TiN Particles and Clusters during Ladle Treatments of Ni-based Alloy 825 using Different Stirring Modes

Abstract

Today, titanium is often used in steelmaking not only for deoxidation but also for micro-alloying and alloying for a wide range of steel grades. Therefore, many studies are focused on investigations on the formation and behavior of Ti-containing non-metallic inclusions (such as oxides, nitrides and carbides) during production of different Ti-containing steels and their effect on final steel properties. This study has examined the behavior of TiN clusters and particles in the melt during the ladle treatment of Alloy 825 containing up to 1.2 wt% of Ti. The industrial trials were performed at the end of the ladle treatment by using argon gas in combination with electromagnetic stirring using an upwards or a downwards stirring direction. Metal samples were taken before and after ladle treatment to enable three-dimensional investigations of non-metallic inclusions and clusters. The composition, size and number of particles and clusters were determined after electrolytic extraction of the metal samples by using SEM in combination with EDS. It was found that agglomerations of TiN clusters and particles in the melt are faster during an upwards stirring in comparison to a downwards stirring. However, the removal of clusters from the melt is more effective when using a downwards stirring direction compared to when using an upwards stirring in combination with gas stirring. It was also found that the Turbulent collision is the dominant factor for the agglomeration of TiN particles in the melt.

1. Introduction

In modern times, titanium is used more and more during steelmaking not only for deoxidation but also for micro-alloying and alloying of an ever increasing range of steel grades. These grades are used for production of automobile parts, plates for ships, engines, high strength micro-alloyed steels etc. The titanium in the Ti-alloyed steels will form non-metallic inclusions (such as oxides, nitrides and carbides), which will affect the properties of the steels and alloys. Their number, size, morphology and composition can have significant negative or positive influences on the properties of the final steel product. For instance, the large size inclusions and clusters are a common origin to cracks and fractures in super alloys at high temperatures and low stresses.¹⁾ On the other hand, small size inclusions inhibit grain growth in steel and promote the formation of a fine metal structure.¹⁾

Several previous studies have proven that the characteristics of non-metallic inclusions in liquid steel can be modified during the ladle treatment.^{2,3,4,5,6,7,8,9,10,11,12,13,14)} Both model calculations^{2,3,4,6,7,8,9,10,12,14)} and experimental investigations^5,11,12,13) have been done to evaluate the growth and flotation of oxide inclusions and clusters during ladle treatment of different steels. During these studies, several parameters influencing the agglomeration and removal of inclusions (such as stirring conditions and inclusion characteristics) were considered and discussed.^{2,3,4,5,6,7,8,9,10,11,12,13,14)} A overview of the obtained results at different stirring modes and conditions were done in a separate article.¹³⁾ It was reported that the turbulent dissipation energy rate (ε), as a parameter of intensity of melt stirring in the ladle, can be varied in very wide range (from 0.0002 to 0.7 m² s⁻²). Furthermore, that it can significantly effect the formation, growth and removal of clusters in the melt during ladle treatment. For instance, based on the results obtained from experimental investigations and model calculations, Chung et al.¹²⁾ concluded that gas stirring (GS) was preferable for mixing of the melt in a 100 t ASEA-SKF ladle but electromagnetic stirring with upwards direction (EMS/U) was preferable for the removal of Al₂O₃ inclusions. However, the combination of the electromagnetic and gas stirring (EMS/U +GS) was superior for both a melt mixing and removal of inclusions.

The most harmful non-metallic inclusions in the Alloy 825 during ladle treatment are Al₂O₃–MgO and TiN, which can form large size clusters that can remain after casting.¹⁵⁾ The behaviors of Al₂O₃–MgO clusters in the melt for different stirring modes during ladle treatment were discussed in separate article.¹³⁾ However, the formation and behavior of TiN inclusions and clusters in the melt of Ni-base alloys have not been considered up to now.

This study focused on the investigations of TiN particles and clusters in the Ni-base Alloy 825 during ladle treatment. Furthermore, on the evaluation of the effect of different stirring modes on the behavior and removal of TiN clusters. The size distributions of clusters and particles in the clusters were investigated after electrolytic extraction of industrial alloy samples taken during ladle treatment.

2. Experimental

2.1. Sample Preparation

The effect of the different stirring conditions on the cluster and particle growth in the Ni-based Alloy 825 was evaluated in industrial trials (Heat A and Heat B) at Sandvik Materials Technology. The composition of the melt at the initial and final moment of the last period of the ladle treatment of the two heats is given in Table 1. Moreover, the slag compositions for the two heats are given in Table 2.

Table 1. Composition of the two heats (in wt%) at the initial and final moment of Period II.

		Fe	Ni	Cr	Mo	Cu	Ti	Mn	Si	Al	C	T.O	S
Heat A	Initial	Balance	38.22	19.78	2.51	1.59	0.76	0.51	0.22	0.17	0.009	0.0056	≤0.001
	Final	Balance	38.27	19.81	2.51	1.59	0.77	0.51	0.23	0.16	0.009	0.0070	≤0.001
Heat B	Initial	Balance	38.14	19.62	2.52	1.67	0.80	0.69	0.18	0.16	0.007	0.0066	≤0.001
	Final	Balance	38.18	19.67	2.52	1.67	0.80	0.69	0.18	0.14	0.007	0.0064	≤0.001

Table 2. Composition of the slag during Period II for Heat A and Heat B (wt%).

	CaO	Al2O3	TiO2	MgO	SiO2	Cr2O3	FeO	MnO	V2O5	P2O5
Heat A	42.83	27.16	13.58	7.42	5.12	2.61	0.84	0.33	0.10	<0.01
Heat B	48.06	26.70	12.82	5.55	4.91	1.07	0.64	0.14	0.12	<0.01

The ladle treatment of the melt was divided into two periods for both heats. Period I is the main period, at which an addition of a synthetic slag and alloying components takes place. Period II is meant for the final homogenization of the melt temperature and composition as well as for the removal of non-metallic inclusions before casting. Stirring of the melt in the 75 t ladle was done by using the electromagnetic stirring with upwards (EMS/U) and downwards (EMS/D) directions in combination with gas stirring (GS), as shown schematically in Fig. 1.

Fig. 1.

Schematic illustration of combined gas stirring and electromagnetic stirring in the ladle with (a) upwards direction (EMS/U) and (b) downwards direction (EMS/D).

Period I of Heat A (53 min) and Heat B (87 min) was done with the same stirring conditions: EMS/U at 800 A. For Period II (10 min), two different stirring modes were used: EMS/U at 800 A combined with GS at 0.04 m³ min⁻¹ argon flow for Heat A and EMS/D at 800 A combined with GS at 0.04 m³ min⁻¹ argon flow for Heat B. For evaluation of TiN clusters in Alloy 825, lollypop samples (Ø 33 mm, 12 mm thickness) were taken manually before and after Period II of each heat.

2.2. Sample Analysis

Specimen for electrolytic extraction (Specimen EE) was cut out from each lollipop steel sample, as shown in Fig. 2(a).

Fig. 2.

Schematic illustrations of metal specimen used for electrolytic extraction (a), measured length and width of a cluster (b) and of particles in a cluster (c) on SEM images.

One surface of this specimen was dissolved in a 10%AA electrolyte by using the following electric parameters: a voltage of ~4 V, an electric current of 40–60 mA, and an electric charge of 1000 Coulombs. About 0.18–0.21 g of the metal was dissolved from the specimen. However, the non-metallic inclusions including TiN are not dissolved during the electrolytic extraction. Thereafter, the electrolyte with undissolved non-metallic inclusions was filtered by using a film filter having an open pore size of 0.4 μm. Then, the TiN clusters and particles in the clusters on the surface of the film filter were investigated by using a scanning electron microscope (SEM) in combination with energy dispersive spectroscopy (EDS). The length and width of the clusters (L and W) and particles in the clusters (l and w) were measured as shown in Figs. 2(b) and 2(c), respectively. The l and w values were used for calculations of the equivalent size of the particles in the clusters as seen in Eq. (1). Overall, the total number of observed clusters for each specimen varied from 70 to 120.   

$$d_{\mathrm{e}\mathrm{q}}=\sqrt{l\cdot w}$$(1)

The number of clusters per unit volume (N_VC) and the number of agglomerated particles in clusters per unit volume (N_VP) were calculated according to Eqs. (2) and (3).   

$$N_{VC}=\frac{n_C\cdot A_{fil}}{A_{obs}}\cdot \frac{{\rho }_M}{\mathrm{\Delta }{W}_M}$$(2)

  

$$N_{VP}=\frac{n_P\cdot A_{fil}}{A_{obs}}\cdot \frac{{\rho }_M}{\mathrm{\Delta }{W}_M}$$(3)

where A_fil is the area of the filter with collected inclusions after the electrolytic extraction; A_obs is the area of filter surface observed in SEM; ρ_M is the density of the metal; n_P is the total number of observed particles in the given size range in all observed clusters and ΔW_M is the weight of the metal dissolved during the electrolytic extraction.

3. Result and Discussion

An earlier study showed that harmful Al₂O₃–MgO and TiN clusters exist in Alloy 825 after casting.¹⁵⁾ It was found that the upwards stirring (EMS/U+GS) promote a faster agglomeration of Al₂O₃–MgO clusters. The numbers of these clusters for upwards stirring were found to decrease for small and medium size inclusions and to increase for large size inclusions. However, for downwards stirring (EMS/D+GS) the number of clusters for all sizes decreased.¹³⁾ Furthermore, more detailed studies (in 2D) have been done on 3 heats for each stirring method. Despite some variations, the obtained results showed similar tendencies for Al₂O₃–MgO and TiN clusters in the heats with the same stirring parameters. However, due to limitations of the article volume, the obtained results will be published and discussed in a separate article. Meanwhile, this study will concentrate on how the stirring affect the TiN clusters.

3.1. Precipitation and Agglomeration of TiN Inclusions in Melt

Usually, the TiN inclusions precipitated only during solidification of most steel grades. However, the Ti content in the Alloy 825 grade is high enough (up to 1.2%) to enable a precipitation of primary TiN in the melt before solidification.¹⁵⁾ By using the TERMOCALC v3.1 software and the TFC7 database (with elements; Fe, Ni, Cr, Cu, Mo, Ti, Al, N, O), the weight fraction of TiN inclusions in the melt of the Alloy 825 grade were calculated depending on the temperature. It can be seen in Fig. 3 that for Heat A, most of the TiN inclusions (almost 80%) precipitated in this melt before solidification at 1385°C. Therefore, most of the TiN clusters formed and grew during the ladle treatment. According to the obtained results, only 20% of the TiN inclusions can precipitate during the melt solidification. Moreover, the size of the secondary TiN inclusions (< 1 μm) in the fast solidified samples (as in the used lollipop samples) are significantly smaller in comparison to the primary inclusions. The latter were formed early and grew considerably before the solidification.

Fig. 3.

Plot of the weight fraction of TiN over the temperature with the approximate sampling temperature and solidification time marked for Heat A.

An evidence of that agglomeration of TiN inclusions occurs has been provided by Xuan et al.¹⁶⁾ based on the determination of the wettability and attractive forces of TiN inclusions in liquid iron and carbon steels. It was reported that the contact angle between TiN/pure Fe and TiN/carbon steel at 1600°C is larger than 100°. Therefore, the TiN inclusions have a relative high possibility to agglomerate and to form clusters.

In this study, the formation and behavior of TiN clusters were evaluated in metal samples of the Ni-base Alloy 825 grade taken from the melt during the ladle treatment.

3.2. Characterization of TiN Clusters in Steel Samples

It was found that the TiN clusters were composed of >90 wt% TiN in all samples from both heats. The size distributions of TiN clusters in metal samples before and after Period II of the ladle treatment are shown in Fig. 4. It can be seen for Heat A where an EMS/U+GS stirring (Fig. 4(a)) was used, that the number of clusters per unit volume of metal (N_VC) has increased by ~20–40%. However, the size distribution peak has not moved compared to the initial moment. It can be explained by the growth of all clusters due to the agglomeration and formation of new TiN particles and clusters due to the “open eye” zone on a surface of the alloy melt during the intensive stirring in the ladle. As a result, the melt is in contact with air and can absorb the nitrogen from the air. For Heat B a combination of EMS/D+GS stirring (Fig. 4(b)) was used. Here the number of small size clusters decreases drastically (>50%) while the N_VC values of large size clusters have a slight increase. It can be explained by the growth of clusters due to the agglomerations of inclusions with clusters and clusters with clusters. In this case, the formation of new TiN inclusions in the melt is limited due to absence of the “open eye” zone and the additional absorption of nitrogen from the air. The difference in the “open eye” zone between Heat A and Heat B can be explained by the different melt velocity in two heats, as can be seen in Fig. 5.¹²⁾

Fig. 4.

Size distributions of TiN clusters when using an upwards induction stirring in combination with gas stirring in Heat A (a) and (b) at downwards induction stirring in combination with gas stirring in Heat B.

Fig. 5.

Schematic illustration of the different velocity components at different directions of electromagnetic stirring for a) Heat A and b) Heat B.

It is apparent that the number of clusters at the initial moment of Period II of the ladle treatment in Heat B is significantly larger compared to that in Heat A. It can be explained by some differences in the melt conditions in the initial moment of Period II (the different stirring times or different amounts of additions during Period I), despite that the stirring parameters in Period I were the same for both heats. However, the steel and slag compositions given in Tables 1 and 2 for the initial moment of Period II are similar for both heats. For elimination of the effect of some differences in initial melt conditions on behavior of TiN particles and clusters during Period II in both heats, changes (in%) of N_VC values for different cluster size ranges (ΔN_VC) were used for the estimation of the effect of different stirring modes on the formation, growth and removal of TiN clusters in the Ni-base alloy. The normalized values of ΔN_VC for each cluster size range were calculated as follows:   

$$\mathrm{\Delta }{N}_{\mathrm{V}\mathrm{C}}=100\%\cdot \left(N_{\mathrm{V}\mathrm{C}\text{-}\mathrm{f}\mathrm{i}\mathrm{n}\mathrm{a}\mathrm{l}}-N_{\mathrm{V}\mathrm{C}\text{-}\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}}\right)/N_{\mathrm{V}\mathrm{C}\text{-}\mathrm{i}\mathrm{n}\mathrm{i}\mathrm{t}}$$(4)

where N_VC-init and N_VC-final are the numbers of clusters per unit volume of metal in the given size range before and after Period II of ladle treatment, respectively.

Based on Swedish standard SS 111116,¹⁷⁾ the following three size ranges of clusters were chosen: small (<5.6 μm), medium (5.6–11.2 μm) and large (>11.2 μm) clusters. It can be seen in Fig. 6 that the numbers of all sizes of TiN clusters in Heat A increased by 15 to 90%. Therefore, it can be concluded that the upward induction stirring in combination with gas stirring in Heat A increases the number of clusters in all size ranges but mostly for clusters larger than 5.6 μm. This confirms the formation of additional TiN inclusions and new clusters during Period II. It should be pointed out that the formation and growth of clusters in this case are significantly faster than the removal and accumulation of clusters to the slag. For Heat B with a downward induction stirring in combination with gas stirring, the ΔN_CV values for clusters smaller than 11.2 μm decreased by ~40 to 60% but the values for large size clusters increased by ~20%. It follows from these results that the processes of growth of clusters due to cluster/cluster collisions and their removal into the slag are dominant at the downwards stirring combination. As a result, the total number of TiN clusters in the Ni-base Alloy 825 grade decreases by ~30% during Period II of the ladle treatment.

Fig. 6.

Chang of the numbers of TiN clusters (ΔN_VC) of different size ranges in samples from Heat A and Heat B.

Table 3 shows the mean size of TiN clusters and particles in clusters found in both heat A and B during Period II. It is apparent that the mean length of the clusters increases by 30% in Heat A and decreases by 7% in Heat B. The obtained results agree well with discussion and conclusions given above. However, the mean size of the particles, d_eq, decreases by 5 to 10% in both heats.

Table 3. Mean size of TiN clusters and particles in clusters in samples from Heat A and B.

Characteristics	Heat A		Heat B
	Initial	Final	Initial	Final
Mean length (L) of clusters (μm)a)	12.8±6.3	16.7±9.6	13.2±9.8	12.3±6.1
Mean size (deq) of particles in clusters (μm)a)	2.0±1.0	1.9±1.1	2.1±1.1	1.9±1.2

^a)   Mean value ± standard deviation (σ).

The size distributions for particles in TiN clusters are shown in Fig. 7 for samples taken before and after Period II of both heats. It can be seen in Fig. 7(a) that the number of particles in clusters increased significantly (on average on ~40%) when using an upwards induction stirring in combination with gas stirring during the final period of ladle treatment in Heat A. This can be explained by the additional precipitation of TiN inclusions and their intensive agglomeration into clusters. This finding is confirmed by an increase of the nitrogen content in metal samples from 0.022 up to 0.032%, due to an existence of the “open-eye” zone on a surface of the melt during this treatment period. On the other hand, the number of TiN particles in clusters decreased significantly (on ~65%) when using a downwards induction stirring in combination with gas stirring in Heat B (Fig. 7(b)). However, the average size of the particles does not change practically. It means that the big clusters, which have the largest number of particles, float out from the melt faster than the TiN cluster growth due to cluster/particle and cluster/cluster collisions. It should be noted that the nitrogen content in the melt when using a downwards induction stirring in combination with gas stirring were not changed and equals 0.022%.

Fig. 7.

Size distributions of particles in TiN cluster in samples of Heat A using an upwards directed induction stirring (a) and Heat B when using a downwards directed induction stirring (b). In both cases, gas stirring was used in combination with the induction stirring.

3.3. Evaluation of Formation and Growth of Clusters at Different Stirring Modes

It is a well-known fact that the number of collisions of clusters and inclusions affect the growth of clusters in the melt. The total number of inclusion/cluster collisions in the melt can be calculated as a summation of Stokes’ collisions (due to different flotation speeds for different inclusions in the melt), Brownian collisions (due to random movements of inclusions and clusters in the melt) and Turbulent collisions (due to turbulent movement of the melt). The effect of different collision types in the melt on the growth of TiN clusters depends on collision volumes for Stokes’ ($W_{ij}^S$), Brownian ($W_{ij}^B$) and Turbulent ($W_{ij}^T$) collisions. The values of $W_{ij}^S$, $W_{ij}^B$ and $W_{ij}^T$ can be calculated as follows:¹⁸⁾   

$$W_{ij}^S=\frac{2\pi g\left({\rho }_M-{\rho }_i\right)}{9{\mu }_M}{\left(r_i+r_j\right)}^3\left|r_i-r_j\right|$$(5)

  

$$W_{ij}^B=\frac{2kT}{3{\mu }_M}\left(\frac{1}{r_i}+\frac{1}{r_j}\right)\left(r_i+r_j\right)$$(6)

  

$$W_{ij}^T=1.3\alpha {\left(r_i+r_j\right)}^3\sqrt{\frac{\pi \epsilon }{v_M}}$$(7)

where ρ_M and ρ_i are the densities of steel melt and inclusion, respectively; g is the gravity acceleration; r_i and r_j are the radius of i-th and j-th inclusions (or clusters), respectively; μ_M is the dynamic viscosity of the melt; k and T are the Boltzmann’s constant and temperature of the melt; α is the agglomeration coefficient; ε is the turbulent energy dissipation rate; ν_M is the kinematic viscosity of the melt (= μ_M/ρ_M). The values of these parameters are given in Table 4.

Table 4. Parameters used in Eqs. (5), (6), (7), (8) and their values.

Parameter	Name		Value	Ref
ρM	Density of liquid Fe	kg m−3	7240	[19]
ρTiN	Density of solid TiN	kg m−3	5220	[20]
g	Gravity acceleration	m s−2	9.81	–
T	Temperature	K	1873	–
k	Boltzmann’s constant	J K−1	1.381·10−23	−
A33	Hamaker constant for liquid Fe	J	2.53·10−18	[21]
A11	Hamaker constant for TiN	J	1.823·10−19	[22]
εA	Turbulent energy dissipation rate for Heat A	m2 s−3	0.0077	[23]
εB	Turbulent energy dissipation rate for Heat B	m2 s−3	0.00092	[23]a)
μM	Dynamic viscosity	m Pa s	0.007	[24, 25]

^a)   Approximated from data of Reference [23].

For evaluation of the energy dissipation rate (ε) at different stirring modes, it was assumed that the gas flow do not play a significant role at the present stirring rates.²⁶⁾ The ε values for upwards (ε_A) and downwards (ε_B) electromagnetic stirring were taken for a 100 t ladle.²³⁾ The α value can be calculated by using Eq. (8):²¹⁾   

$$\alpha =0.727{\left(\frac{{\mu }_M{\left(r_i\right)}^3\sqrt{\frac{\epsilon }{v_M}}}{A_{131}}\right)}^{-0.242}/\sqrt{\pi }$$(8)

where $A_{131}={\left(\sqrt{A_{11}}-\sqrt{A_{33}}\right)}^2$ is the Hamaker constant for TiN in liquid iron.

Based on the results for both heats shown in Figs. 4 and 7, the radiuses for size distribution peaks for TiN clusters (r_i-C = L/2) and particles in clusters (r_i-P = d_eq/2) were taken as 5.0 and 0.75 μm, respectively. According to the Eqs. (5), (6), (7), it was found that the values of collision volumes between cluster-cluster can be 10–100 times larger compared to particle-cluster collisions. Therefore, it can be expected that the cluster growth in the melt will be much faster due to cluster-cluster collisions in comparison to the particle-cluster collisions. However, the number of TiN particles in the melt is much higher than the number of clusters. Therefore, the particle-cluster collision will occur more often and still dominate the inclusion growth.

A comparison of the collision volumes for Stokes’ ($W_{ij}^S$), Brownian ($W_{ij}^B$) and Turbulent ($W_{ij}^T$) collisions are shown in Fig. 8 for TiN particle-cluster collisions at r_i-P = 0.75 μm. Moreover, the Turbulent collision volumes ($W_{ij}^T$) are plotted for an upwards induction stirring in Heat A and for a downwards induction stirring in Heat B. It is apparent that the Brownian collisions have no particular effect on the total collision volume and cluster growth. For TiN particles and clusters with r_j< 20 μm (d_eq < 40 μm), the $W_{ij}^T$ values are ~2–200 times larger than the $W_{ij}^S$ values. Therefore, the turbulent collisions are the most important factor for particle-cluster agglomeration during upward and downward stirring because very few clusters having the length more than 40 μm were observed in all metal samples. Moreover, it is apparent that the Turbulent collision volume for the case with an upwards induction stirring combined with gas stirring (Heat A) is about 3 times larger than that for the case with a downwards induction stirring combined with gas stirring (Heat B). Therefore, it can be expected that the cluster growth will be faster when using an upwards induction stirring compared to when using a downwards induction stirring. These theoretical results agree well with the obtained experimental results shown in Figs. 4 and 6.

Fig. 8.

Collision volumes for Stokes’ ($W_{ij}^S$), Brownian ($W_{ij}^B$) and Turbulent ($W_{ij}^T$) collisions for TiN particles and clusters for an upwards induction (Heat A) and a downwards induction (Heat B) stirring combined with gas stirring during the final period of ladle treatment.

4. Conclusions

This study focused on the evaluation of the influence of the stirring mode on the formation and behavior of TiN inclusions and clusters during the ladle treatment of Ni-based Alloy 825. Specifically, the TiN inclusions and clusters were investigated in metal samples taken from Alloy 825 during the final period of ladle treatment when using an upwards (EMS/U) and a downwards (EMS/D) electromagnetic stirring in combination with gas stirring. Based on the obtained results the following conclusions were drawn:

(1) Most of the TiN inclusions (up to 80%) in the melt of the Ni-based Alloy 825 are precipitated before the start of the solidification of the melt at 1386°C.

(2) The number of small (<5.6 μm) clusters when using an upwards induction stirring (Heat A) increased by ~15% while the number of the medium (5.6–11.2 μm) and large (>11.2 μm) clusters increased by ~60–90%. For a downwards induction stirring (Heat B), the numbers of small (<5.6 μm) and medium (5.6–11.2 μm) clusters decreased with ~40–60% while large size clusters (>11.2 μm) increased by ~20%.

(3) Agglomeration and growth of TiN inclusions and clusters in the melt are faster when using an upwards electromagnetic stirring (Heat A) in comparison to when using a downwards stirring (Heat B). Specifically the former has about 3 times larger values of the Turbulent ($W_{ij}^T$) collision volumes compared to the latter case.

(4) Turbulent collisions ($W_{ij}^T$) due to the stirring of the melt during ladle treatment are the dominant cause for the formation and growth of TiN clusters having sizes smaller than 40 μm.

Acknowledgement

H. Kellner acknowledges the financial support and cooperation from Sandvik Materials Technology, Dalarna University, Jernkontoret, Sandvikens kommun, Länsstyrelsen Gävleborg, Region Dalarna and Region Gävleborg.

References

• 1)   E. S.  Huron and  P. G.  Roth: Superalloys 1996, eds. by R. D. Kissinger, D. J. Deye, D. L. Anton, A. D. Cetel, M. V. Nathal, T. M. Pollock, and D. A. Woodford, The Minerals, Metals & Materials Society, Warrendale, PA, (1996), 359.
• 2)   D. Y.  Sheng,  M.  Söder,  P. G.  Jönsson and  L.  Johnsson: Scand. J. Metall., 31 (2002), 134.
• 3)   L. T.  Wang,  Q. Y.  Zhang,  S. H.  Peng and  Z. B.  Li: ISIJ Int., 45 (2005), 331.
• 4)   Y. J.  Kwon,  J.  Zhang and  H. G.  Lee: ISIJ Int., 48 (2008), 891.
• 5)   S.  Taniguchi,  A.  Kikuchi,  T.  Ise and  N.  Shoji: ISIJ Int., 36 (1996), 117.
• 6)   D. Q.  Geng,  H.  Lei and  J. C.  He: ISIJ Int., 50 (2010), 1597.
• 7)   W.  Luo and  M.  Zhu: Metall. Mater. Trans. B, 44 (2013), 762.
• 8)   J. P.  Bellot,  V. D.  Felice,  B.  Dussoubs,  A.  Jardy and  S.  Hans: Metall. Mater. Trans. B, 45 (2014), 13.
• 9)   M.  Söder,  P. G.  Jönsson and  J.  Alexis: Scand. J. Metall., 31 (2002), 210.
• 10)   H.  Lei,  K.  Nakajima and  J. C.  He: ISIJ Int., 50 (2010), 1735.
• 11)   C.  Xuan,  A.  Karasev and  P. G.  Jönsson: ISIJ Int., 56 (2016), 1204.
• 12)   S. I.  Chung,  Y. H.  Shin and  J. K.  Yoon: ISIJ Int., 32 (1992), 1287.
• 13)   H. E. O.  Kellner,  A. V.  Karasev,  O.  Sundqvist and  P. G.  Jönsson: Steel Res. Int., 88 (2017), DOI: 10.1002/srin. 201700165.
• 14)   M.  Söder,  P. G.  Jönsson and  L.  Jonsson: Steel Res. Int., 75 (2004), 128.
• 15)   H. E. O.  Kellner,  A. V.  Karasev,  O.  Sundqvist,  A.  Memarpour and  P. G.  Jönsson: Steel Res Int., 88 (2017), DOI: 10.1002/srin. 201600024.
• 16)   C.  Xuan,  W.  Mu,  Z. I.  Olano,  P. G.  Jönsson and  K.  Nakajima: Steel Res. Int., 86 (2015), 1.
• 17)  SS 111116: 1987, Bedömning av icke-metalliska inneslutningar - Mikrometoder - Jernkontorets skala II för kvantitativbedömning.
• 18)   L. T.  Wang,  Q. Y.  Zhang,  S. H.  Peng and  Z. B.  Li: ISIJ Int., 45 (2005), 331.
• 19)   H. E. O.  Kellner,  A. V.  Karasev,  A.  Memarpour and  P. G.  Jönsson: Steel Res. Int., 87 (2016), 1461.
• 20)   S.  Zhang and  W.  Zhu: J. Mater. Process. Technol., 39 (1993), 165.
• 21)   C.  Xuan,  A. V.  Karasev,  P. G.  Jönsson and  K.  Nakajima: Steel Res. Int., 88 (2017), DOI: 10.1002/srin. 201600090.
• 22)   S.  Eichenlaub,  C.  Chan and  S. P.  Beaudoin: J. Colloid Interface Sci., 248 (2002), 387.
• 23)   J.  Alexis,  P. G.  Jönsson and  L.  Jonsson: ISIJ Int., 39 (1999), 772.
• 24)   H.  Kobatake and  J.  Brillo: J. Mater. Sci., 48 (2013), 6818.
• 25)   L.  Wang,  H.-G.  Lee and  P.  Hayes: ISIJ Int., 36 (1996), 7.
• 26)   U.  Sand,  H.  Yang,  J. E.  Eriksson and  R. B.  Fdhila: Steel Res. Int., 80 (2009), 441.