Investigation of Porosity and Mechanical Properties of Graphene Nanoplatelets-Reinforced AlSi10 Mg by Selective Laser Melting

In recent years, the increasing demand of geometrically complex metal components has significantly stimulated the advancement of selective laser melting (SLM) technique due to their advantages of high material utilization, short design-to-product cycle, and high manufacturing flexibility. The main characteristic of SLM is the direct and layerwise addition of metal material with focused laser as heat source, to create prototypes or functional components according to predefined three-dimensional computer-aided design data. It has the advantages of fast build speed and high geometrical precision compared to other additive manufacturing techniques. However, SLM still exhibits several drawbacks when compared with traditional manufacturing methods. For example, metal parts produced via laser-assisted additive manufacturing (LAAM) techniques usually have high porosity, significant inhomogeneity and induced residual stresses, and low surface quality. In order to strengthen the metal parts produced by SLM and further inspire the potential of SLM for industrial implementation, we aim to improve the mechanical/physical properties of SLM-produced metal parts by incorporating secondary phase particles as reinforcement.

The target matrix material is AlSi10 Mg alloy. It comprises aluminum alloyed with silicon of weight fraction up to 10% and small amount of magnesium and iron, along with other minor elements. It is a widely used alloy for its excellent weldability and hardenability [1] along with good corrosion resistance and thermal conductivity. The good weldability is attributed to the near-eutectic composition of Al and Si, resulting in a small solidification range. Small amount of Mg is added to react with Al, and Mg₂Si precipitates are formed in the subsequent natural or artificial aging process. In addition, due to the natural formation of an oxide layer on the surface of the aluminum alloy, the material has high corrosion resistance. In literature, fabricating aluminum and Al-based alloy components using LAAM techniques has been widely studied. For example, Louvis et al. [2] used SLM technique to fabricate aluminum components and they found that the process is more difficult to control compared to other iron- or titanium-based alloys, mainly due to the high reflectivity of aluminum powders for laser beam, resulted from the oxidation on workpiece surface of both solid and molten materials. Buchbinder et al. [3] prepared AlSi10 Mg alloy by SLM by varying the laser power from 500 to 900 W and the scanning velocity from 1300 to 2100 mm/s. Under optimum manufacturing parameters, the relative density of product was more than 99.8%. Dinda et al. [4] conducted laser deposition of Al-11.28%Si alloy. By investigating material microstructure, it was found that Al grains were mostly oriented along <100> directions and a periodic transition of microstructural morphology from columnar dendrite to microcellular structure in each layer was seen. This phenomenon was attributed to the thermal history during deposition process. Fatigue performance of Al-based alloy made by LAAM is also an intensively investigated topic. For instance, Tang and Pistorius [5] prepared AlSi10 Mg parts by SLM, and they investigated the effect of hatch spacing and building orientation on material tensile performance and fatigue properties. Oxide particles formed by oxidation of vaporized metal during part manufacture are likely to become the crack initiation points and thus deteriorates fatigue performance. Brandl et al. [6] investigated the effect of platform temperature, building direction, and heat treatment on fatigue life of AlSi10 Mg produced by SLM; it was found that heat treatment had the largest effect. Meanwhile, in order to improve the performance of LAAM-produced metal parts, strengthening particles are often used to strengthen Al/Al-alloy material. For example, Gu et al. [7] fabricated nano-TiC particle-reinforced AlSi10 Mg nanocomposite by SLM process. It was discovered that the addition of TiC reinforcement significantly lowered the coefficient of friction and wear rate, and insufficient laser energy per unit length caused balling effect and formation of residual stress. A similar work from the same group was reported on fabricating SiC particle-reinforced AlSi10 Mg matrix material by SLM process [8]. Ghosh et al. [9] obtained SiC-reinforced Al-4.5Cu-3 Mg alloy matrix composite material using direct metal laser sintering. Microstructure observations indicated that SiC particulates are bonded with the matrix alloy leading to formation of grain boundary. The influence of laser power and scanning speeds on resulted surface roughness, density, and hardness of the SLM-processed Al/Fe₂O₃ composite material was also investigated [10].

By conducting the literature review on LAAM fabrication of metal matrix composite (MMC), it is found that most existing research adopt particulate as reinforcement but other forms of reinforcement (e.g., whiskers, platelets, fibers) are still uncharted in LAAM processes. In this study, we aim to fill this gap by innovatively adopting graphene nanoplatelets (GNPs) as reinforcement to strengthen AlSi10 Mg in SLM process. Graphene refers to all forms of graphitic material from 100 nm down to single carbon atomic layer [11]. Due to the excellent mechanical and unique structured characteristics of graphene, as well as good bonding with metals, fabrication of novel graphene-reinforced MMCs has drawn significant research interests. However, due to the difficulties of dispersing GNPs into metal matrix, graphene-reinforced metals have not been studied as intensely as compared to graphene-reinforced polymers. There are a number of studies in literature focusing on fabrication of GNPs-reinforced MMCs. For example, Wang et al. [12] obtained a multilayer graphene/Al composite using flake powder metallurgy. Compared to pure aluminum from the same metallurgical process, the tensile strength was improved significantly. Tensile enhancement was also reported by Rashad et al. [13], in which GNPs were added into Mg–Al–Sn alloy powders, followed by high energy ball milling and hot extrusion. A novel processing route that combines liquid state ultrasonic processing and solid state stirring was proposed by Chen et al. [14] to fabricate a GNP/Mg composite. Bastwros et al. [15] prepared a graphene-reinforced aluminum composite, and the effect of ball milling parameters on graphene dispersion was investigated. Under favorable condition, an increase of flexural strength was observed compared to the unreinforced material.

To the best of our knowledge, the research of producing bulk GNPs-reinforced light-weighted metals using SLM process remains uncharted. The present study aims to fill these gaps by investigating the feasibility of reinforcing AlSi10 Mg alloy using GNPs through SLM, and the microstructure and tensile properties of obtained materials are characterized, and the related strengthening mechanism is discussed.

GNPs in the form of few-layer-graphene (6–10 layers) were made by chemical vapor deposition method, and the size distribution is between 5 and 50 μm. Before mixing with metal powders, the GNPs were washed with 5% hydrochloric and distilled water for multiple times followed by drying at 60 °C. Commercially available AlSi10 Mg powders were used as matrix phase; they are spherical in shape with a mean diameter of about 40 μm and the chemical composition is listed in Table 1. AlSi10 Mg powders and GNPs are mixed with appropriate amount of ethanol as the mixing agent. Later, the slurry was sealed into a planetary ball milling tank, and the milling process duration is 4 h. Zirconia container and milling balls were used to avoid contamination of powder. The ball-to-powder ratio used is 3.0, and the ball milling was carried out in vacuum environment to avoid oxidation and the temperature was maintained below 70 °C. Finally, the milled slurry was placed in drying oven for 24 h at 60 °C. By following this procedure, metal/GNPs mixture powders with GNPs content of 0.5 wt % was obtained for SLM process. Figure 1 shows the scanning electron microscopy image of obtain mixture powders. It can be seen that overall the AlSi10 Mg powders have spherical shape. Due to the powder refinement effect of ball milling process, some small particles less than 5 μm are also found throughout the observation area. However, GNPs cannot be clearly characterized due to the limited resolution.

Pure AlSi10 Mg powders and GNPs/AlSi10 Mg powder mixture are separately processed by a commercial EOS-280 SLM system to obtain pure AlSi10 Mg and GNPs-reinforced AlSi10 Mg parts. The geometry of tensile test specimens is exactly described by the stereolithography input file, so that the SLM-produced specimens can be directly used for tensile test and no more postmachining is needed. The geometry of tensile specimen is in compliance of ASTM E8 standard, and the computer-aided design drawing is shown in Fig. 2. The SLM parameters were set fixed for both pure and reinforced AlSi10 Mg materials, with 370 W of laser power, 1300 mm/s of scan speed, 0.19 mm of scan space, and 0.03 mm of layer thickness. The entire SLM process was protected by argon to avoid oxidation of powders. Figure 3 shows the schematic of material building strategy. The longitudinal direction of tensile samples is oriented horizontally, so that the tensile direction is perpendicular to SLM build direction. The materials were directly built on a stainless steel substrate without preheating and electrical discharge machining was used to detach specimen from substrate. Figure 4 shows the as-built pure and GNPs-reinforced AlSi10 Mg tensile test samples without polishing.

After fine polishing of the SLM-produced samples, Keller's etchant is used to reveal the microstructure of material. Images were obtained by 40× magnification optical microscope, and a FEI Quanta 200 scanning electron microscope. The tensile performance of obtained unreinforced and GNPs-reinforced AlSi10 Mg parts was characterized by a WDW 20 universal tensile tester at ambient temperature. Before tests, sample surfaces were fine-polished to eliminate any visible scratches and defects on the sample surface. The tension rate was maintained 1 mm/min during the test and the engineering strain–stress curves until fracture were recorded. image pro plus software package was used to process the images obtained by optical microscopy and obtain material porosity and pore size distribution.

Microstructure plays an important role in determining the mechanical and physical properties of material. Microstructure after etching is obtained using optical microscopy, as shown by Fig. 5. The view plane is horizontal, which is parallel to laser scan plane. Laser tracks can be clearly seen and different tracks represents different laser movement vectors. The average individual track width is about 120 μm. Multiple longitudinal laser scanning tracks along different directions are clearly seen; because multidirection scanning strategy is used for building the material in this study, the direction of laser scanning is rotated by a certain angle when building each new layer. SLM with variable scan direction is believed to give larger overlapping between tracks and improve material density. In addition, multidirectional scan strategy is beneficial to obtain more uniform material and reduce anisotropic characteristics on XY (horizontal) plane. For both pure AlSi10 Mg and GNPs-reinforced AlSi10 Mg, a number of pores are visible at the investigated magnification (40×). The pores in pure AlSi10 Mg are generally spherical in shape and distribute in the region of both melt pool interior and melt pool boundary. However, by comparing Figs. 5(a) and 5(b), the porosity is significantly increased when 0.5 wt % GNPs are added, and some pores exhibit irregular shape and are significantly larger in size up to 100 μm. The irregular shaped pores are found to preferentially distribute at the area in the vicinity to melt pool boundaries. The formation of irregular pores is regarded as the result of un-melted powders or insufficient overlapping between laser tracks, and the formation of spherical pores is generally caused by entrapped gasses in powder bed or the voids contained inside powder particles [16]. Many attempts have been made to reduce the porosity of additively produced metal components, such as increasing laser power density [17], using spherical particles with smaller size [18], and postprocessing by heat isostatic pressing [19]. In the present study, the formation mechanism of pores becomes more complicated due to the addition of external reinforcement phase. In order to effectively reduce the porosity and optimize the mechanical performance, further comprehensive investigations on the effects of manufacturing variables and characteristic of GNPs, as well as the GNPs/matrix interaction at both solid and liquid states, need to be carried out in the future. It is well known that porosity significantly influences a series of mechanical properties (e.g., fatigue, tensile strength, compressive strength, ductility) and therefore makes the analysis of strengthening effects of incorporated GNPs difficult. In order to understand the effects GNPs on the tensile performance of obtained composite, the weakening effect of pores on material strength needs to be investigated first.

Room temperature tensile test was conducted to investigate the tensile performance of as-built pure AlSi10 Mg and GNPs-reinforced AlSi10 Mg. Three important performance indicators, namely, ultimate tensile strength, yield strength, and percent elongation, were obtained from stress–strain curves. The percent elongation is the ratio of gage length elongation caused by plastic deformation to original gage length. It was obtained by fitting together specimen after fracture and measuring gage length increase. The comparison summary is presented in Fig. 6. By comparing the results of unreinforced and reinforced AlSi10 Mg, it is not surprising that there is no obvious improvement of material strength, considering the significant porosity increase of GNPs-reinforced sample. The ultimate tensile strength values are 337 and 346 MPa, for pure AlSi10 Mg and GNPs-reinforced AlSi10 Mg, respectively. The yield strength values of the two contrast materials are 234 and 246 MPa, respectively. It should also be noted that the addition of GNPs does not deteriorate the material ductility; on the contrary, the percent elongation slightly increases from 3.0% to 3.2%. It is attributed to the excellent bonding between GNPs and metal matrix.

Figure 7 shows the micrographs of the fracture surfaces of fractured tensile specimens. For both materials, ductile failure characteristics, such as dimple morphology, are not obvious. Both materials underwent brittle failure. The results are consistent with the low percent elongation obtained. In addition, for both materials, a large number of pores with large size distribution ranges can be observed on the fractured surface, with pore size ranging from several micrometers to about 20 μm. It is important to note that the pore density on the fracture surface is significantly higher than that on the planar cross section as shown in Fig. 5. This strongly suggests that fracture preferentially occurs along the pores and voids inside the materials as they provide easier propagation paths for the cracks when the materials fails. Tear ridges are generally observed around pores, especially for large-sized pores, indicating significant stress concentration in the vicinity of pores under loading.

In order to numerically investigate the effect of pores on material properties, the porosity of material must be quantified at first. In order to do this, the planer metallographical images with 2048 × 1536 pixels at 40× magnification are converted into binary black and white format, which only highlights the pores' morphologies while other metallurgical information is eliminated, as shown by Fig. 8. The dark area in the figure represents pores and the white area represents dense material. It is found that the distribution of pores shape in unreinforced sample is overall uniform and the shape is generally spherical. However, for the GNPs-reinforced samples, the pore shape exhibits more irregular characteristic and the distribution of pores is less uniform as they tend to aggregate at certain areas. Since it is reported that irregular pores in SLM-produced metal parts are related to insufficient energy transmission between laser beam and metal powders to be melted [16]. It is inferred that the addition of GNPs alters the interaction between laser beam and metal powders, but further experiments need to be carried out to validate this speculation.

The porosity and number of pores are counted using the software package. Since the pore distribution exhibits random characteristics, in order to obtain statistically significant results, eight metallographical images away from sample surfaces and at least 500 μm apart were processed by following the above-mentioned method for the contrast samples. The results are averaged for analysis. Figure 9 summarizes the comparison of porosity and pore amount between the two materials. It is found that by adding 0.5 wt % GNPs, the porosity of material increases from 1.2% to 5.3%, and the number of pores also increases from 495 to 1782. Figure 10 shows the distribution of pore size. It is observed that the size distributions of pores for both materials are similar, and the number of pores decreases as the pore size increases. The size ranges are also similar as 99% of pores are less than 40 μm in both cases.

In the present study, in order to obtain the value of $σTS0$ and $b$⁠, experimental data from the present study and that of Read et al. [22], who adopted the same process to fabricate AlSi10 Mg samples, are used to fit the curve shown by Eq. (2). The obtained values for $σTS0$ and $b$ are 348.9 MPa and 2.95, respectively. Therefore, the relationship between relative density and porosity for SLMed pure AlSi10 Mg can be depicted, as shown in Fig. 11.

Several factors account for the strength increase when GNPs are added into AlSi10 Mg. First, load transfers from matrix material to GNPs, and applied load can be transferred from metal matrix to the GNPs by interface shear (shear lag) and thus the high stiffness is utilized directly. Our previous study indicated that load transfer effect is the dominating strengthening effect, mainly due to the unconventionally large aspect ratio of GNPs [23]. Second, during SLM, the high mismatch of thermal expansion coefficient between GNPs and AlSi10 Mg results in high-density dislocation when material solidifies. The dislocations around GNPs result in the increase of strain hardening rate and therefore increase material strength. In addition, the Orowan strengthening plays an important role on strengthening material as dislocations movement is hindered by the dispersed GNPs in the matrix; as a result, higher critical stress is required for the dislocations to propagate through matrix.

and the theoretical strength of pure AlSi10 Mg at the resulted porosity (⁠$ε=5.3%$⁠) can be obtained by from Fig. 11, which gives the value of 285.7 MPa. Simple mathematical operation gives the strengthening effect of GNPs: $ΔσGNPs=60.2 MPa$⁠. Considering the very few amount of incorporated GNPs (0.5 wt %), the strengthening effect of GNPs is significant.

In the present study, pure AlSi10 Mg and 0.5 wt % GNPs-reinforced AlSi10 Mg samples are produced using SLM using the same processing parameters. Surprisingly, tensile tests show no significant improvement of strength for the reinforced AlSi10 Mg. Fractography analysis indicates that the dominating fracture mode is brittle failure for both materials, and stress concentration is observed in the areas near pores. Also, observations indicate that adding GNPs results in the increase of porosity. For unreinforced and reinforced samples, the average porosity values are 1.2% and 5.3%, respectively.

It is well known that the increase of porosity leads to significant decrease in material strength. Therefore, it is suspected that the potential beneficial effects from GNPs' addition are offset by the detrimental effect from porosity increase. In other words, the two effects should be decoupled. For this purpose, the quantitative relationship between the relative strength and porosity for SLMed AlSi10 Mg is obtained. By taking into consideration the strength reduction caused by the increased porosity, the strengthening effect of GNPs turns out to be significant, which is 60.2 MPa. In addition, the reason of porosity increase in the case of GNPs-reinforced composite is not fully understood yet. Further research is called for in the future.

• •

  Directorate for Engineering (Grant No. CMMI-1563002).