Tribological Performance of Laser Shock Peened Cold Spray Additive Manufactured 316L Stainless Steel

For many industrialized components, the presence of mechanical abrasion is quite common as they tend to be exposed to a wide array of working environments [1]. This is especially true for standard engineering components such as gears, bearings, pistons, injectors, and metal forming/cutting tools where their surfaces are continually in contact with other surrounding components. The plastic deformation over time will lead to eventual material degradation thus shortening the in-service lifetime of these components [2]. For example, in applications such as nuclear power generation, minimizing friction and wear can reduce the energy consumption of various processes which can improve their efficiencies and overall reliability [3]. Extensive research efforts have been paid to achieve this goal. First, the material of use can be improved either by (1) the replacement of more durable material, (2) employing some type of composite during the fabrication process, or simply by (3) changing the fabrication process altogether. However, these methods suffer from high costs and low feasibility for large-scale applications [4]. In place of this, surface processing technologies have attracted much attention as an alternative method of friction and wear control [5–11].

Among a variety of surface processing technologies, cold spray additive manufacturing (CSAM) is one surface enhancement technique that has seized the attention of many due to its unique solid-state process [12,13]. As a relatively novel technique, CSAM is essentially a low-temperature process that depends on the kinetic energies of impacted particles for material deposition [14–16]. This is quite advantageous in the aspect that the original phase of the particle feedstock is preserved in contrast to traditional thermal-based techniques such as high-velocity oxy-fuel (HVOF) spray and plasma spraying [17]. In its most basic sense, CSAM consists of rapidly launching micrometer-sized particles (typically within the range of 300–1200 m/s [18]) to a working substrate through a de Laval nozzle at supersonic velocities [19]. For these speeds to be achieved, high-pressure gas is applied, typically in the form of helium, nitrogen, or air, which is typically categorized as high-pressure cold spraying (HPCS) [20]. As the particles are rapidly impacted, mechanical bonding is induced by the combination of shear stresses from both the substrate and the coating [21]. Compared to other thermal-spray technologies, there is no particle melting, which will preserve undesirable defects such as grain diffusion, phase transformations, and tensile residual stresses [19].

Given the simplicity and ease of application for CSAM coatings, many industrial sectors such as thermal power plants, nuclear power plants, automotive, aerospace, and marine have greatly benefitted from the prolonged lifespans of their respective components [22,23]. Despite its wide applications, there are some intrinsic disadvantages associated with the CSAM process which has limited its full potential in improving its wear resistance. The surface defects formed during the CSAM process, such as porosity, are quite detrimental to the tribological properties of the coatings [24]. This defect is mainly due to the lack of heterogeneous deformation of the individual particles with respect to the impacted region. Whether it is from the irregularity of the particle morphology/size or from the lack of sufficient plastic deformation, microcracks, and micropores are typically observed within the inter-particle boundaries which can limit the structural integrity of the coating [25]. This in combination with the already brittle nature of these coatings creates early crack propagation which can increase the wear-rate and tendency of third-body wear in the affected region [26]. As consequence, the adherence of cold spray (CS) coatings is negatively impacted by the unmatched tensile and residual stresses at both the top and bottom of the coating which then will further promote early crack initiation [27].

Post-surface processing techniques have been explored as a practice routine to tailor and optimize the tribological properties of CSAM coatings. Although many have studied the effects of post-processing techniques such as heat treatment [28], friction stir processing [29], and shot peening [30] on CSAM coatings, there are few publications that study the effects of laser shock peening (LSP) on CSAM coatings [31,32], especially in the sense of pore reduction. LSP is a laser-based surface processing technique that has been widely utilized to enhance the tribological properties of metallic materials [33–37]. During this process, an intense laser pulse with a duration time of 1–30 ns is typically utilized to generate high shockwave pressure (in an order of GPa) through laser-matter interactions [38]. Throughout this process, a transparent confinement and an ablative coating are typically utilized to absorb the laser energy as well as create a constrained plasma to the surface. However, it should be mentioned that in some cases the transparent confinement and/or ablative coating can be removed during peening [39]. As the laser-induced shockwave propagates through the specimen, the processed material undergoes ultrahigh strain rates (10⁶–10⁷/s) plastic deformation, leading to localized work hardening and grain refinement effect. Additionally, deep compressive stresses are introduced to the material which can enhance the fatigue performance of the sample processed due to its capability to balance external cyclic stresses [40]. From an application standpoint, this will limit the likelihood of early crack propagation which can result in extended operational lifespans of various components [40–43].

Due to its advantages, LSP has been applied as a post-processing technique for a wide variety of materials and applications. Compared to technologies such as shot peening or thermal post-processing techniques (such as laser cladding), LSP can maintain a superior surface quality while preventing the formation of tensile stresses along the surface. This has especially been noted for laser-based additively manufactured (AM) materials that tend to suffer from unwanted pores and voids [44,45]. With many thoroughly studying the effects of LSP on laser-based AM materials, rare research efforts have been paid to investigate the effects of LSP on CSAM coatings. Given its capability to hinder crack propagation and tailor surface roughness, LSP may serve as an efficient approach for CS post-processing. To fill this gap, the feasibility of LSP on CSAM 316L stainless steel (SS) is investigated. Particularly, the surface characteristics and tribological performance will be evaluated. To characterize these changes, a series of experiments were performed on each specimen. These tests include single scratch ball-on-flat tests, microhardness tests, and X-ray diffraction (XRD). The surface morphology of the tested surfaces was also characterized through a scanning electron microscope (SEM) and optical profilometry in order to obtain a better understanding of the relationship between surface morphology and tribological performance.

In this work, the CSAM deposition process was done using an Impact Spray 5/11 CS system (Impact Innovation GmbH, Rattenkirchen, Germany). For the working material, commercially pure (CP) 316L SS powder was employed as a freestanding deposit. From this point onward, the CSAM substrate will be referred to as the CS substrate. A schematic illustration of the operating system is shown in Fig. 1 [15].

The particle size prior to deposition was measured to be 20.17 μm with a standard deviation of 6.64 μm. Micrographs used for the particle measurements as well as a bin distribution of the particle size are shown in Fig. 2 It can be seen that the particles are mainly spherical with the exception of some outlying particles. In the case of non-spherical particles, they may lead to uneven stress concentrations upon impact thus leading to porosity [46]. All other processing parameters for the CSAM deposition process are depicted in Table 1.

The surface morphologies of the CS coatings processed before and after LSP were characterized using a Rtec 3D optical profilometer (Rtec-Instruments, San Jose, CA, USA) with a resolution of 50 nm. A surface area of 1.9 × 1.2 mm² was used to determine the change of surface roughness (SR) parameters induced by LSP. These surface parameters were calculated using the Gwyddion software [52]. To further characterize the change in surface morphology, micrographs of the surface porosity and microstructure were obtained from a JSM-6010LA InTouchScope SEM (JEOL, Tokyo, Japan) for the surfaces of the peened and non-peened specimens. Before microstructural evaluation, all samples were submerged in Carpenter’s Etchant (8.5 g FeCl₃ + 2.4 g CuCl₂ + 122 ml C₂H₅OH + 122 ml HCl + 6 ml HNO₃) [53]. To investigate the possible phase transformation during the LSP process, a Bruker-D2 Phaser (Bruker, Madison, WI, USA) XRD apparatus with Cu-Kα radiation was employed. The excitation voltage and current were set to 40 kV and 25 mA. The 2θ angles spanned from 20 deg to 100 deg with a scan speed of 1 deg/min. The step width was set to 0.02 deg throughout each scan. Surface microhardness (HV) values also were measured post-processing in order to further obtain a better understanding of the surface-strengthening effect of LSP. These readings were measured under the loading conditions of 100-g with a 10-s dwell time (Beuhler-Wilson, Tukon 1202). Each test was replicated ten times to ensure repeatability.

For tribological evaluation, each sample was subjected to a dry single scratch test as per ASTM G133 [54]. These tests were performed using a Rtec multi-function Tribometer (Rtec-Instruments, San Jose, CA, USA) at an ambient room temperature of ∼25 °C. The counter material used was an Al 6061-T6 pin. For each test, a normal load of 10 N was applied for a sliding distance of 10 mm. The moving velocity was set to 0.5 mm s⁻¹. Afterward, the transfer layer formation of the Al pin on the CS flat was studied using an SEM and energy dispersive spectroscopy (EDS) from a JSM-7100FT SEM (JEOL, Tokyo Japan). The transfer layer was then mapped and analyzed. To ensure repeatability, each test was replicated three times.

To understand the influence of LSP on the SR of the CS samples, the change in average surface roughness (S_a) was investigated. The variations of this parameter with respect to laser intensity are shown in Fig. 5. As expected, the influence of LSP at 1.91 GW cm⁻² (i.e., CS + I1) roughened the CS surface by 17.9%. This is to be expected due to the peak pressure shockwave inducing plastic deformation to the surface [55]. However, interestingly, S_a begins to decrease after being subjected to a laser intensity of 2.54 GW cm⁻² (i.e., CS + I2). In fact, relative to the unpeened specimen, the S_a value for the CS + I2 specimen decreased by 4.7%. Upon increasing laser intensity, the CS + I3 and CS + I4 specimens also showed a reduction in S_a by 30.9% and 39.5%.

To obtain additional information regarding the change in surface characteristics with respect to laser intensity, 3D surface profiles alongside their corresponding 2D line profile characterization were effectively analyzed (Fig. 6). It can be seen that across all specimens, there was some presence of surface porosity, as indicated by the listed arrows. This is to be expected due to the intrinsic defects of the CS process, as aforementioned. However, as the laser intensity increased, the surface asperities were effectively deformed and reconstructed, thus sealing most of the surface porosity as a function of laser intensity. This finding is not uncommon, as LSP has been reported to seal surface defects due to the compressive stresses induced by peening [56–60].

In addition to the change in surface morphology, LSP is widely known for its capability of surface grain refinement due to the ultrahigh strain rate-induced dislocation dynamics [61]. Figure 7 depicts the change in microstructure across all specimens. Observing the CS substrate (Fig. 7(a)), the various impacted particles can be clearly differentiated from one another by the black lines. The lengthened morphology of the particles clearly indicates that mechanical and metallurgical bonding (due to the translation of kinetic energy of the rapidly accelerated particles) did occur due to the severe plastic deformation from the deposition process [20]. Post-peening, there is a remarkable distinction between the CS substrate’s structural morphology compared to its original pre-peened state. Across all peened samples (Figs. 7(b)–7(e)), non-uniform deformation bands are densely packed within the grain boundaries with minimal spacing. Upon closer inspection, the microstructure of the CS specimens at lower intensities appears to be largely cellular with homogeneous nanoscale grains. However, at higher laser intensities, the cellular structures appear to be increasingly elongated likely due to the rapid melting from the laser peening. Nonetheless, the microstructural features of all peened specimens are visibly refined from their original state due to the severe plastic deformation from the LSP process.

It can also be noted that the degree of surface porosity also decreased as laser treatment was applied to the specimens. This can be visually seen from the surface micrographs that show the surfaces of the substrates alongside their pore area calculations from the ImageJ software in Fig. 8. For the CS substrate (Figs. 8(a) and 8(b)), various pores can be seen along the surface. However, considering that no confinement media was used, the plasma formation from the laser peen was not confined, which likely resulted in the surface experiencing a form of atomistic diffusion [62–64]. As such, the localized particles are reorganized into a more robust and dense structure as a function of laser intensity, as shown in Figs. 8(c)–8(j) [65,66]. A quantitative analysis of the pore count can be shown in Fig. 9, which confirms that the percentage of porosity decreases due to LSP treatment. For the CS substrate, a density of 99.15% was recorded, whereas the CS + I1, CS + I2, CS + I3, and CS + I4 substrates had densities of 99.35%, 99.54%, 99.73%, and 99.92%.

From a cross-sectional perspective, micrographs of the CS specimen (without peening), CS + I2, and CS + I4 specimens are shown in Fig. 10. It can be seen that for the CS specimen (Figs. 10(a)–10(d), there appears to be a large number of pores dispersed along the cross section. However, when subjected to peening, the porosity along the surface (Figs. 10(b), 10(c), 10(e) and 10(f)) greatly decreases. This is especially evident for the CS + I4 specimen (Figs. 10(c)–10(f)), of which the depth of the pore closure appears to be the greatest when compared to the CS + I2 specimen (Figs. 10(b)–10(e)). To explicate, the surface recoil pressure and vaporization of the plasma were effectively confined in the ablative coating, which likely assisted with surface densification. As the depth increases, the extreme plastic deformation induced by the peak shockwave pressure allowed for compressive stresses to seal the pores, thus allowing for the affected region to be effectively densified. This finding is supported by Glaser et al. [67] as well as other peening-related literature [68].

To further quantify the degree of plastic deformation from LSP, XRD patterns of the pre-and-post-peened specimens were investigated, as shown in Fig. 11. Based on the XRD spectra, all samples exhibit a gamma (γ) structure with no phase changes induced by LSP. Typically in LSP literature, a martensite transformation occurs due to the high degree of plastic strain from peening. According to Chen et al. [69], strain rates as low as 10¹ and 10⁻³ s⁻¹ induced martensite (ε) transformations in stainless steels due to the high degree of dislocation motions. With LSP-inducing strain rates as high as 10⁶ s⁻¹ [70], it would be expected that some degree of a phase transformation would occur. However, the localized surface heating from the laser irradiation process could potentially suppress the γ → ε transformation, as the degree of plastic strain is suppressed [71]. Additionally, due to the strain rates of LSP being much higher than the necessary strain rates for a martensite transformation, the excessive amount of dislocation twins can impede dislocation activation thus impeding the formation of martensite [72].

Despite no phase change, the full width at half maxima (FWHM) of each peak was observed to increase, as shown in Fig. 12(a). It is well known that when a peak is broadened, it indicates that the microstructure of the material is both refined and re-oriented due to the presence of external stresses, which follows suit with the influence of surface peening [73]. Understanding this, the widely utilized Williamson–Hall method [74,75] can be applied to determine the crystallite size, micro-strain, and dislocation density of the specimens. These findings are shown in Figs. 12(b)–12(d) and Table 3. According to these findings, the crystallite size for CS + LSP specimens (Fig. 12(b)) showed a net decrease compared to the unpeened CS specimen. Interestingly enough, the CS + I1 specimen showed the greatest reduction in crystallite size (with a 52.9% reduction), with each following laser intensity resulting in a gradual increase. This likely is a result of the microstructural re-orientation observed in Fig. 7, in which the microstructures at higher intensities appear to be less uniform/refined and more elongated. Also, the intensity of the laser at higher laser intensities likely resulted in greater diffusion, which inhibited the formation of finer crystallite sizes. A similar trend can also be seen with dislocation density (Fig. 12(d)), as crystallite size tends to share an inverse relationship to dislocation accumulation. Nonetheless, across all specimens, the crystallite size and dislocation density yielded improvements over the base CS substrate. For the strain (Fig. 12(c)), there was an initial decrease with a gradual increase as a function of laser intensity. Although it might be counterintuitive, the initial decrease in strain can be attributed to the initial microstructural reorientation (due to diffusion) observed in Fig. 7 [76], as CS substrates in their base form tend to already be extremely work-hardened. However, as the laser intensity increases, the degree of plastic deformation due to the higher strain rates (in addition to the rapid heating) occurs resulting in greater flow stresses. With the following conditions, the strain is increased despite the elongated microstructural evolution [77].

When subjected to tribological loading, the effect of LSP drastically reduced the coefficient of friction (COF) as demonstrated in single scratch tests under dry conditions, as shown in Fig. 14(a). For all specimens, the COF was relatively stable throughout sliding. When subjected to LSP, there were some slight oscillations with the COF during sliding. This was especially evident for specimen CS + I1. However, upon increasing the laser intensity, a more-or-less monotonic behavior was observed after the rapid run-in period, which was largely pronounced within the first few seconds of sliding. This stabilization can also indicate that there is no significant plastic deformation occurring along the surface, thus suggesting less tangential stresses along the tribo-loaded interfaces [82,83]. When considering their averaged values (Fig. 14(b)), the COF does indeed follow a direct decrease as a function of laser intensity. The reduction was quantified by 53.6% for the CS + I1 specimen and 60.6%, 65.5%, and 72.6% for the CS + I2, CS + I3, and CS + I4 specimens.

Observing the wear track of the specimens (Figs. 15(a), 15(c), 15(e), 15(g), and 15(i)), EDS observations (Figs. 15(b), 15(d), 15(f), 15(h), and 15(j)) were made in order to quantify the degree of Al-6061 transfer material that occurred during tribo-loading. Following a similar trend as the aforementioned COF, there was a gradual decrease of transferred Al to the CS surface as the laser intensity increased. However, this was visually not the case for sample CS + I1 (Fig. 15(d)), as there appears to be slightly more Al accumulation than in sample CS + I0 (Fig. 15(b)). Specifically, the CS + I1 specimen had a 5.33% transfer layer, whereas the CS substrate had a 4.61% transfer layer. In the case of the CS + I2 (Fig. 15(f)), CS + I3 (Fig. 15(h)), and CS + I4 (Fig. 15(j)) specimens, the transfer layer concentration, and disbursement were largely decreased by 2.78%, 1.38%, and 0.33%. SEM observations (Figs. 15(a), 15(c), 15(e), 15(g), and 15(i)) also indicate that the presence of the wear track becomes increasingly faint as the laser intensity increases. This finding insinuates that a lesser degree of plastic deformation occurred during tribo-loading, which further supports the aforementioned friction findings. Likely, the smoothening of the surface asperities allows for greater resistance to shearing from the counter interface, which resulted in a wear track largely free from visual deformation. To quantify this change, the width of the wear tracks was measured in order to further understand the change in tribological response across the specimens. The wear track width for the CS, CS + I1, CS + I2, CS + I3, and CS + I4 specimens was recorded at 149.84 ± 12.20 μm, 120.79 ± 10.41 μm, 102.63 ± 4.36 μm, 101.67 ± 4.62 μm, and 95.24 ± 2.81 μm. This change was measured against the change in aluminum transfer layer content, as shown in Fig. 16. It can be seen that the influence of LSP allows for the wear track width and transfer layer content to share the same decreasing trend, which is again attributed to the smoothening and strengthening of the CS surface post-LSP.

To understand the tribological mechanisms, three synergistic factors should be taken into consideration. The first factor lies in the response of the surface morphology to tribo-loaded deformation, whereas the second and third factors lie in the material-related properties and surface texture of the tribo-loaded regions. To elucidate the first factor, it is important to keep in mind the reduction in surface porosity due to the laser treatment (Fig. 7). Under tribo-loading, it is well known that the influence of pores can increase the contact area of the counter material (i.e., Al6061 pin) [53,84,85]. This creates an implication that the degree of deformation along a porous edge would increase, thus creating greater frictional forces compared to a non-porous surface [86]. Since the CS steel substrate also has a higher hardness than the Al6061 pin, this idea is exacerbated, as surface defects (in the form of porosity) can result in a great degree of plowing on the counter surface, thus increasing the localized frictional force [87]. From a material transfer point of view, the surface pores can also act as active sites that would actively accumulate the counter Al material surface. The accumulation of material in the pore sites was quite evident in the EDS maps (Fig. 15), where the distribution of Al along the wear track generally becomes less scattered as the laser intensity increases. This would suggest that the surface pores do indeed work to accumulate the counter Al material to the surface. It should be mentioned that in the case of sample CS + I1, there is a slightly greater degree of Al accumulation than in the CS specimen. To explicate, the surface features should be considered. In the case of sample CS + I1, the size of the pores is somewhat similar to the CS specimen (Fig. 8). With sample CS + I1 also having a greater surface roughness, the degree of surface shearing (and frictional force) to overcome the roughened surface is increased. As such, the plowing forces on the surface-pin interface increase which would result in a greater transfer layer [88]. From another perspective, the porous regions can also be sensitive to elevated stress concentrations, which could result in brittle fracturing thus extending the exposed porous region and resulting in greater Al accumulation [89–91].

The change in material properties also has a large influence on the tribological response of the LSP specimens. This was especially evident with the microstructural features of all peened specimens, as grain refinement due to the ultra-high strain rate of plastic deformation from the LSP allowed for the formation of a denser microstructure. The crystallite size and dislocation density measurements also confirm this finding. According to the findings, the role of crystallite size in the tribological characteristics of the peened specimens was more dominant with lower intensities, whereas the dominant factor at higher intensities is surface densification. Nonetheless, the microstructural features of the CS + LSP specimens were refined. It is commonly known that when a microstructure is refined, the hardness is increased due to the well-known Hall–Petch relationship, with then can improve the adhesion resistance of a material. To further support this notion, Archard’s theoretical equation for wear [92] suggests that harder surfaces result in less surface deformation under tribo-loading. However, it should be mentioned that the tribological characteristics of materials such as polymers do not necessarily follow this rule, as increases in hardness can promote a brittle-like behavior, thus resulting in greater surface deformation [93–95]. Although the brittle-like behavior is not the case for all polymers (depending on their composition), this variation can exist. Nonetheless, when these instances occur for metals (such as in this work), it is expected that there would be a decrease in friction due to the lesser degree of shearing from the tribo-loaded surface. To further explicate, the aforementioned notion essentially suggests that flow stress during tribo-loading is effectively reduced due larger fraction of atoms along the grain boundaries [96]. This explanation further is validated by observing both the change in wear track width and adhesive wear along the wear track surface. Aside from the COF being decreased, the lesser degree of plastic deformation also suggests that fewer asperities are interlocked, thus resulting in a lesser area of contact area and less material transfer [97], as seen in this work.

The overall mechanisms proposed in this work as schematically shown in Fig. 18. As such, this model suggests three key findings. First, LSP is indeed a viable post-processing technique to augment the surface topography of CS specimens. Second, the influence of LSP can indeed reduce the surface frictional response by pore closure, surface smoothing, and surface hardening. A hybrid surface parameter, ζ_F, was introduced to relate this relationship. Third, the transfer layer from tribo-loading can also be controlled through the aforementioned conclusions. Collectively, this work acts as a novel foundation for understanding the frictional response and transfer layer mechanisms of post-processed CS coatings, a necessity that is needed to improve their operational performance.

In this work, the effect of LSP on the surface morphology, microstructure, and tribological performance of CS 316L stainless steel was investigated. The primary findings in this work are as followed:

• As the laser intensity increased, the degree of surface roughness gradually increased due to the plastic deformation from the plasma-induced shockwaves of the laser. However, after utilizing a 2.54 GW cm⁻² laser intensity, the surface roughness decreased. This finding was likely attributed to the high amplitude of compressive stresses along the surface which can result in pore healing.

• Due to the synergistic combination of severe plastic deformation and rapid heating/cooling, homogeneous nanoscale cellular grains were fabricated along the surface. This was especially evident in higher laser processing intensities.

• LSP resulted in a refinement in crystallinity, a net increase in dislocation density, and a net increase in strain due to the influence of severe compressive stresses. However, the crystallinity appeared to have an increasing trend whereas the dislocation density had a decreasing trend. Similarly, the strain initially decreased but began increasing as laser intensity became more intense. These findings are attributed to the structural reformation dominant mechanisms of LSP without a coating, as lower intensities were more deformation-based whereas higher intensities were diffusion-based.

• The COF and degree of Al transfer layer to the CS + LSP specimens were observed to drastically decrease. Findings indicate that this was largely due to the densification of the surface, having desirable surface roughness characteristics (i.e., low surface roughness), and grain refinement due to the LSP process. In an attempt to simplify these key relationships, a new hybrid surface parameter, ζ_F, was introduced to relate these factors to frictional performance.

Based on the findings of this work, it appears that LSP not only acts as a technique that improves the tribological performance of CS deposits but is also a useful technique to densify the surface. With this work being among the first that studies the effect of LSP on the surface morphology and tribological performance of CS substrates, the authors suggest further investigation into this topic should be conducted. By doing so, many industries which utilize CS as a protective coating will greatly benefit.

The authors would like to thank the National Science Foundation (Grant No. CHE-1429768) for allowing the use of the powder X-ray diffractometer.

There are no conflicts of interest.

The data sets generated and supporting the findings of this article are obtainable from the corresponding author upon reasonable request.