Metallic 4D Printing of Laser Stimulation

Abstract 4D printing of metallic shape‐morphing systems can be applied in many fields, including aerospace, smart manufacturing, naval equipment, and biomedical engineering. The existing forming materials for metallic 4D printing are still very limited except shape memory alloys. Herein, a 4D printing method to endow non‐shape‐memory metallic materials with active properties is presented, which could overcome the shape‐forming limitation of traditional material processing technologies. The thermal stress spatial control of 316L stainless steel forming parts is achieved by programming the processing parameters during a laser powder bed fusion (LPBF) process. The printed parts can realize the shape changing of selected areas during or after forming process owing to stress release generated. It is demonstrated that complex metallic shape‐morphing structures can be manufactured by this method. The principles of printing parameters programmed and thermal stress pre‐set are also applicable to other thermoforming materials and additive manufacturing processes, which can expand not only the materials used for 4D printing but also the applications of 4D printing technologies.

lead to an increase in the bending angle of the sample. [2] However, when the laser energy density is high enough (200-300 W, 1600-400 mm/s), the bending angle will be decreased until no bending occurs. Because as the energy density increases, more powder is melting, which will reduce the porosity of the sample and increases the support strength, and makes it difficult for thermal stress to separate them. [3] (ⅳ) The support thickness has a direct impact on the bending angle. When the support thickness is smaller (0.4-0 mm), the substrate has a higher restraint on the support, and when the support thickness is higher (0.4-0.6 mm), the support strength will be increased and the support will not be easily damaged. Both of the above will reduce the bending angle. [4] From the analysis of the properties of the sample itself, for a given printing process, the bending angle can be controlled by the size and geometric shape. (ⅴ) The influence of the sample width on the bending angle is not obvious (3-9 mm). The support area will increase with the increase of the sample width, which leads to an increase in the support strength. [4] At the same time, the laser will scan the surface for more time and generates more thermal stress, which offsets the above effect. (ⅵ) Conversely, the sample length has a significant effect on the bending angle. If the length is too short (3-6 mm), the displacement of the sample along the vertical direction will become more difficult. If the length is too long (6-9 mm), the area of the support along the length will increase. Both of the above will eventually decrease the bending angle. [4] (ⅶ) It can be seen that the bending angle is inversely proportional to the sample angle (15-90 °). Because when the length of the bottom side is constant, the area of the support around the corner becomes larger with the increase of the sample angle, which makes it difficult for the support to be damaged. [4]

S3. Finite element analysis
The finite element analysis is performed on the warping simulation of the sample to show the mechanism of the deformation process. (Figure 3a,b and Movie S2, Supporting Information).
Due to the high computational cost of the track-by-track deposition in FEM (finite element method), quantitatively simulating the whole manufacturing process of each case is unrealistic.
As the parts are thin-wall structure, the strain level on the free surface and near base plane are quite different. Thus, the warping is modelled based on the inherent strain method.
The inherent strain can be expressed mathematically as： [5] (2) Which can also be rearranged as： represents the elastic strain in steady state. [5] Since this simulation aims to qualitatively analyze the deformation trend, the specific history of mechanical strain calculation appears to be unnecessary. By assigning different inherent strain values into the elements of the parts according to its layers, the part distortion can be simulated. Due to the continuous scanning of the laser, the attribution of the inherent strain is changed from tension to compression, and from top surface to the base ground. For each case, the agglomeration approach is utilized to merge several layers together as a whole block with the attribution of different strains assigned to different regions according to their position. In this way the whole part is separated into 4 different blocks and the orthotropic base inherent strain component is set to be (-0.0045, -0.0045, 0.008). [5] Normally, due to the faster cooling rate on the top surface and energy accumulation in the lower layers, the strain state should be tensive on the top and compressive at the bottom. [6] Therefore, the coefficients for the 4 layers are chosen to be 0.03, 0.01, -0.01, -0.03 from top to bottom. According to reference， [5] the way to assign the inherent strains are through the equations for the thermal strains calculation in j th direction： [5] (4) Where is the equivalent thermal expansion coefficient, is the temperature change set to be unity in this work.
Noted: the young's modulus and Poisson's ratio does affect the distortion, since the strain is determined by the attribution, but they will affect the stress level. This work only focuses on the distortion. Figure S1. FEA analysis results of stress and displacement for the base model. The equilateral triangle sample with the laser scans along the X direction (a), and the two rectangular samples with the laser scans along the X (b) and Y (c) directions.

Discussion:
The laser scanning strategy has a direct effect on the residual stress and deformation of the sample. [1][2]7] As shown in the triangular sample (a), the laser scanning along the X direction causes the maximum stress on both sides of the sample along the X direction, and also has a large displacement at the corners. As shown in the rectangular samples (b and c), when the laser scanning along the X direction, that is, along the length direction, the stress on both sides of the sample in the X direction is the largest and has a large displacement; when the laser scanning along the Y direction, that is, perpendicular to the length direction, the stress on both sides of the sample in the Y direction is the largest and there is a small displacement on both sides in the X direction. Therefore, it can be seen from the above simulation that the residual stress of the sample is the largest on both sides of the laser scanning path, and the sample will also be deformed greatly in this direction.   Here, the scanning speed, support thickness, sample width, sample length and sample thickness are fixed, which are 1600 mm/s, 0.4 mm, 6 mm, 6 mm and 0.6 mm, respectively.
Scale bar: 5 mm.       Discussion: This method can be used to print some special structures, such as the English abbreviation of the Jilin University --"JLU". As shown in Figure S11, the designed 2D "JLU" precursor is finally transformed into a 3D structure with a certain bending angle by laser stimulation, which can stand on the desktop. Figure S12. Specific designs of four rectangular (a) and triangular (b) 2D precursors and their shape-morphing 3D structures. Scale bars: 5 mm.

Discussion:
As for the same 2D precursor shape, different specific designs can generate different structural transformation effects. As shown in Figure S12, when the 2D precursor models are all (a) rectangular or (b) triangular, the specific designs of the four 2D precursors can result in distinct 3D structures with the laser stimulation. Figure S13. Model and process parameters of the bioinspired frog tongue (Figure 3c). Here, the overall length is 36 mm, and the left part is rectangular, 16 mm and printed directly; the right part is approximately triangular, 20 mm and printed with support. According to the results in Figure 1d, set the laser power is 200 W, the scanning speed is 1600 mm/s, the sample width is 5 mm, the sample angle is 10 °, and the support thickness is 0.4 mm.  structures. The compression speed is 1 mm/min, and the process parameters used for the two structures are the same: the laser power is 200 W, the scanning speed is 1600 mm/s, and the sample thickness is 0.6 mm. The maximum compressive force that the rectangular structure can bear is 202.137 ± 6.356 N, and the maximum compressive force that the bridge structure can bear is 212.867 ± 40.739 N. Scale bars: 5 mm.