The Oxide Layer of 10Mn5 Medium Manganese Steel for Wear Protection in High-Temperature Friction during Hot Stamping

Abstract

A custom-designed high-temperature sliding-on-sheet-strip (SOSS) tribo-tester was used to simulate the high-temperature friction process of 10Mn5 medium manganese steel bare plate under actual hot stamping conditions. To reveal its high-temperature friction mechanism in the hot forming process, the high-temperature friction behavior of 10Mn5 steel and 22MnB5 steel was compared. The scanning electron microscope (SEM), energy spectrum analyzer (EDS) and X-ray diffractometer (XRD) were used to investigate the structure of the oxide layer, composition of physical phase, wear surface morphology and elemental composition. The results show that the average coefficient of friction of 10Mn5 steel is 12.7% lower than that of 22MnB5 steel. The cross-section of both steel consists of an oxide layer, an alloying element-rich layer and the matrix. The oxide layer of 10Mn5 steel is mainly composed of Fe₃O₄, approximately 63.7%, while that of 22MnB5 is mainly composed of Fe₂O₃, approximately 66.9%. The complete and less flaking scale of 10Mn5 steel provides good wear protection, and the mechanism is abrasive with slight adhesive wear. Meanwhile, oxide particles and fragments are embedded in the 22MnB5 surface thus increasing the wear, and the mechanism evolves into severe abrasive and adhesive wear. The difference in the mechanism between the two steels is mainly caused by different austenitizing temperatures, which for 10Mn5 is lower than 22MnB5, about 100 °C. This makes the thermal stress of 10Mn5 from the temperature difference between the furnace and the environment not enough to break the scale and decrease abrasion.

1. Introduction

Modern automotive anti-collision reinforcements, such as door impact beams, A-pillars and B-pillars, are generally produced by hot stamping [1,2] to overcome the problems of poor formability, serious springback and higher shaping pressure associated with the traditional cold-forming process [3,4,5]. The most widely used hot-stamping steel is 22MnB5, but 22MnB5 also faces problems such as low plasticity (elongation of the industrial product below 7%) and severe oxidation [6]. To solve these problems, Dong et al. [7,8,9] successfully developed the third-generation automobile medium-Mn steel and hot stamping technology. The strength of medium manganese steel parts [10,11,12] manufactured by warm forming can reach 1400 MPa and the elongation can reach 11%. In addition, the austenitizing temperature is reduced by 100 °C–150 °C due to the higher Mn content [13,14], thus effectively alleviating the problem of serious high-temperature oxidation.

The stability of austenite is improved by increasing the content of Mn elements in medium manganese steels. The austenite stability is enhanced to obtain ultra-fine ferrite and 20%–50% ferrite by Austenite Reversion Treatment (ART). Compared with traditional boron steel, medium manganese steel such as 10Mn5 usually has better mechanical properties, the product of strength and ductility, which can reach 30%. The reasons for this can be summarized by the following: (1) Medium manganese steels retain more austenite grains, and austenite can transform to martensite during deformation, which contributes to work hardening and improves mechanical properties, especially elongation [15,16]. (2) As the austenitizing temperature in hot forming decreases, the primary austenite grain size (PAGS) decreases [17]. This can refine the microstructure after quenching, including the refinement of the martensitic layered organization and retained austenite grains, thus improving strength and toughness [18].

The high-temperature oxidation and wear behavior of hot stamping steel has been widely studied. Hardel et al. [19] investigated the friction behavior during hot forming and found that the formation of a continuous Fe₃O₄ oxide layer shape on the surface of 22MnB5 steel, as well as a work-hardened sub-surface layer, led to a 50% reduction in the coefficient of friction (COF). Yanagida [20] found that the thickness of the scale of 22MnB5 steel increased with temperature and the average COF increased also. Ghiotti et al. [21] found that the wear mechanisms of 22MnB5 steel during hot forming were adhesive wear and abrasive wear. Guo et al. [22] investigated the effect of high-temperature oxidation on the subsurface microstructure and properties of medium manganese steel. Results show that the inner oxide layer consists of Mn-Cr, Fe-Cr, Mn-Al and Fe-Si spinel at high oxidation temperature.

Hot stamping involves temperature and phase fields. During the high-temperature friction process, there exist a greater number of uncontrollable factors. The process necessarily involves the effects of deformation, adhesion and oxidation on the friction behavior, making it more difficult for the test to be close to reality. However, most of the above-mentioned tests just heat the sample, mold to a specified temperature and start the test, which is not square with the facts. The reasons mainly include the following: (1) The austenitizing process of the sheet before forming is ignored, and the austenite will transform into martensite during the subsequent pressure quenching, thus affecting the friction behavior. (2) Under practical production conditions, as the number of stampings increases, the mold temperature will increase too, resulting in a gradual decrease in the temperature difference between the mold and sheet. The cooling rate even falls below the critical cooling rate. This makes austenite difficult to fully transform into martensite. It is necessary to consider the state of the mold in the design of the high-temperature friction test program.

Compared to 22MnB5 boron steel, the hot forming process for 10Mn5 medium manganese steel is an emerging technology. The behavior and mechanism of high-temperature friction during forming of the steel is not yet known due to changes in chemical composition, microstructure and reduced austenitizing temperature. In this study, a custom-designed high-temperature sliding-on-sheet-strip (SOSS) tribo-tester is used to evaluate the high-temperature friction performance of 10Mn5 steel and 22MnB5 boron steel during the hot stamping process and find the differences in the high-temperature friction behaviors of the two steels. This result will be helpful to promote the development of hot stamping of medium manganese steels.

2. Materials and Methods

The two sheet steels used in this study are industrially produced. The initial microstructure of as received 10Mn5 steel consists mainly of austenite and ferrite, while 22MnB5 steel consists mainly of pearlite and ferrite (Figure 1).

To cool the sample rapidly, a Φ8 mm circular hole to fix the sample was machined at the front end using a drill machine. To fully simulate the production process, quenched and tempered H13 hot work die steel was used for the friction pair. The surfaces of the mold were wet sanded using silicon carbide papers from 320 grit down to 1000 grit. The surface roughness of the steel plates was maintained in the as received condition (at approximately 0.14 μm). A TR200 roughness meter was used to measure the surface roughness values of the samples. Each sample was measured continuously 5 times, 3 times free to measure and 2 times in the mutually perpendicular direction. Then, the 5 point measurement average was taken. The chemical composition and surface roughness of the sample and friction pair are shown in

Table 1. Chemical composition and roughness of 10Mn5 steel, 22MnB5 steel and H13 steel.

Sample		Mass Fraction/%										Ra/μm	Hardness
	C	Mn	Si	P	S	Ni	Al	B	Mo	Cr	Fe
10Mn5	0.10	5.00	0.13	0.008	0.002	-	0.030	-	-	0.35	Bal.	0.164 ± 0.008	239 ± 11HV0.2
22MnB5	0.26	1.24	0.25	0.021	0.005	-	0.040	0.0030	-	0.10	Bal.	0.141 ± 0.007	182 ± 23 HV0.2
H13	0.39	0.38	0.89	0.015	0.015	0.32	-	0.0030	1.35	4.83	Bal.	0.213 ± 0.005	52.4 ± 1HRC

. The size of the samples is shown in Figure 2c.

The SOSS tribo-tester was used to conduct unidirectional wear tests on 10Mn5 and 22MnB5 steel. This equipment consists of a heat treatment system, a lever loading system, a friction system and a sample moving system, as shown in Figure 2a. The machine is also equipped with a cooling channel, which can cool the sample rapidly, to simulate the actual hot rolling process.

The heat treatment system can heat the sample with a furnace that has a total length of 1000 mm. One end of the furnace is provided with a square hole to allow the sample to move out of the chamber. The furnace is designed with sectional mode heating to ensure the temperature of the chamber. The furnace is also equipped with heat-resisting ceramic balls to support high-temperature softening sample movement.

The loading system uses a lever to apply force to the sample through the upper mold. The contact between the lever and the upper mold is a ball and cambered surface structure, so the force is always normal during the test, as shown in Figure 2d for parts 6, 7, 4, 9 and 12.

The sample moving system consists of a stepper motor kit (including a controller, driver and driver power switch) and a liner guide slide (including a ball screw pair, base and sliding block). The motor is connected to the liner guide slide via coupling. The speed and stroke of the block can be programmed for the moving system.

Before the wear test, the samples are heated to setting temperature in the furnace for 2 h and soaked at complete austenitizing temperature for 5 min to obtain homogeneous austenite. The samples are then pulled from the furnace to the underside of the mold using a stepper motor at a constant speed of 20 mm/s. The transfer process lasts about 3 s. Then, the lever applies a certain load (5N). The samples make one fast pass through the mold, and the distance is 200 mm. The friction process lasts 10s. During the transfer process, the sample experiences a temperature reduction of approximately 100 °C (K-type thermocouple, Nanpu Meter Factory, Shanghai, China), while a decrease of 200 °C was observed during the friction process. One sample and 1 mold are used in each run. The cooling water in the mold ensures the completion of the quenching process of the sample. The COF can be determined based on the data recorded by the force transducer attached to the sample moving system through Equation (1),

$$\mu =\frac{F}{2P}$$

(1)

where the P (N) is the normal load applied by the lever and the F (N) is the tension measured by the force transducer. The average COF can be determined through Equation (2)

$${\mu }_A=\frac{1}{L_s}{{\displaystyle \int }}_{L_0}^{L_S}\mu \mathrm{d}L$$

(2)

where the L_s (mm) is the total wear distance and the L₀ (mm) is the distance entering the steady wear stage.

The specific parameters for the wear test of 10Mn5 and 22MnB5 are listed in

Table 2. High-temperature friction test-specific parameters.

Sample	10Mn5	22MnB5
Austenitizing temperature/°C	830	930
Soaking time/min	5
Normal load/N	270
Speed/(mm·s−1)	20
Mold temperature/°C	75

.

The worn surface and subsurface of the samples were characterized using SEM (ZEISS Sigma300 equipped with an EDS produced by Oxford Instruments plc.) and the operation voltage is 20 kV. To protect the oxide layer, samples used to characterize the subsurface were cut up and inset into hot mounting resin to reveal the cross-sections.

The spallation of the scale during the experiment was ground into particles smaller than 20 μm and analyzed by a Rigaku SmartLab X-ray diffractometer using CoKα radiation. The scanning speed is 10°/min and the scanning angle 2θ ranges from 20–100°. The relative amount of the phases was calculated by the adiabatic principle (Equation (3)) [23],

$$W_i=\frac{I_i}{K_i{\sum }_{i=1}^n\frac{I_i}{K_i}}$$

(3)

where the I is the intensity of the strongest diffraction peak of each phase, K is the K-value (RIR) of each phase and n is the number of phases of the sample. The strongest diffraction peak of each phase was calculated by JADE and the K-value is obtained from the PDF card. The oxide phases include FeO (PDF #77-2355), Fe₂O₃ (PDF #73-2234) and Fe₃O₄ (PDF #89-0688).

3. Results and Discussion

The average COF and COF for the high-temperature friction test of the samples are shown in Figure 3. Each group of friction tests is repeated three times, the error does not exceed 5%. The average coefficient of friction (Figure 3b) of 10Mn5 steel is 0.295, while for 22MnB5 steel it is 0.302, which is 12.7% lower than that of 22MnB5 steel. The difference in the COF (Figure 3a) for high-temperature between 10Mn5 steel and 22MnB5 steel in hot stamping is mainly seen in the following two aspects. Firstly, the COF of the medium manganese steel is lower than that of the 22MnB5 steel during the sliding process. Secondly, the COF curve of the 10Mn5 steel is smooth as the experiment proceeds, while the value of the COF of the 22MnB5 steel changes intensely. The COF fluctuates the most at the initial stage of the friction, with a range of more than 0.08, and continues to fluctuate above and below 0.35 as the sliding distance increases, with a fluctuation of more than 0.06. The fluctuation of the COF is not a transient process, and the sliding distance between the appearance of the peak and the trough can reach 15 mm. The fluctuation of the COF is not caused by the cold welding and broken bonds during the adhesive wear process, but by the destruction of part of the scale during the friction process. When the oxide layer of the mold and the sample are in contact, the oxide layer acts as lubrication, avoiding the direct contact between the mold and the matrix of the sample, reducing adhesive wear, and reducing the shear stress on the worn surface. When the oxide layer is broken, the mold and the sample are in direct contact, increasing the coefficient of friction.

Figure 4 shows the surface scale after austenitizing. The sample would then be transformed a position situated beneath of the mold to start the friction test. The oxide layer forms in the austenitizing process. As shown in Figure 4a, the surface of 10Mn5 steel is complete, and the attachment of the oxide layer is good. In contrast, 22MnB5 has poor surface quality with oxide bulges (Figure 4b). The scale will be broken into fragments under subsequent friction stress, which leads to excessive surface wear.

When metals are exposed to air at high temperatures, an oxide film is generated on the surface. This film plays a crucial role in high-temperature friction. During the high-temperature friction process, a continuous and intact oxide layer will provide good wear protection by reducing the contact area between metals and decreasing the sliding resistance [24]. However, the broken oxide layer may adhere to the metal and mold surface and even cause three-body abrasive wear, which aggravates the wear condition. Therefore, the oxide layer peeling mechanism in terms of oxide thickness, composition and surface morphology is studied. The morphology and thickness of the oxide layer were characterized by observing the cross-section of the samples using SEM. The composition of the oxide layer spalling from the surface of the sample was analyzed by XRD, and the wear surface morphology of the sample and the mold and the composition of the adhesions were analyzed by SEM-EDS.

Figure 5 shows the SEM images of the cross-section of 10Mn5 steel and 22MnB5 steel, and the composition of the area marked by the red box in the figure is analyzed by EDS. The formation of the oxide layer and alloying element enrichment layer was observed in all samples near the surface. However, the thickness of the oxide layer of 10Mn5 and 22MnB5 steel is 15.75 μm and 27.54 μm, respectively. The thickness of the alloying element enrichment layer of 22MnB5 steel is 2.31 μm, mainly composed of Si and O (Figure 5b). The thickness of the alloying element enrichment layer of 10Mn5 steel is 4.21 μm, mainly composed of Cr and O (Figure 5a). This proves that 10Mn5 steel can effectively mitigate the serious problem of high-temperature oxidation. Most of the Mn in 10Mn is distributed in the matrix, while regularities of distribution of that in 22MnB5 are not obvious.

Generally, the thickness of the stable layer depends mainly on the temperature, oxygen partial pressure and the rate of ion diffusion [25]. The addition of Mn can significantly reduce the austenitizing temperature of medium manganese steel. The austenitizing temperature of 10Mn5 steel is only 830 °C, which is 100 °C lower than that of 22MnB5 steel. At lower temperatures, the reaction rate will be reduced, resulting in a slower rate of oxide growth. The slower growth rate results in a thinner oxide layer. Therefore, the lower austenitizing temperature leads to a thinner oxide layer thickness [26]. Fe combines with O atoms in the air to form FeO, Fe₂O₃ and Fe₃O₄, while the atoms that have relatively slow diffusion rates such as Si and Cr will form a compact layer between the scale and matrix. The layer can effectively oppose iron and manganese which have similar atomic radius, atomic motion and prevents further oxidation at high temperature. Due to the higher alloying element content compared with 22MnB5, especially Cr, the alloying enrichment layer of 10Mn5 steel is twice as thick as that in 22MnB5 steel; this also reduces the thickness of the oxide layer. Thinner oxide layer thickness is beneficial to increase critical compressive strain during cooling.

The composition of the oxide layer will cause obvious changes in the wear mechanism. Generally, Fe will form three types of oxide at high-temperature [27,28]. FeO forms above 570 °C, Fe₃O₄ forms between 200 °C and 570 °C and Fe₂O₃ forms below 200 °C. Fe₂O₃ and Fe₃O₄ have a compact structure and small lattice constant, which make it difficult for Fe ions to pass through and prevent further oxidation. FeO is porous and has a simple lattice structure. There are vacancies in the lattice, and Fe ions can easily diffuse through the FeO layer and intensify the oxidation of Fe. In addition, it has been shown [29] that Fe₃O₄ has better wear resistance than FeO and Fe₂O₃. This is because the cubic crystal structure of Fe₃O₄ is easier to shear.

To further study the oxidation behavior of the sample during the austenitizing process and its influence on the friction process, the composition of the scale of the sample was analyzed by XRD. Figure 6 shows that the oxide layers of 10Mn5 and 22MnB5 consist of FeO, Fe₂O₃ and Fe₃O₄. The relative contents of each phase of the oxide were calculated by the intensity of the strongest diffraction peak using the adiabatic principle (

Table 3. The fraction of each phase (calculated by the adiabatic principle) in the surface oxide of the sample (wt.%).

Sample	FeO	Fe2O3	Fe3O4
10Mn5	24.3 ± 0.9	12.0 ± 0.5	63.7 ± 2.5
22MnB5	10.6 ± 0.6	66.9 ± 3.5	22.4 ± 0.9

). The results show that the scale of 10Mn5 is mainly composed of Fe₃O₄ with a small amount of Fe₂O₃ and FeO, and the higher Fe₃O₄ is beneficial to the wear resistance of 10Mn5. Compared with 10Mn5, the scale of 22MnB5 is mainly composed of Fe₂O₃, which has poor wear resistance compared with Fe₃O₄. The oxide of 22MnB5 contains only 10.6% FeO, which is only half of the FeO content in the oxide layer of 10Mn5. Generally, FeO forms above 570 °C and the content increases with the temperature. The austenitizing temperature of 22MnB5 is much higher than that of 10Mn5 Steel. However, the FeO content in the oxide of 22MnB5 is lower. This is because, as the friction progresses, the surface defects of the sample increase. The FeO is labile, and the increase in the surface defects is beneficial to the decomposition of FeO and diffusion of Fe atoms from the surface. Fe atoms will fully combine with O atoms in the air to form Fe₃O₄, which has a higher oxygen content. Compared to 10Mn5 steel, the wear of 22MnB5 steel is more serious, and a large amount of FeO was transformed to Fe₃O₄ during the friction process, so the FeO content in the oxide layer of 22MnB5 steel after the friction test was lower than that of 10Mn5.

As shown in Figure 7a, 10Mn5 steel wear is mild, and the oxide layer is relatively complete and has less exfoliation. The black adhesion indicates that adhesive wear occurred, which is due to the adhesion that occurred on the local contact surface, and the adhesion point subsequently sheared during the sliding process. This leads to the transfer from the sample to the mold surface or damage to the sample surface. The oxide layer can decrease the tendency to develop adhesive wear and lower shear stress between friction pairs, thus effectively protecting the matrix and reducing COF. In addition, it reduces plastic deformation of the sample surface [30,31]. In the high-temperature test, all black adhesion was adhered to the oxide layer, indicating that the oxide on the surface of 10Mn5 steel protects the matrix well and reduces the wear. Figure 7b shows the wear surface of the 22MnB5 steel. Compared with 10Mn5 steel, the image shows a large area of the oxide layer exfoliating and can hardly cover the matrix. The adhesion produced during the wear process is only partly adhered to the oxide layer. This means direct contact occurred between the sample and mold, which leads to severe wear.

SEM images show the worn surface clearly. As shown in Figure 8a, only a small number of cracks appear on the oxide layer. EDS results (Figure 8b) show that the surface mainly consists of Fe and O, which indicates that the oxide can prevent the matrix from contacting the mold directly. When the sample is dragged out of the furnace and contacts cold air, the oxide layer will generate thermal stress. However, the austenitizing temperature of 10Mn5 steel is only 825 °C. The temperature difference between the sample and air is not enough to break the scale. Correspondingly, the austenitizing temperature of 22MnB5 steel reached 930 °C, and a larger temperature difference led to larger thermal stress. Therefore, when the sample was dragged out of the furnace, it is obvious that the scale bulged, and finally broke and spalled off during the wear process. The scale spalling was crushed during the sliding process, generating a large number of oxide particles and fragments embedded in the sample. Unlike the high-temperature reciprocating wear test, the scale that broke up under unidirectional friction conditions was difficult to compact to form a smooth enamel layer and provide wear protection. Instead, the oxide spalling was crushed during sliding, producing a large number of oxide particles and fragments embedded in the matrix (Figure 8c). During the wear process, things will not only cause adhesive wear but also scratch the surface of the material and exacerbate wear, leading to an increase in COF. Figure 8e shows a large amount of oxide spalling off from the surface of 22MnB5 steel, with the surface of the sample covered with a large number of oxide particles and fragments. The results of the EDS analysis are in Figure 8d. They indicate that the point analyzed contains both matrix elements and O, which means the oxide layer has been crushed into oxide fragments, making it difficult to completely cover the surface of 22MnB5 steel. The surface of 22MnB5 steel cannot be fully covered.

In order to clarify the effect of the oxide scale on the high-temperature friction behavior, a schematic of the model of two different steels’ oxidation during the process is shown in Figure 9. The oxide and alloying element enrichment layer will form on the surface of the metal, as shown in Figure 9a. The sample was subjected to thermal stress when it was dragged from the furnace and in contact with the lower temperature mold (about 75 °C) and could not expand freely. The coefficient of thermal expansion of iron oxide varies largely between 600–800 °C [32], with 1.7 × 10⁻⁵/°C for FeO, 1.25 × 10⁻⁵/°C for Fe₂O₃ and 1.5 × 10⁻⁵/°C for Fe₃O₄. The oxide layer of 10Mn5 steel is mainly composed of Fe₃O₄, the coefficient of which is close to that of Fe (1.46 × 10⁻⁵/°C). The austenitizing temperature of 10Mn5 is also lower than that of 22MnB5. This means the scale of 10Mn5 is subjected to lower stress. On the contrary, larger temperature differences and the scale mainly composed of Fe₂O₃ led to higher thermal stress applied on the material. Higher stress makes the bond strength between the matrix and oxide layer weak and causes bulges (Figure 9b). The bulges will be easily broken during the wear process, which will lead to wear speeding up.

The wear mechanisms of 10Mn5 and 22MnB5 steel are further verified by the comparison of the morphology and composition of the H13 mold surface after the wear test. SEM images (Figure 10a,c) show different levels of black adhesion of the two steels. This indicates that oxides formed in samples at high temperatures based on the results of the EDS analysis, and the oxides on the surfaces of the samples were transferred to the mold surface during sliding, indicating that adhesive wear occurred in both samples. Due to the protection of the oxide layer, the surface of the mold in the wear process of 10Mn5 steel has less adhesion, which leads to less tendency to adhesive wear. In addition, the oxide layer spalling was crushed during sliding and turned into oxide particles, leading to abrasive wear. The 22MnB5 steel possesses an extensive oxide layer on its surface, which serves a protective function. However, during the sliding process, the oxide layer undergoes extensive delamination and is fragmented into Fe₂O₃ oxide particles. These particles are transferred from the surface of the sample to that of the mold, exacerbating the wear phenomenon. At this stage, the wear mechanism transforms into a combination of abrasive wear and severe adhesive wear.

In conclusion, the SOSS tribo-tester was used to compare the high-temperature friction properties of the 10Mn5 steel and 22MnB5 steel, and SEM and XRD were also used to analyze the samples and H13 friction pair. The results show that 10Mn5 steel can form a complete and less flaking scale during the unidirectional friction process, which contributes to excellent high-temperature wear resistance compared with 22MnB5 steel.

4. Conclusions

In this study, the high-temperature friction behavior of 10Mn5 medium manganese steel during a hot stamping process is studied by the SOSS tribo-tester, and the comparison between 10Mn5 and the conventional hot forming 22MnB5 steel is analyzed. The main results are as follows.

(1)

During the high-temperature wear test, the COF of 10Mn5 steel is significantly lower than that of 22MnB5 steel, and the wear mechanism is both adhesive wear and abrasive wear, but the wear condition of medium manganese steel is slighter.

(2)

The thickness of the oxide layer of 10Mn5 steel is 15.75 μm, which is thinner than that of 22MnB5 teel (27.54 μm). This is due to the higher content of Mn and Cr in 10Mn5 steel. On the one hand, the Mn element lowers the austenitizing temperature, and on the other hand, the Cr element forms an alloying element with high-temperature oxidation resistance.

(3)

The oxide layer of 10Mn5 is mainly composed of Fe₃O₄, while that of 22MnB5 is mainly composed of Fe₂O₃.The differential coefficient of thermal expansion between Fe₂O₃ and Fe is more significant in comparison to that between Fe₃O₄ and Fe_.. The difference leads to the bulge appearing on the steel surface.

(4)

The austenitizing temperature of 10Mn5 steel is much lower due to the addition of Mn. In consequence, the thermal stress from the temperature difference between the furnace and the environment is not enough to break the scale. An intact scale protects the matrix during friction and reduces the coefficient of friction.