Microstructure evolution during tempering of martensitic Fe–C–Cr alloys at 700 °C
- 6.8k Downloads
The microstructure evolution of two martensitic alloys Fe–0.15C–(1.0 and 4.0) Cr (wt%) was investigated, using X-ray diffraction, electron backscatter diffraction, electron channeling contrast imaging and transmission electron microscopy, after interrupted tempering at 700 °C. It was found that quenching of 1-mm-thick samples in brine was sufficient to keep most of the carbon in solid solution in the martensite constituent. The high dislocation density of the martensite decreased rapidly during the initial tempering but continued tempering beyond a few minutes did not further reduce the dislocation density significantly. The initial martensitic microstructure with both coarse and fine laths coarsened slowly during tempering for both alloys. However, a clear difference between the two alloys was distinguished by studying units separated by high-angle boundaries (HABs). In the low-Cr alloy, M3C precipitates formed and coarsened rapidly, thus they caused little hindrance for migration of HABs, i.e., coarsening of the HAB units. On the other hand, in the high-Cr alloy, M7C3 precipitates formed and coarsened slowly, thus they were more effective in pinning the HABs than M3C in the low-Cr alloy, i.e., coarsening of HAB units was minute in the high-Cr alloy.
As one of the main strengthening constituents in high-performance steels, martensite has been attracting significant attention. The martensitic microstructure with high defect density and units arranged in different hierarchic levels is complex and can be considered to be a low-ductility microstructure. Therefore, it must in general undergo a tempering treatment to improve the toughness before it is put to use [1, 2]. The tempering of martensite has been investigated for a long time, and the development of transition carbides and cementite as well as defect annihilation during tempering of martensitic carbon steels is well known [3, 4, 5, 6, 7, 8, 9, 10, 11, 12]. However, when it comes to highly alloyed steels that are tempered at high temperatures, the literature is scarce [13, 14, 15]. It is in particular difficult to find quantitative studies on the evolution of different microstructural parameters such as the evolution of the size distribution of laths  during tempering of alloyed steels.
Fe–C–Cr is the base system for important steel categories such as hot-work tool steels. Further developments of these steels rely on an improved understanding of the microstructural evolution during tempering and subsequent service which, in general, occurs at a somewhat lower temperature than the tempering treatment. Therefore, it is essential to gain an understanding of softening of the martensitic matrix by defect annihilation and microstructural coarsening, but it is also quite important to understand the precipitation of carbides during tempering. To retain their hot hardness, these steels often rely on a multistage precipitation sequence where the formation of different transition metal carbides at different points in time during service enables preservation of the hardness [1, 2, 3]. Recent improvements in theoretical and experimental analysis of steels, inspired by the trend toward integrated computational materials engineering (ICME) , enable more computationally driven development of tempered martensite. However, further improvements in the predictive capabilities of models predicting microstructure and correlating with properties rely on an improved understanding of the microstructure evolution during tempering. The modeling of precipitation must consider the different potency of nucleation sites and the diffusional mobilities in the bulk and through crystal defects, including their evolution during tempering. Furthermore, models correlating structure and properties must be able to predict both the potential precipitation hardening effect and how this is counteracted by softening of the matrix which occurs by defect annihilation and carbon migration.
The purpose of the present work was, therefore, to investigate the microstructure evolution of martensitic Fe–C–Cr alloys during tempering. Two model alloys both resulting in a lath martensitic microstructure , with varying additions of chromium—a slow diffusing substitutional element compared to carbon, were investigated. The evolution of the martensitic microstructure and precipitation during tempering was evaluated using X-ray diffraction (XRD), electron backscatter diffraction (EBSD), electron channeling contrast imaging (ECCI) and transmission electron microscopy (TEM). The effect of varying Cr composition on the evolution of the microstructure is discussed.
Chemical composition of the investigated alloys (in wt%), and the corresponding calculated Ms temperature using the model proposed in Ref. 
Samples for SEM analyses were prepared by mechanical grinding and polishing, finishing with an alumina slurry with a particle size of 0.05 μm. ECCI images were acquired in randomly selected grains for the evaluation of lath size. The thickness of the laths was evaluated with a linear intercept method using lines drawn perpendicular to the length direction of the laths. For each sample, more than 600 laths in several locations were counted to provide sufficient statistics. Orientation maps obtained through automated EBSD was also utilized to study the hierarchic structure of fresh and tempered martensite. Both EBSD and ECCI measurements were performed using a JEOL JSM-7800F field-emission SEM operated at 15 kV. The EBSD analysis was acquired using a Bruker e-flashHR detector, the scan step size was either 50 or 100 nm, and data acquisition and post-processing were performed using the Bruker Quantax software.
Results and discussion
Dislocation density evolution from X-ray diffraction measurements
The dislocation densities of fresh martensite, evaluated for the two alloys in the present work, are compared to data from the literature in Fig. 3b [10, 21, 25, 26, 27, 28, 29, 30]. It can be seen that the data from the present work are in good agreement with previous studies with similar carbon contents. It should, however, be mentioned that in addition to carbon, other alloying elements will also affect the dislocation density of martensite [9, 25, 28]. This is also seen in the present work where the high-Cr alloy has a slightly higher dislocation density than the low-Cr alloy in the as-quenched condition and is in agreement with our prior observations that Cr has a similar effect to C on the lath martensite microstructure . It may be related to the slight lowering of the Ms temperature (Table 1) when the Cr content increases [14, 17]. It should also be noted that there is a difference in the prior austenite grain size between the two alloys of 40 ± 6 μm for 4% Cr alloy and 65 ± 8 μm for 1% Cr alloy . Though it is generally believed that the prior austenite grain size affects the dislocation density of martensite, the difference in prior austenite grain size between the two alloys here is small and the effect of the difference due to alloying element content is believed to be larger [14, 17]. After long-term tempering, the dislocation density in the high-Cr alloy is slightly higher than that in the low-Cr alloy. It could be related to the difference in Cr content  although the difference in dislocation density between the two conditions is too small to draw any conclusion.
Carbon in solid solution (Css) as determined from the lattice parameters of as-quenched (aAs-q), room temperature aged (aAs-q*) and reference conditions (aRef.) of the martensitic steels
0.1387 ± 0.0235
0.09983 ± 0.0235
Microstructure evolution from ECCI and EBSD
In this section, the evolution of the hierarchical martensitic microstructure, consisting of fine laths, sub-blocks, blocks, packets and prior austenite grain boundaries, is presented. The fine laths are about 50–100 nm thick, whereas the EBSD scan step size used in this study is of the same order. It could be possible to optimize the resolution of the EBSD and use a smaller step size, but the amount of time needed to capture a statistically populated set of data for lath thickness would be costlier in comparison with other competitive techniques like ECCI and TEM. Compared to the quantification of the thickness using TEM, the main advantage of ECCI is its capability to capture a large amount of laths in a quicker way. It has proven to be a powerful tool when combined with EBSD to evaluate microstructural parameters like misorientation distribution . EBSD analysis was specifically used here to evaluate the “unit size” and to perform correlative work with ECCI so as to identify precipitation at certain boundaries. In the present work, the term “unit” represents regions separated by misorientation angles larger than 5° within a packet of martensite  in order to collectively specify features which are otherwise called by various names such as sub-block, block, domain and colony [14, 22, 23].
Evolution of lath thickness
Unit boundary evolution and influence of precipitates
The observations of boundary pinning give an important clue about the precipitation event. In our previous work , a significant amount of precipitates was found in the microstructure for the same alloys after 5-s tempering. This means that the precipitates form early during microstructure coarsening. Thus, the difference in unit size evolution between the alloys considered in this study should depend on the precipitates type, size distribution, number density and their effectiveness in pinning the unit boundaries. The slower coarsening of the units in alloy 0.15C–4.0Cr as compared to alloy 0.15C–1.0Cr is most likely a result of pinning by the precipitates. Hence, these observations give an indication that the M7C3 precipitates found in alloy 0.15C–4.0Cr are more effective in pinning than the M3C precipitates found in alloy 0.15C–1.0Cr. The recovery and dislocation annihilation are not expected to be affected significantly by the small precipitates at the early stages of aging as seen in .
The as-quenched martensitic microstructure usually contains a highly dense network of dislocations. During the initial stages of tempering these dislocations tend to recover and after prolonged tempering recrystallization of the microstructure occurs, a process very much dependent on the alloy composition [1, 2, 37]. From the TEM and SEM microstructural observations, only dislocation recovery was found in all the stages of tempering in the two alloys discussed here. This is evidenced by (1) the stable dislocation density (~ 5 × 10−14 m−2) after about 30 min of tempering; (2) the rather stable distribution of misorientation angle boundaries, i.e., no newly formed high-angle grain boundary due to recrystallization; and (3) no visible appearance of equiaxed grains. The dislocation density even after 5000-h tempering is two orders of magnitude larger compared to the dislocation density of annealed cold-rolled steels (1010 to 1011 m−2) . It is often believed that the recrystallization could occur in low carbon (< 0.2 wt%) martensitic steels at 600–700 °C tempering, but not when alloying elements such as Mn, Cr are present [1, 5, 6]. However, recrystallization in steels containing extra-low/low carbon/medium carbon, tempered at 700 °C, has been reported previously [1, 37, 38]. The reason for these contradicting results in the literature is still unknown. In the present work, it can be speculated that the pinning effect of carbides on boundaries is high, and a large driving force is thus needed for recrystallization to occur. However, the sharp drop in dislocation density even after short-term tempering at 700 °C leads to a reduction in the driving force available for recrystallization. This may explain why no recrystallization is observed here.
The dislocation density of fresh martensite and martensite tempered at 700 °C has been investigated for two alloys: Fe–0.15C–(1.0, 4.0) Cr (wt%). The high dislocation density of the fresh martensite decreases rapidly during the initial tempering of 5 s but tempering beyond a few minutes does not further reduce the dislocation density significantly.
The initial lath martensitic microstructures of both investigated alloys with coarse and fine laths coarsen slowly during tempering. However, a clear distinction between the low- and high-Cr alloys was found regarding the mobility of unit (high-angle) boundaries. In the low-Cr alloy, cementite forms and coarsens rapidly and thus causes little hindrance for coarsening of the units. On the other hand, in the high-Cr alloy, M7C3 forms more densely at unit boundaries and coarsens more slowly, leading to an effective pinning of the unit boundaries.
The work was performed within the VINN Excellence Center Hero-m, financed by VINNOVA, the Swedish Governmental Agency for Innovation Systems, Swedish industry and KTH Royal Institute of Technology. The authors are grateful to Fredrik Lindberg and Niklas Pettersson at Swerea KIMAB for experimental assistance, and to Lindsay Leach for valuable comments on the manuscript. Z. H. acknowledges the support from the China Scholarship Council (CSC), the National Natural Science Foundation of China (Nos. 51674080, 51404155, U1260204).
Compliance with ethical standards
Conflict of interest
The authors declare that the contents have no conflict of interest toward any individual or organization.
- 2.Bhadeshia HKDH, Honeycombe SR (2006) The tempering of martensite, steels, 3rd edn. Butterworth-Heinemann, Oxford, pp 183–208Google Scholar
- 3.Speich GR (1973) Tempered ferrous martensitic structures, metallography, structures and phase diagrams. In: Metals handbook, 8th edn, vol 8, American Society for Metals, Metals Park, pp 202–204Google Scholar
- 4.Jack KH (1951) Structural transformations in the tempering of high carbon martensitic steels. J Iron Steel Inst 169:26–36Google Scholar
- 6.Caron RN, Krauss G (1972) The tempering of Fe–C lath martensite. Metall Mater Trans A 3(9):2381–2389Google Scholar
- 8.Ohmura T, Tsuzaki K, Matsuoka S (2002) Evaluation of the matrix strength of Fe-0.4 wt% C tempered martensite using nanoindentation techniques. Philos Mag A 82(10):1903–1910Google Scholar
- 35.Hedström P, Stormvinter A, Borgenstam A, Gholinia A, Winiarski B, Withers PJ, Karlsson O, Hagström J (2012) In: 1st international conference on 3D materials science, TMS 2012, p 19Google Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made.