Cleanliness distribution of high-carbon chromium bearing steel billets and growth behavior of inclusions during solidification

Chao Gu, Yan-ping Bao^*, Lu Lin

State Key Laboratory of Advanced Metallurgy, University of Science and Technology Beijing, 30 Xueyuan Road, Haidian District, Beijing 100083, P.R. China

*Corresponding author: 15210951549@sina.cn

1. INTRODUCTIONTOP

With the rapid development of high-speed railway technology in China, the quality of bearing steel used in high-speed railway is under severe test. The service environment is complex and the cleanliness requirements are extremely high. Under high-speed repeated load, even tiny quality problem could have large impact on the service life of bearing steel (Hua et al., 2013; Mazzù et al., 2015; Moghaddam et al., 2015). The non-metallic inclusions failed to be removed during melting process remain in the billets of bearing steel and will not disappear in the following heat treatment. These inclusions normally become key factors that shorten the service life of bearing steel. Efforts have been made to investigate the formation mechanism and evolution of inclusions during melting process (Guo et al., 2013; Ma et al., 2014; Yang et al., 2015; Zhang et al., 2015a; Zhao et al., 2015), however few studies have been carried out on the evolution of cleanliness, also the same for the distribution and size evolution of inclusions in billet. Wang et al. (2015) studied the cleanliness evolution in the thickness direction of IF steel billets and found that the inclusion types and cleanliness changed with positions in billets.

In present study, typical high-carbon chromium bearing steel GCr15, which is widely used in China, is set as the research object. The cleanliness evolution and the type, distribution, and size variation of inclusions were investigated. The formation sequence of all kinds of inclusions during solidification process, the inclusion size variation in the thickness-direction and width-direction of billet, the dominant area of total oxygen content, and the control direction of how to optimize inclusions were made clear, which could help guide the industrial production of bearing. Thanks to this, during bearing processing, accumulation regions of larger inclusions and high total oxygen contents in billets could be avoided, and high cleanliness part could be used. It is of great significance for improving the quality of GCr15 bearing steel.

2. EXPERIMENTAL STUDIESTOP

The sectional dimension of the GCr15 billet used in present study is 204×189 mm, the average compositions tested by Thermo spark direct-reading spectrometer are shown in Table 1.

2.1. Analysis of total oxygen contents and nitrogen contentsTOP

Total 17 cylindrical samples (ɸ5×45 mm) were taken on the cross profile of GCr15 billet. The sketches is shown in Fig. 1a. As shown in the figure, at the center line through the width direction of billet cross profile, 9 samples were taken every other 13.75 mm respectively from inner to outer arc; 5 samples were taken every other 37.5 mm from inner to outer arc in the middle of the longitudinal centerline and the edge of billet; 3 samples were taken every other 85 mm from inner to outer arc on the edge of the billet.

The total oxygen contents and nitrogen contents of the 17 cylindrical samples were analyzed with infrared absorption method and thermal conductivity method to study the content distributions of total oxygen and nitrogen on the cross profile of GCr15 billet.

2.2. Analysis of micro-inclusionsTOP

13 metallographic specimens (No.1~13) were taken at the center line of width-direction of billet cross profile and 3 other metallographic specimens (No.14~16) were taken in the center line of thickness-direction. The sketches are shown in Fig. 1b. The specimen size of No. 5, 6, 8, and 9 was 7.5×15×15 mm and the specimen size of other 12 pieces was 15×15×15 mm. These metallographic samples were polished by SiC paper and diamond suspensions to characterize the inclusions through observations in a Scanning Electron Microscope (SEM). The chemical compositions of inclusions were analyzed by Energy Dispersive Spectrometry (EDS) combined with the SEM.

2.3. Observation of secondary dendrite armTOP

The metallographic specimens numbered as 1, 2, 4, and 7 (in the position of inner arc, inner arc side 1/6, inner arc side 1/3, and center of the billet respectively) were corroded with 4% Nital, and then cleaned out with absolute ethyl alcohol and blow-dried. These corroded metallographic specimens were observed through optical microscope to study the variation of secondary arm spacing (SDAS) in GCr15 billet with solidification.

3. RESULTS AND DISCUSSIONTOP

3.1. Variation of total oxygen contents and nitrogen contentsTOP

The color map of total oxygen contents in cross profile of billet is showed in Fig. 2. This color map was plotted using the griddata v4 method with MATLAB (version R2014b) based on the 17 total oxygen test results in different positions shown in Fig. 1a.

From Fig. 2, it is clear that the change rule of total oxygen contents in thickness-direction of billet is as below: the total oxygen contents in inner arc and outer arc are relatively low, and gradually increase from inner arc side 3/16 and outer arc side 1/4. The total oxygen contents in the center of billet show a slight decline, but still higher than those in inner arc and outer arc. The change rule of total oxygen contents in width-direction of billet is that the total oxygen contents in the center are relatively high and those near the edge of billet are lower. Therefore, the region with high total oxygen content in the cross profile of GCr15 billet is shown as the solid border in Fig. 2 (from inner arc side 3/16 to outer arc side 1/4 and from left edge side 5/16 to right edge side 5/16). Generally, symmetry of cleanliness only exists in left side and right side of billet while the cleanliness of inner arc part and outer arc part of billet is different. Oxide inclusions usually form in molten steel, and tend to float during solidification. Consequently, the region with relatively high total oxygen content in the inner arc side is nearer the outer layer of billet compared to that in the outer arc side.

Nitrogen contents in the cross profile of GCr15 billet fluctuated within the range of 34~36×10⁻⁶, which is pretty narrow, and the distribution rule is not apparent.

3.2. Variation of secondary arm spacing and cooling rate of billetTOP

The morphologies of secondary dendrite arm in different positions of cross profile of GCr15 billet are shown in Fig. 3. When the carbon content in steel is over 0.15wt%, the relation among secondary arm spacing, cooling rate, and carbon content is shown in Eq. (1) (Wolf et al., 1982):

where λ₂ is secondary arm spacings (μm); R_c is cooling rate (°C/s). With the observed secondary arm spacings and carbon content in GCr15 bearing steel (1.01 wt%), the cooling rates were calculated and results are shown in Table 2.

3.3. Distribution characteristics of inclusionsTOP

Total 667 micro-inclusions were observed in different position of GCr15 bearing steel billet. There are three main types: MgO-Al₂O₃, MnS, and TiN. Large inclusions will do great harm to the quality of steel. However, smaller inclusions, especially when sizes are under 1 μm, usually are considered harmless, even beneficial to the quality of steel (Yang et al., 2013; Zhang, 2013). Consequently, the observation of inclusions mainly aimed at those whose sizes were above 1 μm in present study.

3.3.1. Formation sequence of inclusionsTOP

Some MgO-Al₂O₃ inclusions observed in GCr15 bearing steel billet were partly or totally wrapped by MnS and few MgO-Al₂O₃ inclusions were wrapped by TiN. Few TiN inclusions were combined with MnS. The morphologies of typical inclusions are shown in Fig. 4.

The reactions related to the generation of MgO-Al₂O₃, MnS, and TiN inclusions are shown as Eqs. (2)~(6) (Sigworth and Elliott, 1974; Kubaschewski et al., 1993; Itoh et al., 1997; Fujii et al., 2000; Zhang et al., 2015b), among which Eq. (2) is the formation reaction of MnS inclusions, Eq. (3) is the formation reaction of TiN inclusions. Assuming the atomic ratio of MgO to Al₂O₃ in MgO-Al₂O₃ type inclusions is 1, the formation reaction of MgO-Al₂O₃, Eq. (7), can be derived from Eqs. (4)~(6).

Assuming the activity of solid product is 1, the formula of practical solubility product Q and equilibrium solubility product K are shown as Eqs. (8)~(13). When the system is in equilibrium, Eq. (14) can be derived.

The solidus temperature (T_L) and liquidus temperature (T_S) of high-carbon bearing steel can be calculated with Eqs. (15)~(16) (Goto et al., 1995). The temperature of solidification front (T  ) can be calculated with Eq. (17) (Takada et al., 1976):

where T_m is the melting point of pure iron; f_s is solidification ratio.

According to Schiel equation, there is Eq. (18) (Ohnaka, 1986):

where C₀ is solute concentration of molten steel; C_L is solute concentration in liquid phase at solidification front; k_M is distribution coefficient of solute in molten steel.

The precipitation sequence of inclusions in molten steel of GCr15 bearing steel (Fig. 5) was calculated through Eqs. (2)~(14) and Eq. (17). The formation condition is met when practical solubility product Q is larger than equilibrium solubility product K. As shown in Fig. 5, it is clear that Q is always larger than K for MgO-Al₂O₃ from the beginning to the end of solidification, thus the formation of MgO-Al₂O₃ inclusions is prior to the beginning of solidification. However, TiN and MnS inclusions both generate during the solidification process of molten steel, and the solidification ratios of molten steel when they first appeared are 0.795 and 0.997 respectively.

Therefore, the formation sequence of inclusions in GCr15 bearing steel is MgO-Al₂O₃ > TiN > MnS, which corresponds with the observed wrapping status of inclusions. During solidification, the initial generating time of MgO-Al₂O₃ inclusions is far earlier than those of TiN and MnS inclusions at a certain point in billet, thus MgO-Al₂O₃ inclusions are often totally or largely wrapped by TiN and MnS. On the other hand, the initial generating times of TiN and MnS are quite close, TiN is often combined partly with MnS rather than wrapped by it. Consequently, relatively soft MnS could only protect the steel matrix from hard MgO-Al₂O₃ inclusions to a certain degree during service of bearing steel, but not for TiN inclusions, which often carry sharp edges and would do great harm to steel matrix. Moreover, due to the earlier possible generating time of TiN compared to MnS, MgO-Al₂O₃ inclusions could possibly be wrapped by TiN firstly, which could not be effectively protected by MnS. And this would increase the probability of quality problems of bearing steel.

3.3.2. Size variation of inclusionsTOP

The variations of average size of MgO-Al₂O₃, MnS, and TiN inclusions observed in the cross profile of GCr15 bearing steel billet with position are shown in Fig. 6. The average sizes are the smallest in the outer layer of billet (inner arc, out arc, and edge of billet), and increase gradually during the solidification process, reaching the maximum size in the region of inner arc side 1/3 and outer arc 1/4 in thickness direction, and edge 1/3 in width direction. The average size of TiN inclusions increases from 2.5~3.0 μm to 4.6~5.7 μm. The average size of MnS inclusions increases from 2.6~2.8 μm to 4.2~4.6 μm. After that the average sizes of these two kinds of inclusions drop slightly in the center of billet, but still larger than those in the outer layer of billet. This result is consistent with the variation tendency of inclusions in IF steel billet reported by Wang et al. (2015). The average size of MgO-Al₂O₃ inclusions fluctuates over the range of 1.6~2.1 μm in the whole billet and there is no obvious change in thickness direction and in width direction.

The growth of TiN and MnS inclusions in steel could be described by Eq. (19). And through integrals over Eq. (19), Eq. (20) can be derived (Ohnaka, 1986):

where r is the radius of inclusion (μm); t is solidification time (s); t is local solidification time (s); M_m is the molar mass of Fe (0.056 kgmol⁻¹); M_S is the molar mass of inclusion (TiN: 0.062 kgmol⁻¹; MnS: 0.087 kgmol^-1); r_m is the density of Fe (7070 kg.m^-3); r_s is the density of inclusion (TiN: 5430 kg.m^-3; MnS: 3990 kg.m^-3); D is the diffusion coefficient of solute in molten steel (S: 3.9×10⁻⁷ cm²s^-1 (Cornelissen, 1986), N: 3.25×10⁻³e^−11500/RT cm²s^-1 (Wang et al., 2014).The restrictive steps of the growth of TiN and MnS are the diffusion of N and S respectively. Local solidification time t in Eq. (19) is obtained though Eq. (21):

Considering the solidification ratios f_s of molten steel are 0.795 and 0.997 respectively when TiN and MnS begin to generate, calculations of the growth of the two kinds of inclusions both begin with f_s. The calculation results are shown in Fig. 7. As shown in Fig.7, the sizes of TiN and MnS keep increasing from inner arc to the center of billet. For the same position in billet, the size of TiN is larger than MnS. This calculation result is consistent with the test result in the region from inner arc to inner arc side 1/3, but is contrary to the test result in the region from inner arc side 1/3 to the center of billet. The cause of this phenomenon is that the secondary arm spacing is smaller in the region from inner arc to inner arc side 1/3, where local solidification time is longer and thus the model error is smaller. While the secondary arm spacing is larger in the region from inner arc side 1/3 to the center of billet, where local solidification time is shorter and thus the model error is larger. Moreover, the initial size of inclusion in the kinetic model shown in Eq. (19) is simplified, which leads to a certain limitation of this model (Liu et al., 2002; Yu and Li, 2015).

4. CONCLUSIONSTOP

ACKNOWLEDGMENTSTOP

This research was financially supported by State Key Laboratory of Advanced Metallurgy Foundation (N° 41614014), National Natural Science Foundation of China (N° 51574019), and Fundamental Research Funds for the Central Universities (FRF-TP-15-008A3). The authors express their appreciation to the foundation for providing financial support that guarantees the study successfully to be carried out.

REFERENCESTOP