Evaluation of Surface Integrity in 18CrNiMo7-6 Steel after Multiple Abrasive Waterjet Peening Process

: Abrasive waterjet peening (AWJP) as an important surface strengthening method can e ﬀ ectively improve surface properties. In this study, after multiple AWJP, the distribution of compressive residual stress and roughness on the surface of 18CrNiMo7-6 steel has been evaluated by an X-ray di ﬀ raction (XRD) method and a 3D surface topography system, respectively. Compared with the single AWJP, multiple AWJP can obviously increase the surface residual stresses ( − 1024 MPa to − 1455 MPa) and the depth of maximum compressive residual stress (100 µ m to 120 µ m), as well as make the stress distribution more uniform. In terms of the surface roughness, multiple AWJP inﬂuences its uniform distribution and reduces the surface roughness ( S a = 0.69 µ m), compared with a single AWJP ( S a = 2.96 µ m), due to the smaller shot balls and a uniform deformation during multiple AWJP. In addition, we have studied the e ﬀ ects of multiple AWJP on the hardness of the surface layer. The results show that multiple AWJP increases the hardness by up to 15.9%, compared to the single AWJP. These studies provide useful insight into improving the surface properties of 18CrNiMo7-6 steel by multiple AWJP.


Introduction
The study of mechanical properties of metals and alloys, including steel, and their resistance to external factors, is a fundamental matter for the development of multiple technological applications [1,2]. 18CrNiMo7-6 steel is widely used in the field of transportation, energy generation and general mechanical engineering, such as bearings, gears and wear pins, due to its three key characteristics: high wear resistance; good fatigue strength; and cost efficiency [3]. However, during the manufacturing processes such as cutting and grinding, tensile stresses always generate, which will deteriorate the fatigue life. In order to improve the fatigue life of industrial components, the residual stresses in the surface layer must be suppressed [4]. In addition, key components require hard surface layers and good surface quality. Surface deformation strengthening methods, such as shot peening (SP) [5,6], laser shot peening (LSP) [7,8], surface mechanical grinding treatment (SMGT) [9,10] and surface mechanical attrition treatment (SMAT) [11,12] can generate a beneficial microstructure and compressive residual stress to induce property enhancements. The surface deformation strengthening technology means to make the surface produce a certain degree of plastic deformation led to the forming of a certain depth of the deformation layer, which can improve the material properties [13].
Abrasive waterjet peening (AWJP) is a metal strengthening technology based on a basic principle like shot peening (SP), but it shoots balls along with high-pressure and high-speed water rather than air. During the process of AWJP, a large number of small balls with high kinetic energy impact on the metal surface, which causes a large elastic and plastic deformation [14,15]. The residual compressive

Materials
The studied samples are manufactured from case-hardened steel 18CrNiMo7-6 (EN 10084). The chemical composition of 18CrNiMo7-6 steel is summarized in Table 1. All the samples are carburized at 920 • C for 34 h in a controllable gas carburizing furnace, then subsequently quenched at 820 • C in oil to room temperature, then followed by tempering at 180 • C for 3 h, and finally cooling in air. Samples with a dimension of 20 × 20 × 20 mm are cut via the wire electrode cutting method, and ground to 800-grit SiC prior to being subjected to the AWJP process.

AWJP Processes
The AWJP experiments are carried out on a waterjet machine (OMAX corporation, Kent, WA, USA), which can generate water pressure from 5 MPa to 430 MPa. The experimental setup and schematic diagram of the AWJP process are shown in Figure 1. The nozzle used in this experiment is a round one with a diameter of 2.0 mm, and the moving speed and feed of the nozzle are 500 mm/min and 0.25 mm, respectively. The incidence angle and standoff distance between the nozzle and the samples are set at 90 • and 10 mm, respectively. The detailed parameters of AWJP progress are shown in Table 2. In this experiment, the abrasive media are two kinds of shot balls, cast steel balls with a hardness of HRC45-50 and ceramic balls with a hardness of HRC50-55. Media materials are commonly used in traditional shot peening. The detailed parameters of multiple AWJP are shown in Table 3. For Metals 2020, 10, 844 3 of 12 example, the sample 'S4 was first AWJP processed by cast steel balls (0.2 mm round) with a water pressure of 300 MPa, and then the second AWJP process was performed by ceramic balls (0.1 mm round) with a water pressure of 200 MPa. Since sample S4 is treated by AWJP twice, the peening time is twice as much as a single AWJP. In the AWJP process, a large amount of water, followed by many balls, beat on the surface of samples with high kinetic energy. Obviously, the kinetic energy was related to the parameters of the AWJP process, and the water pressure is the most important parameter, which can be found in our previous work [24]. After the AWJP process, we selected the center area with 10 × 10 mm 2 to measure the surface properties. Moreover, a corresponding sample was prepared for comparison.
Metals 2020, 10, x FOR PEER REVIEW 3 of 13 Table 3. For example, the sample 'S4′ was first AWJP processed by cast steel balls (0.2 mm round) with a water pressure of 300 MPa, and then the second AWJP process was performed by ceramic balls (0.1 mm round) with a water pressure of 200 MPa. Since sample S4 is treated by AWJP twice, the peening time is twice as much as a single AWJP. In the AWJP process, a large amount of water, followed by many balls, beat on the surface of samples with high kinetic energy. Obviously, the kinetic energy was related to the parameters of the AWJP process, and the water pressure is the most important parameter, which can be found in our previous work [24]. After the AWJP process, we selected the center area with 10 × 10 mm 2 to measure the surface properties. Moreover, a corresponding sample was prepared for comparison.

Residual Stresses and Microstructure Measurements
The XRD analysis was obtained via XRD equipment (Empyrean, PANalytical, Almelo, The Netherlands) using Cu-Ka radiation (λ = 1.54056 . A). The generator settings were 40 kV and 30 mA, the diffraction data were collected over a 2θ range of 30-90 • , with a step width of 0.02 • and a scan speed of 2 • per minute.
The magnitude of the induced residual compressive stress of the samples was measured by an X-ray residual stress analyzer (LXRD; PROTO, Oldcastle, ON, Canada) with Cr Kα radiation (Ni filter), and the XRD system was calibrated by a stress-free sample. The residual stress tests were conducted according to ASTM: E915-10. The voltage and current were 30 kV and 25 mA, respectively. The Cr Kα radiation with a wavelength λ = 2.2897 . A was used to determine the ferrite (211) diffraction peak. The residual stress was determined by the sin 2 ψ method [25]. An average of at least five residual compressive stresses data was recorded. The in-depth profiles of residual stresses were received by removing the surface layers one-by-one via chemical etching with the saturated salt solution.
Moreover, the hardness was measured by a Micro-hardness Tester (HV-1000, Beijing Times Mountain Peak Technology, Beijing, China) with a force of 1.96 N and a loading time of 15 s. In order to obtain the hardness distribution at different depths, the surface layers were removed one-by-one via chemical etching with the saturated salt solution. An average of at least five hardness data was recorded at every depth.
The surface state of the samples was measured by NPFLE 3D surface topography system (Bruker Nano Inc, Karlsruhe, Germany). A measurement area of about 1500 × 1500 µm 2 was covered. Light and dark colors indicated peak and valley regions, respectively. An average of at least five roughness values was used for evaluating the surface roughness (S a ). All the experiments were conducted at room temperature.
To reveal the subsurface microstructure, the cross-sections of the samples were metallographically ground and polished, followed by etching with 4% Nital. The microstructures of the samples were subsequently observed by scanning electron microscopy (Zeiss Auriga FIB-SEM, Jena, Germany).

Residual Stress Analyses
In Figure 2, the depth profiles of residual stress for all the surface treated samples are displayed. As shown, since the samples have been heat-treated before AWJP, compressive residual stress occurred in the near surface region of the corresponding sample S0. However, the maximum compressive stress (−267 MPa) and the depth of compressive region (40 µm) are limited. After single AWJP treatment, the maximum compressive stress of the S1 sample is −1280 MPa and the depth of the compressive region is 250 µm, which were introduced by the ceramic balls peening with a water pressure of 200 MPa. With the increasing of depth, the compressive residual stress (CRS) increases, firstly to the maximum at 65 µm, but then declines rapidly along the layer depths, and finally approaches zero. In comparison, if the single AWJP process was conducted by cast steel balls (0.2 mm round) with a pressure of 200 MPa, the maximum CRS and the depth of the compressive region of S2 sample are −1180 MPa and 760 µm, respectively. Since the steel balls are bigger and heavier, and are prone to cause higher kinetic energy than the ceramic balls, the influence depth is greater. However, the ceramic balls have high hardness and poor plasticity, so the maximum CRS is bigger than that caused by the steel balls.
If we increase the water pressure of the single AWJP process to 300 MPa by the same cast steel balls, the maximum CRS and the depth of compressive region of the S3 sample are −1191 MPa and 760 µm, respectively. This is an expected trend, since the higher water pressure makes the steel balls own the higher energy, and more serious plastic deformation is produced on the surface layer. We can find the similar results in SP treatment. For example, Mohamed et al. [26] observed that the applied air pressure has the largest effect on the shot peening process when compared to other process parameters, where high air pressure induces a thick and uniform compressed surface layer, compared to low pressure. by single AWJP. As can be seen in Figure 2, the surface CRS and maximum CRS of samples S4 and S5 increase compared to the sample S3, after the multiple AWJP processes. However, the depth of compressive region of samples S4 and S5 see almost no change. In the multiple AWJP process, the samples were fully peened at a higher intensity, and then re-peened in a second or third operation at a lower intensity. Therefore, the depth of the compressive region depends on the first step of the multiple AWJP, and it is almost same for samples S3, S4 and S5.   Table 4. As we all know, the standard deviation can reflect the uniform, the smaller standard deviation, the more uniform of the residual stress of distribution. It can be found that the sample S3, which was treated by the steel balls peening with a water pressure of 300 MPa, has the highest standard deviation (63.26), so its distribution of residual stress is the most nonuniform.
We use a ratio (R) introduced by Xie et al. [19] to define the uniformity of residual stress distribution, where R is the ratio of the average residual stress to the standard deviation. It can be found in Table 4 that the ratio of sample S1 is highest (50.41), which means that the single AWJP with ceramic balls produces the most uniform distribution of residual stress. If we process the sample with ceramic balls after the first or second step of the AWJP, the distribution of residual stress will be more uniform compared with the single AWJP. The experimental results of S4 and S5 have verified the validity of multiple AWJP in improving the uniformity of residual stress distribution. Through Kobayashi's research [27], shot peening creates an almost uniform compressive residual stress distribution along a surface after many small steel balls collided with the surface, due to the superposition of residual stress produced by surrounding shots. Therefore, the distribution of the As depicted in Figure 2, the dual AWJP induces a higher CRS than single AWJP in the upper layer of the sample S4. The maximum CRS increases from −1191 MPa to −1321 MPa, although the depth of the compressive region sees no change. This is to be expected, as since the ceramic balls are prone to cause near-surface deformation, the influence depth is limited. After the sample was treated by the first AWJP treatments with the pressure of 300 MPa, the depth of compressive region is 760 µm, so the ceramic balls with low kinetic energy struggle to cause large deformation on the thick hardened layer [19], and just increase the surface compressive stress.
In order to get a high value of compressive residual stress, sample S5 was processed by triple AWJP. The maximum CRS of the sample S5 increases from −1191 MPa to −1455 MPa, and the depth of compressive region of the sample S5 sees almost no change, compared with the sample S3 treated by single AWJP. As can be seen in Figure 2, the surface CRS and maximum CRS of samples S4 and S5 increase compared to the sample S3, after the multiple AWJP processes. However, the depth of compressive region of samples S4 and S5 see almost no change. In the multiple AWJP process, the samples were fully peened at a higher intensity, and then re-peened in a second or third operation at a lower intensity. Therefore, the depth of the compressive region depends on the first step of the multiple AWJP, and it is almost same for samples S3, S4 and S5. Figure 3 shows the distribution of residual stress on the surface of the sample S3, and the related parameters are displayed in Table 4. As we all know, the standard deviation can reflect the uniform, the smaller standard deviation, the more uniform of the residual stress of distribution. It can be found that the sample S3, which was treated by the steel balls peening with a water pressure of 300 MPa, has the highest standard deviation (63.26), so its distribution of residual stress is the most non-uniform.
We use a ratio (R) introduced by Xie et al. [19] to define the uniformity of residual stress distribution, where R is the ratio of the average residual stress to the standard deviation. It can be found in Table 4 that the ratio of sample S1 is highest (50.41), which means that the single AWJP with ceramic balls produces the most uniform distribution of residual stress. If we process the sample with ceramic balls after the first or second step of the AWJP, the distribution of residual stress will be more uniform compared with the single AWJP. The experimental results of S4 and S5 have verified the validity of multiple AWJP in improving the uniformity of residual stress distribution. Through Kobayashi's research [27], shot peening creates an almost uniform compressive residual stress distribution along a surface after many small steel balls collided with the surface, due to the superposition of residual stress produced by surrounding shots. Therefore, the distribution of the residual stress of the sample treated by multiple AWJP is more uniform than that treated by a single AWJP. Moreover, the uniformity of stress distribution is very important for the component's fatigue resistance.
Metals 2020, 10, x FOR PEER REVIEW 6 of 13 residual stress of the sample treated by multiple AWJP is more uniform than that treated by a single AWJP. Moreover, the uniformity of stress distribution is very important for the component's fatigue resistance.   Figure 4 shows the XRD patterns of the top surface of 18CrNiMo7-6 steel before and after AWJP treatments. As depicted in Figure 4, both martensite (α) and austenite (γ) phases are found in the top surface of 18CrNiMo7-6 steel before AWJP treatment. A small amount of austenite is still retained after sample S2 is treated by AWJP treatments by cast steel balls (0.2 mm round) with a pressure of 200 MPa. However, only martensite state is tested after triple AWJP treatment, implying that the retained austenite has transformed to the martensite state by multiple AWJP treatments. AWJP treatments can improve the mechanical properties of the surface of 18CrNiMo7-6 steel, as the retained austenite contents decrease near the surface region. These similar results could be found in SP treatments [5,21].   Figure 4 shows the XRD patterns of the top surface of 18CrNiMo7-6 steel before and after AWJP treatments. As depicted in Figure 4, both martensite (α) and austenite (γ) phases are found in the top surface of 18CrNiMo7-6 steel before AWJP treatment. A small amount of austenite is still retained after sample S2 is treated by AWJP treatments by cast steel balls (0.2 mm round) with a pressure of 200 MPa. However, only martensite state is tested after triple AWJP treatment, implying that the retained austenite has transformed to the martensite state by multiple AWJP treatments. AWJP treatments can improve the mechanical properties of the surface of 18CrNiMo7-6 steel, as the retained austenite contents decrease near the surface region. These similar results could be found in SP treatments [5,21]. Figure 5 shows the 3D surface topographies of samples treated under different AWJP processes, and the values of S a are displayed in Table 5. S a is a commonly used parameter for the evaluation of measured surface roughness, which denotes the arithmetical mean height. Figure 5a is a polished sample's surface morphology, the surface has the texture left by the polishing process, and the sample S0 has a S a value of 0.19 µm. The value of S a increased to 1.19 µm after the S1 samples were AWJP processed by ceramic balls with a pressure of 200 MPa, as shown in Figure 5b. In this process of AWJP, a great amount of small ceramic balls impacted the surface with high kinetic energy. The humps and horns on the surface are clearly visible, which means that the AWJP treatments by ceramic balls induce a large plastic deformation, and at the same time remove a part of the surface metal.  Figure 5 shows the 3D surface topographies of samples treated under different AWJP processes, and the values of Sa are displayed in Table 5. Sa is a commonly used parameter for the evaluation of measured surface roughness, which denotes the arithmetical mean height. Figure 5a is a polished sample's surface morphology, the surface has the texture left by the polishing process, and the sample S0 has a Sa value of 0.19 μm. The value of Sa increased to 1.19 μm after the S1 samples were AWJP processed by ceramic balls with a pressure of 200 MPa, as shown in Figure 5b. In this process of AWJP, a great amount of small ceramic balls impacted the surface with high kinetic energy. The humps and horns on the surface are clearly visible, which means that the AWJP treatments by ceramic balls induce a large plastic deformation, and at the same time remove a part of the surface metal.

Surface Morphology and Roughness
In contrast, if the AWJP process is conducted by cast steel balls (0.2 mm round) with a pressure of 200 MPa, the surface roughness of the S2 sample is increased to 0.66 μm. This is because the use of larger and heavier cast steel balls can produce a higher level of plastic deformation, which will produce an arithmetic mean distributed texture, as shown in Figure 5c.
In Figure 5d, a lot of pits remain on the surface of the sample S3 treated by AWJP. The pits are due to the small indentations, and will lead to stress concentrations on the surface of the part, so we should eliminate the pits on the surface. The surface roughness of the S3 sample is increased to 2.96 μm, which is approximated to be 383% more than that of the S2 sample. Therefore, the state of sample surface is apt to deteriorate, with the increase in the water pressure from 200 MPa to 300 MPa. In order to decrease the surface roughness, the multiple AWJP processes were carried out on the surface of the samples S4 and S5.
As can be seen in Figures 5e,f, the surface roughness of sample S4 and S5 decrease to 0.69 and 1.64 μm, compared with sample S3 after the multiple AWJP processes. The humps on the surface of the sample are removed, and therefore the surface quality is improved after multiple AWJP.    In contrast, if the AWJP process is conducted by cast steel balls (0.2 mm round) with a pressure of 200 MPa, the surface roughness of the S2 sample is increased to 0.66 µm. This is because the use of larger and heavier cast steel balls can produce a higher level of plastic deformation, which will produce an arithmetic mean distributed texture, as shown in Figure 5c.
In Figure 5d, a lot of pits remain on the surface of the sample S3 treated by AWJP. The pits are due to the small indentations, and will lead to stress concentrations on the surface of the part, so we should eliminate the pits on the surface. The surface roughness of the S3 sample is increased to 2.96 µm, which is approximated to be 383% more than that of the S2 sample. Therefore, the state of sample surface is apt to deteriorate, with the increase in the water pressure from 200 MPa to 300 MPa. In order to decrease the surface roughness, the multiple AWJP processes were carried out on the surface of the samples S4 and S5.
As can be seen in Figure 5e,f, the surface roughness of sample S4 and S5 decrease to 0.69 and 1.64 µm, compared with sample S3 after the multiple AWJP processes. The humps on the surface of the sample are removed, and therefore the surface quality is improved after multiple AWJP. Figure 6 shows the hardness profiles of the samples treated under different AWJP processes. For sample S0, the maximum hardness appears on the surface, and then decreases with increasing depth and reaches a constant. However, after the samples are treated by the AWJP, all the samples are found to have a similar trend of hardness; the top surface hardness first increases to a maximum value, then decreases gradually along the depth of the layer, and finally remains constant. The sample S1 is treated by the ceramic balls peening with a water pressure of 200 MPa, and the thickness of its hardened layer is small compared to the others processed by the steel balls. This is to be expected, since because the ceramic balls have a lower impact kinetic energy than steel balls, the depth of the surface work hardening layer is small. The diversification trends of hardness are the same as those of residual stress, which is due to the relevance between the two factors. It was found that the hardness decreases with tensile residual stress, and increases with compressive residual stress [28].  Figure 6 shows the hardness profiles of the samples treated under different AWJP processes. For sample S0, the maximum hardness appears on the surface, and then decreases with increasing depth and reaches a constant. However, after the samples are treated by the AWJP, all the samples are found to have a similar trend of hardness; the top surface hardness first increases to a maximum value, then decreases gradually along the depth of the layer, and finally remains constant. The sample S1 is treated by the ceramic balls peening with a water pressure of 200 MPa, and the thickness of its hardened layer is small compared to the others processed by the steel balls. This is to be expected, since because the ceramic balls have a lower impact kinetic energy than steel balls, the depth of the surface work hardening layer is small. The diversification trends of hardness are the same as those of residual stress, which is due to the relevance between the two factors. It was found that the hardness decreases with tensile residual stress, and increases with compressive residual stress [28].

Hardness
The average maximum hardness of the sample S5 increases up to a maximum of 1145 HV0.3 at 80 μm. The results show that multiple AWJP increase the maximum hardness of sample S5 up by 15.9%, compared to the maximum hardness of sample S3 (988 HV0.3) treated by a single AWJP. Though the maximum hardness of samples treated by steel balls are different, the depth of hardened layers is almost the same. It is evident that the depth of the hardening layer depends on the first AWJP, rather than the following steps of the AWJP. The experimental results demonstrate that multiple AWJP can produce a larger work hardening layer in the 18CrNiMo7-6 steel than a single AWJP.

Microstructure
In Figure 7, the cross-sectional microstructures of sample S2 treated by AWJP are shown. Firstly, the cross-sectional microstructure of the sample S2 displays a typical carburized steel. The casehardened zone consists of an almost martensite microstructure with a very small quantity of retained austenite and finely dispersed undissolved spherical carbide particles [5]. Secondly, it can be found that the plastic deformation layer is built on the surface of sample S2. Comparing Figure 7b with Figure 7c, the plastic deformation-induced grain refinement occurs during the AWJP treatment. The AWJP process projects balls along with high-pressure and high-speed water by the action of an external stress. When the external stress exceeds the critical stress for forming the deformation twins and the gain deformation, the grain refinement will be induced by plastic deformation [29]. The average maximum hardness of the sample S5 increases up to a maximum of 1145 HV 0.3 at 80 µm. The results show that multiple AWJP increase the maximum hardness of sample S5 up by 15.9%, compared to the maximum hardness of sample S3 (988 HV 0.3 ) treated by a single AWJP. Though the maximum hardness of samples treated by steel balls are different, the depth of hardened layers is almost the same. It is evident that the depth of the hardening layer depends on the first AWJP, rather than the following steps of the AWJP. The experimental results demonstrate that multiple AWJP can produce a larger work hardening layer in the 18CrNiMo7-6 steel than a single AWJP.

Microstructure
In Figure 7, the cross-sectional microstructures of sample S2 treated by AWJP are shown. Firstly, the cross-sectional microstructure of the sample S2 displays a typical carburized steel. The case-hardened zone consists of an almost martensite microstructure with a very small quantity of retained austenite and finely dispersed undissolved spherical carbide particles [5]. Secondly, it can be found that the plastic deformation layer is built on the surface of sample S2. Comparing Figure 7b with Figure 7c, the plastic deformation-induced grain refinement occurs during the AWJP treatment. The AWJP process projects balls along with high-pressure and high-speed water by the action of an external stress. When the external stress exceeds the critical stress for forming the deformation twins and the gain deformation, the grain refinement will be induced by plastic deformation [29].

Conclusions
In summary, after multiple AWJP treatments were implemented to 18CrNiMo7-6 steel, the uniformity of the residual stress distribution was studied by XRD. The results showed that the specimens after triple AWJP treatments had a high compressive residual stress of −1455 MPa and a small standard deviation of 22.05 MPa. It is indicated that the multiple AWJP treatments can improve the uniformity of distribution and the values of residual stress. In addition, the multiple AWJP treatments reduced the surface roughness following the single AWJP. The hardness was improved to 1145 HV0.3, which has the same trend with the residual stress. The AWJP by steel balls produced a thicker hardening layer, with higher hardness than the AWJP by ceramic balls. These results confirm that multiple AWJP is a more effective surface treatment than single AWJP. These results can provide a basic experimental foundation for improving the fatigue strength and wear resistance of 18CrNiMo7-6 steel.

Conclusions
In summary, after multiple AWJP treatments were implemented to 18CrNiMo7-6 steel, the uniformity of the residual stress distribution was studied by XRD. The results showed that the specimens after triple AWJP treatments had a high compressive residual stress of −1455 MPa and a small standard deviation of 22.05 MPa. It is indicated that the multiple AWJP treatments can improve the uniformity of distribution and the values of residual stress. In addition, the multiple AWJP treatments reduced the surface roughness following the single AWJP. The hardness was improved to 1145 HV 0.3 , which has the same trend with the residual stress. The AWJP by steel balls produced a thicker hardening layer, with higher hardness than the AWJP by ceramic balls. These results confirm that multiple AWJP is a more effective surface treatment than single AWJP. These results can provide a basic experimental foundation for improving the fatigue strength and wear resistance of 18CrNiMo7-6 steel.

Conflicts of Interest:
The authors declare no conflict of interest.