Process Structure Property Research Articles

Application of machine learning to reveal relationship between processing-structure-property for polypropylene injection molding

Revealing relationship between processing–structure-property using machine learning (ML) is beneficial to shorten the development cycle of polymer materials, which could be a promising solution to break through the traditional mode of trial-and-error. In this work, the quantitative data of processing field, polymorphic structure and mechanical properties of samples are successfully obtained based on microfocus wide-angle X-ray diffraction and mechanical property measurements, which provides the dataset for the application of machine learning. An initial step toward predicting the contents of polymorphic structure (α and β) and mechanical properties of isotactic polypropylene injection molding is taken by applying four algorithms of machine learning to establish regression models. The importance of processing descriptors is further analyzed. The prediction models for the contents of polymorphic structure and mechanical properties demonstrate satisfactory performance (R2max = 0.95). Compared with the prediction for the content of α-crystals, the results of feature importance analysis indicates that the injection pressure and shear rate of filling end in the processing descriptors charge higher contribution to the prediction for the β-crystals. And mold temperature plays an important role in predicting mechanical properties, which is consistent with the results of the signal-to-noise (S/N) ratio analysis.

Evaluation of computational homogenization methods for the prediction of mechanical properties of additively manufactured metal parts

It is well known that the strongly location-dependent microstructures observed in metal parts made using additive manufacturing (AM) processes differ from those found in components produced via traditional manufacturing processes. This is primarily because the microstructures in AM parts strongly depend on spatially-varying cooling rates that are caused by several factors including part geometry and proximity to the build plate. Consequently, it is necessary to take a fresh look into methods used for calculating resulting mechanical properties, as the increased spatial resolution that must be applied to the description of properties of AM parts calls for vast improvements in computational efficiency. Computational homogenization methods are employed for the calculation of these mechanical properties based on the polycrystalline microstructures typically encountered in metallic materials. Such calculations that enable the description of properties as a function of position increase the accuracy of a vital link in the structure–process–property–performance continuum that falls within the Integrated Computational Materials Engineering (ICME) paradigm. In this work we describe and critically review six of these methods and their application to AM microstructures. We find that the Fast Fourier Transform (FFT) method and the Finite Element Method (FEM) are the optimal choices based on several factors including accuracy, efficiency, and applicability to polycrystalline microstructures. Provided the microstructure is amenable to the use of uniform grids and the assumption of periodic boundary conditions, the FFT method has greater computational efficiency. Where unstructured meshes, adaptive remeshing and/or non-periodic boundary conditions are desired, the FEM is the method of choice. Our recommendations are equally applicable to functionally graded AM parts where site-specific microstructures are engineered using multiple materials or different process parameters. They can also be utilized for other material systems and circumstances where increased accuracy is achieved by describing mechanical properties as a function of location.

Non-Metallic Alloying Constituents to Develop a Wear-Resistant CrFeNi-BSiC High-Entropy Alloy for Surface Protective Coatings by Thermal Spraying and High-Speed Laser Metal Deposition

Compositional alterations to high-entropy alloys (HEAs) allow further evolution of these materials by adjusting their property profiles. This way, they can be used for coating technologies and surface-protection applications. In the present work, minor quantities of the non-metallic alloying constituents, BSiC, were added to the CrFeNi base system. The alloy development was carried out in an electric arc furnace in comparison with the nickel-based alloy Ni-600. With regard to the BSiC-free variant, the wear resistance can be significantly increased. The powder was manufactured by inert gas atomization and characterized, followed by processing via high-velocity oxy-fuel spraying (HVOF) and high velocity laser metal deposition (HS-LMD). Depending on the manufacturing conditions, the proportion and shape of the precipitates within the microstructure differ. Compared to both the reference system and the as-cast condition, the coating systems demonstrated comparable or improved resistance to wear. The evaluation of the process–structure–property relationships confirmed the great potential of developing load-adapted HEA systems using non-metallic alloy constituents in the field of surface engineering.

The Importance of Synthesis Conditions: Structure–Processing–Property Relationships

When structure–property relationships are discussed in inorganic chemistry and materials science, the common perception is that materials with the same composition and atomic arrangement have the same properties. In this laboratory experiment, we show that processing─the exact conditions under which materials are made─can in fact significantly alter the properties of materials. Using halide perovskite as a model, we demonstrate that the same set of starting chemicals, when processed differently, can form distinct material platforms–single crystals, thin films, and nanoparticles, the last of these being made of crystallites that are 10,000 times smaller than the first. We then show how this difference in crystal size leads to major changes in the optical properties of the materials, despite the fact that they have the same composition and crystal structure. The experiment provides an example of how the basic concept of crystallization leads to different materials and how processing affects the crystal size and thus influences the structure–property relationship.

Metal AM process-structure-property relational linkages using Gaussian process surrogates

In metal additive manufacturing (AM), with sufficient understanding of process-structure–property (PSP) relational linkages, the control of build parameters can produce parts with previously unattainable properties. Establishing these PSP linkages involves varying process parameters until a desired microstructure, and corresponding properties, are achieved, either experimentally or computationally. However, both methods can have a high acquisition cost and be difficult to sample repeatedly. This work describes a Gaussian process (GP)-based workflow that is capable of predicting melt pool characteristics, microstructure features, and mechanical properties of previously unseen process parameter combinations. The workflow implements multi-fidelity, multi-output, and functional GPs, trained on a limited set of experiments and simulations in order to make melt pool, microstructure, and mechanical property predictions. The established linkage results in approximately 95% accuracy in predicting mechanical properties for previously unseen set of process parameters propagated through the whole framework. The use of GPs in the workflow limits the number of experiments/simulations needed, yields a nearly negligible cost for acquisition of new predictions, and allows for a Bayesian treatment of the PSP linkages that was not previously feasible.

An Investigation of Mechanical Properties of Recycled Carbon Fiber Reinforced Ultra-High-Performance Concrete.

Carbon fiber-reinforced concrete as a structural material is attractive for civil infrastructure because of its light weight, high strength, and resistance to corrosion. Ultra-high performance concrete, possessing excellent mechanical properties, utilizes randomly oriented one-inch long steel fibers that are 200 microns in diameter, increasing the concrete's strength and durability, where steel fibers carry the tensile stress within the concrete similar to traditional rebar reinforcement and provide ductility. Virgin carbon fiber remains a market entry barrier for the high-volume production of fiber-reinforced concrete mix designs. In this research, the use of recycled carbon fiber to produce ultra-high-performance concrete is demonstrated for the first time. Recycled carbon fibers are a promising solution to mitigate costs and increase sustainability while retaining attractive mechanical properties as a reinforcement for concrete. A comprehensive study of process structure-properties relationships is conducted in this study for the use of recycled carbon fibers in ultra-high performance concrete. Factors such as pore formation and poor fiber distribution that can significantly affect its mechanical properties are evaluated. A mix design consisting of recycled carbon fiber and ultra-high-performance concrete was evaluated for mechanical properties and compared to an aerospace-grade and low-cost commercial carbon fiber with the same mix design. Additionally, the microstructure of concrete samples is evaluated non-destructively using high-resolution micro X-ray computed tomography to obtain 3D quantitative spatial pore size distribution information and fiber clumping. This study examines the compression, tension, and flexural properties of recycled carbon fibers reinforced concrete considering the microstructure of the concrete resulting from fiber dispersion.

The influence of crystallographic texture on structural and electrical properties in ferroelectric Hf0.5Zr0.5O2

Ferroelectric (Hf,Zr)O2 thin films have attracted increased interest from the ferroelectrics community and the semiconductor industry due to their ability to exhibit ferroelectricity at nanoscale dimensions. The properties and performance of the ferroelectric (Hf,Zr)O2 films generally depend on various factors such as surface energy (e.g., through grain size or thickness), defects (e.g., through dopants, oxygen vacancies, or impurities), electrodes, interface quality, and preferred crystallographic orientation (also known as crystallographic texture or simply texture) of grains and/or domains. Although some factors affecting properties and performance have been studied extensively, the effects of texture on the material properties are still not understood. Here, the influence of texture of the bottom electrode and Hf0.5Zr0.5O2 (HZO) films on properties and performance is reported. The uniqueness of this work is the use of a consistent deposition process known as Sequential, No-Atmosphere Processing (SNAP) that produces films with different preferred orientations yet minimal other differences. The results shown in this study provide both new insight on the importance of the bottom electrode texture and new fundamental processing-structure–property relationships for the HZO films.

Ternary Organic Solar Cells: Recent Insight on Structure–Processing–Property–Performance Relationships

This article presents recent advances in the ternary organic solar cell (TOSC), such as technological interventions from the material design to the device performance, which led to more than 19% power conversion efficiency (PCE). The research and development in TOSCs reported in the past decade have been inspiring and promising in terms of molecular processing to device properties for low‐power devices and building integrated photovoltaics applications. Many of these are still in the early phase, enabling researchers to explore more and flourish. The research community has made numerous efforts to address the different aspects to enhance the efficiency and stability of the TOSCs. Therefore, the objective of this article is to review the recent advances in TOSCs and present a comprehensive discussion in this regard. This review also identifies and suggests the possible outlook for the critical issues associated with TOSCs, along with future perspectives.

On study of process induced defects-based fatigue performance of additively manufactured Ti6Al4V alloy

Additively manufactured Ti6Al4V are high strength alloys but associated with inferior ductility and inherent defects. These process-induced defects act as a stress raiser and deteriorates the mechanical and fatigue properties of the alloy. Post-processing is generally done to enhance the ductility and fatigue properties but at the expense of strength of the additively manufactured Ti6Al4V alloy. In this study, the effect of defects on mechanical properties and fatigue performance of additively manufactured (selective laser melting) Ti6Al4V alloy has been investigated based on experimentally calibrated defects-based model. The as-built specimens were subjected to different heat treatments in order to optimize their mechanical properties and the most optimum heat treatment was then selected for investigation of changes in fatigue properties. The beneficial effect of heat treatment on mechanical properties was mainly associated with the reduction in fraction of α ’-martensitic microstructure and increase in α -lamellae thickness. Most of the fatigue failures occurred from surface and sub-surface defects, owing to the high stress concentration factor associated with them, and improved fatigue lives as a result of optimum heat treatment. A process-structure–property (PSP) linkage has been established to predict the mechanical properties. Experimentally calibrated defects’ based fatigue model has been formulated to predict the fatigue performance using the experimentally observed data i.e. the process induced defects’ size and their locations. • Different post-processing of SLM-based AM Ti6Al4V was investigated. • Heat treatment was carried out, which improved both mechanical & fatigue properties. • A Process-Structure-Property (PSP) linkage was established. • A new defects-based model was proposed to predict the fatigue properties.

Traceability in Battery Cell Production

The manufacturing of lithium‐ion battery (LIB) cells has been identified as a hotspot addressing growing price competition and the environmental and economic pressures on technologies along the value creation chain. Effective quality management strategies (QMS) are essential to meet these rising challenges. Herein, data‐driven technologies have fostered an improved understanding of existing process–structure–properties and their process chain covering interdependencies. However, there are often limited in their capability due to the lack of availability or in‐depth information about the manufactured product, existing incidences (i.e., product faults), and the possibility to deploy inline QMS due to insufficient knowledge about the history and condition of the identified products. Against this background, this work describes the implementation of a traceability system as part of QMS for battery cell production and presents a developed framework to overcome challenges from an LIB production perspective for traditional traceability approaches. In addition, this article discusses the implementation of technologies, focusing on the identification and traceability on electrode‐sheet level through continuous and feature‐based approaches. With those technologies implemented, the deployment of quality gates within the electrode production is now possible to not only improve quality inline, but also discover new possibilities for future electrode balancing to reduce manufacturing costs and the overall environmental impact.

Integrated Computational Materials Engineering for Advanced Automotive Technology: With Focus on Life Cycle of Automotive Body Structure

Integrated computational materials engineering (ICME) is a simulation‐driven design approach that employs multiscale‐multiphysics modeling and is based on the understanding of process–structure–property–performance relationships. The ICME approach has attracted considerable attention in the automotive industry, considering that it significantly reduces development cost and time in optimization of materials and processes through computational advancement. This study presents a comprehensive review of ICME activities in advanced automotive manufacturing technology, which focuses on the life cycle of the automotive body structure such as structural material design, manufacturing process optimization, and reliability improvement in service quality. Results show that the ICME approach has been widely implemented in automotive manufacturing processes, from development of lightweight materials to performance management. However, further advances and technological strategies should be sought to develop more robust methodologies that can replace conventional experiment‐based design approaches.

Multi-Scale Mechanical Property Prediction for Laser Metal Deposition

The Laser Metal Deposition (LMD) involves extremely complex multi-scale multi-physics and multiple thermal cycles issues, making it difficult to accurately predict the resultant mechanical properties of fabricated components from given process parameters. This research, by proposing a cladding stacking model that uses the structural evolution history of the heat-affected zone, predicts the overall structure of fabricated components, and establishes a process–structure–property multi-scale simulation framework based on this model, a general solution for the abovementioned difficulty. Based on the Hot Optimal Transportation Meshfree (HOTM) method, a platform ESCAAS is developed to simulate the meso-scale Ti-6Al-4V powder evolution process. Based on the Cellular Automaton (CA) method, the micro-scale grain structure in the molten pool is simulated. The macro-scale mechanical property of the fabricated component is calculated based on a polycrystalline Representative Volume Element (RVE) model and the homogenization technology. Experiments including LMD multilayer printings, metallographic observations, and static tension are designed to verify the accuracy of the model and simulations. The results are greatly consistent with the experimental data and the relative error of the final mechanical property prediction is within 5.18%. This work provides a basis for the quantitative analysis of the process–structure–property relationship and the optimization of process parameters.

Multi-scale modeling for prediction of residual stress and distortion in Ti–6Al–4V semi-circular thin-walled parts additively manufactured by laser powder bed fusion (LPBF)

Despite many advantages of metallic additive manufacturing technology (AM), the difficulties in fully observing and modeling the complex nature of the process prevent its wide implementation in different industry sectors. There have been significant efforts during the last decade to develop reliable models to predict temperature field, melting stage, residual stresses and distortion during a material deposition in AM processes, in order to understand the complex process–structure–property relations in different scales, from powder particle to the whole part. In this study, a mechanical and thermo-mechanical computational methods are implemented to trade off the assessment between accuracy and computational cost of the process simulation. By taking advantage of a multiscale approach to reduce the computational cost and time while achieving promising accuracy, two scaled-models and non-scaled models with various levels of fidelity’s discretization were developed. The fidelity level is defined in space by allocating different numbers of physical layers of powder deposition to the finite element and in time by the time increment definition in numerical simulations. The commercial finite element software ABAQUS-implicit associating AM-plugin is utilized for modeling the printing process. The high-fidelity level for scaled part modeling duplicates the process and guides the medium-fidelity level reducing the computational resources needed. Then, by achieving the adequate configuration from the aforementioned level, the non-scaled model with the real dimensions of the experiments is simulated. Subsequently, the low-fidelity level is implemented for the real scale part utilizing the inherent strain method, taking into account the residual strains. The different approaches are compared, showing that the inherent strain method is more efficient in terms of accuracy and computational cost than the multiscale thermo-mechanical method.

A study on developing process-structure–property relationship with molten pool thermal history during laser surface remelting of Inconel 718

In the area of laser material processing, laser surface remelting has been found to be an effective method of improving surface properties such as hardness, wear, and corrosion resistance. However, the scale of improvement depends on the evolving microstructure and phases, which depend on the cooling rates. Therefore, in the present study, laser surface remelting of Inconel 718 was carried out and a process-structure–property relationship with respect to the cooling rates has been developed. During the laser surface remelting process, the molten pool thermal history i.e. cooling rate, molten pool lifetime, and solidification shelf time is monitored and estimated using an IR pyrometer. The evolution of microstructure is later correlated with these parameters. With an increase in scan speed, the cooling rate is found to increase resulting in transformation of microstructure from equiaxed grains to columnar epitaxial growth. Based on the results obtained, a process map is proposed to establish a particular type of microstructure with respect to the cooling rate. Further, the effect of cooling rate and microstructure on the surface hardness and specific wear rate has also been investigated. Both surface hardness and specific wear rate got reduced with decreasing cooling rate at a slower scan speed due to grain coarsening and an increase in elemental segregation or Laves phase formation.

Microscopic Insight into the Structure-Processing-Property Relationships of Core-Shell Structured Dialcohol Cellulose Nanoparticles.

In the quest to develop sustainable and environmentally friendly materials, cellulose is a promising alternative to synthetic polymers. However, native cellulose, in contrast to many synthetic polymers, cannot be melt-processed with traditional techniques because, upon heating, it degrades before it melts. One way to improve the thermoplasticity of cellulose, in the form of cellulose fibers, is through chemical modification, for example, to dialcohol cellulose fibers. To better understand the importance of molecular interactions during melt processing of such modified fibers, we undertook a molecular dynamics study of dialcohol cellulose nanocrystals with different degrees of modification. We investigated the structure of the nanocrystals as well as their interactions with a neighboring nanocrystal during mechanical shearing, Our simulations showed that the stress, interfacial stiffness, hydrogen-bond network, and cellulose conformations during shearing are highly dependent on the degree of modification, water layers between the crystals, and temperature. The melt processing of dialcohol cellulose with different degrees of modification and/or water content in the samples was investigated experimentally by fiber extrusion with water used as a plasticizer. The melt processing was easier when increasing the degree of modification and/or water content in the samples, which was in agreement with the conclusions derived from the molecular modeling. The measured friction between the two crystals after the modification of native cellulose to dialcohol cellulose, in some cases, halved (compared to native cellulose) and is also reduced with increasing temperature. Our results demonstrate that molecular modeling of modified nanocellulose fibers can provide fundamental information on the structure–property relationships of these materials and thus is valuable for the development of new cellulose-based biomaterials.

Optimisation-driven design to explore and exploit the process–structure–property–performance linkages in digital manufacturing

An intelligent manufacturing paradigm requires material systems, manufacturing systems, and design engineering to be better connected. Surrogate models are used to couple product-design choices with manufacturing process variables and material systems, hence, to connect and capture knowledge and embed intelligence in the system. Later, optimisation-driven design provides the ability to enhance the human cognitive abilities in decision-making in complex systems. This research proposes a multidisciplinary design optimisation problem to explore and exploit the interactions between different engineering disciplines using a socket prosthetic device as a case study. The originality of this research is in the conceptualisation of a computer-aided expert system capable of exploring process–structure–property–performance linkages in digital manufacturing. Thus, trade-off exploration and optimisation are enabled of competing objectives, including prosthetic socket mass, manufacturing time, and performance-tailored socket stiffness for patient comfort. The material system is modelled by experimental characterisation—the manufacturing time by computer simulations, and the product-design subsystem is simulated using a finite element analysis (FEA) surrogate model. We used polynomial surface response-based surrogate models and a Bayesian Network for design space exploration at the embodiment design stage. Next, at detail design, a gradient descent algorithm-based optimisation exploits the results using desirability functions to isolate Pareto non-dominated solutions. This work demonstrates how advanced engineering design synthesis methods can enhance designers’ cognitive ability to explore and exploit multiple disciplines concurrently and improve overall system performance, thus paving the way for the next generation of computer systems with highly intertwined material, digital design and manufacturing workflows.Graphical abstract

Metal Additive Manufacturing Process Design based on Physics Constrained Neural Networks and Multi-Objective Bayesian Optimization

The high-dimensionality of the process-structure-property (P-S-P) relationships for metal additive manufacturing (AM) is the major research challenge for process optimization to produce defect-free, structurally sound, and reliable AM parts. The goal of this work is to investigate the feasibility of a hybrid physics-based data-driven process design framework to establish reliable surrogates of process-structure relationships for metal AM process optimization. The process design framework includes a mesoscale multiphysics simulation model to predict microstructure evolution, a physics-constrained neural network to construct the surrogate of the process-structure relationship, and Bayesian optimization for process design. The proposed framework is demonstrated by optimizing the initial temperature and cooling rate for the single dendritic growth of Ti-6Al-4V alloy during rapid solidification in metal AM so that the desired dendritic area and microsegregation level can be achieved. The multi-objective optimization problem is solved with single-objective Bayesian optimization by aggregation. The effects of the aggregation weights on the optimization results are investigated with sensitivity analysis. The results show that the weights of individual objective functions determined from design preferences can help the designer to produce desired microstructure. The Pareto front and processing map for dendritic growth are constructed with the multi-objective Bayesian optimization method. The proposed process design framework is generic and can be potentially applied in materials design and digital twins of metal AM.

A modular framework to obtain representative microstructural cells of additively manufactured parts

The repeating nature of additive manufacturing (AM) translates into quasi-periodic repeating hierarchical microstructures, which vary depending on the working principle of the AM technology. The geometry and internal characteristics of these repeating microstructures affect the geometrical accuracy and mechanical properties of AM parts. We propose a modular and generic framework for AM microstructures to automatically determine a representative microstructural cell with average shape and microstructural information based on micrograph processing and morphological shape analyses. A general microstructural mapping procedure is presented to map and average different microstructural features inside the representative cell. Moreover, an extension of the method is proposed to recover a representative cell with a periodic shape as required for subsequent numerical micromechanical simulations under periodic boundary conditions to unravel process-structure–property relationships. The framework is demonstrated on a virtually generated microstructure and on the metallic and polymeric microstructure of two parts manufactured with different AM processes, showing the framework's versatility for different materials and AM technologies. The three test cases show the framework's modularity, where different procedures were applied to overcome the challenges of each particular case while keeping the primary method the same. We also discuss applications of the framework, such as enabling statistical investigations of spatial variations of the microstructural features across AM parts and providing essential input for microstructural and mechanical numerical simulations.

Skeletal-based microstructure representation and featurization through descriptors

Modeling process–structure–property relationships using machine learning methods has become a valuable enabler for materials design and discovery. However, the machine learning models rely heavily on the featurization of the materials’ structure. This paper introduces a microstructure featurization framework to compute generic topological and morphological descriptors. The framework relies on our skeletal microstructure representation. The representation allows for the seamless calculation of topological descriptors, which is the main focus of this paper.To demonstrate the efficacy of our featurization framework, we couple it with a feature selection method to establish the structure–property model for organic photovoltaics (OPV). For this goal, we identify a salient set of descriptors and construct the structure–property map with high accuracy.

Superconducting niobium nitride: a perspective from processing, microstructure, and superconducting property for single photon detectors

Superconducting niobium nitride (NbN) continues to be investigated decades on, largely in part to its advantageous superconducting properties and wide use in superconducting electronics. Particularly, NbN-based superconducting nanowire single-photon detectors (SNSPDs) have shown exceptional performance and NbN remains as the material of choice in developing future generation quantum devices. In this perspective, we describe the processing–structure–property relationships governing the superconducting properties of NbN films. We further discuss the complex interplay between the material properties, processing parameters, substrate materials, device architectures, and performance of SNSPDs. We also highlight the latest progress in optimizing SNSPD performance parameters.