Effect of heat treatment on the crystalline structure of martensite in iron-, nickel-, manganese- and silicon-doped CoW and CoMo alloys

X-ray investigation of the crystalline structure of martensite of CoWFe, CoMo(Fe, Ni, Mn, Si) single crystals is performed after quenching and preliminary aging in α and β phases. Continuous decomposition of α and β solid solutions at aging at temperatures 500–700°C is observed. In Fe- and Ni-doped alloys the multilayer martensile polytypes are formed. In CoMoFe and CoMoNi single crystals such polytypes are found at the satellite stages of decomposition.     

Lattice changes of iron-carbon martensite on aging at room temperature

The aging behavior of iron-carbon martensite (5.1 at. pct C = 5.3C/100Fe) at about 297 K was studied by analyzing, in particular, the changes in the {002} and {200} diffraction line profiles obtained by X-ray diffractometry. Martensitic specimens were prepared by gaseous carburizing of pure iron in a mixture of H₂ and CO, followed by quenching in brine and subsequently in liquid nitrogen. The aging process can be divided into two stages. First, a redistribution of carbon atoms in the martensite matrix occurs in four ways (aging time < about 50 hours): (1) segregation of carbon to lattice defects (about 0.21C/100Fe), (2) transfer of carbon from thea/b- type toc-type octahedral interstices in the bulk of the martensite (0.13C/100Fe), (3) transfer of carbon fromc-type toa/b- type interstices at locations of significant stress variation, and (4) formation of carbon enrichments (clusters) in the matrix in which the majority of the carbon atoms is involved. Second, coarsening of the clusters occurs (aging time > about 50 hours) leading to diffraction by the matrix independent of the clusters. Within the range of aging times employed (up to 2 years), the diffraction by the retained austenite did not change.     

Martensite lattice changes during tempering

An X-ray diffractometer study of martensite formed in an 18 wt pct nickel, 0.98 wt pct carbon austenite single crystal yields the shapes, positions, and integrated intensities of 200, 020, and 002 peaks. Martensite, which forms below ? 60 °C, was tempered at successively higher temperatures from ?45 to 450 °C. The results show that after subambient aging, during which C atoms in c-oriented octahedral sites have clustered, carbide precipitation starts and small regions (～30Å in the [001]) with negative tetragonality appear. Upon subsequent tempering these are augmented by larger regions which have small positive tetragonality. In this process the “c” lattice parameter changes markedly but the “a” and “b” lattice parameter increase very little. These results indicate the formation of carbon depleted martensite which is coherently strained by the carbide particles. At and just above 100 °C the 200, 020, and 002 peaks all become doublets as the martensite matrix discontinuously breaks free of coherency and becomes highly imperfect ferrite. This change also occurs during the so-called “first stage of tempering.” Further tempering decreases the defect content of this ferrite. The lattice of the martensite is extensively reoriented during tempering just above room temperature. These reorientations probably accommodate the lattice parameter changes described above and may be carried out by movement of twin boundaries.     

On the abnormally large tetragonality of martensite in FeNiC alloys

In order to elucidate the origin of abnormally large tetragonality (c/a) of martensite in the FeNiC alloy system, a large number of alloys of which the Ni content is varied for a constant carbon content have been examined, using low temperature X-ray diffraction, optical and electron microscopy. The results show that the tetragonality of martensite in this alloy system is strongly dependent on martensite morphology. Namely, it is very large for a plate martensite, while it is normal or not so large for a lenticular martensite. In more physical words, the martensite tetragonality is dependent on the mode of the lattice invariant deformation in martensitic transformation. If the lattice invariant deformation is twinning, the resulting c/a is large, while in the case of slip it is small.     

200, 020, and 002 X-ray peaks in tempered Fe-18 Ni-C martensites

Separate 200, 020, and 002 X-ray peaks were recorded for 0.0, 0.4, and 0.8 wt pct carbon (18 pct Ni) martensites after tempering between 25 and 500°C. The carbon bearing martensites studied here have been tempered initially enough to eliminate the “high tetragonality” 002 peak usually recorded for as-quenched martensite and the present results apply to tempered martensite only. The peak maximum is taken to determine the lattice parameter and the peak shape is recorded. At all carbon levels and after all tempering treatments, the “crd parameter is larger than or equal to the “a” or “b”. The relative enlargement is very small (0.08 pct) for the lowest carbon level and for any carbon level after severe tempering (500°C for 15 min). For the two higher carbon alloys tempered at temperatures below 400°C (for 15 min) the “c” parameter is significantly larger than the “a” and “b” and for the 0.4 wt pct C alloy the “b” is significantly smaller than the“a” whereas in the 0.8 pct C alloy the “b” is slightly larger than the “a”. Within experimental error the mean volume of the unit cell does not change during the tempering studied here and is nearly unaffected by the initial carbon content. This indicates that little (at most 0.1 wt pct) carbon is dissolved in tempered martensite. In the low carbon alloy the peaks are symmetric and sharpen symmetrically during tempering. In the higher carbon alloys the peaks are nearly symmetric and sharp after severe tempering. After less severe tempering the 002 peak is asymmetrically broadened toward lower9 values (higher lattice parameters) whereas the 200 and 020 peaks are asymmetrically broadened toward higher 0 values corresponding to lower lattice parameters. This collection of results is tentatively interpreted as being due to strains in martensite due to transformation induced substructure and precipitated carbides.     

Excess Solute Carbon and Retained Tetragonality in Tempered Fe-0.6C-1Mn Martensite and the Effect of Silicon Addition

The tetragonality and carbon distribution in tempered Fe-0.6C-1Mn martensite were investigated by X-ray diffraction and atom probe tomography to elucidate strain relaxation in the tetragonal lattice during tempering and its relationship with the solubility of excess carbon in martensite. Even though tetragonality (c/a) decreased with an increase in the tempering temperature, it persisted at low levels up to 400 °C. Si addition suppressed the decrease in tetragonality at 400 °C by inhibiting recovery in the dislocated matrix. Such persistence implies that dislocation migration is crucial for the complete release of tetragonal lattice strain at such a temperature, in addition to the decrease in the amount of solute carbon in martensite. A low level of tetragonality was observed for martensite containing carbon in the solid solution below the critical value of ~ 0.2 mass pct, at which a bcc structure was predicted. The amount of solute carbon after tempering was linearly correlated with tetragonality in the solute carbon content range of 0.07 to 0.6 mass pct, and the correlation coefficient was similar to those for as-quenched auto-tempered martensite and bainitic ferrite; these results indicate that the amount of excess carbon is simply determined by the amount of tetragonal lattice distortions remaining after carbide precipitation and recovery.

     

200, 020, and 002 X-ray peaks in tempered Fe-18 Ni-C martensites

Separate 200, 020, and 002 X-ray peaks were recorded for 0.0, 0.4, and 0.8 wt pct carbon (18 pct Ni) martensites after tempering between 25 and 500°C. The carbon bearing martensites studied here have been tempered initially enough to eliminate the “high tetragonality” 002 peak usually recorded for as-quenched martensite and the present results apply to tempered martensite only. The peak maximum is taken to determine the lattice parameter and the peak shape is recorded. At all carbon levels and after all tempering treatments, the “crd parameter is larger than or equal to the “a” or “b”. The relative enlargement is very small (0.08 pct) for the lowest carbon level and for any carbon level after severe tempering (500°C for 15 min). For the two higher carbon alloys tempered at temperatures below 400°C (for 15 min) the “c” parameter is significantly larger than the “a” and “b” and for the 0.4 wt pct C alloy the “b” is significantly smaller than the“a” whereas in the 0.8 pct C alloy the “b” is slightly larger than the “a”. Within experimental error the mean volume of the unit cell does not change during the tempering studied here and is nearly unaffected by the initial carbon content. This indicates that little (at most 0.1 wt pct) carbon is dissolved in tempered martensite. In the low carbon alloy the peaks are symmetric and sharpen symmetrically during tempering. In the higher carbon alloys the peaks are nearly symmetric and sharp after severe tempering. After less severe tempering the 002 peak is asymmetrically broadened toward lower9 values (higher lattice parameters) whereas the 200 and 020 peaks are asymmetrically broadened toward higher 0 values corresponding to lower lattice parameters. This collection of results is tentatively interpreted as being due to strains in martensite due to transformation induced substructure and precipitated carbides.     

Phenomena of carbon atom redistribution in martensite

Phenomena related to the redistribution of carbon atoms prior to carbide precipitation (tempering) are considered in this paper. 1) Order-disorder processes are described, and the influence of several factors is considered. It is shown that the lattice conjugation and coherence between different variants of martensite plates as well as between martensite plates and austenite leads to a disappearance of the tetragonal cubic phase transition. 2) A possible mechanism for the formation of “abnormal” martensite having a comparatively low axial ratio is considered to be (011)_M transformation twinning. 3) Diffuse maxima observed in electron diffraction patterns near reciprocal lattice points are discussed. The diffuse scattering is related to the short range order of carbon atoms and a spinodal decomposition of the Fe-C solid solution because of the C-C elastic strain interaction. 4) The influence of irradiation on the carbon atom distribution in martensite is considered, and it is shown that the observed decrease in martensite tetragonality after cooling below 0°C and its restoration upon subsequent heating may be explained by a phase transition involving the condensation of carbon atoms on irradiation-induced defects. 5) A short review on experimental results related to the redistribution of carbon atoms because of the presence of dislocations is given.     

The Aging and Tempering of Iron-Nickel-Carbon Martensites

The aging and tempering of freshly quenched (M_s > RT) and virgin (M_s < RT) martensites with lath and plate morphologies in Fe-Ni-C alloys have been studied to obtain kinetic and structural information. At subambient temperatures, the first change is attributed to isothermal conversion of a small amount of retained austenite or to slight relaxations in the martensite, but this is not a significant part of the martensite aging process. Aging above ?40 °C to about 70 °C is accompanied by the diffusion-controlled clustering of carbon atoms, resulting in an increase in electrical resistivity proportional to the carbon content but independent of the martensitic morphology. This regime is followed above 100 °C by the precipitation of ε-carbide (i.e., the conventional first stage of tempering), which may emerge directly from the carbon-rich clusters. At still higher temperatures, cementite forms separately (i.e., the conventional third stage of tempering) in competition with the ε-carbide. These two precipitation processes overlap, and their kinetics appear to be controlled by iron-atom diffusion away from the growing carbide particles along dislocation paths. No evidence was found in this investigation for a regime reflecting carbon migration to dislocations or other defects, but this possibility is not ruled out by the experimental methods employed.     

Characterization of the Carbon and Retained Austenite Distributions in Martensitic Medium Carbon, High Silicon Steel

The retained austenite content and carbon distribution in martensite were determined as a function of cooling rate and temper temperature in steel that contained 1.31 at. pct C, 3.2 at. pct Si, and 3.2 at. pct noniron metallic elements. Mössbauer spectroscopy, transmission electron microscopy (TEM), transmission synchrotron X-ray diffraction (XRD), and atom probe tomography were used for the microstructural analyses. The retained austenite content was an inverse, linear function of cooling rate between 25 and 560 K/s. The elevated Si content of 3.2 at. pct did not shift the start of austenite decomposition to higher tempering temperatures relative to SAE 4130 steel. The minimum tempering temperature for complete austenite decomposition was significantly higher (>650 °C) than for SAE 4130 steel (～300 °C). The tempering temperatures for the precipitation of transition carbides and cementite were significantly higher (>400 °C) than for carbon steels (100 °C to 200 °C and 200 °C to 350 °C), respectively. Approximately 90 pct of the carbon atoms were trapped in Cottrell atmospheres in the vicinity of the dislocation cores in dislocation tangles in the martensite matrix after cooling at 560 K/s and aging at 22 °C. The 3.2 at. pct Si content increased the upper temperature limit for stable carbon clusters to above 215 °C. Significant autotempering occurred during cooling at 25 K/s. The proportion of total carbon that segregated to the interlath austenite films decreased from 34 to 8 pct as the cooling rate increased from 25 to 560 K/s. Developing a model for the transfer of carbon from martensite to austenite during quenching should provide a means for calculating the retained austenite. The maximum carbon content in the austenite films was 6 to 7 at. pct, both in specimens cooled at 560 K/s and at 25 K/s. Approximately 6 to 7 at. pct carbon was sufficient to arrest the transformation of austenite to martensite. The chemical potential of carbon is the same in martensite that contains 0.5 to 1.0 at. pct carbon and in austenite that contains 6 to 7 at. pct carbon. There was no segregation of any substitutional elements.     

Atomic displacements due to C in Fe Ni C martensite

The ratio of the X-ray intensity in the 002 to that in the 200, 020 in martensite formed in fine-grained 18 pct Ni ～ 1 pct C austenite provides a measure of the strength of the carbon displacement field. Theoretical calculation of the spacial variation of displacement due to carbon in the c-oriented octahedral site coupled with the assumption that such carbon is dispersed randomly then allows computation of the displacement of the two nearest neighbors as 0.6Å in the c-direction, corresponding to a carbon-iron distance of 2.0Å. Aging the martensite for one hour at temperatures above ?40 °C, the temperature of the above measurements, results in a decrease in the 002 intensity with little or no change in the 200 or 020 intensity, and little change in the lattice parameters until +40 °C when precipitation starts. This behavior is expected if clustering of carbon occurs and is contrary to expectations if carbon changes site during aging. If the absence of aging at —40 °C is assumed, then aging corresponds to formation of clusters of two to four carbon atoms.     

Carbon Clustering in Low-Temperature Bainite

The bainitic ferrite phase formed at temperatures below 573 K (300 °C) in high-carbon high-silicon steels holds an amount of carbon well above that expected from the thermodynamic paraequilibrium with austenite. Diffraction experiments have shown that the ferrite lattice is sufficiently Zener-ordered to possess a tetragonal symmetry, which allows the structures to be supersaturated in carbon. It could be expected that carbon undergoes ordering beyond that indicated by the Zener-ordering temperature as in the early stages of tempering of Fe-based martensites. This study examines the formation of cluster arrangements of carbon within bainitic ferrite and their relationship to the tetragonal distortion.     

A study of the early stages of tempering of iron-carbon martensites by atom probe field ion microscopy

The redistribution of carbon atoms during the early stages of ageing and tempering of iron-carbon martensites has previously been studied only by indirect methods. The computer-controlled atom probe field ion microscope permits the direct, quantitative determination of carbon concentrations at the atomic level, and thus all the stages of the martensite decomposition process become amenable to direct study. Analyses of a low-carbon martensite, Fe-1.0 at. pct C, (Fe-0.21 wt pct C), water quenched and tempered for 10 min at 150 °C, showed a matrix carbon content of only 0.14 at. pct. Analysis of a 2 nm diam area centered on a lath boundary showed a local concentration of 2.01 at. pct C. There is some evidence that this carbon level is associated with the presence of a thin film of retained austenite at the boundary. In the case of a higher carbon martensite, Fe-0.64 at. pct Mn, 3.47 at. pct C, (Fe-0.65 wt pct Mn-0.78 wt pct C) water quenched and aged for approximately 24 h at room temperature, analysis of twinned regions showed a matrix carbon level of 2.7 at. pct and a concentration enrichment to 6.9 at. pct in a region 2 nm diam, centered on the coherent twin interface. Assuming the segregated carbon to be located in a single atomic layer at the twin interface, this result indicates that a carbon concentration of 24 at. pct exists locally at the boundary. These results appear to be the first direct demonstration of the segregation of carbon atoms to lattice defects in carbon martensites. Tempering of the higher carbon martensite for 1 h at 160 °C produced further segregation of carbon to the region of twin interfaces. The matrix carbon content fell to 1.5 at. pct and the average carbon content over a 2 nm diam region at the interface rose to 8.7 at. pct. The width of the carbon segregated regions also increased, which seems to imply that incipient carbide precipitation in the plane of the twin boundaries is occurring at this stage of the tempering process. Formerly with the Department of Metallurgy and Science of Materials, University of Oxford     

Morphology transition of deformation-induced lenticular martensite in FeNiC alloys

The morphology and habit planes of deformation-induced lenticular martensite were investigated by optical and transmission electron microscopy in Fe30Ni and Fe30Ni0.11C alloys. Transitions in morphology were observed with progressive deformation levels going from lenticular to butterfly and to compact martensite for the Fe30Ni alloy and lenticular to butterfly and to small butterfly martensite for the Fe30Ni0.11C alloy. The habit planes changed from {225}_f or {259}_f for the thermal lenticular martensite to {111}_f for the strain-induced martensite. The morphology and crystallography of the small butterfly martensites was also investigated. A change in the orientation relationships from K-S to N-W relations was also observed. These changes were attributed to the contribution of mobile dislocations which modified the shear mode from twinning to slip, and to a plastic accomodation of transformation strains.     

The tempering of martensite in an Fe-1.5 pct N alloy

The tempering of Fe 1.5 pct N martensite has been studied at temperatures up to 300°C using X-ray and electron microscope techniques. Stage 1 decomposition occurs below 270°C by the general precipitation, resembling spinodal morphology, of fine τa^" (Fe₁₆ N₂) lamellae on 001 habit planes in both matrix and twin crystals of the partially 112 twinned martensite plates. Yet, gaged by changes in the X-ray spectrum, the reaction is discontinuous, the tetragonal martensite doublets decaying in intensity without change in their Bragg positions. The anomaly and the failure to detect by electron microscopy regions exhibiting fractional stages of the fine scale α^’ → α + α^" reaction is attributed to its occurrence at different times in different martensite (or parts of martensite) plates. It is believed that transformation occurs in this manner because the nucleation of coherent α^" plates is controlled by the prevailing internal stress field. Thus the time exponent “n” for the reaction decays from a normal value between 1 and 0.67 to less than 0.3 as stress relief by recovery dominates the more protracted stages of the reaction. Above 200°C the more stable nitride γ^’ (Fe₄N) forms at an increasing rate as plates on 012 habit planes, accompanied by marked softening.     

Clusters in carbon martensite: Thermodynamics and kinetics

An original method of evaluation of the cluster population in carbon martensite has been developed. Using this method, it is shown that Kurdjumov’s model of carbon redistribution within the different octahedral site sublattices can quantitatively account for both observed normal and abnormal tetragonality in carbon martensite. It is also shown that the existence of the internal strains in martensite constitutes a necessary and sufficient condition for the energetic preference of tetrahedral over the cubic lattice. The presence of the residual tetragonal distortion in the quasi-cubic phase of k-martensite is associated with the presence of the mixed clusters formed of the atoms belonging to O _c sublattice as well as to remaining ones. By using a computer simulation of the dynamical behavior of carbon martensite approaching the thermodynamical equilibrium, it was found that the ultimate state of this system is strongly beyond the thermal equilibrium. Even after long-term aging, the free energy is far beyond the minimum value allowed for this system. The reason for such a behavior and the possible aging processes proceeding in this system are discussed at the molecular level. All of the ordering parameters are affected by the aging process. The evolution proceeds in the distinctly different time intervals for different parameters. At first, the long-range ordering parameter that determines the tetragonality of martensite evolves and reaches the stable value. In the next stage, the formation and then disintegration of two-particle clusters occurs. Disintegration of two-particle clusters coincides with the stage when three-particle cluster formation occurs at a high rate. Threeparticle clusters also disintegrate when some time elapses. The same pattern repeats regarding four-, five-, six-, seven-, and eight-particle clusters. To simplify the calculations, the nine-particle clusters are assumed to be the largest possible and are identified with an existence of superstructure. The formation of 100 pct of nine-particle clusters with no contribution of free atoms in an alloy ceases all aging processes. The evolution of these processes is illustrated graphically in the time range from 16 seconds to 1500 years, as estimated on the basis of experimental data.     

Aging and tempering behavior of iron-nickel-carbon and iron-carbon martensite

The aging at room temperature (RT) and the tempering behavior in the temperature range 293 to 973 K of ternary iron-nickel-carbon martensite (containing 14.4 at. pct Ni and 2.35 at. pct C) was investigated principally by using X-ray diffractometry to analyze changes in the crystalline structure and differential scanning calorimetry to determine heats of transformation and activation energies. These techniques also were used in the parallel study performed in this work of the tempering behavior of FeC martensite (containing about 4.4 at. pct C) in the temperature range 298 to 773 K. Analysis of the structural changes revealed that in both FeNiC and FeC the following processes occurred: (1) formation of carbon enrichments and development of a periodic arrangement of planar carbon-rich regions up to 423 K; (2) precipitation of ε/η transition carbide and transformation of a part of the austenite into ferrite under simultaneous enrichment with carbon of the remaining austenite (between 423 and 523 K); (3) decomposition of the retained austenite into ferrite and cementite between 523 and 723 K (only partly for FeNiC); (4) precipitation of cementite between 523 and 723 K; and (5) for FeNiC, reformation of austenite from ferrite and cementite above 773 K. A short comparative discussion concerning the first stage of martensite decomposition for FeC, FeNiC, FeN, and FeNiN martensites is given.     

The effect of the first and second stages of tempering on microcracking in martensite of an Fe-1.22 C alloy

The effect of tempering on microcracking in the plate martensite of an Fe-1.22 C alloy was investigated by isothermal heat treatments in the temperature range between 180 and 225°C. The second stage of tempering, followed by X-ray measurement of retained austenite, was confirmed to depend upon the diffusion of C in austenite, and the transformation product was found to consist of very closely spaced cementite lamellae in ferrite. Microcracking, despite the volume expansion that accompanies the transformation of the retained austenite, decreased only slightly with time during the second stage. The major decrease in microcracking occurred during the first stage, a result attributed to the plastic deformation that accompanies the dimensional changes caused by the reduction of the lattice tetragonality of the high carbon martensite in the first stage. Metallographic observations of surface relief and etching effects associated with martensite plates provided evidence of the first-stage plastic flow. The authors were formerly Research Assistant and Professor, respectively at Lehigh University, Bethlehem, PA.     

Mössbauer effect studies of tempered martensite

The⁵⁷Fe Mössbauer effect has been widely used to trace the tempering of Fe-C and Fe-N martensitic systems, and has provided much detailed information on the interstitial atom configurations formed during these processes. Carbon and nitrogen atoms are found predominantly to o°Cupy octahedral interstitial positions in the virgin martensite structure. During aging at room temperature, the interstitial atoms tend to agglomerate to regions of high carbon or nitrogen content. Carbon atoms are believed to cluster by jumping from tetrahedral interstitial sites and/or forming regions of ordered Fe₄C. Nitrogen atoms agglomerate to regions of ordered Fe₁₆N₂. On tempering above room temperature, the sequential precipitation of the ε or η, ? and θ carbides is observed in Fe-C systems. The α (Fe₁₆N₂) and γ′ (Fe₄N) nitrides are observed during the decomposition of Fe-N martensite. Mössbauer hyperfine spectra associated with each of these structures are described.     

Mössbauer Effect Studies of Tempered Martensite

The⁵⁷Fe Mössbauer effect has been widely used to trace the tempering of Fe-C and Fe-N martensitic systems, and has provided much detailed information on the interstitial atom configurations formed during these processes. Carbon and nitrogen atoms are found predominantly to occupy octahedral interstitial positions in the virgin martensite structure. During aging at room temperature, the interstitial atoms tend to agglomerate to regions of high carbon or nitrogen content. Carbon atoms are believed to cluster by jumping from tetrahedral interstitial sites and/or forming regions of ordered Fe₄C. Nitrogen atoms agglomerate to regions of ordered Fe₁₆N₂. On tempering above room temperature, the sequential precipitation of the ε or η, χ and θ carbides is observed in Fe-C systems. The α′’ (Fe₁₆N₂) and γ′ (Fe₄N) nitrides are observed during the decomposition of Fe-N martensite. Mössbauer hyperfine spectra associated with each of these structures are described.