- Email: [email protected]
Effect of particle size ratio on microstructure and mechanical properties of aluminum matrix composites reinforced with Zr48Cu36Ag8Al8 metallic glass particles
Author’s Accepted Manuscript Effect of particle size ratio on microstructure and mechanical properties of aluminum matrix composites reinforced with Zr48Cu36Ag8Al8 metallic glass particles Tianbing He, Onur Ertuğrul, Nevaf Ciftci, Volker Uhlenwinkel, Kornelius Nielsch, Sergio Scudino www.elsevier.com/locate/msea
PII: DOI: Reference:
S0921-5093(18)31526-0 https://doi.org/10.1016/j.msea.2018.11.007 MSA37127
To appear in: Materials Science & Engineering A Received date: 5 September 2018 Revised date: 30 October 2018 Accepted date: 1 November 2018 Cite this article as: Tianbing He, Onur Ertuğrul, Nevaf Ciftci, Volker Uhlenwinkel, Kornelius Nielsch and Sergio Scudino, Effect of particle size ratio on microstructure and mechanical properties of aluminum matrix composites reinforced with Zr48Cu36Ag8Al8 metallic glass particles, Materials Science & Engineering A, https://doi.org/10.1016/j.msea.2018.11.007 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Effect of particle size ratio on microstructure and mechanical properties of aluminum matrix composites reinforced with Zr48Cu36Ag8Al8 metallic glass particles
Tianbing He a, *, Onur Ertuğrul b, Nevaf Ciftci c, Volker Uhlenwinkel c, d, Kornelius Nielsch e, f, Sergio Scudino a
a
IFW Dresden, Institute for Complex Materials, 01069 Dresden, Germany
b
c
Leibniz Institute for Materials Engineering IWT, 28359 Bremen, Germany
d
e
Izmir Katip Celebi University, Department of Materials Science and Engineering, 35620 Izimir, Turkey
University of Bremen, Faculty of Production Engineering, 28359 Bremen, Germany
IFW Dresden, Institute for Metallic Materials, 01069 Dresden, Germany
f
TU Dresden, Institute of Materials Science, 01062 Dresden, Germany
*
Corresponding author:
[email protected] [email protected]
Abstract: Aluminum matrix composites reinforced with Zr48Cu36Ag8Al8 glassy particles were synthesized by powder metallurgy using reinforcement particles larger than the matrix. The effect of the matrix to
1
reinforcement particle size ratio (PSR) on the microstructure and mechanical properties was studied. The results show that high densification (relative density > 98%) was achieved and the glassy particles retained the amorphous structure in all the composites. Quantitative particle distribution analysis in the three-dimensional space indicated that the homogeneity distribution index decreases with increasing PSR. The findings suggest that a ratio of 1/3 ≤ PSR ≤ 1 can be used to obtain composites with rather homogenous distribution of the reinforcement particles. Both compressive and tensile yield strengths of the composites are not sensitive to the PSR change (in the range of 1/3 – 1/6), whereas the ultimate tensile strength and the ductility are significantly reduced with decreasing PSR; this behavior is accompanied by the change of the fracture mode. The experimental yield strength was found to be consistent with the quantitative strengthening mechanism calculations, and indicated that the reduced matrix ligament size, the thermal mismatch and load bearing are the main strengthening contributions.
Keywords: Metal matrix composites (MMC); Zr-based metallic glass; Particle size ratio; X-ray computed tomography (micro-CT); Strengthening mechanisms
1. Introduction Discontinuously reinforced aluminum matrix composites exhibit isotropic properties, excellent comprehensive performances (low density, high specific strength and stiffness, good fatigue resistance etc.), as well as relatively simple and low-cost fabrication process, which have been used
2
in many industrial fields, such as aerospace, automobile and electronics [1]. A variety of materials have been utilized as reinforcements in the past decades, where ceramic particles (SiC [2], Al2O3 [3]), short carbon fibers [4], and novel nano materials (carbon nanotubes [5], graphene [6]) are among the most frequently reported ones. However, due to the intrinsic properties of the reinforcements and metallic matrix, some problems during the preparation process and service conditions are standing out and gradually becoming a bottle neck for the development and potential applications of this class of composites. For example, the difference in physical and chemical properties between the aluminum matrix and SiC particles or carbon fibers usually result in relatively weak interface bonding [7, 8]. Besides, carbon and aluminum tend to react with each other at elevated temperature to form brittle intermetallics; the formation of large size aluminum carbides not only weakens the reinforcement but also becomes the preferential crack initiation sites under loading [9, 10]. Although some special techniques, such as reinforcement coating [8] and adjustment of the chemical composition of the matrix [9], have been used to overcome this problem, the overall properties of the composites do not show significant improvements [11, 12]. Metallic glasses are an attractive alternative to ceramic and carbonaceous materials as reinforcements for aluminum matrix composites [11, 13, 14]. Metallic glasses have extremely high strength, hardness and elastic strain limit, as well as good corrosion and wear resistance compared with the corresponding crystalline metallic alloys [15], excellent characteristics as reinforcements in composites. Most importantly, metallic glasses manifest much better compatibility with aluminum because of their metallic nature, thus the interface bonding between matrix and reinforcement can be enhanced [12, 16].
3
As for the ceramic particle reinforced aluminum matrix composites, detailed investigations have been done to explore the influence of particle size (or PSR) on the dispersion/distribution of reinforcements, and the mechanical behavior of the composites. Slipenyuk et al. [17] reported that by decreasing the PSR, higher yield and tensile strengths and better material fabricability can be obtained on the condition of a uniform reinforcement distribution. Fathy et al. [18] found that smaller PSR resulted in more uniform distribution of the SiC particles in the matrix and better tensile properties of the composites, while the agglomeration and voids increased when PSR increases. Since the majority of the studies is limited to PSR > 1, Diler et al. [19] used the size of the SiC particles larger than the aluminum powder; the results showed better densification and improved tribological properties with increasing PSR when the reinforcement volume fraction is below 17.5 %. It can be seen that the PSR is an important parameter determining the microstructure and properties of the composites to a large degree. On the other hand, the studies on aluminum matrix composites reinforced with metallic glass particles are mainly focused on the effect of different classes of reinforcements (Ni-, Zr-, Fe-, Mg-based [14, 20-24]). In addition, the previous works about the strength and plasticity of this new kind of composites are mostly under compressive load. In this study, aluminum matrix composites reinforced with a relatively high volume fraction (20 vol.%) of Zr48Cu36Ag8Al8 metallic glass particles and PSR < 1 are prepared via powder metallurgy. The effect of the PSR on the reinforcement distribution in the three-dimensional space is evaluated by quantitative methods, and the mechanical properties of the composites under both compressive and tensile loads as well as the strengthening mechanisms are investigated.
4
2. Experimental 2.1 Materials and preparation process Pure aluminum powder (purity ≥ 99.5 wt. %) with mean size of 13 μm (Fig. 1) was used as matrix. The Zr48Cu36Ag8Al8 (at.%) metallic glass particulate reinforcement was synthesized by gas atomization using free-fall atomization and subsequently sieved into 20-45 μm, 45-63 μm and 63-90 μm particle class sizes. The mean particle diameter (D50) of 35 μm, 54 μm and 75 μm (corresponding to a PSR of ~1/3, 1/4 and 1/6, respectively) for different particle class sizes were obtained via laser diffraction (Marlvern Matersizer 2000), as shown in Fig. 2. Powder mixtures consisting of the aluminum matrix with 20 vol.% of glassy particles with different sizes were prepared by using a Retsch PM 400 planetary ball mill equipped with hardened steel balls and vials. No process control agent was used. The composite powders were milled at room temperature for 2 h at 100 rpm with a ball to powder weight ratio of 10:1. To avoid or minimize possible atmosphere contamination during milling, the vials were sealed in a Braun MB 150B-G glove box under a purified argon atmosphere (less than 0.5 ppm O2 and H2O). Consolidation of the as-blended powder mixtures was carried out by uni-axial hot pressing under argon atmosphere at 673 K and 640 MPa with a holding time of 10 min. After hot pressing, some of the hot-pressed billets were further hot extruded with extrusion ratio of 9:1 at 673 K. Pure Al bulk specimens were produced by hot pressing or hot extrusion under the same condition as a reference group.
5
Fig. 1 As-received pure aluminum powder: (a) SEM micrograph and (b) particle size distribution.
Fig. 2 Zr48Cu36Ag8Al8 metallic glass powders: SEM micrographs of the glassy particles with average size (a) 35, (b) 54 and (c) 75 μm; (d) particle size distributions.
2.2 Composites characterization The phases of the samples were characterized by X-ray diffraction (XRD) using a Philips PW 1050 diffractometer with Co Kα radiation source operating at 40 kV and 40 mA. The microstructures 6
were investigated by scanning electron microscopy (SEM) using a Zeiss Gemini 1530 field emission microscope equipped with an energy dispersive X-ray spectroscope (EDX). The relative density of the samples was measured using the Archimedeans principle. The X-ray computed tomography (micro-CT) analysis was performed with a GE phoenix Nanotom device, using 110 kV voltage, 100 μA current and 750 ms detector timing. The obtained radiographic images with a spatial resolution of 3.64 μm were reconstructed with the VG Studio Max 2.2 software. Quantitative particle distribution analysis was carried out by dividing the image into 9 grids. The particle concentration (area) in each grid, as well as the standard deviation of the concentration per grid was calculated. Ten images (with sample area represents 2 × 2 mm) acquired from different cross sections (interval of 0.5 mm with a sample length of 5 mm) of the reconstructed CT 3D model were used. The distribution index D was evaluated by D=1−S
(1)
where S is the standard deviation of the particle concentration per grid. A higher distribution index indicates a more homogeneous microstructure of the composites. Compression tests were conducted on cylindrical specimens with 6 mm length and 3 mm diameter. Both ends of the specimens were polished to make them parallel to each other before the tests. Tensile samples with dog-bone cylindrical shape (Fig. 3) were prepared from the hot extruded composites. The mechanical tests were performed by using an Instron 5862 testing facility at room temperature with a crosshead speed of 6 × 10-4 mm/s for compression and 1 × 10-3 mm/s for tension. The compression tests were terminated after the specimen strain reached 20 %.
7
Fig. 3 Schematic representation of the tensile test samples.
3. Results and discussion 3.1 Microstructures Fig. 4 shows the morphology of the as-blended composite powder mixtures with different PSR. The glassy particles are not fractured during the ball milling process and they retain the original as-atomized spherical morphology. Some of the aluminum particles are cold welded with each other and turned into relatively large particle agglomerates. For most of the glassy particles, their surface is partially covered by aluminum, a beneficial aspect for the following consolidation process, where the soft aluminum can act as a binder between the hard-to-sinter metallic glass particles.
8
Fig. 4 Morphology of the as-blended composites powders with PSR (a) 1/3, (b) 1/4, and (c) 1/6; (d) high magnification micrograph corresponding to the white box in (b).
Fig. 5 shows the XRD patterns of the composites after the different preparation stages. The pattern of the as-atomized glassy particles is also shown. The Zr48Cu36Ag8Al8 powder displays the broad diffraction peak (at about 2θ = 45°) characteristic of amorphous materials without any detectable crystalline Bragg peak. The as-blended powders and consolidated composites have similar patterns, consisting of the sharp Bragg peaks corresponding to fcc aluminum along with the amorphous broad maximum of the metallic glass phase. The absence of additional phases indicates that crystallization of the glassy particles or the reaction between glassy particles and aluminum matrix does not occur during the fabrication process.
9
Fig. 5 XRD patterns of the as-atomized glassy particles, as-blended powder mixtures and consolidated composites.
Fig. 6 shows the relative density of the hot pressed unreinforced pure aluminum and composites with different PSR. A relative density of more than 98 % was achieved, indicating a good densification of the materials. Little difference of densification is found when the PSR changes. Fig. 7 presents the microstructure of the hot pressed composites. No porosity can be observed from the images, further corroborating the high densification of the materials. The distribution of the glassy particles is rather uniform irrespective of the PSR. Fig. 7(d) shows the interface between the aluminum matrix and a metallic glass particle. The interface is sharp without any visible reaction products, further supporting that the glassy reinforcement retained the amorphous structure, in agreement with the XRD results.
10
Fig. 6 Relative density of the hot pressed unreinforced pure Al and composites with different PSR.
Fig. 7 SEM micrographs of hot pressed composites with different PSR: (a) 1/3, (b) 1/4, and (c) 1/6; (d) interface between a glassy particle and the aluminum matrix.
The densification capability of the powders is an important factor in powder metallurgy, which determines the final physical and mechanical properties of materials. The densification mechanisms
11
for metal matrix composites reinforced with hard particles can be mainly classified as follows: (1) sliding and rearrangement of matrix powders and reinforcing particles, (2) plastic deformation of matrix powders, (3) localized particles fragmentation and (4) elastic compaction of the bulk nearly nonporous composites [19, 25]. At the initial compaction stage, the arrangement of matrix powders and reinforcement particles are the dominant densification mechanism. With increasing the compaction pressure, the movement of the matrix powders and reinforcement particles is restricted, the compaction energy is spent on the deformation of matrix powders and the friction between powders and particles. The aforementioned densification mechanisms are strongly affected by the volume fraction of the reinforcement particles, and the size and shape of particles and matrix powders. In the present study, the PSR is less than 1, and most of the glassy particles in the composites are isolated from each other (Fig. 7(a) – (c)). This means that the matrix powders can easily be located in the voids between the reinforcement particles, which is responsible for the high densification of the composites. A more realistic representation of the distribution of the glassy particles within the Al matrix can be obtained by X-ray microtomography (Fig. 8). The 3D reconstructions clearly show the difference in the distribution of the glassy reinforcements when the PSR changes. For the hot pressed composite with PSR = 1/3 (Fig. 8(a)), the tomography reveals that the particles are homogeneously distributed in 3D, giving rise to a high distribution index of 96.3 %. With the increase of the PSR to 1/4 and 1/6, the spatial distribution of reinforcements becomes progressively nonuniform (Figs. 8(b) and (c)) and the distribution index decreases to 88.1% (Fig. 8(d)). After hot extrusion, a slightly more homogeneous distribution of the reinforcements is exhibited by the composites compared with their
12
hot pressed counterparts (Fig. 8(d)). In addition, the glassy particle distribution in the hot extruded composites displays the same trend as for the hot pressed ones, both decreasing with the increase of PSR.
Fig. 8 3D views of the glassy particle distribution in hot pressed and hot extruded composites obtained by X-ray microtomography for the PSR: (a) 1/3, (b) 1/4, and (c) 1/6; (d) corresponding spatial distribution index.
The PSR plays an important role in obtaining a homogeneous microstructure of the composites. When the size of the matrix powders is larger than the reinforcement particles (PSR > 1), the homogeneity of the particles distribution in the composites decreases with increasing PSR; a behavior that can be attributed to the geometry dependent clustering [18, 26-28]. The increase of 13
PSR leads to a reduction in the combined surface area of the matrix powders, which becomes insufficient for a uniform arrangement of the reinforcement particles. As a result, the reinforcement particles cluster in the interstices of the matrix powders, leading to their inhomogeneous distribution. The corresponding microstructure of the composites is usually called as “necklace structure” [18, 26]. In this study, the size of aluminum powders is less than the glassy particles (PSR <1); although the bonding surface area of the matrix powders changes when PSR decreases, this seems to have only a limited influence on the variation of the particle distribution. As can be seen in Figs. 7(a) - (c), most of the glassy particles are isolated from each other, and the inhomogeneity mainly results from the nonuniform distribution of the glassy particles in the three-dimensional space (particle rich or deplete regions, Fig. 8(a) – (c)) instead of particle agglomerations in the whole composites. The mixing process of the composite powders can be seen as a circulation flow of the matrix powders and reinforcement particles under an external force (i.e. ball milling), and the flowability of each powder or particle is an important factor affecting this process. The particle shape, the particle size distribution, powder density and humidity determine the powder flowability [29]. Generally, spherical and heavier particles with larger size display better flowability compared with irregular small and light powders. Therefore, the glassy particles (ρ = 7.11 g/cm3) are expected to have superior flowability than the aluminum powders (ρ = 2.70 g/cm3) in the present case. With increasing the size of the glassy particles (i.e. PSR decreases), the discrepancy of flowability between glassy particles and aluminum powder becomes much more remarkable. In other words, during the powder mixing process as well as the following procedures which may involve stirring or blending (e.g. powders transfer), the glassy particles may tend to flow together in the composite powders with
14
decreased PSR, thus particle-rich (or depleted) regions appear. Slipenyuk et al. [30] reported that a PSR of around 3 may be low enough to obtain a composite material with relatively homogenous distribution of reinforcements, provided that the content does not exceed 15 vol. %. Here, the composites show a rather homogeneous microstructure when the PSR is 1/3. This suggests that it may be possible to draw an empirical conclusion that with 1/3 ≤ PSR ≤ 3, a relatively homogenous distribution of reinforcements in the composites can be achieved when the volume fraction of the particles is no more than 20 %, and the closer that PSR is to 1, the more homogeneous is the microstructure. During the extrusion process, the distribution of the reinforcement becomes more homogeneous: the large deformation results in the flow of the reinforcements within the matrix, and the movement of the particles from the rich to depleted zones, promoting the homogenization of the composites microstructure. However, since there is a dramatic rise of the hydrostatic stress in the particle rich regions, which constrains the lateral plastic strain of the specimen [31], the improvement of the homogeneity is rather limited and the particle rich regions are only elongated in the extrusion direction, forming particle-rich bands and a fiber-like texture.
3.2 Mechanical properties Fig. 9 (a) shows the compressive properties of the hot pressed pure aluminum matrix and composites. The compressive yield strength (YS) of the pure matrix and composites with PSR = 1/3 are 90 MPa and 125 MPa, respectively, corresponding to an improvement of about 39 %. When the PSR decreases, the compressive yield strength of the composites does not show significant changes.
15
Fig. 9 (b) shows the compression stress – strain curves of the composite with PSR = 1/3 after hot pressing and hot extrusion. The hot extruded composite exhibits higher yield strength compared to the composite that was only hot pressed. This can be ascribed to the generation of the fiber-like texture during extrusion. The tensile properties of the hot extruded composites with different PSR are shown in Fig. 10. The three composites exhibit tensile yield strength on the same level (between 155 and 160 MPa), while the ultimate tensile strength (UTS) and total extension at fracture present large difference. The composite with PSR = 1/3 has UTS and strain at fracture of 235 MPa and 14 %, respectively, which sharply drop to 203 MPa and only 3 % for the PSR = 1/6.
Fig. 9 Compression properties of: (a) hot pressed composites with different PSR and (b) the hot pressed and hot extruded composites with PSR = 1/3.
16
Fig. 10 Tensile properties of the hot extruded composites: (a) stress - strain curves and (b) corresponding mechanical data as a function of the PSR.
Fig. 11 shows the fracture surface after tensile tests for the composites with PSR = 1/3 and 1/6. The composite with PSR = 1/3 shows a rough fracture surface and a large number of dimples (Fig. 11(a)), indicating a ductile fracture mode operating in the composites. When the PSR is 1/6 (Fig. 11(c)), the morphology turns into relatively flat, comprising several fractured (marked by arrows) and many decohered glassy particles surrounded by ductile regions (termed as “tear ridges”). The larger size of the particles results in easier initiation of cracks since voids nucleate at the matrix/particle interface when dislocations are pushed to the undeformable particle under straining [32]. Secondly, the largest possible defect within a particle is limited by its size and there is a higher statistical probability of there being a critical crack in larger particles, i.e. larger particles are more likely to fracture than smaller ones [31]. The voids, critical defects coalesce with each other when the applied load increases, and final fracture is achieved by rapid crack propagation through the matrix between the reinforcement particles. Moreover, the reduced PSR is detrimental for obtaining the homogeneous spatial distribution of reinforcements, which leads to stress concentration under deformation, further promoting the growth of cracks and fracture of the composites. It should be noted that the matrix dimples present on the surface of the glassy particles (Fig. 11(b) and (d)) indicate a good interface bonding between aluminum and the metallic glass.
17
Fig. 11 Morphology of the fracture surface for the composites with (a) and (b) PSR =1/3, (c) and (d) PSR = 1/6.
3.3 Strengthening mechanisms The strengthening effect of particles in metal matrix composites can be mainly divided into two categories: direct and indirect strengthening [21, 33]. These two strengthening factors are interdependent and act simultaneously, resulting in a combined effect [33]. Direct strengthening considers the load-bearing effect of the reinforcements, in which the load is transferred from the soft matrix to the hard reinforcements via the matrix-particle interface. According to the modified shear lag model [34], the yield strength of the composites reinforced with spherical particles can be described as follows: σ𝑐𝑦 = 𝜎𝑚𝑦 (1 +
𝑉𝑝 2
18
)
(2)
where 𝜎𝑐𝑦 and 𝜎𝑚𝑦 are the yield strength of the composites and matrix respectively, and 𝑉𝑝 is the volume fraction of the reinforcement particles. The indirect strengthening is related to the increase of dislocation density in the matrix due to the introduction of the reinforcements, and the yield strength of the matrix 𝜎𝑚𝑦 can be expressed as [21, 33, 35] 𝜎𝑚𝑦 = 𝜎0 + Δ𝜎𝑑𝑖𝑠 = 𝜎0 + Δ𝜎𝑜𝑟 + Δ𝜎𝑡ℎ𝑒 + Δ𝜎𝑔𝑒𝑜 + Δ𝜎𝑠
(3)
where 𝜎0 is the yield strength of the unreinforced matrix, Δ𝜎𝑑𝑖𝑠 is the yield strength increase caused by the dislocation strengthening, Δ𝜎𝑜𝑟 is the Orowan stress [14, 32, 36, 37], Δ𝜎𝑡ℎ𝑒 is the statistically stored dislocation caused by the thermal mismatch between the matrix and reinforcements during cooling [14, 38-40], Δ𝜎𝑔𝑒𝑜 is the geometrically necessary dislocations [41, 42] and Δ𝜎𝑠 is the strength increase attributed to the decrease of matrix ligament size [14, 21, 32, 33, 36, 37, 43]. The stress increment due to geometrically necessary dislocation can also be written as [44-46] ∆σ𝑔𝑒𝑜 = ∆𝜎𝑖𝑠𝑜 + ∆𝜎𝑘𝑖𝑛
(4)
where ∆𝜎𝑖𝑠𝑜 and ∆𝜎𝑘𝑖𝑛 are the isotropic and kinematic strain gradient contribution to the flow stress respectively. The different contributions to the improvement of the yield strength resulting from dislocation strengthening are summarized in Tab. 1. The strengthening effect caused by the reduced matrix ligament size and the thermal mismatch between matrix and glassy particles give the largest contributions to the increment of yield strength. Tab. 2 lists the calculated (according to Eqs. (2) and (3)) and experimental yield strength of the composites with different PSR. The results of the
19
predicted yield strength are slightly higher than the ones derived experimentally. The reasons for this behavior can be ascribed to: (a) the different dislocation strengthening mechanisms interact with each other and the contributions to the matrix yield strength may not be simply additive (Eq. (3)) [35]; (b) the particle size distribution, nonideal periodical configuration of the particles [36, 43] etc. are not taken into consideration in the mechanical models. Therefore, the present calculations should be viewed as an upper limit with respect to the strength prediction. Tab. 1 Calculated increment of yield strength from dislocation strengthening
PSR 1/3 1/4 1/6
Δσor (MPa) 0.3 0.2 0.1
Δσgeo (MPa) Δσiso Δσkin 0.6 1.4 0.5 1.1 0.4 1.0
Δσthe (MPa) 8.5 7.6 7.0
Δσs (MPa) 27.5 22.1 18.8
Δσdis (MPa) 38.3 31.6 27.3
Tab. 2 The estimated and experimental yield strength of composites with different PSR
PSR 1/3 1/4 1/6
σcy1 a (MPa) calculated experimental 141 125±8 134 122±2 129 116±4
σcy2 b (MPa) calculated experimental 174 159±3 167 152±10 162 157±1
σcy1 a Compression yield strength, hot pressed composites, σcy2 b Tensile yield strength, hot pressed and hot extruded composites.
4. Conclusions Aluminum matrix composites reinforced with different sizes of Zr48Cu36Ag8Al8 metallic glass particles (PSR < 1) were synthesized by powder metallurgy. All the composites show high densification (relative density > 98%) since the soft matrix powder can easily fill the interstices 20
between the reinforcement particles. The glassy particles retained their amorphous structure during the composite fabrication process. Quantitative particle distribution analysis in the three-dimensional space demonstrates that the distribution index decreases with reducing PSR. The comparison of the present results with those reported in literature suggests that a range of 1/3 ≤ PSR ≤ 3 may be necessary for attaining a relatively homogenous distribution of reinforcements; furthermore, when the volume fraction of reinforcement does not exceed 20 %, the closer that PSR is to 1, the more homogeneous is the microstructure. Hot extrusion can further improve the particle spatial distribution. The yield strength of the composites did not show significant variations within the PSR range analyzed here (1/3 – 1/6) either under compressive or tensile loading. However, the ultimate tensile strength and the corresponding ductility are reduced significantly, from 235 MPa and 14 % for PSR = 1/3 to 203 MPa and only 3 % for PSR = 1/6. At the same time, the ductile dimple fracture mode turned into mixed fracture of interface debonding and crack coalescence in the matrix between the particles. Quantitative strengthening mechanisms calculations indicated that the main strengthening contributions in the present composites come from the reduced matrix ligament size of the matrix, the thermal mismatch between matrix and glassy particles and load-bearing of the reinforcements. The calculated yield strength of the composites is consistent with experimental results.
Acknowledgements The authors thank B. Opitz, A. Funk and H. Merker for technical support, and R.N. Shahid and P. Wang for valuable suggestions. T. He gratefully acknowledges the financial support of the China Scholarship Council (CSC).
21
References: [1] C. William, J. Harrigan, Commercial processing of metal matrix composites, Mater. Sci. Eng. A. 244 (1998) 75-79. [2] Y. Cui, L.F. Wang, J.Y. Ren, Multi-functional SiC/Al Composites for Aerospace Applications, Chin. J. Aeronaut. 21 (2008) 578-584. [3] M. Rahimian, N. Parvin, N. Ehsani, Investigation of particle size and amount of alumina on microstructure and mechanical properties of Al matrix composite made by powder metallurgy, Mater. Sci. Eng. A. 527(4) (2010) 1031-1038. [4] B. Bhav Singh, M. Balasubramanian, Processing and properties of copper-coated carbon fibre reinforced aluminium alloy composites, J. Mater. Process. Technol. 209(4) (2009) 2104-2110. [5] T. He, X. He, P. Tang, D. Chu, X. Wang, P. Li, The use of cryogenic milling to prepare high performance Al2009 matrix composites with dispersive carbon nanotubes, Mater. Des. 114 (2017) 373-382. [6] J.L. Li, Y.C. Xiong, X.D. Wang, S.J. Yan, C. Yang, W.W. He, J.Z. Chen, S.Q. Wang, X.Y. Zhang, S.L. Dai, Microstructure and tensile properties of bulk nanostructured aluminum/graphene composites prepared via cryomilling, Mater. Sci. Eng. A. 626 (2015) 400-405. [7] P. Jin, B. Xiao, Q. Wang, Z. Ma, Y. Liu, S. Li, Effect of hot extrusion on interfacial microstructure and tensile properties of SiCp/2009Al composites fabricated at different hot pressing temperatures, J. Mater. Sci. Technol. 27(6) (2011) 518-524.
22
[8] Y. Tang, Y. Deng, K. Zhang, L. Liu, Y. Wu, W. Hu, Improvement of interface between Al and short carbon fibers by α-Al2O3 coatings deposited by sol-gel technology, Ceram. Int. 34(7) (2008) 1787-1790. [9] C. Wang, G. Chen, X. Wang, Y. Zhang, W. Yang, G. Wu, Effect of Mg Content on the Thermodynamics of Interface Reaction in Cf/Al Composite, Metall. Mater. Trans. A 43(7) (2012) 2514-2519. [10] Y. Tang, L. Liu, W. Li, B. Shen, W. Hu, Interface characteristics and mechanical properties of short carbon fibers/Al composites with different coatings, Appl. Surf. Sci. 255(8) (2009) 4393-4400. [11] P. Yu, K.B. Kim, J. Das, F. Baier, W. Xu, J. Eckert, Fabrication and mechanical properties of Ni-Nb metallic glass particle-reinforced Al-based metal matrix composite, Scr. Mater. 54(8) (2006) 1445-1450. [12] P. Yu, L.C. Zhang, W.Y. Zhang, J. Das, K.B. Kim, J. Eckert, Interfacial reaction during the fabrication of Ni60Nb40 metallic glass particles-reinforced Al based MMCs, Mater. Sci. Eng. A. 444(1-2) (2007) 206-213. [13] M.H. Lee, J.H. Kim, J.S. Park, J.C. Kim, W.T. Kim, D.H. Kim, Fabrication of Ni-Nb-Ta metallic glass reinforced Al-based alloy matrix composites by infiltration casting process, Scr. Mater. 50(11) (2004) 1367-1371. [14] S. Scudino, G. Liu, K.G. Prashanth, B. Bartusch, K.B. Surreddi, B.S. Murty, J. Eckert, Mechanical properties of Al-based metal matrix composites reinforced with Zr-based glassy particles produced by powder metallurgy, Acta Mater. 57(6) (2009) 2029-2039.
23
[15] M.M. Khan, A. Nemati, Z.U. Rahman, U.H. Shah, H. Asgar, W. Haider, Recent advancements in bulk metallic glasses and their applications: a review, Crit. Rev. Solid State Mater. Sci.
(2017)
1-36. [16] P. Gupta, S. Pal, N. Yedla, Molecular dynamics based cohesive zone modeling of Al (metal)-Cu50Zr50 (metallic glass) interfacial mechanical behavior and investigation of dissipative mechanisms, Mater. Des. 105 (2016) 41-50. [17] A. Slipenyuk, V. Kuprin, Y. Milman, V. Goncharuk, J. Eckert, Properties of P/M processed particle reinforced metal matrix composites specified by reinforcement concentration and matrix-to-reinforcement particle size ratio, Acta Mater. 54(1) (2006) 157-166. [18] A. Fathy, A. Sadoun, M. Abdelhameed, Effect of matrix/reinforcement particle size ratio (PSR) on the mechanical properties of extruded Al-SiC composites, Int. J. Adv. Manuf. Technol. 73(5) (2014) 1049-1056. [19] E.A. Diler, A. Ghiami, R. Ipek, Effect of high ratio of reinforcement particle size to matrix powder size and volume fraction on microstructure, densification and tribological properties of SiCp reinforced metal matrix composites manufactured via hot pressing method, Int. J. Refract. Met. Hard Mater. 52 (2015) 183-194. [20] D. Markó, K.G. Prashanth, S. Scudino, Z. Wang, N. Ellendt, V. Uhlenwinkel, J. Eckert, Al-based metal matrix composites reinforced with Fe49.9Co35.1Nb7.7B4.5Si2.8 glassy powder: Mechanical behavior under tensile loading, J. Alloys Compd. 615 (2014) S382-S385.
24
[21] Z. Wang, J. Tan, B.A. Sun, S. Scudino, K.G. Prashanth, W.W. Zhang, Y.Y. Li, J. Eckert, Fabrication and mechanical properties of Al-based metal matrix composites reinforced with Mg65Cu20Zn5Y10 metallic glass particles, Mater. Sci. Eng. A. 600 (2014) 53-58. [22] Q. Yang, Y. Zhang, H. Zhang, R. Zheng, W. Xiao, C. Ma, Fabrication of Al-based composites reinforced with in situ devitrified Al84Ni8.4Y4.8La1.8Co1 particles by hot pressing consolidation, J. Alloys Compd. 648(Supplement C) (2015) 382-388. [23] Ö. Balcı, K.G. Prashanth, S. Scudino, M. Somer, J. Eckert, Powder metallurgy of Al-based composites reinforced with Fe-based glassy particles: Effect of microstructural modification, Part. Sci. Technol. (2017) 1-6. [24] F. Bahrami, R. Amini, A.H. Taghvaei, Microstructure and corrosion behavior of electrodeposited Ni-based nanocomposite coatings reinforced with Ni60Cr10Ta10P16B4 metallic glass particles, J. Alloys Compd. 714 (2017) 530-536. [25] P.J. Denny, Compaction equations: a comparison of the Heckel and Kawakita equations, Powder Technol. 127(2) (2002) 162-172. [26] V.V.B. Prasad, B.V.R. Bhat, Y.R. Mahajan, P. Ramakrishnan, Structure–property correlation in discontinuously reinforced aluminium matrix composites as a function of relative particle size ratio, Mater. Sci. Eng. A. 337(1) (2002) 179-186. [27] C. Wu, K. Ma, J. Wu, P. Fang, G. Luo, F. Chen, Q. Shen, L. Zhang, J.M. Schoenung, E.J. Lavernia, Influence of particle size and spatial distribution of B4C reinforcement on the microstructure and mechanical behavior of precipitation strengthened Al alloy matrix composites, Mater. Sci. Eng. A. 675 (2016) 421-430.
25
[28] O. El-Kady, A. Fathy, Effect of SiC particle size on the physical and mechanical properties of extruded Al matrix nanocomposites, Mater. Des. 54 (2014) 348-353. [29] A.B. Spierings, M. Voegtlin, T. Bauer, K. Wegener, Powder flowability characterisation methodology for powder-bed-based metal additive manufacturing, Prog. Addit. Manuf. 1(1) (2016) 9-20. [30] A. Slipenyuk, V. Kuprin, Y. Milman, J.E. Spowart, D.B. Miracle, The effect of matrix to reinforcement particle size ratio (PSR) on the microstructure and mechanical properties of a P/M processed AlCuMn/SiCp MMC, Mater. Sci. Eng. A. 381(1) (2004) 165-170. [31] P.B. Prangnell, S.J. Barnes, S.M. Roberts, P.J. Withers, The effect of particle distribution on damage formation in particulate reinforced metal matrix composites deformed in compression, Mater. Sci. Eng. A. 220(1) (1996) 41-56. [32] K.K. Chawla, M.A. Meyers, Mechanical behavior of materials, Prentice Hall1999. [33] S. Scudino, G. Liu, M. Sakaliyska, K.B. Surreddi, J. Eckert, Powder metallurgy of Al-based metal matrix composites reinforced with β-Al3Mg2 intermetallic particles: Analysis and modeling of mechanical properties, Acta Mater. 57(15) (2009) 4529-4538. [34] V.C. Nardone, K.M. Prewo, On the strength of discontinuous silicon carbide reinforced aluminum composites, Scr. Metall. 20(1) (1986) 43-48. [35] H. Sekine, R. Chen, A combined microstructure strengthening analysis of SiCp/Al metal matrix composites, Composites. 26(3) (1995) 183-188. [36] M.J. Tan, X. Zhang, Powder metal matrix composites: selection and processing, Mater. Sci. Eng. A. 244(1) (1998) 80-85.
26
[37] W.S. Miller, F.J. Humphreys, Strengthening mechanisms in particulate metal matrix composites, Scr. Metall. Mater. 25(1) (1991) 33-38. [38] R.J. Arsenault, N. Shi, Dislocation generation due to differences between the coefficients of thermal expansion, Mater. Sci. Eng. 81 (1986) 175-187. [39] N. Hansen, The effect of grain size and strain on the tensile flow stress of aluminium at room temperature, Acta Metall. 25(8) (1977) 863-869. [40] Y. Zhang, N. Mattern, J. Eckert, Atomic structure and transport properties of Cu50Zr45Al5 metallic liquids and glasses: Molecular dynamics simulations, J. Appl. Phys. 110(9) (2011) 093506. [41] M.F. Ashby, The deformation of plastically non-homogeneous materials, Philos. Mag. 21(170) (1970) 399-424. [42] N.A. Fleck, G.M. Muller, M.F. Ashby, J.W. Hutchinson, Strain gradient plasticity: Theory and experiment, Acta Metall. Mater. 42(2) (1994) 475-487. [43] T.W. Gustafson, P.C. Panda, G. Song, R. Raj, Influence of microstructural scale on plastic flow behavior of metal matrix composites, Acta Mater. 45(4) (1997) 1633-1643. [44] C.W. Nan, D.R. Clarke, The influence of particle size and particle fracture on the elastic/plastic deformation of metal matrix composites, Acta Mater. 44(9) (1996) 3801-3811. [45] L.M. Brown, W.M. Stobbs, The work-hardening of copper-silica v. equilibrium plastic relaxation by secondary dislocations, Philos. Mag. 34(3) (1976) 351-372. [46] L.M. Brown, Precipitation and Dispersion Hardening, Strength of Metals and Alloys, Pergamon1979, pp. 1551-1571.
27
PII: DOI: Reference:
S0921-5093(18)31526-0 https://doi.org/10.1016/j.msea.2018.11.007 MSA37127
To appear in: Materials Science & Engineering A Received date: 5 September 2018 Revised date: 30 October 2018 Accepted date: 1 November 2018 Cite this article as: Tianbing He, Onur Ertuğrul, Nevaf Ciftci, Volker Uhlenwinkel, Kornelius Nielsch and Sergio Scudino, Effect of particle size ratio on microstructure and mechanical properties of aluminum matrix composites reinforced with Zr48Cu36Ag8Al8 metallic glass particles, Materials Science & Engineering A, https://doi.org/10.1016/j.msea.2018.11.007 This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting galley proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Effect of particle size ratio on microstructure and mechanical properties of aluminum matrix composites reinforced with Zr48Cu36Ag8Al8 metallic glass particles
Tianbing He a, *, Onur Ertuğrul b, Nevaf Ciftci c, Volker Uhlenwinkel c, d, Kornelius Nielsch e, f, Sergio Scudino a
a
IFW Dresden, Institute for Complex Materials, 01069 Dresden, Germany
b
c
Leibniz Institute for Materials Engineering IWT, 28359 Bremen, Germany
d
e
Izmir Katip Celebi University, Department of Materials Science and Engineering, 35620 Izimir, Turkey
University of Bremen, Faculty of Production Engineering, 28359 Bremen, Germany
IFW Dresden, Institute for Metallic Materials, 01069 Dresden, Germany
f
TU Dresden, Institute of Materials Science, 01062 Dresden, Germany
*
Corresponding author:
[email protected] [email protected]
Abstract: Aluminum matrix composites reinforced with Zr48Cu36Ag8Al8 glassy particles were synthesized by powder metallurgy using reinforcement particles larger than the matrix. The effect of the matrix to
1
reinforcement particle size ratio (PSR) on the microstructure and mechanical properties was studied. The results show that high densification (relative density > 98%) was achieved and the glassy particles retained the amorphous structure in all the composites. Quantitative particle distribution analysis in the three-dimensional space indicated that the homogeneity distribution index decreases with increasing PSR. The findings suggest that a ratio of 1/3 ≤ PSR ≤ 1 can be used to obtain composites with rather homogenous distribution of the reinforcement particles. Both compressive and tensile yield strengths of the composites are not sensitive to the PSR change (in the range of 1/3 – 1/6), whereas the ultimate tensile strength and the ductility are significantly reduced with decreasing PSR; this behavior is accompanied by the change of the fracture mode. The experimental yield strength was found to be consistent with the quantitative strengthening mechanism calculations, and indicated that the reduced matrix ligament size, the thermal mismatch and load bearing are the main strengthening contributions.
Keywords: Metal matrix composites (MMC); Zr-based metallic glass; Particle size ratio; X-ray computed tomography (micro-CT); Strengthening mechanisms
1. Introduction Discontinuously reinforced aluminum matrix composites exhibit isotropic properties, excellent comprehensive performances (low density, high specific strength and stiffness, good fatigue resistance etc.), as well as relatively simple and low-cost fabrication process, which have been used
2
in many industrial fields, such as aerospace, automobile and electronics [1]. A variety of materials have been utilized as reinforcements in the past decades, where ceramic particles (SiC [2], Al2O3 [3]), short carbon fibers [4], and novel nano materials (carbon nanotubes [5], graphene [6]) are among the most frequently reported ones. However, due to the intrinsic properties of the reinforcements and metallic matrix, some problems during the preparation process and service conditions are standing out and gradually becoming a bottle neck for the development and potential applications of this class of composites. For example, the difference in physical and chemical properties between the aluminum matrix and SiC particles or carbon fibers usually result in relatively weak interface bonding [7, 8]. Besides, carbon and aluminum tend to react with each other at elevated temperature to form brittle intermetallics; the formation of large size aluminum carbides not only weakens the reinforcement but also becomes the preferential crack initiation sites under loading [9, 10]. Although some special techniques, such as reinforcement coating [8] and adjustment of the chemical composition of the matrix [9], have been used to overcome this problem, the overall properties of the composites do not show significant improvements [11, 12]. Metallic glasses are an attractive alternative to ceramic and carbonaceous materials as reinforcements for aluminum matrix composites [11, 13, 14]. Metallic glasses have extremely high strength, hardness and elastic strain limit, as well as good corrosion and wear resistance compared with the corresponding crystalline metallic alloys [15], excellent characteristics as reinforcements in composites. Most importantly, metallic glasses manifest much better compatibility with aluminum because of their metallic nature, thus the interface bonding between matrix and reinforcement can be enhanced [12, 16].
3
As for the ceramic particle reinforced aluminum matrix composites, detailed investigations have been done to explore the influence of particle size (or PSR) on the dispersion/distribution of reinforcements, and the mechanical behavior of the composites. Slipenyuk et al. [17] reported that by decreasing the PSR, higher yield and tensile strengths and better material fabricability can be obtained on the condition of a uniform reinforcement distribution. Fathy et al. [18] found that smaller PSR resulted in more uniform distribution of the SiC particles in the matrix and better tensile properties of the composites, while the agglomeration and voids increased when PSR increases. Since the majority of the studies is limited to PSR > 1, Diler et al. [19] used the size of the SiC particles larger than the aluminum powder; the results showed better densification and improved tribological properties with increasing PSR when the reinforcement volume fraction is below 17.5 %. It can be seen that the PSR is an important parameter determining the microstructure and properties of the composites to a large degree. On the other hand, the studies on aluminum matrix composites reinforced with metallic glass particles are mainly focused on the effect of different classes of reinforcements (Ni-, Zr-, Fe-, Mg-based [14, 20-24]). In addition, the previous works about the strength and plasticity of this new kind of composites are mostly under compressive load. In this study, aluminum matrix composites reinforced with a relatively high volume fraction (20 vol.%) of Zr48Cu36Ag8Al8 metallic glass particles and PSR < 1 are prepared via powder metallurgy. The effect of the PSR on the reinforcement distribution in the three-dimensional space is evaluated by quantitative methods, and the mechanical properties of the composites under both compressive and tensile loads as well as the strengthening mechanisms are investigated.
4
2. Experimental 2.1 Materials and preparation process Pure aluminum powder (purity ≥ 99.5 wt. %) with mean size of 13 μm (Fig. 1) was used as matrix. The Zr48Cu36Ag8Al8 (at.%) metallic glass particulate reinforcement was synthesized by gas atomization using free-fall atomization and subsequently sieved into 20-45 μm, 45-63 μm and 63-90 μm particle class sizes. The mean particle diameter (D50) of 35 μm, 54 μm and 75 μm (corresponding to a PSR of ~1/3, 1/4 and 1/6, respectively) for different particle class sizes were obtained via laser diffraction (Marlvern Matersizer 2000), as shown in Fig. 2. Powder mixtures consisting of the aluminum matrix with 20 vol.% of glassy particles with different sizes were prepared by using a Retsch PM 400 planetary ball mill equipped with hardened steel balls and vials. No process control agent was used. The composite powders were milled at room temperature for 2 h at 100 rpm with a ball to powder weight ratio of 10:1. To avoid or minimize possible atmosphere contamination during milling, the vials were sealed in a Braun MB 150B-G glove box under a purified argon atmosphere (less than 0.5 ppm O2 and H2O). Consolidation of the as-blended powder mixtures was carried out by uni-axial hot pressing under argon atmosphere at 673 K and 640 MPa with a holding time of 10 min. After hot pressing, some of the hot-pressed billets were further hot extruded with extrusion ratio of 9:1 at 673 K. Pure Al bulk specimens were produced by hot pressing or hot extrusion under the same condition as a reference group.
5
Fig. 1 As-received pure aluminum powder: (a) SEM micrograph and (b) particle size distribution.
Fig. 2 Zr48Cu36Ag8Al8 metallic glass powders: SEM micrographs of the glassy particles with average size (a) 35, (b) 54 and (c) 75 μm; (d) particle size distributions.
2.2 Composites characterization The phases of the samples were characterized by X-ray diffraction (XRD) using a Philips PW 1050 diffractometer with Co Kα radiation source operating at 40 kV and 40 mA. The microstructures 6
were investigated by scanning electron microscopy (SEM) using a Zeiss Gemini 1530 field emission microscope equipped with an energy dispersive X-ray spectroscope (EDX). The relative density of the samples was measured using the Archimedeans principle. The X-ray computed tomography (micro-CT) analysis was performed with a GE phoenix Nanotom device, using 110 kV voltage, 100 μA current and 750 ms detector timing. The obtained radiographic images with a spatial resolution of 3.64 μm were reconstructed with the VG Studio Max 2.2 software. Quantitative particle distribution analysis was carried out by dividing the image into 9 grids. The particle concentration (area) in each grid, as well as the standard deviation of the concentration per grid was calculated. Ten images (with sample area represents 2 × 2 mm) acquired from different cross sections (interval of 0.5 mm with a sample length of 5 mm) of the reconstructed CT 3D model were used. The distribution index D was evaluated by D=1−S
(1)
where S is the standard deviation of the particle concentration per grid. A higher distribution index indicates a more homogeneous microstructure of the composites. Compression tests were conducted on cylindrical specimens with 6 mm length and 3 mm diameter. Both ends of the specimens were polished to make them parallel to each other before the tests. Tensile samples with dog-bone cylindrical shape (Fig. 3) were prepared from the hot extruded composites. The mechanical tests were performed by using an Instron 5862 testing facility at room temperature with a crosshead speed of 6 × 10-4 mm/s for compression and 1 × 10-3 mm/s for tension. The compression tests were terminated after the specimen strain reached 20 %.
7
Fig. 3 Schematic representation of the tensile test samples.
3. Results and discussion 3.1 Microstructures Fig. 4 shows the morphology of the as-blended composite powder mixtures with different PSR. The glassy particles are not fractured during the ball milling process and they retain the original as-atomized spherical morphology. Some of the aluminum particles are cold welded with each other and turned into relatively large particle agglomerates. For most of the glassy particles, their surface is partially covered by aluminum, a beneficial aspect for the following consolidation process, where the soft aluminum can act as a binder between the hard-to-sinter metallic glass particles.
8
Fig. 4 Morphology of the as-blended composites powders with PSR (a) 1/3, (b) 1/4, and (c) 1/6; (d) high magnification micrograph corresponding to the white box in (b).
Fig. 5 shows the XRD patterns of the composites after the different preparation stages. The pattern of the as-atomized glassy particles is also shown. The Zr48Cu36Ag8Al8 powder displays the broad diffraction peak (at about 2θ = 45°) characteristic of amorphous materials without any detectable crystalline Bragg peak. The as-blended powders and consolidated composites have similar patterns, consisting of the sharp Bragg peaks corresponding to fcc aluminum along with the amorphous broad maximum of the metallic glass phase. The absence of additional phases indicates that crystallization of the glassy particles or the reaction between glassy particles and aluminum matrix does not occur during the fabrication process.
9
Fig. 5 XRD patterns of the as-atomized glassy particles, as-blended powder mixtures and consolidated composites.
Fig. 6 shows the relative density of the hot pressed unreinforced pure aluminum and composites with different PSR. A relative density of more than 98 % was achieved, indicating a good densification of the materials. Little difference of densification is found when the PSR changes. Fig. 7 presents the microstructure of the hot pressed composites. No porosity can be observed from the images, further corroborating the high densification of the materials. The distribution of the glassy particles is rather uniform irrespective of the PSR. Fig. 7(d) shows the interface between the aluminum matrix and a metallic glass particle. The interface is sharp without any visible reaction products, further supporting that the glassy reinforcement retained the amorphous structure, in agreement with the XRD results.
10
Fig. 6 Relative density of the hot pressed unreinforced pure Al and composites with different PSR.
Fig. 7 SEM micrographs of hot pressed composites with different PSR: (a) 1/3, (b) 1/4, and (c) 1/6; (d) interface between a glassy particle and the aluminum matrix.
The densification capability of the powders is an important factor in powder metallurgy, which determines the final physical and mechanical properties of materials. The densification mechanisms
11
for metal matrix composites reinforced with hard particles can be mainly classified as follows: (1) sliding and rearrangement of matrix powders and reinforcing particles, (2) plastic deformation of matrix powders, (3) localized particles fragmentation and (4) elastic compaction of the bulk nearly nonporous composites [19, 25]. At the initial compaction stage, the arrangement of matrix powders and reinforcement particles are the dominant densification mechanism. With increasing the compaction pressure, the movement of the matrix powders and reinforcement particles is restricted, the compaction energy is spent on the deformation of matrix powders and the friction between powders and particles. The aforementioned densification mechanisms are strongly affected by the volume fraction of the reinforcement particles, and the size and shape of particles and matrix powders. In the present study, the PSR is less than 1, and most of the glassy particles in the composites are isolated from each other (Fig. 7(a) – (c)). This means that the matrix powders can easily be located in the voids between the reinforcement particles, which is responsible for the high densification of the composites. A more realistic representation of the distribution of the glassy particles within the Al matrix can be obtained by X-ray microtomography (Fig. 8). The 3D reconstructions clearly show the difference in the distribution of the glassy reinforcements when the PSR changes. For the hot pressed composite with PSR = 1/3 (Fig. 8(a)), the tomography reveals that the particles are homogeneously distributed in 3D, giving rise to a high distribution index of 96.3 %. With the increase of the PSR to 1/4 and 1/6, the spatial distribution of reinforcements becomes progressively nonuniform (Figs. 8(b) and (c)) and the distribution index decreases to 88.1% (Fig. 8(d)). After hot extrusion, a slightly more homogeneous distribution of the reinforcements is exhibited by the composites compared with their
12
hot pressed counterparts (Fig. 8(d)). In addition, the glassy particle distribution in the hot extruded composites displays the same trend as for the hot pressed ones, both decreasing with the increase of PSR.
Fig. 8 3D views of the glassy particle distribution in hot pressed and hot extruded composites obtained by X-ray microtomography for the PSR: (a) 1/3, (b) 1/4, and (c) 1/6; (d) corresponding spatial distribution index.
The PSR plays an important role in obtaining a homogeneous microstructure of the composites. When the size of the matrix powders is larger than the reinforcement particles (PSR > 1), the homogeneity of the particles distribution in the composites decreases with increasing PSR; a behavior that can be attributed to the geometry dependent clustering [18, 26-28]. The increase of 13
PSR leads to a reduction in the combined surface area of the matrix powders, which becomes insufficient for a uniform arrangement of the reinforcement particles. As a result, the reinforcement particles cluster in the interstices of the matrix powders, leading to their inhomogeneous distribution. The corresponding microstructure of the composites is usually called as “necklace structure” [18, 26]. In this study, the size of aluminum powders is less than the glassy particles (PSR <1); although the bonding surface area of the matrix powders changes when PSR decreases, this seems to have only a limited influence on the variation of the particle distribution. As can be seen in Figs. 7(a) - (c), most of the glassy particles are isolated from each other, and the inhomogeneity mainly results from the nonuniform distribution of the glassy particles in the three-dimensional space (particle rich or deplete regions, Fig. 8(a) – (c)) instead of particle agglomerations in the whole composites. The mixing process of the composite powders can be seen as a circulation flow of the matrix powders and reinforcement particles under an external force (i.e. ball milling), and the flowability of each powder or particle is an important factor affecting this process. The particle shape, the particle size distribution, powder density and humidity determine the powder flowability [29]. Generally, spherical and heavier particles with larger size display better flowability compared with irregular small and light powders. Therefore, the glassy particles (ρ = 7.11 g/cm3) are expected to have superior flowability than the aluminum powders (ρ = 2.70 g/cm3) in the present case. With increasing the size of the glassy particles (i.e. PSR decreases), the discrepancy of flowability between glassy particles and aluminum powder becomes much more remarkable. In other words, during the powder mixing process as well as the following procedures which may involve stirring or blending (e.g. powders transfer), the glassy particles may tend to flow together in the composite powders with
14
decreased PSR, thus particle-rich (or depleted) regions appear. Slipenyuk et al. [30] reported that a PSR of around 3 may be low enough to obtain a composite material with relatively homogenous distribution of reinforcements, provided that the content does not exceed 15 vol. %. Here, the composites show a rather homogeneous microstructure when the PSR is 1/3. This suggests that it may be possible to draw an empirical conclusion that with 1/3 ≤ PSR ≤ 3, a relatively homogenous distribution of reinforcements in the composites can be achieved when the volume fraction of the particles is no more than 20 %, and the closer that PSR is to 1, the more homogeneous is the microstructure. During the extrusion process, the distribution of the reinforcement becomes more homogeneous: the large deformation results in the flow of the reinforcements within the matrix, and the movement of the particles from the rich to depleted zones, promoting the homogenization of the composites microstructure. However, since there is a dramatic rise of the hydrostatic stress in the particle rich regions, which constrains the lateral plastic strain of the specimen [31], the improvement of the homogeneity is rather limited and the particle rich regions are only elongated in the extrusion direction, forming particle-rich bands and a fiber-like texture.
3.2 Mechanical properties Fig. 9 (a) shows the compressive properties of the hot pressed pure aluminum matrix and composites. The compressive yield strength (YS) of the pure matrix and composites with PSR = 1/3 are 90 MPa and 125 MPa, respectively, corresponding to an improvement of about 39 %. When the PSR decreases, the compressive yield strength of the composites does not show significant changes.
15
Fig. 9 (b) shows the compression stress – strain curves of the composite with PSR = 1/3 after hot pressing and hot extrusion. The hot extruded composite exhibits higher yield strength compared to the composite that was only hot pressed. This can be ascribed to the generation of the fiber-like texture during extrusion. The tensile properties of the hot extruded composites with different PSR are shown in Fig. 10. The three composites exhibit tensile yield strength on the same level (between 155 and 160 MPa), while the ultimate tensile strength (UTS) and total extension at fracture present large difference. The composite with PSR = 1/3 has UTS and strain at fracture of 235 MPa and 14 %, respectively, which sharply drop to 203 MPa and only 3 % for the PSR = 1/6.
Fig. 9 Compression properties of: (a) hot pressed composites with different PSR and (b) the hot pressed and hot extruded composites with PSR = 1/3.
16
Fig. 10 Tensile properties of the hot extruded composites: (a) stress - strain curves and (b) corresponding mechanical data as a function of the PSR.
Fig. 11 shows the fracture surface after tensile tests for the composites with PSR = 1/3 and 1/6. The composite with PSR = 1/3 shows a rough fracture surface and a large number of dimples (Fig. 11(a)), indicating a ductile fracture mode operating in the composites. When the PSR is 1/6 (Fig. 11(c)), the morphology turns into relatively flat, comprising several fractured (marked by arrows) and many decohered glassy particles surrounded by ductile regions (termed as “tear ridges”). The larger size of the particles results in easier initiation of cracks since voids nucleate at the matrix/particle interface when dislocations are pushed to the undeformable particle under straining [32]. Secondly, the largest possible defect within a particle is limited by its size and there is a higher statistical probability of there being a critical crack in larger particles, i.e. larger particles are more likely to fracture than smaller ones [31]. The voids, critical defects coalesce with each other when the applied load increases, and final fracture is achieved by rapid crack propagation through the matrix between the reinforcement particles. Moreover, the reduced PSR is detrimental for obtaining the homogeneous spatial distribution of reinforcements, which leads to stress concentration under deformation, further promoting the growth of cracks and fracture of the composites. It should be noted that the matrix dimples present on the surface of the glassy particles (Fig. 11(b) and (d)) indicate a good interface bonding between aluminum and the metallic glass.
17
Fig. 11 Morphology of the fracture surface for the composites with (a) and (b) PSR =1/3, (c) and (d) PSR = 1/6.
3.3 Strengthening mechanisms The strengthening effect of particles in metal matrix composites can be mainly divided into two categories: direct and indirect strengthening [21, 33]. These two strengthening factors are interdependent and act simultaneously, resulting in a combined effect [33]. Direct strengthening considers the load-bearing effect of the reinforcements, in which the load is transferred from the soft matrix to the hard reinforcements via the matrix-particle interface. According to the modified shear lag model [34], the yield strength of the composites reinforced with spherical particles can be described as follows: σ𝑐𝑦 = 𝜎𝑚𝑦 (1 +
𝑉𝑝 2
18
)
(2)
where 𝜎𝑐𝑦 and 𝜎𝑚𝑦 are the yield strength of the composites and matrix respectively, and 𝑉𝑝 is the volume fraction of the reinforcement particles. The indirect strengthening is related to the increase of dislocation density in the matrix due to the introduction of the reinforcements, and the yield strength of the matrix 𝜎𝑚𝑦 can be expressed as [21, 33, 35] 𝜎𝑚𝑦 = 𝜎0 + Δ𝜎𝑑𝑖𝑠 = 𝜎0 + Δ𝜎𝑜𝑟 + Δ𝜎𝑡ℎ𝑒 + Δ𝜎𝑔𝑒𝑜 + Δ𝜎𝑠
(3)
where 𝜎0 is the yield strength of the unreinforced matrix, Δ𝜎𝑑𝑖𝑠 is the yield strength increase caused by the dislocation strengthening, Δ𝜎𝑜𝑟 is the Orowan stress [14, 32, 36, 37], Δ𝜎𝑡ℎ𝑒 is the statistically stored dislocation caused by the thermal mismatch between the matrix and reinforcements during cooling [14, 38-40], Δ𝜎𝑔𝑒𝑜 is the geometrically necessary dislocations [41, 42] and Δ𝜎𝑠 is the strength increase attributed to the decrease of matrix ligament size [14, 21, 32, 33, 36, 37, 43]. The stress increment due to geometrically necessary dislocation can also be written as [44-46] ∆σ𝑔𝑒𝑜 = ∆𝜎𝑖𝑠𝑜 + ∆𝜎𝑘𝑖𝑛
(4)
where ∆𝜎𝑖𝑠𝑜 and ∆𝜎𝑘𝑖𝑛 are the isotropic and kinematic strain gradient contribution to the flow stress respectively. The different contributions to the improvement of the yield strength resulting from dislocation strengthening are summarized in Tab. 1. The strengthening effect caused by the reduced matrix ligament size and the thermal mismatch between matrix and glassy particles give the largest contributions to the increment of yield strength. Tab. 2 lists the calculated (according to Eqs. (2) and (3)) and experimental yield strength of the composites with different PSR. The results of the
19
predicted yield strength are slightly higher than the ones derived experimentally. The reasons for this behavior can be ascribed to: (a) the different dislocation strengthening mechanisms interact with each other and the contributions to the matrix yield strength may not be simply additive (Eq. (3)) [35]; (b) the particle size distribution, nonideal periodical configuration of the particles [36, 43] etc. are not taken into consideration in the mechanical models. Therefore, the present calculations should be viewed as an upper limit with respect to the strength prediction. Tab. 1 Calculated increment of yield strength from dislocation strengthening
PSR 1/3 1/4 1/6
Δσor (MPa) 0.3 0.2 0.1
Δσgeo (MPa) Δσiso Δσkin 0.6 1.4 0.5 1.1 0.4 1.0
Δσthe (MPa) 8.5 7.6 7.0
Δσs (MPa) 27.5 22.1 18.8
Δσdis (MPa) 38.3 31.6 27.3
Tab. 2 The estimated and experimental yield strength of composites with different PSR
PSR 1/3 1/4 1/6
σcy1 a (MPa) calculated experimental 141 125±8 134 122±2 129 116±4
σcy2 b (MPa) calculated experimental 174 159±3 167 152±10 162 157±1
σcy1 a Compression yield strength, hot pressed composites, σcy2 b Tensile yield strength, hot pressed and hot extruded composites.
4. Conclusions Aluminum matrix composites reinforced with different sizes of Zr48Cu36Ag8Al8 metallic glass particles (PSR < 1) were synthesized by powder metallurgy. All the composites show high densification (relative density > 98%) since the soft matrix powder can easily fill the interstices 20
between the reinforcement particles. The glassy particles retained their amorphous structure during the composite fabrication process. Quantitative particle distribution analysis in the three-dimensional space demonstrates that the distribution index decreases with reducing PSR. The comparison of the present results with those reported in literature suggests that a range of 1/3 ≤ PSR ≤ 3 may be necessary for attaining a relatively homogenous distribution of reinforcements; furthermore, when the volume fraction of reinforcement does not exceed 20 %, the closer that PSR is to 1, the more homogeneous is the microstructure. Hot extrusion can further improve the particle spatial distribution. The yield strength of the composites did not show significant variations within the PSR range analyzed here (1/3 – 1/6) either under compressive or tensile loading. However, the ultimate tensile strength and the corresponding ductility are reduced significantly, from 235 MPa and 14 % for PSR = 1/3 to 203 MPa and only 3 % for PSR = 1/6. At the same time, the ductile dimple fracture mode turned into mixed fracture of interface debonding and crack coalescence in the matrix between the particles. Quantitative strengthening mechanisms calculations indicated that the main strengthening contributions in the present composites come from the reduced matrix ligament size of the matrix, the thermal mismatch between matrix and glassy particles and load-bearing of the reinforcements. The calculated yield strength of the composites is consistent with experimental results.
Acknowledgements The authors thank B. Opitz, A. Funk and H. Merker for technical support, and R.N. Shahid and P. Wang for valuable suggestions. T. He gratefully acknowledges the financial support of the China Scholarship Council (CSC).
21
References: [1] C. William, J. Harrigan, Commercial processing of metal matrix composites, Mater. Sci. Eng. A. 244 (1998) 75-79. [2] Y. Cui, L.F. Wang, J.Y. Ren, Multi-functional SiC/Al Composites for Aerospace Applications, Chin. J. Aeronaut. 21 (2008) 578-584. [3] M. Rahimian, N. Parvin, N. Ehsani, Investigation of particle size and amount of alumina on microstructure and mechanical properties of Al matrix composite made by powder metallurgy, Mater. Sci. Eng. A. 527(4) (2010) 1031-1038. [4] B. Bhav Singh, M. Balasubramanian, Processing and properties of copper-coated carbon fibre reinforced aluminium alloy composites, J. Mater. Process. Technol. 209(4) (2009) 2104-2110. [5] T. He, X. He, P. Tang, D. Chu, X. Wang, P. Li, The use of cryogenic milling to prepare high performance Al2009 matrix composites with dispersive carbon nanotubes, Mater. Des. 114 (2017) 373-382. [6] J.L. Li, Y.C. Xiong, X.D. Wang, S.J. Yan, C. Yang, W.W. He, J.Z. Chen, S.Q. Wang, X.Y. Zhang, S.L. Dai, Microstructure and tensile properties of bulk nanostructured aluminum/graphene composites prepared via cryomilling, Mater. Sci. Eng. A. 626 (2015) 400-405. [7] P. Jin, B. Xiao, Q. Wang, Z. Ma, Y. Liu, S. Li, Effect of hot extrusion on interfacial microstructure and tensile properties of SiCp/2009Al composites fabricated at different hot pressing temperatures, J. Mater. Sci. Technol. 27(6) (2011) 518-524.
22
[8] Y. Tang, Y. Deng, K. Zhang, L. Liu, Y. Wu, W. Hu, Improvement of interface between Al and short carbon fibers by α-Al2O3 coatings deposited by sol-gel technology, Ceram. Int. 34(7) (2008) 1787-1790. [9] C. Wang, G. Chen, X. Wang, Y. Zhang, W. Yang, G. Wu, Effect of Mg Content on the Thermodynamics of Interface Reaction in Cf/Al Composite, Metall. Mater. Trans. A 43(7) (2012) 2514-2519. [10] Y. Tang, L. Liu, W. Li, B. Shen, W. Hu, Interface characteristics and mechanical properties of short carbon fibers/Al composites with different coatings, Appl. Surf. Sci. 255(8) (2009) 4393-4400. [11] P. Yu, K.B. Kim, J. Das, F. Baier, W. Xu, J. Eckert, Fabrication and mechanical properties of Ni-Nb metallic glass particle-reinforced Al-based metal matrix composite, Scr. Mater. 54(8) (2006) 1445-1450. [12] P. Yu, L.C. Zhang, W.Y. Zhang, J. Das, K.B. Kim, J. Eckert, Interfacial reaction during the fabrication of Ni60Nb40 metallic glass particles-reinforced Al based MMCs, Mater. Sci. Eng. A. 444(1-2) (2007) 206-213. [13] M.H. Lee, J.H. Kim, J.S. Park, J.C. Kim, W.T. Kim, D.H. Kim, Fabrication of Ni-Nb-Ta metallic glass reinforced Al-based alloy matrix composites by infiltration casting process, Scr. Mater. 50(11) (2004) 1367-1371. [14] S. Scudino, G. Liu, K.G. Prashanth, B. Bartusch, K.B. Surreddi, B.S. Murty, J. Eckert, Mechanical properties of Al-based metal matrix composites reinforced with Zr-based glassy particles produced by powder metallurgy, Acta Mater. 57(6) (2009) 2029-2039.
23
[15] M.M. Khan, A. Nemati, Z.U. Rahman, U.H. Shah, H. Asgar, W. Haider, Recent advancements in bulk metallic glasses and their applications: a review, Crit. Rev. Solid State Mater. Sci.
(2017)
1-36. [16] P. Gupta, S. Pal, N. Yedla, Molecular dynamics based cohesive zone modeling of Al (metal)-Cu50Zr50 (metallic glass) interfacial mechanical behavior and investigation of dissipative mechanisms, Mater. Des. 105 (2016) 41-50. [17] A. Slipenyuk, V. Kuprin, Y. Milman, V. Goncharuk, J. Eckert, Properties of P/M processed particle reinforced metal matrix composites specified by reinforcement concentration and matrix-to-reinforcement particle size ratio, Acta Mater. 54(1) (2006) 157-166. [18] A. Fathy, A. Sadoun, M. Abdelhameed, Effect of matrix/reinforcement particle size ratio (PSR) on the mechanical properties of extruded Al-SiC composites, Int. J. Adv. Manuf. Technol. 73(5) (2014) 1049-1056. [19] E.A. Diler, A. Ghiami, R. Ipek, Effect of high ratio of reinforcement particle size to matrix powder size and volume fraction on microstructure, densification and tribological properties of SiCp reinforced metal matrix composites manufactured via hot pressing method, Int. J. Refract. Met. Hard Mater. 52 (2015) 183-194. [20] D. Markó, K.G. Prashanth, S. Scudino, Z. Wang, N. Ellendt, V. Uhlenwinkel, J. Eckert, Al-based metal matrix composites reinforced with Fe49.9Co35.1Nb7.7B4.5Si2.8 glassy powder: Mechanical behavior under tensile loading, J. Alloys Compd. 615 (2014) S382-S385.
24
[21] Z. Wang, J. Tan, B.A. Sun, S. Scudino, K.G. Prashanth, W.W. Zhang, Y.Y. Li, J. Eckert, Fabrication and mechanical properties of Al-based metal matrix composites reinforced with Mg65Cu20Zn5Y10 metallic glass particles, Mater. Sci. Eng. A. 600 (2014) 53-58. [22] Q. Yang, Y. Zhang, H. Zhang, R. Zheng, W. Xiao, C. Ma, Fabrication of Al-based composites reinforced with in situ devitrified Al84Ni8.4Y4.8La1.8Co1 particles by hot pressing consolidation, J. Alloys Compd. 648(Supplement C) (2015) 382-388. [23] Ö. Balcı, K.G. Prashanth, S. Scudino, M. Somer, J. Eckert, Powder metallurgy of Al-based composites reinforced with Fe-based glassy particles: Effect of microstructural modification, Part. Sci. Technol. (2017) 1-6. [24] F. Bahrami, R. Amini, A.H. Taghvaei, Microstructure and corrosion behavior of electrodeposited Ni-based nanocomposite coatings reinforced with Ni60Cr10Ta10P16B4 metallic glass particles, J. Alloys Compd. 714 (2017) 530-536. [25] P.J. Denny, Compaction equations: a comparison of the Heckel and Kawakita equations, Powder Technol. 127(2) (2002) 162-172. [26] V.V.B. Prasad, B.V.R. Bhat, Y.R. Mahajan, P. Ramakrishnan, Structure–property correlation in discontinuously reinforced aluminium matrix composites as a function of relative particle size ratio, Mater. Sci. Eng. A. 337(1) (2002) 179-186. [27] C. Wu, K. Ma, J. Wu, P. Fang, G. Luo, F. Chen, Q. Shen, L. Zhang, J.M. Schoenung, E.J. Lavernia, Influence of particle size and spatial distribution of B4C reinforcement on the microstructure and mechanical behavior of precipitation strengthened Al alloy matrix composites, Mater. Sci. Eng. A. 675 (2016) 421-430.
25
[28] O. El-Kady, A. Fathy, Effect of SiC particle size on the physical and mechanical properties of extruded Al matrix nanocomposites, Mater. Des. 54 (2014) 348-353. [29] A.B. Spierings, M. Voegtlin, T. Bauer, K. Wegener, Powder flowability characterisation methodology for powder-bed-based metal additive manufacturing, Prog. Addit. Manuf. 1(1) (2016) 9-20. [30] A. Slipenyuk, V. Kuprin, Y. Milman, J.E. Spowart, D.B. Miracle, The effect of matrix to reinforcement particle size ratio (PSR) on the microstructure and mechanical properties of a P/M processed AlCuMn/SiCp MMC, Mater. Sci. Eng. A. 381(1) (2004) 165-170. [31] P.B. Prangnell, S.J. Barnes, S.M. Roberts, P.J. Withers, The effect of particle distribution on damage formation in particulate reinforced metal matrix composites deformed in compression, Mater. Sci. Eng. A. 220(1) (1996) 41-56. [32] K.K. Chawla, M.A. Meyers, Mechanical behavior of materials, Prentice Hall1999. [33] S. Scudino, G. Liu, M. Sakaliyska, K.B. Surreddi, J. Eckert, Powder metallurgy of Al-based metal matrix composites reinforced with β-Al3Mg2 intermetallic particles: Analysis and modeling of mechanical properties, Acta Mater. 57(15) (2009) 4529-4538. [34] V.C. Nardone, K.M. Prewo, On the strength of discontinuous silicon carbide reinforced aluminum composites, Scr. Metall. 20(1) (1986) 43-48. [35] H. Sekine, R. Chen, A combined microstructure strengthening analysis of SiCp/Al metal matrix composites, Composites. 26(3) (1995) 183-188. [36] M.J. Tan, X. Zhang, Powder metal matrix composites: selection and processing, Mater. Sci. Eng. A. 244(1) (1998) 80-85.
26
[37] W.S. Miller, F.J. Humphreys, Strengthening mechanisms in particulate metal matrix composites, Scr. Metall. Mater. 25(1) (1991) 33-38. [38] R.J. Arsenault, N. Shi, Dislocation generation due to differences between the coefficients of thermal expansion, Mater. Sci. Eng. 81 (1986) 175-187. [39] N. Hansen, The effect of grain size and strain on the tensile flow stress of aluminium at room temperature, Acta Metall. 25(8) (1977) 863-869. [40] Y. Zhang, N. Mattern, J. Eckert, Atomic structure and transport properties of Cu50Zr45Al5 metallic liquids and glasses: Molecular dynamics simulations, J. Appl. Phys. 110(9) (2011) 093506. [41] M.F. Ashby, The deformation of plastically non-homogeneous materials, Philos. Mag. 21(170) (1970) 399-424. [42] N.A. Fleck, G.M. Muller, M.F. Ashby, J.W. Hutchinson, Strain gradient plasticity: Theory and experiment, Acta Metall. Mater. 42(2) (1994) 475-487. [43] T.W. Gustafson, P.C. Panda, G. Song, R. Raj, Influence of microstructural scale on plastic flow behavior of metal matrix composites, Acta Mater. 45(4) (1997) 1633-1643. [44] C.W. Nan, D.R. Clarke, The influence of particle size and particle fracture on the elastic/plastic deformation of metal matrix composites, Acta Mater. 44(9) (1996) 3801-3811. [45] L.M. Brown, W.M. Stobbs, The work-hardening of copper-silica v. equilibrium plastic relaxation by secondary dislocations, Philos. Mag. 34(3) (1976) 351-372. [46] L.M. Brown, Precipitation and Dispersion Hardening, Strength of Metals and Alloys, Pergamon1979, pp. 1551-1571.
27