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 9 model for NiTi SMAs have been coupled to a precipitation model that predicts values of characteristic microstruc- tural features (e.g., phase fraction) at discrete points throughout the success- fully printed part [11] . Parameters of the precipitation model are calibrated for a range of dif- ferent heat treatments through a Bayes- ian calibration framework that uses the Ni content of the matrix as a compara- tive metric between the model and ex- periments [12] . Once the phase fraction of the precipitates is known locally, a microstructurally informed SMA mod- el is used to predict the effective local thermomechanical response of the pre- cipitation hardened NiTi SMAs, either AM fabricated or statically aged [13] . The model takes into account the structural effects of the precipitates on the revers- ible martensitic transformation under load as well as the chemical effects re- sulting from the Ni-depletion of the ma- trix during AM fabrication and precipi- tate growth. These structural effects in- clude precipitate volume fraction, elas- tic properties, elastic mismatch be- tween the precipitates and matrix, and coherency stresses due to the lat- tice mismatch between the precipitates and matrix. The thermomechanical re- sponse is predicted based on finite el- ement simulations of representative microstructures. While precipitation plays an im- portant role in modifying the local response in NiTi SMAs [8] , local compo- sitional changes through differential evaporation play an equally import- ant role. The preferential evaporation of Ni over Ti induced by manipulation of the hatch spacing parameter is suf- ficient to realize local variations in the Ni content of the LPBF fabricated sam- ples in the range of 0.5-1.5 at.%. This is significant because only a few tenths of a percent change in Ni composition is sufficient to change the transforma- tion temperature in NiTi SMAs by tens to 100 degrees K [14] . In fact, it is known that at high temperatures encountered during AM, there is significant evapo- ration of alloying elements from the molten pool. Because some alloying el- ements are more volatile than others, selective evaporation of the elemental constituents often results in a signifi- cant change in alloy composition. This has been widely observed in laser weld- ing and AM. Preliminary work by this research team includes implementing a framework to couple thermal models based on Rosenthal’s solution to evap- oration models based on the Langmuir equation [15,16] . This team also uses sur- rogate thermal models based on the high-fidelity thermal model mentioned above to predict the thermal history resulting from complex processing paths and resulting differential evapo- ration in NiTi (Fig. 4). The model lacks some predictive capability because discrepancies remain with the experi- ments, but its advantage is numerical efficiency. The overarching goal of these com- putational and experimental efforts is to translate the systems-level specifi- cation of the spatially-tailored shape memory response into specifications, (a) (b) Fig. 3 — (a) Multi-stage shape recovery in a U-shaped AM NiTi SMA part using laser powder bed fusion (LPBF). A location-dependent active response is achieved when the two arms of the piece activate their shape recovery at different temperatures; (b) the origin of the location- dependent active response is due to a change in the LPBF processing parameters (shown in the differential scanning calorimetry curves on the right) at different sections of the build, which results in differences in the transformation temperatures in corresponding sections [4] . Fig. 4 — (a) Connecting scan strategy to (b) local thermal histories to (c) quantitative models for differential evaporation during LPBF AM of NiTi shape memory alloys.