July_August_AMP_Digital

A D V A N C E D M A T E R I A L S & P R O C E S S E S | J U L Y / A U G U S T 2 0 1 9 1 8 of melt pool geometry in a way that is alloy-specific, by first carrying out pre- dictions of melt pool dimensions us- ing finite element methods within the COMSOL Multiphysics heat transfer module. Simulations of the thermal model are then validated through ex- perimental temperaturemeasurements using in situ pyrometry and SEM char- acterization of single-track melt pool cross sections. The thermal model in- cludes phase-dependent thermo-phys- ical properties of the powder feedstock, e.g., thermal conductivity, specific heat capacity, and absorptivity. These prop- erties are used to approximate heat and mass transport phenomena such as melting, solidification, vaporization, and keyhole formation. By combin- ing best estimates of thermo-physical properties with high-fidelity thermal models, the printability map of the de- sired alloy is generated (Fig. 1). To date, most research on AM of metallic functional materials such as NiTi SMAs has focused on the effects of processing conditions on functional responses, but comparatively little ef- fort has been spent on investigating its printability. To address the issue, this research team has investigated the im- pact of AM process conditions on the printability and transformation behav- ior of NiTi during LPBF [7] . Our results indicate that the linear energy density where P is the laser power and v is the scanning speed (the speed with which the laser moves over the pow- der layer), is a better design parameter for identifying satisfactory printability. In contrast, volumetric energy density , where h is the hatch spacing (which defines the degree of overlap between successive laser paths) and t is the layer thickness, is more relevant for controlling the transformation behav- ior of the processed material (Fig. 2). More recently, the team has also demonstrated a method to achieve lo- cal control of the 3D thermal history in AM. This work shows that by merely changing the local hatch spacing, it is possible to change the local transfor- mation temperatures between 60° and 100°C in NiTi SMA components (Fig. 3) [8] . The multi-stage transformation origi- nates from differences in local thermal histories, which in turn result in chang- es to local composition (due to differ- ential evaporation [7] ) and/or secondary precipitation structure [8] . For the first time, this research shows that the sen- sitivity of NiTi-based SMAs to changes in composition, coupled with the abil- ity to locally manage process parame- ters during AM, enables control—at the voxel level—of the functional response of materials fabricated during AM. While this study demonstrates the potential to control the behavior of NiTi SMAs on a location-by-location basis for 4D printing, there is still consider- able work to be done if one is to lever- age this to design active components with location-based functionality that can be printed. Indeed, in order to use this spatial control of properties in a design context, computational models are needed that can predict the effects of process parameters on functional properties throughout the entire part. Knowledge gained from these compu- tational methodologies can then be used to predict and guide new experi- ments toward more optimal designs lo- cally, or other interesting areas of the design space globally. PRECIPITATE EVOLUTION Understanding precipitate evolu- tion during bulk fabrication and pro- cessing is essential to the 4D printing and microstructural design of NiTi- based SMAs. The formation of metasta- ble Ni 4 Ti 3 precipitates in Ni-rich com- positions leads to appreciable local compositional changes that affect the martensitic transformation character- istics, properties, and performance [9,10] . This issue deserves even more atten- tion for AM due to the complex ther- mal history of the printed parts leading to hard-to-predict precipitation char- acteristics. Therefore, in the work dis- cussed here, the outputs of the thermal Fig. 1 — Printability maps that predict melt pool morphology regions for a Ni-base alloy (Ni-5%Nb). Predicted regions are as follows: 1) good quality; 2) keyholing defect; 3) ball-ing defect; and 4) lack of fusion defect [6] . Fig. 2 — Influence of volumetric energy density on martensite start (Ms) and austenite finish (Af) transformation temperatures of the nondefective AM coupons of the NiTi shape memory alloy [7] . Courtesy of Elsevier.