1 Introduction

There are different techniques that aim to improve contact conditions by modification of surface properties. Just to name a few, (i) surface coatings [1], (ii) surface texturing [2], (iii) ion implantation [3], (iv) plasma treatment [4] and (v) laser surface modification [5] are all successfully investigated techniques.

In this scenario, surface topography plays an important role on the tribological performance of two-rubbing surfaces both in dry and lubricated conditions.

There are several applications in the industrial sector for which the presence of micro dimples can improve contact conditions and consequently reduce friction. With reference to the cutting processes, for example, the presence of microstructured surfaces can be useful in both dry and lubricated conditions. In particular, as stated by Lu and Wood in dry conditions, they promote the reduction of the contact area between the tool and chip, as well as trapping wear debris [2].

With reference to the current industrial scenario, the increasingly stringent requirements in terms of reducing environmental impacts is driving the automotive industry to adopt intensively light alloys (like aluminium, magnesium and titanium alloys) and advanced strength steels (AHSS). However, according to the results discussed by Cruz et al. regarding the contact pressure in sheet metal forming of AHSS, increasing strength and high surface hardness can affect the typical contact pressures between the tool and the material sheet itself from which the final component will be produced; this results in tool wear with an impact on the final quality of the part [6].

In the sheet metal forming processes, the friction conditions between the die and the sheet metal are crucial. In fact, the formability of the sheet metal, as well as the quality of the final component, are strictly related to the conditions at the sheet-die interface. More specifically, the COF, if properly managed, could guarantee those stresses necessary to deform the sheet until the final shape required. On the other hand, if not properly managed, it may cause early rupture of the sheet. Therefore, friction plays a major role in metal forming processes. Surface texturing could be considered a viable way to improve all these aspects; this is especially true when medium–low loads are present throughout the entire stamping process, as stated by Xu et al. [7]. As suggested by several numerical/experimental studies, during stamping processes, depending on the shape complexity and the material adopted, the contact pressure varies up to just over 1 GPa. In this regard, Kilpi et al. studied the behaviour of tool steel-aluminium alloy tribopair at different temperatures and with contact pressures equal to 0.51 and 0.81 GPa; they found that friction coefficient as well as wear were strongly influenced by both contact pressure and temperature [8].

Cruz et al. studied the wear of a die for sheet metal forming processes, finding that the geometry of the die has a strong influence on contact pressures and, consequently, on wear. In this study, the typical contact pressures recorded varied between 0.2 and 0.8 GPa (with an average of about 0.5 GPa) and the most critical areas were generally characterized by reduced radii or complex geometric shapes, whose appropriate modifications could significantly improve wear behaviour [6].

In addition to the previous aspects, in the manufacturing sector, one of the most stringent demands is the need of reducing the pollutants produced during the process itself; it is clear, therefore, that the possibility of reducing, if not eliminating, the adoption of lubricants allows to shorten the production route by eliminating the need of washing the part before the following steps (like welding, for example).

The surface characteristics of a metal sheet affect its press-forming behaviour. In fact, many metal materials used in the fabrication of sheet metal components (especially aluminium alloys) are characterized by a pronounced tendency to adhere to typical tool materials, thus affecting the friction and, in turn, the magnitude of the forming forces. Therefore, the adoption of suitable technologies able to improve the tool-sheet contact conditions is of great interest. An interesting approach might be to specifically texturize the most critical contact areas between the sheet and the tool (the texture might be created, in some areas, during the rolling process, for example).

On the other hand, to improve tribological conditions during the deformation process, Sulaiman et al. demonstrated the capability of the application of a laser-obtained textured surface on steel tools intended for sheet metal forming; moreover, this effectiveness also turns out to be a function of the geometric characteristics (i.e. groove width and the distance between grooves) of the texture itself [9]. In addition, the texturing technique has already been successfully studied in the rolling process; as documented by Song et al., a microtextured surface allows to halve the friction between the contact surfaces between the sheet and rollers [10].

In light of the foregoing, the role of textured surfaces in the industrial field for even dissimilar applications is critical for improving contact conditions; therefore, the study and improvement of methodologies aimed at surface modification of parts is mandatory. This is even more true when dry conditions are present in any technological process. In fact, as obtained by Maruda et al., metal-to-metal contact in dry conditions, even with high loads applied during turning operations, is very severe: surfaces machined under such conditions are affected by both seizing and adhesion phenomena [11]. In another work, Skalante et al. evaluated friction conditions under extreme conditions (cryogenic dryness) by applying pressures and sliding velocities closed to those adopted during the machining of a Ti6Al4V-ELI Grade 5 alloy. A decrease of between 5 and 15% was achieved depending on the contact pressure applied (1 to 2.5 GPa). Therefore, the reduction is less pronounced the smaller the applied loads are [12].

As reported by Etsion, surface texturing has emerged as a reliable method to reduce wear and friction over the past few decades [13]. It entails creating a pattern of dimples or grooves on a material surface in a proper and controllable way, changing the surface topography. The earliest evidence for such an approach, in any case, is much older. In fact, Hamilton et al. introduced the concept of “microirregularities” for the first time in 1966 [14]. They discovered that microirregularities on the surface of lubricated rotary shaft could create hydrodynamic pressure, which in turn improves the load-carrying capability.

Over the years, the attractiveness of these methodologies has meant that different techniques could be developed or simply adapted for obtaining suitably modified surfaces with the aim of improving their tribological conditions. According to this, to obtain the desired surface texture, several techniques, as documented in the available literature, have been employed such as (i) laser surface texturing (LST) [15], (ii) ion etching [16], (iii) vibro-rolling [17], (iv) electrochemical etching [18], and (v) mechanical texturing [2]. Each of these dimple texturing methods has its own unique characteristics; however, the development of effective and high-quality dimple texturing techniques is still an open problem. Among the available techniques, LST has been widely used in different researchers. However, the high reflection of metals like copper and aluminium during laser processing [2], the creation of a heat-affected zone [19], the surface oxidation due to high temperatures [20], and the high costs of the equipment [21] are the main problems limiting the massive application of the LST.

It is clear that the differentiation of the dissimilar techniques that can be adopted for the purpose makes it so that they are equally dissimilar, as well as numerous, causes of a variation in the tribological properties of the modified surfaces. The surface topography of tribological interfaces significantly affects various energy-dissipating and wear phenomena such as those related to adhesion that take place during sliding. The adhesion between moving surfaces can be tuned by surface texturing. In general, the higher COF observed for smooth surfaces is a consequence of the increased plastic flow stress which, in turns, increases the real contact area and adhesion. As reported by Mao et al., for a patterned surface, the COF emerges because of multiple mechanisms, such as the resistance force originating from the indentation edge, the interlocking of asperities at the interfaces of the tribological pair and alterations in surface strength [22].

Previous investigations performed by Bai et al. showed that the surface texture in dry condition plays an important role in capturing debris thus preventing further severe wear [23]. In this regard, Meng et al. improved the tribological performance of AISI 1045 hardened steel surfaces by ultrasonic surface rolling technology under dry condition [24]. In addition, Fang et al. investigated the surface modification effects of different laser shock peening (LSP) treatments on the fretting wear properties of the GH4169 alloy. They showed that LSP technologies can avoid the appearance of adhesive zones and have a decisive effect on material wear due to (i) the formation of a hardened layer, (ii) a high amount of debris produced by the process having a lubricating effect and (iii) the increase in the material’s tangential stiffness [25]. The influence of surface microhardness on tribological properties was also demonstrated by Yuan et al., who investigated the surface morphology of specimens processed with vibration-assisted reinforcement technique by varying both time and static load. The treated surfaces showed an increase in surface microhardness, resulting in improved wear resistance [26].

There is no doubt that texture geometries such as grooves or indentations as well as the size, depth, shape and density of the texture have a strong influence on the lifetime of the final component. Sun et al. investigated the dry wear characteristics of different microtextures produced by laser micro machining on TC11 alloy surfaces at 500 ºC. They showed that the microdimples with relatively small diameters and high densities may improve TC11 alloy’s tribological properties [27]. In another study, Conrady et al. modified the Ti alloy surfaces using a nanosecond laser technique with different textures including lines, crosshatching and dimples. The lowest COF was obtained for the dimple shapes in dry conditions [28]. Maldonado-Cortes et al. evaluated the tribological effect of different textures obtained through laser surface texturing, observing significant differences between texturized and unprocessed samples. As for the type of geometry, a dependence of the behaviour has been observed with respect to discrete or “closed” structured geometries and continuous or “open” geometries. In addition, at low contact pressures, all the structured geometries allowed reductions of the COF with better performance regarding wear for “closed” geometries such as circle, square and triangle [29].

Gimeno et al. evaluated the behaviour of different texture patterns (varying dimple diameter, height and density) on flat 100Cr6 steel specimens and bearings in final service. The textures were produced by ultra-short pulse laser process on the inner ring raceway in tapered roller bearings. The study showed that the geometrical aspects of the texture inferred with the final tribological properties with an improvement in high-speed friction tests of about 18% respect to the untextured surfaces. Among all, the effect of dimple height, which should be reduced as much as possible, emerged strongly [30]. Zhan et al. investigated the dry tribological properties of 40Cr steel specimens with six different types of textures: dimple, grooves and sinusoidal shapes, including single and multi-shaped textures generated by a nanosecond laser [31]. Various textural morphologies were found to have different tribological properties. In the studies conducted by Hu et al. [32] and Kumari et al. [33], the variation of the dimple density was shown to be very important for the reduction of the abrasive wear due to the storage of the wear particles in the dimples.

Focusing on the effect of the void ratio (VR), some works have reported a detrimental effect in terms of COF when using high VR values. Keeping VR low means to create a small number of dimples with long distance among them, whereas high dimple density leads to the reduction of this distance, thus resulting a decrease of the real contact area between the surfaces in different was observed due to the large transverse force. In fact, as reported by Li et al., the surface with too high texture density requires great transverse force for the relative sliding movement and consequently causes high friction coefficient (because the reduction in contact area can lead to a concentration of force on specific points). Moreover, too high texture density indicates limited contact area and great contact pressure, which therefore leads to large deformation and small gap between contact surfaces [34]. On the contrary, in some studies, a significant positive effect of high density on the wear behaviour was observed compared to the low-density case. In general, the ideal density of dimples is highly dependent on the analyzed material, as widely discussed by Wang et al.; in particular, it depends on the stiffness of the sliding material. [35].

In the case of non-conforming contact under dry conditions, as demonstrated by Joshi et al., it was shown that the adoption of a surface texture obtained by means of laser technique can have opposite results. In fact, in comparison with untextured samples, it can lead to a reduction rather than an increase in COF. The motivation of this behaviour can be associated with two different competing mechanisms: (i) stress caused by the edges of the holes that motivates an increase in COF and (ii) capability to trap debris resulting in a decrease in COF [36].

Recently, a new approach has been proposed by some of the authors based on mechanical microindentation for surface texturing. Also, Elias et al. suggested a technique to fabricate microtextures on the flank surface of the cutting inserts using a Vickers microhardness tester and investigated the wear properties of textured cutting inserts under dry and minimum quantity lubrication (MQL) conditions [21]. In another study, Wu et al. employed an indentation method to produce dimple textures on bronze samples surface and studied the tribological properties of indented dimple surfaces. They found that the indented dimple surfaces undergo a lower COF and less wear compared to the non-textured surfaces [37].

As emerged from the previous discussion, the creation of effective microtextures on a surface usually requires advanced and expensive manufacturing processes. There are numerous studies on the effectiveness of microtextures on the tribological behaviour of metallic materials. However, most of these focus on advanced techniques such as laser techniques with their own industrialization problems. Reducing costs and simplifying manufacturing processes is certainly a challenge to bridge the gap with a feasible applicability of such approaches.

Once again, regarding LST techniques, although it is the most widely adopted of all surface modification methods for tribological purposes, they often cause the formation of bulges, burrs or regular contamination around the edge of the dimples. In addition, the material is partially melted due to the high energy density used. Therefore, specimens after processing are usually polished to remove such undesirable effects of the processing itself [38].

Alternatives to laser texturing represent a relatively recent subject of research that has not received much attention. The pursuit of advancements should not be interrupted, and this could be particularly true when referring to the local effects on the microstructure because of the thermal input due to laser techniques (for example softening and, therefore, decreasing of mechanical properties). Moreover, microindentation techniques would be competitive, in terms of manufacturing time and costs, in cases where texturing can be carried out in a single step (or reduced number of steps) using specific tools. From this point of view, microindentation could represent a valid alternative to the techniques based on melting or material removal.

In light of the proposed literature review, it clearly emerges that just as there are many possible techniques that can be used for texturizing (on a nano- or microscale) a surface, there are also many possible causes that affect the tribological properties, especially when metal-to-metal contact are considered, such as (i) reducing the contact area, (ii) trapping any wear debris in the dimples by reducing the rate or even (iii) changing the surface mechanical properties, to name a few.

According to this, the present research aimed at evaluating the use of microindentation techniques for improved dry contact conditions using a common microindenter and applying it to a non-planar surface in order to assess not only its effectiveness but also its versatility degree.

The choice, at least for this first suggested approach, to focus on the dry condition is motivated by the fact that one of the main fields of application of the methodology proposed is that of sheet metal forming. Very often, in this case, working conditions are dry. In fact, while there are sheet metal forming processes where lubricants may be necessary or beneficial, such as in deep drawing operations or when forming complex shapes, dry conditions are often preferred for their predictability, cost-effectiveness, environmental friendliness and ease of handling.

Specifically, the present work investigated the tribological behaviour of a 100Cr6 steel.

Dimples were created on the surface of a sphere using an automatic microhardness tester equipped with a Vickers indenter. In order to reduce the time for preparing the textured surface, since a standard hardness tester was used, the dimples were created on the surface of the sphere used for the ball on disk tests. Therefore, non-conformal dry contact conditions were investigated. Specifically, two different values of the VR (5% and 17%) were investigated. Friction tests at various speeds were performed on textured and un-textured surfaces using a ball-on-disc tribometer. In addition, to evaluate the effect of the dimples size on the tribological behaviour, friction tests were also performed (at a given speed of 4.18 m/s) on microindented surfaces with larger dimples while keeping the VR values unchanged. Textured surfaces were analysed both before and after the friction tests by means of a nanoindenter, a scanning electron microscope (SEM), an atomic force microscope (AFM) and a surface roughness tester.

2 Materials and methods

In Fig. 1, the main phases of the methodological approach adopted are schematized.

Fig. 1
figure 1

Adopted xperimental methodology

Full size image

The surfaces of 100Cr6 steel spheres with 10 mm diameter were textured (step 1) using an automatic microhardness tester (Qness Q10); in particular, as schematically depicted in Fig. 1, a square matrix of footprints was designed in order to have an equal distance along the two principal directions (horizontal X and vertical Y). For this purpose, the ASTM Standard E-384 Vickers method was employed by using a predefined load for 15 s (as will be further detailed below, two different loads were selected). In addition, the different VRs were obtained by changing the distance (X and Y) between two adjacent dimples exploiting the possibility of defining a square matrix directly managed by the Qpix Control2 M Software including inspect microscopy features. Subsequently, the tribological properties of both untextured and textured spherical surfaces were examined in dry conditions through a ball-on-disk CSM high-temperature tribometer using a 100Cr6 rotating sample-disk as counter surface (step 2); the surfaces (before and after the friction tests) were finally analysed both from the point of view of morphology and mechanical properties (step 3).

The 100Cr6 steel was chosen because of its wide use in a countless number of engineering applications among as, to name some: (i) in the manufacture of roller bearings, (ii) in chip removal applications and (iii) in the manufacture of dies. The same alloy was selected as counterpart; this ensures consistency in material properties such as hardness and surface roughness. This consistency can help in reducing variability in test results and allows for more accurate comparisons between different test conditions (size of the dimples and VR value). The chemical composition (expressed as a weight percentages) of the 100Cr6 steel is reported in Table 1.

Table 1 Chemical composition and mechanical properties of the 100Cr6 steel (wt. %)
Full size table

As previously mentioned, the work aimed at experimentally investigating the frictional properties of non-conformal dry contacts. To this end, we decided to texture the sphere instead of the disc as this would be equivalent in terms of friction forces but would shorten the preparation of the textured surface since a standard hardness test machine was used. For this purpose, as depicted in Fig. 2, a proper experimental setup was adopted, thus making the procedure for creating the dimples highly replicable: the sphere was thus kept in place for being mounted in the epoxy resin (Durofast provided by Struers ®) using a steel ring that guaranteed the correct and replicable positioning with respect to the upper surface. The mounting operation was performed using the Struers® Citopress setting the temperature to 180 °C, the pressure to 250 MPa and the total time equal to 5 min (including the cooling step).

Fig. 2
figure 2

Schematic image of a 3D view and b cross section of sphere mounting

Full size image

With the aim to investigate the effect of the textures on the COF and the wear properties, two different values (5% and 17%) of the void ratio (VR), calculated according to Eq. (1), were tested:

$$VR=\frac{A\cdot N}{{A}_{0}}$$
(1)

where \(A\) is the area of a single dimple, \(N\) is the total number of dimples and \({A}_{0}\) is the area of the whole surface [39].

Thus, to ensure a constant contact between the textured surface of the sphere and the disk during the experimental tests, a square matrix area of about 0.8 mm2 was designed. Micrograph in Fig. 3 shows the surface of the sphere after the texturing step.

Fig. 3
figure 3

Micrograph of the textured sphere

Full size image

Friction tests were performed in dry condition both on untextured and textured surfaces. Moreover, such tests were carried out at room temperature (25 ± 2 °C), with a load of 1 N. According to the Hertzian theory, the maximum contact pressure (Pmax) was determined through the Eq. (2)

$$P\text{max}=\frac{1}{\pi } {(\frac{6F{{E}^{*}}^{2}}{{R}^{2}})}^\frac{1}{3}$$
(2)

where R is the sphere radius, F represents the applied force and E.* can be determined, considering both the modulus of elasticity (E) and the Poisson’s ratio (ν), by Eq. (3)

$$\frac{1}{{E}^{*}}=2\frac{1-{\nu }^{2}}{E}$$
(3)

The estimated pressure was 0.469 GPa, where we considered an elastic modulus of 210 GPa and a Poisson’s ratio of 0.3. Such a pressure value, in accordance with what was previously introduced, is in line with various manufacturing processes such as, for example, for metal forming applications. In addition, the calculated value is close to the yield stress value (520 MPa), as summarized in Table 1.

A 30-mm diameter 100Cr6 steel disk with a roughness (Ra) of 0.010 ± 0.001 µm was used as counterpart of the sliding pair. Each disk was properly prepared by grinding and then polished and, finally, degreased with acetone. This procedure made it possible to obtain a disk Ra was very close to that of the sphere (0.045 ± 0.006 µm) and much smaller that the size of each microdimple. This minimizes the effect of surface finishing on frictional properties.

Hardness values were also recorded for both the smooth sphere and the disc: They were found to be comparable to each other (862 HV1 and 874 HV1, respectively).

To identify each single textured surface, we use the acronym VRX-YN where the X is the percentage of the void ratio (VR) and Y is the microindentation load (N). Untextured surfaces are identified with the acronym UT.

Surface analyses, particularly useful for 2D profiling of indented areas, were performed by means of the ZIGO NewView™ 9000 3D optical surface profiler based on CSI (coherence scanning interferometry) and according to ISO 25178 and ISO 4287 standards.

Friction tests were conducted setting six different linear speeds (0.56, 0.84, 2.80, 4.18, 5.59 and 8.37 mm/s); based on these values, the maximum travelled distance was 335 mm. The test durations, as well as the overall covered distance, are consistent with typical sheet metal processes, for example.

To obtain the most robust results possible, each experimental condition was repeated several times; this means that, for each surface condition (untextured and textured with different VRs) and for each value of both load and linear speeds, at least three tests were conducted. Thus, the average values of the COF could be calculated and plotted as a function of the linear speed (as possible to see in Figs. 8 and 14). After each test, the morphology of the worn surfaces was examined by means of a Zeiss EVO-MA10 scanning electron microscope (SEM). In order to evaluate possible process influences in microstructure change, metallographic analyses were performed by means of the Nikon MA200 light microscope with a maximum magnification equal to 700 × . For this purpose, each sample was etched for 5 s with a 5% Nital solution. In addition, to assess very local variations of the material properties determined by the plastic deformation due to microindentations, nanohardness tests were performed using a CSM Instrument (NHT2) equipped with a Berkovich indenter, with the maximum load set to 10 mN. The sequence of nanoindentations in the area between four dimples and a micrograph of the footprints after nanoindentations are reported in Fig. 4: note that nanoindentations were performed in the middle of two adjacent dimples.

Fig. 4
figure 4

A Schematic and b real micrograph of the footprints produced by the nanoindentations

Full size image

Furthermore, the residual stresses induced by the microindentations on the surface of the samples (as well as those of the non-textured surface) were evaluated by X-ray diffraction (XRD) technique according to the Standard UNI EN 15305. To conduct this type of analysis, it was necessary to have planar surfaces; therefore, each texture was fabricated on flat surfaces previously prepared by grinding and polishing processes. In addition, to ensure that the initial state of all substrates was the same, so that the effect of different texture conditions evaluated could be uniquely identified, five spheres were embedded simultaneously (as schematized in Fig. 5) in Bakelite hot mounting resin with carbon filler particularly suitable for SEM analysis.

Fig. 5
figure 5

Definition of local coordinate systems for XRD analysis

Full size image

Residual stress measurements were carried out using a Stress X diffractometer (GNR) equipped with a chromium tube with a wavelength λ = 0.22897 nm. The diffraction patterns for the stress calculation via the sin2Ψ method were recorded by Dectris Mythen2R 1 K detector equipped with 0.02-mm vanadium filter. X-rays were projected onto the surface of the samples using a 0.5-mm diameter circular aperture collimator. Triaxial stress measurements were made at the center point of the textured area along the three different directions identified by a Local Coordinate System (LCS) of each sample, as shown in Fig. 5.

Eleven ψ tilt angles evenly distributed in the angular range (− 45°, 45°) were used to compute elastic strain values in each direction. The diffraction angle (2θ) was set equal to 156.07°, and an exposure time of 120 s was selected. Measurements were based on the peak hkl (211) diffraction reflection of K-α1 radiation assuming the elastic constants S1 and ½ S2 equal to respectively 1.381 × 10−6 and 6.1429 × 10−6 MPa−1.

X-ray diffraction process parameters and data analysis methods are summarized respectively in the following Tables 2 and 3.

Table 2 X-ray diffraction process parameters
Full size table
Table 3 X-ray diffraction data analysis methods
Full size table

3 Results

3.1 Properties of textured surfaces

SEM analyses allowed to evaluate the effectiveness and the replicability of the microindentation technique in order to obtain VR values varying from 5 to 17% when setting the load to 0.5 N. An example of the textured surfaces is shown in Fig. 6.

Fig. 6
figure 6

SEM images of various surface textured specimens: a VR5-0.5N, b VR17-0.5N

Full size image

The results reveal that dimples using the indentation technique are quite regular and their dimensions are highly replicable. The length of the diagonal (L) is about 10 μm, from a macroscopic point of view. As reported in Fig. 6, according to the VR, the distance (D) between two adjacent dimples was largely decreased (more than 40%) when changing the VR from 5% (D = 30 µm) to 17% (D = 17 µm).

In addition, 2D profile of the indented surface referred to the highest void ratio (VR17-0.5N) is reported in Fig. 7b. The surface from which the path was extracted for analysis, on the other hand, is shown in Fig. 7a.

Fig. 7
figure 7

a Analysed path (a) for the 2D profile (b) of surface textured (VR5-0.5N)

Full size image

The 2D profile of an indentation line shows the replicability of the process. In fact, the distance of two adjacent dimples as well as the width and depth of each dimple are invariable along the analysed line. The plots show a distance of 17 µm for a VR of 17%, as expected in the design phase. In addition, as reported in Fig. 7, the depth of each dimple is equal to about 2 µm (this is easy to observe by focusing on the first dimple investigated).

3.2 Friction and wear

Figure 8 shows the obtained values of the COF (average values after three replications) as a function of the linear speed for untextured and textured surfaces.

Fig. 8
figure 8

COF values versus linear speed for the three investigated test conditions

Full size image

COF values of the textured surfaces were significantly lower than the ones of the untextured one (empty markers), thus revealing the effectiveness of the proposed texturing methodology in reducing the friction between the surfaces in contact. More in detail, the test condition VR5-0.5N (green markers) determined an average reduction of the COF about equal to 37.5%; in addition, when VR was increased up to 17% (red markers), the COF was further reduced, thus reaching about a half (53% less) of UT surface’s COF. Note that in both the investigated test conditions (VR5-0.5N and VR17-0.5N) the maximum decrease of COF was obtained at a linear speed of 2.8 mm/s. Note that with an indentation load of 0.5 N the COF decreases as the VR increases. In fact, for the texturing VR17-0.5N, the reduction of COF — compared to the untextured condition — is about 65% at intermediate linear speed (2.80 mm/s) and of 46% at low (0.56 mm/s) or high (8.37 mm/s) linear speeds.

To investigate the contact conditions, the worn surfaces analysed by SEM are presented in Fig. 9.

Fig. 9
figure 9

SEM micrographs of worn surfaces for untextured and textured surfaces: a UT, velocity = 2.8 mm/s; b UT, velocity = 8.37 mm/s; c VR5-0.5N, velocity = 2.8 mm/s; d VR5-0.5N, velocity = 8.37 mm/s; e VR17-0.5N, velocity = 2.8 mm/s; f VR17-0.5N, velocity = 8.37 mm/s

Full size image

The textured surfaces in the figures have been acquired after friction tests conducted setting the speed to 2.8 mm/s (the linear speed at which the largest reduction of COF was obtained) and to 8.37 mm/s (the highest investigated speed). In each micrograph, the sliding direction is highlighted with the white arrow. As visible in Fig. 9a, the untextured sample tested at 2.8 mm/s experienced a significant wear on a large area. As a result of the abrasive wear, parallel grooves and ploughing lines appeared on the wear scar. The wear scar can be also due to localized plastic deformation on the upper edge of the worn surface. At the highest speed (8.37 mm/s), as reported in Fig. 9b, the surface experienced less damage, but the presence of material deposited on the surface can be noted, which could be related to the adhesive wear mechanism.

When considering the test condition VR5-0.5N, some grooves were observed but the samples were characterised by less damage compared to the untextured one, as it is possible to see in Fig. 9 c and d. Moreover, Fig. 9 e and f show that the surface with the high VR value (VR17-0.5N) was scratched much less than the surface with the low VR value (VR5-0.5N), especially if not so high speeds were used. Finally, no accumulation of wear debris was detected inside the observed dimples.

3.3 Mechanical properties

The very local hardening is summarized in Fig. 10 in terms of nanohardness values: For each tested condition, the nanohardness was determined by averaging the results of the four nanoindentation tests conducted close to the dimples, following the scheme proposed in Fig. 4. It is worth noticing that after texturing, the surface becomes harder due to increased dislocation densities [40], thus improving the local mechanical properties (and, in turn, both the wear resistance and the friction behaviour) of the investigated material.

Fig. 10
figure 10

Average nanohardness values concerning the tested surfaces

Full size image

The maximum nanohardness value was recorded when testing the textured surface VR17-0.5N (dotted column in red), which was characterised by a hardness value about double (16.24 GPa) than the UT surface (8.61 GPa). The VR5-0.5N condition, instead, turned out to be intermediate (13.11 GPa) between the other two. Based on these data analysis, the surface of the sample VR17-0.5N might offer the best wear performance if compared with the UT condition (white column).

4 Discussion and comparisons

Local mechanical properties are strongly affected by the imposed surface modifications; in particular, when a high VR was used, such an effect was strongly emphasized. Specifically, the local hardness after microindented texturing, if compared with the UT one, was improved by 52% and 89% for the surfaces VR5-0.5N and VR17-0.5N, respectively. Note that there is a strong inverse correlation between hardness and COF. The surface hardness is, indeed, effective in determining a reduction of COF. Such a hardness increase is due to the plastic deformation caused by the microindentations which produce dislocations in the material’s microstructure [41, 42]. By increasing the hardness of the surface, the adhesion and plastic deformation between the surfaces was reduced, thus leading, in turn, to a better wear resistance of the surface. In fact, it is well known that the surface wear resistance can be explained by the wear volume V according to Archard's law given in Eq. (4), albeit under certain conditions and otherwise applicable to metallic materials [43]

$$V=\frac{K l F}{H}$$
(4)

where K is the wear coefficient, l is the sliding distance, F represents the applied load and H is the surface hardness. Moreover, it was demonstrated in other work that improvement in the mechanical properties (such as surface microhardness) of the material results in growth in the wear resistance of the specimen [44]. On the other hand, as documented by Bhalerao and Lakade, several works have focused on different surface hardening techniques (from thermal to mechanical ones, with or without the presence of special coatings) in order to effectively improve the wear resistance of bearing steel [45].

It is worth noticing that other texturing techniques based on the adoption of a laser source, or on the material removal, do not lead to a modification of the hardness of the material and, therefore, cannot have an additional beneficial effect in terms of friction reduction in the case of dry contacts, as indeed observed by Li et Al., who showed that after texturizing a 304 Stainless Steel by fiber laser, no statistically significant change in the surface hardness of the material itself was recorded [46]. At the same time, Wei et al. showed that, regarding the textures obtained by laser techniques, the dominant mechanism for COF reduction may be associated with contact area reduction [47].

In order to investigate the effect of dimples size, which would be helpful in reducing the texturing time, we also adopted an indentation load of 5 N, while keeping the same VRs used when setting the load to 0.5 N, i.e. 5% and 17%, respectively. For the 5-N load case, we tested the frictional behaviour for a single sliding velocity, i.e. 4.18 mm/s. Figure 11 shows the two surfaces textured using an indentation load of 5 N. The average size of the dimple in this case was about 30 µm (i.e. three times larger compared to the 0.5-N load case).

Fig. 11
figure 11

SEM images of textured surfaces: a VR5-5N and b VR17-5N

Full size image

As done for the low-load indentations, 2D profile of a texture line (referred to the path depicted in Fig. 12a) for the sample indented with the highest value of both VR and Load (VR17-5N) is proposed in Fig. 12b.

Fig. 12
figure 12

a Analysed path (a) for the 2D profile (b) of surface textured (VR5-5N)

Full size image

Texture homogeneity is further evidenced, as already demonstrated in the previous case; all dimples, when compared to those obtained with the lowest load, were more than twice as deep (about 5 µm). High loads led to large dimples, which can be related to the higher value of the local plastic deformation (in fact the hardness values resulted to be increased). Indeed, Fig. 13 shows that the VR17-5N presents the highest hardness values, but the increase compared to the VR17-0.5N and VR5-5N is not very significant.

Fig. 13
figure 13

Nanohardness for all tested conditions

Full size image

This is clearly reflected in terms of coefficient of friction as shown in Fig. 14.

Fig. 14
figure 14

Average COF obtained for a linear speed of 4.18 mm/s

Full size image

Figure 15 shows the SEM images of the damaged surface of VR5-5N and VR17-5N after the friction test at a linear speed of 4.18 mm/s. As before, the sliding direction is highlighted by the white arrows.

Fig. 15
figure 15

SEM images of the surfaces textured using 5N after the friction test setting the linear speed to 4.18 mm/s: a VR5-5N; b VR17-5N

Full size image

The worn part of the surface is very tiny; therefore, the very small area to be investigated made the focus difficult. Nevertheless, as underlined in the two micrographs, some particles were observed on the surface, which shows the capability to trap debris by large dimples. The presence of this debris within the dimples has evidently improved the wear condition since, although parts of the debris are clearly visible in the SEM micrographs, scratches were found to be almost absent of very small. Such evidence was not observed in the case of small dimples (obtained with a load of 0.5 N), as also supported by the micrographs in Fig. 9.

Considering the investigated texturing conditions, the best performances were obtained when setting the load to 5 N and the VR to 5% (in this way the manufacture time was also reduced) and when combining the low load (0.5 N) with the high VR (17%). However, there is no doubt that, although the hardness is almost the same for the conditions VR17-0.5N, VR5-5N and VR17-5N, the surfaces evaluated via SEM after friction tests (Figs. 9 and 15) are very different: In the case of textures obtained with the highest load value, in fact, the scratches are almost absent.

Table 4 reports the main results related to the experimentally measured residual stress values for each of the different investigated surface conditions according to the three directions (0°, 45° and 90°) identified by local coordinate system.

Table 4 Values of residual stress along 0°, 45° and 90°
Full size table

All examined surface conditions exhibit significant compressive residual stresses at the surface, with values exceeding 1000 MPa in absolute value. The UT condition, as expected, shows the lowest (absolute) residual stress value compared to the other four textured conditions. The peak of surface compressive residual stress is observed for the VR17-5N condition. No significant variation of the residual stresses with the direction of measurement is observed; in fact, the values of the coefficient of variation (CV) calculated as the ratio between the sample standard deviation of the stresses in the three directions and their mean value in absolute value is in all cases less than 7% except for the UT condition (12%).

In light of the latter consideration, the plot of stresses of only one direction (σ 90°) was considered end reported in Fig. 16 in order to understand the influence of both applied load and VR value.

Fig. 16
figure 16

Residual stress according to the different investigated conditions (a) and the recorded nanohardness (b)

Full size image

The residual stresses increase consistently with the load and VR, as demonstrated by the plot in Fig. 16a. In addition, the correlation with the previously discussed nanohardness values is also evident (see Fig. 16b). These results are in full agreement with the existing literature; in fact, for example, Yonezu et al. found that, in the case of stainless steel, indentation load is intimately related to residual stress, which increases according to the applied load [48].

Finally, slight shear stresses are also observed in all the samples examined, probably associated with the metallographic preparation to which they were subjected prior to the RXD measurements. In addition to this, it is known that the grinding and polishing steps in the processing of metallic materials like 100Cr6 steel can impact residual stresses through various mechanisms, including mechanical deformation [49]. In support of this, Turley et al. state that the grinding operation on hardened steels produces compressive residual stresses in the surface layer ranging from 400 to 900 MPa depending on the type of grinding procedure [50]. In addition, as demonstrated by Everarets, a maximum residual stress value of about 300 MPa in compression is achieved after grinding and polishing in Ti6Al4V alloy specimens [49].

Therefore, considering that the starting condition for all samples analyzed is the same after the grinding and polishing procedures, the percentage increase in residual stresses ranges from 31 (VR5-0.5N) to 59% (VR17-5N). This increase is nearly linear, except for the last condition (VR17-5N) for which a markedly larger value is recorded.

As is well known, residual stresses often have a major effect on the physical behaviour of materials and, therefore, of the final components; in particular, fatigue life, distortion, dimensional stability and corrosion resistance are the main properties that can be affected by such residual stresses [51]. In general, residual compressive stresses are a desired effect since they help to reduce the incidence of any tensile stress loads. In addition, in most cases, surface compressive stresses contribute to improved fatigue resistance, as well as to stress corrosion cracking. In this regard, Cretu et al. demonstrated both experimentally and numerically the beneficial effect of residual stresses in the elastic–plastic domain. Specifically, investigations of ball bearings with induced residual stresses have demonstrated a longer fatigue life when compared with those without an initial stress level [52].

In general, as stated by several works and previously discussed in Sect. 1, the microindentations on the surface creating small depressions or grooves result in a substantial reduction of the contact area when facing another substrate (even non-textured). This reduction in contact area consequently leads to a decrease in the overall friction force [2, 22, 34]. To this effect on the part of the dimples must be added the capability to trap wear debris, which, in addition to reducing wear effects, can act as a lubricant [2, 22, 35]. The other possible cause of the improved tribological behaviour may be related to work hardening [44] due to mechanical penetration of the indenter in dimple fabrication. According to the experimental results collected in the present work, however, the main causes of COF reduction can be clearly related to both work hardening phenomena and the presence of compressive residual stresses induced by the micro indentation process.

A scheme of these two phenomena is proposed in Fig. 17.

Fig. 17
figure 17

Scheme of the two mechanisms by which textured surface obtained by microindentation can improve the tribological behaviour: a work hardening and b compressive residual stress induced

Full size image

Below the indented area, a dislocation structure is expected. This area represents the plastic zone characterised by an increased dislocation density. Such dislocations are essentially necessary from a geometric point of view [53, 54]. In fact, the material originally below the plastic region is forced into the material as an additional defect cluster. More specifically, as stated by Xu and Tian, the red lines in Fig. 17a schematize possible sliding systems in polycrystalline material through which dislocations have a greater chance of moving [55].

As schematized in Fig. 17b, during the Vickers indentation process, once the load is removed, there is a recovery of elastic deformations resulting in residual stresses induced in this region [56].

Baxevani et al., in this regard, evaluated the cold deformation energy stored below the penetrator in the form of dislocations by correlating it with the residual stresses due to the presence of these lattice defects [57].

A further investigation on the dimples profile according to the two different investigated load (0.5 and 5 N) was conducted by exploiting both an atomic force microscope (AFM) NT-MDT NTEGRA and a surface roughness Mitutoyo SJ-400 tester (according to the DIN 4768 Standard).

This analysis allowed to identify the presence of evident bulges in the case of dimples created using the load of 5 N. As shown in Fig. 18, it is worth noticing that the surface textured using the load of 0.5 N is more regular than the one obtained with the load of 5 N. As evidenced in Fig. 18b, bulges at the perimeter area of the dimple were observed; such bulges are, on the contrary, absent when setting the load to 0.5 N, as evident in Fig. 18a.

Fig. 18
figure 18

Dimples 3D profile obtained by means of AFM: a 0.5 N and b 5 N

Full size image

More in detail, as shown in Fig. 19, the comparison of the profiles of the dimples obtained setting the two different load values reveals that at high load (5 N) produced bulges of about 0.5 µm height at the edges; on the contrary, such a phenomenon is almost absent at low load (0.5 N).

Fig. 19
figure 19

Dimple 2D profile obtained by means of surface roughness: a 0.5 N and b 5 N

Full size image

The measurements conducted with both types of instruments (AFM and surface roughness) are consistent with each other. It is important to point out that the dimple depth for both cases reported in Fig. 19 is less than the surface analyses previously described and illustrated in Figs. 7 and 12. This was expected since the stylus tip dimensions of the surface roughness prevent its contact with the bottom of the dimple (as also stated by the standard ISO 25178–601); conversely, this does not limit the measurement of the peaks (bulges).

Although the creation of bulges on the texture edge has been reported to negatively affect friction and wear performance by some researchers [25, 26], the textured surfaces with ‘‘bulges’’ have better anti-stick–slip effects than the untextured one.

Under the same normal pressure, the smooth surface has a larger contact area and more touchpoints than a textured surface. The number of touchpoints is reduced when textured surfaces have bulges, as surface-to-surface contact is replaced by multiple point contacts. Thus, irrespective the presence of dimples with bulges, COF values resulted to be smaller than the UT surface. These results are also supported by the evidence recorded by Amanov et al. [38] One could expect that the larger superficial hardness than the one determined by small load would lead to further decrease of the friction when using high load, but the presence of bulges probably inhibited such a tribological effect. In fact, as stated by Amanov et al., it was found that bulges are not necessarily a limitation for friction and wear behaviour. However, an increase in the height of such bulges determined a lower effect of texture on the COF.

In support of this experimental evidence, Rapoport et al. showed that especially when the characteristic dimple depth is not high, the presence of the bulges significantly reduces the texture wear rate with an improvement in COF as well [38].

Moreover, hard bulges may strongly affect disk wear. In fact, as reported by the micrographs of the disk’s worn surfaces in Fig. 20, larger wear is observed when the disk is in contact with high-load structured surface, given the same VR (17%). Such an aspect plays a key role in the design phase of the textured surface to be created using the proposed methodology. In fact, the parameter VR largely affects the tribological behaviour of such indented surfaces, confirming the capability to reduce the COF in dry condition, at least when dimples of limited size (obtained with a load of 0.5 N and with an average size of 10 µm) are considered.

Fig. 20
figure 20

Micrographs of worn disk surfaces after friction tests at 4.18 m/s conducted using textured spheres characterised by the same VR (17%) but created using the load of 0.5 N (a) and 5 N (b)

Full size image

On the contrary, the adoption of the large load (5 N) for creating the VR equal to 5% could be attractive from an industrial point of view, since it would allow to drastically reduce the texture manufacturing time, but the presence of bulges may cause worst wear behaviour of the counter surface.

As regard the microstructure of the material before and after texture fabrication, it is possible to observe the following micrographs reported in Fig. 21. Because of the difficulty in placing exactly in the cross section of each dimple, the analysis was conducted by considering the top sections, as shown in Fig. 21.

Fig. 21
figure 21

Micrographs of tested surfaces: UT (a), VR5_0.5N (b), VR17_0.5N (c), VR5_5N (d) and VR17_5N (e)

Full size image

The microstructure of the samples for all conditions investigated consists of martensite (characterised by a beige or brown colouring) with retained austenite (light areas) and globular carbides (small white spots) [58, 59]. As can be seen, the surface modification does not produce microstructural changes (a pronounced black colouration for the micrographs of the dimples obtained with a load of 5 N is due to local corrosion effects after chemical etching), confirming that the main mechanisms involved in the surface modification are essentially the densification of dislocations intimately and the increase in residual compressive stresses. This is also supported by the microscopic analysis of the textured surfaces using polarised light, which revealed visible plastic deformations on the sides of the dimples, as reported in Fig. 22 and highlighted with white arrows (not only more evident in the case of dimples obtained with 5 N load, but also present for the lower applied load).

Fig. 22
figure 22

Micrographs with polarised light of VR5_0.5N (a), VR17_0.5N (b), VR5_5N (c) and VR17_5N (d)

Full size image

Further consideration should be given to the integrity of the dimples produced; in fact, the analysis of the cross section by optical microscopy of the most severe condition (5 N as load) shows the absence of damage due to the process; in fact, as underlined by the black arrows, it is possible to observe the absence of cracks at the boundary of the indented zone (Fig. 23). This result is also consistent with the compressive nature of the residual stresses generated, which preserves crack propagation in the material.

Fig. 23
figure 23

Micrographs of cross section regarding the dimpled area with a load of 5 N

Full size image

5 Conclusions

The present work proposes an innovative methodology for the creation of textured surfaces able to reduce the friction in dry conditions using plastic deformation (microindentation) instead of a material removal technique.

Friction tests using spheres whose surface was textured using the lowest load (0.5 N) revealed that the void ratio (VR) largely affects the tribological behaviour of indented surfaces, confirming the possibility of reducing the COF in dry condition: the higher the VR, the lower the COF. Such a result was found to be strictly related to the very local hardness increase (measured by nanoindentation measurements) of the manufactured surface coupled with compressive residual stresses (measured by means XRD).

We also investigated the tribological properties of the surface textures using the highest load (5 N), by keeping unchanged the VR values. We observed that at the highest load, a significant reduction of COF is obtained when setting the VR value to about 5%; such results are comparable to the friction reduction obtained when setting the load to 0.5 N load and void ratio to 17%.

On the other hand, a further increase of the void ratio up to VR 17% when keeping loading to 5 N does not provide any significant further improvement of the tribological properties. In fact, in this case, the presence of bulges may even cause more severe wear on the counter-surface.

The proposed methodology, by limiting the use of lubricants for reducing the friction at the tool-sheet interface, is able to improve both the sustainability (it allows the reduction of the environmental impact) and the efficiency (no noon of washing components for removing the lubricant) of manufacturing processes.

In addition, the proposed methodology has big potentialities to be exploited for improving the contact behaviour during the use or the production of the component: for example, in order to improve the stamping behaviour, local modifications of the COF at the tool-sheet interface could be obtained by creating microindented textures, even during the rolling operation.