Characterization of surface textures generated on hardened steel parts in high-precision machining operations

This paper presents the comparison of surface textures produced by high-precision cutting and abrasive processes on hardened steel parts. Investigations include surfaces generated by hard turning, belt grinding, and superfinishing (external honing) operations. As a result, several surfaces generated with different process kinematics and the Sa roughness parameter of about 0.05 μm were compared. Apart from the standard 2D and 3D roughness parameters, the fractal, motif, and frequency parameters were considered.


Introduction
Nowadays, the technological shifts in surface metrology allow the surface features generated by modern manufacturing processes (including hard part machining) to be characterized with a higher accuracy using a number of the field parameters (S-parameters and V-parameters sets) [1]. Special attention in this area has been given during two past decades to the standardization of 3D set roughness parameters [1][2][3]. Precision machining with Rz=2.5-4 μm and high-precision machining with Rz<1 μm of hardened steels (45-60 HRC) with ultrahard cutting tool materials have been developed with a special interest to automotive, hydraulic, and die and mold industry sectors [4,5] due to high flexibility, possible complete machining, lesser ecological impact, and higher MRR [6,7]. The discussion platform over the possible replacement of grinding by hard cutting was initiated in [8]. In particular, the capability profiles of cutting and grinding operations against the functionality of the machined surfaces were discussed. This is because hard turning and grinding, as well as other abrasive finishing operations, generate different surface structures which influence distinctly their functional properties. Unfortunately, this attempt includes only 2D height and amplitude parameters and the relevant bearing area curves (BACs) shapes [8]. However, these analyses showed a dissimilarity between the hard turned and ground surface topographies, although the Ra or Rz parameters are nearly the same. The 2D and 3D comparison, more oriented on bearing area parameters, related to precision hard turning and belt grinding is provided by Grzesik et al. [9]. Moreover, it was extended to superfinishing and ball burnishing operations [10] in order to check how nonremoval burnishing process can modify the characteristics of the turned surface. In this aspect, a special focus should be placed on surface finish and surface texture induced by these challenging operations [9,10]. The objective of this study is to comprehensively characterize and compare surface textures of representative hard turned and two abrasively treated surfaces using a number of standardized 3D roughness parameters as well as other characteristics such as fractal dimension, motif, and frequency parameters. The quantitative criterion was that the Sa and Sz roughness parameters are about 0.05 and 1 μm, respectively.   [9] and superfinishing head mounted in a conventional lathe described in [10].
Machining conditions for cutting and abrasive operations performed were as follows: 1. High-precision hard turning (HT) using TNGA 160408 S01030 chamfered CBN insert, cutting speed v c =150 m/ min, feed rate f=0.025 mm/rev, and depth of cut a p = 0.03 mm. Previously, surfaces were turned using the same cutting tool with feed rate of 0.05 mm/rev and depth of cut of 0.05 mm. 2. Two step oscillation belt grinding (BG) using abrasive belts with 30 and 9 μm grains. Rotation speed of the workpiece was 900 rev/min, belt feed was 0.06 mm/rev, oscillation frequency was 12 Hz, oscillation amplitude was ±0.5 mm, and roller pressure was 2 bars. Belt grinding ( Fig. 1) was performed during 9 s with supplying oil mist produced by a MQL system. 3. Superfinish (SF) using 99A320N10V ceramic stone, oscillation frequency of 680 osc/min and amplitude of 3.5 mm, applied force of 40 N, and cooling and lubrication medium 85 % kerosene and 15 % machine oil.
Surface topographies generated by HT, BG, and SF operations were recorded using a 3D contact profilometer with a diamond stylus radius of 2±0.5 μm and an Alicona noncontact device. The determination of 3D roughness parameters and 3D visualization of machined surfaces were performed using a Digital Surf, Mountains® Map package. The characterization of surface topographies was based on four groups of parameters including (a) standardized 3D surface roughness parameters: height, amplitude, horizontal, hybrid, and functional [11]; (b) fractal dimension; (c) standardized motif parameters; and (d) characteristics of frequency spectra recorded.

Characterization of surface topography
Representative surface topographies obtained in hard turning (HT) and abrasive (BG, SF) operations performed are visualized in Fig. 2a-c using a special zooming technique. In terms of the surface quality criterion, both operations can be classified as high-precision machining-Rz≈2 μm [8]. As shown in Fig. 2, the measured values of Sa oscillate between 0.045 and 0.07 μm. As depicted in Fig. 3, Sz parameter ranges from 0.78 for hard turning to 1.20 and 1.33 μm for SF and BG operations, respectively. It means the comparison of the turned and a) Sal=0.01, isotropy: 0.68% b) Sal=0.06 isotropy: 5.19% c) Sal=0.01, isotropy: 12.8% abrasively treated surface textures keeping height Rz (Sz) parameter constant seems to be questionable.
The strong anisotropy of all machined surfaces shown in Fig. 2 is confirmed by characteristic shapes of the autocorrelation function (AACF) presented in Fig. 4. The turned surface is periodic-anisotropic ( Fig. 4a) but the abrasively treated surface is mixed, between anisotropic and random structures (Fig. 4b, c). The values of the fastest decay autocorrelation length (Sal) are equal to 0.01 for hard turned and 0.06 or 0.01 for abrasive treated surfaces, respectively. A larger Sal=0.06 for the belt ground surface (Fig. 4b) suggests that its texture is compounded of low spatial frequency components due to copying of ultra-fine grains embedded into abrasive belts (see Fig. 13a). Figure 5 presents the shapes of 3D BACs and associated Amplitude density function (ADF) curves obtained for the compared machining operations. In particular, as shown in Fig. 6, hard turning produces surfaces with a positive skew Ssk=0.17 (HT) while finish belt grinding generates surfaces with a small negative skew Ssk=−0.07 (BG), which visibly increases when superfinish is performed (Ssk=−1.03 for SF).

Characterization of function-related parameters
Moreover, Fig. 5b suggests that hard turning and abrasive operations produced topographies with distinctly different ADF shapes, which result in various bearing and contact properties. The superior bearing properties corresponding to Ssk= −1.03 were obtained when sharp irregularities produced by hard turning were removed by abrasive stone during superfinishing (BAC #3 in Fig. 5a). It can be noted in Fig. 5b that the ADF curve is similar to a typical bell (Gaussian) curve because the kurtosis value is close to 3 (Sku=2.60). On the other hand, the ADF curves are distinctly sharper when the initial turned surface is modified by belt grinding or even more by superfinish (kurtosis increases to 5.31 and 6.13, respectively).
Additionally, values of the areal material ratio Smr (c), the inverse areal material ratio Sdc (mr) and the peak extreme height Sxp are given in Fig. 5a. In Fig. 9a, they are analyzed in terms of areal (V) parameters including the reduced core (Sk), peak (Spk), and valley (Svk) height or their specific ratios (Spk/Sk, Svk/Sk, Spk/Svk).
Additional information on the fluid retention can be obtained using an original technique of the vectorisation of microvalleys network (Fig. 7) generated on the machined surface [2,3]. The maximum depth of valleys is between 0.35 and 0.55 μm and their widths 0.08 and 0.16 μm depending on the sequence of machining operations performed. Additionally, the average density of valleys is between 650 and 950 cm/cm 2 , respectively (the lower value corresponds to the turned surface and the higher one to the superfinished surface). This comparison indicates that abrasive operations produce surfaces with a larger number of more densely distributed valleys (Fig. 7b, c). As a result, the functional properties of these two groups of surfaces should be different. These data coincide well with the distributions of the volume functional parameter (Vmp and Vvv) shown in Fig. 8.
The functional analysis of the 3D BACs is based on the four volume parameters including the peak material volume (Vmp), the core material volume (Vmc), the core void volume The comparison between the function related parameters [12] are given in Fig. 9. In these case studies, three ratios of various areal material ratio parameters-Spk/Sk, Svk/Sk, and Spk/Svk-were used in order to asses the nature of the particular surface textures. The ratio of Spk/Sk may be helpful to distinguish between two surfaces with indistinguishable roughness average Sa [13]. In fact, surfaces generated by cutting and abrasive tools with the same Sa have slightly different Spk/Sk values of 0.333 (HT), 0.308 (BG), and 0.267(SF) as shown in Fig. 9a. On the other hand, Svk/Sk ratio is equal to 0.208 for turned surfaces and increases vastly to 0.54 and 0.60 for surfaces produced by BG and SF operations. Also Spk/Svk ratio is distinctly different-1.6 for HT and about 0.5 for abrasive operations. As shown in Fig. 9b, the ratio of Spk/Sk correlates well also with the Vmp parameter, whereas the ratio of Svk/Sk with Vvc volume parameter and micro-valleys density. Additional relationships can be observed (Fig. 9a) between the ratio of Spk/Svk and Sdc and Sxp material ratio parameters.

Characterization of spatial and hybrid parameters
The set of 3D parameters includes four spatial parameters, three of which are texture parameters. The belt ground and especially superfinished surfaces contain distinctly more summits within the scanned area-Sds=3421.2 1/mm 2 (SF) versus 1788.6 1/mm 2 (HT). The small texture aspect ratio Str= 0.01-0.05 for turned and ground surfaces indicates stronger directionality (anisotropy); its values less than 0.1 are characteristic for highly anisotropic surfaces [11] (for SF is something higher of 0.13). The texture direction Std close to 90°for HT and BG surfaces indicates that the dominant surface lay is perpendicular to the measurement direction (for SF operation Std = 112.5°which results from crossing lays shown in Fig. 2c). The values of Sal parameter are given in Fig. 4. Values of three 3D hybrid parameters emphasize additional geometrical differences in the compared textures. Very low slopes Sdq of 0.01-0.02 were obtained for all machined surfaces which suggest their high optical quality. The values of the average summit curvature Ssc are between 0.005 and 0.007 μm −1 which are typical for machined surfaces (0.004-0.03 μm −1 given in [11]). The Sdr parameter (the developed interfacial area ratio) of 0.01-0.02 % is obtained.

Motifs and fractals
The motif analysis is performed on the unfiltered surface profile divided into a series of windows [11], as exemplarily shown in Fig. 10. The three parameters-the mean depth of roughness motif R, the mean spacing of roughness motif AR, and the largest motif height Rx were analyzed.
The values of Rx are comparable, although turned surfaces include slightly deeper pits (Rx=0.45 μm) than belt ground and superfinished surfaces (Rx=0.4/0.43 μm) which is in accordance with volume bearing parameters (Fig. 8). Figure 11 shows that the Rx motif parameter is stronger correlated with the Sz parameter rather than Rz, although motifs are based on 2D analysis. On the other hand, the R motif parameter of 0.14-0.25 μm better coincides with the Rz changes.
The values of fractal dimension Sfd determined by means of the method of enclosing boxes are equal to 2.53, 2.54, and 2.71 for turned, belt ground, and superfinished surfaces, respectively.
Functional relationships between fractal dimension Sfd and Sal, Ssc and Sds, spatial and hybrid parameters were revealed. In Fig. 12, the Sfd is strongly correlated with the density of summits (Sds), and Sfd=2.71 corresponds with the maximum value of Sds=3421.2 1/mm 2 determined for superfinished   surface. At the second order, it deals with the arithmetic summit curvature (Ssc) and the autocorrelation length Sal parameter which characterize the uniformity of the texture.

Frequency analysis
The characteristic power spectral density (PSD) obtained for hard turned and ground surfaces are presented in Fig. 13. The PDS is very sensitive for all disturbances of the generated surfaces appearing in the machining system.
It is evident in Fig. 13a that the PSD spectrum contains only one dominant low-frequency component with the wave length something lower than the feed rate of 0.025 mm (25 μm) and the amplitude of 0.035 μm. On the other hand (Fig. 13b), the belt ground surface contains two components with longer wavelengths but distinctly lower amplitudes, below 0.02 μm, than for hard turning. Comparable small amplitudes of about 0.02 μm were recorded for surfaces produced by superfinish, although belt grinding is more stable (dynamic effect after abrasive processes depends on the structures of initial textures).

Conclusions
This study clearly indicates how high-precision machining operations can be performed in order to obtain desired surface texture and functional properties, i.e., resistant to wear, fluid retention ability, resistant to contact loads.
Although attributes of turned and ground and honed surfaces are described by the same Sa parameter of about 0.05 μm, their spatial features and functional properties are distinctly different.
The distributions of the PSD (APSD) function and vectorial maps of micro-valleys suggests that the textures of hard turned and ground and honed surfaces are periodic-anisotropic and mixed periodic-random anisotropic, respectively.
3D BAC curves and appropriate functional parameters indicate that honed hard surfaces have enhanced fluid retention abilities. This is due to negative Ssk value and up to three times higher Vvv volume in comparison to hard turned and belt ground surfaces.
Belt ground and honed textures have comparable Vmp and Spk parameters and similar tribological properties. This is due to minimum Vmp and Spk values in comparison to the turned surfaces.
Open Access This article is distributed under the terms of the Creative Commons Attribution License which permits any use, distribution, and reproduction in any medium, provided the original author(s) and the source are credited.