Amorphous alloys of the (Fe-Co-Ni)-(Cr-Mo-Nb)-B system are promising materials to supply the demands for higher wear resistance components in the petrochemical industry. Since the development of the CALPHAD method, the development of new metallic alloys has been accompanied by thermodynamic modelling and calculations. The prediction of the formation of amorphous alloys requires special care with the modeling of the liquid and or an amorphous phase. As a initial stage in the more complex system, the basic Fe-Nb-B ternary system was selected. In order to predict the stability and tendencies of transformations of these amorphous alloys, the Fe-Nb-B system was reassessed using Ågren's two-state model to describe the liquid phase. The results of the present assessment show very good agreement with the recently reported stable phase diagram. Furthermore, the use of the two-state model for the liquid is more accurate and physically consistent when evaluating transformations from supercooled liquid, as shown it the present work.
The influence of external stresses on the transformation behaviour has been studied for a polycrystalline Fe-3 IMn-SSi alloy exhibiting the y-+& martensitic transformation. Samples have been loaded at temperatures above the M, temperature, cooled with a constant external load to a temperature below the M, temperature and subsequently heated without load to a temperature above the A, temperature of the alloy. Stress, strain and temperature have been continuously measured and transformation temperatures and strains have been determined from the change in strain during the transformation. It has been found that the M, temperature during cooling with load is not influenced by 0.3-2 % plastic pre-deformation of the austenite whereas there is a substantial increase in the transformation strain, i.e. variants with a favourable orientation are enhanced by pre-strain of austenite.
A unified thermodynamic analysis has been developed which allows the combined complex effects of applied stresses and alloy composition on the martensite start temperature, Ms(σ), to be explained and predicted. In this unified analysis, a thermodynamic analysis of the effect of applied stress on the relative stability of the high and low temperature phase is combined with existing thermodynamic descriptions of the alloy system. The calculated results are compared with experimental results for the γ→ martensitic transformation in Fe–Mn–Si shape memory alloys. For single crystals, the comparison shows that the influence of applied stresses can be directly predicted with a reasonable accuracy for a wide composition range. By integrating a polycrystallinity parameter in the analysis it is shown that the calculations are also valid for polycrystalline alloys. Application of this analysis to experimental values of Ms(σ) can also generate useful data for future optimizations of the thermodynamic description of several alloy systems.
The effects of recrystallization and grain size have been studied in an Fe-31%Mn-5%Si shape memory alloy. The amount of martensite formed in samples which have been hot rolled at 1373 K and subsequently homogenized for 24 h at 1273 K followed by a final 30 min anneal at 1323 K is twice as large as in samples which have been hot rolled and homogenized in the same way as the former followed by a compression and a subsequent recrystallization at 1323 K for 30 min. The only difference between the samples is the plastic deformation induced by the compression. We thus conclude that some effect of the compressive deformation remains despite the recrystallization and suppresses the subsequent martensite formation. However, we have not yet been able to identify this effect. This matter will be subject to further investigations.
No significant change is found in the M(S) and A(S) temperatures for the gamma <----> epsilon transformations determined by DSC. When the grain size varies between 40 and 160 mu m in completely recrystallized samples the variation in transformation temperatures is small, less than 10 K.
The effect of training and predeformation on the martensitic transformation is investigated. It is concluded that MS as a function of predeformation strain must have a maximum at around 3% strain. Prestrains less than 5% will enhance the martensitic transformation whereas larger prestrains depress it due to strain hardening of the γ phase.
A phase-field model of sintering and related phenomena in a two-phase system and in a multi-phase system is presented. We consider diffusion of vacancies as the atomic mechanism for redistribution of material and we will use the familiar model of thermal vacancies in a crystal as our energy formulation. The solid material will thus be characterized by a low vacancy content and the surroundings by a very high vacancy content and a very low content of atoms. The surface of the solid body will be characterized by a continuous variation in vacancy content. The temporal development of particles during solid state sintering with effects such as wetting is shown in various simulations.
Recently the present authors presented a Phase-field model of sintering in a multiphase system. In the present contribution the major features of the model are summarized. The model is based on diffusion of vacancies as the atomic mechanism for redistribution of material. The solid material is characterized by a low vacancy content and the surroundings by a very high vacancy content and a very low content of atoms. The surface of the solid body is characterized by a continuous variation in vacancy content. The temporal development of particles during solid state sintering with effects such as wetting has been shown previously and here we discuss the effect of a highly anisotropic interfacial energy on the morphological evolution of the particles.
The Cahn–Hilliard equation is solved with thermodynamic and kinetic input, using the Thermo-Calc and DICTRA software packages rather than simpler models e.g. regular solution. A concentration dependent expression for the gradient energy coefficient is introduced and its effect on simulated decomposition is discussed. Simulations were carried out in 2D and 3D using the FiPy software package modified for non-linear problems.
Creep of 9-12% Cr steels is modeled using the composite model, developed by Blum et al. [R. Sedlacek, W. Blum, Comput. Mater. Sci. 25 (2002) 200], and thermodynamic calculations based on the Calphad approach. The composite model yields a physical description of the deformation behavior of materials that have a pronounced heterogeneous dislocation structure and is briefly surveyed. A few of the input parameters have been thermodynamically calculated using Thermo-Calc and introduced to the main program via a programming interface. This combined approach allows us to simulate the creep deformation behavior with less extensive microstructural investigations. This is a step towards enabling predictions of the creep behavior predominantly based on the nominal composition, heat treatment and mechanical load. Simulation results for two different 9-12% Cr steels are presented.
We describe current approaches to thermodynamic modelling of liquids for the CALPHAD method, the use of available experimental methods and results in this type of modelling, and considerations in the use of atomic-scale simulation methods to inform a CALPHAD approach. We begin with an overview of the formalism currently used in CALPHAD to describe the temperature dependence of the liquid Gibbs free energy and outline opportunities for improvement by reviewing the current physical understanding of the liquid. Brief descriptions of experimental methods for extracting high-temperature data on liquids and the preparation of undercooled liquid samples are presented. Properties of a well-determined substance, B-2 O-3, including the glass transition, are then discussed in detail to emphasize specific modelling requirements for the liquid. We then examine the two-state model proposed for CALPHAD in detail and compare results with experiment and theory, where available. We further examine the contributions of atomic-scale methods to the understanding of liquids and their potential for supplementing available data. We discuss molecular dynamics (MD) and Monte Carlo methods that employ atomic interactions from classical interatomic potentials, as well as contributions from ab initio MD. We conclude with a summary of our findings.
A mean-field model dealing with prismatic grain growth during liquid phase sintering of cemented carbides with a Co-rich binder is presented. The evolution of the size of an assembly of non-spherical grains is obtained using a Kampmann-Wagner approach and by introducing a constant shape factor between the characteristic lengths of prisms. This factor is a function of interfacial energies of the two kind of facets, basal and prismatic, considered. The growth model is based on three different mechanisms, that can be rate limiting, taking place in series: 2D nucleation of a new atomic layer, mass transfer across the interface and long-range diffusion. The driving force for coarsening is distributed between the different facets. These equations are solved numerically, and the simulation results reveal that the specific abnormal grain growth phenomena experimentally observed in cemented carbides may be reproduced with this new more realistic description of the grain shape contrary to the spherical approach developed in the past. It is also shown that the initial powder size distribution, and more specifically its shape has a strong influence on the distribution of the driving force between the different rate limiting mechanisms and thus on the occurrence of abnormal grain growth. In that case, the self-similarity of the normalized grain size distribution over time is not achieved.
Nucleation kinetics in a multicomponent supersaturated solid solution is examined. Attachment rate of atoms to a nucleus of a size close to the critical one is determined combining a thermodynamic extremum principle and the Fokker-Planck equation. Two limiting cases are examined; when bulk diffusion controls the nucleation kinetics and when the process is limited by the interfacial mobility. The mixed regime is also treated. Moreover, the growth law in multicomponent alloys is derived in the general case, when both mechanisms are considered. Additionally, the attachment rate is derived, in the classical framework, from a new macroscopic growth equations and the fundamental role of the interfacial mobility is examined. These new general expressions, for the attachment rates and the growth laws, determined either applying the thermodynamic extremum principle or derived from the classical formalism are found to be consistent.
In the present paper, a general survey of the diffusion-controlled transformations (DICTRA) software is given. DICTRA is an engineering tool for diffusion simulations in multicomponent alloys. The simulations are based on multicomponent diffusion and thermodynamic data, both obtained by analyzing and assessing experimental information. This allows for many different cases to be studied as soon as the underlying data are available. DICTRA is not a complete simulation tool because only geometries that can be transformed into one space variable can be treated, but many well posed problems of practical interest may be solved. The program contains several different models, which are discussed in the present paper. Each model has its own applications and several examples from recent simulations are given in order to demonstrate the usage of the particular models.
Information on the diffusional transformation products of austenite in high-carbon steels is reviewed and supplemented with new microscopic studies. A comparison with transformation products in low-carbon steels indicates that there is a symmetry with pearlite in the middle, where ferrite and cementite are equal partners, and with acicular ferrite or cementite on each side. They both form with a surface relief, and at lower temperatures, each one is the leading phase in a eutectoid microstructure, bainite and inverse bainite, respectively. However, there is an asymmetry because at low temperatures bainite appears in high-carbon steels but inverse bainite never appears in low-carbon steels. At a constant high carbon content, there is another kind of symmetry, which is related to temperature. At intermediate temperatures the eutectoid reaction results in spherical nodules in which the cementite constituent originates from Widmanstatten plates. It turns spiky at both higher and lower temperatures with the leading phase in the spikes being cementite at higher temperatures and ferrite at lower temperatures. In the first kind of symmetry, there is an abrupt change among the three reaction products; in the second kind of symmetry, there is a gradual change. Accepting that all the eutectoid microstructures form by diffusion of carbon, one may explain the existence of both symmetries by the variation of the ratio of the supersaturations of ferrite and cementite with carbon content and with temperature.
Ms may be defined as the temperature below which the formation of martensite starts upon cooling. It may also be useful to define Mg, the temperature below which martensite can grow if it is already nucleated. In order to analyze the mechanism of martensite formation, it is essential to know the difference Mg - Ms. We have tried to evaluate Mg - Ms for an Fe-C alloy with a decarburized surface zone in order to induce nucleation. The samples were studied by means of electron microprobe, serial sectioning and optical microscopy. The results indicate that Mg is surprisingly close to Ms. The possibility that Mg is controlled by growth rather than nucleation is discussed.
There are two paradigms regarding the formation of bainite. One is based on the first stage being rapid, diffusionless growth or acicular ferrite and the subsequent formation of carbide occurring by precipitation from the supersaturated ferrite. All assumption that the first stage occurs as a series of subsequent rapid steps resulting in sub-units plays an important role as an explanation of the not so rapid growth observed macroscopically. The other paradigm is based on the first stage being the formation of acicular ferrite under carbon diffusion and on the subsequent growth of carbide and ferrite side by side. Metallographic observations are presented that support the second paradigm. It is difficult to see how they can be accounted for by the first paradigm, in particular the observation of the shapes of sub-units.
Synthesis and phase separation of (Ti,Zr)C were investigated in the present work. The (Ti,Zr)C phase was synthesized at 2200 C and subsequently aged at 1300 C for different times. The microstructure was investigated using X-ray diffraction and electron microscopy, and supplemented by first-principles calculations. The (Ti,Zr)C phase separates into a lamellar nanostructure with alternating Ti- and Zr-rich face-centered cubic domains as well as non-stoichiometric TiC and ZrC. The lamellar structure is a consequence of phase separation within the miscibility gap that is directionally constrained by high coherency stresses, as indicated by the first-principles calculations. Moreover, the increased hardness due to the phase separation suggests that the mixed carbide could be used as a strengthening constituent in, for example, cemented carbides.
Liquid-phase sintering is an important step in the production of cemented carbides. During sintering, the average WC grain size increases, leading to a coarser structure, which affects the performance of the final product. The coarsening occurs by dissolution of small grains and growth of large grains. In the present work, the effect of high carbon activity during sintering on the WC grain coarsening has been evaluated using electron backscattered diffraction (EBSD) and the results have been compared with a previous work where sintering was performed at a lower carbon activity. A more homogeneous grain size distribution was observed in alloys sintered at a high carbon activity. In addition, the effect of the initial powder particle size distribution was investigated. It was found that the coarsening rate of a WC powder with an initial small average grain size is significantly higher as compared to the coarsening rate for a powder with a larger initial average grain size. The results obtained emphasize the importance of considering the complete particle size distribution in order to predict coarsening.
In the present work, the size distribution and shape of WC grains in cemented carbides (WC-Co), with different Co contents, have been investigated in three dimensions. Direct three-dimensional (3-D) measurements, using focused ion beam serial sectioning and electron backscattered diffraction (EBSD), were performed and a 3-D microstructure was reconstructed. These measurements were supplemented by two-dimensional (2-D) EBSD and scanning electron microscopy on extracted WC grains. The data from 2-D EBSD collected on planar sections were transformed to three dimensions using a recently developed statistical method based on an iterative inverse Saltykov procedure. This stereological analysis revealed that the assumed spherical shape of WC grains during the Saltykov method is reasonable and the estimated 3-D size distribution is qualitatively in good agreement with the actual distribution measured from 3-D EBSD. Although the spherical assumption is generally fair, the WC grains have both faceted and rounded surfaces. This is a consequence of the relatively low amount of liquid phase during sintering, which makes impingements significant. Furthermore, the observed terraced surface structure of some WC grains suggests that 2-D nucleation is the chief coarsening mechanism to consider.
The properties of cemented carbides strongly depend on the WC grain size and it is thus crucial to control coarsening of WC during processing. The aim of this work was to study the effect of sintering at different carbon activities on the final microstructure, as well as the coarsening behavior of the WC grains, including the size distribution and the shape of WC grains. These aspects were investigated for five WC-Co alloys sintered at 1410 C for 1 h at different carbon activities in the liquid, in the range from the graphite equilibrium (carbon activity of 1) to the eta (M6C) phase equilibrium (carbon activity of 0.33). The grain size distribution was experimentally evaluated for the different alloys using EBSD (electron backscatter diffraction). In addition, the shape of the WC grains was evaluated for the different alloys. It was found that the average WC grain size increased and the grain size distribution became slightly wider with increasing carbon activity. Comparing the two three-phase (WC-Co-eta and WC-Co-graphite) alloys a shape change of the WC grains was observed with larger grains having more planar surfaces and more triangular shape for the WC-Co-graphite alloy. It was indicated that in alloys with a relatively low volume fraction of the binder phase the WC grain shape is significantly affected by impingements. Moreover, after 1 h of sintering the WC grains are at a non-equilibrium state with regards to grain morphology.
The microstructure of cemented carbides with a gradient structure at the surface consists of WC, cubic carbonitrides and a binder phase. The carbonitrides can, for example, consist of Ti(C,N)-Zr(C,N) where it is reasonable to believe that there is a miscibility gap with Ti-rich and Zr-rich carbonitrides. In the present work, the effect of the N-2-gas pressure on the equilibrium composition of the miscibility gap in the (Ti,Zr)(C,N) system has been investigated. In the study, the carbonitride system is in equilibrium with: WC, liquid binder, graphite and, N-2-gas of different pressures. Both Fe and Co are used as binder phase to study the effect of the binder phase. The results verify that there is a miscibility gap in the carbonitride system and that the region of the miscibility gap will change when N is introduced. There is a critical N-2-gas pressure lower than 0.1 bar and above that pressure the compositions of the carbonitride are rather constant as a result of the formation of a surface rim.
A new thermodynamic database has been combined with an existing kinetic database to perform coarsening simulations in ternary systems including MC and M7C3 carbides in an fcc matrix. The kinetic database was revised taking into consideration the new experimental information on the Fe-Cr-V-C system obtained in the present work, and available experiments on the ternary Fe-Cr-C and Fe-V-C systems. After revision the agreement between experimental results and simulations was satisfactory. It was found that the interfacial energy of M7C3 was twice as large as that of the MC carbide. The calculations for commercial steels with 6 alloy elements gave results in satisfactory agreement with new experimental measurements. The present coarsening simulations use the calculated equilibrium state and the observed particle sizes as the state for the start of the simulations. All the simulations were performed with the DICTRA software.
New interrupted cooling experiments have been designed to study the kinetics of bainitic ferrite formation starting from a mixture of austenite and bainitic ferrite. It is found that the kinetics of bainitic ferrite formation during the cooling stage is determined by the isothermal holding time. The formation rate of bainitic ferrite at the beginning of the cooling decreases with increasing prior isothermal holding time. An unexpected stagnant stage during the cooling stage appears when the isothermal holding time increases to a critical point. There are two reasons for the occurrence of the stagnant stage: (i) a solute spike in front of the interface; and (ii) kinetic transition. A so-called Gibbs energy balance approach, in which the dissipation of Gibbs energy due to diffusion inside the interface and interface friction is assumed to be equal to the available chemical driving force, is applied to theoretically explain the stagnant stage. A kinetics transition from a fast growth mode without diffusion of Mn and Si inside the austenite-bainitic ferrite interfaces to a slow growth mode with diffusion inside the interface is predicted. The stagnant stage is caused by the transition to a slow growth mode. The Gibbs energy balance approach describes the experimental observations very well.
We propose an approximate growth rate equation that takes into account both cross-diffusion and high supersaturations for modeling precipitation in multicomponent systems. We then apply it to an Fe-alloy in which interstitial C atoms diffuse much faster than substitutional solutes, and predict a spontaneous transition from slow growth under ortho-equilibrium to fast growth under the non-partitioning local equilibrium condition. The transition is caused by the decrease in the Gibbs-Thomson effect as the growing particle becomes larger. The results agree with DICTRA simulations where full diffusion fields are calculated.
In multicomponent systems the diffusion coefficient turns into a matrix. The diagonal elements represent diffusion of a species caused by its own concentration gradient. In a thermodynamically stable binary alloy it is easy to see that this diagonal element must be positive but in a multicomponent system it is less obvious. The sign of the diagonal elements in the general case is discussed in this report. It is shown that the sign of an individual diagonal element has no physical meaning but can be changed by changing the dependent concentration variable. Only the sum of all the diagonal elements need to be positive in a stable system.
Complex equilibria and phase transformations involving diffusion can now be calculated quickly and efficiently. Detailed examples are given for cases which involve varying degrees of non-equilibrium and therefore time-dependence. Despite very good agreement between such calculations and experimental results, many potential end-users are still not convinced that such techniques could be usefully applied to their own specific problems. Friendly graphic interface versions of calculating software are now generally available, so the authors conclude that the most likely source of the reluctance to use such tools lies in the formulation of relevant questions and the interpretation of the results. Although the potential impact of such tools was foreseen many years ago [M. Hillert, Calculation of phase equilibria, in: Conference on Phase Transformations, 1968], few changes in the relevant teaching curricula have taken into account the availability and power of such techniques. This paper has therefore been designed not only as a collection of interesting problems, but also highlights the critical steps needed to achieve a solution. Each example includes a presentation of the "real" problem, any simplifications that are needed for its solution, the adopted thermodynamic formulation, and a critical evaluation of the results. The availability of such examples should facilitate changes in subject matter that will both make it easier for the next generation of students to use these tools, and at the same time reduce the time and effort currently needed to solve such problems by less efficient methods. The first set of detailed examples includes the deoxidation of steel by aluminum; heat balance calculations associated with ladle additions to steel; the determination of conditions that avoid undesirable inclusions; the role of methane in sintering atmospheres; interface control during the physical vapour deposition of cemented carbide; oxidation of gamma-TiAl materials; and simulation of the thermolysis of metallorganic precursors for Si-C-N ceramics and interface reaction of yttrium silicates with SiC-coated C/C-SiC composites for heat shield applications. A second set of examples, more dependent on competitive nucleation and growth, includes segregation and carburization in multicomponent steels and features a series of sophisticated simulatons using DICTRA software. Interfacial and strain energies become increasingly important in defining phase nucleation and morphology in such problems, but relatively little information is available compared to free energy and diffusion databases. The final section therefore demonstrates how computational thermodynamics, semi-empirical atomistic approaches and first-principles calculations are being used to aid filling this gap in our knowledge. (c) 2006 Elsevier Ltd. All rights reserved.
Different geometrical models of allotriomorphic growth of ferrite in undercooled austenite are investigated by means of numerical and analytical treatments of diffusional growth under local equilibrium. The results obtained by the numerical method are compared with analytical solutions for those cases where such solutions may be derived. An excellent agreement is obtained. The numerical method is subsequently applied to simulate the experiments by Aaronson et al. and some more recent experiments by Hougardy et al.. taking into account the concentration dependence of the diffusivity of C in austenite and the most recent thermodynamic assessment of the ferrite/austenite equilibrium. Taking into account the experimental uncertainties we conclude that the growth of allotriomorphic ferrite must be essentially controlled by long-range carbon diffusion in austenite.
Some numerical methods for solving a Stefan problem are discussed and compared with the exact solution. The growth of a planar particle from a supersaturated solution (or solidification from a supercooled liquid) is considered. It is found that the Murray-Landis method, based on a finite difference technique to solve the diffusion equation on a contracting grid, yields a poor accuracy for high supersaturations. The enthalpy method, also based on the finite difference technique and an interpolation formula for obtaining the interface position, shows a satisfactory performance at high supersaturations but a less satisfactory one at low supersaturations. It is demonstrated that the poor accuracy of the Murray-Landis method depends on the application of a less accurate flux-balance equation for finite time increments and the procedure for displacing the grid points. A modification of the Murray-Landis method is developed and is found to have superior numerical performance.
Compound layers developed at 848 K during gaseous nitrocarburizing of iron and iron-carbon specimens were investigated for several combinations of N and C activities imposed at the specimen surface by gas mixtures of NH3, N-2, CO2, and CO. The microstructural evolution of the compound layer was studied by light microscopy and by X-raydiffraction analysis. Composition-depth profiles were determined by electron probe (X-ray) microanalysis. Layer growth kinetics was investigated by layer thickness measurements. The influence Of the N and C activities on the microstructural and compositional evolution and the growth kinetics of the compound layers formed is discussed for the iron substrate. The results indicate that the microstructure is governed by a fast C and a slow N absorption at the surface in an early stage of gaseous nitrocarburizing. The influence of carbon in the substrate on the microstructural and compositional evolutions and on the growth kinetics was evaluated from comparing the results obtained for: a: normalized Fe-0.8C alloy with those for iron under identical nitrocarburizing conditions.
The formation of surface zones with a composition gradient during sintering of WC-Ti(C,N)-Co cemented carbides has been studied experimentally and by computer simulations. The microstructure has been investigated with SEM and EPMA. The simulations are based on a solution of the multicomponent diffusion equations, coupled with thermodynamic calculations using thermodynamic descriptions of the individual phases. The results from the simulations are in good agreement with the experimental results, indicating that diffusion and the thermodynamic properties are the two major factors that control the gradient structure formation.
A general model to treat multicomponent diffusion in multiphase dispersions is presented. The model is based on multicomponent diffusion data and basic thermodynamic data and contains no adjustable parameters. No restriction is placed on the number of components or phases that take part in the calculations, as long as the necessary thermodynamic and kinetic data are available. The new model is implemented into the DICTRA software, which makes use of THERMO-CALC to handle the thermodynamics. The model is applied to carburization of Ni alloys and heat treatment of welded joints between dissimilar materials. In both cases, the diffusion is accompanied by carbide formation or dissolution. A good agreement between experiments and calculations is found, despite the fact that no adjustable parameters are needed.
Nb(C,N) precipitates were studied in a niobium-stabilised stainless steel (AISI 347) statically aged at 700 degrees C. Scanning electron microscopy and energy filtered transmission electron microscopy were used to determine the volume fraction and precipitate size of primary and secondary Nb(C,N) after ageing times between 0 and 70,000 h. The experimental data were correlated with simulations of Nb(C,N) formation based on the assumption that the process is controlled by diffusion. These simulations provide a rationale for the existence of two sets of mobium carbonitrides in commercial tubes of AISI 347. Growth of primary Nb(C,N) occurred essentially during manufacturing, with no significant growth at 700 degrees C. Rapid dissolution and re-precipitation of secondary Nb(C,N) occurred during manufacturing. Coarsening at 700 degrees C of secondary particles was modelled using the Lifshitz-Slyozov-Wagner theory, which overestimated the coarsening rate. These problems were overcome with a model developed by the authors. This model takes both growth and coarsening into account.
Precipitation phenomena in type 347 austenitic stainless steels have been investigated after long-term heat treatment and creep in air at 700 and 800 degreesC. Nitrogen uptake was observed during long-term creep testing at 800 degreesC. No such effect was observed at 700 degreesC although times up to about 70,000 h were used. The major phases precipitated after long time exposure at 800 degreesC were primary Nb(C,N), Z-phase, Cr2N and M23C6, while primary Nb(C,N), secondary Nb(C,N) and sigma-phase were the major phases at 700 degreesC. Z-phase precipitated in both intragranular and intergranular form at 800 degreesC. Large precipitates exhibiting a core/rim structure showed a rim of Z-phase surrounding undissolved primary Nb(C,N). The microstructural evolution during creep deformation in air at 800 degreesC was modelled thermodynamically. The model satisfactorily predicts nitrogen uptake and the essential features of the evolution of the microstructure with time. The precipitation sequence could be qualitatively described, although it was not possible to model the formation of all precipitates.
A comparison was made between two experimental methods to determine the (T-phase volume fraction and three methods to model a-phase growth in a niobium-stabilized stainless steel (AISI 347). The a-phase volume fraction and precipitate size were determined in material statically aged and creep deformed at 700 degrees C with both KOH etched specimens using bright field optical microscopy (OM/BF) (conventional method) and specimens etched with oxalic acid using scanning electron microscopy and backscattered electrons (SEM/BSE) (new method). Both experimental methods used manual thresholding together with digital image analysis. The calculations were made with DICTRA software, using both the TCFE database and the SSOL database with some modification concerning the effect of silicon on the stability of sigma-phase particles. The modeled sigma-phase volume- fractions showed rather good agreement with the measured results from statically aged material using the new method. It was found that the stabilizing effect of silicon on sigma phase should be included in the thermodynamic database used for modeling.
The thermodynamic properties of the Fe-Mn-Si system are analyzed by means of thermodynamic models for the individual phases. Special attention is paid to the γ → ε martensitic transition. A complete set of parameters, from which arbitrary sections of the phase diagram as well as the Ms and As temperatures may be calculated, is given.
Cemented carbide cutting tool inserts with a surface zone depleted of hard cubic carbides and enriched in ductile binder phase have been studied experimentally using scanning electron microscopy (SEM), electron probe micro analysis (EPMA) and transmission electron microscopy (TEM). The results are compared with simulations based on a solution of the multicomponent diffusion equations, coupled with calculations using thermodynamic descriptions of the individual phases. The materials in the study are based on WC-Ti(C,N)-Co, WC-Ti(C,N)-NbC-Co and WC-Ti(C,N)-TaC-Co. The surface zone is formed by creating a gradient in nitrogen activity in the material, leading to an outward diffusion of N. Due to thermodynamical coupling between N and Ti, the outward diffusion of N will lead to an inward diffusion of Ti, and a surface zone depleted in cubic carbides is formed. Additions of elements like Ta or Nb will affect the width of the surface zone. A material with a Ta-containing cubic phase will have a narrower surface zone than a material with a Nb-containing cubic phase. Ta or Nb additions also affect the distribution of the different phases in and adjacent to the surface zone.
To increase the cutting performance of WC-MC-Co cemented carbide tools, it is common to use a high temperature CVD process to coat them with thin wear resistant layers. During the process cracks are unavoidably introduced in the coating. To prevent crack propagation it is of interest to create a tough surface zone in the substrate, enriched in WC and binder phase. A way to create such a zone is to sinter a nitrogen-containing cemented carbide in a nitrogen free atmosphere. This formation of gradient structures has been extensively studied using microscopy and simulations, and it has been shown that the process is driven by diffusion in the binder phase. However, the diffusion paths are partly blocked by the dispersed particles. This effect can be formally handled by considering effective diffusivities by introducing a so-called labyrinth factor, lambda. In prior work it has been assumed that lambda =f(2) where f is the volume fraction of the binder. The validity of this assumption has been studied by simulations and experimental analysis of gradient sintered WC-Ti(C,N)-Co cemented carbides containing 5.0, 6.7, 10.0 and 20.0 vol% binder phase. It was found that by using the labyrinth factor lambda = f(2) instead of a better correspondence between experiments and simulations can be achieved.
Solid state phase transformations in metals, and more precisely the science of transformation interfaces, is a key point to understand the formation of nano/microstructure, and thus, as a result, many physical properties such as mechanical properties, conductivity, thermoelectric and magnetic properties of materials. Steels are by far the most widely used metallic alloys, and a deep understanding of their microstructure is essential to tailor their service properties. The transformation of high temperature parent austenite to ferrite is one of the main issues controlling the final microstructures, and for more than a century, this has driven metallurgists to investigate in detail this solid state transformation, and, particularly, the details of austenite to ferrite interface migration. In this paper, we review the evolution of the different concepts and experiments developed in the last century to investigate this transformation mechanism. After a brief introduction, most of the physical models developed, which reduce the α/γ interface into a mathematical body with its own properties, are reviewed and discussed with regard to experimental data. The increased availability of highly sophisticated experimental and modelling tools in recent decades has considerably clarified the perceptions of transformation interfaces. These recent advances are presented, and their contribution to the field of migrating austenite-ferrite interfaces are highlighted in a third section. In the fourth section, the latest developments in experimental methods, which now allow the quasi atomistic direct characterization of the interface chemistry, are presented. The observed conditions at the interfaces can be compared with model predictions, which is believed to be a critical step for the refinement of the theoretical concepts guiding the understanding of the interface migration. Finally, in the concluding section, the present situation of the field is summarized, and some perspectives regarding the expected future developments are sketched.
Using a CALPHAD approach together with a Cahn-Hilliard model, we describe the microstructure evolution in cubic Ti1-xAlxN including vacancies on the metal sublattice. Our results show that vacancy content has a pronounced effect on the decomposition kinetics. Furthermore, vacancies show a strong tendency to segregate to the coherent AlN-TiN interface regions. We illustrate how vacancies anneal to grain boundaries, and finally, we compare our prediction to experimental differential scanning calorimetry data and attribute the second peak in the thermogram to vacancy depletion.
it has long been known that a cobalt addition increases the resistance to tempering in steels. This may be due to the fact that Co raises the Curie-temperature which retards diffusion. In the present work the effect of Co on coarsening of M23C6 in the 9% Cr steel P92 is studied by computer simulations. The results show that a final average radius of the carbides after 30 000 h at 600 degreesC decreases with 30 % with a Co addition of 10 mass %. This raises the Orowan stress with 30 %. Moreover, ii is assumed that slower particle coarsening also leads to a retarded coarsening of the martensite lath structure.
The influence of oxygen on the sintering behavior of WC-Co has been investigated by Auger electron spectroscopy (AES) and scanning tunneling microscopy (STM). Deposition of Co on the WC(0001) surface and subsequent annealing at 650 degreesC results in a 2x2 reconstructed pre-cursor layer on top of which Co grows in weakly bound islands which can be moved on the surface by the STM rip. Annealing at 850 degreesC removes excess Co and leaves only the 2x2 surface. Oxygen exposure of the 2x2 surface results in a clustered cobalt oxide overlayer which on annealing at 750 degreesC breaks up and restores the 2x2 structure as the metallic Co wets the surface.
Solid state sintering of cemented carbides is studied experimentally by means of dilatometry. Both an ordinary dilatometer and a dilatometer where an axial load was applied on the specimen were used. It is concluded that increased axial load, due to its hydrostatic component, gives larger volumetric shrinkage and that the decrease in the shrinkage rate at isothermal treatment is a result of a rapidly increasing rigidity of the powder compact at isothermal temperatures.
Solid state sintering of cemented carbides is studied experimentally by means of dilatometry and SEM. Surface analysis of the powders is performed by means of ESCA, Auger and SIMS. It is concluded that for pure WC-Co mixtures the major part of densification occurs in the solid state and that high heating rates yield higher densification rates. 200 ppm of Ti, preferentially located at grain surfaces, yield a drastic decrease in sintering rate. Melting of the binder, i. e. transition to liquid phase sintering, only has a minor effect on densification rate.
The Ni-Ru and Al-Ni-Ru systems are assessed with a combined CALPHAD and ab initio approach. Particular attention is paid to the possible existence of a miscibility gap in the B2 phase. Both face-centered cubic and body-centered cubic ordering are analyzed within the compound energy formalism. Ab initio calculations for the B2 phase show a similar trend as calorimetric measurements but the magnitude is much smaller. It is found that the calorimetric measurements cannot be reconciled with any reasonable phase diagram, whereas the ab initio results can. From the parameters obtained, isothermal sections in reasonable agreement with experimental phase diagrams are calculated. We have concluded that there is no miscibility gap in the B2 phase at 1273 K and higher temperatures.
Using CALPHAD methods, Cr2O3 growth on pure Cr is modeled using DICTRA and the vacancy model for diffusion. The results are compared with thermobalance and TEM experiments at 625 and 700 degrees C in O-2. The experimental scatter is significant, leading to a compromise suggestion. With the experimental conditions from the furnace exposures in this work, optimized mobilities are validated with a series of oxidation simulations. Despite the complex microstructure and initial growth rate variations, it is possible to reproduce the experimental oxide thicknesses with good accuracy, allowing for extension to multicomponent systems.
A long missing feature in the diffusion simulation software DICTRA has been the diffusion in oxides. Recently, this capability was implemented and in this report we present some results of the current achievements.DICTRA can now treat diffusion in basically any oxide, provided that there is diffusion data available and that the necessary mobilities have been assessed. The first system to be addressed was the important Fe-O system. It contains three different oxides, from the simple wustite to the complex magnetite, which has the spinel structure. The current approach was successful and the work was continued with the Cr-O system and the technically important Fe-Cr-O. We will present the diffusion models and the results of the assessment, as well as successful simulations of oxidation where layers of oxides grow on top of an Fe or Fe-Cr substrate.