Journal of Materials Science & Technology  2019 , 35 (9): 1860-1868 https://doi.org/10.1016/j.jmst.2019.05.006

Orginal Article

Enhanced tensile properties in a Mg-6Gd-3Y-0.5Zr alloy due to hot isostatic pressing (HIP)

B. Zhouab, D. Wua*, R.S. Chena*, En-hou Hana

a The Group of Magnesium Alloys and Their applications, Institute of Metal Research, Chinese Academy of Sciences, Shenyang 110016, China
b School of Materials Science and Engineering, University of Science and Technology of China, Hefei 230026, China

Corresponding authors:   *Corresponding authors.E-mail addresses: dwu@imr.ac.cn (D. Wu), rschen@imr.ac.cn (R.S. Chen).*Corresponding authors.E-mail addresses: dwu@imr.ac.cn (D. Wu), rschen@imr.ac.cn (R.S. Chen).

Received: 2018-12-27

Revised:  2019-01-25

Accepted:  2019-02-20

Online:  2019-09-20

Copyright:  2019 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

Hot isostatic pressing (HIP) was applied to Mg-6Gd-3Y-0.5Zr (GW63) alloy to reduce shrinkage porosity, thus, to enhance the integrity and reliability of castings. During HIP process, shrinkage porosity was closed by grain compatible deformation and subsequent diffusion across the bonding interface. The amount of initial shrinkage porosity was the key factor for shrinkage porosity closure. HIP was testified to be effective on shrinkage porosity reduction in GW63 alloy due to its relatively narrow solidification range and resultant low content of initial shrinkage porosity in most sections, leading to higher tensile properties both in as-cast and cast-T6 condition. The improvement in tensile properties was mainly because of shrinkage porosity reduction and resultant effective rare-earth (RE) elements homogenization and precipitation strengthening.

Keywords: Hot isostatic pressing ; Mg-Gd-Y-Zr ; Shrinkage porosity reduction ; Tensile properties

0

PDF (5097KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

B. Zhou, D. Wu, R.S. Chen, En-hou Han. Enhanced tensile properties in a Mg-6Gd-3Y-0.5Zr alloy due to hot isostatic pressing (HIP)[J]. Journal of Materials Science & Technology, 2019, 35(9): 1860-1868 https://doi.org/10.1016/j.jmst.2019.05.006

1. Introduction

Magnesium (Mg) alloys are prone to form shrinkage porosity because of their relatively wide solidification range [[1], [2], [3], [4]]. Shrinkage porosity is an intrinsic and unavoidable defect resulting from shrinkage during solidification [5,6]. It is well recognized that shrinkage porosity is deleterious to tensile properties and a major cause of casting rejection [[7], [8], [9]]. In our precious research [10], the Vickers hardness of as-cast Mg-5.02Y-1.92Nd-2.13Gd-0.45Zr (WE54) alloy declined linearly with shrinkage porosity volume fraction increasing and the tensile strength and yield strength declined exponentially as shrinkage porosity volume fraction increased. Besides, Lee [11] found an inverse parabolic relationship between tensile properties and variation of porosity on the fracture surface in both AM60B and AZ91D Mg alloys.

Hot isostatic pressing (HIP) is widely applied to metal powder/ceramic consolidation and to porosity closure in castings in aerospace, marine and automotive industries [[12], [13], [14], [15], [16]]. Uniform density very close to theoretical density, elimination of porosity, improved mechanical properties and decreased scatter of properties can be achieved by HIP [[17], [18], [19]]. For example, fine-grained and transparent Nd:YAG was obtained at SiO2 doping levels as low as 0.02 wt.% by sinter plus HIP approach [20]. Bor et al. [21] reported the microporosity was eliminated by HIP and the tensile strengths and elongations were significantly improved in fine grain MAR-M247 superalloys. In addition, a reduction in fatigue data scattering and an increase in fatigue resistance were obtained by HIP in sand-cast A356 (Al-Si-Mg) and A204 (Al-Cu-Mg) aluminum alloys [22]. However, studies on the application of HIP to Mg alloys were rarely reported. To our knowledge, this may result from ineffectiveness of HIP on commercial Mg alloys, e.g., Mg-9Al-1Zn (AZ91) alloy, which is inherently susceptible to shrinkage because of its considerably wide solidification range of $\widetilde{1}$70 ℃ [2,4].

Recently, Mg-Gd-Y-Zr alloys have been developed as potential candidates for application in aerospace and defence industries [[23], [24], [25]], having a relatively narrow solidification range ($\widetilde{1}$04 ℃) and resultant relatively low tendency of shrinkage porosity formation [26]. In this study, we applied HIP to Mg-6Gd-3Y-0.5 Zr (GW63) alloy to analyze the effectiveness of HIP on Mg alloy with relatively narrow solidification range. Besides, the mechanisms for shrinkage porosity closure and tensile properties improvement were characterized.

2. Experimental procedures

2.1. Material preparation

The investigated Mg-5.87Gd-2.65Y-0.47 Zr (wt.%) alloy, designated as GW63 henceforth, was prepared from pure Mg (99.95 wt.%), Gd (99.5 wt.%), Y (99.5 wt.%), and Mg-30Zr (wt.%) by melting the raw materials in an electric resistance furnace at approximately 780 ℃ under protection with an anti-oxidizing flux. Then, the melt was poured into a sand mold. Experimental samples were cut from symmetrical ingates. One of the ingates was hot isostatically pressed in nitrogen gas with temperature of $\widetilde{4}$80 ℃ and pressure of $\widetilde{1}$00 MPa for 3 h, designated as HIPed sample. Other ingates are designated as non-HIPed samples. Solution treatment was performed at 500 ℃ for 8 h. Then, the homogenized samples were peak-aged at 200 ℃ for 80 h.

2.2. Microstructure observations

For optical microstructure observation, the samples were etched in a 4 vol.% nitric acid alcohol solution after a conventional mechanical grinding and polishing procedure. 3D morphology of shrinkage porosity was characterized by X-ray tomography (XRT) based on the application of Beer-Lambert law [6,27]. Samples for the XRT scanning were cylinders with size of Φ1 mm × 10 mm, and the voxel size was 3 μm. Fractography, element distribution and electron back-scatter diffraction (EBSD) observations were conducted on a JEOL JSM-7001 F field-emission scanning electron microscope (SEM) equipped with energy dispersive X-ray spectroscopy (EDX) and a fully automated EBSD analysis system (Oxford Instruments-HKL Channel 5).

2.3. Tensile tests

Dog-bone-shaped tensile samples with a gauge length of 20 mm, width of 4 mm, and thickness of 2.5 mm were cut from both HIPed and non-HIPed ingates in as-cast condition. For these as-cast tensile samples, uniaxial tensile tests were conducted at room temperature (RT) with strain rate of 1 × 10-3 s-1 on a Sans type tensile testing machine. Tensile bars with a gauge length of 30 mm and diameter of 5 mm were extracted from both HIPed and non-HIPed ingates in cast-T6 condition. For these cast-T6 tensile bars, uniaxial tensile tests were conducted at RT with strain rate of 1 × 10-3 s-1 on a universal testing machine with an extensometer attached to the tensile bars.

3. Results

3.1. Microstructure evolution

Optical microstructures of cast-T6 GW63 alloys are presented in Fig. 1. As can be seen in Fig. 1(a), a few shrinkage porosities (indicated by black arrows) and few second phases (indicated by white arrows) are situated along grain boundaries in non-HIPed samples. As expected, shrinkage porosity is markedly reduced according to optical observation in most HIPed samples (Fig. 1(b)), although high amount of shrinkage porosity is observed in few HIPed samples (Fig. 1(c, d)). Interestingly, an irregular line decorating grain boundaries are clearly observed in Fig. 1(b), which will be discussed later to elucidate its formation and shrinkage porosity closure mechanisms. The irregular lines are designated as shrinkage-porosity-close lines hereafter. Careful examination on HIPed samples with residual shrinkage porosity (Fig. 1(c, d)) reveals that these samples usually have high content of shrinkage porosity. In addition, plastic deformation is observed in Fig. 1(d), i.e., twins in the rectangular area.

Fig. 1.   Optical microstructures of (a) non-HIPed and (b, c, d) HIPed GW63 alloys in cast-T6 condition. Black and white arrows indicate the shrinkage porosities and second phases in Fig. 1a, respectively. Interested microstructures in the rectangles are magnified and embed in Fig. 1(b, d).

Fig. 2 shows the three-dimensional morphology of shrinkage porosities in non-HIPed (Fig. 2(a)) and HIPed (Fig. 2(b)) GW63 alloys in as-cast condition. Obviously, both the morphology and number density of shrinkage porosities are different from each other (Fig. 2(c, d)). As shown in Fig. 2(a), $\widetilde{7}$0 individual shrinkage porosities with equivalent diameter ranging from 4 μm to 78 μm and sphericity ranging from 0.2 to 0.4 are observed in non-HIPed sample. However, only two individual shrinkage porosities with equivalent diameter larger than 100 μm and sphericity larger than 0.5 exist in HIPed sample (Fig. 2(b)). Based on the formation mechanism and morphology (volume size and sphericity), four types of microporosities, including gas-shrinkage pore, gas pore, net-shrinkage and island-shrinkage, were identified in AM60B magnesium alloy [28]. Accordingly, the microporosities observed in non-HIPed sample are characterized as shrinkage porosities (i.e., net-shrinkage and island-shrinkage in Ref [28].), while the microporosities observed in HIPed sample are characterized as gas-shrinkage porosities (i.e., gas-shrinkage pore in Ref [28].). Thereby, HIP is effective in shrinkage porosity reduction, not effective in gas-related microporosities reduction. It should be noted that the main microporosity in GW63 alloy is shrinkage porosity. The existence of low amount of gas-related microporosities is not the main reason for residual microporosity (Fig. 1(c, d)) after HIP process.

Fig. 2.   XRT observation results of (a) non-HIPed and (b) HIPed GW63 alloys in as-cast condition. (c, d) are the partially enlarged view of (a, b), respectively (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article).

3.2. Tensile properties

Fig. 3 depicts the engineering stress‒engineering strain curves of GW63 alloys in various conditions and the results are summarized in Table 1. Comparison between Fig. 3(a) and (b) reveals that higher tensile properties were obtained after HIP process in as-cast condition. The improvement in tensile properties is more evident in cast-T6 condition. As shown in Fig. 3(c) and (d), the average ultimate tensile strength (UTS) and average elongation-to-failure (EL) were increased from 269 MPa to 335 MPa and from 1.21% to 5.29% after HIP process, respectively. Additionally, the reliability and consistency of tensile properties were enhanced after HIP process.

Fig. 3.   Engineering stress-engineering strain curves of GW63 alloys in (a) non-HIPed and as-cast, (b) HIPed and as-cast, (c) non-HIPed and cast-T6 and (d) HIPed and cast-T6 condition. The stresses at which the quasi in-situ tensile test was paused for OM observation are labeled as green stars with stress values in Fig. 3(b) (UTS: ultimate tensile strength, YS: yield strength, EL: elongation-to-failure).

Table 1   Tensile properties of GW63 alloys in various condition.

TemperUTS (MPa)YS (MPa)EL (%)
Non-HIPed + as-cast1471271.03
HIPed + as-cast1851512.5
Non-HIPed + cast-T62692041.21
HIPed + cast-T63352185.29

New window

3.3. Fractography

Fig. 4 shows the fracture morphology and EDX observation results of cast-T6 GW63 alloys with/without HIP process. As can be observed in Fig. 4(a), the fracture microstructure in HIPed sample consists of quasi-cleavages and tearing ridges with homogenous rare-earth (RE) elements distribution (Fig. 4(b)). As expected, Fig. 4(c) exhibits the morphology of shrinkage porosities on the fracture surface of non-HIPed samples. Further inspection of backscatter electron (BSE) image (Fig. 4(d)) reveals non-uniform distribution of RE elements on shrinkage porosity surface, i.e., some shrinkage porosities surface are rich in RE elements, while others are not. Again, the non-uniform distribution of RE elements is verified by EDX analysis in Fig. 4(e) (61.39Mg-20.00Gd-18.61Y, at.%) and Fig. 4(f) (97.92Mg-01.20Gd-00.88Y, at.%).

Fig. 4.   (a) Secondary electron (SE) image and (b) backscatter electron (BSE) image of HIPed sample in cast-T6 condition at the same location. (c) SE image and (d) BSE image of non-HIPed sample in cast-T6 condition at the same location. EDX observation results of shrinkage porosity with (e) high and (f) low content of rare-earth (RE) elements on the surface in non-HIPed and cast-T6 sample.

4. Discussion

4.1. Shrinkage porosity closure

To elucidate the shrinkage porosity closure mechanism, firstly, we need to clarify the formation mechanism of shrinkage-porosity-close lines. The EDX element mappings of shrinkage-porosity-close lines are presented in Fig. 5. Apparently, the shrinkage-porosity-close lines are rich in RE elements (Y and Gd). Here, the RE elements in shrinkage-porosity-close lines are in all likelihood the remains of RE elements on the surface of shrinkage porosity (Fig. 4(d, e)) after shrinkage porosity closure. On the other hand, these RE elements may also result from new precipitates during HIP process.

Fig. 5.   SEM image and EDX element mappings of shrinkage-porosity-close lines in HIPed and as-cast sample (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article).

Further characterization on shrinkage-porosity-close lines was conducted by EBSD and the results are shown in Fig. 6. In the inverse pole figure map (Fig. 6(b)), random grain orientation and clustering of low-angle grain boundaries are observed, which mainly distribute around the shrinkage-porosity-close lines. These low-angle grain boundaries indicate the existence of dislocations and the regions around shrinkage-porosity-close lines may have experienced plastic deformation. Besides, the local misorientation map (Fig. 6(c)) and the line profile of the misorientation angle along the arrow AB (Fig. 6(d)) corroborate the deformation around shrinkage-porosity-close lines. Deformation characteristics were also found in precious researches [29,30], e.g., sub-grain boundaries were found by transmission electron microscope (TEM) in a sand cast unmodified A356-T6 aluminum alloy after HIP process [29]. On the other hand, only few shrinkage-porosity-close lines are observed in cross section of 10 mm × 10 mm, which are apparently less than shrinkage porosity in non-HIPed samples. Hence, shrinkage-porosity-close lines form due to shrinkage porosity closure, rather than new precipitates. Otherwise, deformation will not be concentrated around shrinkage-porosity-close lines and many shrinkage-porosity-close lines will be observed.

Fig. 6.   (a) Optical microstructure and corresponding EBSD observation results of HIPed and as-cast sample containing shrinkage-porosity-close lines: (b) inverse pole figure map, (c) local misorientation map and (d) line profiles of the misorientation angle along the black arrow AB in Fig. 6(b) (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article).

Several mechanisms contributing to densification during HIP were pointed out in Refs. [31,32], including deformation, creep and diffusion. Shrinkage porosity closure occurs in two stages [33]. In the first stage, the closure of shrinkage porosity takes place under an isostatic pressure that is larger than the yield strength (YS) of a material at the HIP temperature. In the second stage, the bonding of the mutually opposing surfaces of the collapsed shrinkage porosity occurs because atoms diffuse in both directions across the interface. Kellett and Lange [34] emphasized that plastic deformation was the dominant mechanism for pore shape change and closure in a two-phase ZrO2/20 vol.% Al2O3 material during HIP and forging. HIP was employed to heal cracks in a Rene88DT superalloy [35], in which the fracture surfaces of the cracks were first mechanically closed by the high temperature creep, then bonded together, and finally diffusion homogenized. In addition, it was reported [30] that pores shrunk via dislocation movement on octahedral glide planes in a single-crystal nickel-base superalloy. The results of modelling (large pore closure rate by plastic flow) further supported the mechanism of plastic pore closure in Ref. [30].

With regard to the formation mechanism of shrinkage-porosity-close lines, considering the twins surrounded by shrinkage porosity (Fig. 1(d)) and the low-angle grain boundaries around the shrinkage-porosity-close lines (Fig. 6(b)), it is speculated that shrinkage porosity close through compatible plastic deformation of its surrounding grains and subsequent diffusion across the bonding interface. Fig. 7 shows the optical microstructure of sample with partially closed shrinkage porosity (Fig. 7(a)) and schematic of shrinkage porosity closure (Fig. 7(b)). For sample with low content of initial shrinkage porosity (region A), the plastic strain for shrinkage porosity closure is small (e.g., Fig. 1(b)). However, for sample with high content of initial shrinkage porosity (region B), the plastic strain for shrinkage porosity closure seems to be too large for Mg alloy to coordinate (e.g., Fig. 1(c, d)). Besides, clustering of shrinkage porosities in region B are probably connected with each other and attached to the surface of sample, i.e., connecting with atmosphere, which makes shrinkage porosity unaffected by HIP [36]. As is the case for Mg alloys, high tendency to form shrinkage porosity [[1], [2], [3], [4]] and low deformation compatibility (insufficient independent slip systems in hexagonal close-packed structure) [37,38] may be two obstacles for the application of HIP.

Fig. 7.   (a) Optical microstructure of sample with partially closed shrinkage porosity and (b) schematic of shrinkage porosity closure.

4.2. Tensile properties improvement

UTS, EL and consistency improvement both in as-cast and cast-T6 condition are mainly because of shrinkage porosity reduction (Figs. 1,2) and subsequent dramatic decrease of stress concentration [39]. In addition, the increase of cohesive strength of grain boundary probably contributes to the improvement of tensile properties [40]. On the other hand, the grain sizes of non-HIPed and HIPed samples were measured to be $\widetilde{9}$3.2 μm and $\widetilde{9}$2.9 μm by a linear intercept method [41], respectively. Almost the same grain size may be related with the retarded diffusion of the alloying elements under high pressure [42]. Therefore, grain size should not be taken into account in tensile properties improvement analysis. Similarly, negligible effect of HIP on grain size was found in sand-cast A356-T6 and A204-T6 aluminum alloys [22].

Since shrinkage-porosity-close lines are rich in RE elements and have large size along grain boundaries, it is necessary to study its effect on tensile properties. Fig. 8 shows the evolution of shrinkage-porosity-close lines during quasi in-situ tensile test in as-cast condition. The stresses at which the quasi in-situ tensile test was paused for OM observation are labeled as green stars with stress values in Fig. 3(b). Apparently, no crack nucleates around shrinkage-porosity-close lines, indicating that no deleterious effect on tensile properties is introduced by shrinkage-porosity-close lines in as-cast condition. As for HIPed and cast-T6 samples with low content of initial shrinkage porosity, the tensile properties are similar with/higher than GW63 alloys prepared by permanent mold casting without initial shrinkage porosity (UTS: 313 MPa, EL: 4.2% in Ref. [43], UTS: 335 MPa, EL: 5.29% in the present study). Thus, shrinkage-porosity-close lines cause no deleterious effect in cast-T6 condition either, indicating that metallurgical bonding is obtained. Similar microstructure to shrinkage-porosity-close lines was observed in an HIPed joint. Nakano et al. [44] investigated the effect of HIPed interface on tensile properties in HIPed 316 L(N)-IG SS and revealed that the HIP process caused no deleterious effects.

Fig. 8.   Evolution of shrinkage-porosity-close lines during quasi in-situ tensile test in as-cast condition. The crack nucleating around second phases is indicated in rectangles. The stresses at which the quasi in-situ tensile test was paused for OM observation are labeled as green stars with stress values in Fig. 3(b).

However, it is noted that cracks nucleated around second phases during quasi in-situ tensile test (indicated in rectangles in Fig. 8), suggesting second phases degraded tensile properties. Therefore, the tensile properties improvement is partially associated with almost no residual second phases in HIPed and cast-T6 samples (Fig. 1(b)). In addition, uniform RE elements distribution and plastic deformation, introduced by HIP, may play an important role in YS enhancement (from 204 MPa to 218 MPa in cast-T6 condition). As shown in Fig. 1(b) and Fig. 4(b), more uniform distribution of RE elements was obtained after HIP process, compared with Fig. 1(a) and Fig. 4(d). Uniform distribution of elements promotes fine and dispersive precipitates which is a main strengthening mechanism in Mg-Gd-Y-Zr alloys [24,45]. Previous reports showed [46,47] that the HIP schemes would influence the microstructure and properties of powder metallurgy Ti2AlNb alloy significantly. Particularly, HIP temperature would affect the volume density and distribution of the porosity. Accordingly, the HIP processing route may affect the shrinkage porosity healing and tensile properties in GW63 alloy, and it is our future work to characterize the influence of HIP schemes.

5. Conclusion

In summary, HIP was firstly applied to a Mg-Gd-Y-Zr alloy with relatively narrow solidification range and resultant low tendency to form shrinkage porosity. HIP was effective in shrinkage porosity reduction in GW63 alloy, although some residual shrinkage porosities were observed in samples with high content of initial shrinkage porosity. The tensile properties of GW63 alloy were enhanced by HIP process both in as-cast and cast-T6 condition. The average UTS, YS and EL were increased from 269 MPa to 335 MPa, from 204 to 218 MPa and from 1.21% to 5.29% in cast-T6 condition, respectively, which was mainly associated with shrinkage porosity reduction and resultant effective RE elements homogenization and aging hardening. During HIP process, the shrinkage porosity was closed by adjacent grain compatible deformation and subsequent diffusion across the bonding interface, and the amount of initial shrinkage porosity was the key factor for shrinkage porosity closure. High susceptibility to shrinkage porosity and low deformation compatibility of Mg alloys may be two obstacles for the application of HIP to Mg alloys.

Acknowledgements

This work was financially supported by the National Science and Technology Major Project of China (No. 2017ZX04014001), the National Key Research and Development Program of China (No. 2016YFB0301104), the National Natural Science Foundation of China (Nos. 51531002, 51301173, 51601193 and 51701218), and the National Basic Research Program of China (No. 2013CB632202).

The authors have declared that no competing interests exist.


/