Effect of Current on Segregation and Inclusions Characteristics of Dual Alloy Ingot Processed by Electroslag Remelting

Experimental procedure

The consumable electrode consisted of NiCrMoV alloy bar (upper part) and CrMoV alloy bar (lower part) by welding was remelted in a laboratory-scale electroslag furnace with different currents under an open air atmosphere. The inner diameter of the copper mold is 120 mm and the transformer capacity is 160 kVA. The diameter and length of both alloy bars is 55 mm and 800 mm, respectively. The chemical composition of the consumable electrode used in each heat is listed in Table 1. The slag composed of 70 wt% CaF₂ and 30 wt% Al₂O₃, the thickness and weight of the slag were about 60 mm and 2.3 kg, respectively.

Three heats were designed to investigate the effects of current on characteristics of inclusions during ESR process of dual alloy ingot. The different currents adopted in ESR experiment are listed as follows: case A (the first heat) – 1500 A, the remelting rate is approximately 32 kg/h; case B (the second heat) – 1800 A, the remelting rate is approximately 39 kg/h; case C (the third heat) – 2100 A, the remelting rate is approximately 44 kg/h. The working voltage was about 30–45 V and the cooling water pressure was 0.22 MPa. The electrode immersion depth was about 10 mm and was controlled by an electrode lifting device powered by an electromotor.

Chemical analysis and microscopic observation

After ESR process, three dual alloy ingots were obtained. The each ingot was evenly divided into two halves along the length direction by wire-electrode cutting. Steel filings were obtained by drilling along the transverse radius every 15 mm for Cr analysis, as shown in Figure 1. Three slices were taken from the upper (NiCrMoV), middle (Transition zone (TZ)), and the lower parts (CrMoV) of each ingot, and three 6 mm×6 mm×6 mm specimens were then sampled from three slices, respectively. The steel filings were obtained from the corresponding position for composition analysis by a drill machine. Carbon, sulfur and oxygen contents were measured by the carbon and sulfur analyzer, and oxygen and nitrogen analyzer. The manganese, silicon, chromium and nickel contents in the ingot were analyzed by the ICP-AES. In order to observe the inclusions, the specimens were grinded and polished, and then examined using SEM with EDS. The number and size distribution of inclusions were counted by Image Pro-Plus 6.0 software with photo from 50 randomly visual fields of the SEM at 2000 magnification. Then the specimens were etched at 75 °C in picric acid solution for metallographic observation by OM.

Figure 1:

Schematic drawing of the dissection of the ESR ingot.

The segregation along the transverse radius

The distributions of Cr concentration along the transverse radius in ingots with different currents are shown in Figure 2. It indicates that Cr content is higher at the center of all ingots than the edge. And the segregation becomes severer with the increase of the current, which is mainly affected by the solute transport [2]. The flows in molten metal pool have a dominant influence on the solute transport [2, 15].

Figure 2:

The distributions of Cr concentration along the transverse radius in ingots processed by electroslag remelting with different currents.

Figure 3 shows the illustration of the flows of the molten metal pool in the ESR process. The flows in the molten metal pool are driven by two forces: thermal buoyancy force induced by the temperature gradient, and inward Lorentz force, which arises from the flow of current through the pool [19, 20]. The distribution of thermal buoyancy force and Lorentz force has been well studied in previous researches [16, 21, 22]. The hot metal floats up and the cool metal sinks down. The clockwise circulations are formed in the vicinity of the mold wall and the oblique solidification front due to the temperature gradient. The cool metal with a higher density will move down along the oblique solidification front and wash out the solidifying mushy zone. The cooling intensity weakens around the base of the molten metal pool. The hot metal rises toward the slag-metal interface and then returns to the mold wall. Furthermore, according to the Faraday’s law of electromagnetic induction, a clockwise circular magnetic field (looking down from the top) would be induced by the downward current. The interaction between the clockwise circular magnetic field and the downward current creates an inward Lorentz force, which also pushes the metal from the edge to the middle.

Figure 3:

Illustration of the flows of the slag bath and metal pool in the ESR process.

The Cr become enriched in the mushy zone of solidification front due to the partition ratio (k_Cr=0.76) [23]. In addition, the density of Cr (ρ_Cr=6900 kg/m³) is lower than that of iron (ρ_Fe=7500 kg/m³) [9]. The Cr would be enriched in the molten metal pool due to the buoyancy force, resulting in the so-called gravity segregation. At the solidification front, the solute-poor metal displaces the solute-rich metal through washing out the mushy zone due to the clockwise circular flow. Inward Lorentz force also pushes the metal from the edge to the bottom center of metal pool. As a result, the Cr accumulates at the bottom center of the pool and the concentration decreases from the middle to the edge. With the current increasing, the flows in metal pool become more intense due to the fiercer thermal buoyancy force and Lorentz force and the molten metal pool becomes deeper, resulting in a severer segregation along the transverse radius.

The number and size distribution of inclusions

Table 2 shows the statistical results of inclusions in electrode and ingots. It can be seen that the number of inclusions observed in electrode is far more than that in each ingot, and the maximum size of inclusion in Elec. NiCrMoV and Elec. CrMoV is 12.01 μm and 6.68 μm, respectively, which is larger than that of ingots. The ratio of total area of inclusions to total area of observed view fields, S_a, is also an essential index. The value of S_a is larger in electrode, implies that an improved cleanliness is obtained in ingots. As seen in Table 2, the number of inclusions and S_a in NiCrMoV zone of each ingot is larger than that in CrMoV zone, implies that cleanliness of ingot is closely related to the electrode. It is interesting that the inclusion size and S_a in transition zone are smaller than that in NiCrMoV zone when the current is 1800 A or 2100 A, whereas, the larger inclusion size and S_a are found in transition zone when the current is 1500 A. The inclusion size has a close relationship with the type of inclusion, which will be discussed in detail later. With the increase of current, the number of inclusions, S_a and total oxygen (T.[O]) content (Table 2) in every zone of ingot increases, indicating a worse cleanliness with higher current.

Figure 4 shows the size distribution of inclusions observed in consumable electrode and ingots. It indicates that size of inclusion in the ingot is finer and more uniform than that in electrode (Figures 4(a-c)). It can be seen from Figures 4(d-f) that the size of inclusions in NiCrMoV zone, transition zone and CrMoV zone of ingot distribute more evenly with a lower current. The proportion of the inclusions with large size (>3 μm) increases with the current increasing, implying that more inclusions with large size exist in the ingot processed by the larger current, which is detrimental to the performance of ingot.

Figure 4:

Size distribution of inclusions observed in consumable electrode and each dual alloy ingot processed by ESR with different currents.

There are two reasons for the deterioration of cleanliness of ingot processed by larger current. One is the electrode surface oxidation, and the other is the remelting rate. With the increase of current, the temperature of the electrode surface is higher and severer electrode surface oxidation occurs due to the faster mass transfer of oxidation reaction. The products of electrode surface oxidation enter into molten slag to increases the FeO content of slag. The FeO in slag can be decomposed into [Fe] and [O] in liquid steel at the steelmaking temperature, resulting in the increase of oxygen content in ingot during ESR. Furthermore, the surface of electrode inserted into the slag pool is liquid, which is a film of molten steel; it will flow downward along electrode surface to form the droplet at electrode tip. The removal of inclusions from steel mainly occurs during this process [14]. Then the droplet will be detached from the electrode tip when the droplet reaches a certain mass. The faster remelting rate makes the formation time of the droplet shorter, namely, the interaction time between slag pool and film of molten steel at the electrode tip is shorter, which is detrimental to the removal of inclusions. Moreover, more droplets drip almost simultaneously just likes rain under the faster remelting rate with a larger current, the size of droplets is larger than that for low current. The contact between the droplets and slag pool is not very abundant for a large current condition, which is also detrimental to the removal of inclusions and harmful elements (sulfur element). To quantitatively analyze the effect of current on the removal of inclusion, the inclusion removal rate was roughly calculated according to the S_a of inclusions in Table 2. With the current increasing from 1500 A to 1800 A and 2100 A, the remelting rate increases from 32 to 39 kg/h and 44 kg/h, the inclusion removal rate in NiCrMoV zone decreases from 85.4% to 56.7% and 39.2%, that in CrMoV zone decreases from 77.8% to 74.3% and 56.1%.

The morphology and types of inclusions

In addition to the size and distribution of inclusions, the chemical composition and morphology of inclusion is also an effective standard to assess steel cleanliness [24], especially the inclusion evolution mechanism during ESR. So the morphology and types of inclusions in the consumable electrode and dual alloy ingots processed under different currents were determined.

Figure 5 shows the SEM images of typical inclusions observed in electrode. It indicates that the inclusions in the electrode are mainly large Al-Ti-Mn-Si-Ca-Mg-Cr oxides (Figures 5(a-c), 5(e-g)) and (Mn,Cr)S (Figures 5(d), 5(h)). Some (Mn,Cr)S inclusions are distributed at the edge of the oxide (Figures 5(a), 5(e), 5(g)). The single long strip (Mn,Cr)S inclusions are also observed in the electrode (Figures 5(d), 5(h)). Figure 6 shows the SEM images of typical inclusions in dual alloy ingot. When the current is 1500 A, the inclusion in NiCrMoV zone of ingot consists of Al-Mn-Cr oxides covered with a layer of (Mn,Cr)S (Figure 6(a)) and no single (Mn,Cr)S inclusion is observed. In transition zone, the inclusions are complex inclusion with two-layer structure and single (Mn,Cr)S inclusion. A thicker outer layer of (Mn,Cr)S and large single (Mn,Cr)S inclusion increase the average size of inclusions and S_a in transition zone, which are larger than that in NiCrMoV zone when the current is 1500 A (As mentioned before). In CrMoV zone of ingot, the inclusions are complex inclusion with two-layer structure and the Al-Cr-Mn oxides. With the current increasing to 1800 A and 2100 A, the single (Mn,Cr)S are observed in NiCrMoV zone of ingot. The type of inclusions in transition zone is similar with that under the condition of 1500 A. In CrMoV zone of ingot, the inclusion is complex inclusion with two layers structure.

Figure 5:

SEM images of typical inclusions observed in consumable electrodes: (a-d) specimen Elec. NiCrMoV, (e-h) specimen Elec. CrMoV.

Figure 6:

SEM images of typical inclusions observed in ESR dual alloy ingot: (a-e) specimen 1500 A, (f-j) specimen 1800A, (k-o) specimen 2100 A.

Figure 7 shows SEM image and element mappings of a typical inclusion in the form two-layer structure observed in the ESR dual alloy ingot. The inclusion is spherical and the size of inclusion is about 4.5 μm (Figure 7(a)). Elemental mappings of the inclusion (Figures 7(b-f)) indicate that the core of inclusion consists of Al and O. The S and Mn are observed on out layer of the inclusion, a small amount of Cr is also observed, which is seen as Al oxides covered with an outer layer of (Mn,Cr)S. The precipitation of (Mn,Cr)S on the surface of complex oxide increases the size of inclusion.

Figure 7:

SEM image and element mappings of a typical inclusion in the form two-layer structure observed in the ESR dual alloy ingot.

Desulfurization and thermodynamics of sulfide inclusions formation during ESR process

The chemical composition of dual alloy ingots is listed in Table 2. It indicates that S content of CrMoV zone of each ingot is lowest and that of NiCrMoV zone is highest due to the lower S content in Elec. CrMoV. With the increase of current, the S content of every zone in ingots increases, implying that the larger current has a negative effect on desulfurization. The desulfurization rate in NiCrMoV zone decreases from 56% to 47%, which in CrMoV zone decreases from 84% to 55% (Tables 1 and 2). During the ESR process, the (Mn,Cr)S is decomposed into the [Mn], [Cr] and [S] in the liquid steel, then the [S] in the liquid is removed by the gasifying desulfurization and the interaction between slag and steel. The chemical reaction for desulfurization by the interaction between slag and molten steel can be expressed as:

where K^Θ is the equilibrium constant of reaction, the equilibrium distribution constant of sulfur between slag and molten steel (L_s) can reflect the desulfurization capacity of slag, which can be expressed as:

where x(S^2-) and w[S] is the mole fraction of sulfur in slag and mass fraction of sulfur in the molten steel, respectively. γO2− and γS2− is the activity coefficient of oxygen ion and sulfur ion in the slag, respectively. fO and fS are the activity coefficients of oxygen and sulfur in the molten steel. The desulfurization mainly occurs at the interface between slag and metal film of electrode tip and the increasing current can increases the temperature [25], which can accelerate the mass transfer of reaction (1) and is favorable for gasifying desulfurization by the interaction between slag and air. Whereas, the large current also results in the severer electrode surface oxidation, which results in a higher oxygen content in steel and a higher FeO content in slag. It can be seen from eq. (1) that desulfurization by the reaction between slag and molten steel is dominated by oxygen content in steel and the activity of oxide ion in slag. The lower oxygen content in liquid steel and higher oxide ion activity in molten slag contribute to deeper desulfurization. FeO in molten slag suppresses desulfurization because it can increase the dissolved oxygen content in molten steel ((Fe²⁺) + (O^2-)=[Fe] + [O]) [26]. In present experiment, the FeO in molten slag contributes less to the increase of oxide ions activity in slag, but leads to an increase in oxygen content in liquid steel, playing a negative role on desulfurization.

The desulfurization rate decreases with the increase of current. As a result, the single (Mn,Cr)S inclusions are observed in NiCrMoV zone of ingot with 1800 A and 2100 A, which are not found in that of ingot with 1500 A. It can be seen from Figure 5 that only a small amount of Cr exists in single (Mn,Cr)S inclusions of NiCrMoV zone and transition zone (TZ). The atomic ratios of Mn and Cr (x(Mn)/x(Cr)) in single (Mn,Cr)S inclusions are large. Shi et al. revealed the precipitation of (Mn,Cr)S is dominant by concentration of Mn, Cr and S [24]. It can be concluded from Ref. [24] that the (Mn,Cr)S with larger x(Mn)/x(Cr) is easier to precipitate from the view of thermodynamics. After the ESR, the single (Mn,Cr)S inclusions may precipitate when the product of [Mn], [Cr] and [S] in the liquid steel exceeds the equilibrium value for (Mn,Cr)S precipitation at certain temperature. To study the formation of sulfide in detail, the single (Mn,Cr)S inclusions with high x(Mn)/x(Cr) in the ingot is studied as pure MnS inclusions due to the lack of relevant reliable data.

The chemical reaction for formation of MnS inclusions in liquid steel can be expressed as [27]:

Where K is the equilibrium constant and T is the temperature (K). The activity of MnS is unity (pure solid MnS as the standard state) in the present study. The equilibrium constant K can be expressed as:

where f_M is the activity coefficient of dissolved element M, which can be calculated using the following equation [26]:

where eij is interaction parameter between elements. Interaction parameter eij at 1600 °C is listed in Table 3 [28]. The liquidus temperature T_L and solidus temperature T_S of the steel can be calculated according to the corresponding eqs. (7–8) [29, 30], which is listed in Table 2.

The equilibrium of Mn-S reaction in liquid steel at calculated liquidus and solidus temperature were calculated according to the relevant data (Tables 2 and 3) related with eqs. (3–6). Only the precipitation of MnS in NiCrMoV zone and transition zone were calculated because the sulfur content in CrMoV zone is low and no single MnS is observed in CrMoV zone. Figure 8 shows stability diagram of MnS precipitation in NiCrMoV zone and transition zone of each ingot. It indicates that MnS inclusion is unable to precipitate in liquid steel with the Mn and S contents of NiCrMoV zone and transition zone above the solidus temperature. However, during the cooling process of liquid steel in the water-cooled mold from the liquid metal pool temperature to the liquidus temperature, and then to the solidus temperature, the solute is rejected into interdendritic liquid phase, which results in the enrichment of Mn and S in residual liquid steel between solid steel dendritic arms at the solidifying front because of the difference in solubility of solutes between liquid and solid phases [31]. The concentrations of [Mn] and [S] in liquid steel during solidification of liquid steel were calculated by eqs. (9) to (12) based on the simple microsegregation model [32].

Figure 8:

Stability diagram of MnS precipitation in the dual alloy ingot: specimen 1500 A, specimen 1800 A, specimen 2100 A.

where T₀, T_L and T_S represent the melting point of pure iron (1809 K (1536 °C)), liquidus and solidus temperature of the studied steel, respectively, C_L is the concentration of a given solute element in the liquid at the solid-liquid interface, C₀ is the initial (nominal) liquid concentration, k (=C_S/C_L) is the equilibrium partition coefficient for that element, f_S is the solid fraction, α is a back-diffusion parameter, D_S is the diffusion coefficient of solute in the solid phase in cm²/s, t_f is the local solidification time in seconds, λ_S is the secondary dendrite arm spacing (SDAS) in cm, C_R is the cooling rate ( °C/s) and can be calculated by eq. (13) [32]. Figure 9 shows the optical microstructures of NiCrMoV and transition zone in each the dual alloy ingot processed by ESR with different currents. The SDAS of NiCrMoV zone from 1500 A to 1800 A and 2100 A obtained from Figure 9, is 61.5 μm, 49.6 μm and 65.3 μm, respectively, which in transition zone is 49.2 μm, 40.3 μm and 49.7 μm.

Figure 9:

Optical microstructures of NiCrMoV zone and transition zone (TZ) in each dual alloy ingot processed by ESR with different currents.

The temperature at the solid-liquid interface, T, is given by eq. (14) [33].

In present study, only the precipitations of MnS in the NiCrMoV zone of ingot with 1500 A and 1800 A and transition zone of ingot with 1500 A were calculated as an example. According to Fe-Cr-C ternary phase diagram, the solidification of the steels in this experiment is in solid (δ) phase. Data of equilibrium partition coefficients and diffusivity of solute elements are shown in Table 4 [34]. The relationship between the product of [%Mn]×[%S] and the solid fraction f_S is shown in Figure 10. The calculated equilibrium value of [%Mn]×[%S] for MnS precipitation and the calculated temperature at the solid-liquid interface through the application of eqs. (3–6, (14)) are also shown in Figure 10. It can be seen that the actual [%Mn]×[%S] increases with the increasing of the solid fraction. When the actual [%Mn]×[%S] exceeds the equilibrium [%Mn]×[%S], MnS would precipitate in the residual liquid phase at the solidification front. The single MnS could not precipitate in the NiCrMoV zone in ingots with 1500 A (Figure 10(a)), whereas, precipitate in that with 1800 A (Figure 10(c)) when the solid fraction (f_s) reaches to 0.985 due to higher S content and solute segregation. The Mn content in Elec. CrMoV is higher than that in Elec. NiCrMoV, whereas, the sulfur is lower in Elec. CrMoV. During the ESR processing, the Mn content decreases from bottom to the top of the ingot and the sulfur content increases from bottom to the top. In the transition zone, the product of [%Mn]×[%S] reaches to the maximum, then the single MnS inclusion would precipitate during solidification. It can be seen from Figure 10(b) that the critical value of solid fraction (f_s) for MnS precipitation is 0.973. The calculated results show that MnS can’t precipitate in the NiCrMoV zone of ingot with 1500 A. The pure MnS precipitates easier than the (Mn,Cr)S [24]. That is to say that (Mn,Cr)S can’t precipitate in the NiCrMoV zone of ingot with 1500 A. Several studies have shown that MnS heterogeneous nucleates easily on some oxides due to lower nucleation energy [35, 36, 37] and the oxides covered a layer of (Mn,Cr)S was found in the dual alloy ingot. These calculated results agree well with the SEM/EDS results. In order to eliminate the (Mn,Cr)S inclusion, a deeper desulfurization is needed.

Figure 10:

Relationship between concentration products of [Mn], [S] and solidification fraction: (a) specimen 1500 A-NiCrMoV, (b) specimen 1500 A-TZ, and (c) specimen 1800 A-NiCrMoV.