Laser-treatment-induced surface integrity modifications of stainless steel

Scanning of a high-power laser beam on the surface of martensitic stainless steel (SS420) has been studied, addressing the effect of scanning rate V on integrity modifications in the near-surface regions. Structural, compositional, and crystallographic characterizations revealed the presence of ablations, surface melting/resolidification, surface oxidations, and austenite (γ-phase) precipitations when V ≤ 20 mm s−1. Melt pool (MP), heat affected zone (HAZ), and base material have been clearly distinguished at the cross-section of the slow-scanned samples. Adjacent MPs partially overlapped when V = 5 mm s−1. The γ-phase precipitations solely occurred in the MPs, i.e., of ∼ 400 μm deep for V = 5 mm s−1, while oxidations dominantly occurred in the surface regions of shallower than ∼30 μm within the MPs. Compositional analysis revealed increased Cr-, Mn-, and Si-to-Fe ratios at the laser-scanned surface but without variations along the surface normal direction. The enhanced surface hardness has been achieved up to 805 HV, and the hardness monotonically decreased when moving deeper (i.e., ∼1000 μm) into the base material. These observations shed new light on surface engineering of metallic alloys via laser-based direct energy treatments.


Introduction
High-power lasers have been widely applied for precisely manufacturing and remanufacturing of metallic components in various industrial areas [1][2][3]. Direct energy deposition of metallic alloys can be realized by lasercladding in both powder-feeding and -bed configurations using the rapid heating and fast cooling characteristics of laser process [4][5][6][7][8][9]. Pressure waves generated from laser-induced plasma have been developed to laser-shock peening (LSP) for surface integrity enhancement and directly shape-forming of metallic alloys [10][11][12][13][14]. The precise power distribution and spot size of the laser beam as well as the pulse configurations of a pulsed laser, together with the digitalizable and programable control, have been recognized as an advanced technique for surface engineering, e.g., patterning, texturing, and hardening, of metallic alloys [15][16][17][18]. All these applications are based on laser-matter interactions, where energy is transferred from the laser beam to the workpiece. Solidto-liquid and/or solid-to-vapor phase changes may occur in the laser-matter interactions [16,19,20].
In general, laser-based processing is noncontact, clean, and high-speed, which can be feasibly integrated on robot and computer numerical control (CNC) platforms with increased functions, accuracy, and efficiency [21]. Along with the rapid developments of semiconductor technologies in the past decades [22][23][24], microchip lasers have now been entering surface enhancement applications for metal alloys, including LSP, texturing, welding, etc [25][26][27][28]. Driven by advanced additive manufacturing in recent years, integration of laser processes has been proposed for innovative manufacturing and/or remanufacturing [29][30][31]. As a result, laser treatment of metallic alloys has been attracting increasing interest. It has been demonstrated that laser-induced structural modifications of metallic alloys could significantly enhance their hardness and resistance against wear and/or corrosion in aggressive conditions [32][33][34][35]. The structural modifications, including grain refinements and precipitations, are greatly affected by the laser processing environment [36][37][38].
Although the effect of laser treatments on the mechanical and anti-corrosion properties of metal alloys has been extensively investigated [32][33][34][35][36][37][38], more efforts are necessary to understand the relationship between laser treatment parameters and the resultant structural modifications. In an earlier work, we have studied morphological, compositional, and crystallographic changes in the near-surface regions of stainless steel (SS420) induced by laser scanning (λ = 1020 nm and 400 W) with varied scanning rate V [16]. It was found that laserablation occurred at V 20 mm s −1 , while surface melting and resolidification occurred at lowered V. Here, we presented more detailed microstructural characterizations obtained from the cross-section of the laser-scanned samples. Melt pool (MP) and heat affected zone (HAZ) have been clearly distinguished. Austenite (i.e., γ-phase) precipitations occurred in the MPs. It has also been confirmed that the increased Cr-, Mn-, and Si-to-Fe ratios probed from the laser-scanned surface were dominantly induced by faster ablations of Fe than the other elements. However, vertical atomic interdiffusions were absent in the MPs except for the surface oxidations.

Sample preparation
Laser scanning was carried out on a CNC platform (LASERTEC 65 3D hybrid) in an air environment with a continuous-wave fiber laser diode as the light source. The laser wavelength and power were 1020 nm and 400 W, respectively. The spot size of the laser beam was 3.0 mm in diameter on the workpiece. Zig-zag scanning was performed with a longitudinal scanning rate, V, and a stepover distance of 2.1 mm. Commercially annealed , and sample F (V = 5 mm s −1 ), respectively. The cross-section of the samples was prepared by electrical discharge wire-cutting along the transverse direction of the laser-scanning. Grinding by SiC sandpaper in a sequence of 500-, 800-, 1200-, and 2400-grit was processed, which was followed by polishing using oxide polishing suspensions [39].

Materials characterization
Morphological and topographic properties of the laser-scanned samples were characterized using scanning electron microscopy (SEM) and step-profilometer (KLA Tencor: P-16+). X-ray fluorescence (XRF, Bruker M4 TORNADO) and energy dispersive x-ray spectroscopy (EDX) were employed for compositional analysis on macro and micro scales, respectively. X-ray diffraction (XRD, λ Cu-Kα ) with a general area detector was used to investigate the crystallographic structures in the near-surface regions. Line scan XRD was collected along the surface normal direction from the cross section to evaluate the crystallographic changes as a function of depth below the laser-scanned surface. The diameter of the x-ray beam was set at 1.0 mm and the scanning step was set at 100 μm [40]. Hardness was measured from the cross section using a micro-indentation with the applied force of 1.0 kgf.

Results and discussion
3.1. Morphology and compositions measured from laser-scanned surface  significantly reduced while Cr, Mn, and Si are apparently increased; an increase in Al also occurred, although Al was a minor alloying element in the studied SS420. The most regular profiles occurred for sample D, where the highest Cr-, Mn-, and Si-to-Fe ratios occurred at the center of the laser-scanned tracks as indicated by the dashed lines.
Figures 2(a)-(c) present the correlations between the compositional profiles measured by XRF from the laser-scanned surface and the topological mapping measured by the step-profilometer of samples D, E, and F, respectively. The topography in figure 2(a) shows that regular ablations occurred at the laser-scanned tracks, which is manifested by the lower surface height at the track centers than those at the track edges. The boundaries between adjacent tracks (indicated by the vertical dashed lines) were intact. In comparison, the track centers are higher than the track edges for samples E and F in figures 2(b) and (c), which indicate occurrence of surface melting and resolidification at the laser-scanned tracks. The correlations between the compositional profiles and the topographies suggest that the modified elemental distributions could be caused by laser ablations and/or atomic diffusions in the near-surface regions, especially in the MPs. It has been reported that although the vapor pressure of Fe is lower than those of Cr and Mn in their respective pure liquid state, reversed comparisons, i.e., higher vapor pressure of Fe than Cr and Mn, occurred on steel alloy under laser spot welding [41,42]. In this regard, the increased Cr-, Mn-, and Si-to-Fe ratios observed on the surface of samples D, E, and F are reasonably attributed to the higher evaporation rate of Fe than Cr, Mn, and Si during the laser-scanning.  respectively. The beam size and the scanning step of the XRF mapping were 25 and 20 μm, respectively. However, the XRF maps do not exhibit any distinguishable compositional variations across the MPs, the HAZs, and the base materials for both samples, indicating the absence of macroscale atomic diffusions.

Compositions and microstructures measured from cross-section
To investigate the microscale compositional distributions in the MPs, EDX maps were further measured from the cross-section of sample F at the near-surface regions. Figure 4(a) is a typical SEM image recorded from the MP near the surface, which does not exhibit any features except for the surface regions within tens of microns. In this light, EDX was mapped from the surface region as indicated in the dashed box in figure 4(a), and the elemental distributions of Fe, Cr, Mn, Si, Al, and O are presented in figures 4(b)-(g), respectively. They show that Fe is poor while Cr, Mn, Si, Al, and O are rich in the surface layer. This comparison is consistent with the presence of chromium and manganese oxides measured by combining XRD and XRF from the laser-scanned surface [16]. Presented in figure 4(h) are depthprofiles of Fe, Cr, Mn, Si, Al, and O collected from the surface regions as deep as 280 μm, i.e., within the MP (∼400 μm deep). They confirm the absence of any distinguishable compositional gradients except for the oxide at the surface regions shallower than ∼30 μm.
A careful look at the surface oxide layers in figures 4(b)-(g) revealed uniform distributions for Fe, Mn, Al, and O but some variations for Cr and Si. EDX measurements were further carried out with a higher magnification at the surface oxide regions to resolve the microstructures, and the results are presented in figure 5. Figure 5(a) presents a typical SEM image taken from the EDX-mapping area, which clearly shows microscale dendritic structures that are absent from deeper regions. The EDX maps in figures 5(b)-(g) revealed that Cr is apparently, while O is slight, rich in the dendrite branches. In contrast, Si is rich in the interdendritic regions. In comparison, Fe, Mn, and Al do not exhibit significant variations across the dendritic structures. The overall EDX analysis in figure 5(h) shows that Fe is significantly reduced from the oxide surface due to its higher evaporation rate than Cr, Mn, Si, and Al in the steel alloy under the laser-scanning (see above discussion) [41,42]. Both the elemental distributions and the compositional comparisons, as well as the EDX analysis in point mode (results not shown for the sake of brevity), suggest that the dendritic structures are dominated by spinel (Mn x Cr 3-x O 4 ) and/or bixbyite-type (Mn x Cr 2-x O 3 ) oxides. In comparison, the interdendritic regions are dominated by SiO 2 and MnO 2 [43]. Both the macroscale XRF and the microscale EDX measurements provide evidence for an absence of vertical atomic interdiffusions within the MPs except for the surface oxide regions of shallower than ∼30 μm. These results confirm that the increased Cr-, Mn-, and Si-to-Fe ratios probed from the laser-scanned surface originated from the higher Fe evaporations than Cr, Mn, and Si in SS420 under the laserscanning. Figure 6 presents XRD measurements from the cross-section of sample F. Figure 6(a) is a photograph employed for in situ locating the area for XRD measurement. The boundaries of MP (MPB) and HAZ (HAZB) are clearly seen in the photograph. Figure 6(b) presents the XRD patterns collected from locations 1-5 indicated in figure 6(a), i.e., moving from the base material across the HAZB and MPB approaching the surface. It is seen that austenite, i.e., γ-phase, only emerges once approaching the HAZ. The appearance of the γ-phase was accompanied by an intensity weakening of the XRD peaks from the α-phase, e.g., see α(200) of pattern 4 in figure 6(b) [40,44] figure 6(c) presents an enlarged fragment of the XRD patterns covering the γ(111) and α(110) peaks collected with a beam size of 1.0 mm diameter moving from the deeper regions approaching the surface at a step size of 100 μm. Intensity distributions of the γ(111), α(110), and γ(200) peaks are presented in figure 6(d) as a function of distance away from the laser-scanned surface of sample F. It is seen that the γ(111) peak appeared in the surface regions of ∼1.5 mm thick. Considering the x-ray beam size of 1.0 mm diameter, we can conclude that the γ-phase precipitates occurred dominantly in the MPs rather than the HAZs and lower regions.

Crystallographic modifications and surface hardening
It can also be seen in figures 6(c) and (d) that the γ (111) and γ(200) peaks slightly shift to higher angles while the α(110) peak does not show apparent shifts when moving towards the surface. This observation suggests that the peak shifts, i.e., the lattice constant changes, were caused by compositional changes or localized micro precipitations rather than residual stress. Nevertheless, more detailed microstructural studies, such as electron backscatter diffraction and transmission electron microscopy, are necessary to understand the laser-scanninginduced crystallographic modifications in the near-surface regions of SS420.
Finally, Vickers hardness was measured with a load of 1.0 kgf on the cross-section of the laser-scanned samples. It is found that the surface hardening was negligible, i.e., from ∼165 to 215 HV, when V 20 mm s −1 , i.e., for samples B, C, and D, in which laser ablations, rather than surface melting/resolidification, occurred. However, when V is reduced to 10 mm s −1 for sample E, the surface hardness is significantly enhanced to ∼630 HV and exhibits a monotonic decrease when moving deeper into the base material. The effectively hardened surface layer has a thickness of ∼400 μm. A further reduction of V to 5 mm s −1 of sample F led to a surface hardness of ∼805 HV. The thickness of the hardened layer was increased to ∼1000 μm, covering both the MP and HAZ layers. The combination between the V-dependent surface hardening and the crystallographic modifications discussed above provides clues that both the surface oxidation and the γ-phase precipitation, especially the former, contributed together with the HAZs to the observed surface hardening.

Conclusion
Morphological, compositional, and crystallographic properties have been investigated for SS420 upon laserscanning with varied scanning rates. It is found that (i) laser ablations dominated the surface modification at high scanning rate (i.e., 20 mm s) −1 and, instead, surface melting/resolidification became dominant at lower scanning rate (i.e., 10 mm s −1 ); (ii) Remarkably increased Cr-, Mn-, and Si-to-Fe compositional ratios occurred on the laser-scanned surface. The lower the scan rate, the larger the ratios, which have been attributed to the higher evaporation rate of Fe than Cr, Mn, and Si from SS420, rather than atomic diffusions, during the laser-scanning; (iii) Laser-induced oxidations occurred in the surface layer shallower than ∼30 μm. The oxides were formed in microscale dendritic structures with the dendritic branches dominated by spinel (Mn x Cr 3-x O 4 ) and/or bixbyite-type (Mn x Cr 2-x O 3 ) oxides while the interdendritic regions dominated by SiO 2 and MnO 2 ; (iv) γ-phase precipitations occurred in the MPs; and (v) Both the surface oxides and the γ-phase precipitations contributed to the surface hardening.