Effect of thermal induction cycling on wear and corrosion performance of a X90CrMoV18 (440B) martensitic stainless steel

2.1 Materials

In this study, X90CrMoV18 (440B) martensitic stainless-steel specimens with dimensions of 35 × 15 × 4 mm³ were prepared, and heat treatment applications were performed. The chemical composition of the material is shown in Table 1. The specimens were kept in the furnace at 950 °C for 30 min, cooled in air, and normalized before induction heat treatment. The normalized specimens were subjected to heat treatment at induction austenitizing temperature at different cycle numbers, followed by oil quenching.

2.2 Induction heating process

In the study, induction heat treatment was performed using a copper coil designed to fit the samples properly, along with an induction device equipped with a medium-frequency unit (45 kHz, 12 kW). A schematic figure of the induction heat treatment is given in Figure 1. The samples were fixed in a glass tube in a position where the laser pyrometer could be measured during the heat treatment period. The samples were heated to the target austenitizing temperature between the coils and cooled by falling into the cooling bath placed under the heating unit after the defined waiting times. Isomax heat treatment (90 °C s⁻¹) oil was used for quenching at 25 °C (room temperature). The temperature of the cooling oil during quenching was kept at an average temperature of 40 °C in order not to affect the cooling performance of the oil.

Figure 1:

Schematic illustration of the induction heat treatment line and heat treatment application.

The details of each step of the heat treatment plan and the nomenclature notation are shown in Table 2, and a graphical representation of the thermal cycle applications is shown in Figure 2. The specified time in the heating process was applied in different cycles of heating. The austenitizing temperature was 1,110 °C, and the holding time was 2 min. The thermal cycles were at a temperature of 950 °C, and the holding time was 30 s. In induction heating, the target austenitizing temperatures were reached by increasing the heating rate by 20 °C s⁻¹. After heat treatment, the samples were cryogenically treated in a cryogenic treatment cabinet at −80 °C for 24 h with uniform cooling. Following cryogenic treatment, the specimens were tempered at 180 °C.

Figure 2:

Applied thermal cycles.

2.3 Microstructural examination and phase analysis

The heat-treated specimens were cut, polished, etched, and microstructure images from the cross sections of the specimens were examined by optical microscopy and SEM analysis. The preparation of the samples started at 60 grit and sanded up to 1,200 grit, and the surfaces were polished. The etching process was carried out with Kalling No. 1 etching solution (33 ml distilled water, 1.5 g CuCl₂, 33 ml HCl, and 33 ml ethanol) for microstructural examinations. Following this, Clemex software, which operated with a Nikon microscope, was used to analyze the matrix phase and carbide structures in the microstructure.

The chemical composition of the matrix and carbides was determined by SEM-EDS analysis, and the phase and carbide types were determined by XRD analysis. XRD analyses were performed on a Rigaku DMAX-SEM-EDS application that was made with a Zeiss brand Gemini 500 model device. 2,200/PC with a scanning speed of 2θ min⁻¹ using a Cu source anode. XRD graphs of heat-treated 440B steel between 1 thermal cycle and 3, 4 thermal cycles were plotted between 30° and 80°. The full width at half maximum (FWHM) values were obtained from the XRD graphs using graphical software. The average crystallite size and dislocation density values were calculated with the aid of peak angle and FWHM using Eqs. (1) and (2). Here, the Scherer constant (K) is assumed to be 0.9 [19], [20], and the wavelength (λ) is 1.5418 Å based on XRD raw data. Δ represents the density of dislocation.

with D: crystal size, K: Scherer constant, λ: X-ray wavelength, β: full width at half maximum (FWHM), and θ: peak angle

2.4 Hardness tests and wear test

The dry sliding wear test of the specimens was performed using a CSM tribometer at room temperature (21 ± 2 °C). According to ASTM G99 standard, the pin-on-disk method was used at a rotational speed of 10 cm  s⁻¹ and a load of 10 N, sliding distance of 500 m in a dry environment. The wear track radius was fixed at 3 mm. The profile of the wear scar (wear lines) was obtained using a Mitutoya SJ301 profilometer; the area of the wear scar was calculated, and the amount of wear was determined volumetrically by multiplying the wear scar by its circumference.

The macrohardness of the specimens was measured using the BMS 200 RB brand Rockwell hardness tester Rockwell C hardness testing method in accordance with ASTM E18-12. Rockwell C hardness tests were conducted to determine the overall hardness of the material throughout the entire sample. Moreover, the hardness results were correlated with the material’s wear loss.

Investigations on hardness depth and hardness distribution were carried out according to ASTM E384-11e1. Microhardness measurements were performed using a Future Tech FM-700 brand microhardness tester by applying a 300-gf load for 10 s of application time, and hardness measurements were taken at 400-μm intervals along the cross-sectional areas. The microhardness distribution of the samples and their hardenability characteristics from the surface to the core were evaluated. Moreover, the results of the microhardness tests were utilized for reliability analysis.

2.5 Corrosion tests

Corrosion tests were conducted by the Metrohm DropSens µStat400 Potentiostat-Galvanostat at room temperature in a 75 ml NaCl (3.5 wt.%) solution. Open circuit potential was monitored for 5 min prior to the corrosion test. A three-electrode corrosion cell was set with a saturated calomel reference electrode, graphite electrode, and work electrode (440B samples). The surface of samples was insulated except for the 25 mm² area that was used for corrosion measurement. Tafel curves were obtained, and all measurements were performed at a scan rate of 5 mV s⁻¹.

2.6 Reliability analysis of Vickers hardness measurements

The micro-Vickers hardness probe scans a very small area and thus plays an important role in reflecting the internal structure data of the relevant area [21]. The negative feature of this phenomenon is that a high number of micro-Vickers hardness measurements are required for reliability purposes. In this study, at least 10 hardness data were taken for each sample from the entire cross section of the samples at 400-μm intervals, and the results were analyzed. The Shapiro–Wilk normal distribution test was applied to all the Vickers hardness test results.

3.1 Microstructure

The microstructure formed in 440B martensitic stainless steel because of 1 cycle heat treatment application was observed as M₂₃C₆ carbides in an α′ martensitic matrix. The phases formed were examined in microstructure examinations with an optical microscope in Figure 3 and SEM images in Figure 4 and confirmed by XRD analysis (Figure 5). OM microstructure photographs are useful to show the distribution of carbide precipitates. Depending on the heat treatment conditions, retained austenite can be observed in the α′+M₂₃C₆ microstructure. Venske et al. confirmed that the microstructure of AISI 440B steel, austenitized at 1,100 °C for 20 min and subsequently tempered at 100 °C and 300 °C for 30 min, consists of α′, M₂₃C₆, and retained austenite [22].

Figure 3:

Optical microscope images of 440B steel specimens, a) 1 cycle, b) 3 cycles, and c) 4 cycles.

Figure 4:

SEM images of heat treated 440B steel specimens, a) 1 cycle, b) 3 cycles, and c) 4 cycles.

Figure 5:

XRD graphics of heat treated 440B steel (α′).

When the optical microscope and SEM images are examined, very fine martensite structure and densely spherical small carbides are observed. Among the features of induction heat treatment applications, there are publications that the grains are finer and smaller in size. Induction heat treatment shows very positive results in terms of smaller grain size and hardness improvement compared to conventional heat treatment methods. In the study conducted by Çivi et al. on quenching and tempering processes because of induction heat treatment applied to AISI 6150 steel, hardness increase and small grain microstructure formation were reported in comparisons between conventional heat treatment methods and induction heat treatment applications [23]. Induction heat treatment applications as well as thermal cycling applications are known to provide smaller austenite grain size formation and accordingly directly affect martensite morphology [14].

It was seen that the size and number of precipitates formed decreased with multiple-cycle heating applications compared to single-cycle heating applications (Figure 3); fine carbide precipitates were formed at the grain boundaries, and thus, the prior austenite grain boundaries were visible under the same etching conditions (Figure 4). In one cycle of heating (Figure 4-a), the precipitates surrounded the grain boundaries in the form of a network and grew bigger, while in the third cycle (Figure 4-b), these grain boundary precipitates were intermittently formed, and in the fourth cycle (Figure 4-c), they were dispersed at the grain boundaries in much smaller sizes. When the martensite structure was examined, lath martensite lamellar was quite prominent in one cycle. As is known, lath martensite shows a hierarchical subgrain structure with packets, blocks, sub-blocks, and slats with specific crystallography [14]. Twenty-four kinds of martensite laths can be formed from a single austenite grain with the Kurdjumov–Sachs (K–S) orientation relationship. Reversed austenite is formed when the martensite phase is reheated to the austenite phase in the second and more cycles after the martensite phase is formed. Reversed austenite contains low-angle package, block, and lath boundaries and has high dislocation density. According to Yeddu et al. recycling occurs by a shear mechanism, and therefore, dislocations are carried over from martensite to reversed austenite [24].

When the reversed austenite is quenched again, the crystallite size and grain size are expected to decrease. SEM images (Figure 4) clearly showed that martensite laths are reduced in the third and fourth heating cycles.

When the thermal cycle applications were examined, very fine martensite structures and white carbide structures were seen in the microstructure images of the samples with one heating cycle (Figure 4-a). Like microscope images, martensite phase and carbide structure profiles were also observed in the SEM analyses. In the samples where thermal heating cycles were applied, changes in austenite grain size and carbide structures were observed in the tendency of grain size reduction (4-b/4-c).

When XRD graphs are examined, the formation of martensite and carbide structure (M₂₃C₆) is observed as dominant peaks [25]. XRD analyses and related thermal cycling applications showed that the martensite phase and carbide phase were formed in 1-cycle samples and confirmed the microstructure images. Zhou et al. reported that induction heating has a significant effect on grain size [26]. Shajari et al. also suggested that a high heating rate is beneficial for grain refinement [27]. The crystallite size was found to be 20 nm on the XRD analysis results (Figure 5) calculated by Scherer’s formula. In our previous studies, we determined that the crystallite size of the CHT sample was 37.23 nm. Maximum full width at half width (FWHM) values and 2θ angle data were extracted from the XRD analyses. The data extracted can be used to predict the dislocation density, crystal size, and microtension values of the martensite phase in the material (Table 3).

Since induction heating provides rapid heating, there is not enough time for grain growth. In contrast to CHT (conventional heating methods), the grain structure is exposed to a magnetic field during induction heating, which can lead to an improvement of the magnetic order within the alloy due to electromagnetic forces [10]. Lu et al. reported that the presence of a magnetic field during induction heating can prevent static recrystallization behavior in nickel-based super alloys [26]. The alternating magnetic field can slow grain growth by inhibiting atomic diffusion. During induction heating, magnetic and thermal effects exhibit opposite effects on the grains. The presence of the magnetic field effectively promotes the refinement of the grain structure [26]. In a study conducted by Çivi et al. hardness improvement and small grain microstructure formation were reported in comparisons between conventional heat treatment methods and induction heat treatment applications [23]. Hidalgo et al. revealed that the packet size and block width decreased as well as the prior austenite grain size as the number of cycles increased in the rapid heat treatment, they applied for 4 cycles of 1,100 °C and subsequent cycles of 800 °C. Although almost the same crystal size was calculated as the 1st cycle sample in the 3rd cycle, a slight increase in grain size was observed in the 4th cycle (+3 nm). However, Hidalgo et al. suggested that the previous austenite, package, and block size distributions become more homogeneous with an increasing number of cycles. Yeddu et al. reported that martensite becomes more stable with increasing number of thermal cycles [24]. The standard deviation values calculated from the crystal size calculations from the results of repeated XRD analyses on different samples are shown in Table 3. When the result obtained here is examined, the homogeneity of the microstructure increased with the increase in the number of cycles. The value of 5.569 obtained in 1 cycle was replaced by 0.924 in 4 cycles. This confirms the idea of a more reliable and homogeneous microstructure. This homogeneity is considered to have significant effects on mechanical properties and corrosion behavior.

3.2 Correlation between hardness and wear

Induction through hardening processes can be achieved with low- and medium-frequency induction devices. As can be shown in Figure 6, a very high hardness value was obtained for the entire material. The hardness decreased at first with thermal cycling applications and then increased and came very close to the previous value. In the study performed by Honda et al. it was revealed from the results obtained that induction heating can be used as an effective method for hardening 13Cr–2Ni–2Mo stainless steel alloy [28].

Figure 6:

Relationship between hardness and wear, a) 1 cycle, b) 3 cycles, and c) 4 cycles.

The graph showing the relationship between volumetric material loss and hardness is given in Figure 6. In general, abrasive wear resistance of materials is associated with their hardness. Puli et al. reported that the hardness and wear performances after coating, hardening, and tempering applied to AISI 410 martensitic stainless-steel material were similar to those of AISI 410 martensitic stainless-steel material with the same hardness without treatment [29].

Materials that have high hardness are supposed to exhibit high abrasive wear resistance. However, in metallic materials containing high-hardness phases such as carbides, the hardness value is not the only determining parameter for wear resistance. Wear mechanisms are also influenced by carbide size, distribution, and proportion. According to Yılmaz’s research, composites with smaller carbide sizes demonstrated a lower resistance for the same carbide content at different carbide fraction sizes [30]. Decreased carbide sizes improve hardness and lower wear rate in materials made of carbides scattered throughout a soft matrix [31].

As can be seen from the figure, the highest hardness was obtained in the 1st cycle sample. A decrease in hardness with the number of cycles and then an increase in hardness were observed in the samples. These results were consistent with microstructure and XRD results. It is thought that increasing the number of cycles homogenizes the internal structure of the material, resulting in an increase in wear resistance. From the perspective of a homogeneous internal structure, this idea was supported by reliability measurements with repeated measurements.

In 1 time cycle application, when the wear scar profile was examined, particle-shaped wear deformation was observed. When 3 and 4 times thermally cycled materials were examined, more superficial abrasive effects were observed in the material. However, when the abrasion loss data were examined, it was seen that 3 times the thermal cycle application gave the same value compared to 1-time thermal cycle application, and there is some decrease in abrasion loss in the material with 4 times thermal cycle application. According to this information, it was thought that thermal cycle applications improved the wear loss due to the homogeneous and balanced heat treatment of the material in terms of wear (Figure 7).

Figure 7:

Worn surface of heat-treated samples, a) 1 cycle, b) 3 cycles, and c) 4 cycles.

3.3 Corrosion rate

The corrosion rate and microstructure images after the corrosion test are given in Figures 8 and 9. The lowest corrosion rate is in the 1 cycle sample, and the microstructure image shows small carbide dissolution. However, some intergranular corrosion was also observed in the 1 cycle applications compared to the other cycle applications. It can be said that this was related to the grain boundary continuous carbide network, as can be seen from the microstructure photographs (Figure 4). This was because the grain boundary carbide had become discontinuous in the 3 and 4 cycle specimens. M₂₃C₆ showed that the matrix adjacent to the carbide forms a chromium-free zone. It is well known that chromium-deprived regions, associated with Cr-rich carbide precipitation, affect the pitting and intergranular corrosion tendencies of stainless steels in chloride solutions [32]. The formation of Cr-poor regions around the grain boundaries with high energy levels and M₂₃C₆ carbides, which exhibit passive properties in terms of corrosion, caused the occurrence of intergranular corrosion (Figure 3).

Figure 8:

Corrosion rate in thermal cycle processes.

Figure 9:

Optical microscope image of samples after corrosion test, a) 1 cycle, b) 3 cycles, and c) 4 cycles.

Polarization curves obtained from corrosion tests are given in Figure 10. From the polarization curves, the corrosion current (I_corr) value, which is an indicated corrosion rate of a metal, is determined. A higher corrosion current (I_corr) value indicates accelerated corrosion [33], [34]. The 1 and 4 cycle samples were close to each other and showed the lowest I_corr values. As can be seen from Equation (1), I_corr is obtained from anode/cathode Tafel slopes, β _c , and polarization resistance R _p . These results are given in Table 4.

Figure 10:

The polarization curves of samples at different cycles.

Polarization resistance R _p is the resistance of the electrode surface under overpotential. Indicates how much the electrode potential deviates from the equilibrium potential. A high R _p value indicates that the electrode surface is more resistant to corrosion. Low values of β _a and β _c indicate that the corrosion rate is low. The corrosion test results showed that the application of multicycle reduced β _a and β _c values. However, since it also decreased the R _p value, the corrosion resistance decreased.

The pitting potential E_pit is the potential at which stable pits nucleate and grow [35]. In this study, the corrosion test results (corrosion rate and I_corr) showed that the application of multicycle heat treatment did not make a significant difference, but the E_pit value showed a significant difference showed the E_pit value is 60 % in 4 cycle sample compared to 1 cycle sample. Multicycle heat treatment provided a homogeneous microstructure, which improved the pitting resistance.

3.4 Results of the reliability analysis of Vickers hardness measurements

The obtained Vickers microhardness values according to cross section are given in Figure 11.

Figure 11:

Hardness distribution of the specimens in cross section at different cycles.

When the hardness distribution given in Figure 11 is examined, it was seen that although the entire section was hardened, the hardness change along the section of the one cycle sample was in a wide range. In contrast, with multicycle application, this distribution became stable. As a result of the normality analysis of these data, the obtained Vickers microhardness values were found in accordance with the normal distribution. It should also be noted that this situation is in accordance with previous studies [36]. In this context, the graphical representation of the obtained normal distribution function and the reliability values obtained by its cumulative statistical examination are given in Figuress 12 and 13.

Figure 12:

Graphical representation of the obtained Vickers microhardness distribution function.

Figure 13:

Graphical representation of the obtained cumulative distribution of Vickers microhardness values.

As seen in Figure 12, when the hardness distribution function on the samples subjected to 1 cycle heat treatment is examined, the function is in a wide range as a reflection of the values. However, when the heat treatment cycle is applied, the hardness distribution function narrows. This fact shows that the microstructure has a more stable structure based on the hardness results. Figure 13 and Table 5 show the Vickers hardness values with 90 % reliability as a reflection of these data.

These data reveal the necessity of evaluating the value ranges and the resulting reliable values instead of taking average values in experimental analyses [37]. Microstructural stability can directly affect mechanical property data, as seen in previous studies [21], [38], [39], [40]. As can be seen from Table 5, while there was only a small increase in the average micro-Vickers hardness values, when the 90 % reliable values were examined, there was a significant increase. As a result of the study, when the entire data set was evaluated together, it was seen that the most important effect of cyclic heat treatment applications was to increase the microstructural stability [41], [42].