The qualitative–quantitative approach to microstructural characterization of nanostructured bainitic steels using electron microscopy methods

The tested material was 9XC grade steel (GOST standard) after isothermal heat treatment, focused on obtaining low-temperature bainite with a high fraction of retained austenite. The high content of silicon (Si) retards carbide precipitation from austenite [3, 4, 7, 20, 21]. In previous research [22], it was determined that 9XC grade steel is susceptible to formation of a nanobainitic structure. The steel in the as-received state is characterized by a microstructure consisting of spheroidal cementite and a ferrite matrix – typical for softening annealing. The chemical composition is presented in Table 1. The time–temperature–transformation (TTT) chart of the tested steel was obtained using phase transformation simulation (JMatPro software), presented in Figure 1. To obtain a refined bainitic structure with varied morphology of the retained austenite (both film-like and blocky), along with a reasonable period of heat treatment, austempering at 300 °C was applied. The simulated time of bainite transformation at this temperature was about 4.5 h (Figure 1). Based on the simulation of phase transformations, the martensite start (M_s) temperature was about 132 °C. This result is comparable to the analytical calculations presented in previous research [22], in which M_s was in the range of 112–160 °C. The heat treatment parameters are shown in Figure 2. To ensure a fully bainitic structure, the samples were isothermally treated for 5 h. The austenitization temperature was 1,000 °C to ensure homogeneous austenite. The hardness after the designed heat treatment was 525±5 HV30.

Fig. 1

Fig. 2

Cuboidal samples of dimensions (10 mm × 10 mm × 20 mm) were used for PAGS and microstructural analysis. In the case of conventional picric acid solution etching, the samples were quenched (water cooled) and tempered. Due to the high Si content, which retards carbide precipitation [23, 24], the tempering process was necessary for an effective etching process. Similar to previously reported research [11], tempering was applied at 550 °C for 1 h. The preparation of sample intended for PAGS analysis consisted of grinding using sandpapers (with number ranging from 120 to 1,200) and polishing using diamond medium (with diamond size ranging from 6 μm to 1 μm). A saturated 5% picric acid solution in water at 60 °C was used for the etching process. PAGS was determined by measuring the diameters of the grains utilizing the ImageJ software. Grain diameters were defined as the arithmetic mean of the two diagonals of the grains.

The microstructural characterization included SEM, EBSD, and TEM analyses. The SEM observations were performed using a JEOL JSM7200F scanning electron microscope after etching with 3% nital. The topographic contrast mode (secondary electron detector), an accelerating voltage of 15 kV, and a working distance of 10 mm were applied. The EBSD analysis was performed using the JEOL JSM-7200F scanning electron microscope equipped with EBSD and detector (Hikari Plus, EDAX). The data obtained during the investigations were processed utilizing the EDAX/TSL software. Investigations were performed on 25.3 μm × 25.4 μm areas with a step size of 60 nm. Confidence index standardization and single-iteration grain dilation cleanup procedures were conducted. The grain size data were obtained considering a grain tolerance angle of 5°. The minimum grain size was determined at 2 pixels. The cleanup procedure did not exceed >5% of the pixels.

Microstructure observations were also carried out using a HITACHI H-800 transmission electron microscope. Preparation of samples intended for TEM investigations involved the mechanical prethinning of 0.5 mm plates to a thickness of 90–110 μm. Then, the stamped disks with ø=3 mm were electrochemically polished using the Struers TenuPol and ion polished using the GATAN DuoMill.

Conventional etching with a solution of picric acid reveals the grain boundaries of the prior austenite from martensitic structures, described in the European Standards of the International Organization for Standardization (EN-ISO) 643:2019 standard. A solution of picric acid etches the primary carbides (probably epsilon carbide), which are released at the grain boundary of the prior austenite. The main problem with this method for bainitic steels is the presence of a high concentration of Si, which effectively inhibits the precipitation of carbides, including epsilon carbide. Thus, often the grain boundary envelope is not visible without specific tempering processes directed toward precipitation processes. The results of conventional etching of 9XC grade steel are presented in Figure 3A. Moreover, to determine the PAGS by this method, a quenching process should be carried out for similar austenitization parameters; therefore, we should consider this measurement as indirect. On the other hand, EBSD analysis enables direct measurement because this analysis can be performed on samples after isothermal heat treatment aimed at obtaining a bainitic structure. Inside the grains of the prior austenite, the bainitic sheaves are characterized by various crystallographic orientations. It has been found that the angles, primarily in the range of 20°–40° in the case of martensitic structures, correspond to the boundaries of the prior-austenite grains [25]. This range of misorientation angles was also used for the bainitic structure (Figure 4). However, it should be highlighted that the misorientation angles in the mentioned range did not reveal a continuous envelope at the prioraustenite grain boundaries (Figure 4A,B,C). However, the obtained results enable manual PAGS reconstruction based on both 20°–40° misorientation angles and bainitic sheaf orientation resulting from the inverse pole figure map (Figure 4D).

Fig. 3

Fig. 4

However, the obtained measurement results for the PAGSs (Table 2) and their distributions (Figures 3B and 5) indicate some discrepancies. Both the average diameter value and the median were significantly lower for the EBSD measurements (Table 2). The standard deviation (SD) was at a comparable level, while in the case of the semi-interquartile range (½IQR), these values were also significantly different (±8.0 μm for the conventional method and ±4.0 μm for the EBSD technique). Due to the lower sensitivity to outliers, the median and IQR more accurately reflect the actual grain sizes. The distribution of grain diameters using the conventional method exhibits a slight right skewness (Figure 3B), while this cannot be found for the distribution obtained with the EBSD method, which is approximately normal (Figure 5).

Fig. 5

Due to the nanometric scale of the microstructural constituents of bainitic steel, the microstructural characterization should be performed using electron microscopy methods rather than with light microscopy. The first microstructure analysis includes SEM observations at relatively low magnifications to assess the homogeneity of the structure and the general morphology of bainite (Figure 6A). At higher magnifications, it is possible to identify film-like and blocky austenite morphologies and bainitic ferrite laths (Figure 6B). On the other hand, the unambiguous identification of bainitic sheaves should also include observations by TEM methods in the bright field (Figure 6C). Thus, it was found that after thermal heat treatment, the 9XC steel sample consists of bainitic sheaves, which consist of ferritic bainite and film-like austenite laths, as well as a significant blocky austenite fraction.

Fig. 6

A clear distinction between the film-like austenite and the cementite plates occurring in the upper bainite was determined by selected area electron diffraction (SAED) from the bainitic sheaves area (Figure 7). The crystallographic orientation between bainitic ferrite and film-like austenite was consistent with the Kurdjumov–Sachs (K–S) orientation relationship (OR) (Figure 7C): (1) [1¯01]γ[11¯1]α;(111)γ(011)α {[\bar 101]_\gamma }{[\overline {11} 1]_\alpha };{(111)_\gamma }{(011)_\alpha }

Fig. 7

The dark field image from the austenite reflex ( 11¯1 1\bar 11 ) also confirms the presence of retained austenite with a film-like morphology (Figure 7). After isothermal heat treatment, the tested steel – despite the high concentration of Si – also showed precipitation of cementite (Figure 8). The cementite precipitations are partially coherent with the ferritic and austenitic matrix (Figure 8A); however, they are not clearly visible in the bright field image (Figure 8B). Thus, the identification of the occurring phases was carried out using a qualitative approach. This approach is important in the context of the correct characterization of the microstructural constituents and is the basis for quantitative analysis.

Fig. 8

To determine the morphology of the presented structural components, quantitative EBSD analysis was performed based on the misorientation angles. It was identified that bainitic sheaves were characterized by different crystallographic orientations (Figure 9A). Besides, the presence of both film-like and blocky austenite was confirmed (Figure 9B). However, it should be emphasized that the EBSD technique does not detect the strongly refined austenite films that were identified by TEM. Quantitative analysis of the misorientation angles enables quantitative evaluation of both the crystallographic orientations and the related structural constituents. It can be noticed that in the tested area, the high-angle boundaries (HABs) in the range of 42°–47° were distinctly dominant (Figure 9D). In the case of nanobainitic steels, the misorientation angle of 46° corresponds to the misorientation angle between bainitic ferrite and retained austenite, which is related to the ideal angle of the Nishiyama–Wassermann (N–W) OR – 45.99° [26]. However, the ideal angle of the K–S OR is 42.85°. The well-known K–S and N–W ORs define the typical relations between ferrite and austenite; therefore, in the case of bainite steels, they indicate the presence of bainitic ferrite and retained austenite. This was also confirmed on the phase distribution maps with the K–S and N–W ideal angles (Figure 9C). It was also identified that the dominant orientation was the N–W OR (approximately 25%), rather than the K–S OR (approximately 16%). Another peak of misorientation angle was found in the range of 55°–60°. The misorientation angles between adjacent sheaves typically are 55° and 60° [25, 27]. It may also indicate the presence of martensite of lower bainite [27, 28, 29]; however, neither martensite nor lower bainite was confirmed in the TEM observations. Thus, these angles probably refer directly to the boundaries between the bainitic sheaves. A low fraction of low-angle boundaries (LABs) in the range of 2.5°–8° corresponds to lath-like subgrains inside bainitic ferrite laths [30].

Fig. 9

To determine the degree of refinement of the microstructural constituents, the shortest distance perpendicular to the longitudinal dimension of the bainitic ferrite laths and film-like austenite was measured (Figure 10). Because the measurements were performed on random areas, stereological corrections were applied. The correction proposed in previous research [31, 32] was applied, based on which the obtained measurements were corrected by the factor of π2 {\pi \over 2} . The measurement results are presented comparatively for both phases in Figures 11 and 12 and Table 3. The width distribution of the bainitic ferrite laths and the austenite films exhibits a slight right skewness. It can also be seen that the overall thickness of the bainitic ferrite was more than that of the film-like austenite (Figure 11). The width of the film-like austenite was in the nanometric range (x̄: 44.6 nm ± 20.3 nm and median: 40.0 nm ± 10.4 nm; Table 3), while the width of bainitic ferrite laths slightly exceeds 100 nm (x̄: 106.8 nm ± 40.7 nm and median: 102.0 nm ± 27.7 nm; Table 3). The scatter resulting from the measurements is clearly visible on the boxplots, which indicates that many more outliers occur for film-like austenite (Figure 12). In summary, the tested 9XC steel after isothermal heat treatment can be classified at the limit of the nanometric scale.

Fig. 10

Fig. 11

Fig. 12

Due to the significant role of the retained austenite morphology on the mechanical properties of nanobainitic steels, the quantification of both film-like and blocky morphologies is particularly important. Only based on microscopic observations is it possible to estimate their ratio. Two methods of quantifying both morphologies are proposed. One of them is based on the graphical analysis of SEM images in a topographic contrast (secondary electron mode), presented in Figure 13. The amount of film-like austenite was determined using threshold and area fraction measurements (Figure 13C). Due to the similar contrast of blocky austenite and bainitic ferrite, the blocky austenite regions were defined manually (Figure 13D). The second method was to consider separately the blocky austenite, determined based on the EBSD technique (phase distribution map), and the austenite with film-like morphology, calculated from the bright field images (TEM method). It should be emphasized that to correctly determine the blocky austenite in the EBSD technique, appropriate magnification should be used. Thus, the measurement result depends on the selected research areas, which should be representative and comparable. Graphical analysis of images in the bright field (TEM) was performed using the threshold and manual tracing of austenite with film-like morphology (Figure 14). Because very small areas are analyzed in this method and since the measurement may also be affected by the different orientations of the bainitic sheaves, this method requires a series of measurements. Table 4 compares the results obtained with the two described methods. The results are generally comparable between the methods but should be regarded as an estimate. In both methods, the fraction of film-like austenite was slightly higher than that of blocky austenite.

Fig. 13

Fig. 14