Wear resistance of an additively manufactured high-carbon martensitic stainless steel

Microstructural and mechanical properties

The microstructure of the EBM-processed HCMSS consists of a homogenous network of carbides surrounded by a matrix (Fig. 2a,b). The EDX analysis shows that the grey coloured and dark coloured carbides are Cr-rich and V-rich carbides, respectively (Table 1). As calculated via image analysis, the volume fraction of carbides is estimated to be ~ 22.5% (~ 18.2% Cr-rich carbides and ~ 4.3% V-rich carbides). The average grain sizes with standard deviation are 0.64 ± 0.2 μm and 1.84 ± 0.4 μm for V-rich and Cr-rich carbides, respectively (Fig. 2c,d). The V-rich carbides tend to be more circular with a shape factor (± standard deviation) of ~ 0.88 ± 0.03 since a shape factor with a value close to 1 corresponds to a circular carbide. In contrast, the Cr-rich carbides are not entirely circular, having a shape factor of ~ 0.56 ± 0.01, possibly due to the agglomeration. Martensite (α, BCC) and retained austenite (γ′, FCC) diffraction peaks are detected in the XRD pattern of the HCMSS, as shown in Fig. 2e. Further, the XRD diffractogram shows the presence of secondary carbides. The Cr-rich carbides are identified as M3C2 and M23C6 type carbides. Diffraction peaks of VC carbides have been reported at ≈ 43° and 63° according to the literature36,37,38, it is assumed that the VC peaks have been masked by the M23C6 peaks of Cr-rich carbides (Fig. 2e).

Microstructures of EBM-processed high carbon martensitic stainless steel (a) at low magnification and (b) at high magnification showing Cr-rich, V-rich carbides, and stainless-steel matrix (back-scattered electron mode). Histograms revealing grain size distribution of (c) Cr-rich and (d) V-rich carbides. XRD pattern showing the presence of martensite, retained austenite, and carbides within the microstructure (d).

Full size image

Full size table

The average micro-hardness is 625.7 + 7.5 HV5, showing a relatively high hardness compared to non-heat treated conventional processed martensitic SS (450 HV)1. The nanoindentation hardness of the V-rich carbides and Cr-rich carbides has been reported ranging between 12 and 32.5 GPa39 and 13–22 GPa40, respectively. Thus, the high hardness of EBM-processed HCMSS is attributed to the high carbon content that promoted carbide network formation. In conclusion, the EBM-processed HCMSS presents promising microstructural characteristics and hardness without any additional post-heat treatment.

Friction performance

The mean coefficient of friction (CoF) curves of the samples at 3 N and 10 N are presented in Fig. 3; the semi-transparent shading indicates the range of the minimum and maximum friction values. Each curve demonstrates running-in and steady-state stages. The running-in stage ends at 1.2 m with a CoF (± standard deviation) of 0.41 ± 0.24 at 3 N, while it ends at 3.7 m with a CoF of 0.71 ± 0.16 at 10 N, and then the steady-state stage occurs where the friction does not change that rapidly. The friction forces rapidly increase in the running-in stages at both 3 N and 10 N due to the small contact area and the initial plastic deformation of the asperities41, where higher friction forces and an extended sliding distance occur at 10 N possibly due to the higher surface damage compared to that of 3 N. The CoF in the steady-state stage is 0.78 ± 0.05 and 0.67 ± 0.01 for 3 N and 10 N, respectively. The CoF is almost stable at 10 N, while it gradually increases at 3 N. In the limited literature, the CoF of L-PBF-processed SS against ceramic counterbodies at low applied loads has been reported ranging between 0.5 and 0.728,20,42, agreeing well with the measured CoF values of this study. The decrease of CoF (around 14.1%) with increasing load in the steady-state could be attributed to the surface degradation that occurred at the interface between the worn surface and counterbody, which is further discussed through the surface analysis of the worn samples in the following sections.

The coefficient of friction against the sliding distance of EBM-processed HCMSS samples at 3 N and 10 N; steady-state stages are annotated for each curve.

Full size image

Wear performance

The specific wear rate of HCMSS (625.7 HV) was estimated as 6.56 ± 0.33 × 10–6 mm3/Nm and 9.66 ± 0.37 × 10–6 mm3/Nm at 3 N and 10 N, respectively (Fig. 4). Thus, the wear rate increased with increasing load, agreeing well with the existing studies on L-PBF-processed austenitic and PH SS17,43. The wear rate at 3 N is lower around one fifth of the value of an L-PBF-processed austenitic SS (k = 3.50 ± 0.3 × 10−5 mm3/Nm, 229 HV) under the same tribological conditions, as reported in a previous study8. Furthermore, the wear rate of HCMSS at 3 N is significantly lower than conventional processed austenitic SS; more specifically, it is lower around one sixth and one seventh of the value of a high isotropic pressing—(k = 4.20 ± 0.3 × 10−5 mm3/Nm, 176 HV) and a cast—(k = 4.70 ± 0.3 × 10−5 mm3/Nm, 156 HV) processed austenitic SS, respectively8. The improved wear-resistance of HCMSS compared to those studies in the literature is attributed to high carbon content and formed carbide network, resulting in a higher hardness than those AM-processed and conventionally processed austenitic SS. To further examine the wear rate of the HCMSS samples, similarly processed high carbon martensitic tool steel (HCMTS) samples (with a hardness of 790 HV) were tested under similar conditions (at 3 N and 10 N) for comparison; surface profile maps of HCMTS included in the supplementary material (Supplementary Fig. S2). The wear rate of HCMSS (k = 6.56 ± 0.34 × 10–6 mm3/Nm) was almost the same compared to the wear rate of HCMTS at 3 N (k = 6.65 ± 0.68 × 10–6 mm3/Nm), indicating an exceptional wear resistance. This performance was primarily attributed to the microstructural features of HCMSS (i.e. a high carbide content, the size, shape and distribution of carbide particles within the matrix as described in Sect. 3.1). As previously reported31,44, the carbide content influences the width and depth of the wear track as well as the micro-abrasive wear mechanisms. However, the carbide content was insufficient to protect the matrix at 10 N, resulting an increase in the wear rate. In the section that follows, worn surface morphologies and topographiesare used to explain the dominant wear and deformation mechanisms affecting the wear rate of the HCMSS. The wear rate of HCMSS (k = 9.66 ± 0.37 × 10–6 mm3/Nm) was higher compared to the the wear rate of the HCMTS (k = 5.45 ± 0.69 × 10–6 mm3/Nm) at 10 N. Comparatively, these wear rates are still quite high: chromium-based and Stellite coatings exhibit lower wear rates than the HCMSS under similar testing conditions45,46. Finally, the wear rate of the counterbody was negligible due to the high hardness of alumina (1500 HV), and there were signs of material transfer from the sample to the alumina balls.

Specific wear rates of EBM-processed high carbon martensitic stainless steel (HCMSS), EBM-processed high carbon martensitic tool steel (HCMTS) and L-PBF-, cast- and high isotropic pressing (HIP)-processed austenitic stainless steel (316LSS) at different applied loads. The scatter bars show the standard deviation of the measured values. The data for austenitic stainless steel was retrieved from8.

Full size image

Despite the fact that hardface coatings, such as chromium-based and Stellite coatings, can provide higher wear-resistance than AM-processed alloy systems, AM enables (1) microstructural refinement, particularly with alloys having a constituent with big differences in density, (2) the reduction of subtractive operations on a final part, and (3) the production of novel surface topologies, such as built-in hydrodynamic bearings. Moreover, AM offers geometric design flexibility. This study is particularly novel and significant as it is critical to reveal the wear behaviour of these newly developed metal alloys via EBM, where the current literature is very limited.

Wear mechanisms

The worn surface morphologies and topography of the worn samples at 3 N are shown in Fig. 5, where the dominant wear mechanism was abrasion followed by oxidation. First, the steel matrix was plastically deformed, and then the steel matrix was removed, causing grooves with a depth ranging between ~ 1 and 3 μm as shown in the surface profile map (Fig. 5a). The removed material remained at the interface of the tribosystem, forming a tribolayer consisting of small Fe-rich oxide islands around Cr-rich and V-rich carbides (Fig. 5b and Table 2) due to the friction heat from the continuous sliding, as also reported for L-PBF-processed austenitic SS15,17. Figure 5c indicates the intense oxidation that occurred in the centre of the wear track. Thus, either the material removal was accelerated due to the fracturing of the tribolayer (i.e., oxide layer) (Fig. 5f) or the material removal progressed in the weak regions within the microstructure, promoting the formation of the tribolayer. In both cases, the fracture of the tribolayer generated wear debris at the interface, which may be the reason for the increasing trend of the CoF in the steady-state at 3 N (Fig. 3). Further, there were signs of three-body abrasion caused by the oxide and loose wear particles on the wear track, eventually forming micro-scratches on the matrix (Fig. 5b,e)9,12,47.

Surface profile map (a) and micrographs of worn surface morphologies (b–f), the cross-section of wear track (d) in BSE mode for EBM-processed high carbon martensitic stainless steel at 3 N and the optical microscopy of worn surface of the alumina ball at 3 N (g).

Full size image

Full size table

Slip bands were formed on the steel matrix, indicating the plastic deformation due to wear (Fig. 5e). Similar results were also reported in a study on the wear behaviour of L-PBF-processed austenitic SS47. The re-orientation of the V-rich carbides also indicated the plastic deformation of the steel matrix during the sliding (Fig. 5e). The cross-section micrograph of the wear track revealed the existence of minor circular pits surrounded by micro-cracks (Fig. 5d), possibly due to the excessive plastic deformation of the near-surface. There was limited material transfer to the alumina ball, while the ball remained undamaged (Fig. 5g).

The wear width and depth of the samples increased with increasing load (at 10 N), as shown in the surface topography map (Fig. 6a). Abrasion and oxidation were still the dominant wear mechanisms, while the increased number of micro-scratches on the wear track suggests that three-body abrasion was also significant at 10 N (Fig. 6b). The EDX analysis showed the formation of Fe-rich oxide islands. The Al peaks in the spectrum confirmed that material transfer occurred from the counterbody onto the sample (Fig. 6c and Table 3) at 10 N, which was not observed at 3 N (Table 2). Three-body abrasion was caused by wear debris particles from the oxide islands and the counterbody, where detailed EDX analysis revealed material transfer from the counterbody (Supplementary Figure S3 and Table S1). The development of the oxide islands was associated with high depth pits, as also observed at 3 N (Fig. 5). Carbide cracking and fragmentation occurred mainly for Cr-rich carbides at 10 N (Fig. 6e,f). Furthermore, V-rich carbides detached and abraded the surrounding matrix and then further caused three-body abrasion. In the cross-section of the track (Fig. 6d), there was also a pit (highlighted with a red circle) with a similar size and shape with the size of V-rich carbides (please see carbide size and shape analysis in Sect. 3.1), showing that V-rich carbides were potentially dislodged from the matrix at 10 N. The circular shape of V-rich carbides promoted the pull-out effect, while the agglomerated Cr-rich carbides were susceptible to cracking (Fig. 6e,f). This fracturing behaviour suggested that the matrix's capacity to withstand plastic deformation was already exceeded, and the microstructure did not provide adequate toughness at 10 N.Vertical cracking in the subsurface (Fig. 6d) indicated the intensity of the plastic deformation that occurred during the sliding. Some material was transferred from the wear track onto the alumina ball with increasing load (Fig. 6g), which may be the underlying reason for the decreased CoF values at 10 N in the steady-state (Fig. 3).

Surface profile map (a) and micrographs of worn surface morphologies (b–f), the cross-section of wear track (d) in BSE mode for EBM-processed high carbon martensitic stainless steel at 10 N and the optical microscopy of the worn surface of the alumina ball at 10 N (g).

Full size image

Full size table

Mechanical properties of wear affected zone

During sliding wear, the surface is subjected to compressive and shear stresses induced by the counterbody, resulting in significant plastic deformation beneath the worn surface34,48,49. Consequently, strain-hardening within the subsurface can occur due to the plastic deformation, influencing wear and deformation mechanisms that govern the wear behaviour of materials. Thus, in the present study, cross-sectional hardness mapping (as detailed in Sect. 2.4) was performed to identify the development of a plastically deformed zone (PDZ) beneath the wear track as a function of load. Since, clear signs of plastic deformation below the wear track were observed (Figs. 5d, 6d), specifically at 10 N, as discussed in the previous sections.

In Fig. 7, the cross-sectional hardness maps of the wear track of EBM-processed HCMSS at 3 N and 10 N are given. It is noteworthy to state that these hardness values are used as an indicator to evaluate the strain-strengthening effect. The hardness variation beneath the wear track was between 667 and 672 HV at 3 N (Fig. 7a), indicating that strain hardening was insignificant. Presumably, the applied hardness measurement method was not able to detect any hardness change due to the low resolution (i.e. spacing between the idents) of micro-hardness mapping. By contrast, a PDZ zone having hardness values between 677 and 686 HV and the maximum depth of 118 μm and length of 488 μm was observed at 10 N (Fig. 7b), correlating well with the width of the wear track (Fig. 6a). Similar findings on the variation of the size of PDZ as a function of load was reported in a study on the wear behaviour of L-PBF-processed SS47. It was shown that the presence of retained austenite played a role in the plasticity of AM-processed SS3,12,50 and that the retained austenite transformed to martensite under plastic deformation (transformation induced plasticity effect), enhancing the strain-hardening of steels51. As the HCMSS samples contain retained austenite according to the previously discussed XRD pattern (Fig. 2e), it is assumed that the retained austenite within the microstructure may have transformed to martensite during the contact, increasing the hardness in the PDZ (Fig. 7b). Further, the slip formation that occurred on the wear track (Figs. 5e, 6f) also indicates the plastic deformation by dislocation slipping under shear stresses was caused during the sliding contact. However, the shear stress generated at 3 N was insufficient to obtain a high dislocation density or transform the retained austenite to martensite on a scale observable by the methods employed; thus, the strain hardening is only observed at 10 N (Fig. 7b).

Cross-sectional hardness maps of the wear track of EBM-processed high carbon martensitic stainless steel at 3 N (a) and 10 N (b).

Full size image