Creep Strength of Magnesium-Based Alloys KOUICHI MARUYAMA, MAYUMI SUZUKI, and HIROYUKI SATO The high-temperature creep resistance of magnesium alloys was discussed, with special reference to Mg-Al and Mg-Y alloys. Mg-Al solid-solution alloys are superior to Al-Mg solid-solution alloys in terms of creep resistance. This is attributed to the high internal stress typical of an hcp structure having only two independent basal slip systems. Although magnesium has a smaller shear modulus than aluminum, the inherent creep resistance of Mg alloys is better than that of Al alloys. The creep resistance of Mg alloys is improved substantially by the addition of Y. Solid-solution hardening is the principal mechanism of the strengthening, but the details of the mechanism have not been elucidated yet. Forest dislocation hardening in concentrated alloys and dynamic precipitation in a Mg-2.4 pct Y alloy also contribute to the strengthening. An addition of a very small amount of Zn raises the dislocation density and significantly improves the creep resistance of Mg-Y alloys.
I. INTRODUCTION
MAGNESIUM has the lowest density among the elements widely used as structural materials. In addition to their high specific strength, magnesium alloys have excellent castability and better recyclability than plastics. Owing to these benefits, they are prospective candidates for applications in automotive components. Automotive engine components made of Mg alloys are used at elevated temperatures from 423 to 473 K and are required to have sufficient creep resistance. However, it is often claimed that Mg alloys are inferior to Al alloys in creep resistance.[1] The objective of this article is to evaluate the inherent creep resistance of Mg alloys in comparison with Al alloys and to discuss how to improve their creep resistance. In the first part of this article, the creep rates of Mg-Al solid-solution alloys having an hcp lattice structure are compared with those of Al-Mg solidsolution alloys with an fcc structure. The merit and demerit of Mg alloys will be discussed in terms of creep resistance. It has been reported recently that the addition of rare-earth (RE) elements significantly improves the mechanical properties of Mg alloys at elevated as well as at ambient temperatures.[2–6] The strengthening by the RE elements is more substantial than any other elements that have been added to Mg alloys. However, the strengthening mechanism of MgRE alloys has not been fully understood in the creep regime and will be discussed in the second part of this article, with special attention paid to Mg-Y alloys. The experimental results on Mg-Y and Mg-Al alloys have been reported in more detail elsewhere.[6,7]
KOUICHI MARUYAMA, Professor, and MAYUMI SUZUKI, Assistant Professor, are with the Department of Materials Science, Graduate School of Engineering, Tohoku University, Sendai, 980-8579, Japan. HIROYUKI SATO, Associate Professor, is with the Department of Intelligent Machines and System Engineering, Faculty of Science and Technology, Hirosaki University, Hirosaki 036-8561, Japan. This article is based on a presentation made in the symposium entitled “Defect Properties and Mechanical Behavior of HCP Metals and Alloys” at the TMS Annual Meeting, February 11-15, 2001, in New Orleans, Louisiana, under the auspices of the following ASM committees: Materials Science Critical Technology Sector, Structural Materials Division, Electronic, Magnetic & Photonic Materials Division, Chemistry & Physics of Materials Committee, Joint Nuclear Materials Committee, and Titanium Committee. METALLURGICAL AND MATERIALS TRANSACTIONS A
II. COMPARISON OF CREEP RESISTANCE BETWEEN Mg AND Al ALLOYS In order to develop an Mg alloy with better creep resistance than Al alloys, we ought to recognize the origins of the so-called poor creep resistance of Mg alloys. In the case of solid-solution alloys, the melting temperature of the alloy, solute concentration (N ), and atomic-size misfit (␦ ) of the solute atom in the solvent are the most important parameters determining creep resistance. These parameters are required to be as similar as possible when comparing the inherent creep resistance between Mg and Al alloys. The melting temperatures of Mg and Al are 924 and 934 K, respectively, and are similar to each other. The values of ␦ are ⫺12 pct for Al in Mg and ⫹12 pct for Mg in Al,[8] and the absolute values are the same in the two alloys. Therefore, Mg-Al and Al-Mg alloys were selected to examine the relative creep resistance of Mg and Al alloys. Creep specimens of Mg-Al alloys were cut from hot-extruded rods. They were annealed at 800 K for 7.2 ks in an argon gas atmosphere and had average grain sizes of 160 m. Tensile creep tests were conducted under constant stress in air. The experimental procedure has been reported in more detail elsewhere.[9] A. Creep Behavior of Mg-Al Alloys Before making the comparison, the fundamental creep properties of Mg-Al alloys are summarized in this section. Figure 1 shows representative creep curves of the Mg-Al binary solid-solution alloys tested at 600 K under 10 MPa. The curves reveal the typical three creep stages. The creep rate decreases in the primary creep stage, takes the minimum creep rate, and then accelerates toward rupture. The minimum creep rate is taken as a measure of creep resistance in the present article. It is to be noted that the minimum creep rate decreases, the rupture life becomes longer, and the rupture elongation decreases with increasing Al concentration. In general, the creep rate (˙ ) of metals and alloys is described by the following equation as functions of grain size (d ) and applied stress ( ):[10] p
G
n
冢冣冢 冣
DGb b ˙ ⫽ A kT d
[1]
where D, G, b, k, and T are the diffusion coefficient, the shear VOLUME 33A, MARCH 2002—875
Fig. 1—Representative creep curves of Mg-Al solid solution alloys tested at 600 K under 10 MPa.
Fig. 3—Concentration dependence of minimum creep rate of Mg-Al alloys.
A similar change in creep mechanism has been confirmed in Al-Mg solid-solution alloys.[12] Figure 3 shows the concentration dependence of minimum creep rate in the Mg-Al alloys at 600 K. The concentration exponent in Eq. [2] takes different values, depending on the creep mechanisms. At 10 MPa, the exponent m is unity, and this is typical of the solute-atmosphere dragging mechanism.[12] At 20 MPa, the value of m increases from unity to 2 at a high creep rate, corresponding to the change in stress exponent from n ⫽ 4 to 7. In the high-creep-rate (lowconcentration) region of m ⫽ 2, dislocations can escape from the solute atmosphere, resulting in the change in creep mechanism. B. Comparison of Normalized Creep Rate Fig. 2—Stress dependence of minimum creep rate of Mg-Al alloys.
modulus, the length of the Burgers vector, the Boltzmann’s constant, and the absolute temperature, respectively. The terms p and n are the grain-size and stress exponents, respectively, and A is a material constant. Structural materials are usually used in the dislocation-creep regime. In this regime, p ⫽ 0, and the equation can be modified to the following form: D G b ⫺m N ˙ ⫽ A⬘ kT G
n
冢冣
[2]
where A⬘ is a material constant. The effect of solute concentration (N ) is taken into account in the equation, and m is the concentration exponent. Minimum creep rates of the Mg-Al alloys are plotted in Figure 2 as a function of applied stress. The power law described by Eq. [2] holds, and the two regions with different values of stress exponent appear in the stress range of the figure. The stress exponent n is 4 at low stress and increases to 7 at high stress. The change in stress exponent corresponds to the change in creep-deformation mechanisms from an alloy type (n ⫽ 4: controlled by the glide of dislocations dragging a solute atmosphere) to a metal type (n ⫽ 7: controlled by recovery of the dislocation substructure).[9,11–13] 876—VOLUME 33A, MARCH 2002
Equation [2] points out that several material constants affect creep resistance. The diffusion coefficient, shear modulus, and length of the Burgers vector (atomic size) are the most important ones, and their values are different between Mg-Al and Al-Mg alloys. The diffusion coefficient and the length of the Burgers vector are larger and the shear modulus is smaller in magnesium than those in aluminum. In order to separate the effects of D, b, and G from the other inherent differences in creep resistance between the two alloy systems, the normalized creep rate defined by the following equation is used in the present comparison: ˙ k T ⫽ A⬘ N ⫺m DGb G
n
冢冣
[3]
Magnesium and aluminum have an hcp and fcc lattice structure, respectively. The contribution of the different lattice structures appears in the constant A⬘ in the present comparison. The value of D to be used is the interdiffusion coefficient ˜ ) when the solute-atmosphere dragging mechanism con(D trols creep deformation, or the self-diffusion coefficient (Dsd) when the recovery process determines creep rate. The normalized creep rates of the two alloy systems are given in Figure 4 as a function of the normalized creep stress ( /G). The data of Al-Mg alloys reported in Reference 12 were employed for the comparison. The interdiffusion coeffi˜ was used to normalize the creep rates, since the cient D METALLURGICAL AND MATERIALS TRANSACTIONS A
Table I. The Ratio of Internal Steels i to Creep Stress , ˜ , Length of Burgers Vectors b, and Interdiffusion Coefficient D Shear Modulus G at 600 K* Item r ⫽ i/ * (1 ⫺ r)r 2 ˜ (10⫺16 m2/s) D b (nm) G (GPa)
Mg-3 Pct Al
Al-1 Pct Mg
0.98 0.02 4.3 0.321 14
0.8 0.13 3.7 0.286 21
*The internal stress was measured at ⫽ 30 MPa in Mg-3 pct Al and at 15 MPa in Al-1 pct Mg.
Fig. 4—Normalized creep rate of Mg-Al and Al-Mg alloys as a function of normalized creep stress.
inter- and the self-diffusion coefficients are close to each other in both alloy systems. The values reported in the following literature were used for the normalization: the G and ˜ values of Mg-Al alloys as reported in Reference 11, and D those of Al-Mg alloys as reported in References 14 and 15. The transition from a low value of stress exponent to a high value is confirmed in all the alloys, and the transition stress increases with increasing alloy concentration. In the lowstress region, n ⫽ 4 in the Mg-Al alloys and n ⫽ 3.3 in the Al-Mg alloys, and the difference in the normalized creep rate between the two alloy systems increases with decreasing creep stress. The largest difference in creep rate is more than one order of magnitude between the two alloy systems, when the creep rates are compared at a constant solute concentration. This fact points out that the hcp structure is superior to the fcc structure in its inherent creep resistance. When dislocations drag a solute atmosphere and move in a viscous manner, the creep rate is represented by the Orowan equation, ˙ ⫽ b
[4]
where and are the dislocation density and the dislocation velocity, respectively. The dislocation velocity is proportional to the effective stress ( ⫺ i) and is given by[11,16] ˜ kT D ⫽ 0 2 5 ( ⫺ i) [5] G b N 앚␦앚2 where 0 is a constant, and ␦ is the size misfit between solute and solvent atoms. The internal stress (i) originates primarily from the dislocation substructure and is related to the dislocation density by the following equation:
i ⫽ ␣ M G b 冪
[6]
where ␣ is a numerical constant, and M is the Taylor factor. The following equation is obtained from Eqs. [4] through [6]:
0 k2 T 2 ˙ k T ⫽ ˜ Gb D ␣2 M2
冢
冣 冢N 앚␦앚 冣 冢G b 冣 (1 ⫺ r)r 1
3
2
1
5
2
7
r ⫽ i /
[7]
The first three terms on the right-hand side of this equation are essentially the same between the Mg-Al and Al-Mg alloy systems, but the other two terms are different. METALLURGICAL AND MATERIALS TRANSACTIONS A
Fig. 5—Comparison of minimum creep rates of Mg-Al and Al-Mg alloys at 600 K.
The internal stress was measured by the stress-transient dip technique in Mg-3 pct Al and Al-1 pct Mg alloys, and the results are listed in Table I. The ratio r of internal stress to applied stress is larger in the Mg-Al alloy than in the AlMg alloy. Five independent slip systems are provided by the 具110典{111} slip systems in fcc crystal, whereas the 具1120典(0001) slip systems in hcp crystal provide two independent slip systems only. It is not easy to relax the strain concentration when the number of available slip systems is small. The larger internal stress of the Mg-Al alloy is attributed to the limited number of slip systems available in hcp crystal. The normalized creep rate of the Mg-Al alloy, estimated from the last two terms of Eq. [7], is 1/20 of that of the Al-Mg alloy. This result explains most of the difference in the normalized creep rate between the two alloy systems. C. Comparison of Absolute Values of Creep Rate Although the comparison of the normalized creep rates is valuable from the scientific viewpoint, a comparison of the absolute values of creep rate at a constant creep stress is more relevant from the engineering point of view. The data points shown in Figure 4 are represented in Figure 5 in a modified form: the creep rate is plotted against creep stress. The values of DGb in Eq. [3] are 2.2 ⫻ 10⫺15 J/s in Mg-Al alloys and 1.9 ⫻ 10⫺15 J/s in Al-Mg alloys. The difference between these two values is not significant, but the shear modulus of magnesium is 2/3 of that of aluminum. VOLUME 33A, MARCH 2002—877
Table II. Chemical Composition of the Mg-Y Alloys Studied (in Mol Percent) Alloy 0.2Y 0.3YZ 0.9YZ 1.1Y 1.6YZ 2.4Y
Y
Zn
Al
Ca
Fe
Mn
Si
0.20 ⬍0.001 0.002 ⬍0.001 0.005 ⬍0.001 0.016 0.30 0.023 ⬍0.001 ⬍0.001 0.005 ⬍0.001 0.005 0.88 0.024 0.002 ⬍0.001 0.005 ⬍0.001 0.006 1.12 ⬍0.001 0.003 ⬍0.001 0.007 ⬍0.001 0.005 1.62 0.015 0.007 ⬍0.001 0.008 ⬍0.001 0.005 2.41 ⬍0.001 0.004 0.001 0.005 ⬍0.001 0.013
Mg bal bal bal bal bal bal
Fig. 6—Comparison of creep rates of several Mg alloys. The open symbols are solid solution alloys, and the solid symbols contain precipitates. Creep rates measured at several temperatures were converted to 550 K by using an activation energy of 140 kJ ⭈ mol⫺1.
Because of the lower value of the shear modulus in magnesium, the difference in creep rates between the two alloy systems is reduced to less than one order of magnitude in Figure 5 at a constant creep stress. However, the Mg alloys are still superior to the Al alloys in terms of creep resistance. This finding suggests that it is possible to develop commercial Mg alloys with better creep resistance than Al alloys. III. STRENGTHENING MECHANISM OF Mg-Y ALLOYS A. Relative Creep Resistance of Mg-Y Alloys In Figure 6, a comparison of creep rates is made among several Mg alloys: AZ91 (Mg-9 pct Al-1 pct Zn)[17,18] and AS21 (Mg-2 pct Al-1 pct Si)[17,18] commercial alloys, in addition to Mg-Al (single-phase, 160 m in grain size),[9] Mg-Mn (200 m in grain size),[19] and Mg-Y (50 m)[6] binary alloys. The AZ91 and AS21 alloys are commercial high-purity alloys produced by pressure die casting, and their dendrite size was 10 to 20 m. The Mg-Mn alloy contained ␣ -Mn particles homogeneously distributed within grains, and their spacing was 0.5 m. The Mg-2.4 pct Y alloy had grain-boundary precipitates, and dynamic precipitation on dislocations took place during creep (Section III– C–3). Data points measured at several temperatures are converted to 550 K by using an activation energy (Q) of 140 kJ⭈mol⫺1. The values of Q from 135 to 143 kJ⭈mol⫺1 have been reported in literature.[11,17,18] The open symbols represent solid-solution alloys, and the solid symbols are materials having precipitates. Except AS21 and the Mg-Y alloys, the data points of the other alloys are close to each other, suggesting that the presence of precipitates does not always provide prominent improvement in creep resistance. On one hand, a substantial improvement in creep resistance is achieved even in the Mg-0.2 pct Y dilute solid-solution alloy. The creep rate of the Mg-2.4 pct Y alloy decreases by four orders of magnitude as against AZ91. The addition of Y gives enormous strengthening in Mg alloys. 878—VOLUME 33A, MARCH 2002
Fig. 7—Stress dependence of minimum creep rate of Mg-Y alloys at 550 K.
It is known that an addition of a RE element improves the mechanical properties of Mg alloys at elevated as well as at ambient temperatures.[2–6] Several explanations have been proposed to explain the role of RE elements on the basis of solid-solution hardening and precipitation-hardening mechanisms.[20] In Mg-Y alloys, pseudoequilibrium  ⬙ and  ⬘ phases have been reported in addition to the equilibrium  phase (Mg24+xY5). The fine and uniformly dispersed  ⬙ precipitates are effective in strengthening at ambient temperatures.[20,21] However, the  ⬙ precipitates are not thermally stable and readily coarsen at elevated temperatures. The precipitation hardening by  ⬙ is simply not applicable to Mg-Y alloys in the creep regime. In this section, the strengthening mechanism of Mg-Y alloys is discussed in the creep regime. The chemical compositions of the alloys to be used in the following examination are listed in Table II. The materials were cast and then hot rolled (40 pct reduction) at 723 to 823 K. Specimens were solution treated, and the average grain size after the treatments was about 50 m in all the alloys. Compressive creep tests were carried out under constant stress in air. The crept specimens were water quenched under load for microstructural observations. The experimental procedure has been reported in detail elsewhere.[6] B. Creep Behavior of Mg-Y Alloys The typical three-stage creep curve similar to the ones shown in Figure 1 was always observed in the Mg-Y alloys of the present study. Minimum creep rates of the Mg-Y alloys are plotted as a function of applied stress in Figure 7. The change in stress exponent from 4 to 7 is known in METALLURGICAL AND MATERIALS TRANSACTIONS A
Fig. 8—Concentration dependence of minimum creep rate of Mg-Y and Mg-Al solid solution alloys together with an Mg-Mn precipitation-hardened alloy. The solid symbols represent alloys having precipitates. Fig. 9—Dislocation substructures observed in (a) Mg-1.1 pct Y and (b) Mg-2.4 pct Y alloys crept to minimum creep rate at 550 K under 100 MPa.
Figure 2 on the Mg-Al alloys. A similar change in stress exponent is recognized in Figure 7, but the exponent increases from 6 to 12 in the Mg-Y alloys. Because of the higher stress range employed in the creep tests of Mg-Y alloys, the region of n ⫽ 4 is absent in Figure 7. In the region of n ⫽ 6 (corresponding to the region of n ⫽ 7 in the Mg-Al alloys), dislocations can detach from the solute atmosphere with the aid of applied stress. A similar high stress exponent greater than 10 has been reported at high stress in an Mg-RE alloy.[5] The minimum creep rates of Mg-Y alloys are plotted against Y concentration (N ) in Figure 8 on the basis of Eq. [2]. The concentration exponent is 2, which is typical of the region of n ⫽ 6, as depicted in Figure 3. The Mg-2.4 pct Y alloy has precipitates at 550 K, and, owing to the precipitation hardening, its creep rate is lower than the straight line extended from the dilute alloys. Precipitates of ␣ -Mn are uniformly dispersed within grains in the Mg-Mn alloy.[19] The difference in creep rate between the Mg-0.6 pct Al solid-solution alloy and the Mg-0.5 pct Mn alloy is always one order of magnitude, independent of the testing temperature. On the other hand, the amount of the creep-rate decrease is sensitive to testing temperature in the Mg-Y alloys. The creep rate of the Mg-Y alloys decreases by one order of magnitude as against the Mg-Al alloys at 650 K, and the amount of creep-rate decrease becomes three orders of magnitude at 550 K. C. Strengthening Mechanism of Mg-Y Alloys 1. Solution hardening As evident in Figure 8, the Mg-Y alloys reveal the prominent strengthening without precipitates, suggesting solution hardening as their principal strengthening mechanism. The METALLURGICAL AND MATERIALS TRANSACTIONS A
solution hardening is represented by the following equation in the case of a simple mechanism based on the soluteatmosphere dragging:[11,16] ˜ /N 앚␦앚2 [8] ˙ ⬀ D This equation points out that the atomic-size misfit and the interdiffusion coefficient are the major parameters to be considered when discussing the solution hardening. The value of ␦ is ⫺12 pct for Al in magnesium and ⫹11 pct for Y.[6] The absolute values of ␦ of the two elements are close to each other and cannot explain the prominent difference in creep rate between Mg-Al and Mg-Y alloys, although the information on the effects of Y on the interdiffusion ˜ is not available at present. Ninomiya et al.[22] coefficient D have proposed another solution-hardening model based on a discrete variational-X␣ molecular orbital calculation, but the details of the solution-hardening mechanism of Mg-Y alloys have not been elucidated yet. 2. Forest dislocation hardening Figure 9 shows dislocation substructures of Mg-1.1 pct Y and Mg-2.4 pct Y alloys crept to the minimum creep rate at 550 K under 100 MPa. The incident-beam direction and the diffraction condition are [2110] and g ⫽ 0111, respectively. Dislocations having a Burgers vector other than 1/3[2110] are visible with the diffraction condition. Most of the dislocations are aligned parallel to the trace of basal planes in the Mg-1.1 pct Y alloy shown in Figure 9(a). They are a ⫽ 1/3具1120典–type dislocations moving on basal slip planes. The preferred slip on basal planes at 550 K is consistent with the result reported in Reference [23]. Most of the dislocations have the a-type Burgers vector also in the Mg2.4 pct Y alloy shown in Figure 9(b). However, many straight VOLUME 33A, MARCH 2002—879
Fig. 10—Densities of total dislocations, a-type dislocations, and nonbasal dislocations as a function of Y concentration in Mg-Y alloys. The Mg-0.2 pct Y alloy was crept at 60 MPa and the others at 100 MPa.
dislocations are inclined to the trace of basal planes in this alloy, suggesting that they were moving on nonbasal planes. The nonbasal dislocations act as obstacles to the motion of basal dislocations. The dislocation densities measured at 550 K in the MgY alloys are plotted against Y concentration in Figure 10. The densities were evaluated by the line-interception method, taking account of the foil thickness. The total dislocation density increases with increasing Y concentration, and most of the dislocations have the a-type Burgers vector. “Nonbasal” refers to the density of dislocations that are not laid on basal planes. It should be noted that the density of nonbasal dislocations increases significantly with Y concentration. The internal stress due to the forest (nonbasal) dislocations is described by Eq. [6] using the forest dislocation density (f) instead of . Loginov and Predcoditelev[24] have discussed the forest dislocation hardening in magnesium and proposed that ␣ ⫽ 1.21 for the constant in Eq. [6]. Taking M ⫽ 3, G ⫽ 14 GPa, and b ⫽ 0.321 nm, Eq. [6] gives
i / ⫽ 0.52 (Mg-1.1 pct Y) and ⬵1 (Mg-2.4 pct Y) This result points out that the internal stress due to the forest dislocations accounts for a significant part of creep stress, and the effective stress decreases with increasing Y concentration. This finding points out that the forest dislocation hardening plays an important role in the improvement of creep resistance in the Mg-Y alloys. 3. Dynamic precipitation during creep The equilibrium phase diagram[25] predicts that the  phase precipitates at 550 K in an Mg-Y alloy having a Y concentration higher than 1.3 mol pct. However, homogeneous precipitation of  phase is not easy, and Mg-Y alloys usually remain in a supersaturated solid-solution state. Heterogeneous precipitation readily occurs in the supersaturated alloys when nucleation sites (dislocations) are provided, resulting in dynamic precipitation during creep. Figure 11 shows a change in the transmission electron microstructure during creep of Mg-2.4 pct Y at 550 K under 100 MPa. The incident-beam direction is parallel to [0002]. There is no
880—VOLUME 33A, MARCH 2002
Fig. 11—Dynamic precipitation during creep of Mg-2.4 pct Y at 550 K under 100 MPa: (a) just after loading and (b) after creep for 72 ks to a minimum creep rate.
precipitate within grains immediately after loading (Figure 11(a)). Precipitates appear in the primary creep stage, and many precipitates are observed along dislocation lines at the minimum creep rate (Figure 11(b)). The precipitates are the pseudoequilibrium  ⬘ phase, and their number density increases with increasing creep strain. The dynamic precipitation was observed in Mg-2.4 pct Y only in the present study. Similar dynamic precipitation on dislocation lines has been reported in an Mg-Nd alloy.[26] The precipitates act as obstacles to dislocation motion. Therefore, the creep rate of Mg-2.4 pct Y in Figure 8(a) is lower than the straight line extrapolated from the dilute alloys.
METALLURGICAL AND MATERIALS TRANSACTIONS A
Fig. 12—Effect of impurity Zn on minimum creep rate of Mg-Y alloys.
D. Strengthening by Impurity Zn The Mg-Y alloys used in Sections III–A through III–C do not contain Zn. In Figure 12, the minimum creep rates of Mg-Y alloys containing a small amount of Zn (about 0.02 mol pct) are compared with those of the alloys without Zn. The chemical compositions of all the alloys used in the figure are listed in Table II. The presence of Zn as an impurity in Mg-Y alloys remarkably improves their creep resistance, especially at 650 K: the reduction in creep rate is by one order of magnitude. The same amount of reduction is confirmed at 550 K in Mg-0.3 pct Y, but the strengthening by impurity Zn disappears in the concentrated alloy (Mg-1.6 pct Y) at this temperature. Figure 13(a) and (b) show dislocation substructures observed after creep deformation to the minimum creep rate at 650 K under 30 MPa. The micrographs are a view from the [2110] direction. Prismatic slip is enhanced in magnesium at elevated temperatures, for example, at 650 K. All the dislocations in Figure 13(a) are not parallel to the trace of the basal plane, confirming the preferential slip on prismatic planes. The alloy containing Zn, shown in Figure 13(b), reveals a different dislocation substructure. There are a lot of bowed-out dislocations trailing straight dislocation segments parallel to the trace of basal planes. The bowed-out dislocations were moving on prismatic planes, as is the case in Figure 13(a). The trailed dislocation segments are laid on the intersection of the basal and prismatic planes. As is evident in Figure 13(a) and (b), the dislocation density is higher in the alloy with Zn, probably due to the presence of the trailed dislocation segments. The higher dislocation density is true in the other Zn-containing alloys, and the dislocation densities of Zn-containing alloys are twice as high as those of the alloys without Zn. The high dislocation density is one of the origins of the improved creep resistance of the Zn-containing alloys.
METALLURGICAL AND MATERIALS TRANSACTIONS A
Fig. 13—Dislocation substructures observed in (a) Mg-1.1 pct Y and (b) Mg-1.6 pct Y-0.015 pct Zn alloys crept to a minimum creep rate at 650 K under 30 MPa.
IV. CONCLUSIONS 1. The normalized creep rates of hcp Mg-Al solid-solution alloys decrease by more than one order of magnitude vs fcc Al-Mg solid-solution alloys, when the normalized creep rate, defined by Eq. [3], is compared at a constant normalized creep stress of /G. The lower creep rate is attributed to the high internal stress of the hcp structure, having only two independent basal slip systems. 2. Primarily because of a smaller shear modulus of magnesium, the difference in creep rate between Mg-Al and Al-Mg alloys is reduced to less than one order of magnitude, when the absolute value of the creep rate is compared at a constant-creep stress. However, creep rates of Mg-Al alloys are still lower than those of Al-Mg alloys, suggesting that the inherent creep resistance of Mg alloys is better than that of Al alloys. 3. Creep rates of Mg-Y alloys decrease by three orders of magnitude vs Mg-Al alloys when compared at a constant solute concentration. The strengthening by the Y addition is more significant than precipitation hardening in an MgMn alloy. 4. Solid-solution hardening is the principal strengthening mechanism of the Mg-Y alloys. However, the simple Cottrell atmosphere-dragging model cannot explain the prominent strengthening, and the detailed mechanism of the solid-solution hardening is not clarified yet. 5. Forest dislocation hardening in high concentration alloys and dynamic precipitation in an Mg-2.4 pct Y alloy also playing an important role in the strengthening of MgY alloys.
VOLUME 33A, MARCH 2002—881
6. An addition of a small amount of Zn significantly improves creep resistance of Mg-Y alloys. A higher dislocation density in the Zn containing alloys is a cause of the strengthening. ACKNOWLEDGMENTS This research was partly supported by Grant-in-Aid for Scientific Research (Grant No. 10650685 and 13750643) from The Ministry of Education, Science, Sports and Culture, Japan. The support of the Light Metals Educational Foundation is also acknowledged. REFERENCES 1. T. Ito and H. Shirai: J. Jpn. Inst. Light Met., 1992, vol. 42, pp. 707-19. 2. W. Henning and B.L. Mordike: in Strength of Metals and Alloys, H.J. McQeen, J.-P. Bailon, J.I. Dickson, J.J. Jonas, and M.G. Akben, eds., Pergamon Press, Oxford, United Kingdom, 1985, pp. 803-08. 3. H. Karimzadeh, J.M. Worrall, R. Pilkington, and G.W. Lorimer: Magnesium Technology, The Institute of Metals, London, 1986, pp. 138-41. 4. J.F. King, G.A. Fowler, and P. Lyon: in Light Weight Alloys for Aerospace Applications II, E.W. Lee and N.J. Kim, eds., TMS, Warrendale, PA, 1991, pp. 423-38. 5. M. Ahmed, G.W. Lorimer, R. Pilkington, and P. Lyon: in Magnesium Alloys and Their Applications, B.L. Mordike and F. Hehmann, eds., DGM Information, Oberursal, FRG, 1992, pp. 251-57. 6. M. Suzuki, H. Sato, K. Maruyama, and H. Oikawa: Mater. Sci. Eng. A, 1998, vol. 252, pp. 248-55. 7. H. Sato, M. Suzuki, K. Maruyama, and H. Oikawa: Key Eng. Mater., 2000, vol. 171-4, pp. 109-14. 8. H.W. King: J. Mater. Sci., 1966, vol. 1, pp. 79-90. 9. H. Sato and H. Oikawa: in Strength of Metals and Alloys, D.G. Brandon, R. Chaim, and A. Rosen, eds., Freund Publishing House, London, United Kingdom, 1991, pp. 463-70.
882—VOLUME 33A, MARCH 2002
10. A.K. Mukherjee, J.E. Bird, and J.E. Dorn: Trans. ASM, 1969, vol. 62, pp. 155-79. 11. S.S. Vagarali and T.G. Langdon: Acta Metall., 1982, vol. 30, pp. 1157-70. 12. H. Oikawa: in Hot Deformation of Aluminum Alloys, T.G. Langdon, H.D. Merchant, J.G. Morris, and M.A. Zaidi, eds., TMS, Warrendale, PA, 1991, pp. 153-80. 13. H. Sato, K. Maruyama, and Hiroshi Oikawa: in Aluminum Alloys, Their Physical and Mechanical Properties, T. Sato, S. Kumai, T. Kobayashi, and Y. Murakami, eds., The Japan Institute of Light Metals, Tokyo, 1998, pp. 1355-60. 14. H. Oikawa, H. Sato, and K. Maruyama: Mater. Sci. Eng. A, 1985, vol. 75, pp. 21-28. 15. H. Oikawa, N. Matsuno, and S. Karashima: Met. Sci., 1975, vol. 9, pp. 209-12. 16. A.H. Cottrell: in Dislocations and Plastic Flow in Crystals, Oxford University Press, Oxford, United Kingdom, 1953, pp. 136-39. 17. W. Blum, B. Watzinger, and P. Weidinger: in Magnesium Alloys and Their Applications, B.L. Mordike and K. U. Kainer, eds., WerkstoffInformationsgessellschaft mbH, Frankfurt, Germany, 1998, pp. 49-60. 18. B. Watzinger, P. Weidinger, F. Breutinger, W. Blum, R. Ro¨sch, H. Lipowsky, and H.-G. Haldenwanger: in Magnesium Alloys and Their Applications, B.L. Mordike and K.U. Kainer, eds., Werkstoff-Informationsgessellschaft mbH, Frankfurt, Germany, 1998, pp. 259-64. 19. M. Suzuki, H. Sato, and H. Oikawa: in Strength of Materials, H. Oikawa, K. Maruyama, S. Takeuchi, and M. Yamaguchi, eds., The Japan Institute of Metals, Sendai, 1994, pp. 555-59. 20. G.W. Lorimer: Magnesium Technology, The Institute of Metals, London, 1986, pp. 47-53. 21. T. Sato: Materia Jpn. (MTERE2), 1999, vol. 38, pp. 294-97. 22. R. Ninomiya, S. Kamado, M. Morinaga, and Y. Kojima: Materia Jpn. (MTERE2), 1999, vol. 38, pp. 305-09. 23. H. Yoshinaga and R. Horiuchi: Trans. JIM, 1964, vol. 5, pp. 14-21. 24. B.M. Loginov and A.A. Predcoditelev: Phys. Status Solidi (a), 1982, vol. 72, pp. 69-77. 25. T.B. Massalski, H. Okamoto, P.R. Subranamian, and L. Kacprzak: Binary Alloy Phase Diagrams, 2nd ed. ASM INTERNATIONAL, Metals Park, OH, 1990, pp. 2566-69. 26. T.J. Pike and B. Noble: J. Less-Common Met., 1973, vol. 30, pp. 63-74.
METALLURGICAL AND MATERIALS TRANSACTIONS A