Additive manufacturing of ultralong and architected single crystals

a, Model set-up. O represents the origin of coordinates for the heat source model. b–e, The temperature (T) distribution and melt-pool shape of the upper region (marked by the dashed blue box in a) at different times marked in the current–time curves. The 3D view is used to verify the melt-pool shape, whereas the mid-section view is used to reveal the solid–liquid interface characteristics. The dotted white lines indicate the solid–liquid interface. b, Deep-squat pulsed heat input mode with peak current I P = 200 A and base current I B = 20 A. The simulated melt-pool shapes match well with experimental observations at times such as t 1 and t 3 . The mid-section at t 2 shows a shallow, wide melt pool with a convex solid–liquid interface, which helps eliminate grains. During cooling, although the edge region may not keep a perfectly convex shape (as seen at t 3 ), it will remelt during the deposition of the next layer owing to the thin layer thickness (about 0.15 mm), permitting effective grain elimination. c, For a constant intermediate current of 29 A, as deposition continues, the temperature field gradually exhibits quasi-steady-state characteristics, so no specific time is marked. The melt-pool morphology aligns with experiments. Compared with b, the melt pool is deeper and maintains a nearly flat (concave) solid–liquid interface. d, The current is kept constant at its peak value I P and excessive heat input causes structural collapse. e, The current is kept constant at its base value I B . The wire cannot be melted smoothly, even though the wire diameter and feeding speed are the same as b. The phenomena of melt-pool collapse in d and unmelted wire in e are in good agreement with the infrared experimental results.