Christopher H. Zenk

Hierarchical coherent precipitate-strengthened ferritic superalloys offer a low-cost, low-density alternative to Ni-based alloys, yet their high-temperature environmental embrittlement remains unexplored. Here we examine a B2/L2₁ precipitate-strengthened Fe-based superalloy exposed discontinuously at 650 °C to dry air and to air containing 20 vol.% water vapor. The alloy gains more mass in dry air, yet sub-surface cracking occurs exclusively in moist air. Combining Raman spectroscopy, electron microscopy, selected area electron diffraction (SAED), atom probe tomography (APT), thermodynamic modeling and classical nucleation theory, we establish a mechanistic framework linking oxidation to cracking and subsequent microstructural evolution. Oxidation in moist air generates atomic hydrogen, while oxide growth imposes tensile stress in the underlying alloy, together satisfying the conditions for hydrogen embrittlement. Crack advance exposes pristine metal that undergoes further oxidation and hydrogen generation. Alumina is found within all analyzed crack wakes, indicating that cracking and oxide growth are coupled. Oxidation-driven Al depletion supplies the chemical driving force for C14 Laves phase formation at alloy/oxide interfaces, and the alumina/matrix interface catalyzes heterogeneous nucleation, supported by estimated nucleation barriers. All analyzed cracks arrest upon intersecting Laves precipitates ahead of the tip. SAED and APT reveal a local C14 Laves to µ-phase transformation at these crack tips, consistent with defect-mediated transformation under the crack-tip stress field, providing a mechanism for crack-tip shielding and arrest. These results establish a unified oxidation–hydrogen–precipitation–fracture mechanism in which the environment governs not only oxide evolution and hydrogen availability but also second-phase precipitation and its response to propagating cracks.

Additive Manufacturing (AM), particularly electron beam powder bed fusion (PBF-EB), offers multiple advantages compared to conventional processing routes, but alloy development for AM is associated with high costs, as significant powder volumes must be produced and printed to enable mechanical property evaluation. This work evaluates whether conventionally processed substitute microstructures can provide meaningful mechanical property data for the early-stage screening of new AM alloy candidates. Substitute microstructures were produced from the γ′ Ni-based superalloys Alloy 247 and Alloy 247 LC using arc-melting, conventional casting, directional solidification, and controlled recrystallization, and were compared to two PBF-EB processed microstructures. Given the limited data on the recrystallization behavior in high γ′ Ni-based superalloys, a systematic recrystallization study was conducted to identify processing parameters capable of producing fine-grained microstructure with controlled grain boundary densities. Microstructural characterization was performed using scanning electron microscopy and electron backscatter diffraction. Compressive creep tests were conducted at 980 °C and 200 MPa. Although no substitute fully replicates the complex AM microstructure, selected microstructures exhibit comparable grain boundary characteristics and similar creep behavior in specific loading directions. Such well-controlled substitutes enable meaningful comparative creep testing and first-order extrapolation, offering a practical, resource-efficient route for early mechanical screening of new AM alloys.

Using FIB–SEM tomography, we reveal an anomalous three-dimensional internal alumina network that forms during high-temperature oxidation (1100 °C) of an electron-beam powder bed fusion (PBF-EB) Ni-based superalloy 247. This network develops beneath the external scale and preferentially forms around an interconnected skeleton of refined carbides. While conventional 2D cross-sections suggest isolated oxide–carbide co-precipitates, three-dimensional reconstructions show that all internal oxides are topologically connected, forming a continuous network. Time-resolved exposures (5 min, 30 min, 50 h) demonstrate that directional carbide agglomeration during isothermal holding builds the skeleton for subsequent alumina growth, as carbide/matrix interfaces act as high-diffusivity oxygen pathways and are enriched in Al. Based on these observations, we propose a mechanism of internal oxidation governed by carbide dissolution and directional agglomeration. A commercially cast counterpart of alloy 247, oxidized for 50 h at 1100 °C, exhibits no internal oxide network, directly linking the emergence of the 3D alumina network to the unique, fine and spatially uniform carbide distribution produced by PBF-EB. These findings uncover a previously unrecognized oxidation pathway in additively manufactured superalloys and highlight the critical role of microstructural design in controlling internal oxidation resistance.

Computational superalloy development is a powerful alternative to the conventional experimental approach. Based on thermodynamic databases and the CALPHAD method, it is possible to estimate the properties of a large number of potential alloys and select the most promising ones. However, the accuracy of the databases and complementary property models can be unsatisfying. The accuracy of two mass density and lattice parameter models and the TTNI8 and TCNI10 databases is analyzed in detail on the experimental basis of computationally optimized single‐crystalline superalloys. Various properties are measured and compared to the results of the property models and databases. Neither of the databases is superior to the other and especially the solvus temperature is not accurately described in both. The new mass density model, a linear regression based on the molar mass, is more reliable for low‐density alloys. Both lattice parameter model versions slightly overestimate the room‐temperature lattice parameter. The lattice parameter, however, is more accurately calculated using the new model version. The results of this study can be readily used to improve a multicriteria alloy optimization tool for computational superalloy design.

One impeding factor for faster solidification in typical investment casting processes is the limited heat removal capability. This study compares the classical Bridgman technique and the fluidized carbon bed cooling (FCBC) casting method. In particular, the influence of the thermal design is investigated in terms of microstructural and compositional homogeneity in single‐crystalline CMSX‐4. These investment casting processes mainly differ in the baffle type used to insulate the hot and cold zones inside of the casting vessel. While the Bridgman process is based on radiation cooling, the FCBC casting process relies on a dynamic baffle made of expanded graphite that is buoyant on fluidized bed of glassy carbon beads. Such a dynamic insulation layer can adapt to the cross‐sectional variation of ceramic shell molds and provide a better thermal insulation, which results in more uniform solidification conditions. This enables a reduction in the primary dendrite arm spacing by up to 26%, an increase in the homogeneity of the microstructure and composition, as well as a reduction in overall porosity and mean pore size of 41%.

Cost‐ and time‐demanding trial‐and‐error methods have been the historical route for alloy development. A combinatorial approach can significantly simplify and accelerate the development process by characterization of composition dependent properties on material libraries, which are specimens or sets of specimens that map out a certain composition space, often employing composition profiles. Herein, a promising production method for such a material library is proposed: laser‐remelting of stacked blocks of different compositions is evaluated for its suitability to produce material libraries, using the ternary CrCoNi system for a proof‐of‐concept. The composition profiles of the successfully created CrCoNi material library were measured by electron probe micro analysis. The intermixing has a length of about 2.5 mm. An analytical model describing the intermixing process is proposed and shows value in the estimation of the intermixing length after the first melting step. The comparison of the experimental microstructure observations from this work and from literature shows mostly good agreement with some deviations related to microsegregation and finite quenching cooling rates, which is supported by thermodynamic calculations regarding phase stability. In the single‐phase region, the mechanical properties as measured via microhardness indentation are discussed as potential candidates for model validation.