Computational Design of a Reactive MoO2/Mo/Cr/Ni-Cr Functionally Graded Composite Interlayer for Sapphire–Inconel 718 Laser Powder Bed Fusion Bonding

Abstract

This archive contains the computational framework and the measurement data used to design and evaluate a reactive eleven-layer MoO2/Mo/Cr/Ni-Cr functionally graded interlayer for joining c-plane sapphire to Inconel 718 by laser powder bed fusion, as described in the accompanying paper. Every number and every figure in the paper and its Supplementary Information is regenerated by the scripts included here. The framework keeps two quantities apart. The interface-step metric measures the mismatch between adjacent layers and is the objective the composition gradient controls; the membrane stress is what each layer carries against the substrate, and each layer is assigned the stress-free reference state its fabrication history gives it. Expansion coefficients are means over each thermal excursion rather than room-temperature values. Modulus-compensated grading equalises the six gradient steps at 0.58 GPa under a bounding 1200 K excursion, 53 per cent below equal composition steps; the shield-to-barrier step of 1.8 GPa lies outside the graded zone and is reported as a limit of the design. The archive implements: the analytical engine for the step metric, the membrane stress, the gradient optimisation and the fluence budgets; a generalised plane strain finite-element model of the stack on sapphire with per-layer reference temperatures and temperature-dependent seed plasticity; edge-crack decks and the steady-state multilayer energy release rate of Hutchinson and Suo; a one-dimensional transient enthalpy conduction model of the bonding pass with phase change, which shows the interlayer to be thermally thin under an 80 microsecond pulse; a diffusion-dose evaluation of the chromium barrier with lattice and grain-boundary transport; and an axisymmetric thermoelastic model of the sapphire under the hot spot. The finite-element models use the free CalculiX solver; everything else is pure Python with NumPy, SciPy and Matplotlib. The measurement data are the raw X-ray diffraction scans of a deposited witness stack of the lower three zones and the photoelectron spectra of the oxide and metallic layers, from which the crystallite sizes and the oxide phase assignment in the paper are taken. The code is released under the MIT licence and the measurement data under CC BY 4.0. Version 1.0.1 of this archive accompanied the first submission of the manuscript and is superseded by this version; README.md lists what changed.