Elsevier

Acta Materialia

Volume 158, 1 October 2018, Pages 297-312
Acta Materialia

Full length article
Formation of eta carbide in ferrous martensite by room temperature aging

https://doi.org/10.1016/j.actamat.2018.07.071Get rights and content

Abstract

For several decades, the formation of carbon(C)-rich domains upon room temperature aging of supersaturated martensite has been a matter of debate. C-rich tweed-like patterns are observed to form after short aging times at room temperature and coarsen upon further aging. Here, we present a systematic atomic-scale investigation of carbide formation in Fe-15Ni-1C (wt.%) martensite after two to three years of isothermal room temperature aging by a combination of atom probe tomography and transmission electron microscopy. Owing to the sub-zero martensite start temperature of −25 °C, a fully austenitic microstructure is maintained at room temperature and the martensitic phase transformation is initiated during quenching in liquid nitrogen. In this way, any diffusion and redistribution of C in martensite is suppressed until heating up the specimen and holding it at room temperature. The microstructural changes that accompany the rearrangement of C atoms have been systematically investigated under controlled isothermal conditions. Our results show that after prolonged room temperature aging nanometer-sized, plate-shaped η-Fe2C carbides form with a macroscopic martensite habit plane close to {521}. The orientation relationship between the η-Fe2C carbides and the parent martensite grain (α′) follows [001]α’//[001]η, (1¯10) α’//(020)η. The observation of η-Fe2C–carbide formation at room temperature is particularly interesting, as transition carbides have so far only been reported to form above 100 °C. After three years of room temperature aging a depletion of Fe is observed in the η carbide while Ni remains distributed homogenously. This implies that the substitutional element Fe can diffuse several nanometers in martensite at room temperature within three years.

Graphical abstract

The direct correlation of TEM (a) and APT (b) enables to identify the nm-scale, carbon-rich features forming after two years of room temperature aging in martensitic Fe-15Ni-1C wt% as eta carbides.

Image 1
  1. Download : Download high-res image (356KB)
  2. Download : Download full-size image

Introduction

Carbon(C)-containing martensitic steels are inexpensive and widely used for high-strength applications, e.g. in the automotive, energy and aerospace industries [[1], [2], [3]]. Martensite forms by a diffusionless phase transformation during quenching or mechanical loading of austenite. For Fe-C alloys the C solubility in austenite is a factor of ∼100 times higher than in martensite. Practically all high strength steel alloys lie in the range between these solubility limits, so that martensite is initially always highly supersaturated with C in solid solution; at least in the moment immediately after the martensitic transformation. Thermal activation rapidly initiates C partitioning among the matrix and defects such as dislocations and internal interfaces, cluster formation and carbide precipitation [[4], [5], [6], [7], [8], [9], [10], [11], [12], [13], [14], [15], [16], [17], [18], [19], [20], [21]]. Regarding the distribution of C atoms, two major effects are responsible for the strength of martensite: solid solution strengthening in concert with defect decoration and precipitation hardening by carbides. Hence, understanding the mechanism of carbide precipitation in C-supersaturated martensite is of basic scientific importance for a wide range of steels. In particular, the rearrangement of C atoms in martensite within several years at room temperature (RT) is of crucial relevance, as most martensitic steels face these conditions within industrial applications.

In contrast to the common assumption of ‘frozen-in’ diffusion kinetics at RT, carbon atoms are mobile even at RT and diffuse [9]. The subject of C-redistribution in martensite at RT has already been addressed for several decades [9,10,[22], [23], [24], [25], [26], [27], [28], [29], [30], [31]]. From a fundamental understanding point of view, this effect would ideally be investigated in the binary Fe-C system. However, the lowest martensite start temperature that can be achieved in this system is about 130 °C for the Fe-1.4 wt% C composition [2] (alloying with more carbon leads to the formation of pearlite). Thus, autotempering inevitably occurs in Fe-C alloys which already leads to C clustering, segregation, and the precipitation of carbides, and subsequently the depletion of C in the matrix during further quenching to room temperature after martensitic transformation [8,24]. For this reason, it is not possible to investigate the redistribution of C in martensite under well-defined isothermal conditions at RT in the binary Fe-C system or in other low-alloyed steels. However, the addition of Ni as an austenite-stabilizing element is a widely-applied approach that allows reduction of the martensite start temperature to below RT [9]. C-supersaturated martensite is then created by quenching in liquid nitrogen, where any C diffusion remains kinetically suppressed until heating up to RT. Using transmission electron microscopy (TEM), Taylor et al. [9] observed nanoscale cross-hatched patterns on {203} habit planes in martensite, already after 5.5 h at RT in a Fe-15Ni-1C (wt.%) alloy, which they attributed to the redistribution of C. These patterns were observed to coarsen upon further aging from a spacing of 1.7 nm after 0.75 h to 5.0 nm after 720 h 1D atom probe measurements of this Fe-15Ni-1C (wt.%) alloy indicated that C-rich features containing ∼11 at.% C had formed after 1580 h at RT. Contradictory results were obtained by Mössbauer spectroscopy measurements of martensite having composition Fe-1.86 wt% C, which indicated the formation of Fe4C (20 at.% C) after 2160 h of RT aging [32]. More recent 3D atom probe measurements in the years 2009 [33] and 2013 [34] revisited this disagreement in carbon content and confirmed the earlier 1D atom probe results. Hence, the discrepancy between Mössbauer spectroscopy and atom probe tomography (APT) results remains unresolved.

Up to now, this topic was investigated primarily by either only TEM or by only APT. Here, we instead use both techniques and by a direct-correlative approach combine them on the exact same specimen to univocally correlate structural features measured by TEM with chemical features measured by APT. In the present work, we especially focus on the later stages of RT aging, between two to three years, in an Fe-15Ni-1C (wt.%) alloy.

Section snippets

Materials and methods

An 800 g ingot of Fe-15Ni-1C (wt.%) alloy was cast using electrolytically cleaned raw material in a 60 mm × 25 mm x 90 mm Cu crucible in a vacuum induction furnace. The ingot was hot rolled at 1150 °C from 60 to 6 mm. The rolled band was cut in 10 cm segments and the oxide layers were mechanically ground off. The segments were stacked and wrapped into Mo foil that served as a diffusion barrier, and further encapsulated in boxes of Thermax steel sheet that were sealed air-proof by welding. The

Results and discussion

The results and discussion section of this manuscript is organized in the following way. First an overview of the microstructure of lenticular martensite is given. Then we identify which features found by APT correspond to those observed by TEM. This is realized by applying both techniques to exactly the same specimen location [51]. Next, we identify the crystal structure of the features of interest by conducting SAD along different zone axes, followed by a habit plane analysis based on APT and

Conclusions

Carbon redistribution in the martensitic phase of an as-quenched Fe-15Ni-1C (wt.%) alloy has been investigated within a time frame of 2–3 years of RT aging. The main findings are as follows:

  • Nanometer-sized η-Fe2C-type carbides were observed by direct-correlative TEM/APT analysis.

  • According to APT analysis, these η-Fe2C carbides have a chemical composition of 26.8 ± 2.5 at.% C, 60.1 ± 3.8 at.% Fe and 13.1 ± 4.0 at.% Ni. However, due to size-related APT artifacts, the real C-content must be higher

Acknowledgements

The authors acknowledge E. Welsch for support with the TEM measurements and T. Meiners for the STEM-EDS measurements. The authors also thank B. Gault for fruitful discussion and APT reconstruction support. The authors are grateful to U. Tezins and A. Sturm for their support with the FIB and APT facilities at MPIE. MH acknowledges the Federal Ministry of Education and Research [Bundesministerium für Bildung und Forschung (BMBF)] for funding through grant 03SF0535.

References (92)

  • R. Naraghi et al.

    Thermodynamics of stable and metastable structures in Fe–C system

    Calphad

    (2014)
  • W.K. Choo et al.

    Mössbauer measurements on the aging of iron-carbon martensite

    Acta Metall.

    (1973)
  • C. Zhu et al.

    Carbide characterization in low-temperature tempered steels

    Ultramicroscopy

    (2009)
  • M. Herbig et al.

    Combining structural and chemical information at the nanometer scale by correlative transmission electron microscopy and atom probe tomography

    Ultramicroscopy

    (2015)
  • K. Babinsky et al.

    A novel approach for site-specific atom probe specimen preparation by focused ion beam and transmission electron backscatter diffraction

    Ultramicroscopy

    (2014)
  • C.H. Liebscher et al.

    A hierarchical microstructure due to chemical ordering in the bcc lattice: early stages of formation in a ferritic Fe–Al–Cr–Ni–Ti alloy

    Acta Mater.

    (2015)
  • J. Takahashi et al.

    Quantitative analysis of carbon content in cementite in steel by atom probe tomography

    Ultramicroscopy

    (2011)
  • R.K. Marceau et al.

    Understanding the detection of carbon in austenitic high-Mn steel using atom probe tomography

    Ultramicroscopy

    (2013)
  • M.J. Yao et al.

    Combined atom probe tomography and density functional theory investigation of the Al off-stoichiometry of κ-carbides in an austenitic Fe–Mn–Al–C low density steel

    Acta Mater.

    (2016)
  • Y. Hirotsu et al.

    Crystal structure and morphology of the carbide precipitated from martensitic high carbon during the first stage pf tempering

    Acta Metall. Mater.

    (1972)
  • F. Li et al.

    X-ray diffraction and Mössbauer studies of the (Fe1-xNix)4N compounds (0≤x≤0.5)x

    J. Magn. Magn Mater.

    (1995)
  • A. Shibata et al.

    Substructures of lenticular martensites with different martensite start temperatures in ferrous alloys

    Acta Mater.

    (2009)
  • M. Herbig

    Spatially correlated electron microscopy and atom probe tomography: current possibilities and future perspectives

    Scripta Mater.

    (2018)
  • B. Hutchinson et al.

    Microstructures and hardness of as-quenched martensites (0.1-0.5%C)

    Acta Mater.

    (2011)
  • F.G. Caballero et al.

    Carbon supersaturation of ferrite in a nanocrystalline bainitic steel

    Acta Mater.

    (2010)
  • Y.J. Li et al.

    Evolution of strength and microstructure during annealing of heavily cold-drawn 6.3 GPa hypereutectoid pearlitic steel wire

    Acta Mater.

    (2012)
  • M.G.H. Wells

    An electron transmission study of the tempering of martensite in a Fe-Ni-C alloy

    Acta Metall. Mater.

    (1964)
  • N. Zhong et al.

    Orientation relationships between ferrite and cementite by edge-to-edge matching principle

    J. Mater. Sci. Technol.

    (2011)
  • J. Cermak et al.

    Carbon diffusion in carbon-supersaturated ferrite and austenite

    J. Alloy. Comp.

    (2014)
  • Y.J. Li et al.

    Atomic-scale mechanisms of deformation-induced cementite decomposition in pearlite

    Acta Mater.

    (2011)
  • M. Thuvander et al.

    Quantitative atom probe analysis of carbides

    Ultramicroscopy

    (2011)
  • Y.J. Li et al.

    Atomic scale investigation of redistribution of alloying elements in pearlitic steel wires upon cold-drawing and annealing

    Ultramicroscopy

    (2013)
  • G. Miyamoto et al.

    Quantitative measurement of carbon content in Fe–C binary alloys by atom probe tomography

    Scripta Mater.

    (2012)
  • M.K. Miller et al.

    Local magnification effects in the atom probe

    Surf. Sci.

    (1991)
  • X. Zhu et al.

    Improved resistance to hydrogen embrittlement in a high-strength steel by quenching–partitioning–tempering treatment

    Scripta Mater.

    (2015)
  • G. Miyamoto et al.

    Effect of partitioning of Mn and Si on the growth kinetics of cementite in tempered Fe–0.6 mass% C martensite

    Acta Mater.

    (2007)
  • C.M. Fang et al.

    Structure and stability of Fe2C phases from density-functional theory calculations

    Scripta Mater.

    (2010)
  • S. Allain et al.

    Toward a new interpretation of the mechanical behavior of as quenched low alloyed martensitic steels

    ISIJ Int.

    (2012)
  • H.K.D.H. Bhadeshia et al.

    Steels Microstructure and Propeties

    (2006)
  • P.C. Chen et al.

    Matensite lattice changes during tempering

    Metall. Trans. Metall. Mater. Trans. A

    (1980)
  • M.K. Miller et al.

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

    Metall. Mater. Trans.

    (1981)
  • L. Chang et al.

    Ageing of Fe-Ni-C martensite

    J. Phys. Colloq.

    (1984)
  • L.I. Chang et al.

    Ageing and tempering of ferrous martensites

    J. Phys. Colloq.

    (1986)
  • K.A. Taylor et al.

    Spinodal decomposition during aging of Fe-Ni-C martensites

    Metall. Mater. Trans.

    (1989)
  • K.A. Taylor et al.

    Carbide precipitation during stage 1 tempering of Fe-Ni-C martensites

    Metall. Mater. Trans.

    (1989)
  • A. Bottger et al.

    Tempering of iron-carbon-nitrogen martensites: (Re) distributin of interstitial atoms

    Mater. Sci. Forum

    (1999)
  • Cited by (51)

    • Application of DFT calculation in the study of iron-based catalyst for Fischer-Tropsch synthesis

      2023, Ranliao Huaxue Xuebao/Journal of Fuel Chemistry and Technology
    View all citing articles on Scopus
    View full text