. Grifiths, C.J. Davidson, in: J. Hryn (Ed.), Proceedings of the Magnesium Technology, TMS, Warrendale, PA, 2001, pp. 161–168. [14] A.M. Gokhale, G.R. Patel, in: J. Hryn (Ed.), Proceedings of the Symposium Magnesium Technology, TMS, Warrendale, 2001, pp. 195–199. [15] A.M. Gokhale, G.R. Patel, in: M. Skillinberg, S. Das (Eds.), Aluminum 2002: Proceedings of the TMS Symposium on Automotive Alloys, TMS, Warrendale, 2002, pp. 65–73. [16] S.G. Lee, G.R. Patel, A.M. Gokhale, A. Sreerangathan, M.F. Horstemeyer, Scripta Mater. 53 (2005) 851–856. [17] M.K. Surappa, E.W. Blank, J.C. Jaquet, Scripta Metall. 20 (1986) 1281–1286. [18] A.M. Gokhale, G.R. Patel, Mater. Sci. Eng. 392 (2005) 184–190. [19] A.M. Gokhale, G.R. Patel, Mater. Character. 54 (2005) 13–20. [20] A.M. Gokhale, G.R. Patel, Scripta Mater. 52 (2004) 237–241. [21] Z. Shan, A.M. Gokhale, Mater. Sci. Eng. A 361 (2003) 267–274. [22] A.M. Gokhale, ASM Metals Handbook, vol. 9, Metallography and Microstructures, ASM International, Materials Park, Ohio, 2004, pp. 428–438.
262
S.G. Lee et al. / Materials Science and Engineering A 427 (2006) 255–262
[23] A.M. Gokhale, in: R.J. Shipley, W.T. Becker (Eds.), ASM Metals Handbook, vol. 11, Failure Analysis and Prevention-2002, ASM International, Materials Park, Ohio, 2004, pp. 538–556. [24] J.P. Weiler, J.T. Wood, R.J. Klassen, E. Maire, R. Berkmortel, G. Wang, Mater. Sci. Eng. A 395 (2005) 315–320. [25] S.G. Lee, A.M. Gokhale, A. Sreeranganathan, Mater. Sci. Eng., in press. [26] M.O. Perkguleyuz, A.A. Kaya, in: A. Luo (Ed.), Proceedings of the Symposium Magnesium Technology, TMS, Warrendale, 2004, pp. 281–287.
[27] C.H. Caceres, B.I. Selling, Mater. Sci. Eng. 220 (1996) 109–116. [28] R.J. Bourcier, D.A. Koss, R.E. Smelser, O. Richmond, Acta Metall. 34 (1986) 2443–2453. [29] Z. Marciniak, K. Kuczinski, Int. J. Mech. Sci. 9 (1967) 609–620. [30] A.L. Gurson, J. Eng. Mater. Mech. 99 (1977) 2–12. [31] V. Tvergaard, Int. J. Fract. 33 (1985) 1079–1087. [32] F.A. McClintock, J. Appl. Mech. 35 (1968) 363. [33] L.M. Brown, J.D. Embury, Proceedings of the Third International Conference on the Microstructure and Design of Alloys, vol. 1, Institute of Metals, London, UK, 1973, pp. 164–173.