Optimising Two-Stage Vacuum Heat Treatment for a High-Strength Micro-Alloyed Steel in Railway Spring Clip Application: Impact on Microstructure and Mechanical Performance

In addition, Figure 4 shows that granular carbides are dispersed and fine in the matrix while the size of carbides was increased with an increase in tempering temperature. To further observe the high magnification of the granular carbides, EDS detections were also applied in this study. It has been proposed that granular carbides grow with the elevation of tempering temperatures, leading to a decrease in dislocation density, and the distribution in carbide precipitation is mainly concentrated on the austenite grain boundaries [ 37 ]. This phenomenon illustrates that under a relatively lower temperature, the fine carbide precipitation is generated at austenite grain boundaries to hinder the movement of the austenite grain boundary, as to prevent the growth of austenite grains. Thus, its main function is to refine the austenite grain; the microstructure after the austenite transformation is also refined.

To further observe the microstructural evolution under different tempering conditions, the SEM graphs are shown in Figure 3 . It is apparent that all the microstructures are composed of granular cementite and polygonal α ferrite, while the martensite still maintains its directional distribution and acicular shape. This is because the granular carbide precipitation formed in martensite and lath edges during the tempering process [ 35 36 ]. Except for this, the granular carbide precipitation also formed at the austenite grain boundary. Accordingly, the coherent relationship was broken, leading to the spheroidisation of sheet cementite and the growth/aggregation of granular cementite.

Based on the orthogonal testing scheme, there are nine sets of experiments that were processed by different quenching technologies. Obviously, the quenched microstructure mainly consisted of fine acicular and lath martensite (see Figure 2 ). In addition, the microbands (marked by yellow arrows) along the rolling direction are exhibited in Figure 2 . For medium carbon steel, the first-formed martensite during quenching was dark due to the influence of self-tempering, while later-formed martensite exhibited light colour without the influence of self-tempering. According to the JB/T 9211-2008 standard [ 28 ], quenched martensite is rated as grade 3. The fine acicular and lath martensite is normally identified by high hardness together with excellent wear resistance and tensile strength [ 29 31 ]. The martensite needles are rated to be grade 1 because their length is less than 15 μm. To compare the influence of quenching temperature and quenching hold time on the final microstructure, it can be observed from Figure 2 a–i that the martensite structure becomes coarser and the needle length becomes longer with the increase in quenching temperature and quenching hold time. Furthermore, it is evident that the acicular martensite needles do not exhibit parallel alignment to each other. Reports [ 32 33 ] revealed that the acicular plates exhibited high misorientation angle boundaries, which were greater than 45 deg. In an austenite grain, the first-formed martensite normally runs through the entire austenite grain and splits it in half, resulting in limits to the length of martensite structures. Therefore, the later-formed martensite presents a relatively smaller grain size. The relationship between quenching temperatures/hold times and the length of martensite needles can be established through statistical analysis of the prior austenite (PA) grain size distribution. This is because the size of the lamellar martensite depends on the PA grain size: an increase in the PA grain size leads to a corresponding increase in the length of the martensite needle [ 34 ].

3.2. Mechanical Evolution of Specimens after Heat Treatment

m), specified non-proportional elongation strength (Rp0.2), Charpy impact toughness (ak), and elongation after fracture (EL).

R

) at each factor level. The calculated results are shown in

The heat-treated specimens were processed into standard tensile and impact samples. The tested mechanical properties of specimens after heat treatment include ultimate tensile strength (R), specified non-proportional elongation strength (R), Charpy impact toughness (ak), and elongation after fracture (EL). Table 4 shows all the abovementioned mechanical properties of the heat-treated samples. In this study, intuitive analysis was applied to analyse the mechanical properties test results and calculate the average deviation and range () at each factor level. The calculated results are shown in Table 5

m, which could reach 24 MPa. Considering the Rm index, a tempering temperature of no higher than 460 °C can ensure a relatively high Rm value, which also provides a certain safety reserve for the spring clips. In addition, under the conditions of a quenching temperature of 870 °C and a quenching medium of 12% PAG aqueous solution, Rm is comparatively higher than that for the other parameters. The fatigue performance of spring steel is, to some extent, correlated with its Rm. Therefore, enhancing the Rm can have a substantial impact on prolonging the fatigue life under specific conditions [

The influence of each factor on tensile strength can be obtained through intuitive analysis of orthogonal test results, as shown in Figure 5 . Clearly, the tempering temperature and holding time during tempering are the primary influential factors on the value of Rm. This is evident from the respective influence levels of Rm is 344 and 34 Mpa, respectively. In addition, the quenching medium also exerted a non-negligible influence on R, which could reach 24 MPa. Considering the Rindex, a tempering temperature of no higher than 460 °C can ensure a relatively high Rvalue, which also provides a certain safety reserve for the spring clips. In addition, under the conditions of a quenching temperature of 870 °C and a quenching medium of 12% PAG aqueous solution, Ris comparatively higher than that for the other parameters. The fatigue performance of spring steel is, to some extent, correlated with its R. Therefore, enhancing the Rm can have a substantial impact on prolonging the fatigue life under specific conditions [ 38 ].

p0.2 value in this heat treatment test, with an influence level of 301 and 34 MPa, respectively. Moreover, the quenching time and quenching medium also have a certain influence on the Rp0.2 value. When the tempering temperature is 460 °C, the Rp0.2 of the specimen is close to the maximum stress value for spring clips. Generally, if the ultimate service stress of the products equals one-half yield strength of the material, the S-N curve almost approaches the horizontal line. Although such an infinite life product is not economical, it is necessary to retain appropriate strength reserve for the material, so that the fatigue life of the spring clips can reach the standard requirement of 5 million times. In principle, the maximum stress of the spring fastener should be less than the yield limit of the material. However, the maximum stress always occurs on the surface of the spring fastener during operation. If a small plastic zone appears on the surface of the local area, the yield of the material in this area does not cause the fracture in the spring fastener, and the safe working conditions of the spring fasteners can still be ensured. Furthermore, benefitting from the strain hardening, both the yield limit and strength of the material in the plastic zone are improved. Although the appearance of the plastic zone leads to a certain residual deformation, both the deformation and the buckle pressure loss are quite small when the local plastic zone of the spring fastener is very small, which improves the fatigue life of the spring fastener.

Likewise, according to the intuitive analysis of Rp0.2, depicted in Figure 6 , it can be observed that the tempering temperature and tempering holding time are the two factors that greatly influence the Rvalue in this heat treatment test, with an influence level of 301 and 34 MPa, respectively. Moreover, the quenching time and quenching medium also have a certain influence on the Rvalue. When the tempering temperature is 460 °C, the Rof the specimen is close to the maximum stress value for spring clips. Generally, if the ultimate service stress of the products equals one-half yield strength of the material, the S-N curve almost approaches the horizontal line. Although such an infinite life product is not economical, it is necessary to retain appropriate strength reserve for the material, so that the fatigue life of the spring clips can reach the standard requirement of 5 million times. In principle, the maximum stress of the spring fastener should be less than the yield limit of the material. However, the maximum stress always occurs on the surface of the spring fastener during operation. If a small plastic zone appears on the surface of the local area, the yield of the material in this area does not cause the fracture in the spring fastener, and the safe working conditions of the spring fasteners can still be ensured. Furthermore, benefitting from the strain hardening, both the yield limit and strength of the material in the plastic zone are improved. Although the appearance of the plastic zone leads to a certain residual deformation, both the deformation and the buckle pressure loss are quite small when the local plastic zone of the spring fastener is very small, which improves the fatigue life of the spring fastener.

2, respectively (see

Through the intuitive analysis of the orthogonal test results, the factors that have greater influences on the ak value are tempering temperature, quenching temperature, and quenching holding time, with an influence level of 12.08, 4.58, and 2.60 J/cm, respectively (see Figure 7 ). The effects of quenching temperature and quenching holding time on the ak mainly result from the PA grain size, while the tempering temperature mainly affects the ak by changing the distribution, shape, and size of the precipitated carbides. When the tempering temperature is 500 °C, the ak reaches its peak value. Nevertheless, the ak value seems to be less susceptible to the tempering time and quenching medium.

Based on the intuitive analysis of the orthogonal test results, the most influential factors affecting the EL after fracture are tempering temperature and tempering holding time. Their influence levels regarding EL are calculated to be 1.77% and 0.70%, respectively. In addition, the quenching medium and quenching holding time also have an influence on the EL after fracture ( Figure 8 ).

From the analysis of the experimental results, the EL value after fracture of the materials tempered by different processes all acquired a sound level, which is more than 9%. Furthermore, according to the Vicker hardness results after quenching exhibited in Table 4 , the sample quenched by PAG aqueous solution possesses a significantly higher hardness value than that for the sample quenched by mechanical oil. The hardness of the quenched samples increases with the elevation of quenching temperature, while the quenching holding time has little effect on the hardness after quenching. According to the relevant data in Table 5 , among the designed experimental parameters, the tempering temperature is the factor that has the greatest influence on the hardness value for the tempered samples. Among them, the hardness values of samples obtained at 440 and 460 °C are reasonable, while other factors have no obvious influence on the hardness values.

From what has been discussed above, the tempering temperature is often the most influential factor on the mechanical properties index, while other factors have varied influences on the mechanical properties, which is consistent with the expectations before the test design. When the tempering temperature is set as 440–460 °C, the tensile strength of the material is greater than the maximum stress value (1496 MPa) at the active service of the spring fastener. When the tempering temperature is 460 °C, the Rp0.2 of the material is close to the maximum stress value for the spring fastener. Therefore, a sound match of impact toughness and elongation after fracture could be achieved.