The thermodynamics of martensitic transformation

The thermodynamic approach to the understanding of martensitic transformations can provide insight into the mechanism of transformation, serve as a generalization technique among the alloying systems, and as a framework for the synthesis of data on kinetic, structural, crystallographic, and other aspects of the transformation. The quality and utility of existing thermochemical models and data are examined. The approach requires long, unverifiable extrapolation of the thermodynamic functions and must be used with circumspection. On the other hand, much that is known is derived from a thermodynamic viewpoint. The present authors are not hopeful of great breakthroughs in the quality of the thermodynamic data. More significant, it appears, would be attempts to formally cast the areas of crystallographic analysis (transformation strain energy) and the like into the thermodynamic framework.     

On the nature of the martensitic transformation

A novel conceptual approach to martensitic transformation is presented. The current concepts are discussed briefly, and a more general qualitative approach to this type of phase change is proposed. It is suggested that the transformation originates in lattice vibration modes, and that once initiated it proceeds adiabatically with no discernible division into nucleation and growth. The transformation does not occur through a dislocation mechanism, and the high dislocation densities observed as associated with martensite formation consist of immobile dislocations. These form to accommodate the transformation strains which are highly localized, but are also self-accommodating and on completion of the transformation are macroscopically nondirectional. All of the anomalies in the mechanical properties reported in the past are ascribed to the effects of an applied stress on the transformation. It is suggested that investigations of such effects on transformation, particularly in single crystal material, should be most fruitful in providing new fundamental insight into the mechanism of the transformation.     

On the nature of the martensitic transformation

A novel conceptual approach to martensitic transformation is presented. The current concepts are discussed briefly, and a more general qualitative approach to this type of phase change is proposed. It is suggested that the transformation originates in lattice vibration modes, and that once initiated it proceeds adiabatically with no discernible division into nucleation and growth. The transformation does not occur through a dislocation mechanism, and the high dislocation densities observed as associated with martensite formation consist of immobile dislocations. These form to accommodate the transformation strains which are highly localized, but are also self-accommodating and on completion of the transformation are macroscopically nondirectional. All of the anomalies in the mechanical properties reported in the past are ascribed to the effects of an applied stress on the transformation. It is suggested that investigations of such effects on transformation, particularly in single crystal material, should be most fruitful in providing new fundamental insight into the mechanism of the transformation.     

The effects of applied stress on the martensitic transformation in TiNi

The deformation behavior of TiNi at 20°C has been investigated as a function of composition and of the prior heat treatment. Wide mechanical property variation and significant differences between the effects of tensile and of compressive loading were observed. Under some conditions anelastic behavior, characterized by a broad hysteresis loop, was reproducibly obtained. The effects of heating the deformed materials aboveA _f on the subsequent stress-strain behavior indicate the anomalies observed to be directly related to the martensitic transformation. The effects of stress application on the martensitic transformation are discussed. It is shown that, under some conditions, the stress-assisted transformation structures may be unstable on the removal of the stress. At test temperatures outside theM _f toA _f range, this can result in anelastic behavior. More complex behavior expected at temperaturesM _f < T < A_f, is discussed in some detail. It is shown that both the anelastic behavior and the “shape memory” can be accounted for by the effects of applied stress. It is also shown that the mechanical properties in this temperature range can vary markedly with the prior heat treatment, even at temperatures not normally considered of significance. Though based on observations made on TiNi, the phenomena discussed are inherent in materials undergoing a martensitic transformation over a narrow range of temperatures.     

Molecular-dynamic simulations of martensitic transformation of cobalt

With a potential-energy function of Co described by the embedded-atom method (EAM), molecular-dynamics (MD) simulations were performed for a series of initial fcc configurations with different types of dislocations or preset hcp embryos. The gliding process of a Shockley dislocation on a closepacked plane has been observed, which starts from the origin of the dislocation and proceeds at a high speed of 280 m/s toward a certain direction. An atom which has been swept by the dislocation line was detected to contribute a displacement close to the Burgers vector of a Shockley dislocation. It is in this way that a new stacking sequence is produced and an hcp lamella grows in the fcc structure. A similar gliding process has been observed in the case where an intrinsic stacking fault is preexisting in the fcc structure. The transformation is, again, toward forming a local hcp region. These results prove that a special dislocation in the fcc structure can act as an embryo of the hcp, as described in many dislocation mechanisms of the martensitic transformation. The fcc → hcp phase-tranformation process of Co has been further reproduced by a simulation initiated from an fcc/hcp two-phase configuration. It yields a single hcp crystal as the final transformed product.     

Homogeneity range and the martensitic transformation in TiNi

It is shown that the structure homogeneity range of the B2 TiNi compound is very restricted below 500°C. A new phase, containing 58 at. pct Ni and structurally related to Ni₃Ti, forms peritectoidally at 625° ± 20°C. The precipitation in nickel-rich TiNi is very slow, and a metastable, supersaturated B2 structure can be readily retained. At equilibrium the martensitic transformation takes place at ∼60°C, independent of the overall composition. Quenching from 600°C and above raises the transition temperature in titanium-rich, and lowers it markedly in nickel-rich compositions. The resulting martensite structure in the latter is also believed to change with increasing the prior anneal temperature. The observations reported previously on the transformation behavior must be considered as obtained on nonequilibrium, metastable materials, which may account for the discrepancies in the observed property variation. Close control of the processing variables is necessary to ensure reproducibility, particularly if nonequilibrium transformation behavior is desired.     

Solidification mechanism of martensitic stainless steel

Abstract

In an unidirectional solidification experiment, an 8 kg stainless steel ingot with the composition 0·25%C, 17%Cr, and 1%Mn was solidified under continuous casting conditions. The dwell time of primary cooling was varied, followed by secondary spray cooling. Metallographic investigation, heat transfer, and segregation were carried out to study the solidification mechanism. The partition ratio of the elements present in ferrite and in austenite (martensite) was determined. It was indicated that the solidification follows: L → L + δ → L + δ + γ → δ + γ + carbides. Under high cooling rates γ austenite solidifies as a leading phase. The beginning of spray cooling has the main effect in controlling the obtained microstructures. Carbide thickening is observed in the rapidly cooled zone between the ferrite and the martensitic matrix. Tempered martensite increases by lowering the cooling rate, which gives more time for carbide dissolution and for carbon to diffuse into the ferrite, eventually increasing the austenite (martensite) fraction in the final matrix at the expense of ferrite.     

An estimate of the nucleation time in the martensitic transformation

A method for the determination of nucleation times for martensitic transformation is described. The method utilizes a shock wave that, upon being reflected at a free surface, generates a tensile wave with a pulse duration that increases as it moves away from the surface. Once the duration of the reflected pulse is large enough for nucleation to occur, transformation can take place. The width of the martensite free layer adjoining the surface is measured and compared with wave predictions. A nucleation time can be obtained. The method requires that the temperature, pulse amplitude, and alloy composition be such that only the reflected tensile wave induce martensite transformation. For the experimental conditions used by Snell, Shyne, and Goldberg¹⁰ the nucleation time is found to be less than 55 nanoseconds. MARC A. MEYERS, formerly Assistant Professor, Department of Metallurgical Engineering, South Dakota School of Mines and Technology     

Stress,deformation, and martensitic transformation

Martensite formation in ferrous alloys is inextricably connected with local deformation. In order to assess the role of this plastic work in the transformation, the dynamic responses of Fe?Ni?C austenites to imposed changes in deformation variables are investigated. For both stable and metastable compositions direct measurements are made of the separate and combined influences of prestrain and applied stress, and of strain-rate and dynamic strain aging. In addition, new martensite-start temperatures are defined and examined conceptually. A differential form of an energy balance that describes our viewpoint of austenite transformation at stress is then used to interpret the results. Some consequences concerning the importance of plastic work, its structure sensitivity, and the pervasive influence of local stress states are deduced and discussed.     

Internal friction due to thermoelastic martensitic transformation

A physical model and phenomenological treatment are developed to derive an expression for internal friction associated with the thermoelastic martensitic transformation. Special attention is given to the effect of pseudoelastic hysteresis(h), which has never been previously addressed. The amplitude dependence of the internal friction is discussed in various ranges of stress amplitude with respect to the critical stresses (σ_cl, σ_c2) for stress-induced martensitic transformation. In the case where the applied stress amplitude is much higher than the critical stresses, the internal friction, due to stress-induced martensitic transformation, is found to be linearly dependent upon the reciprocal of the stress amplitude, in good agreement with the experimental results.     

A general mechanism of martensitic nucleation: Part I. General concepts and the FCC → HCP transformation

Consideration of the martensitic nucleation process as a sequence of steps which take the particle from maximum to minimum coherency leads to the hypothesis that the first step in martensitic nucleation is faulting on planes of closest packing. It is further postulated that the faulting displacements are derived from an existing defect, while matrix constraints cause all subsequent processes to occur in such a way as to leave the fault plane unrotated, thus accounting for the observed general orientation relations. Using basic concepts of classical nucleation theory, the stacking fault energy is shown to consist of both volume energy and surface energy contributions. When the volume energy contribution is negative, the fault energy decreases with increasing fault thickness such that the fault energy associated with the simultaneous dissociation of an appropriate group of dislocations (e.g. a finite tilt boundary segment) can be zero or negative. This condition leads to the spontaneous formation of a martensitic embryo. For the specific case of the fcc → hcp martensitic transformation in Fe-Cr-Ni alloys, the defect necessary to account for spontaneous embryo formation at the observedM _s temperatures may consist of four or five properly spaced lattice dislocations. Such defects are considered to be consistent with the known sparseness of initial martensitic nucleation sites. This paper is Part I of a three-part series based on a thesis submitted by G. B. Olson for the degree of Sc.D. in Metallurgy at the Massachusetts Institute of Technology in June 1974.     

Kinetics of strain-induced fcc→hcp martensitic transformation

The kinetics of the strain-induced γ (fcc)→ε (hcp) transformationi.e. the amount of phase transformationvs applied strain were determined by density measurements at various temperatures. The transformation curve has a sigmoidal shape and approaches saturation below 100 pct transformation. Assuming that ε-platelets form from stacking faults, the volume fraction can be expressed as an implicit function of strain. The saturation value is constant and can be evaluated from quantitative metallography. The approach to saturation is determined by only one temperature-dependent parameter related to the stacking fault energy. Good agreement with experimental results was obtained. The model was also applied to transformation kinetics after a prestrain inducing both slip and twinning. The prestrain stabilizes austenite with respect to the strain-induced transformation through a block-refining of austenite by the substructure. In addition the nucleation is enhanced through the introduction of stacking faults. This effect vanishes at high applied strains but causes the shape of the transformation curve to become parabolic. It is concluded that decreasing the size of the ε platelets provides a simple means for reducing the temperature dependence of the transformation kinetics.