Computationally efficient thermo-kinetic modeling for predictive control of melt pool chemistry in additive manufacturing of high-volatile-disparity alloys

The control of chemical composition remains the single most formidable challenge restricting the industrial adoption of Additive Manufacturing (AM) for high-volatility-contrast multicomponent alloys like TiAl. The extreme energy density utilized in powder bed fusion or directed energy deposition processes, exemplified here by the electron beam directed energy deposition (EB-DED) technique, triggers severe selective evaporation of volatile components, fundamentally compromising alloy design fidelity. This study presents a paradigmatic shift in AM process modeling by establishing a robust composition-prediction framework. We first analytically establish that Al selective evaporation is the dominant mechanism (Al vapor pressure is 1-2 orders of magnitude higher than Ti), and that the mass loss rate is kinetically controlled by the gas-liquid interface reaction ( K _m << β _m ). To resolve the non-uniformity inherent to the AM melt pool, we developed an advanced model that integrates fluid-dynamic predictions of melt flow with interface kinetics. This model uniquely accounts for the formation of distinct surface regimes: flow-induced solute-enriched regions (laminar flow) and a surrounding compositionally uniform region (vortex flow). The resulting model, incorporating both the non-uniform temperature field and the heterogeneous solute concentration, successfully forecasts the final alloy chemistry. The model demonstrates exceptional predictive robustness, achieving an experimental deviation of less than 2 at. % when compared to measurements. Based on this validated framework, an optimal processing window (20– 30 mA beam current; 1.5– 2.5 mm/s deposition speed) is defined. This work provides the necessary theoretical foundation and predictive toolset for achieving high compositional fidelity in the additive manufacturing of all alloys susceptible to severe elemental volatilization.