Journal of Materials Science & Technology  2019 , 35 (7): 1298-1308 https://doi.org/10.1016/j.jmst.2019.01.015

Orginal Article

Effects of rare earth elements on inclusions and impact toughness of high-carbon chromium bearing steel

Chaoyun Yangab, Yikun Luana*, Dianzhong Lia, Yiyi Lia

aInstitute of Metal Research, Chinese Academy of Sciences, Shenyang, 110016, China
bSchool of Materials Science and Engineering, University of Science and Technology of China, Hefei, 230026, China

Corresponding authors:   *Corresponding author.E-mail address: ykluan@imr.ac.cn (Y. Luan).

Received: 2018-04-21

Revised:  2018-06-30

Accepted:  2018-08-14

Online:  2019-07-20

Copyright:  2019 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

High-carbon chromium bearing steels with different rare earth (RE) contents were prepared to investigate the effects of RE on inclusions and impact toughness by different techniques. The results showed that RE addition could modify irregular Al2O3 and MnS into regular RE inclusions. With the increase of RE content, the reaction sequence of RE and potential inclusion forming elements should be O, S, As, P and C successively. RE inclusions containing C might precipitate in molten steel and solid state, but the precipitation temperature was significantly higher than that of carbides in high-carbon chromium bearing steel. For experimental bearing steels, the volume fraction of inclusions increased steadily with the increase of RE content, but smaller and more dispersed inclusions could be obtained by 0.018% RE content compared with bearing steel without RE, whereas the continuous increase of RE content led to an increasing trend for inclusion size and a gradual deterioration for inclusion distribution. RE addition could improve the transverse impact toughness and isotropy of bearing steel, and for modified high-carbon chromium bearing steel by RE alloying, the increase of RE content continuously increased both transverse and longitudinal impact toughness until excessive RE addition.

Keywords: High-carbon chromium bearing steel ; Rare earth ; Inclusions ; Impact toughness

0

PDF (4910KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Chaoyun Yang, Yikun Luan, Dianzhong Li, Yiyi Li. Effects of rare earth elements on inclusions and impact toughness of high-carbon chromium bearing steel[J]. Journal of Materials Science & Technology, 2019, 35(7): 1298-1308 https://doi.org/10.1016/j.jmst.2019.01.015

1. Introduction

GCr15 bearing steel has been widely used for rolling bearings of automobile and machinery subjected to extreme working conditions, i.e., heavy load and high speed. Thus some properties, such as wear resistance, high fatigue resistance and certain impact toughness, are urgently required. As is well known, rolling contact fatigue (RCF) failure is the main failure mode for bearings. Non-metallic inclusions, commonly acting as stress raiser to initiate fatigue crack, play an essential role in RCF life. Numerous studies [[1], [2], [3], [4]] on bearing steel have shown that inclusion composition, size, morphology and distribution could be important factors for RCF life. On the other hand, there have been many publications [5,6] manifesting that rare earth(RE) in steel could purify liquid steel and modify inclusions effectively. Therefore, RE addition could be a promising method to obtain small, dispersed and regular RE inclusions, thereby making it possible to improve the RCF life of bearing steel.

Despite the fact that RE addition may potentially prolong the RCF life of bearing steel, impact toughness and wear resistance of bearing steel should also be taken into consideration to avoid early failure. In the previous researches, methods of alloying, austempering and cold deformation were developed to improve the impact toughness of bearing steel [[7], [8], [9], [10]]. The optimization of wear resistance mainly involved different gaseous atmospheres and temperature, surface treatment, lubrication and additive [[11], [12], [13]]. Moreover, studies on fatigue of bearing steel concentrated on the establishment of models and simulations [[14], [15], [16]] as well as possible crack origins [[17], [18], [19], [20]]. However, the improvemtent methods of certain property were seldom extended to other mechanical properties of conventional GCr15 bearing steel in the previous investigations. In view of excellent combination of mechanical properties, nanostructured bainite steel, which can be obtained by low-temperature austempering of modified bearing steel, could be potentially used for bearings according to previous publications [[21], [22], [23], [24]]. Nevertheless, a comprehensive study on its mechanical properties has not been reported yet, and complicated process together with precise parameter control also restricts its development.

Therefore, it would be highly advantageous to bearing industry if RE addition can improve impact toughness on the basis of modifying inclusions. The purpose of present work is to investigate the effects of RE on inclusions and impact toughness. We prepared GCr15 bearing steel samples with different RE contents to reveal its effects on inclusions and impact toughness.

2. Experimental

GCr15 bearing steels with different RE contents were prepared by induction melting of pure materials and then cast into ingots after 1 h at 1550 °C under vacuum condition. The RE used was La-Ce mischmetal with the mass ratio of 1:2 around. To investigate the effect of different RE contents on bearing steel, the experimental ingots were forged into square rods with cross section of 65 mm × 65 mm in the temperature range of 950-1180 °C under the same forging process. Table 1 shows the chemical compositions of the experimental steels.

Table 1   Chemical compositions of the experimental steels (wt%).

NumberCSiMnCrSAsPLaCeO
10.990.260.301.450.004<0.0010.01------0.0008
20.990.250.361.430.0052<0.0010.0090.00590.0120.0008
30.990.240.351.420.0037<0.0010.0090.010.0250.0007
40.990.250.361.450.006<0.0010.010.0240.0520.0008
50.990.240.351.450.007<0.0010.010.0460.0940.0014
60.950.230.351.410.0013<0.0010.0110.180.40.0019

New window

To observe the inclusions in bearing steels with different RE contents, the steel samples were directly cut from forged steels by wire electrode discharging, and longitudinal cross-sections were ground to 2000 grit using SiC abrasive papers, then polished with diamond paste. The chemical compositions and morphology of inclusions were analyzed through a scanning electron microscope (SEM) combined with an energy dispersive spectroscope (EDS), and an optical microscope (OM) along with image analysis software Image-Pro Plus was used to evaluate the size, distribution and volume fraction of inclusions. For future study on inclusions of No.6 steel, the polished sample of tempered No.6 steel was etched with alkaline potassium permanganate solution (1 g KMnO4, 0.3 g NaOH, 25 ml H2O) and an electron probe micro-analyser (EPMA) was employed to investigate the element distribution of inclusions and carbides. In addition, solution treatment (oil quenched by austenitizing at 1000 °C for 1.5 h) was performed on the forged sample of No.6 steel and the comparison of inclusions before and after solution treatment was made.

Charpy impact test was performed to study the impact toughness of the test bearing steels with different RE contents after heat treatment. Sheet samples with size of 65 × 65 × 14 mm perpendicular and parallel to the axis of forged square rods first underwent isothermal spheroidizing with upper temperature of 790 °C for 4 h and lower temperature of 720 °C for 3 h, and then they were oil quenched by austenitizing at 840 °C for 2 h and tempered at 160 °C for 3 h. Fig. 1 shows the temperature profile of isothermal spheroidizing heat treatment used in our experiment. The transverse and longitudinal impact toughness was evaluated at room temperature by using unnotched charpy impact specimens with dimensions of 10 mm × 10 mm × 55 mm. The fractured surfaces of the specimens after testing and hazardous element distribution of impact specimens were examined by SEM-EDS and secondary ion mass spectroscopy (SIMS), respectively.

Fig. 1.   Diagram of isothermal spheroidizing heat treatment.

3. Results and discussion

3.1. Effects of RE on inclusions

Fig. 2(a) and (b) shows that Al2O3 and MnS are the main inclusions in Al-killed GCr15 bearing steel without RE, and energy spectrum peaks of Ca and Mg in Fig. 2(a) also indicate the existence of small amount of CaO and MgO. The typical complex inclusion and isolated MnS are of irregular shape. Elongated or angular Al2O3 tends to initiate micro-cracks under fatigue loading and elongated MnS with large size probably leads to lamellar tearing together with anisotropy, both of which are detrimental to the mechanical properties of bearing steel due to the differences of thermal expansion property and hardness between matrix and inclusions. In addition, Al2O3 formed early in molten steel could act as heterogeneous nucleus of MnS, so isolated Al2O3 was hardly observed in the low oxygen and rather high sulfur bearing steel.

Fig. 2.   SEM images and EDS analysis results of inclusions in test steels with different RE contents: (a, b) No.1 steel; (c, d) No.2 steel; (e, f) No.3 steel; (g) No.4 steel; (h) No.5 steel; (k) No.6 steel.

Owing to the active chemical property of RE elements, small amount of RE addition can modify irregularly shaped Al2O3 and elongated MnS into spherical or ellipsoidal RE inclusions, mainly RE oxysulfide and RE sulfide as shown in Fig. 2(c) and (d). With more RE addition to bearing steel, Fig. 2(e) and (f) manifests relatively regular and complicated RE-S-As and RE-S-As-P inclusions that still have smooth edges with concave-convex arc in No. 3 steel. When RE content increases to 0.076 wt%, Fig. 2(g) indicates that RE-S-As-P inclusions still remain, but their morphology starts to become irregular, usually with sharp edge, which is primarily attributed to the growth of RE inclusions caused by more inclusion forming element in molten steel. Excessive RE addition results in RE-C interaction and then forms RE inclusions containing C as depicted in Fig. 2(h) and (k). These inclusions are of large size and reveal the tendency of aggregation-grown.

From the SEM observations and EDS analyses of inclusions mentioned above, it can be concluded that RE in molten steel would choose to react with O and S first and form RE oxysulfide and RE sulfide, which coincides with thermodynamic calculation results of previous investigations [25,26] considering the content of O and S in No.2 steel. In view of the appearance of RE-S-As and RE-S-As-P inclusions and the absence of RE-S-P inclusions in No. 3 steel, the reaction range should be expected to extend to As and P successively with more RE addition to conventional bearing steel. Fig. 3 displays the distribution of S, As and P in different RE inclusions. Unlike RE elements occupying the whole inclusions, high S concentration in the inner part and high As content in the outer edge are observed by SEM-mapping, as shown in Fig. 3(a). In contrast, Fig. 3(b) shows that the concentrations of S and As can also exist in almost the same area of inclusions. P enrichment area approximatively forms a closed P-enriched band by surrounding the central area of high S and As content, which means that P and As concentrations seem to be incompatible and complementary. Liu et al. [5] obtained similar observations in the study on the distributions of As, P, Pb and Sn caused by Ce addition. The results from Fig. 3 indicate that RE may interact with As in two possible forms of correlating with S and As simultaneously and reacting with As later than S, while RE-P reaction inclines to follow that of RE-As in No.3 steel. Plus, as Fig. 2(h) and (k) presents, RE will bond with C when the RE content increases to 0.014 wt% or more. Therefore, through synthetical consideration of known thermodynamic calculation results, possible thermodynamic equilibrium state together with the fact that the P content is about ten times of As, it can be deduced that the reaction sequence of RE and potential inclusion forming elements should be O, S, As, P and C successively with the increase of RE content in experimental bearing steels. However, more RE addition increases the probability of the reactions between RE and other potential inclusion forming elements, such as As, P and C, so strange types of inclusions may also form if kinetic conditions cannot match thermodynamic equilibrium conditions. That is why RE-O-As-P and RE-As inclusions without S can be obtained although they seem to fail to obey the reaction sequence inferred by thermodynamic equilibrium. Besides, in order to enhance the microalloying effect of RE elements by increasing the amount of RE solid solution, the method of adding RE in batches should be considered to avoid the reactions between RE and more elements.

Fig. 3.   SEM images and EDS maps of element distribution of inclusions in No.3 steel: (a) RE-S-As inclusion; (b) RE-S-As-P inclusion.

Compared with bearing steel without RE, small amount of RE addition can form spherical or ellipsoidal RE oxysulfide and RE sulfide. The hardness of Ce2O2S is around 360 HV according to first principle calculation and the coefficient of thermal expansion of RE sulfide is about 12 × 10-6/°C [27], so these inclusions probably possess better hardness and thermal expansion property to match the matrix than Al2O3 and MnS [1,27,28], which improves the anisotropy and fatigue property of bearing steel. The increase of RE content, bringing about gradual deterioration of inclusion morphology, expands the reaction range of RE elements and leads to the formation of inclusions containing As and P. As and P are harmful residual elements for steels. They come from scrap and usually segregate at grain boundaries. RE addition, whether by forming inclusions containing As and P or by its grain boundary segregation, suppresses the grain boundary segregation of hazardous residual elements and enhances the cohesion of grain boundaries, which may be beneficial to the properties related to grain boundary, such as toughness and hot ductility. However, excessive RE addition easily causes the aggregation of precipitated inclusions and generates large inclusions. Large inclusions make the incompatibility between matrix and inclusions more obvious, which is certainly detrimental to steels.

RE inclusions containing C differ from conventional inclusions that are usually composed of metallic elements and impurity elements. RE can react with basic alloying element C of bearing steel and form large inclusions containing C and RE elements after excessive RE almost completely consumes impurity elements in molten steel. In order to investigate RE inclusions containing C, Fig. 4 presents the element distribution of inclusions and carbides of No.6 steel detected by EPMA. The result shows that the elements of C, O, La, Ce and Ca almost occupy the whole inclusions obviously. By contrast, S concentration is relatively low. The difference of element concentration in different areas of inclusions suggests that inclusion forming elements can precipitate on existing inclusions in different proportions when thermodynamic condition and kinetics condition are satisfied, and different areas may have different crystal structures. The overall distribution of C and the existence of O and S eliminate the probability that inclusions form in the solid state absolutely. Meantime, the absence of Cr in inclusions and carbides without RE elements not only confirms the difference between inclusions containing C and carbides, but also illustrates that Cr can hardly participate in the formation of inclusions and RE cannot clearly form carbides yet. However, Hufenbach et al. [29] has studied the effect of Ce addition on high strength cast steel. The results show that carbide forming elements (V, Mo and Cr) are able to be enriched in Ce-rich carbo-oxides and Ce can also contribute to the formation of carbides. Fig. 5 displays the statistical results of inclusion size before and after solution treatment in No.6 steel. It is evident that solution treatment can refine RE inclusions containing C, which means that these inclusions may precipitate in the soild state as well. Given the absence of RE elements in carbides of experimental bearing steel, it can be inferred that the precipitation temperature of RE inclusions containing C is significantly higher than that of carbides in the experimental bearing steel, which is conducive to explaining the investigations by Hufenbach et al. [29]. Precisely because of the high precipitation temperature of VC and Mo2C and the facilitation for Cr provided by V and Mo, RE elements and carbide forming elements (V, Mo and Cr) infiltrate into carbides and RE-rich carbo-oxides separately. For deeper understanding of RE inclusions containing C in experimental bearing steel, more details need to be figured out, such as chemical compositions, crystal structure and whether the mechanism of refining inclusions involves the solution of O or not.

Fig. 4.   EPMA images and element distribution maps of inclusions and carbides in No.6 steel.

Fig. 5.   Average size and maximum size of inclusions in No.6 steel before and after solid solution treatment.

The size distribution and volume fraction of inclusions for experimental bearing steels with different RE contents are evaluated and demonstrated in Fig. 6. As shown in Fig. 6(a), the size of inclusions first decreases by small amount of RE in No.2 steel, and then with the increase of RE content in the experimental bearing steels, it shows an increasing trend. On the other hand, Fig. 6(b) indicates that the volume fraction of inclusions continuously increases with more RE addition (The enlarged view of statistic results from No.1 steel to No.5 steel is inside Fig. 6(b)), which is mainly caused by strong affinity of RE elements to other inclusion forming elements and similar density of RE inclusions to liquid steel. Of course, the increase of oxygen content in No.5 and No.6 steels, probably derived from the reactions between RE and the crucible, is another factor for high volume fraction of inclusions. In contrast with bearing steel without RE, where Al2O3 and MnS tend to aggregate and form large inclusions due to high interfacial energy between inclusions and molten steel, small amount of RE addition unlikely gives rise to aggregation of RE inclusions. Therefore, the volume fraction of inclusions larger than 5 μm is lower than that of bearing steel without RE although the volume fraction of inclusions in No.2 and No.3 steels increases. More RE addition promotes the precipitation and growth of inclusions further and increases the probability of aggregation of inclusions, but No.4 and No.5 steels still have relatively lower volume fraction of inclusions larger than 5 μm compared with No.1 steel. When RE content increases to 0.58 wt% as No.6 steel, there is a great tendency for aggregating inclusions, so volume fraction of inclusions larger than 5 μm is significantly higher than that of bearing steel without RE. Large inclusions could easily lead to stress concentration at interface, acting as the origin of micro-cracks which eventually cause the fatigue fracture of bearing steel. Accordingly, large inclusions, especially with irregular shape and high hardness, should be avoided.

Fig. 6.   Statistic results of (a) size distribution and (b) volume fraction of inclusions in test steels with different RE contents.

Fig. 7 shows the OM images of inclusions in the experimental bearing steels containing different RE contents. It is apparent that No.2 and No.3 steels with no aggregation of inclusions display smaller and more dispersed inclusions in comparison with No.1 steel, especially No.2 steel with lower RE content. More RE addition gradually deteriorates the distribution of inclusions from No.4 steel to No.6 steel by increasing the trend of aggregation of inclusions and the formation of more large inclusions, which is in accordance with the results of Fig. 6. Based on Fig. 6, Fig. 7, it could be concluded that the increase of RE content causes the evolution of main growth mechanism of inclusions from precipitation-grown to aggregation-grown [30], which is responsible for the distribution of inclusions for experimental bearing steels.

Fig. 7.   Overview of inclusions in (a) No.1 steel, (b) No.2 steel, (c) No.3 steel, (d) No.4 steel, (e) No.5 steel and (f) No.6 steel.

3.2. Effects of RE on impact toughness

In order to evaluate the effect of RE on impact property, the transverse and longitudinal impact absorbed energy of experimental bearing steels after quenching and tempering are measured, as described in Fig. 8(a). Obviously, RE addition can continuously improve the transverse impact absorbed energy of bearing steel from 21.7 J of No.1 steel to 37.9 J of No.5 steel. By comparison, the longitudinal impact absorbed energy of experimental bearing steel without RE is fairly high, small amount of RE addition in No.2 steel results in a dramatic decrease, and then with the increase of RE content, it gradually increases from No.3 steel to No.5 steel, but still slightly lower than that of No.1 steel. Excessive RE addition in No.6 steel deteriorates the transverse and longitudinal impact property of bearing steel seriously. Fig. 8(b) exhibits the isotropy of experimental bearing steels assessed by the ratio of transverse impact absorbed energy to longitudinal impact absorbed energy. It is found that RE can easily improve the isotropy of bearing steel from the ratio of 0.37 for No.1 steel to higher than 0.6 for RE bearing steel. On basis of the results given by Fig. 8, from the perspective of impact property, appropriate RE addition is extremely beneficial to bearing steel.

Fig. 8.   Effects of RE content on (a) impact absorbed energy and (b) the ratio of transverse impact energy to longitudinal impact energy.

To shed light on the reason why experimental bearing steels possess such impact properties as Fig. 8 shows, the fracture surfaces of longitudinal impact specimens with different RE contents are examined by SEM-EDS after impact test. Fig. 9 displays the fracture surface of No.1 steel and the fractured complex inclusion, with Al2O3 wrapped by MnS, is observed. MnS has excellent plasticity and there are many elongated MnS inclusions with large size existing in No.1 steel. Elongated MnS, whether in the form of isolated MnS or complex inclusions, has a central axis perpendicular to the fracture surface of longitudinal impact specimen, so it will be inevitably cut off during the process of crack initiation and propagation, thus consuming much energy. On the contrary, relatively low energy consumption for transverse impact sample is obtained because elongated MnS or complex inclusions with large size can accelerate crack initiation and propagation due to their central axes parallel to the fracture surface. That is why No.1 steel has rather high longitudinal impact absorbed energy and pretty low transverse impact absorbed energy, which leads to low isotropy and bad impact property.

Fig. 9.   SEM fractograph of No.1 longitudinal impact specimen (a) and EDS of fractured complex inclusion (b).

RE addition can make elongated MnS be replaced with brittle RE inclusions, thereby improving the transverse impact absorbed energy and isotropy of No.1 steel, and both transverse and longitudinal impact absorbed energy of experimental bearing steels has an increasing trend with the increase of RE content until excessive RE addition to No.6 steel. Fig. 10 shows the SEM fractographs of crack initiation regions for experimental bearing steels with different RE contents. It can be seen that intergranular fracture dominates the fracture morphology of crack initiation region, which is attributed to weak grain boundary that more easily initiates cracks and acts as the path of micro-crack propagation. The area of intergranular fracture first decreases from No.2 steel to No.5 steel, then increases sharply for No.6 steel, as indicated by red dotted lines in Fig. 10. Considering that the area of intergranular fracture reflects the strength of grain boundary and the energy consumption of crack initiation, it may be inferred that the strength of grain boundary together with the energy consumption of crack initiation region should increase from No.2 to No.5 steel, and then decrease dramatically for No.6 steel.

Fig. 10.   SEM fractographs of crack initiation regions of (a) No.2, (b) No.3, (c) No.4, (d) No.5 and (e) No.6 longitudinal impact specimens.

It is generally accepted that the grain boundary segregation of S, As and P has a bad effect on the strength of grain boundary. RE addition facilitates the formation of RE inclusions by consuming S, As and P and influences the distribution of impurity elements in the steel matrix. Therefore, with the increase of RE content, grain boundary will be purified by alleviating the grain boundary segregation of S, As and P gradually because more and more RE inclusions containing S, As and P form as mentioned above, which is conducive to the strength of grain boundary. Moreover, according to previous investigations, solute RE can enhance grain boundary cohesion through its segregation at grain boundary [31,32], and the RE content in solid solution probably increases with the increase of RE content, thereby resulting in an increasing trend for the strength of grain boundary. However, large inclusions caused by excessive RE addition easily lead to great stress concentration and the existence of intergranular crack during heat treatment due to the difference of thermal expansion property between matrix and inclusions, which significantly deteriorates the strength of grain boundary of No.6 steel. In order to demonstrate that the increase of RE content helps alleviate grain boundary segregation, samples from No.2 and No.4 steels are taken to be examined by SIMS. As can be seen from Fig. 11, the distribution of P changes from grain boundary segregation to mainly concentrating in RE inclusions, so the strength of grain boundary of No.4 steel is correspondingly higher than that of No.2 steel. In view of low As content and the possibility that the grain boundary segregation of S and RE elements could be covered up by RE inclusions, the grain boundary segregation of S, As and RE does not show a clear difference between No.2 and No.4 steels (not shown here). It is to be noted that the increase of the strength of grain boundary will decrease the possibility and the area of intergranular fracture and in turn contribute to transgranular fracture, which causes more energy consumption of crack initiation region.

Fig. 11.   SIMS images and the distribution of P in (a) No.2 and (b) No.4 impact specimens.

Fig. 12 exhibits crack initiation and slow propagation regions of fracture surfaces examined by SEM-EDS. It is founded that RE inclusions exist between intergranular fracture and transgranular fracture in No.3 and No.4 steels, indicating that RE inclusions may induce the change of fracture mode from intergranular fracture to transgranular fracture, thereby changing the path of crack growth and increasing energy consumption. Of course, RE inclusions could also have an auxiliary effect on the area of intergranular fracture compared with the strength of grain boundary through inducing transgranular fracture. Nevertheless, large inclusions in No.6 steel tend to accelerate crack growth in intergranular fracture mode due to the existence of intergranular cracks, which certainly increases the area of intergranular fracture and decreases energy consumption. The similar role of inclusions for energy consumption applies to crack fast propagation region, as shown in Fig. 13. It is evident that there are RE inclusions located at the bottom of steps on the fracture surface in No.3 and No.4 steels, which implies that RE inclusions can change the path of crack propagation and increase the concave and convex extent of fracture surface, thereby increasing the area of fracture surface and energy consumption. Conversely, large inclusions, probably with intergranular cracks around, still play a main role in facilitating intergranular fracture and crack propagation, so large area of intergranular fracture and low energy consumption are obtained in the crack fast propagation region. Based on Fig. 12, Fig. 13, it can be concluded that RE inclusions with appropriate size range are able to change the path of crack growth and increase energy consumption in the crack initiation and propagation regions, whereas large inclusions reduce energy consumption by promoting intergranular fracture. According to the above results about the size, volume fraction and distribution of inclusions in the experimental bearing steels with different RE contents, it could be deduced that energy consumption, from the perspective of inclusions, should increase from No.2 steel to No.5 steel and suddenly decrease for No.6 steel in the crack initiation and propagation regions.

Fig. 12.   SEM fractographs of crack initiation and slow propagation regions from (a) No.3, (b) No.4 and (c) No.6 longitudinal impact specimens.

Fig. 13.   SEM fractographs of crack fast propagation region from (a) No.3, (b) No.4 and (c) No.6 longitudinal impact specimens.

Fig. 14 shows fracture morphology close to crack source in a relatively large area. As with the analysis results above, the area of intergranular fracture decreases from No.2 steel to No.5 steel and then increases sharply for No.6 steel on account of the opposite variation of the strength of grain boundary and auxiliary influence of inclusions, whereas the concave and convex extent of fracture surface has an increasing trend from No.2 steel to No.5 steels followed by a sudden decrease for No.6 steel due to the effect of RE inclusions on crack growth path. Smaller area of intergranular fracture and larger concave and convex extent contribute to more energy consumption, so energy consumption should increases from No.2 steel to No.5 steels and decrease dramatically for No.6 steel. Considering the stress concentration at the crack tip in high strengthen bearing steel during crack growth, the energy consumption of these regions as shown in Fig. 14 should play an important role in impact absorbed energy for experimental bearing steels with different RE contents, respectively. In addition, the effects of RE inclusions and the strength of grain boundary on energy consumption can further be extended to larger area of fracture surface. Therefore, impact absorbed energy of experimental bearing steels continually increases from No.2 steel to No.5 steel followed by a sharp decrease for No.6 steel, as shown in Fig. 8(a).

Fig. 14.   Fracture morphologies close to crack source of (a) No.2, (b) No.3, (c) No.4, (d) No.5 and (e) No.6 longitudinal impact specimens.

Based on impact absorbed energy and the analysis result of fracture surfaces together with inclusion composition, size and morphology of experimental bearing steels containing different RE contents, Fig. 15 illustrates the effect of inclusions on crack growth of test impact specimens with different RE contents by modeling impact process. Fig. 15(a) shows that crack growth cuts off elongated MnS or complex inclusion for No.1 longitudinal impact specimen, so more energy consumption and high impact absorbed energy are obtained due to excellent plasticity of MnS. Conversely, elongated MnS and complex inclusion promote crack growth of No.1 transverse impact specimen and result in low impact absorbed energy, as depicted in Fig. 15(b). Fig. 15(c) indicates that RE inclusions within certain size range change the path of crack growth. Larger inclusions correspond to bigger change of crack path and larger area of fracture surface, thus consuming more energy. That is why impact absorbed energy continually increases from No.2 steel to No.5 steels. No.6 steel has many large RE inclusions containing C with large aspect ratio. Fig. 15(d) reveals that crack grows fast along parallel inclusions with intergranular voids around, which significantly reduce energy consumption. In contrast, cutting off brittle RE inclusions containing C, especially incompletely aggregate inclusions, does not cause much change of energy consumption. Therefore, No.6 steel has extremely low impact absorbed energy.

Fig. 15.   Models of crack growth influenced by inclusions: (a) No.1 longitudinal impact specimen; (b) No.1 transverse impact specimen; (c) No.2-5 impact specimens; (d) No.6 impact specimen.

4. Conclusions

(1)RE addition could modify irregular Al2O3 and MnS into relatively regular RE inclusions because of its active chemical property. With the increase of RE content, the reaction sequence of RE and potential inclusion forming elements should be O, S, As, P and C successively. The capture of As and P by RE inclusions alleviated grain boundary segregation of harmful residual elements, which benefits the strength of grain boundary. RE inclusions containing C could form easily when RE content was higher than 0.14 wt%. These inclusions might precipitate in molten steel and solid state, but the precipitation temperature was significantly higher than that of carbides in high-carbon chromium bearing steel.

(2)For experimental bearing steels, the volume fraction of inclusions increased steadily with the increase of RE content, but small amount of RE addition could make inclusions smaller and more dispersed compared with conventional high-carbon chromium bearing steel. The continuous increase of RE content led to an increasing trend for inclusion size with the growth mechanism of inclusions changing from precipitation-grown to aggregation-grown. Meanwhile, the distribution and morphology of inclusions gradually deteriorated as well. However, the volume fraction of large inclusions larger than 5 μm did not show a clear increase until excessive RE addition.

(3)Elongated MnS of large size, whether in the form of isolated MnS or complex inclusions, resulted in low transverse impact toughness and isotropy of high-carbon chromium bearing steel, which could be greatly improved by RE addition. For the modified high-carbon chromium bearing steel by RE alloying, the increase of RE content within a certain range enhanced the strength of grain boundary and the variation of crack growth path, so both transverse and longitudinal impact toughness had an increasing trend. However, excessive RE addition seriously deteriorated the impact toughness due to rapidly increasing volume fraction of large inclusions.

(4)From the perspective of impact toughness, the optimum content of RE elements was aout 0.14 wt% for experimental bearing steels. However, considering the detrimental effect of inclusions on fatigue property, RE content should be controlled around 0.018 wt%, where impact property was still better than that of bearing steel without RE.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (Nos. U1508215 and U1708252), the National Key Research and Development Program (No. 2016YFB0300401).

The authors have declared that no competing interests exist.


/