Erosion development in AISI 316L stainless steel under pulsating water jet treatment

Engineering Science and Technology, an International Journal 50 (2024) 101630 Available online 26 January 2024 2215-0986/© 2024 THE AUTHORS. Published by Elsevier BV on behalf of Karabuk University. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/). Erosion development in AISI 316L stainless steel under pulsating water jet treatment Sergej Hloch a , b , * , Jakub Poloprudský c , d , Filip ˇ Siˇ ska c , Tom´ aˇ s Babinský c , Akash Nag e , Alice Chlupov´ a c , Tom´ aˇ s Kruml c a Faculty of Manufacturing Technologies TUKE with a seat in Preˇ sov, Slovakia b The Czech Academy of Sciences, Institute of Geonics, Czech Republic c Institute of Physics of Materials, The Czech Academy of Sciences, Brno, Czech Republic d Faculty of Mechanical Engineering, Brno University of Technology, Brno, Czech Republic e Faculty of Mechanical Engineering, VSB –Technical University of Ostrava, Ostrava, Czech Republic ARTICLE INFO Keywords: Erosion Wear Pulsating water jet AISI 316L Stainless steel Surface integrity Subsurface hardening Microhardness ABSTRACT Erosion of solids by liquid droplets is a phenomenon which is a compromise between mechanical properties of the material and droplet hydrodynamic parameters. While a number of studies deal with the deformation of drops, the deformation evolution inside the material has not yet been revealed, mainly from the point of view of the time action of the impinging drops The mechanical response of AISI 316L was investigated under gradually increasing numbers of impingements of liquid droplets, with a droplet volume of V d ≅0.9 mm 3 , generated by an ultrasonic pulsating water jet with the frequency f=40 kHz from 1 to 20 s. The surface roughness and the wear rates were determined using a laser profilometer. The cross-section of the selected samples was subjected to microhardness measurement with a load of 0.150 N in a 2D grid, which included the entire perimeter of the deformed area. The minimal microhardness measurement grid under the groove had dimensions of 15 ×15 indents, equal to an area of approximately 450 ×600 µm. A maximum hardness increase was observed at the lowest measured depth of 30 µm. An increase in hardness was observed at 300 µm below the surface. The hardening in the deeper subsurface area was most likely caused by shear stress. This shows the high degree of similitude between the solid and liquid droplet impingements. The results indicate that the currently accepted theory on the development of erosion over time has shortcomings, as demonstrated in this work by the ratio between the utilised droplet diameter and the grain size of the material. 1. Introduction The gradual deformation of a material by the repeated action of liquid droplets in natural and technical situations causes a change in structural integrity and results in material loss. The physical background, from the point of view of the droplets’deformation when hitting flooded and non-flooded surfaces, was described in the 19th century [1]; this study is considered to be the very first. It was followed by studies concerned with the impact pressure of the so-called hammer effect [2,3], which were indirectly related. These two theoretical streams were later united by practical problems associated with the degradation of the aerodynamic shape of the blades in steam turbines. In this case, there was a volumetric removal of material and a destruction of the aerodynamic shape, leading to a reduction in the required nominal power [4]. Historically, this is how efforts to understand the interaction of water droplets with solids began. A year later, in 1928, Cook [5] derived a one-dimensional relationship describing the impact pressure acting on a solid surface in the form of a liquid droplet. The impact pressure p i depends on the pressure of the liquid, which depends on the droplet velocity v i at the moment of impact, the density, and the speed of the elastic wave cin the liquid (propagated supersonically). Later, Heynmann [6] extended the equation to include the variant nature of the droplet velocity v i and the constant k, whose value depends on the properties of the liquid, as shown in Eq. (1): pi= ρ •c•vi•(1+k•vi ci)(1) However, there are works that revised and improved the impact * Corresponding author. E-mail address: [email protected] (S. Hloch). Contents lists available at ScienceDirect Engineering Science and Technology, an International Journal journal homepage: www.elsevier.com/locate/jestch https://doi.org/10.1016/j.jestch.2024.101630 Received 1 September 2023; Received in revised form 15 December 2023; Accepted 17 January 2024 Engineering Science and Technology, an International Journal 50 (2024) 101630 2 pressure Eq. [7] which represents the instant that the droplet impacts a solid surface. During the liquid–solid interaction, a pressure transition is created, generating an elastic wave that propagates in both the material and the droplet directions. As a result, the density at the bottom of the water droplet increases for a short time [8]. According to Huygens’ construction, the condensed area is bounded by a surface, and a shock wave is formed by an envelope [9]. This phenomenon transfers energy into the material, and the interatomic bonds are disturbed by the intensive hydrodynamic load [10,11]. Subsequently, the accumulated energy in the contact region of the droplet is released in the form of radial outflow. Droplet hydrodynamic characteristics such as impact pressure followed by lateral outflow were studied by several authors and extensively summarised in [12]. The lateral outflow [9] causes additional damage by shearing exposed grains or by microtunnelling [13]. The penetration reaches greater depths than the measurable depth of the groove or trace [14]. The material response to the hydrodynamical load in the form of repeated drops, either periodically [15,16] or nonperiodically [17,18], depends on the mechanical properties [18] of the material. The erosion stages can be described in terms of exposure time or number of impacts of water droplet [19]. The erosion stages depend on different parameters, such as drop impact velocity, impact duration, the cumulative density of the droplets, and the impact angle; these parameters are described in detail in [20]. The phenomena associated with the action of external forces in the form of repetitive water droplet impingements on materials are the subject of many research questions. The impacts of water droplets lead to the occurrence of microstructural changes in a material [15,20,21], which alter its mechanical properties [21]. The study of such manifestations is necessary not only to clarify the basic phenomena concerning the time intervals of erosion stage evolution, but also for practical needs [22,23]. The possibilities of utilising the hammer effect [24 24] associated with water droplet interactions can be used in the treatment of welded joints [25], the roughening of biocompatible materials or peening to improve fatigue response, or as an alternative to sandblasting or laser shot peening. Although much attention has been paid to material deformation, the evolution of subsurface deformation due to the time-dependent accumulated deformation by impacts from multiple droplets is still not sufficiently understood [26]. Studies show that erosion damage accumulation is caused by the periodic and cumulative action of two phenomena: impact pressure and lateral outflow. At the same time, it also depends on the frequency of impacts. In the initial stage, the exposed area gradually strengthens, which increases the resistance [27] to the external load for a period [14]. Due to the additional impact of the droplets, the hardened layer begins to disintegrate. Pits begin to appear in more deformed areas or in areas with a lower level of integrity. The reinforced layer is opened, and the water penetrates the subsurface areas. Microtunnelling can occur in unreinforced areas of the material [28]. Based on the summarization of previous research on the phenomena associated with the action of water droplets on solid surfaces of materials, this work presents a way how to observe erosion evolution more precisely from the point of view of mechanical response of the material. The innovative approach presented in the work is two-dimensional grid mapping of the hardness gradient. Such detailed approach can allow to observe the prevalence of the erosion agents –the impact pressure and lateral outflow for erosion stages. The aim of this study is to bridge this research gap and to assess the hypothesis how different time exposures affect the erosion evolution in the material in surface and subsurface areas. Erosion evolution was investigated over a series of exposure times (t=1–20 s), keeping all the technological conditions constant throughout the runs within the experiments. The collision epicentre was investigated to identify deformation in the surface and subsurface areas. Two-dimensional grid mapping of sub-surface microhardness was measured and assessed to detect the development of local areas and the progression of subsurface deformation. To exclude the effect of initial surface roughness [29], the sample surfaces were polished using an oxide polishing suspension. 2. Materials and methods 2.1. Material Austenitic stainless steel AISI 316L was selected for the current study for several technological and measurement reasons. The material consists mostly of austenite (face-centred cubic lattice) and a small percentage of ferritic bands (body-centred cubic lattice) as seen on Fig. 1a. The material is highly homogenous which enhances the repeatability and reliability of the microhardness measurement of the selected material. Due to its superior properties, AISI 316L is commonly used in various application domains, such as aerospace, power plant turbines, and the inner lining of hydraulic pipes at various temperatures [30–34]. Presently, stainless steel 316L is widely used also for implants due to its biocompatibility [35]. AISI 316L is also corrosion-resistant in humid tap water environments, such as the environment that occurs during the processing of the material by the PWJ technology. Chemical composition of selected experimental material AISI 316L in wt. % is as follows: Cr 16.63, Ni 10.00, Mo 2.04, Mn 1.26, Si 0.38, C 0.02, N 0.04, P 0.032, S 0.001, Fe rest. [36]. The main properties of AISI 316L are listed in Table 1. The grain size was assessed using electron backscatter diffraction (EBSD) using an equivalent circle diameter to interpret the results (Fig. 1b). The 10◦was selected as grain boundary conditions and special boundaries, including mostly twins were considered as separate grains. The grain orientation and shape are shown in inverse pole figure (IPF) map on Fig. 1b. The hardness of as-received material was evaluated using Vickers method on Duramin hardness tester and the resulting value is in Table 1. Elastic modulus and tensile yield strength, as well as tensile ultimate strength, were evaluated via a tensile test which was performed using an MTS 810 servo-hydraulic testing machine. The phase composition was measured using a 2θscan on an XRD Empyrean device (Malvern Panalytical, UK) with Co-K α radiation. The results were evaluated using HighScore Plus software with an ICSD database. 2.2. Experimental method The impact of liquid droplets was achieved using the patented device called an ultrasonically generated PWJ [37], depicted schematically in Fig. 2a. The advantage of the PWJ is its ability to create drops with different kinematic parameters. The basis is an acoustic generator that generates a signal with a frequency of 40 kHz. The electrical energy is converted into mechanical energy using piezoceramics. The amplitude is amplified by a stepped sonotrode, which is located in a chamber where it causes periodic impulses in high-pressured water in the form of pressurised fluctuations. Water is, in fact, an extension of the sonotrode in the form of a discontinuous hydrodynamic system (Fig. 2b). For each flow, it is necessary to adjust the chamber length to the resonance [38] in an impedance region. For this experiment, the chamber length was adjusted to l c =10 mm (Fig. 2a). The density of the PWJ corresponds to the appropriate standoff distance from the surface of the material that allows the most efficient utilisation of the impact pressure p i . The effective standoff distance z=60 mm for the water flow rate Q=2.15 L/ min (Table 2) was adjusted using the stair trajectory to eliminate the Doppler effect [39]. During the experiment, the hydraulic parameters were kept constant at the supply pressure p=50 MPa, with the nozzle diameter d=0.4 mm. In these conditions, the water jet impacts the surface of the material with subsonic speed v w =285 m/s. The only variable during the experiments was the exposure time of the PWJ (Fig. 2c) to control the strain hardening, erosion, and wear accumulation damage. Under the same conditions, a control group of samples was treated using the CWJ (Table 2). The entire experimental run mentioned in Table 2 was repeated four times for better statistical validity of the results. S. Hloch et al. Engineering Science and Technology, an International Journal 50 (2024) 101630 3 2.3. Characterization 2.3.1. Evaluation of erosion depth and volume The impacted areas after the action of PWJ were scanned using a MicroProf FRT (FormFactor, USA) noncontact profilometer. The scanned data was then analysed using Mountains software to determine the erosion depth and volume removed for every exposure time. The surface roughness parameters, such as the arithmetical mean height (R a ) and maximum height of the profile (R z ), were also calculated from the scanned data. 2.3.2. SEM After the application of the water jet for a selected number of water impingements, the specimen surface was observed using the Tescan Lyra3 XMU FEG/SEMxFIB (Tescan, Czech Republic) scanning electron microscope (SEM). For the assessment of the grain size and the crystallographic orientation of the samples, the EBSD measurement technique was used. For subsurface observation of the dislocation arrangement in the transmission electron microscope (TEM), lamellae were prepared from the central area of the eroded region using a focused ion beam (FIB) in the SEM. 2.3.3. TEM Two lamellae were extracted from the specimen exposed to the PWJ for 5 s, and the microstructures were compared with the untreated material. The lamellae were observed using a transmission electron microscope (TEM) JEOL JEM-2100F (Jeol, Japan). The TEM operated at 200 kV and allowed imaging in scanning mode, i.e., scanning transmission electron microscopy (STEM). STEM imaging was used for the visualisation of the microstructure on nano and micrometric scales. The dislocation density measurements were performed on the micrographs with suitable magnifications using the method by Ham [Ham 1961]. The method considers drawing random lines in the micrograph, counting the number of intersections of lines and dislocations, and eventually estimation of the dislocation density ρ as follows ρ =2N/Lt, where N is number of intersections L is length of the lines and t is thickness of the foil. The thickness of the foil was estimated using the CBED technique [Hirsh 1965, Morniroli 2002]. 2.3.4. Metallography After the surface observations, the centre of the water droplet-treated area was determined using a combination of optical microscopy and the knowledge of the distances between the treated areas (controlled by a robot). Then, metallographic cross-sections 0.2 mm away from the centre of the affected area were prepared using precise electrical discharge machining. The cross-sections underwent a standard metallographic procedure, consisting of hot mounting in Polyfast (Struers, Denmark) black Bakelite with carbon filling mounting resin at 150 ◦C for Fig. 1. EBSD analysis of as received material showing a) phase distribution in the as received material and b) IPF map containing crystallographic orientation of individual grains. Note number of grains containing twin boundaries. Table 1 Mechanical properties of AISI 316L. Material Elasticity modulus [GPa] Tensile yield strength [MPa] Tensile ultimate strength [MPa] Average grain size [µm] Hardness HV0.2 Phase composition [%] AISI 316L 198 322 625 10.2 ±6.7 184 ±10 98.2 FCC 1.8 BCC Note: HV0.2 refers to the Vickers hardness test with 200 g (0.2 kg) weight. S. Hloch et al. Engineering Science and Technology, an International Journal 50 (2024) 101630 4 10 min, grinding, and polishing with increasing fineness up to a 0.25 μ m diamond suspension. 2.3.5. Microhardness The hardness of materials was evaluated at two levels. The macrohardness of as-received material was evaluated using a Duramin hardness tester (Struers, USA) with a load of 200 g (HV0.2). This result is presented in Table 1.The micro-hardness of untreated, CWJ treated and PWJ treated cross-sections was measured using the universal instrumented nanomechanical testing machine Zwick ZHN (Zwick Roell, Germany). The force/time curve of the microhardness measurement is shown in Fig. 3a. The measurement was conducted using a Vickers-type Fig. 2. Experimental setup: a) ultrasonically generated PWJ; b) interaction with the surface of the sample; c) defined exposure time. The experiment was repeated four times to determine the potential variance due to the stochastic nature of erosion. Table 2 Technological conditions of the experiments. Run (n =4) Supply pressure p[MPa] Frequency f[kHz] Ultrasonic power P[W] Acoustic chamber length lc [mm] Nozzle diameter d[mm] Flow velocity v w [m/s] Flow rate Q[l/min] Standoff distance z[mm] Exposure time t[s] PWJ 50 41.00 253 10 0.4 284.89 2.15 60 1–20 s CWJ 0.00 0 * Discharge coefficient cd =0.92 for Hammelmann nozzle insert. S. Hloch et al. Engineering Science and Technology, an International Journal 50 (2024) 101630 5 indenter with a loading force of 15 mN“on a 2D grid (Fig. 3b). The example of 2D grid of indents is depicted on Fig. 3c. The example of results of nano-hardness measurement are in Fig. 3d in the form of a heat map. The treated surface was measured starting at a depth of 30 µm from the surface. The minimal measurement grid under the treated surface consisted of 15 ×15 indents. When the apparent width of the groove was wider than 600 µm, additional columns were added, as was the case at t =15 s and 20 s with the PWJ. The distance between the indents was 30 µm in the depth direction and 40 µm in the direction parallel to the surface (Y and X coordinates, respectively, as shown in the schematics in Fig. 3b). The distance between the indents was measured on an unaffected surface. In the case of the potential collision of the indenter with an empty groove, the indent was not taken, in order to protect the measuring device. Five columns were taken under the untreated surface. 3. Results and discussion 3.1. Erosion behaviour Fig. 4 shows the nonlinear erosion evolution using a box-andwhiskers chart, which provides descriptive data analysis of the effect of different exposure times of multi-droplet impingements on the erosion depth (µm) and volume removal (mm 3 ). From the point of view of erosion development, the graph can be divided into several areas. At 1 and 2 s, no irregularities were detected. However, traces of the action at those timepoints are visible in Fig. 5a and Fig. 5b. In this time interval, the surface was roughened, and compressive stresses were introduced. In a suitable technological setting, this stage can be used for the mechanical strengthening of subsurface layers using the hammer effect. The reconstruction of the linear erosion development in the initial stages in the time interval of 1–7 s is shown in Fig. 5. We took this time into consideration because the first depth data were obtained at 7 s (Fig. 6) The time interval of 1–6 s was evaluated using surface profile roughness parameters (Ra and Rz). The erosion depth curve fits a cubic power model for the selected experimental range of exposure time with a higher coefficient of determination, R 2 =0.9658. Whereas the volume removal trend shows an exponential trend for the same exposure time domain with a R 2 =0.9379. The difference in nature of both trends can be attributed to the erosion mechanism occurring at longer exposure times. For longer exposure time, the water layer present in the erosion Fig. 3. A) force applied during the indentation with respect to time, b) indentation patterns used for nanohardness measurements on sectioned aisi 316 l exposed to pwj/cwj, c) sem image of the measured area, d) example of heat map showing hardness in the individual zones for sample with erosion time t =10 s. Fig. 4. Nonlinear influence of erosion time of PWJ in 1 s increments on wear rates: volume removal and depth (n =4). S. Hloch et al. Engineering Science and Technology, an International Journal 50 (2024) 101630 6 Fig. 5. Reconstruction of the erosion development using SEM images: a) t =1 s, b) t =2 s, c) t =3 s, d) t =4 s, e) t =5 s, f) t =6 s, g) t =7 s, with the indication of the centreline (white dotted line) and the edge of the water nozzle (yellow dotted lines) to highlight the nonuniform distribution of stresses induced by multi-droplet impingements. In e) two FIB lamellas for the TEM analysis were extracted in the areas marked by yellow squares. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.) Fig. 6. The erosion depth and profile evolution, data visualisation processed by MicroProf FRT and SPIP. a) Irregularities expressed by surface profile parameters Ra and Rz. b) The first measurable trace was identified after 7 s. Areas exposed to PWJ with erosion time 1–6 s was evaluated by surface roughness profile parameters. S. Hloch et al. Engineering Science and Technology, an International Journal 50 (2024) 101630 7 crater already formed by the initial droplets resists the further incoming water droplets to interact efficiently with the material, leading to the generation of erosion depth. However, in this phase, the lateral jetting becomes dominant, which increases the surface area of the crater opening, making the overall crater volume increase. These differences in the erosion mechanism are mainly responsible for the different trends in erosion depth and volume removal. However, it must also be emphasized that these predictive models are statistically valid for the selected experimental domain and can vary for longer exposure time. Fig. 5 shows the erosion development caused by the high frequency and the periodic impact of the water droplets on the electrochemically polished surface of the AISI 316L. The captured development spans from 1 to 7 s, at which point the cracks and voids became more apparent. The image in Fig. 5 is divided vertically into four areas. The inner area is the area that corresponds to the diameter of the water nozzle that was used, which had a diameter d=0.4 mm. This region is divided by an axis that indicates the centre of the nozzle exit and, thus, the potentially most effective core of the PWJ. However, the action radius of the impact pressure is larger than the diameter of the water nozzle, which is noticeable after even 1 s. Erosion manifestations in the outer region beyond the diameter of the water jet appear after a time delay. It is known that, in ideal cases [40], stress waves are induced in both the material and the water droplets during the repeated action of the water droplets, and the waves are transmitted according to Huygens construction in spherical wavefronts [41].Fig. 5 can be considered a representative case. We can refer to the so-called pre-incubation stage, i. e., before the first signs of deformation appear. Up to this time, no surface deformities are observable. The material can withstand impingement without external and internal changes of shape and microstructure, as shown in Fig. 4 and Fig. 5a. However, it is necessary to consider the inhomogeneity of the material since there may exist areas with a high probability of early damage. Regarding the action of external forces in the form of water droplets, it is necessary to consider the Weber number. In our case, the Weber number was higher due to the theoretical velocity of the stream at the exit of the nozzle, v=284.89 m/s, which was lower than the velocity of individual water pulses. This observation indicates the accumulation of erosion action on the surface and towards the material itself. Although we consider these to be shortcomings, it is still possible to assume gradual erosion development (Fig. 6) The first part of Fig. 6 shows the influence of exposure time (from t=1 to 6 s) on the generated surface roughness. The threshold at which to consider a sample surface roughened or eroded was selected as Ra =3µm because this corresponded to the recording of the first inequality. Therefore, it was observed that the PWJ with an exposure time of t≤6 s generated surface depressions which could be defined as roughness; after this, it was considered to be erosion depth. The graphs (Fig. 6a) corresponding to Ra and Rz show an increasing trend in the surface roughness with the longer exposure time due to the increasing number of impingements per unit area of the sample. A minimum average surface roughness of Ra =0.505 µm was measured with an exposure time of t=1 s, and a fivefold increase in the value was observed at t=6 s, when Ra =2.53 µm. The generated surface roughness can be attributed to the surface depressions created by the repetitive action of the PWJ deforming the sample surface. These surface depressions also act as microsurface asperities, which interact with the lateral outflow of the jet and result in further deepening and enlarging of the roughened area. A similar observation was also depicted in the work of Huang et al. [35], where it was shown that the lateral outflow of the jet caused damage to the grain boundary regions of the material due to induced surface shear stresses, which led to the formation of surface irregularities and microvoids (see Fig. 5). The irregularities created during the exposure time t=1––6 s (Fig. 5) increased with the increase in successive impingements of the water droplets over the same region of impact, as seen in Fig. 6a. On the other hand, these observations related to the generation of variable surface roughness open a new potential application of the PWJ in the form of the utilisation of controlled water droplet impingements to generate desired surface roughness for the preparation of materials for various engineering and medical applications [42]. One of the possible applications is the roughening of implant surfaces to increase the osseointegration rate, which depends on the surface conditions [15,43]. Fig. 6a shows that the maximum surface roughness height parameter (Rz) follows the same trend as Ra, with Rz =2.15 µm and 9.84 µm for t= 1 and 6 s, respectively. For both Ra and Rz, the spread of the calculated values is larger for the surfaces generated with an exposure time of t=6 s. This is related to the generation of nonuniform roughness due to the stochastic formation of cracks and pits of various dimensions distributed over the entire surface of the material (Fig. 5). The erosion development from the point of view of emerging irregularities on the surface arises due to the deviation of grains from their equilibrium position. This was evident by the sixth second. The first recorded depth penetration was recorded after 7 s (Fig. 6b). To cover the whole experiment concerning erosion evolution under PWJ treatment, the times t=5 s, t=10 s, t=15 s, and t=20 s were selected for deeper analysis. These areas are highlighted in Fig. 6 for easy association with the following SEM analysis and induced subsurface deformation. 3.2. Surface observation results PWJ exposure times of 5, 10, 15, and 20 s were selected for inclusion in Fig. 7 to describe the entire experimental time interval, as compared with the continuous water jet exposition times of 10 and 20 s in Fig. 8. This comparison was made to visualise the effect of the hammer effect, which is generated by the pulsating water jet, and the stagnation pressure, which is created by the action of the continuous water jet [44]. Fig. 7 shows examples of surfaces after exposure times ranging from 5 s to 20 s. An overview of a treated surface after an exposure time of t=5 s is given in Fig. 7a. The treated surface is heavily deformed at the centre of the impacted area. In the region highlighted in detail (Fig. 7a1), there is a visible concentration of grain tilting, as was also observed in studies [20,45]. After 10 s (Fig. 7b), the area of the jet footprint shows significant signs of both plastic deformation and material removal and is similar to that presented in [46]. The centre area shows that microcavities were transformed or connected into longer cracks, which presumably propagated along the grain boundaries. With increasing distance from the centre of the jet footprint, the surface slowly changes to a morphology like that achieved after 5 s of the PWJ and then from a slightly deformed up to an undeformed surface. Due to the repeated action of water droplets with high frequency and low time intervals, it is possible to observe an area with a changed surface morphology after 5 s (Fig. 7a1). Surface depressions correspond to the profile of water droplets generated using ultrasonic PWJ technology [16]. The shape of the water pulses [47]consists of a core where most of the water droplets are concentrated and of a peripheral part that is deformed due to the air drag. In these water droplet complexes, there are satellite droplets connected to the central part by ligaments. If the depth of the crater is greater than the size of the droplet, the lateral flow interacts with the side wall and erodes it. The lateral jet speed is approximately 10 times higher and causes subsurface cavities. The subsurface cavities are more pronounced in the case of the high-frequency impingements of the concentrated action of water droplets on an epicentre of collision [48]. Fig. 7c shows a surface treated with an exposure time of t=15 s. The treated area shows a deep groove with significant material removal. The groove is rimmed with a sharp transition from the removed material to the almost unaffected material outside of the treated area. This applies especially to the upper part of the groove, which can be seen in detail in Fig. 7c1. The groove shows some level of asymmetry; mainly, the bottom part shows a significantly slower transition from the groove bottom to the region of the deformed surface. The surface generated after an exposure time of t=20 s (Fig. 7d) exhibits a further deepening of a groove. The pile-up and the transition region distribution around the groove seem to be the same. A sharp transition is visible around the S. Hloch et al. Engineering Science and Technology, an International Journal 50 (2024) 101630 8 whole groove (see Fig. 7d1), except for the bottom part of the groove—denoted by the blue area. Fig. 8 shows surfaces created by the CWJ after the exposure times of t=10 and 20 s. The CWJ uses the same hydraulic parameters as the PWJ. Fig. 8a shows the grain tilting after the exposure time of t=10 s. The detail (Fig. 8a1) also shows that, inside the preferably oriented grains, twinning starts to accommodate plastic deformation in the form of grain tilting. The twinning was further investigated by TEM analysis (Fig. 11). After a CWJ exposure time of t= 20 s (Fig. 8b), there was still no significant material removal observed. The detail (Fig. 8b1) shows that the rimmed grain boundaries appear Fig. 7. The SEM observations of areas impinged by PWJ water elements for erosion times of a) 5 s, b) 10 s, c) 15 s, d) 20 s. Fig. 8. The SEM observations of areas affected by CWJ with exposure times of a) 10 s and b) 20 s. Note that CWJ represents the control group samples. S. Hloch et al. Engineering Science and Technology, an International Journal 50 (2024) 101630 9 more apparent on the surface. This detail also shows the incubation stage of the cavities at some of the grain boundaries. The twins appear more apparent when the hammer effect of individual droplets prevails than after the CWJ with an exposure time of t=10 s. An interesting feature is the inhomogeneous transition (Fig. 8 a, b), which turns into a short and steep transition when the exposure time is increased (Fig. 8 a1, b1). 3.3. Subsurface observation results To thoroughly analyse the subsurface deformation, a FIB lamella was lifted out of the representative sample exposed to PWJ for t=5 s (Fig. 5e). The lamella was oriented transversally to the surface at the centre of the area (Fig. 5e), and its microstructure was compared with the lamella extracted from the as-received material. In the as-received material dislocation pile-ups at the grain boundaries were frequently found (Fig. 9a). Also, dislocations forming nets without any further arrangement were observed(Fig. 9b). The as-received material (Fig. 10a) contained a relatively high density (~10 13 m −2 ) of dislocations, gliding on multiple slip systems. In the hardened subsurface area observed in the TEM, the dislocations were distributed relatively homogeneously (Fig. 10b). (Fig. 10 provides two magnifications levels Fig. 10 a, b and Fig. 10c, d. Both micrographs show a clear increase in dislocation density following the PWJ treatment. Fig. 11a shows the surface of the lamella corresponding to the surface impacted by the PWJ.The microstructure below the impacted region differs from the as-received material mainly by a substantial increase in dislocation density. The high number of mechanical twins created due to the water impacts is also a characteristic microstructural feature of the impact zone. Due to the stress fields, the contrast of the micrographs is not good; however, two systems of micro-twins are visible in Fig. 11b. The diffractogram in the inset shows extra spots from the twins along two different {111} planes. A detailed view of the microstructure in the grain at about 2 μ m under the surface shows a homogeneous distribution of dislocations without any signs of the arrangement of the dislocations into walls or cells. The evaluation of dislocation density gives an estimation of 7 ×10 14 m −2 . The generally accepted theory of erosion development due to the action of water droplets divides the overall process into incubation, acceleration, maximum erosion rate, deceleration, and termination [27]. The classification is based on the graph of cumulative erosion over time [49]. These findings were interpreted based on experiments in which the experimental equipment used in the study did not allow the water droplets to concentrate in one specific place. Their distribution was more dispersed and less concentrated [20], which is not the case with the PWJ [24,38]. The concentrated cumulative effect of the PWJ in one place creates three different erosion manifestations. We consider erosion damage evolution as a wave that radiates from the centre to the outer areas [45]. As the exposure time increases [50], the degree of damage in the cross-section varies locally, i.e., areas with different erosion stages are created close to each other (see, e.g., Fig. 5g). This means that the additional induction locally moves the material to the area of advanced erosion, while the other part can remain in the incubation stage despite the same exposure time. This insight is especially important with regard to the stochastic distribution of water droplets with variable exposure time and parameters, such as viscosity, droplet size, speed, impact angle, and overlapping factor, and with regard to whether the surface is wetted or not [26]. Regarding the prevailing forces, the source of which is the combined effect of impact pressure and lateral flow, it is possible to deduce that, in the initial stages [45], the impact pressure is more dominant. 3.4. Nanohardness results The impact pressure generates changes on the surface and in the subsurface in the form of compressive stresses in the subsurface layers of the material [51]. These stresses were assessed using nanohardness measurements for erosion times of t=5 s, t=10 s, t=15 s, and t=20 s. Fig. 12 shows the deformation gradient map of the subsurface area after the PWJ treatment for various exposure times. There is an increase in hardness at a depth of 30 µm from the surface (i.e., –30 µm in the Y coordinate) for the PWJ exposure times of t=5 and 10 s. There are visible areas of local hardness increments under the eroded material (exposure times t=15 and 20 s). These local increments are situated in a heavily eroded area. The subsurface under PWJ treatment for 5 s (see Fig. 12) shows a noticeable hardening at –30 µm and a relatively homogenous area underneath represented by a blue colour. Similar, but more pronounced, results are shown in Fig. 12 for the area treated with the PWJ for 10 s. The layer from –30 µm to –80 µm shows a significant hardening; the area between –80 µm and –280 µm shows a visible softening; and, finally, the area from –350 µm and deeper exhibits further signs of hardening. A similar trend can be seen after the action of the jet with a exposure time of t=15 s (Fig. 12). The area after the exposure time of t=20 s presented in Fig. 12 shows a localised hardened layer close to the bottom and side of a groove. The hardness properties in the subsurface after the application of the PWJ for t=20 s are inhomogeneous. The local increase in hardness in the area on the right could have been caused by a local strain imposed during sample Fig. 9. TEM micrograph of as-received material showing details of a) dislocation pile-ups close to a grain boundary and b) net of closely spaced dislocations. S. Hloch et al.