Influence of aging treatment on mechanical properties and wear resistance of medium manganese steel reinforced with Ti(C,N) particles

In this study, the hot rolled medium manganese steel containing titanium was solution treated at 1,000 °C and followed by aging treatment at 500, 550, and 600 °C. The influence of aging treatment on mechanical properties and wear resistance of medium manganese steel reinforced with Ti(C,N) particles was investigated. It was found that the matrix of medium manganese steel was austenite. The austenite grain size was refined, and Ti(C,N) particles were precipitated after aging treatment. Compared to that of the as-hot rolled sample, the initial hardness of 500 °C aged sample increased by 9.5% to 312.86 HV, whose impact energy was more than doubled to 148.5 J. As the aging temperature raised to 600 °C, the initial hardness changed slightly. However, the impact energy dropped significantly to 8 J due to the aggregation of Mn at the grain boundaries. In addition, the main wear mechanisms of the samples were fatigue wear and abrasive wear. It was worth noting that 500 °C aged sample exhibited the best wear resistance under a 300 N applied load, whose wear loss was just half of the as-hot rolled sample. The relationship between wear loss and mechanical properties indicated that the wear resistance of medium manganese steel was independent of the initial hardness. The large difference in the wear resistance was predominately due to the outstanding work hardening ability of 500 °C aged sample, whose strengthening mechanisms were contributed from transformation induced plasticity (TRIP) effect, dislocation strengthening, twinning induced plasticity (TWIP) effect, and precipitation strengthening.


Introduction
Hadfield steel is typical wear-resistant steel because it has excellent work hardening capacity under high applied load conditions [1][2][3]. During the wear process, the material's surface can undergo phase transformation to produce a hardened layer with high hardness to resist wear. In contrast, the inner part of the material still maintains high toughness. Therefore, Hadfield steel has been widely used in the wear materials of engineering fields, including mining, steelmaking, railroad, and cement products [4][5][6][7][8]. References [9,10] have mainly focused on high manganese steel with elements of 1.0-1.4 wt% C and 10-18 wt% Mn achieved a better combination of high ductility and strength and performed favorable wear resistance under severe impact conditions. However, it was found that in the practical application, the high manganese steel often faced the challenge of poor wear performance due to the shortcoming of insufficient work hardening capacity under mild service conditions [11,12]. In the face of this defect, the researchers believed that changing the Mn content and favourable heat treatment could obtain ameliorative work hardening properties. References [13][14][15][16] have been reported on the medium manganese (8-12 wt%) steel. In addition, the mechanical properties and wear resistance were improved by adding Ti, Cr, Mo, and V elements into the steel to obtain the solution and precipitation strengthening effects [17][18][19][20][21][22]. Thus, the alloying, heat treatment, work hardening, and wear performance of medium manganese wear-resistant steel have been systematically studied [23][24][25][26][27]. The results showed that the deformation induced martensitic transformation of medium manganese steel had a self-strengthening effect under medium or low applied loads, and the medium manganese steel had better wear resistance than high manganese, bainitic, and martensitic steels.
High melting points and high hardness characterized the Ti(C,N) particles. When fine Ti(C,N) particles were dispersed in the matrix, they played a role in the precipitation strengthening effect. Therefore, they were widely used to improve the properties of materials [28,29]. In recent studies, many researchers have analyzed the effect of Ti(C,N) particles additive on the wear-resistant steel. Prava Dalai et al. [30] indicated that adding hard TiC particles into the austenitic manganese steel and secondary processing were beneficial for elastic modulus, hardness, and wear resistance. Srivastava and Das [31][32][33] demonstrated that austenite matrix with Ti(C,N) and (Ti,W)C precipitates could effectively protect the sliding surface. Huang et al. [34] studied low alloy abrasion-resistant steel reinforced with self-generated TiC particles developed to improve its wear performance without increasing its hardness, and the results showed that the carbide precipitates were significantly helpful in resisting abrasive embedding and micro-cutting during abrasive wear. Since the precipitated phase can improve the mechanical properties, protect the matrix in the wear process, and reduce the generation of cracks, the precipitation strengthening method may be the breakthrough point in improving the wear resistance.
In the present study, we added the Ti element into medium manganese steel, and the precipitates were obtained through the solution and aging treatment. The effects of aging treatment on mechanical properties and wear resistance of medium manganese steel reinforced with Ti(C,N) particles were investigated. Moreover, the influences of precipitates' morphology, distribution, and quantity on the microstructure and mechanical properties were analyzed, the relationship between mechanical properties and wear mechanisms was described, and the influences of strengthening mechanisms on the work hardening were explicated.

Materials and methods
The nominal chemical compositions of medium manganese steel were 0.98 wt% C, 7.39 wt% Mn, 0.67 wt% Si, 2.0 wt% Cr, 0.3 wt% Mo, 0.17 wt% V, 0.1 wt% Ti, and balanced Fe. Cr and Mo elements were added to medium manganese steel to modify the strength loss caused by the decrease of the Mn element [35]. V and Ti elements refined the grain and inhibited the distribution of carbides along austenite grain boundaries [36,37]. Significantly, the addition of Ti element produced Ti(C,N) particles, which can bring significant precipitation strengthening effect on medium manganese steel. The medium manganese steel was melted in a vacuum melting furnace (ZG-100B, Shanghai Chenhua Science Technology Co., Ltd., China) with a capacity of 100 kg, and then forged into 100 mm × 100 mm × 120 mm ingots. The sample was heated to 1,250 °C for 2 h, and then rolled for 7 passes into 12 mm thick plates, followed by air cooling. The air-cooled plates were austenitized at 1,000 °C for 1 h, followed by water quenching to room temperature. Then the quenched samples were aged at 500, 550, and 600 °C for 1 h, followed by air cooling.
The reciprocating sliding wear tests were performed on medium manganese steel using a highly integrated multifunctional friction and wear test machine (MFT5000, Rtec Instruments, USA). A tungsten carbide (WC) ball with a diameter of 6.35 mm and hardness of 90 HRA (~1,400 HV) was selected as the counterbody because its hardness was much higher than that of the test's material. In the wear test process, the WC www.Springer.com/journal/40544 | Friction ball was pressed on the flat plate, and the reciprocating sliding was executed under the normal loads of 100, 200, and 300 N, a stroke length of 6 mm, and a frequency of 5 Hz. The test time was 1,800 s, and each sample was subjected to three wear tests. Before the wear test, each specimen was ground and mechanically polished to achieve a mirror finish. After the wear test, the wear loss was determined as a volume loss by measuring the size of the wear track with a confocal laser scanning microscope (MFT5000, Rtec Instruments, USA).
Metallographic samples (10 mm × 10 mm × 15 mm) were cut using electro-discharge machining (EFH86S, Shanghai Esuntek Machinery Co., Ltd., China) and etching the samples with 20% NaHSO 3 after being polished. A Vickers hardness tester (HVS-1000, Shanghai Yizong Precision Instrument Co., Ltd., China) measured the hardness of medium manganese steel at a load of 0.98 N and holding time of 15 s, including initial hardness, wear surface hardness, and crosssection hardness. The impact energy test was carried out according to GB/T229-2007, and the impact sample size was 10 mm × 10 mm × 55 mm. V-notch was opened on the drawing machine, and then polished the surface clean. For the measurement accuracy of the test, the values were determined by averaging the results of multiple tests. The optical microscopy (OM; DM1750M, Leica, Germany) and field emission scanning electron microscopy (FESEM; ULTRA 55, ZEISS, Germany) analyzed the microstructures and wear mechanisms. Gwyddion software v2.58 was used to measure the three-dimensional (3D) topography of the wear surface, involving surface roughness (R a ), depth, and width of scratches. The field emission transmission electron microscopy (FETEM; F200, JEOL, Japan) observed the strengthening mechanisms.

Microstructure
The optical micrographs (Figs. 1(a)-1(d)) show the microstructures of the as-hot rolled sample and 500-600 °C aged samples. The microstructural constituents of the four samples were composed of austenite and a small number of precipitates. It was measured that the average grain sizes of austenite were 39.5, 28.4, 29.7, and 30.2 μm. It was evident that the difference in the grain size among the aging treated samples was insignificant. However, austenite of the aging treated samples was refined and more equiaxed than that of the as-hot rolled sample. It can be inferred that the recrystallization of austenite occurred at 1,000 °C, and the formation of precipitates inhibited the growth of austenite grains. However, the diffusion coefficient of carbon increased with the increase in aging temperature, resulting in a significant difference in precipitates. The precipitates of the as-hot rolled sample were mainly concentrated at the austenite grain boundaries. In contrast, the precipitates of 500 °C aged sample were smaller and precipitated inside the austenite grain. With the increase in aging temperature, the precipitates gradually precipitated at the austenite grain boundaries. When the temperature raised to 600 °C, the number of precipitates at grain boundaries increased significantly. Although the aged samples had similar grain sizes, the precipitation behavior was different. Therefore, the effect of aging treatment on the precipitation behavior was studied in the following two paragraphs.
In order to study the influence of aging treatment on the precipitation behavior, the distributions, morphologies, and element contents of precipitates were analyzed by the scanning electron microscopy (SEM; ULTRA 55, ZEISS, Germany)-energy dispersive spectroscopy (EDS), and the characteristics of precipitates were determined. According to the EDS results ( Fig. 2), it was confirmed that the particles were Ti(C,N) particles. As shown in Figs. 2(a)-2(d), the Ti(C,N) particles with different morphologies were uniformly distributed on the matrix. The Ti(C,N) particles of the as-hot rolled sample were mainly triangular, columnar, and polygonal with large sizes. After aging treatment at 500-600 °C, the Ti(C,N) particles were circular, rectangular, and refined. Significantly, more fine particles were observed in the 500 °C aged sample. The fractions of precipitated particles of the four samples estimated by Image Pro Plus software (v6.0, Media Cybernetics, USA) were 0.36%, 0.48%, 0.46%, and 0.40%. The size distributions of precipitates are illustrated in Fig. 3.
It is indicated that aging treatment had a significant effect on the precipitation behavior of Ti(C,N) particles, especially the Ti(C,N) particles with a diameter of 0-1 μm. For the aged samples, some Ti(C,N) particles were remelted into the matrix during the solution treatment at 1,000 °C, and they were precipitated in the form of fine particles during aging at 500-600 °C, changing from large particles such as irregular triangles and polygons to regular round fine particles. Based on the statistics, the fine precipitates (0-1 μm) occupied a fraction of 44% in the as-hot rolled sample, and the large precipitates (≥ 1 μm) accounted for a fraction of 56%. In contrast, for 500 and 550 °C aged samples, the fraction of coarse precipitates decreased, and the number of fine precipitates significantly increased. However, when the aging temperature reached 600 °C, the small precipitates dissolved due to Ostwald ripening phenomenon, and the large precipitates continued to grow, leading to a significant increase in the size of precipitates. It was widely accepted that the strengthening mechanism of the  www.Springer.com/journal/40544 | Friction second phase was Orowan mechanism [38,39], which was inversely proportional to the size of the hard particles and directly proportional to the precipitation amount. Therefore, it can be inferred that the aging strengthening effect contributed from Ti(C,N) particles of 500-550 °C aged samples is better than that of the as-hot rolled and 600 °C aged samples. The effect of aging treatment on mechanical properties will be studied in Section 3.2.

Mechanical properties
The hardness and impact energy are the best indicators for evaluating the material properties that affect the service life and wear efficiency of wear-resistant products [40][41][42]. Figure 4 shows the hardness and impact energy of medium manganese steel after different heat treatment. As demonstrated in Fig. 4(a), the hardness of the as-hot rolled sample was the lowest, and those of the samples aged at the temperature of 500-550 °C increased slightly, which was probably attributed to the increase in the number of Ti(C,N) precipitates. However, for 600 °C aged sample, a large number of carbides precipitated at the grain boundaries rather than intragrain ( Fig. 1(d)), leading to smaller hardness than that of 550 °C aged sample. Although the hardness of the four samples were similar, there was a distinct difference in impact energy. As shown in Fig. 4(b), the toughness of 500 °C aged sample was significantly improved, more than twice as much as that of the as-hot rolled sample. However, the toughness of 550 °C aged sample was even lower than that of the as-hot rolled sample, and the impact energy of 600 °C aged sample was the lowest, only 7.5 J. In order to investigate the reasons for the difference in impact energy, the fracture morphologies and the distributions of elements were observed by the SEM-EDS. Figure 5 shows the fracture surface morphologies of the impact sample. It was obvious that the fracture morphology of 500 °C aged sample was different from those of the other three samples, characterized with dimples, which was consistent with high impact energy [43]. The as-hot rolled sample and 550 °C aged sample showed a typical cleavage plane, and some microcracks were also observed, corresponding to poor impact energy. 600 °C aged sample with the   | https://mc03.manuscriptcentral.com/friction worst impact energy was featured by a large number of intergranular cracks, which were likely attributed to the large Ti(C,N) particles precipitated at the austenite grain boundaries, leading to the initiation and propagation of intergranular cracks. Although the precipitation behaviors of the four samples were different, it was unconvincing that there was such a discrepancy in the impact energy. Thus, in order to explicate the big difference in the impact energy and fracture morphology, the Mn distribution was analyzed by the SEM-EDS. Figure 6 illustrates the SEM micrographs and the corresponding Mn distributions. As shown in Fig. 6, the strip Mn segregation was dominant in the as-hot rolled sample. In contrast, the Mn segregation was insignificant in 500 °C aged sample due to the effect of solution treatment. However, a large amount of Mn was enriched to the grain boundaries and gradually diffused to become islands due to that the diffusion coefficient of Mn atoms increased with aging temperature. It was proposed that Mn is one of the representative elements causing grain boundary segregation, and it reduced grain boundary cohesion and led to brittle intergranular fracture according to the degree of grain boundary segregation [44]. Therefore, besides the precipitations, the difference in Mn segregation significantly affected the impact energy.

Wear resistance
According to Refs. [2,16,45], the loads adopted in dry sliding wear of this study are 100, 200, and 300 N. The wear losses of the samples under different applied loads are shown in Fig. 7. It is distinct that the applied load had a significant effect on the wear loss. Under the applied load of 100 N, the difference in the wear losses of four samples was subtle. As the load increased to 200 N, the difference in the wear loss changed slightly. However, under the applied load of 300 N, 500 °C aged sample exhibited the best wear resistance, and its wear loss was nearly half that of the as-hot rolled sample and far less than those of the samples aged at 550-600 °C. Thus, we will focus on the case of 300 N applied load with great difference in wear resistance. It is worth noting that there is no direct correlation between the wear resistance and the initial hardness at 100, 200, or 300 N applied loads. Therefore, we attributed the significant difference in the wear resistance under the load of 300 N to the microstructure, wear surface, and wear mechanism.

Figures 8(a)-8(d)
show the local wear 3D morphologies and the R a under the wear tests with an applied load of 300 N, which reflected the damage and shape changes of the wear surface. The wear surface of 500 °C aged sample was the smoothest, the R a was 0.18 μm, and the wear loss was only 0.09 mm 3 . However, severe plastic deformation was observed on the wear surfaces of the other three heat-treated To obtain a better understanding, Fig. 8(e) shows the cross-section profiles of the wear tracks under an applied load of 300 N. It was found that the wear track width of 500 °C aged specimen was 1.3 mm, and the maximum wear depth was 46 μm, which were the smallest among the four samples. However, the full width and depth of the as-hot rolled sample with the maximum wear loss were 1.8 mm and 70 μm, respectively, which were much larger than those of the 500 °C aged sample. With the increase in the R a , width, and depth of the scratches, the 3D morphology exhibited pronounced grooves and pits, indicating that the wear resistance became worse, which was consistent with the wear loss, as shown in Fig. 7. Based on the abovementioned discussion, we can infer that the wear resistance was well related to the wear mechanisms, which will be further described in Section 3.5.

Wear characteristics
As shown in Fig. 9, 500 °C aged sample showed a large amount of cutting and ploughing, accompanied by a small number of abrasives, and no fatigue cracks were observed, which is typical abrasive wear characteristic [46]. This is because Ti(C,N) particles evenly distributed inside the grain instead of at the boundaries, which effectively alleviated the stress concentration and fatigue crack generation, corresponding to a smooth wear surface and the lowest R a . In contrast, the other three samples showed plenty of delamination, pits, and debris accompanied by a lot of cracks, which is typical fatigue wear characteristic [46]. High R a was due to bad cohesion, resulting in a coarse wear surface. Correspondingly, the friction resistance increased, leading to severe oxidation wear during reciprocating friction (EDS maps in Fig. 9). By analyzing the wear surface, we found that the wear quantity was closely related to the wear mechanism. Abrasive wear showed lower wear quantity, while fatigue wear showed higher wear quantity. In order to explicate the relationship between wear mechanism and microstructural evolution during sliding wear, the cross-section morphology of wear marks will be investigated. of experimental steel after wear test. It can be seen that ε-martensitic transformation occurred on the subsurface under the sliding friction of applied load, and a certain thickness of the hardening layer was developed. The existence of ε-martensite was proved that the transformation induced plasticity (TRIP) effect occurred in these samples [13]. The formation of hardened layer indicated that as soon as the surface was worn away in the sliding wear process, the newly exposed surface will come into work hardening status, which maintained a constant layer to resist the sliding abrasive wear.

Cross-section morphology
In addition, many spalling pits and fatigue cracks were detected on the cross-section surfaces of the as-hot rolled, 550 °C aged, and 600 °C aged samples, indicating low fatigue crack resistance. These microcracks were initially studied by Suh et al. [47,48], and they were regarded as fundamental experimental evidence for surface delamination. They gradually spread from subsurface to the wear surface and finally formed spalling pits. These samples were more prone to destruction under the action of reciprocating dry sliding wear, which was manifested as fatigue wear mechanism at macro level. However, it was evident for the 500 °C aged sample by analyzing the precipitation behavior of Ti(C,N) particles that the precipitates were fine and abundant, which had better precipitation strengthening effect on hardness and impact energy. When subjected to external wear, www.Springer.com/journal/40544 | Friction almost no spalling pits and fatigue cracks were found in 500 °C aged sample, indicating that the wear resistance is better. It was reasonable to infer that the microcracks were effectively decreased when a good combination between the hardness and the toughness of the surface material.
The variation of hardness along the hardened layer is shown in Fig. 10(e). It was worth noting that the depth and hardness of the hardened layer were different among the four samples, responding to the significant difference in work hardening effect. The hardened layer of the as-hot rolled sample was the smallest, only about 210 μm, and the surface hardness was 570 HV. The hardened layer was the thickest for 500 °C aged sample, about 510 μm, which was more than twice that of as-hot rolled sample; at the same time, the surface hardness increased by >100-675 HV. Combined with the wear loss, as shown in Fig. 7, it can be found that the wear loss decreased gradually with the increase in hardened layer depth and surface hardness. Therefore, it can be inferred that the sample with a better work hardening ability had better wear resistance. As was proposed, a thicker hardened layer provided better stress support for the wear surface. It was reported that the work hardening effect depended on the stacking fault energy (SFE) of austenite, which had a major impact on the mechanisms, involving the TRIP effect, twinning induced plasticity (TWIP) effect, and dislocation movement [45,[49][50][51]. Therefore, it is essential to explore the strengthening mechanism by studying the SFE.

Strengthening mechanisms
According to Refs. [52,53], the strengthening mechanism is closely related to the SFE. Martensitic transformation occurs when the SFE is less than 12 mJ/m 2 , the combined action of martensitic transformation and deformation twins occurs between 12 and 18 mJ/m 2 ; there are deformation twins when the SFE is between 18 and 35 mJ/m 2 , and pure dislocation glide occurs above 35 mJ/m 2 [54,55]. According to the change of C and Mn atoms, the SFE of the as-hot rolled sample and 500-600 °C aged samples have been calculated to be 13.36, 17.14, 15.31, and 14.46 mJ/m 2 [51,56,57]. Therefore, the strengthening mechanisms of these samples were mainly martensitic transformation and deformation twins. In contrast, the TRIP effect was the primary strengthening mechanism for the as-hot rolled sample. However, 500 °C aged sample with higher SFE was apt to produce deformation twins accompanied by martensitic transformation under the applied load. Thus, TRIP and TWIP effects were the primary strengthening mechanisms. With the increase in aging temperature, the corresponding SFE of austenite decreased gradually, leading to a weakened TWIP | https://mc03.manuscriptcentral.com/friction effect. Therefore, the strengthening mechanism resulted in the difference in thickness and depth of hardened layers.
In order to further explicate the influence of SFE on the strengthening mechanisms, the microstructures of the as-hot rolled sample with the lowest SFE and 500 °C aged sample with the highest SFE after tests were compared by the TEM. As shown in Fig. 11(a), it was observed that a large number of dislocations were generated in the as-hot rolled sample. When the deformation reaches a definite degree, the stacking faults confirmed by the high-resolution TEM (HRTEM) (Fig. 11(b)) will appear in these areas of high density dislocations [16]. The solute atoms moved to the direction of the stacking faults and formed Suzuki atmosphere [58]. The Suzuki atmosphere had an obvious pinning effect on dislocation movement, which was beneficial to work hardening. In addition, as shown in Fig. 11(c), the dislocation density near the precipitates was significantly high, meaning that the particles have a pinning effect on dislocation sliding. Moreover, a small number of twins are observed in Fig. 11(d), indicating the occurrence of TWIP effect [53]. In summary, for the as-hot rolled sample, the deformation mechanisms mainly consisted of TRIP effect and dislocation strengthening, accompanied by precipitation strengthening and TWIP effect.
In contrast, many deformation twins with dislocations were observed in the 500 °C aged sample ( Fig. 12(a)), indicating the stronger TWIP effect. It is worth noting that compared with the as-hot rolled sample, the twins were denser and intersected ( Fig. 12(b)). The interaction between twins and high density dislocations is shown in Fig. 12(c). Because the twin boundaries cut off the original continuous slip systems, the dislocation needed to change the slip systems continuously. Therefore, the dislocation movement of 500 °C aged sample was more difficult than that of the as-hot rolled sample. In addition, Fig. 12(d) confirms that the high density dislocation region was caused by the pinning effect of Ti(C,N) particles. However, combined with the behavior of Ti(C,N) particles (Fig. 3), the precipitation strengthening effect of the 500 °C aged sample was more obvious than that of the as-hot rolled sample. Due to the combined action of TRIP effect, dislocation strengthening, TWIP  effect, and significant precipitation strengthening, the work hardening ability and wear resistance of the 500 °C aged sample were better than those of the other samples.

Conclusions
In this study, the as-hot rolled medium manganese steel containing titanium was subjected to aging treatment at 500, 550, and 600 °C. The influence of aging treatment on mechanical properties and wear resistance of medium manganese steel reinforced with Ti(C,N) particles was systematically studied. The main conclusions were as follows: 1) The microstructure constituents of the as-hot rolled sample were composed of the coarse austenite matrix (39 μm) and a small amount of Ti(C,N) particles with large sizes and irregular shapes clustered at the austenite boundaries. In contrast, the aged sample was characterized with a refined austenitic structure (28-30 μm), which was because that Ti(C,N) particles inhibited the growth of austenite grain during solution treatment and achieved grain refinement strengthening. In addition, Ti(C,N) particles in the aged samples showed smaller and more regular shapes due to Ostwald ripening phenomenon. The precipitation behavior of Ti(C,N) particles indicated that the size of 0-1 μm had an excellent precipitation strengthening by pinning effect.
2) Compared to that of the as-hot rolled sample, due to that the Ti(C,N) particles precipitated inside the grains rather than the boundaries, the initial hardness of 500 °C aged sample increased by 9.5% to 312.86 HV, and the impact energy was more than doubled to 148.5 J. With the increase in aging temperature, the segregation of Mn element and Ti(C,N) particles at the grain boundaries reduced the boundary cohesion for intergranular fracture, resulting in a slight change in initial hardness but a significant decrease in impact energy to 7 J.
3) Fatigue wear was the main wear mechanism of the as-hot rolled sample, while the 500 °C aged sample exhibited abrasive wear. The poor impact energy of the as-hot rolled sample easily led to microcracks in the subsurface, and then developed into delamination, pits, and abrasives on the wear surface during wear progress, which were manifested as fatigue wear. However, a good combination of hardness and toughness of the 500 °C aged sample was effectively alleviated fatigue crack generation.
4) The wear resistance of experimental steel was independent of the initial hardness but mainly depended on the work hardening ability during wear. It was verified that the 500 °C aged sample with a better work hardening ability had better wear resistance due to the thicker hardened layer, providing better stress support. In addition, the enhanced strengthening mechanisms of the 500 °C aged sample were reinforced by the TRIP effect, dislocation strengthening, TWIP effect, and considerable precipitation strengthening, which had more favorable effects on work hardening.