Simultaneously enhanced strength and ductility of AlSi7Mg alloy fabricated by laser powder bed fusion with on-line static magnetic field

ABSTRACT This work studied the effects of an on-line static magnetic field on the defects, microstructures, and mechanical properties of AlSi7Mg samples fabricated by laser powder bed fusion (LPBF). Process experiments were carried out on a self-developed LPBF equipment with an on-line static magnetic field generating system, where magnetic field intensity was adjustable from 0 to 0.3 T. With the action of static magnetic field, the relative density of samples increased from 96.9% to 98.6%. Furthermore, the solidification front of the columnar grain in the mushy zone was broken. With the increase of magnetic field intensity, the crystallographic orientation changed from strong <001> to <001>, <101> and <111> uniform distribution and the average grain was gradually refined from 8.35 to 7.22 μm. Based on the above optimisation, the ultimate tensile strength increased from (326.67 ± 5.31) MPa to (382.00 ± 2.45) MPa. Simultaneously, the elongation at break increased from 8.48% ± 0.20% to 11.78% ± 0.20%. In general, the reduction of pores, the refinement of grains and the increase of Mg2Si precipitates contributed to the simultaneous enhancement of strength and toughness together. This study could provide a new idea for laser additive manufacturing of excellent performance aluminum alloys.


Introduction
Laser powder bed fusion (LPBF), as a kind of metal additive manufacturing (AM) technique, can manufacture near-full dense components through laser melting of metal powder and stacking layer by layer.This process makes it possible to manufacture parts with extremely complex shapes that would be different or impossible to produce using conventional manufacturing processes (Yap et al. 2015;Song et al. 2015;Zhao et al. 2022b).Furthermore, much better mechanical properties could be obtained by LPBF than conventional casting technologies due to the rapid cooling rate of 10 5 -10 7 K/s (Herzog et al. 2016;Zhao et al. 2019).Aluminium alloy, as an indispensable advanced material for aerospace, electronics and other industrial fields, has received a lot of researches in the field of LPBF.So far, many researches have shown that better mechanical properties, fewer defects and finer grain size could be obtained than that fabricated by conventional processes (Prashanth et al. 2014;Thijs et al. 2013;Zhang et al. 2019;Li et al. 2016).However, high laser energy input in this process could cause a violent flow of liquid, which could lead to gas trapped in the melt pool, producing many pore defects and impairing the mechanical properties (Galy et al. 2018).In addition, rapid cooling rate and large thermal gradients in the melt pools can induce large residual stress/distortion, resulting in parts failure (Li et al. 2018;Parry, Ashcroft, and Wildman 2016).The top-down directional heat dissipation during the forming process also promoted epitaxial growth of columnar grains and formed strong crystallographic texture (Liu and To 2017).All these factors posed a great challenge to the enhancement of additively manufactured aluminium alloy performance.
Several researches have been conducted to reduce the defects and improve the performance of LPBFprinted aluminium alloy parts from material modification, process optimisation and post-treatment, etc.By adding trace elements, the performance was greatly improved by promoting grain refinement and generating special strengthening phases, which has been confirmed in Zr-modified Al-Cu-Mg alloy (Zhang et al. 2017), Er-modified 7075 aluminium alloy (Zhang et al. 2021), and Zr-modified 7075 aluminium alloy (Yu et al. 2022).For example, the addition of 2 wt.%Er could reduce the cracking in the LPBF-printed AA7075 and promote the generation of Al 3 Er precipitates that facilitate grain refinement (Zhang et al. 2021).Process optimisation could also overcome the metallurgical defects and improve the performance, which mainly focused on the laser process parameters (Gupta et al. 2020), powder characteristics (Muniz-Lerma et al. 2018) and substrate preheating (Buchbinder et al. 2014).Gupta et al. (2020) investigated the effect of layer rotation on the properties of fabricated AlSi10Mg alloy, and pointed out that appropriate layer rotation method could effectively decrease the number of pores and residual stresses.Additionally, post-treatment, mainly including the heat treatment and hot isostatic pressure, provides another easy and effective solution to close pores, refine grains and improve mechanical properties (Uddin et al. 2018;Wang et al. 2019;Sun et al. 2021).Wang et al. (2019) annealed the LPBF-printed AlSi7Mg parts, and found that the average grain size decreased from 3.09 to 2.90 μm, contributing to the increase of elongation at break from 9.21% to 12.87%.
Although the above methods could reduce metallurgical defects and improve properties, the additional cost requires and the composition of material also changed.In recent years, it has been found that applying static magnetic field (SMF) is an effective method to affect metal solidification and improve formability in casting (Hu et al. 2021;Dong et al. 2020;Li et al. 2009b), welding (Huang et al. 2020a;Huang et al. 2020b) and heat treatment (Akhbarizadeh, Amini, and Javadpour 2012) processes.Based on electromagnetic theory, both Lorentz force and thermal electromagnetic force (TEMF) could be generated with SMF.Various interesting solidification phenomena have been identified, such as changes in the crystal growth method, defects, texturing and orientation.Moreover, the convection of liquid phase could also be effectively influenced by SMF.Li et al. (2009a) introduced a high magnetic field in the process of directional solidification of the Al-0.85 wt.% Cu and Zn-2.0 wt.% Cu alloys, which contributed to the breakdown of a planar interface into cellular undulations and the formation of an irregular shape and the enrichment of the solute Cu element in the diffusion boundary layer.Huang et al. (2020b) found that the profile of the resistance spot welding (RSW) nugget became more regular and freer of the internal shrinkage defects and hot cracks with the action of the external magnetic field.To date, some studies have been conducted to introduce magnetic fields into LPBF and to investigate their mechanism of action (Du et al. 2019;Zhao et al. 2022a;Kang et al. 2017a;Zhou et al. 2022;Kang et al. 2017b;Zhu et al. 2021).Zhou et al. (2022) found that the epitaxial growth and <001> crystallographic texture along the building direction were inhibited and cellular dendrites were deflected from the solidification direction by TEMF.Zhao et al. (2022a) regulated the microstructure with twisted prior-β grains and discontinuous α grain boundaries by means of SMF.However, the research on the application of SMF during the LPBF process still remained scarce.Before this paper, AlSi7Mg has not been used as a material for the relevant research.Additionally, the mechanisms of the effect of SMF on defects, columnar-to-equiaxed transition, and performance optimisation were studied not deep enough.
As mentioned above, on-line SMF could play a beneficial role in the liquid solidification.AlSi7Mg is capable of producing significant thermoelectric effect due to its excellent electrical conductivity and thermal conductivity, which is very suitable for researching the role of on-line SMF during the LPBF process.In summary, the objective of this study was to investigate the effects of on-line SMF on the defects, microstructures and mechanical properties of AlSi7Mg alloy fabricated by LPBF; to reveal the extent and mechanism of the effect of on-line SMF.This study could provide a new method for quality control of LPBFprinted aluminium alloy and other alloys.

Materials preparation
The gas-atomized AlSi7Mg alloy powder was used (Jiangsu Vilory Advanced Materials Technology Co., Ltd, China).The powder morphologies were in a spherical shape (Figure 1 (a)), examined by scanning electron microscopy (SEM, Quanta650 FEG, FEI, USA).The powder size measured by laser diffraction particle size analyzer (Mastersizer 3000, Malvern, UK) ranged 13-59 μm in diameter with an average diameter of 27.4 μm (Figure 1(b)).The chemical composition of the powder, which was determined by Xray fluorescence (XRF-1800, Shimadzu, Japan), was listed in Table 1.In summary, the morphology and properties of the alloy powder met the LPBF manufacturing requirements.

Experimental equipment
A self-developed LPBF150 machine with a 150 × 150 mm 2 building area was used in this experiment, equipped with a continuous wave IPG YLR-500 singlemode fibre laser, a German SCANLAB galvanometer and a two-dimensional focusing F-θ lens.The laser beam had a Gaussian profile with a spot diameter of 100 μm and a wavelength of 1060 nm.An on-line SMF device was set up on this machine to generate a transverse magnetic field (parallel to the X-direction in Figure 2(a)).Several permanent magnets (40 × 20 × 10 mm 3 ) were placed on the setting plate and fixed symmetrically on both sides of the substrate which was fixed in the centre of the setting plate (Figure 2

LPBF process
Cube-shaped samples (6 × 6 × 5 mm 3 ) and tensile samples (Figure 2(d)) were fabricated with the specific process parameters, that is, laser power of 450 W, laser scanning speed of 600 mm/s, layer thickness of 30 μm, and hatch distance of 75 μm.The raster pattern was rotated by 67°following the deposition of each layer (Figure 2(c)).The substrate had no preheating and high-purity (99.99%) argon gas was continuously maintained into the fabrication chamber to prevent sample oxidation.Notably, the process parameters used were not the most appropriate, but were in the vicinity of them.This operation was intended to make it easier to zoom in and analyse the mechanism of SMF during the LPBF process.Furthermore, the detailed reasons are as follows: (1) the optimal process parameters hardly reflected the effect of on-line SMF on microstructure solidification, because almost no metallurgical defects were generated; (2) higher temperature gradient was more conducive to be obtained by our specific process parameters (higher laser energy input), and thus more significant thermoelectric magnetic effect could be induced.

Characterisations and mechanical testing
The relative density of the sample was tested by Archimedes' principle and the theoretical density of AlSi7Mg is 2.70 g/cm 3 (Muniz-Lerma et al. 2018).The phase analysis of the samples was identified by X-ray diffractometer (XRD-7000S, Shimadzu, Japan) with a Cu tube at 40 keV and 30 mA.The 2θ varied from 20 to 100°with a step size of 10°/min.Surface morphologies were observed using optical microscopy (OM, Leica DM750M, Germany) and SEM.Internal defects morphologies, distributions and volume fraction were analysed by using micro-computed tomography (micro-CT, ZEISS XradiaContext MicroCT1, Germany) with an isotropic voxel size of 7 µm.The micro-CT scan images were processed on Avizo software.Prior to microstructure observation, the samples were etched with Keller reagent (95 vol.%H 2 O, 2.5 vol.%HNO 3 , 1.5 vol.%HCl and 1 vol.%HF).The crystallographic orientation, size and texture were determined by electron backscattered diffraction (EBSD).The EBSD data were processed on AztecCrystal software.Moreover, every grain shape was fitted as an ellipse, which facilitated the study of the mechanism of magnetic field on the grain morphology    and columnar-to-equiaxed transition (CET) by analysing the aspect ratio (E) of the ellipse.Field emission transmission electron microscope (FTEM, Tecnai G2 F30, Netherlands) was utilised to further analyse the crystallographic structure of precipitates, lattice spacing and dislocations an accelerating voltage of 300 kV.Further precise identification of the precipitated phase was carried out by the selected area electron diffraction (SAED), high-angle annular dark-field (HAADF) and high-resolution TEM (HRTEM).Tensile tests were conducted with a strain rate of 1.11 × 10 −3 /s at room temperature using a mechanical testing machine (Zwick/Roell Z020, Germany).Three samples per magnetic field intensity were tested to obtain the ultimate tensile strength (UTS) and elongation at break.

Pores and microstructures
Figure 3 presents the cross-sectional morphologies of the LPBF-printed AlSi7Mg observed by OM.The results show that the severity of pore defects reduced significantly with the magnetic field application.A large number of pore defects, which could be divided into micropores and keyholes (Xing et al. 2019;Zhao et al. 2020), were observed inside the sample without SMF.Interestingly, when SMF was applied, the number of pores reduced and the size of pores became tiny with the increase of magnetic field intensity.In particular, near-absence of pores was observed inside the sample fabricated with 0.3 T SMF.Moreover, the gradual increase in relative density of samples from 96.9% to 98.6% also provided sufficient evidence for the reduction of pore defects.Furthermore, micro-CT was used to observe the transition of the pore defects in three-dimensional space, including morphologies, distributions and volume fraction.As shown in Figure 4(a), the pores of the sample fabricated without SMF were mainly keyholes with a maximum volume of 0.070 mm 3 , which exhibited spoon-like morphology.Significantly, in Figure 4(b), the pores of the sample fabricated with 0.3 T SMF were evolved from keyholes to micropores with a maximum volume of 0.019 mm 3 .Meanwhile, the volume fraction of pores decreased from 2.7% to 0.5% with the application of SMF.
Figure 5 illustrates the microstructures of AlSi7Mg samples prepared by LPBF without and with SMF observed by SEM.The arc-shaped melt pool boundaries were obviously observed.Since the solidification rate at the melt pool boundary was greater than that inside the melt pool, a large number of coarse grains were presented near the boundary.Similar results are seen in nickel-based superalloys (Sun et al. 2021), copper alloys (Li et al. 2022), etc.In order to ensure the reliability of the experimental results, the observation area with high magnification was selected from the centre of the melt pool.Unlike a mass of Al-Si eutectic structure fabricated by casting process (Fan et al. 2013), part of Si atoms were directly dissolved into the Al matrix during LPBF, while the rest of Si gathered together to form a network structure distributed at grain and subgrain boundaries due to the rapid cooling rate (Dong et al. 2015;Birol 2007).The shape of eutectic Si network gradually changed from columnar to equiaxed after applying SMF, and the degree of variation was more obvious with the increase of magnetic field intensity.The shape change of the eutectic Si network might be caused by the transition of grains.In other words, CET of grains occurred with the action of SMF.
Figure 6 displays the inverse pole figure (IPF) colour maps and pole figures (PFs) from the YZ plane of samples observed by EBSD.The grain growth direction was obviously changed (Figure 6(a-d)), as well as the microstructures were clearly controlled with the application of SMF.Due to the bottom-up heat dissipation characteristics of the melt pools, the grains without SMF grew mainly along <001> direction, which was caused by the excellent thermal conductivity of the substrate.Obviously, the number of grains along <001> direction decreased with 0.1 T SMF, whereas the number of grains growing along <110>,<111> directions increased.Especially, this trend became more obvious with the increase of magnetic field intensity.It can be seen that the grains grew almost uniformly along <001>,<110>, <111> directions when the magnetic field intensity was up to 0.3 T. This was further confirmed by the weakening of the crystallographic texture, as shown in Figure 6(e-h).The texture index reduced from 3.44 to 2.58, up to 25%, with the increase of magnetic field intensity.This also indicated that the crystallographic orientation gradually became uniform, which was consistent with the analyses of IPF colour maps.
Furthermore, grain size and degree of CET were analysed as shown in Figure 7. Figure 7(a-d) illustrate the grain size distribution histograms derived from the above EBSD results.The maximum grain size exceeded 70 μm, as well as the average grain size was ∼8.35 μm without SMF.When SMF was applied, the maximum grain size was down to 50 μm, and the average grain size was reduced to ∼8.07 μm, ∼7.35 μm and ∼7.22 μm with the magnetic field intensity from 0.1 to 0.3 T, respectively.Additionally, Figure 7(e-h) illustrate the fit ellipse aspect ratio (E), which could directly manifest the effect of the magnetic field on grain morphology and the degree of CET by fitting the grain morphology to an ellipse.Here, equiaxed grains were present when E < 3, whereas columnar grains occurred when E ≥ 3. When there was no magnetic field, approximately 56% of grains were equiaxed.Nevertheless, when SMF was applied, the proportion of equiaxed grains gradually increased to 58%, 64% and 68% with the magnetic field intensity from 0.1 to 0.3 T, respectively.The results directly indicated that SMF could effectively promote CET, and this transition effect increased with the increase of magnetic field intensity.The reasons for these would be discussed detailedly in Section 4.

Phase identification
Figure 8 identifies the phase compositions of the samples prepared without and with SMF analysed by XRD.For samples without SMF, it was mainly composed of Al, Si and a small amount of Mg 2 Si phase that is an important strengthening phase in AlSi7Mg alloy.It was noteworthy that the phase composition was not changed by the application of the SMF.However, by magnifying the peak, it could obviously be observed that the peak (311) of Al matrix shifted to the right after applying SMF, and moved to the right continuously with the increase of magnetic field intensity (Figure 8 (b)).According to the Bragg's law (Li et al. 2015), the right shift of the peak angle means interplanar spacing decreases.In other words, the introduction of the magnetic field promoted the precipitation of Si and Mg atoms out of the matrix and formed Mg 2 Si, which might play a key role in the strengthening of the alloy.
Figure 9 shows the TEM results for the LPBF-printed AlSi7Mg alloy with and without SMF in the YZ plane.In the bright-filed images (Figure 9(a,c)), nano-precipitates were observed around the subgrain boundaries.High-density dislocation tangles were concentrated around the subgrain boundaries and nano-precipitates.After applying SMF, it was worth noting that more nano-precipitates were precipitated around the subgrain boundaries.The corresponding SAED image (Figure 9(b)) indicated that the matrix was the Al phase with an FCC lattice structure.The microstructures of phases interface between the matrix and nano-precipitates were observed by HRTEM (Figure 9(d)).According to the Fast Fourier transform (FFT) and inverse FFT, the interplanar spacings of these two phases were measured as 0.233 nm for (111) Al and 0.317 nm for (200) Mg 2 Si.Furthermore, the energy-dispersive X-ray spectroscopy (EDS) maps (Figure 9(f)) also demonstrated that amount of Mg 2 Si was precipitated from the Al matrix around subgrain boundaries.Additionally, smaller magnitude Mg 2 Si clusters (5-10 nm) were detected to be dispersedly and uniformly distributed inside the cells (Figure 9(e)).Its formation might be due to the rapid solidification characteristics of the LPBF process and the application of SMF.The dispersed distribution of nanoclusters was beneficial to improve mechanical properties.

Tensile properties
The results of tensile properties among the samples printed with and without SMF are shown in Figure 10.Obviously, the application of magnetic field enhanced the strength and ductility simultaneously.The ultimate tensile strength of the sample fabricated without SMF was (326.67 ± 5.31) MPa, which was worse than the sample fabricated with the optimised parameter (Wang et al. 2019).As mentioned in Section 2.3, the specific process parameter chosen in this study could contribute to investigate the effect of SMF but would impair the tensile properties.Compared with the samples fabricated without SMF, the samples prepared with SMF exhibited an excellent combination of strength and ductility.The ultimate tensile strength was enhanced by 9.2%, 13.1%, and 16.9% with the increase of the magnetic field intensity, respectively.Simultaneously, the elongation at break also improved by 27.2%, 30.7%, and 38.9%.According to the above results of the relative density, microstructure and tensile properties of AlSi7Mg fabricated under different magnetic field intensity, the transitions were concluded in Table 2.   To further investigate the tensile properties enhancement mechanism, the fracture surfaces of samples fabricated without and with 0.3 T SMF were observed by SEM, as shown in Figure 11.Some un-melted powder and pores could be observed in the fracture surface without SMF (Figure 11(a,b)).These defects could become fracture sources and severely diminish the tensile properties.A local magnification in Figure 11(c) shows that the fracture morphology consisted of a large area of fracture cleavage plane and a small proportion of dimples, indicating that the tensile fracture mode of the samples without SMF was brittle failure.In contrast, the number of defects on the fracture surface decreased obviously when SMF was applied (Figure 11(d,e)).In addition, the local magnification (Figure 11(f)) shows that dimples were dominant and the cleavage plane was almost invisible, which indicated that the tensile fracture mode of the sample with SMF was mainly ductile failure.

Discussion
The above results indicated that the application of SMF could significantly contribute to the reduction of pore defects, the promotion of CET, and the improvement on both strength and ductility.These enhancements were the result of the combined effect of Lorentz force and TEMF produced by the induced currents in the magnetic field.The induced currents were caused by the complex flow and temperature fields in the melt pool during solidification.It was worth noting that SMF has a certain effect on both liquid phase and solid phase, which was discussed in detail below based on the above results.

Effect of on-line SMF on the flow and temperature fields of liquid phase
When SMF was applied, on one hand, according to the Seebeck effect (Uchida et al. 2008;DiSalvo 1999), induced thermal currents would be generated at the solid-liquid interface and between liquid phases due to the high temperature gradient during the LPBF process (Figure 12(b,c)).These induced thermal currents through the liquid phase created TEMF with the action of SMF, which could create thermal electromagnetic convection (TEMC) that facilitated the liquid flow (Zhu et al. 2021).On the other hand, nevertheless, the flow of liquid would be impeded due to the Lorentz force with the action of magnetic field (Du et al. 2019).The Lorentz force and TEMF acted in opposite aspects, competing with each other.Li et al. (2009b) and Lehmann et al. (1998) defined the Hartmann constant (Ha) to determine which factor was dominant, as described below: where B, L, σ L and μ, denote magnetic field intensity, characteristic length, electrical conductivity and dynamic viscosity, respectively.When Ha > 10, the effect of Lorentz force dominates, conversely, the effect of TEMF dominates.For the liquid phase of AlSi7Mg alloy, σ L is approximately equal to 3.65 × 10 6 Ω −1 m -1 when the temperature is at 650°C calculated by JMatPro software (Figure 13), and μ is equal to 1.146 × 10 Willers et al. 2008).In this study, L can be simply set to 1 × 10 −4 m, which is consistent with the width of melt pool.When the magnetic field intensity B reaches the maximum, that is, B = 0.3 T, it could be calculated from the formula (1) that Ha = 1.5 (less than 10).This indicated that the effect of TEMF dominated under this condition, which could promote the flow of liquid regularly.In this way, the flow field would become more orderly and the temperature field became relatively uniform in the melt pool (Figure 12 (b,c)).Zhu et al. (2021) verified the same results when magnetic fields were applied to LPBF processing Inconel 625 superalloy by numerical simulation.
As mentioned above, the number of pore defects was significantly reduced with the application of on-line SMF.The formation of micropores might be caused by gas, which was from the hollow powder or protective atmosphere, trapped in the melt pool during solidification.The keyhole, a deep and narrow pore, caused by the strong recoil pressure from the rapid evaporation of metal which could push the surrounding melt liquid downward (Zhao et al. 2020).Not only the micropores but also the keyholes were affected by the Marangoni convection, which was a mass transfer along the interface between two fluids due to the gradient of surface tension.Chen et al. (2018) and Verhaeghe et al. (2009) found that higher laser power input could generate more drastic Marangoni convection.In this study, severe Marangoni convection was caused by the high laser energy density, leading to a large number of pores.As shown in Figure 12(b,c), when the SMF was applied, the liquid flow intensity was stronger but orderly so that less gas was trapped into the melt pool.Moreover, bubbles were more easily diffused to the surface of the melt pool and escaped from it.Hence, the application of SMF during the LPBF process can help to reduce the number of pores correspondingly.
Furthermore, the introduction of the magnetic field also affected the precipitation of strengthening phase.Due to the rapid cooling rate during the LPBF without SMF, Si and Mg atoms have no sufficient time to precipitate from Al matrix.But when the SMF was applied, the temperature gradient became lower and the cooling rate decreased because of the stirring effect of TEMF, which provided more time for Si and Mg atoms to escape from Al matrix and form Mg 2 Si.This precipitation of Si and Mg atoms created many vacancies causing lattice distortion, which led to the peak (311) of Al matrix being shifted to the right (Figure 8(b)).

Effect of on-line static magnetic field on CET in the mushy zone
According to Fleming's rule (i.e.left-hand rule), the induced thermal currents flowing through the columnar grains would produce TEMFs acting nearly perpendicular to these columnar grains.Moreover, these TEMFs could twist or fracture columnar grains in turn (Figure 14).The value of TEMF could be estimated by formula (2): where F TE is the TEMF density, J TE is the current density, B is the magnetic field intensity, σ L is the electrical conductivity of liquid phase, σ S is the electrical conductivity of solid phase, f L is the volume fraction of liquid phase, f S is the volume fraction of solid phase, η L is the thermoelectric power of liquid phase, η S is the thermoelectric power of solid phase, and ΔT is the temperature gradient.According to the simulation results (Figure 13), σ L ≈ 3.65 × 10 6 Ω −1 m -1 (T = 650°C), σ s ≈ 1.05 × 10 7 Ω −1 m -1 (T = 400°C), η L ≈ 1.05 × 10 −6 V/K, η S ≈ 1.5 × 10 −6 V/K, f S = f L = 50% and ΔT ≈ 10 6 K/ m.Then, F TE could be calculated to be 2.03 × 10 6 N/m 3 when B is 0.1 T. Obviously, this force was strong enough to destroy the solidification front of the columnar grain in the mushy zone, which has been confirmed in previous studies by numerical simulation (Du et al. 2019;Zhu et al. 2021).Subsequently, this crushing effect would become more pronounced with the increase of magnetic field intensity, due to the positive correlation between F TE and B.
As mentioned above (Figure 6), the crystallographic orientation gradually changed from <001> to <110> and <111>.On one hand, TEMF could not only change  the characteristic of the heat dissipation from up to bottom, but also reduce the temperature gradient by the stirring action of TEMC, which effectively restrained the epitaxial growth of columnar crystals along <001> direction and reduced the texture index.On the other hand, as shown in Figure 14(c), the TEMF acting on the columnar grains was strong enough to twist or fracture columnar grains, which led to the crystallographic orientation changed from <001> to <110> and <111>.Meanwhile, remarkable CET also occurred with the application of on-line SMF.Based on the above discussions, it is clear that a large number of grain fragments were created by the broken effect.Due to the liquid phase with higher temperature, the grain fragments split into more tinier fragments which would become the Nuclei-formation particles of heterogeneous nucleation, promoting the columnar-to-equiaxed transition of grains (Figure 14(b-d)).

Performance enhancement mechanisms
The samples prepared by LPBF with SMF exhibited higher both strength and ductility compared with those without SMF.Figure 15 is a series of schematic illustrations showing the YZ plane morphologies for the LPBF-printed AlSi7Mg with and without SMF at multiple scales.The simultaneous increase in strength and ductility may be attributed to the combined effects of pores elimination, columnar-to-equiaxed transition and Mg 2 Si precipitation strengthening.
As mentioned above, the liquid phase flow in the melt pool was more orderly with the action of TEMF, which suppressed the formation of the pore defects (Figure 15(a,d)).The pores defects not only reduced the cross-sectional area leading to an increase in actual stress, but also caused stress concentrations near the pores.The stress concentration mechanism near pores caused initial cracks, which was the source of fracture.The crack propagated around the pores during the tensile process, resulting in the decrease of tensile strength.Hence, the pores elimination due to the application of SMF contributed to enhance both the strength and ductility.
Generally, the slip and accumulation of dislocations are the main mechanisms of macroscopic plastic deformation of the alloy.Grain boundaries or precipitated phases are the main obstacles to dislocation slip, and a large amount of plugging in their vicinity can effectively increase the strength.In this study, the TEMF acting on the columnar grains after the application of SMF was sufficient to twist or break them, which induced the obtaining of sufficiently refined grains and uniform grain orientation (Figure 15(b,e)).In other words, this CET mechanism increased the number of grain boundaries, which could impede dislocation motion and thus lead to enhanced mechanical properties.
Additionally, the temperature field in melt pool became more uniform with the action of SMF, promoting the reduction of solid solution of Mg and Si in the Al matrix.More Mg 2 Si precipitates were generated from the Al matrix and distributed at grain and subgrain boundaries.Furthermore, nano-Mg 2 Si clusters (5-10 nm) were precipitated and dispersedly distributed inside the cells (Figure 15(c,f)).As shown in TEM images (Figure 9), non-coherent interfacial relations in (111) Al and (200) Mg 2 Si phases would contribute to dislocation plugging during plastic deformation according to Orowan strengthening mechanism.The dislocations would bypass the precipitated phase when slipping to the phase interface, generating a dislocation loop near the precipitated phase.Obviously, these Mg 2 Si  precipitates also contributed a lot in enhancing the mechanical properties, especially the diffusely distributed nano-Mg 2 Si clusters.

Conclusions
Simultaneously enhanced strength and ductility of AlSi7Mg alloy were prepared by LPBF with static magnetic field.The effects of magnetic field with different intensities on microstructures and tensile properties were studied thoroughly.The mechanisms of pore inhibition, Mg 2 Si phase precipitation and CET were discussed in depth.The main results and conclusions are summarised as follows: (1) The number of pores in samples prepared by LPBF with on-line SMF decreased obviously, which was related to the effect of static magnetic field on the liquid flow in the melt pool.According to Hartmann constant, TEMF generated with the action of SMF dominates when the magnetic field intensity ranges from 0.1 to 0.3 T. Consequently, the liquid flow was strengthened but orderly, which was more conducive to the overflow of the involved gas, thus reducing the pore defects.
(2) The crystallographic orientation changed from <001> to <110> and <111> under the stirring action of TEMC that was driven by TEMF.Consequently, the texture index decreased from 3.44 to 2.58.Meanwhile, on-line static magnetic field could refine the grain size.The average grain size decreased from 8.35 to 7.22 μm and the percentage of equiaxed grains increased from 56% to 68%.TEMF acting on the columnar grains fractured them, promoting the CET of grains.Additionally, no new phase was generated with the action of SMF, but more Si and Mg atoms precipitated from Al matrix and formed Mg 2 Si, which was attributed to the relatively uniform temperature field in the melt pool caused by the stirring action of TEMC.
(3) The samples prepared by LPBF with SMF exhibited higher strength and ductility compared with those without SMF.The ultimate tensile strength respectively enhanced by 9.2%, 13.1%, and 16.9% with the increase of the magnetic field intensity.Simultaneously, the elongation at break improved by 27.2%, 30.7%, and 38.9%.The simultaneous increase in strength and ductility might be attributed to the combined influences of defect elimination, CET and Mg 2 Si precipitation strengthening.
(a,b)).By adjusting the width of the substrate (d, i.e. the spacing between the permanent magnets on both sides) and the number of the permanent magnets (n), magnetic fields with different intensities (B) can be obtained (such as, d = 15 mm, n = 10, B = 0.3 T; d = 25 mm, n = 10, B = 0.2 T; d = 25 mm, n = 4, B = 0.1 T; d = 15 mm, n = 0, B = 0 T).The magnetic field intensities at the centre of the substrate were measured by a gauss metre (TD8620, Tunkia, China) before the printing experiment to ensure the accuracy of the data.

Figure 1 .
Figure 1.Microscopic properties of AlSi7Mg powder: (a) SEM images of the powder morphologies; (b) particle size distribution in volume.

Figure 2 .
Figure 2. On-line SMF-LPBF equipment and tensile samples: (a) two-dimensional illustration of on-line SMF-LPBF equipment; (b) physical map of on-line SMF-LPBF equipment; (c) laser scanning strategy; (d) the tensile sample dimensions.

Figure 3 .
Figure 3. OM morphologies in the YZ plane of the LPBF-printed AlSi7Mg with and without SMF: (a) 0 T; (b) 0.1 T; (c) 0.2 T; (d) 0.3 T, where yellow and blue arrows depict micropores and keyholes respectively.

Figure 6 .
Figure 6.EBSD results of the LPBF-printed AlSi7Mg with and without SMF: (a-d) IPF colour maps from YZ plane without SMF of 0 T and with 0.1, 0.2, 0.3 T, respectively; (e-h) PFs from the YZ plane of a-d.

Figure 7 .
Figure 7. Grain size distribution histograms and fit ellipse aspect ratio distribution histograms of the LPBF-printed AlSi7Mg with and without SMF: (a-d) grain size distribution histograms; (e-h) fit ellipse aspect ratio distribution histograms.

Figure 8 .
Figure 8. XRD analysis of the LPBF-printed AlSi7Mg with and without SMF: (a) spectra data of 2θ from 20.0 to 100.0°;(b) spectra data of 2θ from 77.5 to 79.0°.

Figure 9 .
Figure 9. TEM images showing phases and element distributions of the LPBF-printed AlSi7Mg (a and b without SMF, c-f with 0.3 T SMF): (a) bright-field image; (b) SAED patterns of the yellow circle region in (a); (c) bright-field image of 0.3 T sample; (d) HRTEM image of the Al matrix and Mg 2 Si at subgrain boundary, and the inset images show the IFFT applied filtering in the corresponding regions; (e) HRTEM image inside the cell; (f) HADDF image, and EDS maps for Al, Si, and Mg.

Figure 10 .
Figure 10.Tensile properties of the LPBF-printed AlSi7Mg with and without SMF: (a) engineering stress-strain curves; (b) average ultimate tensile stress and elongation at break.

Figure 11 .
Figure 11.SEM morphologies of the fracture surface of the LPBF-printed AlSi7Mg with and without SMF: (a-c) 0 T; (d-f) 0.3 T.

Figure 13 .
Figure 13.Electrical conductivity of AlSi7Mg alloy at different temperatures calculated by JMatPro software.

Figure 14 .
Figure 14.Schematic illustrations of the SMF effect on columnar-to-equiaxed transition during the LPBF process: (a) columnar grain growth without SMF; (b) generated TEMF acting on columnar grain; (c) the process of columnar-to-equiaxed transition with the action of TEMF; (d) the consequence of the action of SMF acting on columnar grain.

Table 2 .
Transition of relative density, microstructures and tensile properties of AlSi7Mg printed under different static magnetic field intensity.