To save content items to your account,
please confirm that you agree to abide by our usage policies.
If this is the first time you use this feature, you will be asked to authorise Cambridge Core to connect with your account.
Find out more about saving content to .
To save content items to your Kindle, first ensure firstname.lastname@example.org
is added to your Approved Personal Document E-mail List under your Personal Document Settings
on the Manage Your Content and Devices page of your Amazon account. Then enter the ‘name’ part
of your Kindle email address below.
Find out more about saving to your Kindle.
Note you can select to save to either the @free.kindle.com or @kindle.com variations.
‘@free.kindle.com’ emails are free but can only be saved to your device when it is connected to wi-fi.
‘@kindle.com’ emails can be delivered even when you are not connected to wi-fi, but note that service fees apply.
The development of a consistent framework for Calphad model sensitivity is necessary for the rational reduction of uncertainty via new models and experiments. In the present work, a sensitivity theory for Calphad was developed, and a closed-form expression for the log-likelihood gradient and Hessian of a multi-phase equilibrium measurement was presented. The inherent locality of the defined sensitivity metric was mitigated through the use of Monte Carlo averaging. A case study of the Cr–Ni system was used to demonstrate visualizations and analyses enabled by the developed theory. Criteria based on the classical Cramér–Rao bound were shown to be a useful diagnostic in assessing the accuracy of parameter covariance estimates from Markov Chain Monte Carlo. The developed sensitivity framework was applied to estimate the statistical value of phase equilibria measurements in comparison with thermochemical measurements, with implications for Calphad model uncertainty reduction.
This article presents a brief review of our case studies of data-driven Integrated Computational Materials Engineering (ICME) for intelligently discovering advanced structural metal materials, including light-weight materials (Ti, Mg, and Al alloys), refractory high-entropy alloys, and superalloys. The basic bonding in terms of topology and electronic structures is recommended to be considered as the building blocks/units constructing the microstructures of advanced materials. It is highlighted that the bonding charge density could not only provide an atomic and electronic insight into the physical nature of chemical bond of materials but also reveal the fundamental strengthening/embrittlement mechanisms and the local phase transformations of planar defects, paving a path in accelerating the development of advanced metal materials via interfacial engineering. Perspectives on the knowledge-based modeling/simulations, machine-learning knowledge base, platform, and next-generation workforce for sustainable ecosystem of ICME are highlighted, thus to call for more duty on the developments of advanced structural metal materials and enhancement of research productivity and collaboration.
The software package ESPEI has been developed for efficient evaluation of thermodynamic model parameters within the CALPHAD method. ESPEI uses a linear fitting strategy to parameterize Gibbs energy functions of single phases based on their thermochemical data and refines the model parameters using phase equilibrium data through Bayesian parameter estimation within a Markov Chain Monte Carlo machine learning approach. In this paper, the methodologies employed in ESPEI are discussed in detail and demonstrated for the Cu–Mg system down to 0 K using unary descriptions based on segmented regression. The model parameter uncertainties are quantified and propagated to the Gibbs energy functions.
Functionally graded materials (FGMs) in which the elemental composition intentionally varies with position can be fabricated using directed energy deposition additive manufacturing (AM). This work examines an FGM that is linearly graded from V to Invar 36 (64 wt% Fe, 36 wt% Ni). This FGM cracked during fabrication, indicating the formation of detrimental phases. The microstructure, composition, phases, and microhardness of the gradient zone were analyzed experimentally. The phase composition as a function of chemistry was predicted through thermodynamic calculations. It was determined that a significant amount of the intermetallic σ-FeV phase formed within the gradient zone. When the σ phase constituted the majority phase, catastrophic cracking occurred. The approach presented illustrates the suitability of using equilibrium thermodynamic calculations for the prediction of phase formation in FGMs made by AM despite the nonequilibrium conditions in AM, providing a route for the computationally informed design of FGMs.
Phase stability, elastic, and thermodynamic properties of (Co,Ni)3(Al,Mo,Nb) with the L12 structure have been investigated by first-principles calculations. Calculated phonon density of states show that (Co,Ni)3(Al,Mo,Nb) is dynamically stable, and calculated elastic constants indicate that (Co,Ni)3(Al,Mo,Nb) possesses intrinsic ductility. Young’s and shear moduli of the simulated polycrystalline (Co,Ni)3(Al,Mo,Nb) phase are calculated using the Voigt–Reuss–Hill approach and are found to be smaller than those of Co3(Al,W). Calculated electronic density of states depicts covalent-like bonding existing in (Co,Ni)3(Al,Mo,Nb). Temperature-dependent thermodynamic properties of (Co,Ni)3(Al,Mo,Nb) can be described satisfactorily using the Debye–Grüneisen approach, including heat capacity, entropy, enthalpy, and linear thermal expansion coefficient. Predicted heat capacity, entropy, and linear thermal expansion coefficient of (Co,Ni)3(Al,Mo,Nb) show significant change as a function of temperature. Furthermore the obtained data can be used in the modeling of thermodynamic and mechanical properties of Co-based alloys to enable the design of high temperature alloys.
Currently there are essentially two methods in use for the first-principles calculations of phonon frequencies: the linear response theory and the direct approach. The linear response theory evaluates the dynamical matrix through the density functional perturbation theory. In comparison, one advantage of the direct or mixed-space approach over the linear-response method is that it can be applied with the use of any code capable of computing forces. The direct approach is also referred to as the small displacement approach, the supercell method, or the frozen phonon approach.
However, none of the previous implementations of the supercell approach are able to accurately handle long-range dipole–dipole interactions when calculating phonon properties of polar materials. The problem has been solved by the parameter-free mixed-space approach, which makes full use of the accurate force constants from the supercell approach in real space and the dipole–dipole interactions from the linear response theory in reciprocal space. The mixed-space approach is the only existing method that can accurately calculate the phonon properties of polar materials within the framework of the supercell or small displacement approach.
The program YPHON is written in C++ and can be downloaded at http://cpc.cs.qub.ac.uk/summaries/AETS_v1_0.html. The precompiled executable binaries should work for most Linux and Windows systems. Recompiling YPHON requires the GNU Scientific Library (GSL), which is a numerical library for C and C++ programmers. If one is just interested in phonon dispersions, the phonon density-of-states (PDOS), or the neutron scattering cross-section weighted PDOS the so-called generalized phonon density-of-states (GPDOS), this is enough. YPHON also makes it a lot easier to plot phonon dispersions and PDOS. In this case, it is required that Gnuplot be installed.
The static energy and force constants from first-principles calculations need to be formatted to the YPHON input formats (text formats as detailed later). At present, YPHON works closely with VASP.5 or later. The mixed-space approach has built up a unique base of the supercell approach to polar materials and has been adopted in a number of software tools such as CRYSTAL14 by R. Dovesi, ShengBTE (a solver of the Boltzmann transport equation for phonons) by W. Li, J. Carrete, N. A. Katcho, and N. Mingo, the Phonopy package by Atsushi Togo, and the Phonon Transport Simulator (PhonTS) by Chernatynskiy and Phillpot.
The CALPHAD modeling of thermodynamics was pioneered by Kaufman and Bernstein  and has been reviewed in detail by Saunders and Miodownik  and Lukas, Fries, and Sundman . Information on features of software tools for CALPHAD modeling can be found in two series of publications in the CALPHAD journal [50, 51]. The key feature of the CALPHAD method is the modeling of the Gibbs energy of individual phases using both thermodynamic and phase equilibrium data. The main significance of the CALPHAD method is the following.
i. It enabled the development of the concept of lattice stability, i.e. the energy difference between the stable and non-stable crystal structures of a pure element.
ii.The Gibbs energy expression of each phase covers the full temperature, pressure, and composition spaces including both stable and non-stable regions of the phase. This enables the evaluation of the Gibbs energy of a system as a function of non-equilibrium state, i.e. with ξ as an independent variable.
iii.Thermodynamic data are usually obtained by measurements of heat such as the enthalpy of transition and heat capacity, as discussed in Section 4.2, which have large uncertainties typically in the range of kilojoules per mole-of-atoms. On the other hand, phase equilibrium data as discussed in Section 4.1, though more accurate, only contain information on compositions of phases at equilibria, i.e., the relative Gibbs energy of phases at equilibrium. The combination of these two sets of data is foundational in CALPHAD modeling and allows for the accurate modeling of thermodynamic properties of individual phases and reliable calculations of phase stability and driving forces.
iv. CALPHAD provides a framework to model thermodynamic properties of multi-component systems of industrial importance, enabling computational materials design. It has also been extended to model a range of properties of individual phases in multi-component systems such as diffusion coefficients, elastic coefficients, and thermal expansion, supplying input data for computational simulations of phase transformations during materials processing.
In this chapter, the basics of CALPHAD modeling of the Gibbs energy of individual phases are presented. For detailed implementations in various software packages and modeling procedures, readers are referred to the references given above.
Importance of lattice stability
For modeling of the Gibbs energy of individual phases, it is necessary to define the values of °Gi in Eq. 2.48.
A system is heterogeneous if some properties have different values in different portions of the system when the system is at equilibrium. Two scenarios may exist, where variations of the properties can be either continuous or discontinuous. In the scenario of continuous variations, the gradients of the variations must be coupled so that the system remains at equilibrium. The number of independent variables is thus reduced. These gradients must also be constrained along the boundaries between the system and the surroundings. This type of constrained equilibrium is not discussed in the book as it involves heterogeneous boundary conditions between the system and the surroundings and depends on the morphology of the system.
In the second scenario, with discontinuous variations, the properties have different values in different portions of the system, but remain homogenous within each portion. The system is in equilibrium as each portion is in equilibrium with all other portions of the system. The homogeneous portions represent different phases in the system, with the properties in each phase being homogeneous at equilibrium. In the previous chapter, it was been shown that all potentials are homogeneous in a homogeneous system.
For a heterogeneous system, the same conclusion can be obtained. If the temperature is inhomogeneous, heat can be conducted from high temperature locations to low temperature locations, and this process is irreversible based on the second law of thermodynamics because it increases the internal entropy of the system. If the pressure is inhomogeneous, the amounts of lower molar volume phases will increase to reduce the internal energy of the system. If the chemical potential of a component is inhomogeneous, the chemical potential difference of the component will drive that component to locations with a lower chemical potential in order to decrease the internal energy of the system. Therefore, it can be concluded that all potentials are homogeneous in a heterogeneous system at equilibrium, and the variables that are not homogeneous are thus their conjugate molar quantities. Under certain special circumstances, to be discussed later in this book, some molar quantities may also have the same values in difference phases.
In the previous chapter, the experimental techniques used to obtain the thermochemical and phase equilibrium data that were the inputs for the thermodynamic modeling of a system were summarized. However, experimental data are not always available. This is due to the fact that (i) the experiments are expensive, especially when they involve developing new materials, and (ii) the experiments cannot reliably access the non-stable phases in most cases. The alternative approach is to predict the thermochemical data by first-principles calculations. The prediction of material properties, without using phenomenological parameters, is the basic spirit of first-principles calculations. In particular, the steady increase of both computer power and the efficiency of computational methods have made the first-principles predictions of most thermodynamic properties possible, including both enthalpy and entropy as a function of temperature, volume, and/or pressure.
By definition, the term “first-principles” represents a philosophy that the prediction is to be based on a basic, fundamental proposition or assumption that cannot be deduced from any other proposition or assumption. This implies that the computational formulations are based on the most fundamental theory of quantum mechanics, the Schrödinger equation or density functional theory, and the inputs to the calculations must be based on well-defined physical constants – the nuclear and electronic masses and charges. In other words, once the atomic species of an assigned material are known, the theory should predict the energies of all possible crystalline structures, without invoking any phenomenological fitting parameters.
This chapter is organized in sequence from thermodynamic calculations to fundamental theory, to help those readers who are more interested in realistic calculations using existing computer codes. Detailed theoretical discussions follow the subsections on thermodynamic calculations for those readers who are also interested in the derivation of the formulations used in the thermodynamic calculations. The subsections are arranged accordingly in the order: (i) examples of the commonly adopted calculation procedures for thermodynamic properties using the elemental metal nickel as the main prototype; (ii) derivation of the Helmholtz energy expression under the first-principles framework; (iii) introduction of the solution to the electronic Schrödinger equation within two well-developed frameworks – the quantum chemistry approach and the density functional theory; (iv) detailed description of the procedure on how to solve the Schrödinger equation for the motions of atomic nuclei by means of lattice dynamics; and (v) First-principles approaches to disordered alloys.
This unique and comprehensive introduction offers an unrivalled and in-depth understanding of the computational-based thermodynamic approach and how it can be used to guide the design of materials for robust performances, integrating basic fundamental concepts with experimental techniques and practical industrial applications, to provide readers with a thorough grounding in the subject. Topics covered range from the underlying thermodynamic principles, to the theory and methodology of thermodynamic data collecting, analysis, modeling, and verification, with details on free energy, phase equilibrium, phase diagrams, chemical reactions, and electrochemistry. In thermodynamic modelling, the authors focus on the CALPHAD method and first-principles calculations. They also provide guidance for use of YPHON, a mixed-space phonon code developed by the authors for polar materials based on the supercell approach. Including worked examples, case studies, and end-of-chapter problems, this is an essential resource for students, researchers, and practitioners in materials science.
The most widely used thermodynamic modeling technique is the CALPHAD (CALculation of PHAse Diagram) method to be discussed in detail in Chapter 6. The input data in the evaluations of thermodynamic model parameters came primarily from experiments and estimations until first-principles calculations based on the density functional theory  became a user tool in the later 1990s. Experimental data include both thermodynamic and phase equilibrium data; the first-principles calculations, which provide thermodynamic data for individual phases, are discussed more extensively in Chapter 5.
Three fairly recently published books summarize the methods commonly used for experimental measurements of the thermodynamic properties of single  and multiple phases  and phase diagrams . The methods are briefly discussed here, and readers are referred to these books for details. The main techniques for crystal structure analysis include X-ray diffraction, electron backscatter diffraction (EBSD), electron diffraction in transmission electron microscopy, neutron scattering, and synchrotron scattering, which are not discussed in this book.
Phase equilibrium data
The most common method to determine phase equilibria is to use equilibrated materials. This method typically involves material preparation through high temperature melting or powder metallurgy, homogenization heat treatment, isothermal or cooling/heating procedures, and identification of crystal structures and phase compositions. It is important to avoid macro-inhomogeneity as it can be difficult to remove the inhomogeneity in subsequent treatments. It is also important to use starting materials of the highest purity and to minimize the loss and contamination of materials during the entire experiment using a protective atmosphere of inert gas or vacuum. Typical melting techniques include high temperature furnaces with crucibles, arc melting, and induction melting. Attention needs to be paid to possible reactions between materials and crucibles/containers, which can be avoided by levitating the materials by electromagnetic fields or other means. In addition to using pure elements as raw materials, master alloys with well-controlled compositions are often utilized because the compositions and melting properties of master alloys are usually much closer to those of the final materials than the pure elements. For materials with very high melting temperature or volatile components, the powder metallurgy method can be used where compacts are made, capsulated, and sintered.
Homogenization during subsequent heat treatment is achieved through diffusion, in which time and temperature are two important parameters.