The Microstructure and Mechanical Properties of Ferritic-Martensitic Steel EP-823 after High-Temperature Thermomechanical Treatment

Abstract

The effect of high-temperature thermomechanical treatment (HTMT) with plastic deformation by rolling in austenitic region on the microstructure and mechanical properties of 12% chromium ferritic-martensitic steel EP-823 is investigated. The features of the grain and defect microstructure of steel are studied by Scanning Electron Microscopy with Electron Back-Scatter Diffraction (SEM EBSD) and Transmission Electron Microscopy (TEM). It is shown that HTMT leads to the formation of pancake structure with grains extended in the rolling direction and flattened in the rolling plane. The average sizes of martensitic packets and ferrite grains are approximately 1.5–2 times smaller compared to the corresponding values after traditional heat treatment (THT, which consists of normalization and tempering). The maximum grain size in the section parallel to the rolling plane increases up to more than 80 µm. HTMT leads to the formation of new sub-boundaries and a higher dislocation density. The fraction of low-angle misorientation boundaries reaches up to ≈68%, which exceeds the corresponding value after HTMT (55%). HTMT does not practically affect the carbide subsystem of steel. The mechanical properties are investigated by tensile tests in the temperature range 20–700 °C. It is shown that the values of the yield strength in this temperature range after HTMT increase relative to the corresponding values after THT. As a result of HTMT, the elongation decreases. A significant decrease is observed in the area of dynamic strain aging (DSA). The mechanisms of plastic deformation and strengthening of ferritic-martensitic steel under the high-temperature thermomechanical treatments are also discussed.

1. Introduction

Ferritic-martensitic steels with a chromium content of 9–12% are considered as the candidate structural materials for nuclear and thermonuclear reactors of a new generation. This is due to the achieved complex of physical and mechanical properties, such as high values of heat resistance, low swelling (compared with austenitic steels), high resistance to radiation and helium embrittlement and other advantages [1,2,3]. The efficiency of a nuclear reactor and the possibility of increasing its operating temperature are determined, among other things, by the long-term strength properties of its structural materials. Both long-term and short-term strength properties of ferritic-martensitic steels are controlled by the elemental composition, microstructure features (dimensions of tempered martensite packets, volume fraction of ferrite, dislocation density), composition, dimensions and spatial distribution of carbide (carbonitride) phases [4,5,6,7,8,9,10,11]. Steels of this class are most thoroughly investigated in the structural state after normalization and tempering; this processing is referred to here as a traditional heat treatment (THT). After this treatment, the microstructure of the steels represents tempered martensite and ferritic grains with coarse carbides of M₂₃C₆ type (M–Fe, Cr, Mn) and fine submicron particles of carbonitrides of MX type (M–V, Nb, Ta; X–C, N) [4,5,6,7,8]. It was shown [4,5,6,7,12,13,14,15,16,17,18,19,20,21] that various thermomechanical treatments, including those using plastic deformation in the austenitic region, make it possible to effectively change the above-mentioned parameters of the microstructure and the carbide subsystem of ferritic-martensitic steels. These changes mainly consist of a smaller width of martensitic packets and lamella, an increased dislocation density, a decreased size of coarse phases of the M₂₃C₆ type and an increased dispersion of nanoscale carbides (carbonitrides) of the MX type [11,12,15,16,17,22,23,24,25,26,27,28].

The 12% chromium steel EP-823 is a representative of the ferritic-martensitic class of steels. This steel is currently being considered as one of the candidates for Generation IV nuclear reactors [29,30,31]. It is planned to be used as a material for fuel cladding shells for the Russian innovative fast neutron reactor core BREST OD-300 [32,33]. The influence of thermomechanical treatments on its microstructure and mechanical properties remains poorly understood. This paper presents the results of investigation of the effect of high-temperature thermomechanical treatment (HTMT) of this steel on the features of its grain and defect microstructure and the carbide subsystem, as well as short-term mechanical properties under tensile testing in a wide temperature range in comparison with THT.

2. Materials and Methods

The starting material was a Russian 12% chromium ferritic-martensitic steel EP-823 with a chemical composition shown in

Table 1. Elemental composition of EP-823 steel, Fe—base, wt. %.

Element	C	Cr	Mn	Mo	Nb	V	W	Ni	N	Si	Ce	Ti	B	Al
Comp.	0.14	11.56	0.58	0.74	0.40	0.34	0.68	0.68	0.03	1.09	0.10	0.01	0.006	0.02

. The THT was performed by normalization at 1100 °C for 1 h followed by tempering at 720 °C for 3 h. The HTMT consisted of heating at 1100 °C, holding for 1 h, hot plastic deformation by rolling to a value of ε ≈ 50% per 1 pass, and subsequent water quenching followed by tempering at 720 °C for 1 h. The initial sample thickness was 2 mm. Heating was carried out in a tubular electric furnace of the T-40/600 type (Plant “LETO”, Tula, Russia). The rolling mill was maintained at room temperature. The time before the transfer of the sample from the furnace to the rolling mill was about 5 s, so the actual deformation temperature was lower than that in the furnace.

The microstructural characterization was performed using a Tescan MIRA 3 LMU field emission scanning electron microscope (SEM) (TESCAN ORSAY HOLDING, Brno, Czech Republic) and an Oxford Instruments Nordlys F electron back-scatter diffraction (EBSD) detector (Oxford Instruments, High Wycombe, UK). Some samples were etched in a Vivela etchant to reveal the structure and dispersed particles. The SEM samples for EBSD were prepared by mechanical polishing followed by ion milling as a final polishing step. Ion polishing was performed using a Technoorg Linda SEMPrep 2 system (Technoorg Linda Co., Ltd., Budapest, Hungary). After HTMT, the EBSD images were taken in the section parallel to the rolling plane (perpendicular to normal direction (ND)) and in the longitudinal section (perpendicular to transverse direction (TD)). The EBSD data were obtained with the step sizes of 200 and 50 nm in the section parallel to the rolling plane and 175 nm in the longitudinal section. The microstructure was analyzed using the Oxford Instruments AZtec software (version 3.1, Oxford Instruments, High Wycombe, UK). It is assumed in this paper that the martensite packets are mainly separated by high-angle misorientation boundaries. Therefore, the equivalent circle diameter obtained from the EBSD data for high-angle misorientation boundaries corresponds to the average martensite packet size and the average ferrite grain size. The high angle boundaries are the boundaries with a misorientation angle of >15°.

TEM investigations were conducted using a JEOL JEM-2100 electron microscope (JEOL Ltd., Akishima, Tokyo, Japan) at an accelerating voltage of 200 kV. Thin foils were prepared from the sections parallel to the rolling plane by electropolishing in an electrolyte containing 450 mL of orthophosphoric acid and 50 g of chromic anhydride.

Mechanical tensile tests at 300–700 °C were performed in a vacuum of ≈2.7 × 10⁻³ Pa by the method of static tensile deformation in the rolling direction at a strain rate of ≈2 × 10⁻³ s⁻¹ on a NIKIMT 1246R-2/2300 high-temperature vacuum testing machine (Experimental Plant “NIKIMP” (Research and Design Institute of Testing Machines, Instruments and Measuring Instruments), Moscow, Russia). We used dog-bone samples with a gage length of 13 mm and a gage section of 2 × 1 mm². Samples were cut parallel to the rolling plane. The scheme of microstructure observation and cutting of samples for tension tests is presented in Supplementary file, Figure S1.

3. Results

3.1. SEM EBSD

The results of the EBSD studies and their statistical analysis are presented in Figure 1, Figure 2 and Figure 3. The microstructure of steel after THT (Figure 1) consists of martensitic packets and ferritic grains. The nearly equiaxed former austenite grains (about 10 to 80 microns in size) with martensite packets inside are observed on the orientation and phase maps. The neighboring martensite packets are basically separated by high-angle misorientation boundaries. The low-angle misorientation boundaries observed in Figure 1b appear to be mainly the boundaries of the martensitic lamellae blocks making up the packets. At the scanning step of 200 nm, not all lamellae are detected. Higher magnifications, smaller scanning steps (Figure 1d), and TEM methods (see below) allow detecting the fine structure of the martensitic packets. The predominance of blue and yellow colors on the Kernel Average Misorientation (KAM) map (Figure 1c) suggests that the misorientations on the subgrains are quite small (less than 5°).

The distribution of grains by misorientations (Figure 1e) clearly shows two peaks corresponding to the low-angle and twin (and similar) misorientations. The latter may be the boundaries between different variants of packet martensite [34]. The fraction of low-angle boundaries after THT is ≈55%, and that of twin boundaries is about 15%.

On the phase map, one can see sufficiently coarse (up to several microns) particles of MX type (M = V, Nb, Mo; X = C, N), the volume fraction of which is ≈0.3%. No nanoscale particles are resolved by the SEM EBSD method. Individual rounded grains of bcc-Fe are also detected, which appear to be δ-ferrite (Figure 1d).

After THT, the average grain size of the martensite packets and ferritic grains (equivalent circle diameter) is 3.1 µm. The minimum size is 0.7 μm; the maximum size is more than 30 μm. It should be noted that in the grain size distribution (Figure 1f), a significant fraction of grains has submicron and micron sizes, and single grains have sizes of tens of microns.

The microstructure of steel after HTMT in the section parallel to the rolling plane (Figure 2) has significant differences from that after THT. Large extended regions with similar orientations are seen on the orientation and phase maps (Figure 2a,b). Inside these regions, there are numerous fine grains (martensitic packets) with an orientation different from that in the coarse-grain regions. The density of low-angle misorientation boundaries increases significantly compared to that after THT (Figure 2b). The fraction of these boundaries (sub-boundaries) in the grain distribution by misorientation angles increases (relative to the THT case) and reaches up to 68% (Figure 2e). The phase map also shows MX-type particles (Figure 2b). Their volume fraction does not change compared to the THT case. The predominance of green and yellow colors in the KAM map (Figure 2c) indicates an increase in the misorientation angles at the low-angle boundaries compared to the THT case. The fraction of twin boundaries decreases to about 10%.

Apparently, the sub-boundaries in these structures are not only martensitic lamellae sub-boundaries but also dislocation substructures formed as a result of work hardening under hot deformation conditions.

After HTMT, in the section parallel to the rolling plane, the average grain size (martensite packets and ferrite grains) is 2.1 µm, which is about 1.5 times smaller than that in the case of THT. At the same time, the maximum grain size reaches 84.9 µm, which is almost three times larger than the corresponding sizes after THT. An analysis of the images (Figure 2a,d) and the grain size distribution (Figure 2f) suggests that the grains of micron and submicron sizes and somewhat larger grains of tens of microns are formed in this section.

Investigations of the steel microstructure after HTMT in the longitudinal section (Figure 3) showed the presence of grains elongated in the rolling direction and significantly flattened in the direction normal (ND) to the rolling plane. In ND, their sizes do not exceed 10 µm. Many micron and submicron martensitic packets, or ferrite grains, are observed inside coarse grain and at their boundaries, just as in the section parallel to the rolling plane. Some of them have nearly equiaxed shapes. (Figure 3b).

The analysis of statistical data indicates that the average grain size (equivalent circle diameter) in this section is 1.4 µm, the minimum size is 0.6 µm, and the maximum size is 19.6 µm. These values are smaller than the corresponding values in the section parallel to the rolling plane. The submicron grains dominate in the grain size distribution. In the longitudinal section, the fraction of low-angle boundaries is ≈52%, which is smaller than the corresponding values in the section parallel to the rolling plane.

Thus, based on the SEM EBSD studies in two sections, it is justified that the hot deformation under HTMT results in the formation of a pancake structure presented by the martensite packets and ferrite grains flattened in the rolling plane and elongated in the rolling direction. The average grain size of this structure (martensite packets and ferritic grains) is 1.5–2 times smaller than the corresponding value after THT. The maximum size (equivalent circle diameter) of individual grains in the section parallel to the rolling plane increases significantly (to more than 80 microns), while in the section perpendicular to the rolling plane, the maximum size does not exceed 20 microns. As a result of high-temperature deformation, besides the boundaries between the martensitic lamellae, many new low-angle boundaries are formed, and the misorientation at the sub-boundaries of the martensitic lamellae increases compared to the corresponding values after THT.

3.2. SEM of Etched Samples

The studies of the microstructure features of the etched steel samples after THT and HTMT showed the presence of a large number of coarse-dispersed carbide particles mainly of submicron sizes on the steel surface (Figure 4). The SE images (in secondary electrons) show the particles that are detectable by the characteristic contrast as a result of their selective etching. In BSE images (in back-scattered electrons), one can distinguish between “gray” particles precipitated predominantly at the grain and subgrain boundaries and “white” particles inside the grains. A local elemental analysis (see Supplementary file, Figure S2) showed that the Cr content in the particles of the first type exceeds the corresponding value for the matrix (Table 1). Based on these data and [12,15,20,22,23,24], the particles are identified as M₂₃C₆ carbides ((Fe, Cr)₂₃C₆). The particles of the second type are MX-type carbonitrides (M–Nb, Mo, X–C, N). The estimates of the volume fraction of micron and submicron particles based on SEM images showed that above processing has almost no effect on their volume fractions. After THT and HTMT, the volume fraction of M₂₃C₆ particles is ≈5.8 and 5.6%, and the volume fraction of MX type particles is ≈0.11 and 0.15%, respectively. For the latter particles, those values are somewhat lower than that determined by the SEM EBSD (≈0.3%). The M₂₃C₆ type particles are difficult to distinguish by this method because they have submicron sizes and, in many cases, precipitate at the boundaries of structural elements (grains and subgrains).

It should be noted that not all carbide and carbonitride particles could be detected by SEM of the etched samples. Nanosized (from 5 to 100–200 nm) particles are indistinguishable by this method. Moreover, the contrast on the etched particles can include wider areas, capturing the matrix. This can lead to an overestimation of the volume fractions of particles. Largely this applies to the M₂₃C₆ type particles. These particles locate predominantly along the grain and subgrain boundaries, and the contrast in the particle image may include the contrast at the etched boundaries and sub-boundaries.

3.3. TEM

TEM and STEM images of steel microstructure after THT and HTMT are shown in Figure 5 and Figure S3 in Supplementary file. The images contain tempered martensite lamellae, coarse-dispersed (50–150 nm in size) particles of M₂₃C₆ type and fine-dispersed particles of MX type (5–20 nm). The type of the particles was determined by the interplanar spaces in the selected area electron diffraction (SAED) patterns and by the local elemental analysis (Figure S3, Supplementary file). Inside the martensitic lamellae, a dislocation substructure is found (Figure 5a,b).

Nanoscale MX type particles are observed to pin the dislocations. These structural elements are characteristic of ferritic-martensitic steels [12,15,17,25]. The microstructure of EP-823 steel after THT and HTMT is qualitatively similar. The differences mainly concern the morphology of structural elements, their size, dislocation density increased after HTMT, as well as the location and morphology of coarse carbides of the M₂₃C₆ type. THT is characterized by the preferential location of M₂₃C₆ carbides at the boundaries of martensitic lamellae and ferrite grains (Figure 5a), [15,17,24]. These carbides have a predominantly elongated lenticular shape. There were no differences in the average sizes of these particles after both treatments. The M₂₃C₆ particles precipitate during tempering after hot deformation and can nucleate on dislocation sub-boundaries formed as a result of HTMT.

In the present work, no differences in the sizes and volume fractions of nanosized particles of the MX type after THT and HTMT were detected. A quantitative estimation of the particle volume fractions is difficult due to their small (5–20 nm) size and their possible observation in the thin foil regions only. In spite of this, the increase in the volume fraction of dispersed particles as a result of HTMT of ferritic-martensitic steels is noted in many works [11,12,15,16,17].

3.4. Mechanical Properties

The results of the studies of short-term mechanical properties of steel EP-823 after THT and HTMT under tensile testing in the temperature range from 20 to 700 °C are shown in Figure 6.

An analysis of these curves shows a qualitatively similar temperature dependence of the yield strength and elongation to failure for THT and HTMT, except for some special features. The temperature intervals of the variation of mechanical properties are similar for many ferritic-martensitic steels [4,16,24,35]. On the curves of dependence of the yield strength and elongation on the tensile test temperature for steel EP-823, several characteristic temperature intervals can be distinguished:

–

20 to 300 °C—the interval of gradual (compared with the interval of higher temperatures) decline of the yield strength for both treatments. The elongation at this decrease is insignificant for THT and falls down after HTMT relative to the corresponding value at 20 °C;

–

300 to 500 °C—the interval of a slight decrease in the yield strength with increasing temperature for THT and an increase in the yield strength at 400–450 °C (with a maximum at 450 °C) for HTMT. Elongation to failure in this temperature range is minimal (≈4–5%) after HTMT. After THT, the minimum values of elongation are observed at 350–450 °C, ≈7–7.5%, which exceeds the corresponding values after HTMT;

–

500 to 700 °C—the interval of a significant decrease of the yield strength and an increase of elongation to failure with increasing temperature for both treatments.

After HTMT, there is an increase in the yield strength in the temperature range from 20 to 650 °C relative to the corresponding values after the THT. At 700 °C, the yield strength values after both treatments are almost the same. The maximum difference in the yield strength is observed in the temperature range of 400–450 °C and is found to be 80–120 MPa. The elongation after HTMT is several percent lower than the corresponding values after THT, except for the values at 20 and 650 °C.

For ferritic-martensitic steels [4,6,35], there is a plateau on the yield strength plot versus the temperature in the range of 300–500 °C, which is characterized by its almost invariable value with increasing test temperature. The elongation to failure in this case decreases relative to the values at other tensile temperatures. This yield strength and elongation behavior is characteristic for dynamic strain aging (DSA) [4]. Figure 7 shows the tensile stress curves of steel in the temperature range of 400–450 °C after the two treatments. After THT, a serrated flow is observed only in the tensile curve at 450 °C; at 400 °C, the flow curve is without any features. HTMT has a significant effect on the behavior of EP-823 steel in the DSA region, which shows itself in an increase in the yield strength by 80–100 MPa at 400–450 °C. The tensile curves at these temperatures clearly show a serrated flow. The elongation of steel after HTMT at these temperatures has minimal values.

4. Discussion

The features of microstructure formation under HTMT of the EP-823 steel are typical for ferritic-martensitic steels under this type of treatment [4,6,15,16,17]. Notably, the microstructure of these types of steel was mainly investigated in one section. In the present work, the features of structure and substructure in two sections-parallel to the rolling plane and in the longitudinal section are demonstrated. It is shown that the tempered martensite packets and ferritic grains are pancake-shaped, elongated in the rolling direction and flattened in the rolling plane. The average size of the martensite packets and ferritic grains after HTMT decreases by 1.5–2 factors (relative to the corresponding value after THT), but their maximum sizes in the section parallel to the rolling plane increase. As a result of HTMT, the misorientation angles on the low-angle boundaries of martensitic lamellae increase, new low-angle boundaries are formed, and the dislocation density increases (as can be seen in the KAM maps, Figure 2c compared to Figure 1c).

The increase in the volume fraction of dispersed particles as a result of thermomechanical treatments has been noted in many works [6,15,16,17]. In the present work, no qualitative and quantitative differences in the features of the carbide subsystem of EP-823 steel after HTMT and THT were revealed. The volume fractions and sizes of both coarse-dispersed carbides M₂₃C₆ and MX-type carbonitrides, as well as fine-dispersed MX-type particles, do not practically depend on the treatment regime. This may be due to both the features of the elemental composition of steel (the content and type of carbide-forming elements) and the HTMT parameters.

The features of grain structure and substructure of EP-823 steel after HTMT control the increase in strength and the decrease in plastic properties practically in the whole temperature range of 20–700 °C compared to the corresponding values after THT. At the same time, the main mechanisms of steel hardening are the reduction of the average size of martensitic packets and ferritic grains and the substructural strengthening. Dispersed nanoscale particles pining the dislocation substructure also play an important role in maintaining the high strength properties of steel. An increase in the strength in a wide range of test temperatures as a result of HTMT has been noted for many ferritic-martensitic steels [4,14,15,16,17,18,19,23], while some authors [4] report an increase in ductility due to an increase in grain sizes.

It was noted [4] that the yield strength values in the DSA region reach a plateau and practically do not depend on the test temperature of ferritic-martensitic steels. In the EP-823 steel after HTMT, in the DSA temperature range (400–450 °C), there is an increase in the yield strength. The features of the structural-phase state of steel, apparently, cause a significant effect on the plastic flow of steel in the DSA region after HTMT. The high density of the substructure formed as a result of hot deformation may facilitate the nucleation of dispersed particles under tensile deformation in the DSA temperature interval. In this case, the temperature interval for DSA after HTMT is shifted relative to the corresponding temperatures after THT. The serrated flow after HTMT is observed at 400 °C and is absent after THT at this temperature. At tensile test temperatures of 300–500 °C, the precipitation of dispersed particles contributes to an increase in the yield strength and a decrease in the ductility because the increased density of particles is an obstacle to the dislocation motion. At higher tensile temperatures (600–700 °C), the gliding dislocations become capable of overcoming dispersed particles as a result of climbing. This leads to a decrease in the yield strength and an increase in the elongation at these temperatures. At 700 °C, the effect of the high substructure density created by HTMT is practically leveled.

5. Conclusions

The features of the grain structure and substructure of ferritic-martensitic steel EP-823 after HTMT with plastic deformation in the austenitic region compared with similar features after THT have been studied. It has been shown that HTMT leads to the formation of a pancake structure with the grains elongated in the rolling direction and flattened in the rolling plane and a 1.5–2 times decrease in the average size of martensite packets and ferrite grains (compared to the corresponding values after THT). The misorientations at the sub-boundaries of martensitic lamellae increase, new dislocation substructures are formed, and the dislocation density increases. The sizes and volume fractions of the coarse and fine dispersed particles of carbide and carbonitride do not practically depend on the processing regime. The microstructural features revealed in this study determine the increase in the yield strength and the elongation reduction practically in the whole temperature range (20–700 °C) of the tensile tests. The HTMT has a significant effect on steel behavior in the DSA region. In this temperature range, there is an increase in the yield strength, and the ductility decreases significantly while maintaining satisfactory values. The features of dispersed particles precipitation in the microstructure with the increasing dislocation density under tensile tests in the DSA temperature range determine this behavior.