Optimal Control of Inclusions to Prevent “Sand-Hole” Surface Defects in Deep Cold-Drawn Battery Cups for Electrical Vehicles

Abstract

The present work was conducted to elucidate the influence of inclusionson surface quality of deep cold-drawn battery cups for electrical cars. The obtained results revealed that, to prevent the surface defects of sand-holes in battery cups in deep cold-drawing process, attentions should be paid to steelmaking and casting process for optimal control of inclusions. Because sand-holes were often caused by large Al₂O₃ clusters and CaO–Al₂O₃ particles. Most important, a worthy finding was that these inclusions were not safe when they were over 100 µm and bigger than 27 µm, respectively, which was important for process optimization. By reducing (FeO) contents in ladle slag and tundish covering flux to about 5% or lower, together with optimal fluid flow of molten steel in tundish, such large inclusions can be well decreased to prevent the occurrences of sand-holes in battery cups in industrial production.

1. Introduction

In recent years, to reduce CO₂ emission, manufacturing of electrical vehicles aroused great attentions all over the world.^1,2,3,4,5) To supply sufficient power, thousands of cylindrical battery cells are tightly linked, packaged in a case and equipped to the chassis of a car. A battery cell unit includes a body cup, electrolyte, cathode, anode and a cap.⁶⁾ The cup is often made of deep-drawn steel, in which a flat cold-rolling steel sheet of only 0.30–0.35 mm in thickness is put under a die and subject to multi-pass of stampings in short time. Because of sharp deformation, surface defects often occurred to the cups.^7,8,9)

Ultra-low-carbon steel is often used to produce battery cups for its good drawability. This type of commercial steel is featured by extra-low carbon contents of 20–30 ppm, and is produced by a long industrial process routine of steelmaking, continuous casting, hot-rolling, cold-rolling and cold-drawing, as schematically given in Fig. 1. The process can be briefly described as following:

Fig. 1.

Schematics of industrial production process routine of battery shells. (Online version in color.)

i) At the end of BOF (basic oxygen furnace) primary steelmaking, liquid steel is tapped into a ladle.

ii) Steel is then sent to RH degasser for deeper carburization, during which oxygen is often blown to help remove carbon by promoting [C]-[O] reaction under vacuum atmosphere.

iii) After deep decarburization, oxygen in steel would be removed by adding Fe–Al ferrous alloy, in which the chemical reaction of Al–O occurred and Al₂O₃ inclusions are largely produced.

iv) After chemical composition adjustments of the bulk, molten steel would be held in ladle for about 10 minutes to help floating out inclusions.

v) Liquid steel would be afterwards transferred for continuous casting and slabs are then subjected to hot rolling and cold rolling. So, cold-rolled coils about 0.30–0.35 mm in thickness would be produced.

vi) After 7-passes of cold-drawing, cold-rolling sheets are drawn into battery cups with a wall thickness only about 0.12 mm.

During the cold-rolling and cold-drawing of ultra-low carbon steel, inner defects like inclusions would be gradually exposed to surface or subsurface of the ever-thinner steel to cause surface problems. In the past, surface defects like slivers, blisters, blow-holes etc. in cold-rolled sheets were intensively discussed, and big inclusions like alumina clusters, entrapped mold powder and captured Ar bubbles (often together with inclusions) were often complained.^{10,11,12,13,14,15,16,17,18,19,20)} Recently, in the cold-drawing of battery cells, surface defects of sand-holes were observed in the industrial production. Despite of fine sizes, such defects can possibly cause electrolyte leakage of battery thus must be rejected. Whereas, formation mechanism of sand-holes and the way to effectively prevent them are rarely discussed. Hence, this study was carried out with careful observations of sand-holes, characterizations of inclusions in steel and process optimization to prevent sand-holes in industrial production.

2. Experiments, Samplings and Methodology

The “problem-orientation” method was taken in this study. Firstly, cold-drawing of batteries was tracked to collect battery cups with “sand-hole” defects for careful observations of defects. Then, industrial trials were done in the pyrometallurgy process of steelmaking and casting, to elucidate its influences on the occurrences of sand-hole defects by detailed samplings and analysis of steel and slag. Steel was sampled in ladle after Al deoxidation and holding step, in casting tundish and from as-cast slab. Slag in ladle was taken at the stage of BOF tapping, RH arrival, the end of decarburization, the end of holding and the end of casting. At the end of casting sequence, slag in casting tundish was also sampled. Based on obtained results, the process was optimized and verified by another industrial experiment. So, in the present study, two heats of experiments (labelled as heat 1 and heat 2) were conducted for conventional process, and three heats of experiments (labelled as heat A, heat B and heat C) were carried out for optimal process.

Sand-holes in battery cups were observed by the scanning electron microscope with energy dispersive spectrometry (SEM-EDS). Chemical compositions of sampled slags were analyzed by an X-ray fluorescence spectrometer. Steel samples were prepared for chemical composition analysis and inclusion inspections. Total oxygen (T.O) and [N] contents were tested by the fusion and infrared absorption method. Steel samples were also grinded and mirror-polished for inclusions detection by an automatic SEM-EDS machine (Phenom Particle X) to master the natures (types, sizes, number density etc.) of inclusions, in which accelerating voltage was set as 15 kV.

3. Results and Discussion

3.1. Analysis on the Occurrences of Sand-Holes in Cold-Drawn Battery Cups

3.1.1. Observation of Sand-Holes

The collected battery cups with sand-holes in cold-drawing are typically shown in Fig. 2. As can be seen, sand-holes often occurred in the wall of battery cups (marked by black-colored circles). These tiny defects can be seen by eyes and often looks like punctures in the skins of human. To have more detailed information of sand-holes, they were observed under the SEM-EDS machine.

Fig. 2.

Collected cold-drawn battery cups with sand-holes. (Online version in color.)

It was found that sand-holes can be mainly divided into two types. As shown in Fig. 3(a), many sand-holes were initiated with the formation of concavities in the wall of battery cups. While some other sand-holes were looked like scars and without concavities, as shown in Fig. 3(b). It can be inferred that large stress concentrations were initiated at locations of the 1st type of sand-holes to tear steel matrix and cause cavities. The inference can be verified by SEM observations at larger magnifications in Fig. 4. As can be seen, the 1st type of sand-holes was either caused by Al₂O₃ inclusions or by CaO–Al₂O₃ inclusions. In Fig. 4(a), many Al₂O₃ particles about 5–6 μm scattered along the sand-hole. Moreover, the steel matrix was torn, crinkled and curled. It revealed that stress concentration should be formed near the locations of concavity. As it known, inclusions can easily cause such problems in steel because of different deformability, hardness, thermal expansion and so on.^{21,22,23,24,25)} Moreover, it is accepted that inclusions smaller than 10 μm seldom cause problems in steel. As a result, it was inferred that these scattered Al₂O₃ particles probably originated from a crushed Al₂O₃ cluster. In Fig. 4(b), a group of crushed calcium aluminate particles scattered in an area with a length about 60–70 μm, locating along and inside the sand-hole cavity. Naturally, it can be reasonably inferred that original size of the calcium aluminate inclusion should be smaller than 60–70 μm.

Fig. 3.

The observed two types of sand-holes in cold-drawn battery cups under SEM: (a) with the formation of concavities; (b) without the formation of concavities but looked like scars. (Online version in color.)

Fig. 4.

The 1st type of sand-hole in battery cups (initiated by inclusions): (a) alumina inclusions, (b) calcium aluminate inclusions. (Online version in color.)

The 2nd type of sand-holes was caused by iron oxide, as shown in Fig. 5. Although the iron oxide scale carved in steel matrix was as big as 500–600 μm, no concavity was initiated around it. It can be inferred that stress concentrations caused by iron oxides were much smaller than Al₂O₃ and calcium aluminate inclusions. As iron oxides cannot be formed in liquid steel with Al deoxidation, they should be formed in the reheating of hot-rolling and retained to cause problems in cold-drawing.

Fig. 5.

The 2nd type of sand-hole in battery cups (initiated by iron oxide scales). (Online version in color.)

3.1.2. Initiations of Sand-Holes by Inclusions

From SEM observations, formation of sand-holes in cold-drawn battery cups were schematically illustrated by Fig. 6 and briefly described as following.

Fig. 6.

Schematic illustration on the formation of sand-holes in battery cups during cold drawing. (Online version in color.)

(1) During cold-drawing, steel experienced sharp deformations with ever decreased thickness. Inclusions hidden inside steel were gradually exposed to surface or sub-surface.

(2) Because of sharp deformation, stress suffered by steel would be transferred to inclusions. At inclusions-steel interface, stress concentration would be caused by inclusions. Because they were distinctive from steel in thermal expansion, hardness, deformability etc. For larger inclusions with poor deformability, the initiated stress concentration would be more severe. If the concentrated stress was huge enough, inclusions can be deformed or crushed.

(3) From the classical study of Kiessling,²⁶⁾ it is known that Al₂O₃ and CaO–Al₂O₃ inclusions can be poorly deformed at hot-rolling temperature, let alone at room temperature. It accounts for the observations of crushed Al₂O₃ and CaO–Al₂O₃ inclusions near sand-holes.

(4) Hence, it was understandable that sand-holes caused by Al₂O₃ and CaO–Al₂O₃ inclusions are featured by concavities in steel matrix and with inclusion particles scattering around.

(5) From the study of Kiessling,²⁶⁾ it is also known that iron oxide is more deformable than Al₂O₃ and CaO–Al₂O₃, which means that stress at iron oxide and steel interface would be smaller. As a result, when iron oxides caused sand-holes, the battery cups were free of tears and concavities.

From above results, it can be sum that the critical size of Al₂O₃ inclusions initiating sand-holes were hard to estimate, as they would be crushed into small pieces during the deformation of steel. By contrast, CaO–Al₂O₃ inclusions would only be deformed to some extent. It can be estimated that the CaO–Al₂O₃ inclusion should be originally about 60–70 μm or so. As a result, it can be pointed out that the safe size of a CaO–Al₂O₃ inclusion should be at least smaller than 60–70 μm. As a result, attentions were paid to Al₂O₃ and CaO–Al₂O₃ during the inclusion inspection. On the other hand, as iron oxide scales were originated either from as-cast slabs or hot-rolling plates and can be avoided by careful surface cleaning before cold-rolling and cold-drawing. Therefore, how to prevent sand-holes initiated by iron oxides was not discussed in details in this study.

3.2. Analysis of Conventional Steelmaking and Casting Process

3.2.1. Compositions of Slag and Cleanliness of Steel Melts

Slag compositions of the two heats of industrial trials for conventional process were shown in Table 1, which was featured by strong oxidation ability. In steelmaking, FeO contents in slag always exceeded 8% (concentrations in this paper are all in weight percentages, unless specially noted). At the end of casting, FeO content in ladle slag was decreased to about 5.8%. It meant that molten steel can be easily re-oxidized by top slag. The covering flux in casting tundish was mainly composed of CaO–Al₂O₃ system, with (FeO+MnO), SiO₂ and MgO less than 6%.

Table 1. Chemical compositions of ladle slag and tundish flux under conventional process.

Heat	stage	Compositions of lade slag or tundish flux (wt%)
		CaO	Al2O3	MgO	SiO2	FeO	MnO
Heat-1	BOF tapping	49.42	7.30	5.42	6.01	15.11	2.74
	RH arrival	49.56	13.33	6.86	6.57	11.37	2.89
	End of De–C	46.57	13.92	9.27	6.81	12.11	2.97
	End of RH	44.17	18.69	9.36	6.78	9.71	3.45
	End of casting	43.64	24.34	9.91	7.15	5.85	3.43
Heat-2	BOF tapping	44.23	12.77	6.69	6.27	14.66	2.50
	RH arrival	43.57	17.14	7.77	6.56	13.65	2.75
	End of De–C	43.09	17.77	7.93	6.81	12.94	2.92
	End of RH	41.84	21.34	7.91	6.67	11.12	3.18
	End of casting	42.27	28.07	8.45	6.77	5.85	3.60
Tundish flux		44.92	33.90	4.06	5.90	3.69	3.42

Note: ① As heat 1 and heat 2 were consecutive heats of one casting sequence, tundish flux was considered as the same. ② After De–C and before the end of RH, Al deoxidation and alloying have been carried out to steel.

Cleanliness of steel melts were shown in Fig. 7. At the end of RH refining, T.O and [N] contents in the heat 1 were 0.0030% and 0.0029%, respectively. In casting tundish, T.O and [N] contents sharply increased to 0.0058% and 0.0056%, respectively. In industrial practice, pick-ups of oxygen and nitrogen pick-ups in steel often occurred to the 1st heat of a casting sequence, during which molten steel was poured to an empty tundish thus can be easily contaminated by retained air. Hence, pick-ups of T.O and [N] in heat 2 were much relieved at the initial stage of casting. Nevertheless, it was noticed that T.O contents in the heat 2 were always higher in casting than at the end of RH refining. It should be attributed to much higher (FeO+MnO) about 14.30% in ladle slag of heat 2 to cause continuous re-oxidization of steel in teeming ladle before liquid steel was poured into the casting tundish.

Fig. 7.

Contents of T.O and [N] in steel melts at the end of RH refining and continuous casting under conventional process: (a) heat 1; (b) heat 2.

3.2.2. Inclusions in Steel

(1) Inclusions in Ladle and Casting Tundish

Many inclusions were Al₂O₃ clusters while some others (occupying a small fraction) were CaO–Al₂O₃ particles, as shown in Fig. 8. Strong tendency of Al₂O₃ into aggregates was attributed to its poorer wettability in steel.²⁷⁾

Fig. 8.

Al₂O₃ and CaO–Al₂O₃ inclusions in teeming ladle and casting tundish under conventional process: (a) Al₂O₃: in teeming ladle (1–4), in casting tundish (5–8); (b) CaO–Al₂O₃: in teeming ladle (1–4), in casting tundish (5–8). (Online version in color.)

(2) Inclusions in As-Cast Slab

As it known, inclusions can experience important changes during solidification of steel, by collisions and agglomerations into larger ones. If they were captured by solidifying slab shell, surface defects in cold-rolled coils can be caused. So, in the present work, careful inspections of large inclusions locating at varied depths beneath slab surface were carried out, by the automatic SEM-EDS machine. It was found that big Al₂O₃ clusters (as shown in Fig. 9) often existed at a depth of 0–5 mm under slab surface. Whereas, CaO–Al₂O₃ inclusions were seldomly observed in subsurface areas of slab, and only one CaO–Al₂O₃ inclusion about 13 μm was seen at the depth of 0.5 mm beneath slab surface, as shown in Fig. 10.

Fig. 9.

Large Al₂O₃ clusters locating at varied depths beneath slab surface under conventional process: (a-1)–(a-2)--0.5 mm, (b-1)–(b-2)-1.0 mm , (c-1)–(c-2)--2.0 mm, (d-1)–(d-2)--3.0 mm. (Online version in color.)

Fig. 10.

CaO–Al₂O₃ inclusion inspected 0.5 mm beneath slab surface under conventional process. (Online version in color.)

Previously, surface defects in cold-rolled ultra-low carbon steel by inclusions aroused great attentions.^{28,29,30,31,32,33,34,35,36)} From these studies, it was known that large inclusions should be avoided as much as possible. Emi et al. and Ikeda et al. pointed out that the safe size of an inclusion in slab to avoid surface defect in cold-rolled coils and deep-drawn ironed (DI) cans should be ≤240 μm and ≤50 μm, respectively.^34,35,36) Kobayashi et al. reported that inclusions causing “flange” in DI cans (with five-passes of stamping) were often composed of CaO–Al₂O₃ and the critical size of them to avoid tearing in flanging of the cans was about 50 μm.³⁷⁾ While some other works believed that, to ensure good surface of DI cans, inclusions in slab should be ≤20 μm.³⁸⁾ Despite of discrepancies on the safe size of inclusions for DI cans, there is no doubt that ever-minimized sizes of inclusions would be more desirable to reduce surface defects. In this study, maximum sizes of detected big inclusions at varied depths under slab surface were summarized in Table 2. It can be seen that the largest Al₂O₃ clusters at the depths of 0.5 mm, 1.0 mm, 2.0 mm and 3.0 mm beneath slab surface was 216 μm, 745 μm, 124 μm and 251 μm, respectively. By contrast, fewer big CaO–Al₂O₃ inclusions located in the sub-surface depths of slab and the observed one was only about 13 μm. So, it was understandable that sand-holes were much more frequently caused by Al₂O₃ than CaO–Al₂O₃ inclusions.

Table 2. Maximum sizes of observed inclusions at slab sub-surface regions under conventional process.

Location of inclusions	Maximum size of inclusion		Scanned area
	Al2O3 cluster	Calcium aluminate
0.5 mm beneath surface	216 μm	13 μm	102.51 mm2
1.0 mm beneath surface	745 μm	Not detected	184.53 mm2
2.0 mm beneath surface	124 μm	Not detected	262.44 mm2
3.0 mm beneath surface	251 μm	Not detected	184.53 mm2

3.2.3. Estimation on the Safe Size of Inclusion

Because of tiny thickness about 0.11 mm, it is widely accepted that DI cans have the most stringent cleanness specifications. As battery cups (about 0.12 mm) are thinner than auto panels (about 0.6 mm) while a little thicker than DI cans (about 0.11 mm), it can be theoretically inferred that the safe size of inclusions in battery cups would be smaller than auto panels while approximate to DI cans. Despite of that, whereas, battery cups suffer from tougher cold-drawing than DI cans. So, the safe size of inclusions in battery cups is worthy of quantitative evaluation. However, this topic had been rarely discussed in the past and it was thus furtherly clarified in this study.

During inclusion inspection by the automatic SEM-EDS machine, information of all the detected inclusions was recorded in an excel data sheet. Except shapes, 2-dimensional sizes, compositions etc. of inclusions, perpendicular depths of inclusions in steel were also obtained and listed as “D_perp” in a data sheet, as shown in Fig. 11. As a result, the volume of near-spherical-shaped inclusions can be estimated by their 2-dimensional sizes together with the value of “D_perp”, viz., the 3-dimensional size of CaO–Al₂O₃ inclusions in this study could be reasonably deduced. Naturally, this estimation is hard to conduct for cluster-shaped Al₂O₃ inclusions.

Fig. 11.

The data sheet of inclusions to track calcium aluminates. (Online version in color.)

It was found that CaO–Al₂O₃ inclusions in steel were mostly smaller than 5 μm and the biggest one was about 25 μm (“D_max” in datasheet refers to maximum size of the inclusion), while the maximum “D_perp” of inclusion was always within 5 μm. So, volume of CaO–Al₂O₃ inclusions in the above Fig. 4(b) can be estimated as following:

i) As the crushed CaO–Al₂O₃ covering a 2-dimensional area about 70 μm in length and 30 μm in width, with “D_perp” taken as 5 μm, volume of the inclusion can be calculated as V=70 μm (length) × 30 μm (width) × 5 μm (perpendicular depth) = 10500 μm³.

ii) As the CaO–Al₂O₃ inclusion was originally in near-spherical or spherical shape, radius of it in the as-cast slab could be cautiously deduced as 13.5 μm (viz. with a diameter of 27 μm) by formula (1), ignoring its deformation in hot-rolling and cold-rolling.   

$$V=\left(4\pi r^3\right)/3$$(1)

So, to prevent sand-holes in battery cups, CaO–Al₂O₃ inclusions in slab should be at least ≤27 μm. As battery cups are thicker than auto panels and thinner than DI cans, this estimation was consistent with previous works, in which inclusions ≤100 μm and ≤20 μm were harmless in cold-rolled auto panels (about 0.6–0.7 mm in thickness) and cold-drawn DI cans (about 0.10 mm in thickness), respectively.

4. Process Optimization for Improved Control of Inclusions

Considering important effects of large Al₂O₃ and CaO–Al₂O₃ inclusions on surface quality of battery cups, formation mechanisms of the two types of inclusions were clarified to optimize conventional process.

4.1. Formation Mechanisms of Large Al₂O₃ and CaO–Al₂O₃ Inclusions

At the end of RH refining in steelmaking, when oxygen gas was blown into RH vacuum chamber for deep decarburization, dissolved oxygen ([O]) in steel would be high. Chemical reaction (2) would proceed to increase (FeO) contents in slag, which means that oxygen was not only contained in steel bulk but also “hidden” in ladle slag.   

$$\left[\text{Fe}\right]+\left[\text{O}\right]={\left(\text{FeO}\right)}_{\text{slag}}$$(2)

When FeAl alloy was added into steel for Al deoxidation, chemical reaction (3) occurred to reduce [O] and cause large formations of Al₂O₃ inclusions.³⁹⁾ Thermodynamic equilibrium of reaction (3) can be evaluated by Eq. (4). As Al₂O₃ inclusions were poorly wetted by steel, they agglomerated into bigger clusters.   

$$2\left[\text{Al}\right]+3\left[\text{O}\right]={({\text{Al}}_2{\text{O}}_3)}_{\text{in}}$$(3)

  

$${\text{lgK}}_3^{°}=\frac{64\mathrm{\hspace{0.17em} }000}{\text{T}}-20.57$$(4)

Because of sharply reduced [O] in steel by Al deoxidation, equilibrium of reaction (2) would be broken and the reaction would reversely proceed as reaction (5) to supply oxygen from slag to steel, which can be evaluated by Eq. (6).³⁹⁾ The re-oxidation of steel by (FeO) in slag caused re-formation of Al₂O₃, as expressed by chemical reaction (7). It can be expected that many new Al₂O₃ clusters would be form in steel. As reaction (8) was deduced from reactions (3) and (5), Eq. (9) can be obtained from Eqs. (4) and (5) to quantitatively evaluate the negative effects of (FeO) in slag on control of Al₂O₃ inclusions. Comparing the reaction quotient Q₇ in Eq. (9) with equilibrium constant, occurrence of reaction (7) can be expected.   

$${\left(\text{FeO}\right)}_{\text{slag}}=\left[\text{Fe}\right]+\left[\text{O}\right]$$(5)

  

$${\text{lgK}}_5^{°}=\frac{-6\mathrm{\hspace{0.17em} }150}{\text{T}}+2.064$$(6)

  

$$3{\left(\text{FeO}\right)}_{\text{slag}}+2\left[\text{Al}\right]={({\text{Al}}_2{\text{O}}_3)}_{\text{in}}+3\left[\text{Fe}\right]$$(7)

  

$$lg{K}_7^{°}=\frac{45\mathrm{\hspace{0.17em} }550}{T}-14.348$$(8)

  

$$Q_7=\frac{a_{\left[Fe\right]\times a_{A{l}_2{O}_3}}^3}{a_{\left[Al\right]}^2\times a_{FeO}^3}$$(9)

Referring to slag compositions in Table 1, activities of (Al₂O₃) and (FeO) in slag were calculated by the Factsage software (with FToxid 7.3 database) for slag with CaO, Al₂O₃, SiO₂, MgO about 40–45%, 18–20%, 6–7% and 8–9%, respectively. Activity coefficients of [Al] and [Fe] were considered as 1 while [Al] content was taken as 0.075%.

The calculation results were given in Table 3. As can be seen, only when (FeO) was decreased to 5% or lower, re-oxidation of steel by top slag can be well relieved. This calculation agreed with experimental data in Table 1, in which contents of (FeO) in ladle slags of the two heats of industrial trials were reduced to 5.85% at the end of casting. It meant that even when steel in teeming ladle was all casted, re-oxidization of steel by ladle slag was still going on. Hence, to stably target (FeO) contents in teeming ladle slag to ≤5% before casting was very important and set as a goal for the optimal process.

Table 3. Effect of (FeO) contents in slag on the re-oxidation of steel.

FeO content	5%	10%	15%
aFeO	4.03×10−2	8.19×10−2	1.30×10−1
aAl2O3	9.63×10−4	1.09×10−4	1.52×10−3
Qi	2.60×103	3.51×102	1.24×102
K	6.20×109	6.20×109	6.20×109

During the continuous casting, exposure of steel to air often happened to the 1st heat of a casting sequence, in which oxygen dissolved into steel by reaction (10), accounting for pick-ups of T.O and [N] in heat 1 in industrial trials under conventional process. Then, re-formations of Al₂O₃ clusters would occur by reaction (3).   

$${\mathrm{O}}_2\left(\mathrm{g}\right)=2\left[\mathrm{O}\right]$$(10)

So, formation mechanisms of big Al₂O₃ clusters degrading the surface quality of battery cups were proposed as following:

(1) Because of oxygen blowing for deep decarburization in RH degasser, dissolved oxygen ([O]) in steel and contents of (FeO) in ladle slag were very high.

(2) In Al deoxidation, [O] was decreased with large formation of Al₂O₃ inclusions. Because of poor wettability, these primary Al₂O₃ inclusions agglomerated into big clusters and many of them were possibly be floated out during the holding of steel before casting.

(3) However, if FeO in ladle slag was as high as 11.79–13.72%, re-oxidation of Al-deoxidized steel by the slag would happen and continuously proceed till the end of casting, witnessing large and continuous re-formations of Al₂O₃, as shown in Fig. 12(a).

Fig. 12.

Schematics on the continuous re-oxidation of steel by oxidized slag: (a) causing large re-formation of Al₂O₃, (b) agglomerations of re-formed Al₂O₃ into dangerous clusters. (Online version in color.)

(4) Only when (FeO) in slag was reduced to 5% or lower, re-oxidation of steel by the oxidizing ladle slag would be well relieved.

(5) Besides, during continuous casting, re-oxidation of steel often occurred to the 1st heat of a casting sequence, during which the injected steel into the empty tundish can be easily polluted by the air. Whereas, this type of re-oxidation of steel is often temporary and indicates pick-ups of T.O content in 1st heat.

(6) As the newly formed Al₂O₃ inclusions by re-oxidation have shorter time to float out, as shown in Fig. 12(b), many of them went into casting mold together with the pouring steel and agglomerated into bigger clusters. When they were captured by solidifying slab shell, troubles can be probably caused in cold-drawing of battery cups.

(7) In a word, decreasing large Al₂O₃ clusters in steel is critical to improve surface quality of cold-drawn battery cups. And the key point is to stably keep (FeO) in ladle slag and tundish flux to about 5% or even lower.

Intrinsic formation of CaO–Al₂O₃ inclusions in steel was also estimated. During steelmaking, [Al] in steel can reduce (CaO) in the basic and reducible top slag to supply [Ca] into steel bulk, as expressed by chemical reaction (11). The reduced [Ca] would in turn react with Al₂O₃ inclusions to modify them into CaO–Al₂O₃. However, with the rise of (FeO) content in slag, reaction (7) would compete with reaction (11).   

$$3{\left(\text{CaO}\right)}_{\text{slag}}+2\left[\text{Al}\right]={({\text{Al}}_2{\text{O}}_3)}_{\text{slag}}+3\left[\text{Ca}\right]$$(11)

Only when the priority of reaction (11) was over reaction (7), CaO–Al₂O₃ inclusions could be intrinsically formed by chemical reactions in steel. Based on reaction (12) and reaction (5), reaction (14) can be obtained, together with deduced Eq. (14) by Eqs. (6) and (13).³⁹⁾   

$$\left[\text{Ca}\right]+\left[\text{O}\right]={\left(\text{CaO}\right)}_{\text{in}}$$(12)

  

$$lg{K}_{12}^{°}=\frac{7\mathrm{\hspace{0.17em} }220}{T}+3.292$$(13)

  

$${\left(\text{CaO}\right)}_{\text{slag}}+\left[\text{Fe}\right]={\left(\text{FeO}\right)}_{\text{slag}}+\left[\text{Ca}\right]$$(14)

  

$$lg{K}_{14}^{°}=\frac{-1\mathrm{\hspace{0.17em} }070}{T}-5.36$$(15)

So, reaction (14) was thermodynamically evaluated to predict intrinsic formation of CaO–Al₂O₃ inclusions. From Eq. (15), it could be known that equilibrium constant of reaction (14), $lg{K}_{14}^{°}$, was very small at 1873 K, which meant that the reaction would be hard to proceed in steelmaking. Therefore, it can be reasonably pointed out that CaO–Al₂O₃ inclusions causing sand-holes in battery cups should be mainly exogenous particles. As covering flux in casting tundish was mainly composed of CaO–Al₂O₃ system, it should be the most probable origin of negative CaO–Al₂O₃ inclusions. Moreover, during the inspection of CaO–Al₂O₃ inclusions, it was noticed that the element Si was not detected, as shown in above Figs. 4(b) and 10. It is possible that, although some SiO₂ was contained in tundish flux, SiO₂ in the entrapped slag particle can be possibly reduced by [Al] during their long retainment in steel.

4.2. Process Optimization for Improved Control of Inclusions

For improved control of inclusions, the conventional process was optimized. The key point was to stably target ever-decreased (FeO) in ladle slag and tundish flux to about 5% or lower, to relieve continuous re-oxidation of steel by slag. Also, exposure of steel in casting tundish to air was also more carefully prevented especially for the 1st heat of casting sequences, together with intended preventions of (FeO) pick-ups in tundish flux. By these precautions, large Al₂O₃ clusters over 100 μm were effectively decreased. In order to reduce the frequency of entrapped CaO–Al₂O₃ flux in tundish, fluid flow of steel in tundish was also optimized and a refractory screener was equipped inside the tundish.

To verify positive effects of these optimizations, industrial trials were carried out. 3 heats (heat A, B and C) of steel in a 5-heat casting sequence were sampled. As can be seen from Table 4, very low (FeO) contents in teeming ladle slag were achieved before casting, about 5.49%, 7.24% and 3.18% in heat A, heat B and heat C, respectively. (FeO) contents in tundish flux were also extremely low in casting process, about 0.88%, 1.18% and 1.68% for heat A, heat B and heat C, respectively. It was found that much higher cleanliness of steel melts was realized in ladle and casting tundish after the process optimization, as can be seen in Fig. 13.

Table 4. Chemical compositions of ladle slag and tundish flux under optimized process.

Heat	stage	Compositions of lade slag or tundish flux (wt%)
		CaO	Al2O3	MgO	SiO2	FeO	MnO
A	BOF tapping	48.79	2.35	6.24	14.19	11.81	4.42
	RH arrival	46.01	27.73	9.27	7.40	4.51	1.23
	End of De–C	41.97	26.92	10.73	7.40	6.33	3.03
	End of RH	39.47	31.67	10.15	6.99	5.91	3.04
	End of casting	39.93	33.78	9.68	6.54	5.49	2.70
	Tundish flux	46.57	42.27	3.22	4.46	0.88	0.58
B	BOF tapping	46.71	1.59	6.06	14.59	14.12	4.67
	RH arrival	40.78	8.97	5.47	4.06	16.79	3.87
	End of De–C	42.58	22.08	6.28	4.68	13.28	3.96
	End of RH	40.08	27.24	6.24	4.58	11.28	4.31
	End of casting	42.04	33.40	6.41	4.64	7.24	0.87
	Tundish flux	48.27	41.69	2.26	3.86	1.18	0.49
C	RH arrival	53.95	27.96	5.93	4.88	3.01	1.43
	End of De–C in RH	50.80	28.32	7.68	5.20	3.58	1.75
	End of RH	48.74	30.19	8.32	5.05	3.64	1.84
	After holding	48.48	31.81	8.46	4.92	3.18	1.72
	Tundish flux	46.73	40.81	3.22	4.02	1.68	0.65

Fig. 13.

T.O and [N] contents in steel melts at the end of RH refining and in casting tundish under optimized process: (a) T.O, (b) [N].

In the optimized process, large inclusions locating 0.5–3.0 mm under slab surface were also inspected by the automatic SEM-EDS machine. Large inclusions under different depths of slab surface were shown in Fig. 14. It was found that the observed big inclusions were all Al₂O₃ clusters while CaO–Al₂O₃ inclusions were not seen. Maximum sizes of inspected Al₂O₃ clusters were summarized in Table 5, together with scanned area of steel samples under the SEM-EDS. It can be clearly seen that, after the process optimization, Al₂O₃ inclusions in subsurface regions of as-cast slab were all smaller than 100 μm, within the safe size previously proposed for cold-drawing products.

Fig. 14.

Large inclusions at different depths beneath slab surface under optimized process. (Online version in color.)

Table 5. A summary on the maximum size of inclusions at varied depths under slab surface after the optimization of process.

Heat	Depth under slab surface	Maximum size of inclusion	Scanned area of steel
A	0.5 mm beneath surface	94.3 μm	563 mm2
	1.0 mm beneath surface	43.1 μm	593 mm2
	2.0 mm beneath surface	44.6 μm	555 mm2
	3.0 mm beneath surface	82.7 μm	569 mm2
B	0.5 mm beneath surface	61.8 μm	527 mm2
	1.0 mm beneath surface	48.6 μm	547 mm2
	2.0 mm beneath surface	29.5 μm	581 mm2
	3.0 mm beneath surface	38.7 μm	537 mm2
C	0.5 mm beneath surface	51 μm	659 mm2
	1.0 mm beneath surface	75.8 μm	546 mm2
	2.0 mm beneath surface	43.8 μm	538 mm2
	3.0 mm beneath surface	37.8 μm	564 mm2

Most important, in the afterwards cold-drawing process, no occurrences of “sand hole” were claimed by customers, revealing that the process optimizations were successful.

5. Conclusions

This study focused on how to prevent the surface defects of sand-holes in the cold-drawing of battery cups for electrical vehicles. By careful observations on sand-holes and characterizations of inclusions in steelmaking and casting, relationship between sand-holes and inclusions in steel was elucidated. It was found that large inclusions, mainly Al₂O₃ clusters over 100 μm and CaO–Al₂O₃ particles over 27 μm, were responsible for sand-holes in cold-drawn battery cups. Large Al₂O₃ clusters were mainly formed during the continuous re-oxidation of liquid steel by slag with high (FeO) contents in teeming ladle and casting tundish. While the CaO–Al₂O₃ particles were mainly entrapped tundish flux composed of CaO–Al₂O₃ system. By stably targeting (FeO) in slag to about 5% or lower, large Al₂O₃ clusters in as-cast slab were efficiently decreased. With optimal control on the flow of molten steel in casting tundish and the use of a refractory screener equipped inside tundish, CaO–Al₂O₃ inclusions were also well prevented. As a result, battery cups were produced with decreased occurrences of sand-holes in cold-drawing process.

Acknowledgements

Authors sincerely appreciate the financial supports from National Key Research and Development Program (2021YFB3401001) and Fundamental Research Funds for Central Universities (FRF-DF-20-08).

Conflict of Interest Statement

All authors state that there is no conflict of interests.

Data Availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

References

• 1)   M. M.  Titirici: Adv. Energy Mater., 11 (2021), 2003700. https://doi.org/10.1002/aenm.202003700
• 2)   L. L.  Liu,  N.  Solin and  O.  Inganäs: Adv. Energy Mater., 11 (2021), 2003713. https://doi.org/10.1002/aenm.202003713
• 3)   D. G.  Mackanic,  M.  Kao and  Z. N.  Bao: Adv. Energy Mater., 10 (2020), 2001424. https://doi.org/10.1002/aenm.202001424
• 4)   Y. E.  Durmus,  H.  Zhang,  F.  Baakes,  C.  Desmaizieres,  H.  Hayun,  L. T.  Yang,  M.  Kolek,  V.  Küpers,  J.  Janek,  D.  Mandler,  S.  Passerini and  Y.  Ein-Eli: Adv. Energy Mater., 10 (2020), 2000089. https://doi.org/10.1002/aenm.202000089
• 5)   J.  Apt,  S. B.  Peterson and  J. F.  Whitacre: Science, 333 (2011), 823.
• 6)   I.  Morse: Science, 372 (2021), 780.
• 7)   J. S.  Li: Hebei Metall., 271 (2018), 25.
• 8)   F.  Xue,  J. L.  Liu,  S. K.  Ji and  Y. H.  Fan: Baosteel Technol., 247 (2014), 67.
• 9)   H. B.  Zhang,  M. L.  Li and  M. Y.  Li: Hebei Metall., 249 (2016), 1.
• 10)   J.  Kubota,  K.  Okimoto,  A.  Shirayama and  H.  Murakami: Proc. 1991 Steelmaking Conf., ISS, Warrendale, PA, (1991), 233.
• 11)   T.  Miyake,  M.  Morishita,  H.  Nakata and  M.  Kokita: ISIJ Int., 46 (2006), 1817. https://doi.org/10.2355/isijinternational.46.1817
• 12)   W. H.  Emling and  B. V.  Lakshminarayana: Iron Steelmaker, 27 (2000), 73.
• 13)   J. S.  Powers,  W. H.  Emling and  C. E.  Tomazin: Iron Steelmaker, 16 (1989), 37.
• 14)   H. T.  Tsai,  W. J.  Sammon and  D. E.  Hazelton: Proc. 1990 Steelmaking Conf., (Detroit), ISS, Warrendale, PA, (1990), 49.
• 15)   T.  Ishikawa,  N.  Yukawa,  Y.  Yoshida and  Y.  Tonohata: Tetsu-to-Hagané, 89 (2003), 1142 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.89.11_1142
• 16)   W. F.  Gao: Adv. Mater. Res., 402 (2012), 221.
• 17)   G. R.  Wu: Adv. Mater. Res., 339 (2011), 134.
• 18)   S.  Das,  S.  Roy,  S.  Nayak,  T.  Bhattacharyya and  S.  Bhattacharyya: Eng. Fail. Anal., 44 (2014), 95.
• 19)   P. K.  Tripathy,  S.  Das,  M. K.  Jha,  J. B.  Singh,  A. M.  Kumar and  A. K.  Das: Ironmaking Steelmaking, 33 (2006), 477.
• 20)   M.  Zeze,  A.  Tanaka and  R.  Tsujino: Tetsu-to-Hagané, 87 (2001), 85 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.87.2_85
• 21)   U.  Zerbst,  M.  Madia,  C.  Klinger,  D.  Bettge and  Y.  Murakami: Eng. Fail. Anal., 98 (2019), 228.
• 22)   H. A.  Tinoco,  S.  Fintova,  I.  Heikkila,  D.  Herrero,  T.  Vuoristo,  I.  Dlouhy and  P.  Hutar: Mater. Sci. Eng. A, 856 (2022), 144009.
• 23)   S.  Norasethasopon and  K.  Yoshida: Mater. Sci. Eng. A, 422 (2006), 252.
• 24)   S.  Mukherjee,  A.  Kundu,  P. S.  De,  J. K.  Mahato,  P. C.  Chakraborti,  M.  Shome and  D.  Bhattacharjee: Mater. Sci. Eng. A, 776 (2020), 139029.
• 25)   Y. J.  Wu,  J.  Uusitalo and  A. J.  DeArdo: Mater. Sci. Eng. A, 828 (2021), 142070.
• 26)   R.  Kiessling: Non-metallic Inclusions in Steel, The Institute of Metals, London, (1989), 93, 46.
• 27)   H. B.  Yin,  H.  Shibata,  T.  Emi and  M.  Suzuki: ISIJ Int., 37 (1997), 936. https://doi.org/10.2355/isijinternational.37.936
• 28)   K.  Kumai,  H.  Matsunaga,  K.  Asano and  T.  Saeki: Tetsu-to-Hagané, 60 (1974), 1325 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.60.9_1325
• 29)   H.  Yasunaka,  R.  Yamanaka,  T.  Inoue and  T.  Saito: Tetsu-to-Hagané, 81 (1995), 529 (in Japanese). https://doi.org/10.2355/tetsutohagane1955.81.5_529
• 30)   O.  Dankert,  J. P.  Zijp,  A.  van Veen and  A. A.  Kamperman: ISIJ Int., 42 (2002), 1225.
• 31)   Y.  Miki and  S.  Takeuchi: ISIJ Int., 43 (2003), 1548. https://doi.org/10.2355/isijinternational.43.1548
• 32)   M.  Hanao and  M.  Kawamoto: ISIJ Int., 48 (2008), 1210. https://doi.org/10.2355/isijinternational.48.1210
• 33)   A.  Kumar,  S. K.  Choudhary and  S. K.  Ajmani: ISIJ Int., 52 (2012), 2305. https://doi.org/10.2355/isijinternational.52.2305
• 34)   Y.  Sahai and  T.  Emi: Tundish Technology for Clean Steel Production, World Scientific Publishing Co. Ltd, Singapore, (2008), 19.
• 35)   T.  Emi: Proc. Metal Separation Technologies beyond 2000, ed. by K. C. Liddel and D. J. Chaiko, TMS, Warrendale, PA, 1999, 207.
• 36)   M.  Ikeda: History of Continuous Casting Technology of Steel in Japan, ISIJ, Tokyo, (1996), 482.
• 37)   H.  Kobayashi,  T.  Kurokawa,  T.  Shimomura,  K.  Matsudo and  S.  Miyahara: Trans. Iron Steel Inst. Jpn., 23 (1983), 410.
• 38)   Z. Z.  Liu and  K. K.  Cai: Iron Steel, 35 (2000), 64 (in Chinese).
• 39)  Japan Society for the Promotion of Science: Recommended Values of Equilibrium Constants for the Reactions in Steelmaking, ed. by 19th Committee, Japan Society for the Promotion of Science, Tokyo, (1984), 99.