Evolution of Inclusions in Incoloy825 during Electroslag Remelting

Abstract

Fifty kilogram-scale electroslag remelting (ESR) experiments using slag with different TiO₂ contents in an electroslag furnace were performed to investigate the size, amount and types of inclusions in an electrode and remelted ingots. The results show that the contents of aluminum and titanium increased and decreased, respectively, compared to those in a consumable electrode. The inclusions in the consumable electrode were TiS, TiN and Al₂O₃ surrounded by a TiN layer, and the inclusions in the remelted ingots were TiN, MgO-Al₂O₃ and MgO-Al₂O₃ surrounded by a TiN layer. With the increase in the height of the ingots, the average size of inclusions increased from 2.5 to 4.4 μm. Increasing the TiO₂ content in the slag promoted the formation of MgO-Al₂O₃ and made the inclusions larger in the remelted ingots. To make the size of inclusions in remelted ingots of Incoloy825 smaller, the TiO₂ in the slag should be decreased. The TiS in the electrode was removed during ESR. Part of the TiN dissociated during the reaction between the liquid metal and slag by molten slag, and most of the TiN inclusions originated from the consumable electrode. The Al₂O₃ inclusions surrounded by a TiN layer in the consumable electrode were finally transformed to MgO-Al₂O₃ and MgO-Al₂O₃ surrounded by a TiN layer in the remelted ingots.

1. Introduction

Electroslag remelting (ESR) has been increasingly applied to produce special steels and super alloys because of its great advantage in removing non-metallic inclusions, which are gravely harmful to the quality of the product [1,2,3]. Incoloy825 is one of the typical high-temperature Ni-based alloys and, due to its exclusive, unique advantages, is widely used for chemical processing, pollution control and oil and gas recovery [4,5]. Regarding the existence of Ti and Al in this alloy, there are Ti-containing and Al-containing inclusions. To maximize the quality of Incoloy825, it is vital to investigate the evolution mechanism of inclusions in this alloy during the production process.

In order to understand the removal mechanism of inclusions, different inclusions have been investigated under specific production conditions. Shi et al. [6] studied the evolution of MnO-SiO₂-Al₂O₃ inclusions in Si-Mn-killed steel during protective atmosphere ESR and found they were removed by two methods and formed MgAl₂O₄ as a poor metal. Wang et al. [7] researched the formation and evolution of non-metallic inclusions in calcium-treated H13 steel. Li et al. [8] studied heating austenitic alloy at 1273 K and investigated the evolution of TiS and TiN inclusions. Several researchers [9,10,11] investigated the inclusions through simulation. Many researchers [12,13,14] reported that the inclusions in Ti-containing alloy and steel are related to the chemical composition and deoxidation. However, the evolution mechanism of inclusions in Incoloy825 alloy by ESR has not yet been reported.

To the authors’ best knowledge, the slag used in the ESR process contains high amounts of fluoride. The evaporation of fluoride not only pollutes the environment but also changes the composition of the slag. The exploitation of low fluoride-containing slag for ESR has been studied in previous research. Shi et al. [15] indicated that the TiO₂ in slag can reduce viscosity in place of CaF₂. Ju et al. [16] proposed that increasing CaO/Al₂O₃ in low-fluoride slag can reduce viscosity. Moreover, the TiO₂ in slag can also decrease the loss of Ti in an alloy [17,18]. However, less research of the effect of TiO₂ in low fluoride-containing slag on Ti-containing inclusions has been conducted.

The present study was undertaken to investigate the evolution of inclusions in Incoloy825 during the ESR process. The effect of TiO₂ in slag on inclusions was also revealed. The evolution of the mechanism of inclusions in Incoloy825 was clarified through the transition of inclusions during ESR in combination with thermodynamic calculation.

2. Experimental

2.1. Raw Material for ESR Procedure

The consumable electrode (85 mm in diameter and 850 mm in length) was produced by a 150 kg-capacity vacuum induction melting (VIM) furnace. The chemical composition of the consumable electroslag is shown in

Table 1. Chemical composition of the consumable electrode (mass%).

C	Mn	Si	P	S	Cr	Mo	Ni	Cu	Al	Ti	Fe	O	N
0.10	0.11	0.13	0.01	0.01	20.62	3.18	38.88	1.66	0.12	1.00	33.74	0.0026	0.0047

. The premelted slag was prepared by an electric arc furnace, whose composition is listed in

Table 2. Chemical composition of the slag used in ESR (mass%).

Experiments	CaF2	CaO	Al2O3	MgO	TiO2	FeO	SiO2
S1	18.60	35.59	37.24	1.51	3.47	1.15	2.44
S2	18.15	33.51	35.85	1.61	7.24	1.16	2.48
S3	18.09	30.95	33.53	1.67	11.84	1.27	2.65

. Before the ESR process, the oxide layer on the consumable electrode’s surface was polished and the dummy bars were welded together. The slags used for ESR were heated at 673 K for at least 6 h to remove the moisture.

2.2. ESR Procedure

Three experiments were carried out using a furnace with the capacity to refine a 50 kg ingot. Before arcing, 5 kg of premelted slag was added into the mold, and high-purity argon with a flow of 200 L·min⁻¹ was introduced into the mold to isolate it from the outside air. The alternate current and secondary voltage applied during the normal refining process were 1800 A and 32 V, respectively. No deoxidizer was added into the slag pool during the refining process. The average melting rate was about 1.2 kg/min, and the total refining process lasted about 900 s. The ingots were 160 mm in diameter and 205 mm in length.

2.3. Chemical Analysis and Inclusion Characterization

In order to observe the inclusion characteristics of ingots of different heights, four samples were taken from each ingot, numbered P1-1 to P1-4, P2-1 to P2-4 and P3-1 to P3-4, from top to bottom, with heights of 25 mm, 75 mm, 125 mm and 175 mm, respectively, as shown in Figure 1. Every sample was 15 × 15 × 15 mm³, spaced 50 mm apart. The contents of Al, Si, Ti, Cr, Ni, C, Mn, Mo, Cu and Fe in the electrode were measured by a spark direct reading spectrometer (METAL-LAB75/80J, GNR, Novara, Italia). The contents of total Ca, MgO, TiO₂, SiO₂ and Al₂O₃ in the slag were measured by wavelength dispersive X-ray fluorescence spectrometry (GoldXpert, Olympus, Tokyo, Japan), while the FeO in the slag was determined by a potassium dichromate volumetric method. The total oxygen and nitrogen contents in the samples were determined by an ONH analyzer (G8GALILEO, Bruker, Bremen, Germany). The preparation of samples for the size and morphology of the inclusions was conducted by grinding and polishing. Then, the inclusions were observed by an energy-dispersive X-ray spectrometer (EDS, Gemini 300, Bruker, Bremen, Germany).

3. Results and Discussion

3.1. Compositional Change of the Ingot

The contents of the elements in the inclusions that changed significantly in the remelted ingots are listed in

Table 3. Chemical composition of the samples from ingots (mass%).

Sample	Al	Ti	Mg	Si	S	O	N
P1-1	0.253	0.608	0.0016	0.19	0.01	0.0064	0.0047
P1-2	0.272	0.533	0.0013	0.2	0.01	0.0061	0.0047
P1-3	0.251	0.472	0.0015	0.22	0.01	0.0059	0.0059
P1-4	0.181	0.366	0.0012	0.26	0.01	0.0053	0.0058
P2-1	0.205	0.62	0.0019	0.2	0.01	0.0115	0.0059
P2-2	0.219	0.561	0.0014	0.21	0.01	0.0119	0.0053
P2-3	0.194	0.488	0.0013	0.24	0.01	0.0114	0.0056
P2-4	0.147	0.384	0.0012	0.27	0.01	0.0092	0.006
P3-1	0.196	0.625	0.0015	0.21	0.01	0.0116	0.0068
P3-2	0.202	0.606	0.0016	0.22	0.01	0.0115	0.0065
P3-3	0.184	0.528	0.0014	0.25	0.01	0.0114	0.0062
P3-4	0.14	0.4	0.0011	0.29	0.01	0.0112	0.0062

. With the increase in the height of the ingots, the aluminum contents in ingots first increased and then decreased slightly. The titanium contents increased along with the height of the ingots, and the contents of Mg and Si changed slightly. The contents of aluminum and titanium increased and decreased in the inclusions, respectively, comparable to those in the consumable electrode. With the increase in the TiO₂ contents in the slag, the increase and decrease in inclusions were reduced. This is mainly due to the exchange reaction of Equation (1). The ΔG calculated in the three experiments was 924.2 J, −2098.9 J and −8608.4 J, indicating that the reaction happened in P2 and P3. The silicon contents increased compared to those in the consumable electrode, mainly due to the reduction of SiO₂ in the slag. The oxygen contents exhibited a different increase in the three ingots compared to the electrode, mainly because no deoxidant was added during the ESR processes.

$${4\left[\mathrm{Al}\right]+3[\mathrm{TiO}}_2{]=3[\mathrm{Ti}]+2(\mathrm{Al}}_2{\mathrm{O}}_3),\log K=\frac{\mathrm{25,545}}{T}-5.4(\mathrm{J}·{\mathrm{mol}}^{-1})$$

(1)

$$\Delta G^{\theta }=-RT\ln K,\Delta G=\Delta G^{\theta }+RT\ln \frac{{[\mathrm{Ti}]}^3{f}_{\mathrm{Ti}}^3·a_{{\mathrm{Al}}_2{\mathrm{O}}_3}^2}{{[\mathrm{Al}]}^4{f}_{\mathrm{Al}}^4·a_{{\mathrm{TiO}}_2}^3}$$

(2)

$$\lg f_i=\mathsf{\Sigma }(e_i^j[\%j]+r_i^j{[\%j]}^2),$$

(3)

where $f_{\mathrm{Ti}}$ and $f_{\mathrm{Al}}$ are the activity coefficients of Ti and Al, respectively. $a_{{\mathrm{TiO}}_2}$ and $a_{{\mathrm{Al}}_2{\mathrm{O}}_3}$ are the activity of TiO₂ and Al₂O₃, respectively. $e_i^j$ and $r_i^j$ are the first-order and second-order interaction parameters. The first-order interaction parameters are listed in

Table 4. First-order interaction parameters $e_i^j$ used in the present study data from [21,22,23,24,25,26,27,28].

${\mathit{e}}_{\mathit{i}}^{\mathit{j}}$	C	Ni	Cr	Ti	Al	Si	Mn	Mg	S	O	N
Ti	−0.19	−0.015	0.055	0.042	0.024	−0.025	−0.0043	−1.27	−0.27	−3.4	−2.06
O	−0.421	0.006	−0.032	−0.34	−1.17	−0.066	−0.021	−0.396	−0.133	−0.17	−0.14
Al	0.091	−0.0173	0.03	0.016	0.0045	0.0056	0.035	−0.3	0.03	−1.98	−0.058
Mg	0.15	−0.012	0.05	−0.51	−0.12	−0.096	-	-	-	−0.602	-
S	0.11	-	-	−0.6	−3.9	−0.131	−0.021	-	−0.133	−0.27	0.01
N	0.13	0.01	−0.046	−0.6524	-	0.047	−0.02	-	-	-	0

. The available second-order interaction parameters are summarized as follows: $r_{\mathrm{Al}}^{\mathrm{C}}=-0.004$, $r_{\mathrm{Al}}^{\mathrm{Al}}=-(0.0011+0.17)/T$, $r_{\mathrm{Al}}^{\mathrm{Si}}=-0.0006$, $r_{\mathrm{Al}}^{\mathrm{Ni}}=0.000164$, $r_{\mathrm{Ti}}^{\mathrm{Ti}}=-0.0001$, $r_{\mathrm{Ti}}^{\mathrm{Ni}}=0.0005$ [19,20].

3.2. Characteristics of Inclusions

3.2.1. Number and Size Distribution of Inclusions in Electrode and Ingots

To investigate the evolution mechanism of the inclusions, the number, size and distribution of inclusions in each sample were analyzed by randomly observing 30 fields of P1-3, P2-1, P2-2, P2-3, P2-4, P3-3 and the consumable electrode by SEM. The size range of inclusions in the consumable electrode was mainly 0–3 μm, though some of the inclusions exceeded 3 μm, and almost all TiN inclusions were less than 2 μm. Most of the TiN inclusions in the remelted ingots were also less than 2 μm, while the other inclusions grew larger. The size distribution of inclusions detected in P2-1 to P2-4 is shown in Figure 2. The statistical results show that the average sizes of inclusions in P2-1 to P2-4 were 4.4 μm, 3.8 μm, 3.6 μm and 2.5 μm. The size of the inclusions tended to become larger as the height of the ingots increased from bottom to top, because the increase in the temperature of the slag as the process progressed caused a longer cooling time. The longer cooling time further increased the size. The size distribution of the inclusions detected in P1-3 to P3-3 is shown in Figure 3. The TiO₂ contents after ESR are listed in

Table 5. The TiO₂ contents after the ESR process (mass%).

Experiment	S1	S2	S3
TiO2	5.63	8.04	13.34

. It can be seen that the sizes of inclusions increased with the TiO₂ contents.

3.2.2. Characteristics of Inclusions in Consumable Electrode

Figure 4 and Figure 5 show the SEM images and EDS spectra of typical inclusions in the consumable electrode. Three types of inclusions were observed: TiS, TiN and Al₂O₃ surrounded by a TiN layer. Figure 4a–c show the SEM images of typical inclusions of TiS, TiN and Al₂O₃ surrounded by a TiN layer in the consumable electrode, respectively. The TiS inclusions were irregular strips, and the TiN inclusions were the shape of a rectangle or parallelogram, whereas the Al₂O₃ inclusions surrounded by TiN were circular.

3.2.3. Characteristics of Inclusions in Remelted Ingots

There were four types of inclusions found in the samples from ingots, namely, TiN, MgO-Al₂O₃ and MgO-Al₂O₃ surrounded by a TiN layer. Figure 6 and Figure 7 show the SEM images of typical inclusions of TiN, MgO-Al₂O₃ and MgO-Al₂O₃ surrounded by a TiN layer. The MgO-Al₂O₃ inclusions were circular.

3.3. Evolution Mechanism of TiS Inclusions during the ESR Process

It was found that few TiS inclusions existed in the consumable electrode, and they were removed in the remelted ingots. The standard Gibbs free energy can be calculated by Equation (4) [29].

$${\mathrm{TiS}}_{\mathrm{inclusion}}=\left[\mathrm{Ti}\right]+\left[\mathrm{S}\right],\Delta G_{\mathrm{TiS}}^{\mathsf{\theta }}=\mathrm{186,400}-75.4T(\mathrm{J}·{\mathrm{mol}}^{-1})$$

(4)

$$K_{\mathrm{TiS}}=a_{\mathrm{Ti}}·a_{\mathrm{S}}=f_{\mathrm{Ti}}[\%\mathrm{Ti}]·f_{\mathrm{S}}[\%\mathrm{S}],$$

(5)

where $f_{\mathrm{Ti}}$ and $f_{\mathrm{S}}$ are the activity coefficients of Ti and S, respectively. $a_{\mathrm{Ti}}$ and $a_{\mathrm{S}}$ are the activity of Ti and S, respectively. $K$ is the equilibrium constant.

There was only one type of sulfide inclusion in the electrode, and the size of the TiS inclusions was mainly 2~3 μm. Through Equations (3)–(5), the Gibbs free energy could be calculated as −170,150 J. The $\Delta G_{\mathrm{TiN}}^{\mathsf{\theta }}$ < 0 indicated the feasibility of TiS dissolution from thermodynamics. Jin et al. [29] calculated the actual concentration of TiS and MnS ($Q_{\mathrm{TiS}}$ and $Q_{\mathrm{MnS}}$) through Equations (6) and (7), and they proposed that the precipitation of TiS in the experimental steel was higher than that of MnS at the same S content. The present study confirms their finding.

$$Q_{\mathrm{TiS}}={[\%\mathrm{Ti}]}_{\mathrm{L}}·{[\%\mathrm{S}]}_{\mathrm{L}}=\frac{{[\%\mathrm{Ti}]}_0·{[\%\mathrm{S}]}_0·{(1-x_{\mathrm{S}})}^{k_{\mathrm{Ti}}-1}}{1-(1-k_{\mathrm{S}})·x_{\mathrm{S}}},$$

(6)

$$Q_{\mathrm{MnS}}={[\%\mathrm{Mn}]}_{\mathrm{L}}·{[\%\mathrm{S}]}_{\mathrm{L}}=\frac{{[\%\mathrm{Mn}]}_0·{[\%\mathrm{S}]}_0·{(1-x_{\mathrm{S}})}^{k_{\mathrm{Mn}}-1}}{1-(1-k_{\mathrm{S}})·x_{\mathrm{S}}},$$

(7)

where ${[\%i]}_{\mathrm{L}}$ and ${[\%i]}_0$ are the concentrations of i at the solidification front and in the liquid steel at the beginning of solidification, respectively; $k_i$ is the equilibrium partition coefficient of i in the solid steel or liquid steel; $x_i$ is the solid fraction.

3.4. Evolution Mechanism of TiN Inclusions during the ESR Process

TiN inclusions were obtained both in the consumable electrode and the remelted ingots. The equilibrium phase precipitation in the consumable electrode was calculated by Jmat pro7.0 (functional material performance simulation software, which can be used to calculate a variety of metal material properties), as shown in Figure 8. It can be seen that the precipitation temperature was about 1790 K, and the liquidus temperature was calculated to be about 1668 K. Thus, the TiN inclusions could not be dissociated during liquid metal film formation at the electrode tip, which corresponds to Zheng’s [30] result. During the ESR process, the temperature of slag can reach 1972 K and even higher. The TiN inclusions could be dissociated in this period of the process. Yang et al. [31] indicated that TiN inclusions were all removed in the solid–liquid two-phase region at the electrode, and the TiN in the remelted ingots was secondary TiN generated during the solidification of steel.

The amounts of inclusions in the electrode, P1-3, P2-3 and P3-3 are shown in Figure 9. There were about 79.4%, 34.9%, 29.3% and 31.9% TiN inclusions in the consumable electrode, P1-3, P2-3 and P3-3, respectively. This indicates that part of the TiN inclusions was dissociated during liquid metal–slag reaction processing.

During the ESR process, the Gibbs free energy of the dissolution of TiN inclusions could be calculated as −127,202 J through Equation (8) [32]. The $\Delta G_{\mathrm{TiN}}^{\mathsf{\theta }}$ < 0 indicated the feasibility of TiN dissolution from thermodynamics. The precipitation of TiN mainly depends on the content of Ti and N. The variation in the N content before and after the ESR process changed little in this study, as shown in Table 3. Therefore, the Ti content has a greater influence in the formation of TiN. In Incoloy825, a typical high-Ti and low-Al alloy, the loss of Ti is profound. This can be seen in Table 3: the loss of Ti was 50% on average, mainly because of the exchange reaction [33,34]. The Ti is more likely to react with oxide in the slag, and the formation of TiN is more difficult. Therefore, most of the TiN inclusions originated from the consumable electrode.

$${\mathrm{TiN}}_{inclusion}=\left[\mathrm{Ti}\right]+\left[\mathrm{N}\right],\Delta G_{\mathrm{TiN}}^{\mathsf{\theta }}=\mathrm{379,000}-149T(\mathrm{J}·{\mathrm{mol}}^{-1})$$

(8)

3.5. Evolution Mechanism of Oxide Inclusions during the ESR Process

The obtained oxide inclusion in the electrode was Al₂O₃ surrounded by a TiN layer, and the oxide inclusions after the ESR process were MgO-Al₂O₃ and MgO-Al₂O₃ surrounded by a TiN layer. The formation of MgO-Al₂O₃ in the inclusions (both MgO-Al₂O₃ and MgO-Al₂O₃ surrounded by a TiN layer) is due to the reaction of the solution of Al in the alloy, MgO in the slag and Al₂O₃ in the inclusions. The (MgO) in the slag was first transformed into [Mg] in the steel, as shown in Equation (9), then [Mg] reacted with the (Al₂O₃) in the inclusions, as shown in Equation (10), and the (MgO) in the inclusions appeared. Combining Equations (9) and (10), we can find that the (MgO) in the slag transformed into (MgO) in the inclusions through (Al₂O₃), as shown in Equation (11), and the Gibbs energy was 0 [35]. Therefore, the inclusions of Al₂O₃ transformed to a spinel solid solution. Many researchers [36,37] have found that MgO-Al₂O₃ has a high potential for acting as a TiN nucleation site because of the low disregistry. The existing form of Al₂O₃ inclusions was Al₂O₃ with a TiN layer in this study, which means that the formation of an inclusion of MgO-Al₂O₃ surrounded by a TiN layer was from an inclusion of the Al₂O₃ surrounded by a TiN layer, as shown in Equations (9)–(11). With regard to the MgO-Al₂O₃ inclusion, it originated from MgO-Al₂O₃ surrounded by a TiN layer. When the TiN was removed, the MgO-Al₂O₃ inclusion surrounded by a TiN layer changed to MgO-Al₂O₃.

$${2\left[\mathrm{Al}\right]+3\left(\mathrm{MgO}\right)}_{\mathrm{slag}}{=3\left[\mathrm{Mg}\right]+(\mathrm{Al}}_2{\mathrm{O}}_3{)}_{\mathrm{slag}},$$

(9)

$${3\left[\mathrm{Mg}\right]+(\mathrm{Al}}_2{\mathrm{O}}_3{)}_{\mathrm{inclusion}}{=2\left[\mathrm{Al}\right]+3\left(\mathrm{MgO}\right)}_{\mathrm{inclusion}},$$

(10)

$${3\left(\mathrm{MgO}\right)}_{\mathrm{slag}}{+(\mathrm{Al}}_2{\mathrm{O}}_3{)}_{\mathrm{inclusion}}{=3\left(\mathrm{MgO}\right)}_{\mathrm{inclusion}}{+(\mathrm{Al}}_2{\mathrm{O}}_3{)}_{\mathrm{slag}},$$

(11)

From Figure 5, it can be seen that there is little difference in the proportion of different inclusions from the remelted ingots. This means that the concentration of TiO₂ in slag has little effect on the relative proportion of inclusions in remelted ingots. However, the sizes of inclusions increased with an increase in TiO₂ in the slag, as shown in Figure 3. The size of TiN inclusions in the consumable electrode and remelted ingots was 0 to 2 μm, indicating that the larger inclusions were MgO-Al₂O₃ and MgO-Al₂O₃ + TiN, caused by TiO₂. This might be because the Al₂O₃ increases with an increase in TiO₂ in the slag due to the reaction (1), and Al₂O₃ causes the transformation of Mg to MgO, as shown in Equation (9) [30]. The increase in Al₂O₃ and MgO causes the formation and growth of MgO-Al₂O₃. To make the size of inclusions in remelted ingots of Incoloy825 smaller, the TiO₂ and MgO contents in the slag should be controlled.

4. Conclusions

Three 50 kg-scale electroslag remelting experiments were performed with different TiO₂ contents in the slag. The evolution of inclusions in Incoloy825 alloy was investigated, and the conclusions are summarized as follows:

(1).

The inclusions in the consumable electrode were TiN, TiS and Al₂O₃ with a surrounding TiN layer, and the sizes were mainly 1 to 3 μm. After the electroslag remelting process, the inclusions were TiN, MgO-Al₂O₃ and MgO–Al₂O₃ surrounded by TiN, and the sizes were mainly 1 to 4 μm.

(2).

The sizes of the inclusions increased with the TiO₂ contents in the slag, but the types of inclusions did not vary with them. When the TiO₂ contents were constant, the average size of inclusions increased from 2.5 to 4.4 μm with the increase in the height of the ingots.

(3).

The inclusions of MgO-Al₂O₃ surrounded by a TiN layer and MgO-Al₂O₃ formed as a result of the inclusions of Al₂O₃ surrounded by a TiN layer. The MgO and Al₂O₃ in the slag played an important role in generating spinel.

(4).

TiS inclusions in the consumable electrode dissolved during the ESR process. The TiN inclusions could not be removed from the tip of the electrode, but part of the TiN dissociated during the reaction between the liquid metal and slag. Most of the TiN inclusions originated from the consumable electrode.

(5).

The electroslag remelting process using a high-temperature Ni-based alloy and low-fluoride slag CaF₂-CaO-Al₂O₃-TiO₂-MgO-SiO₂-(FeO) was investigated. The results show that controlling the MgO and TiO₂ contents could decrease the generation of inclusions to meet the industry’s needs.