Mechanical components are usually exposed to fluctuating stresses that produce nucleation of microfractures, which generate catastrophic failures when propagated over time [1]. This failure mode is called fatigue. There are three methods for estimating fatigue failure: stress-life, or S-N, which is responsible for predicting life for more than 10,000 cycles [1]; deformation life, ε-N, or low cycle fatigue, for up to around 10,000 cycles [20] ; and the da/dN method that is based on fracture mechanics and aims to predict service life in the presence of cracks [21], [22]. Moreover, it can be stated that the vast majority of components are designed with the S-N method [1], i.e. leaf springs are designed for lives of at least 100,000 cycles [3], [16].

This cold process consists of impacting a surface with high-speed particles (between 20 and 120 m/s) to increase its hardness [3] to generate superficial plastic deformations and residual compression stresses, affecting only a thin layer. (0.1 to 0.5mm) [23]. In automotive components, i.e., leaf springs subjected to bending stresses, a beneficial effect is generated in the area subjected to positive stresses [4]. This treatment is commonly applied to engineering components that have complex geometry [1], [2] or are subjected to highly localized fluctuating stresses that can suffer from fretting [4], [23]. Shot peening is fast and suitable for mass production [1], [2].

On the other hand, shot peening can negatively impact the behavior of metals in corrosive environments since it decreases the potential for degradation of the passive layer [1]. Additionally, non-significant changes in AISI/SAE 5160 steel fatigue behavior have been reported when alternate bending tests were performed at three-point bending [4] but significant changes in low-carbon steels were reported under aggressive environments [24].

Furthermore, the Almen intensity measures the energy transmitted by the impact to the processed component [2], [24]. This scale measures the deflection of the arc formed in a normalized SAE 1070 steel strip when the pellets impact only one face. The procedure for this measurement is standardized by the SAE 442 [25] and SAE J443 [5] standards. The parameters of the shot peening process are material and diameter of the pellets, angle of the jet, distance from the jet to the surface, air pressure, exposure time, and shot flow [3], [23]. The surface appearance of a sample subjected to SP can be seen in Figure 1. Despite having a roughness that could favor crack initiation, the compressive stress induced by deformation generates beneficial behavior under an alternating load.

Figure 1 Source: [24].

One can place a standard metal strip next to the to-be-treated component to quantify the shot peening. The strip forms an arch, and its height, H, is measured. By plotting the deflection H of the arc formed against exposure time, a curve is generated, described by a parabola that opens towards the time axis. Therefore, increasing the exposure does not generate a more significant change in the curvature or residual stresses. Ideal shot peening is achieved when doubling the shot blast exposure time t results in only an increase in arc deflection, zlH, which does not exceed 10%. This is known as the saturation curve [25] and can be seen in Figure 2.

Figure 2 Source. Adapted from [25].

Residual stresses are internal stresses that remain in a solid material after the action that originated them has ceased [26]. These stresses can occur due to a variety of mechanisms, including cyclic plastic deformations [21], [22], elevated temperature gradients [20] or microstructural changes (i.e., phase transformation) [1]. It is a common but generally undesirable occurrence.

In a sample subjected to bending, the stress changes linearly with distance from the neutral axis. The imposed residual stress leaves compression closer to the surface but creates a tension zone to balance forces. The net stress is the superposition of applied and residual stress represented by the red dashed line in Figure 3.

Figure 3 Source: authors.

X-ray diffraction (XRD) is based on Bragg's Law to determine the diffusion between atoms in crystalline materials, which allows such separation to act as diffraction gratings and lead to the generation of patterns that can be characterized [3]. The procedure for establishing residual stresses by XRD is described in ASTM E2860 [6]. The XRD technique allows measurements of variations in the parameter of the crystal lattice to calculate residual stresses.

On the other hand, the blind hole technique, standardized by ASTM E837 [27], is based on the principle that stresses on free surfaces are zero. A strain gauge placed on a surface measures strain (exx, eyy, exy), and by making incremental perforations right in the center of the rosette that is formed due to the location of the gauges, strains away from the surface can be measured to a depth that can reach up to 15 mm [8].

Recently, Aguado et al. [14] presented a model based on the sinusoidal decay function that described the residual stress profile to minimize fatigue crack growth in a component. It is shown in.

where A, λ, θ, and ω are constants to calibrate for each component, and z is the coordinate from the treated surface to the center of the component. An example of two theoretical residual stress profiles, called a¡ and 02, obtained from equation (1) and using arbitrary constants is shown in Figure 4, in which the vertical axis is normalized residual stress. Such profiles exhibit a similar shape of what Hernandez et al. [3] measured by XRD and electropolishing. Recently, [7] reported similar residual stresses profiles for a heat-treated 100Cr6 steel subjected to different cutting speeds as well for a 15-5PH [28] subjected to compressive stress by roller burnishing.

Figure 4 Source: authors.

The variation of the potential and current density of the anodic and cathodic curves should be linear at the Tafel slopes, intersecting at the point E_corr/i_corr. However, there are linear deviations in the vicinity of Ecorr. These curves have regions with a linear ratio between the potential and the logarithm of the current density, an area called the Tafel region [18]. Figure 5 shows a portion of the curves obtained by applying a different potential to Ecorr. It can be seen that the closer one gets to the value of E to that of Ecorr, the value of w tends to zero.

Figure 5 Source: authors.

The microstructure for steel SAE5160H used in this study is shown in Figure 8a at 1000x and in Figure 8b at 2000x. A microstructure corresponding to tempered, thin, and acicular martensite is observed, corresponding to high hardness, while carbide precipitates can be seen in the dark regions.

Figure 8 Source: authors.

The properties conferred to the AISI 5160H alloy are essential for the required applications in the automotive industry, allowing for the delay of fatigue crack nucleation that tends to occur on the surface.

Carbides are formed due to the segregation of elements such as chromium, vanadium, or molybdenum during the cooling and solidification process. The presence of carbides can have a significant impact on the mechanical properties of the steel, as they can contribute to strengthening and improving abrasion resistance.

Measurements were made at six points for two As-it-is samples and one SP sample, with 9 measurements at each point. A schematic of the location of the measured points is shown in Figure 9, where the roughness of a surface modified by SP can be appreciated. The samples were 10 mm wide x 15 mm long and 1/4" thick. The residual stress value at each point results from sweeping 9 measurements at different angles. Then the residual stress result delivered at each point is the average of the 9 measurements made at different angles. Each measurement point is swept by X-rays at angles between -30° and +30° measured from a line perpendicular to the specimen. The goniometer makes 9 measurements: 4 before the zero angle, 1 at zero, and another 4 after the zero angle. The reliability of the measurement is above 80%. A positive stress means traction, whereas a negative value is compression.

Figure 9 Source: authors.

Figure 10 shows the absolute value of the surface residual stresses measured by the XRD technique. The sample with SP residual stresses presents residual compression stresses of average value -365.8 ± 78 MPa while the sample As-it-is presents average positive residual stresses of 54.2 ± 54.3 MPa. An increase in average compressive residual stress of 420 MPa on the As-it-is samples can be observed. Moreover, when comparing the residual surface stress with the flexural strength (2170 MPa), an increase of 16.8% is obtained, which is within the values accepted by the industry for the process [3], [16]. However, the XRD technique delivers results of residual surface stress. Therefore, the depth of this change and how these stresses vary are unknown.

Figure 10 Source: authors.

Figure 11 shows the Vickers microhardness results for a shot-peened sample taken at three measured depths (0.07, 0.17, and 0.27 mm) from the processed surface. HV values were estimated according to equation (2).

Figure 11 Source: authors.

From the microhardness results, the average and standard deviation for each Z depth was calculated. Table 2 shows the average values of Vickers microhardness and standard deviation for samples with shot peening at the 4 distances measured. It is observed that in the SP sample, the HVN varies between 525 on the surface, reaching a minimum at 374 MPa at 0.07 mm and then recovering to 390 at 0.17 and 418 at 0.27 mm. [11]; The greatest increase occurs not at the surface but at a depth of about 0.3 mm from the peened surface.

From the microhardness results, the average and standard deviation for each Z depth was calculated. Table 2 shows the average values of Vickers microhardness and standard deviation for samples with shot peening at the 4 distances measured. It is observed that in the SP sample, the HVN varies between 525 on the surface, reaching a minimum at 374 MPa at 0.07 mm and then recovering to 390 at 0.17 and 418 at 0.27 mm [11]; The greatest increase occurs not at the surface but at a depth of about 0.3 mm from the peened surface.

From the microhardness results, the average and standard deviation for each Z depth was calculated. Table 2 shows the average values of Vickers microhardness and standard deviation for samples with shot peening at the 4 distances measured. It is observed that in the SP sample, the HVN varies between 525 on the surface, reaching a minimum at 374 MPa at 0.07 mm and then recovering to 390 at 0.17 and 418 at 0.27 mm [11]; The greatest increase occurs not at the surface but at a depth of about 0.3 mm from the peened surface.

Table 2

HVN	SD	Z [mm]
525.6	92.7	0
374.1	36.9	0.07
390.5	34.5	0.17
418.2	25.8	0.27

On the other hand, Table 3 shows the average values of Vickers microhardness and standard deviation for samples without shot peening at the 4 distances measured. It is observed that the HVN varies between 433 and 344. Despite showing a couple of extreme points in both samples, the SD is between 39.5 and 48.2 HV (9.1 to 14%) for As-it-is and between 92.7 and 25.8 HVN (17.6 to 6.2%) for the SP sample.

Although the literature cites SP as a heterogeneous process, we see a variation in the effect of the process from the dispersion of data.

Table 3

HVN	SD	Z [mm]
433	39.5	0
344.1	48.2	0.16
370.1	47.1	0.26
352.7	40.6	0.36

Source: authors.

Table 3, SP sample, shows more homogeneous values of Vickers micro-hardness compared to the values observed in Table 2, As-it-is sample, which is also reflected in the SD of the measurements. This may be related to the compression effect of the surface when treated with SP, confirmed by the results in Figure 10, where the sample subjected to the SP process showed residual compression stresses. The drop in some Vickers microhardness values observed in the As-it-is sample of Figure 12 may be a consequence of the effects of the microstructure on the measured points since areas of retained austenite are observed that are typically softer than the martensite that is also present in the microstructure, see Figure 8.

Figure 12 Source: authors.

Moreover, Figure 13 shows an example of an indentation imprint. The photograph is taken at 40x with a field view of 61.70 x 176.8 um. The imprint's symmetric shape is appreciated, noting that it is much larger than the size of the microstructural features shown in Figure 8.

Figure 13 Source: authors.

Hence, the measured microhardness is not affected by the local microstructure.

On the other hand, the variability of the diagonal measurement for each measured depth is shown in Table 4. For the As-it-is sample, the SD between diagonals ranges between 0.48 and 0.71% of the average hardness of each diagonal, while in the SP sample, the SD between diagonals ranges from 0.21 to 1.06% of the average hardness of each diagonal.

To establish whether there is a significant difference between the hardness measurements taken at different depths, see the nomenclature in Figure 7, a null hypothesis (H0: the hardness averages are equal) and an alternative hypothesis (H1: there are significant differences in the hardness averages) were proposed for each sample. The analysis of variance (ANOVA) showed that the null hypothesis (H0) is accepted with a 96.89% probability for samples without shot peening (As). On the other hand, the ANOVA indicated that the null hypothesis H0) should be rejected for SP samples, giving a probability of 83.95%. This indicates that the hardness does not change with the depth in the As-it-is samples, unlike in the SP samples. One might think that the lateral restriction imposed by the thickness influences how the hardness measurement might be affected. However, according to the ANOVA, that does not occur. That effect might be attributed to the shallow imprint not being affected by thick or thin surroundings.

Finally, Figure 14 shows results for electrochemical tests. For the As-it-is samples the potential is -620 mV, while for SP is -750 mV. Therefore, from the thermodynamics point of view, the SAE 5160H steel with SP has a higher corrosion rate because the surface treatment performs a plastic deformation on the surface of the material, generating empty spaces which are not impacted the pellets. These empty spaces act as interstices, in which agents that generate oxides lodge more easily. Thus, increasing the progressive degradation of steel by corrosive factors.

Table 4

Z [mm]	SD d1	SD d2
As
0.16	2.45	2.31
0.26	2.33	2.23
0.36	1.82	1.69
SP
0.07	2.67	3.98
0.17	2.07	1.13
0.27	1.09	0.89

Source: authors.

Figure 14 Source: authors.