The Ni-based superalloy Haynes 282 exhibits rapid γ′ precipitation kinetics, and experimental studies have shown the formation of in-process precipitates in samples produced using the direct energy deposition additive manufacturing method, laser metal deposition (LMD). Understanding how these precipitates form and influence the final microstructure is essential for predicting and controlling the mechanical properties of processed Haynes 282. In this study, the particle size and volume fraction of γ′ precipitates formed in LMD samples are simulated using classical nucleation and growth theory (CNGT). To account for the thermal history during manufacturing, the precipitate model is implemented as a multi-scale framework integrated into a finite element software. Calphad thermodynamic and diffusion data descriptions are used as input to the CNGT model to simulate the precipitation kinetics during different heat treatments. The simulated results are compared with experimental data obtained from small- and wide-angle X-ray scattering, as well as from atom probe tomography. The simulations show good agreement with experimental findings, demonstrating that thermodynamic databases can be used to accurately simulate precipitate evolution in LMD-processed Haynes 282 using CNGT.

Grain coarsening inhibition in hard metals is regarded as controlled by formation of interface complexions. To date, however, direct experimental insights into the presence and evolution of interface complexions during sintering of hard metals have been lacking. We here present in-situ small-angle neutron scattering (SANS) experiments up to 1500 degrees C and provide first-hand evidence on the thickness and volume fraction evolution of (V,W) Cx interface complexions in V-doped hard metals at various sintering temperatures. The experimental data is complemented by simulations using a thermodynamic-based model to understand the mechanisms behind the nanostructure evolution. We show that there indeed exist (V,W)Cx interface complexions at liquid-phase sintering temperatures; and their thickness and volume fraction are strongly related to the presence of bulk (V,W)Cx precipitation, the V activity in the Co-rich binder phase, and the temperature. The thermodynamics-based model, including the geometry of the investigated material system, reveals that the formation of (V,W)Cx bulk precipitates is energetically favorable over the thickening of complexions in the stability range of bulk precipitation. This, explains the reduction in complexion volume fraction and thickness with increasing temperature up to the dissolution of bulk precipitates. Upon dissolution of bulk precipitates, enhanced interfacial layer formation occurs through the formation of new layers of lower thickness, leading to better coverage of WC grains. The provided understanding of the nanostructure evolution during sintering is expected to foster the further development of representative modelling tools.

Among several separate challenges, the major one for replacing cobalt in cemented carbides is the difficulty to obtain alternative binder materials with a C-window broad enough to be robustly processed under conventional industrial control on the C content. The C-window is defined as the C content range for which phases that are detrimental to the mechanical properties are avoided. The present paper has two main objectives: first, to show that the processing C-window of Fe-Ni based systems is in fact wider than what thermodynamic equilibrium calculations predict, and that its width can be controlled moderately by tweaking the initial WC grain size and the cooling rate used in the material's processing. Secondly, in case those detrimental phases are not avoided, this work gives insight on how to make their appearance less detrimental for the mechanical properties. The morphology, volume fraction and particle size distribution of the detrimental phases, specifically η6-carbides at low C contents, are investigated to explore desirable combination of hardness and toughness of alternative binder cemented carbides. During this study it was also discovered that in samples with carbon contents below the low-C limit of the C window a carbide with hexagonal lattice known as κ, not commonly seen in cemented carbides, appeared and formed the core of a core-rim structure together with the more common η6-phase. It is believed that the κ-carbide form due to local high concentrations of tungsten during solid state sintering and that it has an impact on the precipitation characteristics of the η6-phase.

There are several semi-empirical models available in literature that correlate the intrinsic hardness of cemented carbides' constitutive phases and certain microstructural parameters, such as mean WC grain size and Co volume fraction, with the hardness of the cemented carbide. Nonetheless, such empirical relations fall short on predicting the behavior of materials other than WC-Co which they were fitted to, limiting their applicability on materials with diverse particle size distributions, alternative binder systems or with additional carbides (γ-carbides). Additionally, current models are limited to the prediction of room temperature hardness. Framed in an Integrated Computational Materials Engineering (ICME) approach, this work proposes several models to be integrated into an already validated semi-empirical approach to describe the hardness of cemented carbides as a function of temperature. First, new microstructural descriptors on the particle and binder size distributions are proposed to enable a better understanding of the influence of polydispersity and of the addition of γ-carbides on the hard-to-soft phase reinforcement. Second, a validated Peierls-Nabarro-based model is used to describe the intrinsic softening of the hard phases with temperature. And finally, the importance of the microstructural changes happening under stress at high temperatures is highlighted and its effect on hot hardness is introduced into the model. These upgrades increase the theoretical and physical base of the modeling tool providing a physical meaning to all the modeling parameters, lowering the need for numerical fitting, making the model more generic and bringing additional information into the micromechanics involved in the softening of cemented carbides.

Link prediction between two nodes is a critical task in graph machine learning. Most approaches are based on variants of graph neural networks (GNNs) that focus on transductive link prediction and have high inference latency. However, many real-world applications require fast inference over new nodes in inductive settings where no information on connectivity is available for these nodes. Thereby, node features provide an inevitable alternative in the latter scenario. To that end, we propose Graph2Feat, which enables inductive link prediction by exploiting knowledge distillation (KD) through the Student-Teacher learning framework. In particular, Graph2Feat learns to match the representations of a lightweight student multi-layer perceptron (MLP) with a more expressive teacher GNN while learning to predict missing links based on the node features, thus attaining both GNN's expressiveness and MLP's fast inference. Furthermore, our approach is general; it is suitable for transductive and inductive link predictions on different types of graphs regardless of them being homogeneous or heterogeneous, directed or undirected. We carry out extensive experiments on seven real-world datasets including homogeneous and heterogeneous graphs. Our experiments demonstrate that Graph2Feat significantly outperforms SOTA methods in terms of AUC and average precision in homogeneous and heterogeneous graphs. Finally, Graph2Feat has the minimum inference time compared to the SOTA methods, and 100x acceleration compared to GNNs. The code and datasets are available on GitHub1.

Cemented carbides are widely used materials in industrial applications due to their remarkable combination of hardness and toughness. However, they are exposed to high temperatures during service leading to a reduction of their hardness. A common practice to damp this softening is to incorporate transition metal carbides in cemented carbide compositions, which keeps the hardness relatively higher when temperature increases. Understanding the underlying mechanisms of this softening is crucial for the development of cemented carbides with optimal properties. In this work, atomic-scale mechanisms taking place during plastic deformation are analyzed and linked to the effect that they have on the intrinsic macro-scale softening of the most common TMC used in cemented carbides grades (TiC, ZrC, HfC, VC, NbC and TaC). The proposed model uses the generalized stacking fault energy obtained from density functional theory calculations as an input to Peierls-Nabarro analytical models to obtain the critically resolved shear stress needed for deformation to occur in different slip systems. Subsequently, this information is used to predict the hardness variation across the temperature service range experienced by cemented carbides in wear applications. In addition to the prediction of hot-hardness for TMC, the obtained results also offer valuable insights into the intrinsic mechanisms governing TMCs deformation. The results facilitate the identification of dominant dislocation types influencing plasticity within distinct temperature regimes, define energetically favorable slip systems, and predict the brittle-ductile transition temperature in these materials. For instance, for group IV carbides at low temperatures, the slip system with a lower GSFE is {110}<11̅0> and around 30% of their melting temperature, the GSFE of partial slip in {111}<12̅1> becomes lower, changing the dominant slip mechanism and characterizing the Brittle-Ductile transition.

Integrated Computational Materials Engineering (ICME) has proved to be an efficient tool for understand-ing the process-structure-property relationships and helping us to design materials. For instance, in cemented carbides manufacturing, one of the most critical parameters is the C-window. It is defined as the C content range for which phases detrimental to the mechanical properties are avoided. This pro-cessing window has been traditionally defined using applied thermodynamics methods. However, the deviation between equilibrium calculations and real manufacturing conditions requires big additional empirical efforts to precisely define the C-window. In this work, an ICME-based approach is proposed to redefine the processability limits of cemented carbides taking the cooling rate and the material's initial powder size into consideration. The method relies on the interactive coupling of several adapted models and tools, to not only set the processability boundaries, but also to study the complex mechanisms inter-play happening along microstructural evolution. A better understanding of these underlaying mecha-nisms leads to new inputs that can be used in the design of cemented carbides. In this regard, it is observed that faster cooling rates or coarser WC grades could be effectively used to prevent nucleation of the detrimental phases enlarging the C-window towards lower C contents. 

Astaloy™ 85 Mo is a pre-alloyed, water-atomized 0.85% Mo steel powder. The aim of the present investigation is to study the influence of porosity, controlled by both mechanical and thermal processing, on the mechanical properties in a bainitic microstructure of a pressed and sintered steel. To achieve this, uniaxial tensile and compression testing is performed, together with Vickers macro- and microhardness experiments. Microhardness testing is carried out in order to determine the behavior of the matrix material at a scale where porosity influence is minimized. Both the influence from size and shape of the pores is investigated and compared with relevant mechanical analyses for porous solids. Such mechanical analyses are pertinent to both elastic and plastic properties, where in the latter case the well-known Gurson-Tvergaard model for solids with spherical pores is relied upon. It is shown that assuming a spherical pore shape is not sufficient in order to achieve good agreement between predictions and experimental results and will be further investigated in future studies.

Abstract: Quantitative modelling of precipitation kinetics can play an important role in a computational material design framework where, for example, optimization of alloying can become more efficient if it is computationally driven. Precipitation hardening (PH) stainless steels is one example where precipitation strengthening is vital to achieve optimum properties. The Langer–Schwartz–Kampmann–Wagner (LSKW) approach for modelling of precipitation has shown good results for different alloy systems, but the specific models and assumptions applied are critical. In the present work, we thus apply two state-of-the-art LSKW tools to evaluate the different treatments of nucleation and growth. The precipitation modelling is assessed with respect to experimental results for Cu precipitation in PH stainless steels. The LSKW modelling is able to predict the precipitation during ageing in good quantitative agreement with experimental results if the nucleation model allows for nucleation of precipitates with a composition far from the equilibrium and if a composition-dependent interfacial energy is considered. The modelling can also accurately predict trends with respect to alloy composition and ageing temperature found in the experimental data. For materials design purposes, it is though proposed that the modelling is calibrated by measurements of precipitate composition and fraction in key experiments prior to application. Graphic abstract: [Figure not available: see fulltext.]. 

Integrated Computational Materials Engineering (ICME) has proved to be an efficient tool for understanding the process-structure-properties relationships and help us design materials. In the case of cemented carbides, ICME can be used to study how well-established relationships are affected by new industry challenges, like the replacement of Co by alternative binder materials. The C-window is one of the most critical parameters in cemented carbides and is defined as the range in C-content for which phases detrimental to the mechanical properties are avoided. For alternative binder materials, the processing window may become very narrow, and a good understanding of its limits becomes crucial. The low-C side of the C-window is limited by the precipitation of η-carbides (M6C/M12C). Here, a systematic and integrated study on the formation of η-carbides as a function of binder chemical composition, cooling rate and microstructure is presented.