3.1. Microstructural Characterization and XRD Analysis
FeB phase was not observed at 850 and 900 °C in
Figure 2a,b; however, XRD results proved the presence of FeB at all boriding processes. FeB and Fe
2B phases can be viewed obviously in
Figure 3. Although HMS was high alloying steel (14% Mn and 2.75% Si), the saw-tooth borided morphology unexpectedly occurred on its surface. The saw-tooth structure commonly occurs at the borided low carbon and low-alloy steel [
31,
32,
33]. In addition, Sinha reported that manganese flattened out the saw-tooth morphology in carbon steel and prevented the boron diffusion [
32]. The flat boride morphology appears on the surface of high alloying steel [
3,
9,
24], since the presence of alloying elements in grain boundaries blocks the diffusion of boron atoms from the surface to the inside of the high alloy steel [
34,
35].
Figure 3 exhibits that Mn is densely accumulated in the saw-tooth boron layer. Martini et al. explained that the saw-tooth boride layers were observed in steels because the iron borides prefer to grow in a crystallographic direction [001]. The iron borides contact neighbor crystals and grow inside the metal as an acicular (saw-tooth) shape. This layer grows in-depth, leading to a strong (002) preferred orientation [
36]. This analysis may also be suitable for manganese borides. Ma et al. reported that MnB adopted an orthorhombic Pnma (space group) structure, isotropic with FeB [
37]. Hence, the similarity of crystal structures of MnB and FeB can cause a saw-tooth morphology on borided HMS.
An evident boundary line that was not seen in many studies—particularly in studies exhibiting the saw-tooth morphology [
16,
31,
38]—was also observed in
Figure 2a,b and
Figure 3. The boundary line separated the borided layer and transition zone. During boriding C and Si atoms diffuse away from the boride layer to the matrix and form boro-cementite (Fe
3(B, C)) and iron-silico-borides as a separate layer under the Fe
2B layer [
32].
Several studies have found that the three regions are boride layer (BL), transition zone (TZ), and BM matrix in borided steel [
17,
29,
39]. BM matrix was zone unaffected by heat or boron. TZ formed below the boundary line and was distinguished by the hardness different from that of the BM.
Figure 3 also shows that there is a silicon-rich zone (SRZ) in the boride layer. Therefore, SRZ can be accepted as the fourth region of boride layer.
The most striking result to emerge from the data is shown in
Figure 3. Since iron borides and manganese borides prevented the diffusion of Si from the metal core towards the surface of HMS, Si concentrated strongly between the borided layer (BL) and transition zone (TZ). Taktak [
39] and Gök et al. [
17] determined Si diffusion with the EDX line. In this study, SRZ was confirmed by the EDX mapping. SRZ is seen obviously in
Figure 3 due to the high Si content of the HMS. As the borides formed, they push the Si atoms towards the steel core. Additionally, Si atoms in steel move towards the surface with increasing temperature. Si atoms cannot reach the surface because Taktak [
39] reported that Si could not soluble in iron borides, concentrating effectively at the interface of steel. Si atoms accumulate between BL and TZ and SRZ occurs. Since this formation was not given any name in the literature, it was termed “compact transfer of silicones (CTS)”.
The SEM micrograph of sample 904 and its EDX point analyses are shown in
Figure 4 and
Table 3, respectively. The significant data in
Table 3 revealed that Si and Al could not dissolve in iron borides and MnB. Al and Si ratios increased in SRZ due to their insolubility or solubility limits in the boron layer. The differences between BL and SRZ, where neither B nor Si was detected, respectively, are highlighted in
Table 3. Moreover, it was determined that aluminum presence in SRZ has increased compared to BL and TZ. Although Al and B form intermetallics, such as AlB2 and AlB12, they are not observed as they are unstable at room temperature [
40].
Figure 5 shows that the presence of Fe
2B (JCPDS 00-003-1053), FeB (JCPDS 00-002-0869), SiC (JCPDS 00-002-1042), and MnB (JCPDS 03-065-5149) phases are detected in XRD analysis. Although FeB was not seen in SEM micrographs (
Figure 2a,b), XRD results revealed its presence. XRD analysis revealed that the predominant phases were FeB and Fe
2B. The aforementioned MnB adopted an isotropic orthorhombic Pnma structure with FeB [
37]. This situation was discovered in
Figure 3. Since Mn formed borides with a lattice constant similar to that of iron borides, it tended to dissolve in Fe
2B and FeB phases. SiC can be formed during boriding due to the high level of Si in HMS.
3.2. Thicknesses of Boride Layers and Microhardness
Figure 6 shows that the thicknesses of boride layers range from 31.41 to 117.65 µm depending on treatment temperature and time. Minimum and maximum boride layer thicknesses were observed at samples 852 and 956, respectively. The thickness measurements indicated that the thickness of the boride layer increased with increasing process time and temperature. The comparison of boride layer thicknesses of different steels between this study and the other studies in the literature is shown in
Table 4. It shows that HMS has the second-highest borided layer thickness in high alloy steel. Although Sinha reported that manganese reduced the boride layer thickness in carbon steel [
32], the thickness measurements show that Mn facilitates boron diffusion in HMS.
Figure 7 exhibits the cross-sectional microhardness measurement profile of borided HMS samples. The hardness of unborided HMS was 532 HV
0.05. The highest hardness value was seen at 902 (1915 HV
0.05). The hardness of the boride layers is approximately three to four times higher than that of the BM matrix due to the presence of FeB, Fe
2B, and MnB phases which are significant for increasing the hardness of the surface. In the studies of Gök et al. [
17] and Kayali [
29], the high Cr content in steel caused chromium borides formation, which are harder than manganese borides, to reach hardness values of boride layer above 2000 HV. On the other hand, Duran et al., who boronized Inconel 718, was able to reach almost 1300 HV boron layer hardness due to the lower hardness of nickel borides than manganese borides [
14]. Compared with unborided HMS, the hardness of the BM matrix in borided steel decreases with the effect of time and high temperature. However, in the literature, there is a common ratio between hardness and boriding time and temperature increases [
17,
29,
31,
41], which was not observed in this study. The changing distribution of various phases (MnB, FeB, Fe
2B) in the boride layer can cause fluctuations in the hardness plot.
3.4. Rockwell-C Adhesion Properties
Rockwell-C indentation was applied to analyze the adhesion properties of boride layers on HMS. The test is simple, low cost and can be suitable to identify the failures of borided layers. VDI 3198 Rockwell-C indentation test was carried out to cause damage on the borided layer under 1471 N load. The damage to the boride layer was compared with the quality map in [
28]. The indentation craters of borided HMS formed after the adhesion test were evaluated by using SEM. The adhesion strength quality HF1–HF4 defines strong interfacial, whereas HF5 and HF6 define poor interfacial adhesion between the coating and the substrate [
28]. Applied load and the contact geometry cause shear stresses at the interface. Suitable coatings manage to resist these stresses and prevent extended circular delamination, however, extended delamination at the crater around specifies a poor interfacial adhesion [
28,
46,
47]. Three indentations were deployed for each specimen and intended surfaces were evaluated by SEM.
In conventional steels containing more than 0.8% Si, Si generates a very soft ferrite zone between the base material and the boron layer during boriding. At higher surface pressure, a quite brittle and hard boride layer is significantly damaged softer intermediate layer due to its penetration. Therefore, Si reduces the wear resistance of boride layers. The case is called the egg-shell effect [
32,
33]. However, Si affects to refine ε-martensite plates in HMS [
48]. Since ε- martensite was a harder phase than ferrite, no egg-shell effect was detected in borided HMS in this study. Related results are shown in
Figure 9.
Figure 9 shows that there are radial cracks at the circumference of indentation craters without any flaking or delamination on surfaces of borided HMS. The adhesion strength quality of surfaces of all samples used in this study match with HF1 and HF2. The VDI 3198 indentation test shows that there are strong interfacial bonds between the borided surface and HMS. Zong et al. [
41], Taktak and Tasgetiren [
46] observed HF5-HF6 adhesion quality at 1000 and 950 °C, respectively. Additionally, Zong et al. determined HF4-HF5 adhesion quality at 950 °C [
41]. Both studies attribute that FeB is more prone to cracking and spalling due to tensile residual stresses under mechanical strain than Fe
2B. However, in both studies, high chromium steels were used, and we think that hard, brittle chrome borides would have caused these adhesion damages, since in this study no delamination was observed on the surface as a result of the adhesion test, despite the boriding process at 950 °C for 6 h. The high content of MnB in boronized HMS, which has less hardness than chromium borides [
37], may have caused this result.
3.5. Roughness, COF and Reciprocating Dry Sliding Wear Tests
Figure 10a–c demonstrate COF plots recorded during the wear tests of all samples, that were carried out under 5, 10, and 15 N loads, respectively. Moreover,
Table 6 shows the mean value of COF results of all samples. The COFs of the BM were lower than those of the borided samples at all three test loadings. Although the COFs of sample 954 were lower than BM under the load of 5 and 10 N, the BM had lower COF than sample 954 under the load of 15 N. The COF can be affected by many parameters, such as the adhesion strength of the coating, hardness, roughness and distribution of phases occurred on the substrate surface [
35]. Svahn et al. found that rougher surfaces have higher COF [
49]. The low surface hardness of the substrate can cause low COF [
20,
50]. Costa-Aichholz et al. [
20] reported that in the low hardness unborided sample, when in contact with the counter material plastic deformation occurs; being these deformations a result of ease to shear surface that leads to a low COF, according to the borided sample. Peaks show very high COF for borided morphology due to the high roughness of borided samples (902, 854, 956 in
Figure 10a–c, respectively) in
Figure 10. This could be owing to sharp asperities causing abrasive behavior leading to infrequent high COF [
51]. This situation causes three-body wear between the sliding surfaces.
Table 6 shows that the COF of the BM is lower than that of all borided samples. The surface roughness may have affected the COF results. The effect of high roughness is to distribute the load over asperities contact leading to higher frictional resistance and so a higher value of the COF can be obtained.
The volumetric wear results from dry sliding wear tests are shown in
Figure 11. It exhibits that an increase in the applied load increases the volumetric wear losses of all samples. All borided samples performed lower volumetric wear loss than BM for each wear condition. The lowest volumetric wear losses were observed at sample 954 under the load of 5 and 10 N and sample 856 under the load of 15 N. Under 15 N load, sample 954 exhibited the second lowest volumetric wear loss. According to the literature, as boriding temperature and time increase, volumetric wear loss occurs [
17,
35]. This interpretation was related to the thickness and hardness of the boride layers obtained. In this study, a steady volumetric wear loss was not seen from tribological results of borided samples depending on the time and temperature increase. There are many parameters of material loss from the contacting surfaces under the loading such as work hardening tendency, applied load, type of relative movement, sliding speed, interfacial contact properties, and test environment, determining the contact stresses at the interface and material properties [
30]. Each parameter might have caused this unsteady volumetric wear loss due to the complex morphology formed on the surface. In addition, reciprocating wear tests can affect the results of wear volume loss because of presences of wear debris at the sliding interface. Therefore, asperities might lead to different wear losses on the surfaces of each borided sample.
The wear rates of samples are shown in
Figure 12. The lowest wear rate was obtained in 954, 954, 856 under the load of 5, 10 and 15 N, respectively. The highest wear rates were observed at BM for each load. The hardness of the boride layer is significant for the improvement of wear resistance [
52]. Due to the hardness of the FeB, Fe
2B, and MnB phases, borided HMS showed more resistance to wear. The wear rate of the borided steels is more than six times lower than BM under 15 N load. Both wear rate and wear volume loss test results show that the boriding process significantly increases the wear resistance of HMS.
EDS analyses were carried out on the worn area after the dry sliding wear test. The results of EDS line and mapping analyses aimed at determining the changes in the amount of existing elements on the surface of borided samples and BM after the dry sliding wear tests.
Figure 13a,b shows EDX line analyses of samples 854 and BM, respectively. The x axis indicates the wear track width, and the y axis defines the wear ball movement direction during the tribotest.
Figure 13a shows that the amount of Mn, S, Si, and K significantly decreased after the wear test. K arose due to KBF
4 in the boriding powder. B decreased after the wear test, but it was not as much as the elements mentioned above. There was no significant decrease in iron, however, a significant increase in oxygen along the line indicates that oxide compounds are formed there. It is seen that the regions where oxygen elements increase are in dark color in
Figure 13b. The quantification results indicate the decrease in iron and the increase in oxygen in these dark regions. Most likely, iron oxide occurred on the surface after the wear test.
Figure 13b shows that except Al, C, and Si, no significant decrease in other elements actualized.
The most striking result to emerge from the data is that changes in Mn and S are noticed when comparing
Figure 13a,b.
Figure 13a shows that Mn and S significantly decreased after the wear test. It was determined that the amount of both elements—especially S—in the scale on the left side in the elemental analysis, increased significantly in the boriding process. As a result of the wear test in
Figure 13b, a strong relationship between Mn and S does not appear in
Figure 13a. MnS has a very low hardness, like 142 Vickers [
53]. Therefore, Mn and S could decrease rapidly on the surface of borided HMS after the wear test. MnS formation may have adversely affected the wear volume results of the boronized layer because of its low hardness. However, it is not considered to be overly effective on wear resistance of borided HMS.
Figure 14 shows the cross-sectional view near the surface of HMS before the boriding process. MnS formation was not observed in
Figure 14. EDS mapping analysis confirms the absence of MnS formation on the surface of HMS in SEM image.
Figure 15 provides additional evidence concerning MnS formation on the surface of HMS during boriding. The structures circled in
Figure 15 are assumed to be MnS, probably formed by the effect of high temperature and low cooling kinetic that encourage its nucleation and growth during boriding.
Due to boriding powder, K was detected in the EDS mapping analysis of borided sample surface in
Figure 15a,b. In
Figure 15b, it is determined that oxides are formed like a shell. When oxide shells were broken due to the worn ball, K filled in these spaces (
Figure 15a,b). As mentioned above, it is most likely that K stuck to the WC ball and filled these gaps by the movement of the ball.
Figure 15c confirms the oxidation layer analysis performed in
Figure 13b. The oxide layers are seen in dark color. Penetration of carbon atoms on the edge of the oxide layer is shown in
Figure 15c.
The surface morphologies of the worn samples are given in
Figure 16. It is seen that the oxide layer (dark region) partially delaminates under repeated loads because of plastic deformations in
Figure 16a. Micro-cracks also occurred on the oxide layer. In the wear test, it is observed that the oxide layers formed on the surface disappeared with the increase of the applied load in
Figure 16b. The debris and grooves occurred on the surface of BM. Almost the entire surface of borided HMS had smooth wear tracks. Micro-cracks on the oxide layer and pits on the borided surface as a consequence of surface fatigue [
50] can be observed in
Figure 16c,d.
Figure 16d shows that particle impact-induced brittle fracture caused spalling off of the oxide layer. Oxidative type local delamination was observed in borided HMS. The oxide layer is spalled off in an area much larger than the impact crater, while the substrate is almost unaffected [
54].