2024 Volume 64 Issue 1 Pages 134-141
Samples treated using conventional plasma nitriding have good surface hardness, wear resistance, corrosion resistance, and fatigue strength, but their friction coefficients are not low enough. This study presents a novel method of reducing the friction coefficient of AISI H13 tool steel through a hybrid treatment consisting of atmospheric-pressure plasma nitriding and superheated steam treatment. The surface structures and tribological and mechanical properties of a hybrid-treated sample were investigated. Results showed that atmospheric-pressure plasma nitriding had no effect on the formation of Fe3O4, which improves the corrosion and tribological properties of tool steel. The surface of the hybrid-treated sample had an oxide layer separated into two layers. The nitrided and non-nitrided samples had nearly the same thickness of the oxide layer. The outermost layer of the hybrid-treated sample contained almost no Cr, a large amount of nitrogen, and small amounts of Fe and O. From its outermost surface to its base material, this sample had a three-layer structure consisting of a nitride layer, a Fe3O4 layer, and a Cr-rich oxide layer. The depth of the diffusion layer of the hybrid-treated sample was greater than that of a sample treated using atmospheric-pressure plasma nitriding only. The outermost surface of the hybrid-treated sample was softer than its inner part, and the hybrid-treated sample had the lowest friction coefficient among all samples. Overall, the hybrid treatment reduced the friction coefficient and improved the wear resistance of AISI H13 tool steel.
Plasma nitriding is extensively used for the study and application of surface modification to various steels.1,2,3,4) Plasma nitriding improves the surface hardness, wear resistance, fatigue strength, and other characteristics of materials while maintaining their core properties.5,6,7,8) An advantage of plasma nitriding over conventional nitriding is that the former is a clean, nontoxic process that involves a shorter nitriding time than gas nitriding.9,10) Plasma nitriding using glow discharge is the most common method of plasma nitriding, but this requires a vacuum system to generate plasma.11) Therefore, most traditional plasma nitriding methods have higher initial system costs compared with other nitriding techniques.12,13,14) Moreover, limited products can be processed via conventional plasma nitriding due to its vacuum atmosphere. Therefore, we are developing a plasma nitriding method that is conducted at atmospheric pressure.13,15) However, the friction coefficients of atmospheric-pressure-plasma-nitrided samples are higher than those of samples treated using other nitriding methods. A mechanical part that contacts two surfaces requires a low friction coefficient.
Superheated steam treatment has long been used for cutting tools.16) Superheated steam is a type of vapor produced by adding heat to saturated water steam.17) A complex oxide layer with Fe2O3 and Fe3O4 structures forms on the steel surface, improving its corrosion and tribological properties.18,19) Conventional plasma nitriding and steam treatment cannot be performed in the same chamber due to the vacuum atmosphere in plasma nitriding. In this study, superheated steam treatment was performed on an atmospheric-pressure-plasma-nitrided sample.
The atmospheric-pressure plasma oxynitriding of tool steel was performed via a hybrid treatment consisting of atmospheric-pressure plasma nitriding and superheated steam treatment. The surface structures and tribological and mechanical properties of a sample oxynitrided through the hybrid treatment were investigated.
The chemical composition of the sample material, AISI H13 tool steel, is shown in Table 1. Each sample was a disc with a 20 mm diameter and a 5 mm thickness. Quenching was performed at 1030°C in a vacuum for a holding time of 80 min; subsequently, the samples were tempered to 540°C for 240 min. The hardness of the heat-treated samples was 580 HV. The sample surfaces were ground and polished in several stages. Final surface finishing was performed through buffing using 0.3 μm alumina. The samples were polished to a mirror finish, and their average surface roughness was Ra = 12 nm.
| C | Si | Mn | P | S | Cr | Mo | V | Fe |
|---|---|---|---|---|---|---|---|---|
| 0.37 | 0.94 | 0.43 | 0.022 | 0.003 | 5.28 | 1.21 | 0.80 | Bal. |
In this paper, the plasma nitriding of tool steel was performed using the dielectric barrier discharge method at atmospheric pressure. A schematic diagram of the plasma treatment apparatus is shown in Fig. 1. Plasma was generated by the potential difference between a pulsed-voltage-biased electrode and the sample. Borosilicate glass was placed around the electrode as a dielectric barrier to generate glow discharge. In this method, the temperature of the material after plasma irradiation is insufficient. The sample must be heated during plasma nitriding to dissolve the nitrogen in the sample. Therefore, a heater was placed beneath the sample to heat it during plasma irradiation. The dielectric and the sample were positioned parallel to each other, and the distance between them was kept constant. Tubes were used to flow the gas and steam from above to the area between the electrode and the sample. A glass chamber was used so that the status of the plasma treatment could be confirmed from the outside.
The electrode in the chamber must be negatively or positively biased to generate plasma. The sample was used as a ground electrode in our setup, and the sample bias voltage was a bipolar pulse of Vp-p = 3.5 kV. The voltage was set such that arcing did not occur and stable glow discharge occurred. Each sample underwent the dielectric barrier discharge method for 3 h. The mass flow rates of argon (99.9%), nitrogen (99.9%), and hydrogen (99.9%) were 5, 3, and 1 standard liters per minute (slm), respectively. The distance between the dielectric and sample surfaces during plasma nitriding via dielectric barrier discharge was 1 mm. The treatment temperature during nitriding was maintained at 500°C using the heater installed beneath the sample. The size of treatment chamber was inner diameter 60 mm × length 80 mm. Nitrogen was introduced through the nozzle at 5 slm for 5 min before plasma generation to purge residual oxygen from the chamber and avoid sample oxidation. However, this chamber is not a vacuum chamber; it is a simple chamber, and the inflow of air during processing was unavoidable. The effect of this air inflow on oxidation did not change considerably, even when the purging time was extended. Therefore, the purging time was set to 5 min. Thus, oxidation by air and reduction by hydrogen plasma occurred simultaneously during nitriding.
In the steam treatment, the sample temperature was set to 500°C by controlling the heater beneath the sample. The superheated steam was flowed to the chamber at 250°C using a superheated steam generator. The sample was oxidized at a steam temperature of 100°C to 250°C to clarify the effect of the steam temperature. The thickness and structure of the oxide layer were modified by changing the steam treatment time (0.5 and 2 h).
The surface and cross-sectional hardness of the nitrided samples was investigated using a micro-Vickers hardness tester (HM-102, Mitutoyo Corporation, Kanagawa, Japan). A hardness test was performed using a 0.098 N (0.01 kgf) load to measure the limited range of each sample. Surface analysis was performed via X-ray diffraction (XRD) with Cu-Kα radiation (ATX-G, Rigaku Corporation, Tokyo, Japan). The surface morphologies and cross-sectional micrographs of the oxidized and oxynitrided samples were investigated using scanning electron microscopy (SEM) (JSM-7800F, JEOL Ltd., Tokyo, Japan). Elemental analysis was conducted on the oxidized and oxynitrided samples using energy-dispersive X-ray spectroscopy (EDS) (JSM-7800F, JEOL Ltd., Tokyo, Japan). A digital microscope (RH-2000, Hirox Co. Ltd., Tokyo, Japan) was used to examine the cross sections of the nitrided samples. The friction coefficients and wear rates of the nitrided samples were investigated using 10 mm Al2O3 balls in a dry-atmosphere ball-on-disk reciprocating tribometer (Heidon 40, Shinto Scientific Co. Ltd., Tokyo, Japan). The applied load was 0.49 N; the samples were rubbed for 5000 cycles at a speed of 5000 mm/min and a sliding amplitude of 5 mm. Images of the wear tracks were captured using the digital microscope to clarify the wear mechanism. The surface profile of each sample and the cross-sectional profile of its wear track were investigated using a stylus-type roughness meter (SURFCOM NEX 001 SD-12, Tokyo Seimitsu Co. Ltd., Tokyo, Japan). The measurement length and velocity of the surface profile were 1.25 mm and 0.06 mm/s, respectively, and those of the cross-sectional profile were 3 mm and 0.06 mm/s, respectively.
Figure 2(a) shows the XRD patterns of the untreated sample and the oxidized samples. The diffraction pattern of the untreated sample showed α-Fe phase peaks only. Fe3O4 and Fe2O3 peaks were detected in the XRD pattern of the sample oxidized at 100°C for 0.5 h. However, Fe2O3 peaks were not detected in the patterns of the samples oxidized at 250°C. The Fe3O4 peak intensities of the samples oxidized at 250°C were higher than that of the sample oxidized at 100°C. Therefore, superheated steam is more suitable than saturated steam for the formation of a Fe3O4 layer with superior tribological properties. The Fe3O4 peak intensity of the sample oxidized for 2 h was higher than those of the samples oxidized for 0.5 h. We assumed that the oxide layer grew as the treatment time increased. Figure 2(b) shows the XRD patterns of the untreated sample and the hybrid-treated samples (oxidized after atmospheric-pressure plasma nitriding). The highest peak intensities of Fe4N, Fe3N, and CrN in nitrided AISI H13 appeared at 2θ values of 41.2°, 43.4°, and 43.7°, respectively.13,20) Fe4N, Fe3N, and CrN peaks, which are detected in the presence of a compound layer, were not detected under all experimental conditions when analysis was conducted before steam treatment. Therefore, no compound layers were generated on the surfaces of the samples nitrided using plasma at atmospheric pressure, or compound layers formed but were too thin to be detected. This behavior may be due to the samples’ nitrogen concentration; such a compound layer is related to the nitrogen solute concentration in a sample.21,22) The Fe3O4 peak intensities increased with the treatment temperature and treatment time, similar to those of the steam-treated-only samples. Thus, the nitriding treatment had no effect on the formation of Fe3O4. The α-Fe peak intensities of the hybrid-treated samples were lower than those of the steam-treated-only samples, and the peak widths were broader. These results were attributed to the formation of nitrided layers.
Figures 3(a) and 3(b) show the SEM images of the samples treated using only steam treatment for 0.5 and 2 h, respectively. Oxide crystals were observed on the surfaces of both samples. The oxide whisker crystals on the sample treated for 2 h were larger compared with those on the sample treated for 0.5 h. Figure 3(c) shows the SEM image of the sample treated via atmospheric-pressure plasma nitriding and steam treatment for 2 h. Its surface structures were not only oxide crystals but also nitride crystals.
Figure 4 shows the cross-sectional SEM images of non-nitrided and nitrided samples. Each sample surface had an oxide layer separated into two layers. The non-nitrided sample subjected to steam treatment for 0.5 h had an outer-layer thickness of approximately 550 nm and an inner-layer thickness of approximately 370 nm. The non-nitrided samples did not show any change in the outer-layer thickness with the increase in the treatment time, but the inner-layer thickness increased. Thus, the outer layer formed first, and the inner-layer thickness increased with the treatment time. The oxide layers of the nitrided and non-nitrided samples had almost the same thickness. Therefore, the nitriding treatment did not affect the thickness of the oxide layer.
Figure 5 shows the EDS elemental mapping results for O, Cr, Fe, and N. The outer layers of the non-nitrided samples steam treated for 0.5 and 2 h contained less Cr and more Fe and O than that of the nitrided sample. Cr was detected from the inner layer toward the inside of the base material. The inner oxide layer also showed a loss of Fe. The missing Fe may have been consumed via outward diffusion toward the interface between the outer oxide layer and steam during exposure, forming a Fe-rich outer oxide layer.23,24) Unlike the Fe content, the Cr content of the inner oxide layer was nearly the same as that of the metal matrix. Therefore, the outer layer was Fe3O4, which is composed of Fe and O, and the inner layer was iron oxide containing alloying elements. The outermost layer of the nitrided sample was rich in nitrogen, contained almost no Cr, and contained little Fe and O. The N concentration on the outermost surface was high (450 nm), but it gradually decreased toward the inside of the sample. Nitrides could not be detected via XRD because the layer was very thin. This was because the outermost surface was a compound layer and nitrogen was dissolved in the lower part. Almost no Fe was detected on the outermost surface, and a large amount of Fe was detected in the outer layer, as in the non-nitrided samples; the inner layer had more Fe than the outer layer. Almost no Cr was detected in the outer layer, as in the non-nitrided samples. However, almost the same amount of Cr was detected from the inner layer to the base metal. These results suggested that the nitrided sample, from the outermost surface to the base metal, had a three-layer structure consisting of a nitride layer, an iron oxide layer, and a Cr-rich oxide layer.
According to the results of the current study and a previous paper,25) detailed oxidation mechanisms are proposed in Fig. 6. When the tool steel was exposed to the testing environment (Figs. 6(a) and 6(b)), an outer Fe3O4 oxide layer formed on the original sample surface through external oxidation (outward diffusion). Meanwhile, the inward diffusion of oxygen through the crystal lattice led to the formation of oxygen precipitates initially. Because the outer oxide layer consisted of small columnar oxide grains, both the outward diffusion of ionic Fe and the inward diffusion of oxygen remained available through the high-density network of grain boundaries. After longer exposure, the size of the internal oxide precipitates increased, and new internal oxides precipitated deeper into the base metal. Nitriding produced nitrogen compounds in proportion to the nitrogen concentration. Consequently, a compound layer formed on the outermost surface, and a diffusion layer formed inside the sample. The nitrogen-rich layer on the outermost surface consisted of nitride crystals, as seen in the SEM images of the surface (Fig. 3(c)), and nitrogen was diffused in the lower layer (Fig. 6(c)). When steam treatment was performed after nitriding, the phenomenon in Figs. 6(a) and 6(b) occurred. However, the nitrogen compound was not oxidized; it remained on the outermost surface, forming a nitrogen-rich layer.
We investigated the mechanical and tribological properties of the untreated, nitrided-only, steam-treated-only (500°C for 0.5 h), and hybrid-treated samples. Their surface roughness is shown in Fig. 7. The surface roughness Ra and Rz of the untreated sample were 8 and 53 nm, respectively. The surface roughness of the nitrided-only sample was similar to that of the untreated sample. The nitrided-only sample’s retention of its surface roughness was due to its lack of a rough compound layer. However, because the solid solution formed at a high temperature, distortion occurred in the nitrided-only sample. At atmospheric pressure, physical sputtering does not occur because of the small kinetic energy caused by the short mean free path.26,27) Here, the slight increase in surface roughness may have been due to the distortion of the sample. In addition, sample oxidation is hard to avoid in the case of atmospheric-pressure plasma. Thus, the nitrided-only sample may have oxidized, resulting in a slightly higher surface roughness. The surface roughness of the steam-treated-only sample was greater than that of the untreated sample; this was the effect of the oxides formed on the surface, such as Fe3O4. The higher surface roughness of the hybrid-treated sample over the steam-treated-only sample may have been due to the larger amount of nitrides or oxides formed on the surface of the former sample.
Figures 8(a) and 8(b) show the cross-sectional photographs of the nitrided-only and hybrid-treated samples, respectively. Both samples had diffusion layers but no compound layers. This finding was consistent with the XRD patterns of these samples. The depth of the diffusion layer of the nitrided-only sample was approximately 40 μm. However, the depth of the diffusion layer of the hybrid-treated sample was approximately 70 μm. The hybrid treatment increased the depth of the diffusion layer more than just atmospheric-pressure plasma nitriding. This was either because the material was further heated during superheated steam treatment after atmospheric-pressure plasma nitriding or because oxygen was solid dissolved.
Figure 9 shows the cross-sectional hardness of the nitrided-only and hybrid-treated samples. The surface hardness of the nitrided-only sample and the hybrid-treated sample was approximately 1220 HV and about 1000 HV, respectively. Therefore, an oxidized layer formed on the surface, or nitrogen diffused inside the sample. The hardened layers of the nitrided-only sample and the hybrid-treated sample were 40 and 80 μm, respectively. In the hybrid treatment, the sample was additionally heated via superheated steam treatment after atmospheric-pressure plasma nitriding. This further heating of the sample promoted the diffusion of nitrogen. Therefore, the heating effect of superheated steam treatment affected the formation of the hardened layer. The nitrided-only sample had a harder outermost surface, whose hardness decreased gradually with an increase in the distance from the surface. The outermost surface of the hybrid-treated sample was softer than its inner part, which had higher hardness. Hence, the shear resistance and real contact area of this sample would decrease during a friction test.
Figure 10 shows the friction coefficients of the untreated sample and the samples treated under various conditions. The friction coefficient of the untreated sample was approximately 0.65, which was higher than those of the nitrided-only and steam-treated-only samples. This result indicated the effects of surface hardness and surface roughness. The friction coefficient of the hybrid-treated sample was approximately 0.28, which was the lowest among all samples. In a previous report, the friction coefficient was low when an oxide film formed on the nitrided layer.15) The friction coefficient of the hybrid treatment is lower than that of the steam treatment, and the reasons are considered to be as follows. Due to the cross-sectional hardness distribution, the top surface is soft and the bottom is hard, which effectively reduces the shear force and the real contact area and controls the adhesive friction. The surface asperities of all samples may have reduced at sliding distances of 30–50 m. Therefore, unevenness due to surface roughness did not have to be considered before the friction test. Wear volume is inversely proportional to hardness;28,29) here, the nitrided-only and hybrid-treated samples had the lowest wear volume and thus reduced contact areas.
Figure 11 shows the cross-sectional profiles of the wear tracks of all samples. The untreated sample had a concave surface with a depth of 1.3 μm. However, the nitrided-only sample did not have such a surface, suggesting that it had high wear resistance. Its wear resistance was high because its surface hardness was twice that of the untreated sample, as shown in Fig. 9. The depth of the wear track on the steam-treated-only sample was almost the same as that of the untreated sample. This suggested the influence of three-body abrasive wear. The wear track on the hybrid-treated sample was shallower than that of the steam-treated-only sample. This may have been because nitrides and oxides are less likely to chip off and the occurrence of three-body abrasive wear is suppressed. Therefore, the hybrid treatment not only reduced the coefficient of friction but also improved the wear resistance of the sample.
We observed the sample surfaces after the friction and wear tests to clarify their wear mechanisms. Figures 12(a)–12(d) show the surface photographs of the worn samples. The load was mainly supported by the adhesion part of the asperity due to the lack of lubrication. This result was supported by the friction coefficients. Wear tracks were observed in all samples, and abrasive wear was particularly apparent on the surface of the untreated sample. Traces of surface adhesion were also identified. The color of the adhesion parts of the samples differed from the colors of the other parts. In particular, the untreated sample showed a significant color change. The temperature of a friction surface rises in the absence of lubrication.30) In this study, this high temperature was attributed to sample oxidation. Hence, the wear of the untreated sample was caused by abrasive and adhesive wear. The color of the wear tracks on the nitrided-only and steam-treated-only samples almost did not change. This was because nitrides and oxides have excellent oxidation resistance. Almost no trace of abrasive wear was observed in the nitrided-only sample; by contrast, the steam-treated-only sample showed almost the same level of abrasive wear as the untreated sample. Thus, the nitriding of tool steels effectively prevents abrasive and adhesive wear. The hybrid-treated sample had the narrowest slide mark and no abrasive wear. Therefore, the hybrid-treated sample exhibited excellent friction characteristics and wear resistance.
The surface structure and tribological and mechanical properties of an AISI H13 tool steel sample oxynitrided using atmospheric-pressure plasma nitriding and superheated steam treatment were studied. The results led to the following conclusions:
(1) The XRD peak intensity of Fe3O4 increased with the treatment temperature and treatment time for both steam-treated-only and hybrid-treated samples. The atmospheric-pressure plasma nitriding treatment had no effect on the formation of Fe3O4.
(2) An oxide layer separated into two layers was observed on the surfaces of the steam-treated-only and hybrid-treated samples. The outer-layer thickness of the non-nitrided samples did not increase with the increase in the treatment time, but their inner-layer thickness increased. The thickness of the oxide layer was almost the same between the nitrided and non-nitrided samples.
(3) Less Cr and more Fe and O were detected in the outer layers of the non-nitrided samples. Cr was detected from the inner layer toward the inside of the base material. Hence, the outer layer was Fe3O4, composed of Fe and O, and the inner layer was iron oxide containing alloying elements. The outermost layer of the nitrided sample was a nitrogen-rich layer containing almost no Cr and less Fe and O. Therefore, the nitrided sample, from the outermost surface to the base material, had a three-layer structure consisting of a nitride layer, an iron oxide layer, and a Cr-rich oxide layer.
(4) The surface roughness of the nitrided-only sample was similar to that of the untreated sample. The surface roughness of the steam-treated-only sample exceeded that of the untreated sample. This resulted from the oxides formed on the sample surface, such as Fe3O4.
(5) The nitrided-only and hybrid-treated samples had diffusion layers but no compound layers. However, the hybrid-treated sample had a deeper diffusion layer than the nitrided-only sample.
(6) The hardened layers of the nitrided-only and hybrid-treated samples were 40 and 80 μm, respectively. The nitrided-only sample had a harder outermost surface, whose hardness decreased gradually with an increase in the distance from the surface. The outermost surface of the hybrid-treated sample was softer than its inner part, which had higher hardness.
(7) The hybrid-treated sample had the lowest coefficient of friction, which was approximately 0.28.
(8) The wear track on the hybrid-treated sample was shallower than that on the steam-treated-only sample. Therefore, the hybrid treatment not only reduced the coefficient of friction but also improved the wear resistance of the material.
(9) The hybrid-treated sample had the narrowest slide mark and did not exhibit any abrasive wear.
This work was financially supported by a research grant from the NSK Mechatronics Koudoka Gijyutu foundation.
