Effect of Acid Slag Treatment on the Inclusions in GCr15 Bearing Steel

Abstract By laboratory slag/steel reaction equilibrim experiments, the viriation of oxygen content, inclusion compositions and inclusion sizes were studied. The effect of acid slag treatment on the transition mechanisms of D-type inclusions and the precipitation of TiN inclusions in GCr15 bearing steel were explored. The obtained results showed that the dominant inclusions in steel were plastic and smaller Al2O3-SiO2-MnO. The melting point were lower than 1400°C treated by the acid refining slag of 35.1%CaO-15%Al2O3-43.9%SiO2-6%MgO and there was no TiN found. The evolution of MgO·Al2O3 inclusions is: MgO·Al2O3→ MgO·Al2O3·SiO2·MnO→ Al2O3·SiO2·MnO. Mg and Al from MgO·Al2O3 inclusions were displaced by [Si] and [Mn] in steel liquid , and formation of plastic Al2O3-SiO2-MnO inclusions finally, whose compositions distribution were uniform. Mg and Si, Mn were complementary in inclusions as to the spatial distribution.


Introduction
As one of the most commonly used high-chromium bearing steels, GCr15 has been widely used in manufacturing bearing ring, ball screw and other mechanical components [1]. There are three main factors affecting contact fatigue life of bearing steel: hardness, inclusions and hydrogen content in steel [2]. The development of bearing steel in the twentieth century was decreasing T.[O] [3], and the work highlight to improve fatigue life of bearing steel was reducing the sizes of inclusions [4]. Meanwhile, many researchers had approved and verified that quantities, compositions, morphologies and sizes of inclusions in steel are the main factors which affect the fatigue life of bearings. In general, the quantities and sizes of inclusions in steel could be reduced by controlling the process of steelmaking [5]. Therefore, it is important to explore suitable slag compositions to control the inclusions in bearing steel.
Although there is low oxygen content in GCr15 after high basicity refining slag treatment, which achived high cleanliness, most were D-type inclusions including MgO·Al 2 O 3 , MgO and calcium-aluminate [7]. And the inclusions such as MgO·Al 2 O 3 , calcium-aluminate and christobalite had no plastic deformation ability under the conventional hot working temperature of steel. It can not be in good shape with the base body when rolling. The stress concentration can be caused between the steel matrix and the interfacial, which caused fatigue cracks and fatigue life reduction of the bearing steel [8]. Jiang [9] found that the inclusions with low melting point had good deformation ability, and they were found in liquid form with special shape which were easier to remove. Ders' theoretical calculation [10] showed that lowering the melting point of inclusions (softening inclusions) could not only effectively increase the plastic deformation ability, but also eliminate the stress concentration. Bernard [11] reviewed that the deformation ability of inclusions had an important relationship with its melting temperature: there were good plastic deformation when the melting temperature of the inclusion was less than 1673K.
In order to reduce the harmfulness of D-type inclusions in GCr15, it is necessary to be transformed into inclusions with low melting point and plastic deformation ability. How the acid slag modified the inclusions in bearing steel was seldom reported. The treatment effect of acid slag basicity and compositions on D-type inclusion were analyzed. The viriation of oxygen content, inclusion sizes and compositions, the effect of acid slag treatment on precipitation of TiN-type inclusion in steel and the evolution mechanism of MgO·Al 2 O 3 inclusion during the treatment process were investigated.

Experimental procedure
Experiments were carried out in a tubular resistance furnace. The billet pieces of GCr15 bearing steel, which treated by high basicity refining slag, were used as start material. The chemical composition of GCr15 is listed in Table 1 and the compositions of the high basicity refining slag is listed in Table 2. The billet pieces were melted in Al 2 O 3 crucible under argon protective atmosphere. The argon volume flow and pressure were monitored by a flow meter and a pressure gauge.
The detailed experimental process is described as follows: (1) 70g billet blocks and 10.5g acid slag were placed into the Al 2 O 3 crucible to melt in the tubular resistance furnace. (2) The furnace was heated up to 1600 ∘ C gradually under argon protective atmosphere with a volume flow of 4 L·min −1 . (3) After melting clear, the temperature was maintained at 1600 ∘ C for one and half an hour to homog-enize the chemical composition. (4) The samples was put out and placed into a copper plate, and then cooled down by water quenching. There are four types of acid slag, and the compositions of them are listed in Table 3.
To explore the transitional process, the sample with best acid slag treatment was picked for different holding time at 1600 ∘ C, such as 10min, 20min, 40min, 60min and 80min. Finally, the samples were cooled down by water quenching The concentration of carbon(C) and sulfur(S) in the steel samples were analyzed by infrared carbon and sulfur analyzer ( The morphologies of inclusions in the samples observed by scanning electron microscope and the chemical compositions of inclusions were analyzed with energy dispersive X-ray spectroscopy (SEM-EDS, Model: MLA250). The quantities and sizes of the TiN-type inclusions in the steel samples were analyzed by automatic inclusion analysis system (Model: EVO18-INCAsteel).

Effect of different acid slags treatment
The main inclusions in the steel samples before acid slag treatment were D-type inclusions, including MgO-Al 2 O 3 , TiN, MgO and CaO-MgO-Al 2 O 3 . The types of inclusions in different acid slags treatment steel samples are listed in Table 4. The main inclusions in S1, S2, S4 are CaO   The inclusions of S1, S2 and S3 were CaO and MgO, and (CaO+MgO)≈ 5%, therefore, they could be projected to Al 2 O 3 -SiO 2 -MnO-5%MgO phase diagram. While the inclusions of S3 should be projected to Al 2 O 3 -SiO 2 -MnO phase diagram (see Figures 1-4).
The melting point of inclusions in S1, S2, S3 and S4 showed a downward trend after acid slag treatment, and all of them had some inclusion projection points distributing in the low melting area less than 1500 ∘ C. This kind of inclusion was liquid state in molten steel, which benefit to colliding, growing up and floating out [13]. There were 10% inclusion projection points in S2 and 40% in S3 distributing in the low melting area less than 1400°C. And the distribution of the inclusion projection points in S3 were concentrated, while they were dispersed in S1, S2 and S4.
Inclusion sizes of the start material, S1, S2, S3 and S4 were summarized on the basis of the statistics of 50 inclusions, and the total oxygen is shown ( Figure 5). The results showed that the sizes of the inclusions before treatment distributed in 1~10 µm, while the sizes of the inclusions after treatment distributed in 1~5 µm. The inclusion sizes in S1, S2, S3 and S4 were decreased, and the average inclusion size of S3 was the smallest, which was 2.05 µm. The total oxygen of S1 and S4 increased much, while it's almost invariant in S3. Above all, the S3 had the best treatment effect.

Effect of acid slag treatment on TiN-type inclusion precipitation
With the oxide inclusions greatly reduced in bearing steel, the harmfulness of TiN inclusions gradually revealed, which is second only to calcium aluminate. The generating condition of TiN-type inclusions could obtain by the following thermodynamic analysis [14,15]: [Ti] + [N] = TiN (s) ,   Where, K in Equation (4) represents the equilibrium constant of Equation (3); α TiN (s) represents the activity of TiN in the slag; α [Ti] represents the activity of Ti in steel; α [N] represents the activity of N in steel; f Ti represents the activity coefficient of Ti in steel; f N represents the activity coefficient of N in steel; [%Ti] represents the mass fraction of Ti in steel; [%N]represents the mass fraction of N in steel. Take the logarithm of both sides of the Equation (4) and combine the Equation (5), leading to the Equation (6). The other elements in steel are little comparison with Fe, therefore it could ignore the effect of the activity coefficient of other elements on Ti and N in steel. The Equation (6) could simplify to Equation (7) and calculate the minimum required content of N and Ti to precipitate the TiN-type inclusions in the steelmaking and solidification process.
If there is 0.005%Ti in steel, the minimum required content of N is 0.448% at 1600 ∘ C; the minimum required content of N is 0.448% at 1500 ∘ C; the minimum required content of N is 0.143% at 1460 ∘ C. Therefore, it can't precipitate the TiN-type inclusions in the steelmaking process above the liquid line temperature(1450 ∘ C).
In the steelmaking and solidification process, the solubility of N and Ti in steel decreased with the temperature decreasing, Equation (8) represents. The stability interrelationship is shown in Figure 11.
To research the effect of acid slag S3 on the TiN-type inclusions, 60g piece of steel sample and 9g slags were put   Table 5. Most TiN inclusions are about 5µm, and some bigger than 8µm before treatment. There is no TiN inclusion found after treatment, but it can't be the direct absorption of the acid slag S3.
In the steel-slag balance experiment of alkaline slag, a layer of solid-state CaTiO 3 containing Ti were produced in the steel/slag interface, which prevent Ti diffusing to slag through steel/slag interface [16].
The SEM observations and thermodynamic calculations showed that calcium titanate is formed by the oxidation of dissolved titanium according to SiO 2 + CaO + Ti → CaO.TiO 2 + Si (9) There is no evidence for the direct reaction at the slag/inclusion interface SiO 2 + CaO + TiN → CaO.TiO 2 + Si + N (10) Reaction (9) is only observed when the CaO content of the slag is high enough to enter the CaTiO 3 field of crystallisation in the CaO-TiO 2 -SiO 2 phase diagram, as already described by Kishi et al. [17]. There was not such solid-state CaTiO 3 in the S3 slag/steel interface, and Ti could diffuse to S3 slag. Therefore, the content of Ti in steel decreased much (Table 6). TiN inclusions could be controlled by reducing the content of Ti in steel with acid slag S3 treatment.
With the high alkalinity slag refining, although the total oxygen is very low, there is a significant increase in the occurrence of D-type point inclusions in steel. Acid slag treatment after LF refining, not only reduce the harmfulness of D-type point inclusions to bearing steel, still has a well control to precipitation of TiN-type inclusions.