Objective AlSi10Mg alloy, prepared by selective laser melting, is one of the most widely investigated aluminum alloys recently. At present, most studies focus on the tensile strengths and fatigue properties of as-built or heat-treated alloys. However, fracture toughness is reported rarely. A study on as-built alloys indicated that fracture toughness was anisotropic. The plane-strain fracture toughness (KO is the lowest when the crack surface is parallel to the building direction, whereas it is the highest when the crack surface is perpendicular to the building direction. Heat treatment is beneficial to reduce or even eliminate the microstructural and property anisotropy and the residual stress. However, the anisotropy of fracture toughness and intrinsic principle for AlSi10Mg alloy after annealing have not been reported. Therefore, to explore the anisotropy of fracture toughness of annealed AlSi10Mg alloy prepared by selective laser melting, the alloy annealed at an optimum temperature is used to measure fracture toughness for different opening directions and analyze the reasons for fracture toughness anisotropy. Methods AlSi10Mg alloy is fabricated on XLine 1000R Concept Laser equipment using atomized alloying powders. The selective laser melting process is performed by a checkerboard pattern scanning strategy under an argon atmosphere with the volume fraction of oxygen controlled below 0.1 %. Then, as-built blocks are annealed at 275 degrees C for 2 h. The sample blocks, ground on a series of diamond sandpaper, are polished on the LectroPol-5 electrolytic machine. The microstructures, fracture morphologies, and grain boundary distribution are observed via optical microscopy, field emission scanning electron microscopy, and electron backscattered diffraction, respectively. The room temperature tensile properties in the X, Y and Z directions, analyzed via tensile tests according to GB/T 228. 1 2010, are used to calculate K-1c. The compact tensile specimens with a width (W) of 70 mm in different opening directions are prepared and tested according to GB/T 4161-2007. The load and crack opening displacement during tests are recorded. The conditional values of K-ic(K-Q) for different compact tensile specimens are calculated, and their validity are evaluated according to GB/T 4161-2007. Finally, the J-integral value and crack tip opening displacements are computed for different opening direction samples. Results and Discussions Both K-Q and K-Ic are invalid for all compact tensile specimens because the specimen thicknesses (B), pre-crack lengths (a), ligament lengths (b), and values of F-max/ F-Q do not meet the requirements of K-Q and K-IC, evaluation criteria stipulated by GB/T 4161-2007. Therefore, the fracture toughness and its anisotropy of annealed alloy are estimated by J-integral values and crack tip opening displacements. Results show that the J-integral value of the X-Y opening direction sample is approximately 430 kJ/m(2) (Table 2), which is the same as the Y-Z opening direction sample. Meanwhile, it is 250 kJ/m(2) for the Z-Y opening direction sample. Similarly, the crack tip opening displacements of the X-Y and Y-Z opening direction samples are also almost equal, approximately 0.8 mm. However, it is significantly lower for the Z-Y opening direction specimen, with a value of 0.47 mm, indicating that the fracture toughness of the annealed AlSil0Mg alloy is also anisotropic and similar to the as-built alloy. The results of microstructural observation indicate that the annealed alloy still exhibits the characteristics of "fish-scale" melt pools stacking layer by layer in parallel to the building direction, whereas it presents an interwoven morphology of melt pools as the structure is perpendicular to the building direction, indicating that the difference in fracture toughness in different directions is related to microstructural anisotropy. Because the structures near molten pool boundaries are relatively coarse and the ratio of low angle grain boundary is high, the cracks of specimens in the Z-Y opening direction tend to propagate along molten pool boundaries, resulting in lower fracture toughness. However, the internal structures of the molten pool, existing with a higher ratio of high angle grain boundary, are relatively fine, inducing good fracture toughness when cracks propagate through the interior of molten pools for the X-Y and Y-Z opening direction samples. Conclusions The microstructures and fracture toughness of annealed AlSi10Mg alloy are anisotropic. When a crack surface is parallel to the building direction, the fracture toughness is high, J-integral value and crack tip opening displacement are 430 kJ/m(2) and 0. 8 mm, respectively. Meanwhile, it is lower when the crack surface is perpendicular to the building direction, and the J-integral value and crack tip opening displacement are just 250 kJ/m(2) and 0.47 mm, respectively. The microstructural anisotropy and diversity between molten pool boundaries and the internal structure of the molten pools are the reasons for fracture toughness anisotropy. Because the fracture toughness of the annealed AlSi10Mg alloy manufactured by selective laser melting is relatively high, it is difficult to obtain effective K-1c value according to the general test method, such as GB/T 4161-2007. It may be necessary to use a special linear elastic plane-strain fracture toughness test method, for instance, ASTM B645.
