*3.1. Microstructure*

SEM images (Figure 2) show the microstructures of the Vanadis 6 steel obtained by the conventional quenching and by different sub-zero treatments. The steel after conventional room temperature quenching is composed of martensitic matrix with the presence of a relatively large retained austenite amount, and of three carbide types (Figure 2a). The carbides are, according to the classification reported in [14]: eutectic carbides (ECs), secondary carbides (SCs), and small globular carbides (SGCs). The character of the matrix microstructure does not change significantly with the application of sub-zero treatment, at the magnification used for the SEM observations (Figure 2b–e). The formations of retained austenite become almost invisible after SZT, due to the significant reduction of *γ*<sup>R</sup> amount due to this kind of treatment. Alternatively, sub-zero-treated steel differs from that after CHT in terms of the number and population density of carbide particles (compare the SEM micrograph in Figure 2a with micrographs in Figure 2b–e). This concerns mainly the SGCs: the population density of these particles is much higher after SZT, while the SZT does not alter the population densities of ECs and SCs, as demonstrated recently [4,13,19].

**Figure 2.** Scanning electron microscopy (SEM) micrographs showing the microstructure of Vanadis 6 ledeburitic steel after conventional room temperature quenching (**a**) and after SZT at −75 ◦C (**b**), −140 ◦C (**c**), −196 ◦C (**d**), and −269 ◦C (**e**).

Figure 3 summarizes the population densities of SGCs, which were obtained by different combinations of SZT and tempering. It is shown that the tempering after CHT does not influence the population density of these particles, while the tempering after SZT makes their population density lower than that obtained by SZT without tempering. Despite this fact, the population density of SGCs is much higher than what can be obtained by CHT and tempering. It is also shown that the application of SZT at −140 ◦C acts in the most effective way in the enhancement of the SGCs population density, and the population densities obtained by other sub-zero treatments are much lower at equal tempering temperatures.

Figure 4 shows the TEM images of specimens after CHT and after CHT with subsequent SZT at −196 ◦C. CHT produces a needle-like martensitic microstructure with a relatively large amount of retained austenite. The width of martensitic needles is typically in the range of 100–200 nm (Figure 4a). The retained austenite formations are located at the interfaces of martensitic domains. Their width range is much wider than that of the martensite: some of the austenite formations have a width of a few tens of nanometres, while others are much greater, as illustrated in Figure 4b. The presence of *γ*<sup>R</sup> is confirmed

by diffraction patterns in Figure 4c. The martensite produced by SZT manifests considerable refinement compared to that developed by CHT. The typical width of martensitic domains ranges between 50 and 150 nm (Figure 4d). It is also shown here that the size of the martensitic domains manifests a great level of variability—there are some coarser domains visible in the micrograph, while in some sites, the domains are much smaller. A much higher amount of dislocations are generated within the martensite, as a result of plastic deformation that takes place during the isothermal hold at cryotemperatures. The retained austenite amount is significantly reduced by application of SZT. Additionally, the size of *γ*<sup>R</sup> formations is much smaller compared to the state after CHT (Figure 4e). The presence of retained austenite at the interfaces of martensitic domains is confirmed by diffraction patterns in Figure 4f.

Figure 5 shows X-ray diffraction spectra obtained on the conventionally quenched specimen and on specimens with SZT at −140 and −269 ◦C. All the spectra contain the peaks of the martensite, retained austenite, and major carbides. In the diffraction profile of the CHT specimen, there are the peaks of (111)γ, (200)γ, (220)γ, and (311)γ, all well-visible. On the other hand, the last two peaks disappear almost completely from the XRD spectra of SZT steel, suggesting a significant reduction of retained austenite amounts after this kind of treatment.

**Figure 3.** Population density of small globular carbides (SGCs) for differently sub-zero treated (SZT) specimens as a function of tempering temperature: (**a**) conventional heat treatment (CHT), (**b**) SZT −75 ◦C, (**c**) SZT −140 ◦C, (**d**) SZT −196 ◦C, (**e**) SZT −269 ◦C.

**Figure 4.** Transmission electron microscopy (TEM) micrographs of the CHT specimen (**a**–**c**) and the specimen that was subjected to SZT at −196 ◦C (**d**–**f**). (**a**) Bright-field image showing martensitic needle microstructure with retained austenite at needle interfaces, (**b**) corresponding dark-field image showing the retained austenite, (**c**) diffraction patterns of the retained austenite, (**d**) brightfield image showing martensitic microstructure with a small amount of retained austenite at the interfaces of martensitic domains, (**e**) corresponding dark-field image, (**f**) diffraction patterns of the retained austenite.

Variations in the retained austenite amounts for CHT steel, and for steel samples subjected to different SZTs, as a function of tempering temperature are summarised in Figure 6. It is shown that the application of SZT reduces the *γ*<sup>R</sup> amount to one tenth one fourth compared to the state after room-temperature quenching. Tempering at low temperatures does not influence the *γ*<sup>R</sup> amounts. Alternatively, tempering in the secondary hardening temperature range evokes almost complete retained austenite decomposition in the case of CHT steel. The application of SZT, in addition, accelerates the decomposition of *γ*R. Therefore, the amounts of metastable retained austenite lie below the detection limit of XRD after tempering at 450 ◦C and higher, in the case of the steel SZT in either liquid nitrogen or in liquid helium.

**Figure 5.** X-ray diffraction spectra of CHT steel and steel after SZT at −140 and −269 ◦C.

**Figure 6.** Retained austenite amounts for different SZT specimens, as a function of tempering temperature: (**a**) CHT, (**b**) SZT −75 ◦C, (**c**) SZT −140 ◦C, (**d**) SZT −196 ◦C, (**e**) SZT −269 ◦C.

A detailed description of the precipitation behaviour of differently sub-zero-treated Vanadis 6 steel exceeds the scope of the present paper. This topic has been extensively studied recently, and the main results can be summarised as follows: Application of sub-zero treatments evokes acceleration of precipitation of transient M3C carbide at low tempering temperatures [19,25]. In CHT steel, the M7C3 phase precipitates during tempering within the secondary hardening temperature range, while the precipitation of M7C3 is inhibited in the case of SZT steel, and only growth of M3C particles takes place.

#### *3.2. Corrosion Behaviour*

Figure 7 summarizes the potentiodynamic curves of CHT specimen and specimens after different SZTs, in un-tempered state. The specimens with SZT at −140 ◦C manifest the highest corrosion potential (*Ecorr*) (the most anodic), and the *Ecorr* decreases in the order: steel SZT at −269 ◦C, SZT at −75 ◦C, SZT at −196 ◦C, and CHT steel. This suggests that the SZT lowers the steel tendency towards oxidation, and the most convenient SZT temperature for lowering the oxidation tendency was found to be −140 ◦C. Additionally, the pitting potentials, *Epit*, are shifted to the more anodic values (see also Table 3), indicating slightly better resistance of SZT steel to the stable pitting corrosion. As follows from the values of corrosion current, *Icorr* (Table 3), the CHT specimens manifest the highest corrosion rate, and the corrosion rate (dissolution) decreases in the order: SZT at −75 ◦C, SZT at −269 ◦C, SZT at −196 ◦C, and SZT at −140 ◦C. These results indicate that the application of SZT improves the corrosion resistance of the Vanadis 6 steel in the prior-to-tempering material state.

**Figure 7.** Potentiodynamic polarisation curves of CHT and different SZT specimens in un-tempered state.

The potentiodynamic curves of CHT steel in un-tempered state as well as in the states after different tempering treatments are shown in Figure 8. The un-tempered specimen manifests the highest corrosion potential (*Ecorr*), and the *Ecorr* decreases with the increasing tempering temperature. This indicates that the tempering treatment deteriorates the resistance of conventional room-temperature-quenched Vanadis 6 steel against oxidation in 3.5% water solution of NaCl. The variations in the pitting potentials, *Epit* (see also Table 4), do not manifest a clear tendency with respect to the level of tempering temperature. The specimens treated at either low (170 ◦C) or high (530 ◦C) tempering temperatures manifest more anodic behaviour as compared with un-tempered steel. Conversely, the pitting potentials, *Epit*, of specimens tempered at intermediate temperatures indicate higher susceptibility to the stable pitting corrosion. The corrosion current, *Icorr* (see also Table 4), does not manifest significant changes after tempering at 170 ◦C. However, it raises rapidly after tempering at 330 ◦C and higher, suggesting that the corrosion rate of the material increases

dramatically. In other words, tempering treatment induces considerable worsening of the corrosion resistance of CHT Vanadis 6 steel.

**Table 3.** Corrosion current, *Icorr*, corrosion potential, *Ecorr*, and pitting potential, *Epit*, values acquired from corrosion tests in 3.5 mass % NaCl water solution, for CHT and different SZT specimens in the prior-to-tempering state.


**Figure 8.** Potentiodynamic polarisation curves of CHT Vanadis 6 steel in un-tempered state and in the states after different tempering treatments.



Potentiodynamic curves of the steel that was subjected to sub-zero treatment at −140 ◦C, un-tempered and tempered at different temperatures, are shown in Figure 9. The *Ecorr* decreases with tempering (in a similar way to CHT steel, see Figure 8). However, it is also obvious that the values of *Ecorr* (for the same tempering regimes) are shifted to higher potentials, suggesting that the SZT improves the resistance of the Vanadis 6 steel towards oxidation, not only in un-tempered state, but also after tempering treatments. The pitting potentials, *Epit* (see also Table 5), are practically the same for differently tempered

specimens, indicating almost no effect of tempering on the steel susceptibility to the pitting corrosion with SZT at −140 ◦C. The *Icorr* follows a very similar tendency to what was recorded for CHT steel. However, it is obvious (see also Table 5) that the values of *Icorr* are lower than what was recorded for the CHT steel tempered at the same temperatures. It should be noted that similar variations in corrosion characteristics were recorded for other SZTs. One can thus surmise that the SZT makes an overall amelioration of corrosion behaviour of the Vanadis 6 steel when tested in the 3.5 mass % NaCl water solution.

**Figure 9.** Potentiodynamic polarisation curves of Vanadis 6 steel subjected to SZT at −140 ◦C, in un-tempered state and in the states after different tempering treatments.

**Table 5.** Corrosion current, *Icorr*, corrosion potential, *Ecorr*, and pitting potential, *Epit*, values acquired from corrosion tests in 3.5 mass % NaCl water solution, for specimens after SZT at −140 ◦C and different tempering regimes.


Figure 10 shows the potentiodynamic polarisation curves for the CHT steel specimen and steel specimens after SZT at −75, −140, −196, and −269 ◦C, after tempering at 530 ◦C. As shown, the corrosion potential of SZT specimens is higher (more anodic) as compared to the material state after CHT. Further, it is evident that the SZTs at −75, −140, and −269 ◦C produced higher *Ecorr* than the treatment at −196 ◦C. The pitting potentials, *Epit* (see also Table 6), on the other hand, are shifted to the more cathodic values. This suggests that the application of SZTs provides the examined steel with higher susceptibility to the stable formation of pitting, in the state after tempering at 530 ◦C. The variations of corrosion current well-follow the changes in *Ecorr*: the samples after CHT have the highest dissolution rate (the highest *Icorr*), and the *Icorr* decreases in the order of SZT −196 ◦C, SZT −75 ◦C, SZT −140 ◦C, and SZT −269 ◦C. Here, a very important finding is that the tendency of the corrosion behaviour improvement, due to the SZT, is partly maintained after tempering at temperatures normally used for the secondary hardening.

**Figure 10.** Potentiodynamic polarisation curves of CHT Vanadis 6 steel, and Vanadis 6 subjected to different SZTs, after tempering at 530 ◦C.

**Table 6.** Corrosion current, *Icorr*, corrosion potential, *Ecorr*, and pitting potential, *Epit*, values acquired from corrosion tests in 3.5 mass % NaCl water solution, for CHT and different SZT specimens, after tempering at 530 ◦C.


Figure 11 provides an overview of the dependence of the corrosion rate on the tempering temperature for CHT specimens and for the specimens that were subjected to different SZTs. There are two main tendencies apparently shown. The first one is that the tempering treatment accelerates the corrosion rate (CR), and the higher the tempering temperature, the more accelerated the CR. The second general tendency is that SZTs retard the CR, and that the SZTs at −140 and −269 ◦C act most effectively in this way.

The surfaces after the potentiodynamic measurements of differently heat-treated specimens are presented in Figure 12. It is shown that the carbide/matrix boundaries are preferentially attacked by corrosion. This mainly concerns the boundaries between coarser eutectic and secondary carbides, as they differ from the matrix in terms of their chemistry. The M7C3 carbides (secondary carbides, SCs), for instance, contain 37.2 ± 0.8 mass % of Cr, 46.4 ± 0.5 mass % of Fe, and 12.7 ± 0.3 mass % of V (there are only metallic elements considered). The eutectic MC carbides are formed mainly by vanadium (73.6 ± 0.9 mass %), but they also contain limited amounts of chromium (9.4 ± 0.5 mass %) and iron (0.8 ± 0.1 mass %). On the other hand, the chromium content in the matrix is around 5.5 mass % only, and that of vanadium is correspondingly much lower (up to 1%). The steel also contains small globular carbides, after SZT in particular, see Figure 3. SEM micrographs in Figure 12c,d clearly delineate that the boundaries between SGCs and matrix are less intensively attacked by corrosion; alternatively, they remain well-embedded in the matrix and assist to lower the corrosion rate at the sites where they are present in sufficiently high amounts.

**Figure 11.** Corrosion rate in dependence on tempering temperature for CHT specimens and for specimens after application of different SZTs.

**Figure 12.** SEM micrographs showing the surfaces after the potentiodynamic polarisation measurements, CHT (**a**), CHT + tempering at 530 ◦C (**b**), SZT at −140 ◦C (**c**), SZT at −140 ◦C + tempering at 530 ◦C (**d**).

The SEM image in Figure 13 presents the corrosion-attacked specimen that was subjected to the SZT at −140 ◦C. The same features as in Figure 12 are visible. This mainly concerns the behaviour of different carbides. These carbides are highlighted in EDS maps of chromium (SCs) and vanadium (ECs). The other EDS maps (chlorine, oxygen, sodium) clearly demonstrate that the corrosion-exposed surface is covered by products of this process. However, it is also seen that the corrosion products' layer is not uniform. While the matrix is fully covered, the carbides are attacked by the corrosion environment to a much lower extent.

**Figure 13.** SEM micrograph showing the surface of the specimen that was subjected to the SZT at −140 ◦C, after the potentiodynamic polarisation measurements, and corresponding EDS maps of Cr, V, Na, Cl, and O.

Figure 14 shows the topography and work function mapping of the phases on the clean surface of the examined steel specimen that was sub-zero-treated at −140 ◦C. The white spots on the topography image, in Figure 14a, correspond to the carbide particles. The corresponding work function map in Figure 14b undoubtedly confirms that these sites have higher potential than the matrix. The difference is almost fully consistent through

austenite (matrix microstructure).

**Figure 14.** KPFM images of the examined steel sample that was subjected to the SZT at −140 ◦C. (**a**) Topography and (**b**) work function mapping. Image size 50 <sup>×</sup> <sup>50</sup> <sup>μ</sup>m2.

#### **4. Discussion**

The obtained results infer that the corrosion resistance of the Vanadis 6 steel is generally improved by the application of sub-zero treatments. Additionally, it was demonstrated that ameliorations in corrosion behaviour are maintained after tempering treatment, even though the tempering generally deteriorates the corrosion behaviour. The mentioned variations in corrosion behaviour are the topic of the following discussion.

the whole measured area, and is about 50–60 mV. In other words, these measurements indicate a more noble behaviour of carbides than the mixture of the martensite and retained

It has been summarised recently [26] that the application of SZTs evokes a significant reduction of retained austenite amount, refines the martensite, induces an acceleration of the precipitation rate of transient carbides, and produces an enhanced number and population density of small globular carbides.

### *4.1. Retained Austenite Amount and Characteristics*

The first factor that makes a clear difference in microstructures of SZTs Vanadis 6 steel and the CHT material is the retained austenite amount. Figures 4 and 6 provide clear evidence on the reduction of this phase in SZT steel, by 70–90%, depending on the temperature of SZT. The mentioned results appear counterintuitive with regard to the corrosion behaviour of the steel at the first glance because it has been experimentally proven that the austenite manifests a more noble behaviour than the ferrite (or martensite) [27–29]. The more noble behaviour of the austenite was attributed to a weaker internal stress state and to lower amounts of defects in the gamma phase compared to either ferrite or martensite [30].

However, it was demonstrated recently that sub-zero treatment evokes an introduction of high compressive stresses into the retained austenite of different steels [4,31], due to the combined effect of different thermal expansions of the austenite and the martensite, and to the volumetric effect of the martensitic transformation.

Beneficial effects of compressive residual stresses in the *γ*<sup>R</sup> on the corrosion resistance can be assumed based upon the obtained results of recent investigations. Peyre et al. [32], for instance, reported increased corrosion performance of the AISI 316L austenitic stainless steel, as a result of compressive stresses that were introduced into the surface by either laseror shot-peening. Takakuwa and Soyama [33] investigated the effect of various surface finish techniques (and thereby different stress level) on the corrosion resistance of the same steel grade in 5 mass % aqueous solution of H2SO4, and they arrived at very similar

findings. Moreover, they claimed that the principal explanation of improved corrosion behaviour may be based on the fact that the reduction of interatomic spacing due to the compressive stress on the surface facilitates the growth and maintenance of the passivation film. It is also interesting to note that the introduction of compressive stresses enhances the corrosion behaviour not only for ferrous alloys, but, for instance, also for aluminium alloys [34]. Therefore, one can conclude that the high state of compression in the retained austenite contributes to the overall improvement of corrosion resistance of SZT Vanadis 6 steel, even though the *γ*<sup>R</sup> amount was significantly reduced by this kind of treatment.
