Effects of Alloying Elements on Microstructure and Properties of Magnesium Alloys for Tripling Ball D.H. XIAO, Z.W. GENG, L. CHEN, Z. WU, H.Y. DIAO, M. SONG, and P.F. ZHOU In order to find good candidate materials for degradable fracturing ball applications, Mg-AlZn-Cu alloys with different contents of aluminum, zinc, and copper were prepared by ingot metallurgy. The effects of aluminum, zinc, and copper additions on the microstructure, compressive strength, and rapid decomposition properties of the alloys have been investigated using scanning electron microscopy, compressive tests, and immersion tests. The results show that the addition of high contents Al (15 to 20 wt pct) in pure magnesium promotes a large number of network-like b-Mg17All2 phases, which helps produce more micro-thermocouples to accelerate the corrosion process in 3 wt pct potassium chloride (KCl) at 366 K (93 °C). Adding different Zn contents improves the compressive properties of Mg-20Al alloys drastically. However, it decreases the decomposition rate in 3 wt pct KCl at 366 K (93 °C). Small amount of Cu will slightly reduce the compressive strength of Mg-20Al-5Zn alloy but dramatically increase its decomposition rate. DOI: 10.1007/s11661-015-3053-7 Ó The Minerals, Metals & Materials Society and ASM International 2015

I.

INTRODUCTION

HYDRAULIC fracturing is considered necessarily for economically efficient oil gas production in low natural permeability shale resources.[1] During the fracturing, some controlling devices such as the tripling balls are usually used. One main advantage of using the balls is to eliminate the lost production. Lost production occurs as a result of the pressure in the stage below the ball being lower than the pressure in the stage above the ball. The differential pressure of the stages on either side of the tripling ball causes the ball to remain sealed on the ball seat, and block the production from all stages below that point. So the tripling balls must be strong enough to bear the high pressure in the practical engineering process. After the oil gas production, these in-flow path devices must be disposed by some methods such as drilling out or flowing back to open the flow path.[1] The tripling balls to pile up in the well will create a debris barrier if the production velocity is not high enough, and the process of removing the remnant of the controlling devices might be complex and costly. It is an ideal and relatively inexpensive way to get rid of the tripling balls by decomposing them in the well with high temperature corrosion media which is often 3 wt pct potassium chloride (KCl) at 366 K (93 °C).[1] The decomposition rate or corrosive rate of tripling balls D.H. XIAO and M. SONG, Professors, and Z.W. GENG, L. CHEN, and P.F. ZHOU, Ph.D. Candidates, are with the State Key Lab of Powder Metallurgy, Central South University, Changsha 410083, P.R. China. Contact e-mail: [email protected] Z. WU, Research Scientist, and H.Y. DIAO, Ph.D. Candidate, are with the Department of Materials Science and Engineering, The University of Tennessee, Knoxville, TN 37996. Manuscript submitted December 29, 2014. Article published online July 10, 2015 METALLURGICAL AND MATERIALS TRANSACTIONS A

should be fast and controllable in a certain period of time. Conventional materials for the tripling balls are usually plastics, which cannot meet the requirements of the decomposition rate and strength. Magnesium alloys are usually served as anode materials for reserved batteries, due to their high energy density, high electrode potential, low density, and good corrosive performance.[2–8] Generally, the main alloying elements in magnesium alloys are aluminum and zinc.[9–15] When the content of Al is increased from 1 to 9 wt pct, the hardness of the magnesium alloys increases from 36.3 to 50.1 HV, due to the solid solution strengthening and fine grain strengthening.[9] Zinc is another effective alloying element in magnesium alloys for its dual function of solid solution strengthening and precipitation hardening.[10] However, the microstructure of Mg-Zn alloys usually coarsens because their crystallization temperature range is very large.[11] The Mg-Zn-Zr, Mg-Zn-RE, and Mg-Zn-Zr-RE alloys are studied to search for a better composition of alloys.[11–15] AZ91D alloy containing about 9 wt pct Al and 1 wt pct Zn is one of the most widely used magnesium alloys. The alloy has excellent mechanical properties, but the corrosion rate is only 0.8 mg h1 cm2 in 3 wt pct KCl at 366 K (93 °C).[16] Our previous studies showed that the properties of magnesium alloys can be improved greatly by changing the proportion of aluminum and zinc.[17] Additionally, the corrosion velocity of magnesium alloys can be affected when certain amount of Cu is added.[18] In this paper, new types of magnesium alloys with high fracture strength and reliable decomposition rates were developed as tripling ball. Unlike the traditional magnesium materials containing low contents of Al and Zn,[19] considerable amount of Zn, Al, and Cu will be added into the magnesium alloys. The effects of the alloying elements on the microstructures, mechanical VOLUME 46A, OCTOBER 2015—4793

properties, and corrosion rate of the alloys will be systematically studied and discussed.

II.

EXPERIMENTAL

The alloys with nominal compositions given in Table I were prepared in an electric resistance furnace with a steel crucible protected by RJ-2 flux, using commercially pure magnesium (99.9 pct), aluminum (99.6 pct), zinc (99.2 pct), and Al-49.5Cu (99.0 pct) as the raw materials. The melt was purged for 5 min at 993 K (720 °C). The casting temperature and the mold temperature were set as 893 K (710 °C) and 573 K (300 °C), respectively. The melt was poured into iron molds to produce billets with 100 mm in length and 60 mm in diameter. Specimens with a dimension of 10 9 10 9 10 mm were cut from the middle of the as-cast alloys, and mechanically grind using 1000, 1500, and 2000 grit silicon carbide papers and then polished using 0.5 lm alumina powder in 99.5 vol pct ethanol to mirrorbright. The D/max 2550pc X-ray diffractometer with Cu Ka radiation (k = 0.154 nm) was used to identify the main phases in materials. The microstructure observation of the samples as well as the compositional analysis of some certain phases was conducted using a Nava Nano 230 scanning electron microscopy (SEM) equipped with the Oxford X-ray energy dispersive spectroscopy. The compression properties of the alloys were investigated under compression at room temperature on a fully digital 3356 Instron testing machine at an engineering strain of 1 mm/min. The specimens were machined to /5 9 10 mm. The specimens were machined to /5 9 10 mm, and all faces were ground through 600-grit SiC paper. To determine the effects of Al, Zn, and Cu addition on the corrosion behavior of the alloys at different temperatures, a series of immersed corrosion tests were carried out. The samples were cut into small cylindrical ones (/17 9 10 mm) before being prepared in epoxy. Then, the only exposed surface of the samples is polished by alumina powder in 99.5 vol pct ethanol. Table I.

The Nominal Composition of the Alloys (Weight Percent)

Samples

Al

Zn

Cu

Mg

Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy

12 15 20 25 20 20 20 20 20 20 20 20

0 0 0 0 1.5 5 10 5 5 5 5 5

0 0 0 0 0 0 0 0.2 0.5 1 1.5 2.5

balance balance balance balance balance balance balance balance balance balance balance balance

1 2 3 4 5 6 7 8 9 10 11 12

4794—VOLUME 46A, OCTOBER 2015

Each epoxy-prepared sample was weighed prior to the tests. Several 500-ml beakers filled with 3 wt pct potassium chloride (KCl) were prepared in the HXS-H, which was set to 298 K and 366 K (25 °C and 93 °C), respectively. All of the beakers were equipped with plastic mesh barriers that were hung in one third from the bottom of the beakers to make sure that the fallen corrosion products would not prevent the reaction. When the temperature was increased to the set point, the samples were put into the separate beakers, and removed from the solution after 3, 9, 11, and 14.5 hours, respectively. The samples were dried and weighed after being cleaned using ultrasonic oscillation in the mixed solution of CrO3 (66 g/L) and AgNO3 (3.3 g/L) to get rid of the reserve corrosion products. The differences between the weight of the initial samples and the dried ones were the actual weight loss. The average corrosion rates (it can reflect the decomposition rate of magnesium alloys) were calculated using the following equation m1  m2 ; ½1 Average corrosion rate ¼ st where m1 and m2 are the original mass and the dried mass in microgram, respectively; t is time in hours; s = 1.54 cm2 is the exposed area in square centimeters.

III.

RESULTS

Figure 1 shows the microstructures of as-cast Mg-xAl (x = 12, 15, 20. 25) alloys named Alloy 1, Alloy 2, Alloy 3, and Alloy 4, respectively. The alloys are all composed of the phases a-Mg (the dark areas) and bMg17All2 (the white areas) which are tested by X-ray diffraction (Figure 2). However, the volume fraction and distribution patterns of each phase change, as the aluminum content increases. For Alloy 1, where the Al content is 12 wt pct, the b-Mg17All2 phase has a random distribution, while a-Mg phase exhibits a flocculentcoagulation-like distribution. Increasing Al content to 15 wt pct in Alloy 2, the volume fraction of the b-Mg17All2 phase is increased. At the same time, the b-Mg17All2 phase distributes as a mesh-like structure, resulting in the distribution of the a-Mg phase in a petal form. However, for Alloy 3, whose Al content is 20 wt pct, the petal form is partly destroyed to fragmental block structure and the mesh skeleton structure of b-Mg17All2 phase becomes bulky. For Alloy 4 with 25 wt pct Al, the b-Mg17All2 becomes a main phase and the a-Mg phase is distributed in the b-Mg17All2 phases, which is in agreement with the previous studies.[20] Figure 3 shows the X-ray diffraction patterns of ascast Mg-20Al-yZn (y = 0, 1.5, 5, 10) alloys, named Alloy 3, Alloy 5, Alloy 6, and Alloy 7, respectively. All the alloys with different contents of Zn contain the phases a-Mg and b-Mg17Al12. It can be seen that the intensity of the diffraction peaks changes as the Zn content increases. Compared with X-ray diffraction pattern of Alloy 3, the peak intensity of the b-Mg17Al12 phase in Alloy 5 with 1.5 wt pct Zn decreases and that of a-Mg phase increases. When the METALLURGICAL AND MATERIALS TRANSACTIONS A

Zn concentration reaches to 5 wt pct (Alloy 6), the peak intensity of b-Mg17Al12 phase decreases slightly, indicating the decrease of the volume fraction of b-Mg17Al12 phases. Furthermore, the additional peaks in XRD appear, which were identified to be s-Mg32(Al,Zn)49 phase. When the Zn content increases to 10 wt pct (in Alloy 7), the volume fraction of the b-Mg17Al12 phase does not increase as expected. Figure 4 shows the microstructures of Mg-20Al-yZn (y = 0, 1.5, 5, 10) alloys (i.e., Alloy 3, 5, 6, and 7, respectively) with different Zn contents. All of this information learned from the X-ray diffraction patterns, such as the change of the volume fraction of the phases b-Mg17Al12 and aMg with Zn concentration and the nucleation and growth of the s-Mg32(Al,Zn)49 phase in Mg-20Al-10Zn alloy, can be further confirmed by the scanning electron images taken from the Zn-contained alloys. Furthermore, the s-Mg32(Al,Zn)49 phases were found to be nucleate within b-Mg17Al12 phases. Figure 5 shows the X-ray diffraction patterns of Mg20Al-5Zn-zCu (z = 0, 0.2, 0.5, 1, 1.5, and 2.5) alloys named Alloy 6, Alloy 8, Alloy 9, Alloy 10, Alloy 11, and Alloy 12, respectively. Compared with the Cu-free Mg20Al-5Zn (Alloy 6), the new T phase (Al7Mg8Cu3Zn1) in the alloys appears with the addition of Cu. The volume fraction of b-Mg17Al12 phase does not change greatly

with the addition of Cu up to 0.5 wt pct. However, the diffraction peak intensity of b-Mg17Al12 phase decreases obviously, when Cu content reaches to 1 wt pct (Alloy 10). The above-mentioned observations can also be confirmed by the scanning electron images (Figures 6 and 7). Similarly, the microstructures of the Cu-contained

Fig. 2—X-ray diffraction patterns of the Mg-xAl alloys.

Fig. 1—Microstructures of Mg-xAl alloys. (a) Alloy 1; (b) Alloy 2; (c) Alloy 3; and (d) Alloy 4.

METALLURGICAL AND MATERIALS TRANSACTIONS A

VOLUME 46A, OCTOBER 2015—4795

Mg-20Al-5Zn also mainly consist of the phases a-Mg and b-Mg17All2 (Figure 6) whose volume fractions and distribution are not significantly changed with Cu contents. T phase is found to precipitate in b-Mg17Al12 phases and mainly locates at grain boundaries of a-Mg

phases (Arrowed in Figure 6). The volume fraction of T phase increases as Cu concentration increases. In Mg-20Al-5Zn-1.5Cu alloy (Alloy 11), b-Mg17Al12 phase

Fig. 3—X-ray diffraction patterns of Mg-20Al-yZn alloys.

Fig. 5—X-ray diffraction patterns of Mg-20Al-5Zn-zCu alloys.

Fig. 4—Microstructures of Mg-20Al-yZn (y = 0, 1.5, 5, 10) alloys. (a) Alloy 3; (b) Alloy 5; (c) Alloy 6; (d) Alloy 7. 4796—VOLUME 46A, OCTOBER 2015

METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 6—Microstructures of Mg-20Al-5Zn-zCu alloys. (a) Alloy 6; (b) Alloy 8; (c) Alloy 9; (d) Alloy 10; (e) Alloy 11; (f) Alloy 12.

connects to each other and forms a fine network and the grain size of a-Mg phase decreases. When the content of Cu increased to 2.5 wt pct, T phase starts to aggregate and the size of a-Mg grains increases. Table II shows the compressive properties of the ascast alloys tested at room temperature. By comparing the values for Alloys 1, 2, 3, and 4, we will be able to appreciate the effects of Al addition on the mechanical properties of pure Mg alloys. It shows that Al METALLURGICAL AND MATERIALS TRANSACTIONS A

content has a significant influence on both the yield strength (r0.2) and fracture strength (rf) of the investigated Mg alloys. When Al content was increased from 12 to 25 wt pct, the yield strength was almost doubled. The fracture strength reaches its maximum of 414 MPa at Mg-15Al, which was reduced to 364 MPa in Mg-25Al alloy. From Table II, the fracture strain (d) of Mg-xAl alloys with different Al contents is over 5 pct. VOLUME 46A, OCTOBER 2015—4797

Fig. 7—The energy spectrum analysis of T phase in Mg-20Al-5Zn-2.5Cu alloy.

Table II. Sample Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy Alloy

1 2 3 4 5 6 7 8 9 10 11 12

Mechanical Properties of the Alloys

rf (MPa)

r0.2 (MPa)

ef (Percent)

394 414 406 364 428 468 485 444 439 414 462 439

189 229 327 352 312 363 391 335 333 310 358 332

11.2 7.2 6.5 5.3 5.2 5.0 4.6 6.1 6.2 6.8 5.5 6.3

By comparing the values for Alloy 3, 5, 6, and 7, both r0.2 and rf show positive dependencies on Zn content (Table II). The fracture strength of the alloys significantly improves, when the Zn content increases from 1.5 to 10 wt pct. The fracture strength of Alloy 7 (10 wt pct in Zn) reaches the peak value of 485 MPa. However, the yield strength of the alloys does not have the same trend. Adding 1.5 wt pct Zn makes the yield strength of Alloy 5 drop slightly, from 327 to 312 MPa. Furthermore, Alloy 7 with 10 wt pct Zn has the highest yield strength of 391 MPa. Compared with Al and Zn, Cu has much more complex effects on the mechanical properties of Mg20Al-5Zn alloys (Table II). Both the r0.2 and rf show an initial drop, when the Cu content increased to 1 wt pct (Alloy 10). However, when the Cu content increases to 1.5 wt pct, the r0.2 and rf of Alloy 11 return to the level of Alloy 6 with free Cu. The fracture strength and yield strength of Alloy 12 decreases to 439 and 332 MPa, respectively. Figure 8 shows the results of immersion corrosion results of Mg-xAl alloys at 366 K, 343 K, and 298 K (93 °C, 70 °C, and 25 °C), respectively. It can be seen that the average corrosion rates at 366 K (93 °C) are much higher than those at 343 K and 298 K (70 °C and 25 °C), due to the increased activity of the molecules, ions, and atoms in the liquid with increasing the temperature. Moreover, increasing Al contents from

4798—VOLUME 46A, OCTOBER 2015

12 to 15 wt pct leads to a significant improvement in the corrosion rate. When the Al content reaches 20 wt pct, the average corrosion rates of Alloy 3 after immersion for 14.5 hours are 27.5, 3.6, and 0.75 mg cm2 h1 at 366 K, 343 K, and 298 K (93 °C, 70 °C, and 25 °C), respectively. However, the average corrosion rate of Alloy 4 with 25 wt pct Al drops by 45.5, 58.3, and 80.1 pct at 366 K, 343 K, and 298 K (93 °C, 70 °C, and 25 °C), respectively, compared with those of Alloy 3 with 20 wt pct Al. Figure 9 shows the average corrosion rate of Mg20Al-yZn (y = 0, 1.5, 5, 10) alloys at 366 K, 343 K, and 298 K (93 °C, 70 °C, and 25 °C), respectively. Although the average corrosion rate at 366 K (93 °C) differs obviously from the results at 343 K and 298 K (70 °C and 25 °C), the average corrosion rate of the alloys with different Zn contents at different temperatures shows similar evolution trend with time. When adding 5 wt pct Zn, the average corrosion rate of Alloy 6 increases dramatically to the peak at 40 mg cm2 h1 [after immersion for 14.5 hours at 366 K (93 °C)], 7.5 mg cm2 h1 [for 14.5 hours at 343 K (70 °C)], and 2.1 mg cm2 h1 [immersion for 14.5 hours at 298 K (25 °C)], respectively. However, the average corrosion rate of Alloy 7 at 366 K (93 °C) decreases to 30 mg cm2 h1 and to 7.5 mg cm2 h1 at 343 K (70 °C). At 298 K (25 °C), the average corrosion rate of the Alloy 7 shows a relatively large reduction to 2 mg cm2 h1. When the addition of Zn is 1.5 wt pct, the tested results of Alloy 5 at 298 K and 343 K (25 °C and 70 °C) are between the corrosive rate of Alloy 3 and Alloy 6. It indicates that the addition of 1.5 wt pct Zn changed the corrosion behavior, but the increased rate is still lower than that of Alloy 6 with 5 wt pct Zn under the same situation. However, it can be seen that the corrosion rates of Alloy 6 after immersion for 9 hours or 11 hours are slightly lower than that of Alloy 3 tested at 366 K (93 °C). Figure 10 shows the average corrosion rate of Mg20Al-5Zn-zCu (z = 0, 0.2, 0.5, 1, 1.5, and 2.5) alloys tested at 366 K, 343 K, and 298 K (93 °C, 70 °C, and 25 °C), respectively. The addition of Cu accelerates the corrosion process of the alloys. When the Cu contents increase from 0 to 1.5 wt pct, the average corrosion rate

METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 8—Immersion corrosion rates and the polarization curves at 366 K, 343 K, and 298 K (93 °C, 70 °C, and 25 °C) of Mg-xAl (x = 12, 15, 20, and 25) alloys.

increases 52.5 pct from 40 to 61 mg cm2 h1 at 366 K (93 °C). However, when the content of Cu is 2.5 wt pct, the average corrosion rate of Alloy 12 decreases to 55 g cm2 h1. The average corrosion rate of the Mg20Al-5Zn-zCu alloys at 298 K (25 °C) is slightly slower than those at 366 K (93 °C). In general, the average corrosion rate increases with increasing the Cu contents from 0 to 1.5 wt pct, but when the Cu addition is 2.5 wt pct, the average corrosion rate of Alloy 12 decreases.

IV.

DISCUSSION

A. Effects of Al Addition on the Microstructures and Properties of Pure Mg Figures 1 and 2 show that adding high Al contents from 15 to 20 wt pct in pure magnesium promotes the formation of the b-Mg17All2 phase with a network-like distribution. The compressive strength and the corrosion rate of Mg-20Al alloy reached the maximum values. However, when the Al content further increases to 25 wt pct, the b-Mg17All2 phases in Alloy 4 become much thicker, leading to a decrease of the compressive strength and the average corrosion rate. The microstructure of the alloys changed from the columnar to dendrite after adding higher contents of Al METALLURGICAL AND MATERIALS TRANSACTIONS A

(as shown in Figure 1). When the Al content is 12 wt pct, zigzag secondary dendrites nucleated at the boundaries of the original a-Mg phases. However, when the Al concentration is low, the dendrites in the alloys could not grow up due to the low surpercooling degree of composition. While increasing the Al content to 15 wt pct, the dendrites grow rapidly and form the petal tissue, due to the larger supercooling degree. Further the addition of Al to the Mg-xAl alloys, e.g., 20 and 25 wt pct, leads to more grain boundary migration of the Al atoms during the solidification. It does not only make the skeleton coarsen, but also decrease the liquid melting point in the micro-zone to the eutectic solidification point, which helps to stop the development of the dendrites in the alloys. Therefore, small block-like phases instead of the petal structure are generated. If the content of Al is 25 wt pct, the crystallization could be faster due to the larger supercooling degree. The newborn compounds surround the original a-Mg crystal to form the scattered small blocks.[20] Compared to the traditional AZ magnesium alloys (Mg-Al series), the content of Al in the current investigated alloys increases to 12 to 25 wt pct. The microstructure of the alloys changes from a-Mg phase to eutectic mixture (a-Mg+b-Mg17Al12). The b-Mg17Al12 is the effective strengthening phase, which can promote the formation of dislocation tangles, pin dislocation movements, and grain boundary sliding, and thus VOLUME 46A, OCTOBER 2015—4799

Fig. 9—Immersion corrosion rates and the polarization curves at 366 K, 343 K, and 298 K (93 °C, 70 °C, and 25 °C) of Mg-20Al-yZn (y = 0, 1.5, 5, and 10) alloys.

improve the strength of the alloys (as shown in Table II). Figure 8 shows that the initial reaction rate of the Mg-xAl alloys is slow, and subsequent reaction velocity becomes higher during all immersion testing. This may be the fact that magnesium alloys are at first in the state of self-corrosion or weak anodic polarization which leads to the corrosion hysteresis effect. The effect can also be reflected by the rate of hydrogen evolution. Moreover, the covering film also affects significantly the corrosion behavior. The film includes three layers: Al2O3 (the innermost layer), MgO (the interlayer), and Mg (OH)2 (the outermost and dominant layer).[21] The layers could help to protect the alloy from being corroded in the salt liquor, resulting in a slow corrosion rate at the beginning. However, when the local corrosion occurs, the integrity of the film is severely damaged, and the incompletion could become more serious with the increase of the polarization potential caused by the broken film or the non-film part. The a-Mg phase serves as the positive pole where the ‘‘anode hydrogen evolution’’ occurs and the reaction site locates mainly in the middle of the gains. Moreover, the corrosion potential of the b-Mg17Al12 phase, which is the main second phase of the alloys, is much higher than that of a-Mg phase, and the electric current caused by cathode polarization is higher than that of a-Mg phase, and thus results in the b-Mg17Al12 phase being an effective cathode.[21–24] 4800—VOLUME 46A, OCTOBER 2015

Figure 8 also shows that the average corrosion rate of Mg-xAl alloys at both 366 K and 298 K (93 °C and 25 °C) in 3 wt pct KCl increases gradually and reaches the maximum values when the Al content is 20 wt pct, and then the corrosion rate decreases. This is mainly due to the fact that the change of the microstructure of the alloys affects the corrosion properties. Figure 1 shows that the volume fraction and distribution patterns of the b-Mg17Al12 phases change as the aluminum content increases. The redistribution of the bMg17Al12 phases helps to establish more micro-galvanic batteries, since the b-Mg17Al12 phase has a much higher self-corrosion potential, and thus it is much easier to perform hydrogen evolution on the phase.[21] There are also some factors which could slow the corrosion process. The increased Al content in the film can strengthen the film and then protect the alloy more effectively. However, the film could not provide enough protection in the solution containing chloride ions. When 25 wt pct Al is added to the base alloy, the corrosion rate of the Alloy 4 drops sharply, due to the film covering the phases (Al,Mg)m(OH)n and (Mg,Al)[21] The film is rich in aluminum and stable, and xOy. maybe leads to the passivation of metals. In addition, the net-like b-Mg17Al12 phases are so thick that the aMg phases are divided into small pieces and almost completely isolated. The b-Mg17Al12 phase cannot be corroded easily either. Thus, the corrosion process into METALLURGICAL AND MATERIALS TRANSACTIONS A

Fig. 10—Immersion corrosion rates and the polarization curves at 366 K, 343 K, and 298 K (93 °C, 70 °C, and 25 °C) of Mg-20Al-5Zn-zCu (z = 0, 0.2, 0.5, 1, 1.5, and 2.5) alloys.

the interior of the alloy is blocked by b-Mg17Al12 phase. B. Effects of Zn Addition on the Microstructures and Properties of Mg-20Al Alloy As shown in Figure 4, the microstructures of Mg20Al-yZn alloys vary dramatically with increasing the Zn content. The volume fraction of the b-Mg17Al12 phases decreases when the Zn content increases from 0 to 1.5 wt pct. It should be noted that the addition of Zn to Mg-20Al alloy reduces the solid solubility of Al in Mg,[18] and thus maybe leads to the increase of the volume fraction of b-Mg17Al12 phase. However, the alloying elements in Mg-Al-Zn alloys tend to be gathered and distributed in a-Mg and b-Mg17Al12 phases to strengthen the alloy. When the Zn content is 1.5 wt pct and the mass ratio of Zn/Al far below 1/2, the dominant second phase in the Alloy 5 is the b-Mg17Al12 phases.[19,25,26] However, the non-equilibrium distribution might cause the local enrichment of zinc, resulting in the high values of ratio of Zn/Al (e.g., 1/2). As a result, small block s-Mg32(Al,Zn)49 phases will thus be produced in the alloy.[26] Owing to the limited quantity and volume fraction of the phase, it cannot be detected by XRD, but it can be observed clearly by SEM,

METALLURGICAL AND MATERIALS TRANSACTIONS A

as shown in Figure 4(b). The formation of the s-Mg32(Al,Zn)49 phase maybe decreases the volume fraction of the b-Mg17Al12 phase. When the content of Zn is 5 wt pct, the volume fraction of s-Mg32(Al,Zn)49 phases increases and that of the b-Mg17Al12 phases decreases, due to the consumption part of aluminum element. Moreover, Alloy 6 becomes easier to get constitutional supercooling, which greatly promotes the dispersion of phase b-Mg17Al12.[27] The a-Mg phases are separated into smaller blocks. Adding 10 wt pct Zn in the Mg-20Al alloy forms more s-Mg32(Al,Zn)49 phases. Most s-Mg32(Al,Zn)49 phases are incoherent transition phases with a thin rod or dot shape,[27] and the orientation of the phases could help to pin the dislocation movements. The newly formed phases are distributed uniformly in the b-Mg17Al12 phase, and the size of a-Mg phase becomes slightly smaller. At the same time, adding Zn significantly affects the mechanical properties (Table II), partially due to the refined microstructures. The solid solution strengthening of Zn in magnesium alloys also helps improve the mechanical properties of the Mg-20Al-yZn alloys. It should be noted that the effect of the Zn contents on the mechanical properties is complicated. When the Zn content is 1.5 wt pct, the volume fraction of b-Mg17Al12 phases decreases greatly, as well as the yield strength.

VOLUME 46A, OCTOBER 2015—4801

Adding 5 wt pct Zn or 10 wt pct Zn leads to the increase of the volume fraction of s-Mg32(Al, Zn)49 phases, which improves the mechanical properties dramatically. The s-Mg32(Al, Zn)49 phase is mixed with the b-Mg17Al12 phase, and helps isolate the a-Mg phase into smaller pieces. All these effects are in favor of the tangled dislocations and thus improve the mechanical properties effectively. Additionally, the average corrosion rate of the Mg20Al-yZn alloys is accelerated, due to the addition Zn element (Figure 9). The results disagree with some previous investigations.[28] It was shown that the potential of magnesium shifts toward positive caused by Zn.[28] Adding Zn to magnesium alloys helps improve the solubility limit of impurity, which resists the tendency of local corrosion.[4] The disagreement may be partially caused by the high Al content in magnesium alloys. The aluminum-rich magnesium alloys have large volume fraction of b-Mg17Al12 phase which will connect together to form network structure. Even though adding 1.5 wt pct Zn in the alloy makes the volume fraction of b-Mg17Al12 phase decrease, the thinner distribution of b-Mg17Al12 phase and the formation of s-Mg32(Al, Zn)49 phase lead to the generation of more microthermocouples. The corrosion rate is thus increased. When 5 wt pct Zn is added in Alloy 6, the volume fraction of the b-Mg17Al12 phase decreases obviously, and the phase becomes dispersive and fine, which contributes to the improvement of the corrosion rate. However, the corrosion rates of Alloy 7 with 10 wt pct Zn seem to be relatively much lower. Though the newborn phases help to create more micro-thermocouples to accelerate the corrosion velocity, the reduction of the volume fraction of the a-Mg phase makes the corrosion velocity of Alloy 7 decrease. Apart from the protection effect of b-Mg17Al12 phase, the protective films maybe play an important role during the corrosion process. C. Effects of Cu Addition on the Microstructures and Properties of Mg-20Al-5Zn Alloy Figures 6 and 7 show that Cu addition has a great influence on Mg-20Al-5Zn alloy. The formation of T (Al7Mg8Cu3Zn1) phase in the alloys may be due to the decomposition of the b-Mg17Al12 phase during the solidification process and the nucleation of T phase from b-Mg17Al12 phase. When the Cu content is low (e.g., 0.2 or 0.5 wt pct), the volume fraction of T (Al7Mg8Cu3Zn1) phase is too small to influence the quantity and distribution of the phases a-Mg and b-Mg17Al12. When the Cu content reaches to 1 wt pct, the consumption of b-Mg17Al12 phase in Mg-20Al-5Zn1Cu alloy leads to the growth of T (Al7Mg8Cu3Zn1) phase. Adding 1.5 wt pct Cu to Mg-20Al-5Zn alloy results in the continuous growth of the b-Mg17Al12 phases to develop a network structure. When the Cu content is 2.5 wt pct, the volume fraction of the b-Mg17Al12 phase in alloy 12 increases and the a-Mg grains grow. Generally, the addition of Cu reduces the compression performance because of the appearance of the loose structure, due to the precipitation of the T phase from b-Mg17Al12 phase. When the Cu content reaches to 4802—VOLUME 46A, OCTOBER 2015

1 wt pct, however, the compressive strength and the yield strength experience a dramatic increase to the maximum values in the alloys containing 0 to 2.5 wt pct Cu. One possible reason could be the refinement of the microstructure, as shown in Figure 6. Another reason could be the prevention of dislocation motion by the new-born phase. Moreover, the T (Al7Mg8Cu3Zn1) phase nucleating along the dislocations will strengthen the alloys through Orowan mechanism.[18] However, adding 2.5 wt pct Cu in Alloy 6 causes the connection of existing phases, and hence the decrease of compression performance. According to Table II and Figure 10, the Mg-20Al5Zn-1.5Cu alloy (Alloy 11) has the good decomposition rate in 3 pct KCl liquor [61 mg cm2 h1 at 366 K (93 °C) for 14.5 hours] as well as the compression performance (the fracture strength of 462 MPa and yield strength of 358 MPa). The performances reach the engineering standard of material for tripling ball application. The alloy may be used as the tripling ball material.

V.

CONCLUSIONS

1. Adding high contents of Al (12 to 20 wt pct) in pure magnesium can promote the generation of bMg17Al12 phases, and so strengthen the alloys and produce more micro-thermocouples so as to accelerate the decomposition process. 2. Adding different contents (0 to 10 wt pct) of Zn in Mg-20Al alloy generally improves the compression performance of the alloys, but reduces the decomposition rate in 3 pct KCl liquor. The fracture strength and the yield strength of alloys rise to the maximum of 468 and 363 MPa, respectively, when the content of Zn is 5 wt pct. 3. When minor Cu is added to Mg-20Al-5Zn alloy, the compression performance experiences a minor reduction, but the decomposition rate increases dramatically. The alloy containing 1.5 wt pct Cu has the highest average corrosion rate, 143 mg cm2 h1 at 366 K (93 °C), and the fracture strength and the yield strength of the alloy are 462 and 359 MPa, respectively. 4. In this paper, the Mg-20Al-5Zn-1.5Cu alloy has good decomposition rate in 3 pct KCl liquor [61 mg cm2 h1 at 366 K (93 °C) for 14.5 hours] as well as the compression performance (the fracture strength of 462 MPa and yield strength of 358 MPa). The performances reach the engineering standard of material for tripling ball. The material may be used as the tripling ball material.

ACKNOWLEDGMENTS This work is supported by the National Natural Science Foundation of China (51021063), the Fundamental Research Funds for Central Universities METALLURGICAL AND MATERIALS TRANSACTIONS A

(2011JQ021), and the Fund of China Scholarship Council (201306375039).

REFERENCES 1. X.Y. Xu, G. Agrawal, and B. Salinas: US Patent, 2011, US 2011/ 0135953 A1. 2. M. Yuasa, X.S. Huang, K. Suzuki, M. Mabuchi, and Y. Chino: Mater. Trans., 2014, vol. 55, pp. 1202–07. 3. N.T. Kirkland, N. Birbilis, and M.P. Staiger: Acta Biomater., 2012, vol. 8, pp. 925–36. 4. N.T. Kirkland, G. Williams, and N. Birbilis: Corros. Sci., 2012, vol. 65, pp. 5–9. 5. A.S. Hamdy and D.P. Butt: Electrochim. Acta, 2013, vol. 108, pp. 852–59. 6. G. Song and A. Atrens: Adv. Eng. Mater., 2003, vol. 5, pp. 837–35. 7. X.B. Chen, N. Birbilis, and T.B. Abbott: Corrosion, 2011, vol. 67, pp. 1–16. 8. T.D. Gregory and R.C. Winterton: Electronchem. Soc., 1990, vol. 137, pp. 775–80. 9. D.X. Xiao, M. Song, F.Q. Zhang, and Y.H. He: J. Alloys Compd., 2009, vol. 484, pp. 416–21. 10. J.F. Nie: Metall. Mater. Trans. A, 2012, vol. 43A, pp. 2891–2939. 11. D.H. Sthohn, M.A. Easton, M. Qian, and J.A. Taylor: Metall. Mater. Trans. A, 2013, vol. 44A, pp. 2935–49. 12. T.J. Lee and W.J. Kim: J. Alloys Compd., 2014, vol. 617, pp. 352–59. 13. J.Y. Lee and H.K. Lim: Mater. Sci. Eng. A, 2007, vols. 449–451, pp. 987–90.

METALLURGICAL AND MATERIALS TRANSACTIONS A

14. Y.M. Zhu, A.J. Morton, and J.F. Nie: Acta Mater., 2012, vol. 60, pp. 60–619. 15. P. Dobroa, J. Balik, F.E. Chmelik, K. Lllkova, and J. Bohlen: J. Alloys Compd., 2014, vol. 588, pp. 628–32. 16. L.Q. Bai, K.Y. Shu, and D. Li: J. Aero Mater., 2010, vol. 30, pp. 62–66. 17. D.H. Xiao and X.X. Li: China Patent, 2013, No.20130284659. 18. M.M. Avedesian and H. Baker: ASM Specialty Handbook—Magnesium and Magnesium Alloys, ASM International, Materials Park, OH, 1999. 19. J. Zhang, Z.X. Guo, F.S. Pan, Z.S. Li, and X.D. Luo: Mater. Sci. Eng. A, 2007, vol. 456, pp. 43–51. 20. H.X. Wang, W. Liang, Q. Zhao, Y.P. Yin, and X.G. Zhao: Rare Met. Mater. Eng., 2008, vol. 37, pp. 2004–07. 21. G.L. Song, A. Atrens, X. Wu, and B. Zhang: Corros. Sci., 1988, vol. 40, pp. 1769–91. 22. G.L. Song, A. Atrens, and M. Dargusch: Corros. Sci., 1999, vol. 41, pp. 249–73. 23. H.S. Kim and W.J. Kim: Corros. Sci., 2013, vol. 75, pp. 228– 38. 24. G.L. Song and A. Atrens: Adv. Eng. Mater., 1999, vol. 1, pp. 11– 33. 25. J. Zhang, R.L. Zuo, Y.X. Chen, and F.S. Pan: J. Alloys Compd., 2008, vol. 448, pp. 316–20. 26. G.Q. Li, G.H. Wu, and Y. Fan: Foundry Technol., 2005, vol. 26, pp. 922–26. 27. S.K. Das, D.H. Kang, and I.H. Jung: Metall. Mater. Trans. A, 2014, vol. 45, pp. 5212–18. 28. Y. Fan, G.H. Wu, and H.T. Gao: J Electrochem. Soc. B, 2006, vol. 8, pp. 283–88.

VOLUME 46A, OCTOBER 2015—4803