Effect of Hot Metal on Decarburization in the EAF and Dissolved Sulfur, Phosphorous, and Nitrogen Content in the Steel

ISIJ International, Vol. 55 (2015), No. 3, pp. 491–499

Baek LEE and Il SOHN*

Materials Science and Engineering, Yonsei University, Seodaemungu, Seoul, 120-749 Korea. (Received on June 7, 2014; accepted on December 8, 2014)

The effect of hot metal additions on the decarburization and dissolved sulfur, phosphorous, and nitrogen content in the steel of a DC EAF was investigated. The addition of hot metal of maximum 36% provided heat into the EAF allowing faster melting and reduced power on times of the furnace. The increased melting rate lowered the FeO in the slag and seems to allow faster kinetics for CaO dissolution into the slag. Hot metal utilization into the slag resulted in higher phosphorous distribution ratios compared to hot metal free heats achieving comparable phosphorous levels at the ladle metallurgical furnace even though the initial input phosphorous was much higher than hot metal free heats. The effect of hot metal on the desulfurization was not pronounced and no apparent difference could be ascertained compared with the 100% scrap charge. With the addition of carbon saturated iron units in the EAF, the evolution of CO and foaming was promoted, which inhibited the infiltration of nitrogen and lowered the overall partial pressure of nitrogen resulting in lower nitrogen levels in the steel. In addition, the dilution effect of the tramp elements Cu and Sn with hot metal ensured the critical defect index to be less than 10% resulting in a 50% reduction of the quality defect index.

KEY WORDS: hot metal; steelmaking; steel chemistry; phosphorous; sulfur; nitrogen.

which has significant sensible heat for reduced power con- 1. Introduction sumption in the EAF. With the on-going demands for energy efficiency, envi- Zinurov et al.2) described the effects of hot metal utiliza- ronment, low cost, and flexibility for steelmaking operation in 180 ton/ch arc furnace, where substitution of 1% of tions, while allowing the production of high value-added scrap by hot metal at 1 553 K (1 280°C) lead to power sav- steels, there has been significant development of the mini- ings of 3.5–4.5 kWh/t. However, hot metal introduction into mill process route.1) This operation unlike the integrated the furnace was an issue as problems arose from the radiant steel mills typically utilizes an electric arc furnace (EAF) heat loss and graphite liberation from the hot metal, which with secondary steelmaking facilities and a thin slab caster deposited around the plant equipments increasing mainte- connected to a continuous hot strip mill, where the iron feed nance and particulates in the plant. Duan et al.3) showed 35 is from scrap. However, due to increased consumer demands to 40% hot metal additions into a 50 ton/ch EAF after pre- for steel-cleanliness and low-cost, there has been increased treating the carbon saturated iron source reduced tap-to-tap pressure on the EAF steel producers to provide effective times of the furnace by 5 to 10 minutes and decreased the means for satisfying customer needs, as existing feed mate- electricity consumption by 35–50 kWh/t while maintaining rials of scrap continues to degrade in quality. In addition, the chemical compositional requirements and quality of the thin slab casters, which can be highly productive when opti- steel. In addition, although desulfurization potential was not mized, require better internal quality of the steels compared significantly affected by hot metal additions, the de- to conventional slab casters due to its high casting speed and phosphorization was enhanced with hot metal. Others4–6) high heat extraction directly being charged into the contin- have also indicated significant costs savings for plants that uous hot strip mill. In order to meet these critical demands, have hot metal flexibility to add into the EAF as a source the EAF steelmakers have attempted to utilize iron sources for both high quality iron and energy. Gottardi et al.7) from various virgin iron sources including direct reduced described oxygen injection practices to increase hot metal iron, pig iron, and also hot metal. In particular, as electricity usage in the EAF without significant increases in the tap-to- costs begin to rise, there has been increased emphasis in tap times, which can affect the productivity of the furnace. possible utilization of hot metal for facilities having excess However, few studies have described in detail the chemical hot metal especially during downturns in the steelshop, composition changes with hot metal additions as carbon saturated iron units are added into the EAF to the knowledge * Corresponding author: E-mail: [email protected] of the present authors. DOI: http://dx.doi.org/10.2355/isijinternational.55.491 In this study, the effect of hot metal additions to the EAF

491 © 2015 ISIJ ISIJ International, Vol. 55 (2015), No. 3 in terms of chemical composition changes has been studied. the average tap temperature observed. Effect of hot metal additions up to approximately 36% in the 130 ton/ch DC-EAF on the decarburization rate, phospho- 2.2. Chemical Composition of the Materials Input and rous, sulfur, and nitrogen content has been specifically Analysis examined. In addition, the inclusion chemistry with hot met- The hot metal chemical and thermal characteristics of the al additions has also been discussed. major components in the present study are given in Table 2. The hot metal is de-siliconized and hot metal pre-treated to lower excess slag formation in the EAF and lower the 2. Experimental Method and Procedure overall input sulfur into the steel melt. The incoming tem- 2.1. Hot Metal Supplement Procedure into the EAF perature of the metal was on average 1 595.5 K (1 322.5°C). The schematic of the charging sequence of raw materials Typical scrap types vary with market conditions and opera- input into the EAF is shown in Fig. 1. EAF cover is opened tional parameters such as power input and oxygen lancing and the scrap bucket containing 90–105 tons is placed above are slightly modified depending on the size, density, and the the EAF. The scrap is dropped into the furnace containing chemistry of the incoming scrap. Although the present study the hot heel of the previous charge with the light scrap ini- has attempted to unify the type of scraps used in the trials, tially loaded into the furnace to protect the refractory. The it was not possible to completely control the type of scraps heavy scrap follows the light scrap and the small scrap piec- used in the span of over 200 heats. Table 3 provides the es including the shredded scrap are charged into the furnace average composition of the scrap used. The cold pig iron at the end to protect the boring electrode during the initial units have also been applied as a secondary source of car- power on time. With the start of arcing the electrode, addi- bon, which have been compared with heats utilizing hot tional fluxes of 2 tons of quick lime and 0.5 tons of dolomite is added at the top of the scrap and the electrode bores into the scrap as it descends into the furnace. At the end of the boring cycle, the electrode is raised and 50–65 tons of hot metal is added into the crater created during the boring cycle of arcing. The hot metal is added slowly for about 3–4 min. into the furnace through a charging ladle after the EAF cover is removed. Excessively fast pouring can cause splashing and unnecessary vaporization resulting in metal losses. After the hot metal pouring is completed, cover is returned onto the furnace and arcing resumes and additional fluxes of 1.5 tons of quick lime and 0.5 tons of dolomite is added through the Fig. 1. Schematic of the scrap and hot metal charging into the 130 ton/ch DC-EAF. charging chute. Further melting and refining is continued until tapping the furnace. The typical target temperature range of the steel at tap is Table 1. General characteristics of the 130 ton/ch DC-EAF used between 1 853 to 1 903 K with an average of 1 878 K using in the present study. an R-type (Pt/13mass%Rh-Pt) thermocouple inserted into a Item Details Remarks combined sensor, which is discussed later. Steel temperature is measured after complete meltdown of the input materials Furnace type DC-EAF Twin shell and is taken twice. Once at approximately 80% of the tap- Capacity 130 ton/ch to-tap time and before final tap. Temperature as high as Transformer capacity 102 MVA (51 MVA·2sets) 1 943 K have been recorded for heats requiring additional Shell dimensions Dia. = 7 000 mm, heat or due to overshooting. But these cases are anomalies H. = 3 950 mm that been filtered in most cases. In the present study, the Tapping type EBT Eccentric bottom tapping temperature condition for the thermodynamic evaluations has been set at 1 873 K for simplification, which is close to Typical tapping time 50 min

Table 2. Hot metal chemical and thermal characteristics of the major components in the present study.

Element C Si Mn P S Ti Cu Sn Temp (K) Range 4.2–4.5 0.22–0.59 0.20–0.36 0.077–0.113 0.012–0.034 0.023–0.174 0–0.005 Trace 1 554–1 637 Avg. 4.35 0.405 0.28 0.095 0.023 0.0985 0.0025 Trace 1595.5

Table 3. Average composition of the input scrap, cold pig iron units, and the tapped steel composition for the present study.

Element Fe C P S Cu Ni Cr Sn N Balance Scrap 98.10 0.11 0.02 0.03 0.26 0.10 0.07 0.02 0.0075 1.30 Pig iron 94 4.35 0.1 0.035 – – – – 0.0045 1.515 Tapped steel 99.6 0.03 0.013 0.015 0.058 0.026 0.018 0.003 0.0050 0.237

Table 4. Average and standard deviation of the slag composition with no hot metal and hot metal utilization of approximately 36%.

T.Fe SiO2 CaO Al2O3 MgO MnO Component Avg. Stdev. Avg. Stdev. Avg. Stdev. Avg. Stdev. Avg. Stdev. Avg. Stdev. All scrap 40 8.5 10.88 2.86 25.23 5.51 2.9 0.88 6.28 2.32 3.8 0.83 36% hot metal 35.8 7.35 9.54 2.38 30.43 5.15 2.88 1.04 7.2 4.03 3.72 0.83 metal. The chemical analysis of the hot metal, scrap, and steel samples was analyzed by an OES (optical emission spectrometer; ARL4460, Thermo-scientific, FL, USA)8) and the carbon and sulfur through the LECO C/S analyzer (CS600, LECO, MI, USA). Nitrogen analysis of steel samples was carried out using the N/O analyzer (TC500, LECO, MI, USA). The soluble oxygen and the temperature within the steel melt was measured using the combined oxygen and temperature sensor of Heraeus Electro-Nite, which is based on a stabilized zirconia electrolyte cell and an R-type thermocouple, respectively. Details of the working mechanism of the oxygen sensor can be found in past literature.9)

2.3. Slag and Inclusion Chemistry with Hot Metal Additions Fig. 2. Slag compositional changes with hot metal utilization in the From the slagmaking operation, a typical CaO–SiO2–FeO CaO–SiO2–FeO ternary phase diagram adapted from slag system was formed within a CaO/SiO2 averaging 2.4±0.56, atlas. when scrap alone was used, and a CaO/SiO2 averaging 3.3±0.69, when hot metal of approximately 36% was used. there is significant sensible heat within the hot metal.13–15) The slag composition was verified using an XRF. The Possible reasons for the scatter of the data is speculated to approximate chemical composition of the slag is provided in be caused by the difference in hot metal chemistry and also Table 4. Figure 2 shows the changes in the slag composi- the in-coming hot metal temperature. The chemical compo- tion from its initial scrap melt down composition and oxy- sition of the steel likely changes the thermal conductivity gen blowing projected onto the CaO–SiO2–FeO ternary and heat capacity of the melts, which have been documented phase diagram obtained from slag atlas, when scrap and hot elsewhere.16–18) The in-coming hot metal temperatures are metal is used in the charge of the EAF. The initial melt down between 1 554–1 637 K, as shown in Table 2. In addition, slag composition is within region ‘A’ and as the FeO content unexpected process issues such as breakouts at the continu- increases with oxygen lancing the CaO additions dissolve ous caster, cold ladles, or prolonged LMF (ladle metallurgical into the molten slag resulting in a composition of region ‘B’, furnace) processing can result in extended power-on-times when only scrap is used. With hot metal utilization, the car- that can exacerbate plant data scatter. bon content in the steel is significantly higher and the slag During the hot metal utilization in the EAF, the power on contains less FeO during the oxygen lancing period and time is significantly decreased, but the decarburization of more efficient dissolution of CaO, which rapidly increases the steel increases and thus the oxygen lancing time deter- the basicity (CaO/SiO2) of the slag towards region ‘C’. mines the total steelmaking time of operation. Furthermore, Inclusions from selected steel samples taken from the the hot metal additions allows rapid formation of the steel melt tundish were analyzed by the slime extraction method.10,11) and at the final stages of boring approximately 4 000 Nm3/h The resulting oxide inclusions were then analyzed using of oxygen is lanced for 8 min. and depending on the foam- EPMA for elemental Ca and Al to determine the calcium- ing behavior, the oxygen flowrate is slowly increased. When aluminate inclusions present with use of hot metal. the melt is sufficiently formed and slag begins to foam, the oxygen lance is submerged deeper into the slag layer and the flowrate is increased to 6 000 Nm3/h for 5 min. This allows 3. Results and Discussions an optimal foaming condition of the slag and the oxygen 3.1. Effect of Process Parameters with HMR (Hot flowrate is again increased to 8 000 Nm3/h for 8 min. fol- Metal Ratio) lowed by an oxygen increment to 9 000 Nm3/h for the next From over 200 heats of hot metal additions from 20 to 5 min. Near the target decarburization content, slag foaming 36%, the power-on-time (min/ch) in Fig. 3(a) shows a is significantly decreased and the oxygen flowrate is decrease with a slope of 0.34 min/%HM assuming a linear decreased to 5 000 Nm3/h. The soluble oxygen content at the regression of the data. Power on time is the total amount of final stages of decarburization is targeted at 400 ppm oxy- time within the tap-to-tap time, where power through the gen. The oxygen flowrate pattern is illustrated in Fig. 3(b). transformer is provided and excludes the loading, prepara- Typical EAF operations have approximately a 10 min. tion and tapping times of the EAF.12) refining stage and must operate in an excess oxygen opera- This decrease in the consumed power is expected since tion for increased kinetics. This results in more oxygen dis-

Fig. 3. Effect of hot metal additions on the (a) power on time in the EAF and (b) oxygen flowrate (Nm3/hr).

Fig. 4. (a) Dissolved oxygen content as a function of carbon content and (b) decarburization rate with hot metal additions into the EAF specified as input carbon. solution in the steel compared to the C/CO equilibrium of towards the equilibrium line much faster than scrap alone. reaction (1). The 1 mass% standard state was taken for the And as long as excessively high amount of oxygen, much calculation of the activities of the dissolved carbon and oxy- higher than the necessary input carbon, is avoided the dis- gen in iron. The interaction parameters used for reaction (1) solved oxygen content will be closer to the equilibrium val- are the calculated values derived from the experimental ue. It should be mentioned that the number of data utilized results of Banya et al.19) in Fig. 4(a) and subsequent figures do not correspond com- @1873 K pletely to Fig. 3(a). This is due to the availability of plant data and in some cases with defective sensor measurements 1160 15)...... (1) COCO+= log K = +2 . 003 or sampling errors during operation. All available data per- T taining to the present study have been utilized for analysis. The above thermodynamic calculation is carried out at The decarburization rate with the input carbon in the EAF 1 873 K for simplification considering the average tap tem- charge is shown in Fig. 4(b). Higher input carbon required perature at the EAF was close to 1 873 K. higher decarburization rates for a final carbon target of Figure 4(a) shows the dissolved carbon and oxygen 300 ppm. before tap for heats with and without hot metal. The dis- Figure 5 shows the soluble oxygen content of the steel solved oxygen for the heats with 100% scrap is typically melt and T.Fe in the slag at tap. Hot metal additions into the higher than the equilibrium value, but the heats utilizing hot furnace allowed lower turndown FeO in the slag compared metal is slightly closer to the equilibrium value than heats to the hot metal free heats. The additional carbon available with all scrap. It would initially seem that the utilization of in the melt provided better FeO/C reactions at the slag/metal hot metal would require higher decarburization rates com- interface resulting in lower FeO in the slag. This is also evi- pared to all scrap heats in order to maintain a constant steel- dent from the vicinity of the T.Fe results of the hot metal making process time, which would require higher oxygen with the equilibrium line compared to the heats without hot flowrates and thus possibly increased oxygen dissolution in metal utilization. The T.Fe (%) and O (ppm) equilibrium the steel. However, if optimized for the input carbon content curve at 1 873 K was calculated from the Fe/FeO equilibri- from the hot metal, this increased decarburization of the um using the thermodynamic data provided in reference 19) melt will increase the formation of CO and CO2 gases pro- databook The oxygen activity in liquid Fe equilibrated viding better slag foaming and increased mixing in the fur- with FeO in slag was calculated from the recommended nace resulting in higher reaction kinetics and thus moves interaction parameters and the activity of FeO in slag was

Fig. 6. Phosphorous content at the ladle metallurgical furnace with phosphorous input into the EAF. Fig. 5. The T.Fe in slag as a function of the soluble oxygen in the steel for hot metal free and hot metal added heats. estimated by assuming an ideal liquid solution. The T.Fe in slag was recalculated by dividing the mass percent of FeO with 1.286. The mathematical formulation is beyond the scope of the present study. A more complex equation derived by Basu et al.20) for the activity of FeO in slag is available, but is not significantly different from the calculation of the activity assuming an ideal slag solution in the range of FeO in the present study.

3.2. Effect of HMR on the Phosphorous Content By utilizing hot metal, the time to obtain the molten steel pool is decreased and allows the operator much longer steel refining times before tap, while maintaining the past tap-to- tap times. Depending on the oxygen partial pressure, phosphorous removal from steels through the slag-metal inter- Fig. 7. Effect of T.Fe in slag on the dephosphorization distribution 3− ratio (LP) in the EAF. face can occur by either phosphate (PO4 ) at high O2 or 3– phosphide (P ) at low O2, as expressed in reaction (2) and 21–23) (3). Considering the oxygen lancing during decarbur- the hot metal and the scrap feed and P@LMF is the measured ization and the amount of FeO present in the slag, the phos- P content at the LMF (ladle metallurgical furnace). Figure phate is likely dominant. It should also be mentioned that for 6 shows the De-P% to be on average 50% for 100% scrap most steel decarburization processes at 1 873 K result in an and with 36% HM utilization the De-P% to be on average oxygen partial pressure of higher than 10–16 atm and thus 75%. phosphorous removal by phosphate is dominant. The HM utilization into the EAF can also effect the T.Fe in the slag due to the higher carbon content and its subse- 5 3 2−−3 PO++2 O =() PO4 ...... (2) quent reaction with the FeO, as previously mentioned. 4 2 According to Fig. 5, the use of HM allows better control of soluble oxygen dissolved in the steel and is closer to the 3 23−−3 PO+=+() P O2 ...... (3) oxygen content achieved from the C/CO equilibrium at 2 4 steelmaking temperatures resulting in lower T.Fe in the slag. With 36% HM input into the EAF, the initial input This lower T.Fe and subsequently FeO content and amount of P increases by 0.02%, but due to longer refining improved reaction efficiency towards the thermodynamic times and better CaO dissolution kinetics into the slag, the equilibrium value increases the P distribution of the slag 24) distribution ratio of P between the metal and slag (LP) comparable to the results of Balajiva and Vajragupta. increases with HM additions, which result in comparable P Figure 7 describes the effect of T.Fe in the slag on the phos- contents in the final tapped steel with 100% scrap feed EAF phorous distribution plotted together with the results of processes. If the de-phosphorization percentage (De-P%) is Balajiva and Vajragupta.24) Thus, lower FeO in the slag with defined as Eq. (4). hot metal utilization likely improves the phosphorous removal similar to the results obtained from past litera- 23,25) ()PPInput− @ LMF ture. Furthermore, comparison of the phosphorous De−= P% ×100 ...... (4) removal in the present study at basicity of 3.3 with Balajiva PInput and Vajragupta24) at basicity of 2.4 shows the basicity in the where PInput is the overall total P input into the EAF from present slag compositional range to be insignificant for

495 © 2015 ISIJ ISIJ International, Vol. 55 (2015), No. 3 phosphorous, which has also been suggested by others.23,25) rich elements interact with a slag composition with compar- atively high FeO, MnO, SiO2, and low basicity, which have 3.3. Effect of HMR on the Sulfur Content lower desulfurization potential.28) Figure 8 shows the effect of desulfurization with HM of 36% utilization for approximately 200 heats analyzed at tap. 3.4. Effect of HMR on the Nitrogen Content The removal of sulfur at the slag-metal interface can be Although typical EAF processes utilize a top cover and described by two mechanisms of sulfide (S2–) and sulphate apparently sealed slag door during steelmaking, its effect to 2− 23,26,27) (SO4 ), as expressed in reactions (5) and (6). From seal atmospheric gases from entering the EAF is not abso- the CaSO4–CaS equilibrium relationship, the critical oxygen lute. Thus, significant infiltration of nitrogen through the partial pressure can be estimated to be approximately 10–5 cover and crevices of the slag exit door is possible. As the atm assuming CaO flux is the dominant desulfurizer in the arc is formed between the electrode and the scrap, the nitro- slag. At the oxygen partial pressure of the primary steelmak- gen gas is ionized and rapidly travels towards the surface of ing process, the dominant desulfurization mechanism is well the molten pool, which can dissolve into the steel increasing known to be the sulfide reaction of (5) and the sulfate reac- the nitrogen content.29) A schematic of the typical nitrogen tion of (6) to be negligible, but the low basicity of the pres- content variations in steels during the EAF operating stages ent slag seems to counteract the oxygen partial pressure are provided in Fig. 9.30) During the 1st melting period in effects and inhibit significant removal of sulfur in the metal. region B, where the small molten steel pool forms, nitrogen Furthermore, the lower carbon after decarburization results significantly increases above 170 ppm N. As the final in reduced activity of the sulfur through the changes in the charge is melted and a slag layer is formed, N pickup is interaction coefficient of sulfur, which decreases the driving inhibited. Heating to the steelmaking temperatures with a force for desulfurization. fully covered slag protects the steel from atmospheric nitrogen pickup. When region E is reached, significant boiling 22−−1 SO+=+ S O2 ...... (5) and foaming of the slag occurs, which drives the nitrogen 2 out of the steel and also significantly lowers the partial pressure of N2, resulting in significantly lower N. During tap- 3 2−−2 SOO++=2 () SO4 ...... (6) ping, nitrogen pickup typically occurs through the stream of 2 the steel. According to the results, significant correlation of the sul- The nitrogen content in steels is a function of tempera- fur with HM additions could not be observed. It is speculat- ture, steel composition and nitrogen partial pressure accord- ed that desulfurization in the EAF is not efficient due to the ing to Eq. (7). counteracting effects of basicity and oxygen partial pressure 1 050 pressures during hot metal utilization. Although faster dis- log[]N =− −0..%. 815 − 0 125 ×[]C − 0 065 solution of CaO in the slag is observed for hot metal utili- T ... (7) zation, the high oxygen flowrates counteract the effects of × %.%Si+×002 Mn + 05 .l oog ⎡P ⎤ [] [] ⎣ N2 ⎦ the basicity. Thus, hot metal utilization of the present slag system does not promote the removal of sulfur at the final Equation (7) was derived from the equilibrium reaction target carbon content of less than 0.03%. between gaseous nitrogen (N2) and dissolved N in liquid Fe. For some of the heats, excessive slag carryover from the The activity of the dissolved nitrogen was calculated using EAF during tapping and un-optimized CaO dissolution into the 1 mass% standard state. The recommended interaction the slag can cause resulfurization at the LMF. This reversion parameters from the steelmaking data sourcebook19) with of sulfur from the slag to the steel can take place if the S- respect to the steel composition of C, Si, and Mn for calculation of the activity coefficient have been utilized.

Fig. 8. Comparison of the sulfur content at the ladle metallurgical furnace with sulfur input into the EAF with and without hot metal. Note the resulfurization of the elements in some of Fig. 9. Schematic of the nitrogen content of the steel during the the heats. operating practice of the EAF adapted from Pilliod.30)

Thus to effectively lower the nitrogen in steels, lower metal is likely more affected by the evolution of CO gases temperatures, active CO boiling and subsequent decrease in during decarburization and its subsequent lowering of the the nitrogen partial pressure, and an increase in the slag nitrogen partial pressure from comparison between the cold foaming to decrease the possible infiltration into the bath pig iron and hot metal heats. Furthermore, the addition of melt should be pursued. Figure 10 shows the comparison of carbon containing iron units into the steel melt provides a the effect of N in steels between hot metal and cold pig iron more rapid slag formation and foaming practice that can unit additions into the EAF. A baseline N content for typical inhibit the atmospheric infiltration into the melt and also EAF operations would be between 80 to 130 ppm N during provide a more steady arcing stream, which can lower the tap as can be verified from the plant trials with no carbon nitrogen dissolution into the melt. The negligible effect in saturated iron units within the furnace. For region 1, the region 1 is speculated to be due to the counteracting effect additions of carbon saturated cold pig iron into the furnace of the improved CO evolution and the increased power con- did not significantly lower the nitrogen content up to sumption of the slow melting of the pig iron units. approximately 22%. Although the higher C content in the cold pig iron was expected to improve CO gas formation 3.5. Effect of HMR on the Inclusion Chemistry and subsequent lowering of the N content in the steel, results For HMR of 36% utilization, the steel is affected in three showed that the nitrogen content did not deviate much from aspects of internal quality. First, the input material contains the 100% scrap operation at low amounts of additions. This significant amounts of carbon forming large amounts of CO was speculated to be the increased power input in Region 1 during the melting and refining stages of the steelmaking of the dense cold pig iron units compared to 100% scrap process, which prevent nitrogen infiltration into the furnace operation due to its higher density and slower melting rate and also improved foamy slags inhibiting atmospheric gas during operation. Increased additions of carbon saturated penetration into the steels. This result in about 1/3 of soluble containing iron units showed comparable lowering effects of nitrogen in the hot metal utilized steels compared to hot nitrogen content due to the significantly higher CO evolu- metal free heats and possibly decreased nitride inclusions. tion and subsequent slag foaming that would suppress the Second, sufficient carbon in the steel at the end of the melt infiltration and dissolution of nitrogen into the steel. This down cycle of the EAF allow lower free oxygen in the steels critical ratio of 22% carbon saturated iron unit additions was after the refining stages lowering the oxygen content by obtained from the intercept of the average nitrogen value about 100–150 ppm. Third, the hot metal dilutes the amount with no carbon saturated iron units at 105 ppm N extended of tramp elements such as Cu, Ni, Sn within the scrap,13) towards the addition of carbon saturated iron units and the which significantly lowered visible crack defects of 2 mm linear regression line of data points with carbon saturated and longer (defect ratio %) by 50% on the hot coil during iron units. As the amount of carbon saturated iron units casting and rolling. exceeds 22% of the charge, the overall trend seems to indi- Due to the mini-mill operation of the present study, where cate that the addition of carbon containing iron units into the a thin slab caster and continuous hot strip mill process is EAF decreases the nitrogen content in the tapped steel. In adopted, the process characteristic requires significant con- particular, the addition of hot metal slightly decreases the N trol of inclusions for production of highly value-added steel content compared to the cold pig iron unit additions due to products. Figure 11 shows the quality comparison of iden- the decrease in the power on time of the EAF and subse- tical steel products produced through 100% scrap and 36% quent decrease in the nitrogen ionization through the plasma HM utilization in the DC-EAF with comparable secondary arc. In particular the region B and C in Fig. 9 would signif- metallurgy practices. The defect index is typically obtained icantly decrease with hot metal additions. However, it from the visual inspection results of the surface of the prod- should also be noted that the nitrogen dissolution into the uct, which is determined from the length, depth, and number

Fig. 10. Effect on the nitrogen content by additions of carbon saturated iron units including cold pig iron and hot metal at Fig. 11. Changes in the defect ratio index with 36% hot metal utili- steel tap from the EAF. zation in the EAF.

497 © 2015 ISIJ ISIJ International, Vol. 55 (2015), No. 3 of the defect. The combination of these variables determines range of acceptable steel chemistry, it may have provided the defect index (%), which is a comparison to the reference beneficial effects to the steel. It has been well-defined in value set-up for each individual steel grade from ultra-low past publications32,33) that Ni content approximately half of carbon to high carbon steels. The details of the reference the copper can inhibit surface defects in steels. Shibata et value and the equation are proprietary information to the al.34) has also shown that higher Si content and P up to specific company, but have become more stringent as cus- 200 ppm can lessen the impact of the surface hot shortness tomer demands become more critical. For example, the ref- and crack susceptibility of the steel sample. Thus, various erence value for a crack sensitive peritectic grade may be steel chemistries may have assisted with lowering the defect 10 cracks with 4 mm length by 2 mm depth, but a low car- index, although the Cu+10Sn exceeded 0.1. According to bon grade may be 5 cracks with 3 mm length by 2 mm depth the quality results, a 50% reduction in the average quality within the plant. Typically, the total content of (Cu+10Sn) defect index was observed, when hot metal is used as pre- is an essential control criterion for product quality. Increased viously mentioned. amounts of Cu and Sn in the steel tend to result in signifi- The dominant inclusion chemistry from analysis of the cant downstream quality issues such as surface hot shortness total Ca and Al content in the steel was identified to be clos- 31,32) cracks and are constrained below a total of 0.1 er to the C3A (3CaO·Al2O3) with the inclusions using hot (Cu+10Sn). With hot metal additions of 36%, the tramp ele- metal or scrap alone as shown in Fig. 12. Using the thermo- ments of Cu and Sn is correspondingly diluted and ensures dynamic data and relationships in reaction (8) and (9) from the critical defect index to be less than 10%. The critical Turkdogan,35) the equilibrium line between Al and Ca for defect index is the limit, where quality codes are changed the calcium-aluminate type inclusions expected of forming on the product of the steel. Below the critical defect index at steelmaking temperatures of 1 823 K could be ascertained of 10%, products are not downgraded and above this critical and plotted in Fig. 12. At low solute concentrations of Al value downgrades of the products are typically required. For and Ca in the steel, the activities of Ca and Al were taken heats that have been added with hot metal of 36%, the equivalent to the mass% of Ca and Al, respectively. Al2O3, Cu+10Sn value is less than 0.1% resulting in critical defect CaAl2O4, and Ca3Al2O6 were considered to be in the pure index values lower than 10%. For typical all scrap heats standard states. with no hot metal, the Cu+10Sn values exceed 0.1 and some 4 2 of the inspections of products resulted in defect index above Ca+= Al23 O CaAl 24 O + Al 3 3 ...... (8) the critical defect index of 10% resulting in product down- o grades and subsequent loss in quality. The critical defect ΔGTJmol=−×57 354 85.(/) 3 index of 10% at Cu+10Sn is an empirical value that has 32Ca+= Al O Ca Al O +2 Al been developed within the plant, which has been found to 23 3 26 ...... (9) o be free of steel product downgrades. Not all identical steel ΔGTJmol=−×207 471..(/) 76 229 52 grades in different heats have exactly the same steel chem- From the results of the inclusion analysis, the impact of istry since there is a range of chemical compositions accept- hot metal on the effect of inclusion chemistry does not seem able for qualification. The direct cause of the steel samples to have significant impact as the inclusion modification exceeding Cu+10Sn of 0.1 for all scrap heats, but lower than treatment has been appropriately optimized for the dissolved the critical defect index of 10% has yet to be fully under- oxygen content and Al in the steels before the steel is sent stood. It may be speculated that other elements in the steel to the tundish for casting. The inclusion composition seems such as Ni could have provided increased Cu-solubility in to indicate slightly excess Ca treatment for the steels, but is the austenite to lower the propensity for hot cracks to form. comparable for both the hot metal free and 36% hot metal And if the Ni content was relatively high, but within the added heats. The typical 3CaO·Al2O3 inclusion morphology

Fig. 12. Inclusion chemistry analysis of tundish samples after the LMF with hot metal utilization of 36% and no hot metal utilization in the EAF. Note the EPMA analysis of Ca and Al on the surface of separated inclusions using the slime method is shown.

© 2015 ISIJ 498 ISIJ International, Vol. 55 (2015), No. 3 observed after treating with the slime method is shown adja- REFERENCES cent to Fig. 12 labeled C3A. Although not thermodynami- cally stable for the mass% of Ca and Al in the steel sample, 1) J. P. Birat: Rev. Métall., 97 (2000), 1347. 2) I. Y. Zinurov, A. M. Shumakov, S. G. Ovchinnikov, S. N. Pigin and Al2O3 inclusions are also observed within the melt labeled Yu. F. Mamenko: Stal, 7 (2009), 35. A. It is likely that some Al2O3 inclusions could not effec- 3) J.-P. Duan, Y.-L. Zhang and X.-M. Yang: Int. J. Min. Metall. Mater., 16 (2009), 375. tively interact with the Ca injected at the LMF for the given 4) P. Argenta and M. B. Ferri: La Metallurgia Italiana, 97 (2005), 41. processing time and fluid flow conditions during secondary 5) T. P. Wandekoken, B. T. Maia, P. Hopperdizel and J. R. de Oliveira: refining and thus remained as Al O in the melt. This 43rd Seminario de Aciaria – Internacional Belo Horizonte Brazil, 2 3 Brazilian Metallurgical, Materials and Mining Association Brazil, suggests that the reactions involved during inclusion modi- Brazil, (2012), 75. fication in secondary refining do not reach complete ther- 6) H. Liu, J. Li and C. Ma: Adv. Mater. Res., 572 (2012), 243. 7) R. Gottardi, S. Miani, A. Partyka and M. Suber: Iron Steel Technol., modynamic equilibrium in actual operations. 9 (2012), 61. 8) J.-Y. Park, S.-M. Jung and I. Sohn: Metall. Mater. Trans. B, 45 (2014), 329. 4. Conclusions 9) J. W. Fergus: http://www.tms.org/pubs/journals/jom/0010/fergus/ fergus-0010.html, (accessed 2014-10-10). In this study, the effect of hot metal additions into a DC 10) J. Wakita and S. Xizoguchi: Tetsu-to-Hagané, 66 (1980), 5862. 11) H. Hoef, H. Lessing and G. Masing: Stahl Eisen, 76 (1956), 1442. EAF was investigated. The dissolved oxygen for heats with 12) World Steel Association: http://www.steeluniversity.org/content/html/ 100% scrap was typically higher than the equilibrium value, eng/default.asp? catid=25&pageid=2081271946, (accessed 2014-08- 20). but the heats utilizing hot metal was slightly closer to the 13) C. P. Manning and R. J. Fruehan: JOM, 53 (2001), 20. equilibrium value. Hot metal additions into the furnace 14) A. Orth, N. Anastesijevic and H. Eichberger: Miner. Eng., 20 (2007), allowed lower turndown FeO in the slag compared to the hot 854. 15) T. Jiemin, M. B. Ferri and P. Argenta: Ironmaking Steelmaking, 32 metal free heats. With 36% hot metal input into the EAF, (2005), 191. although initial P was higher than the 100% scrap heats, hot 16) T. El Gammal and X. Li: Mater. Sci. Forum, 73–75 (1991), 37. 17) M. Susa, M. Watanabe, S. Ozawa and R. Endo: Ironmaking metal utilization resulted in longer refining times with better Steelmaking, 34 (2007), 124. CaO dissolution kinetics into the slag increasing the distri- 18) M. J. Peet, H. S. Hasan and H. K. D. H. Bhadeshia: Int. J. Heat Mass bution ratio of P between metal and slag, which attained Transf., 54 (2011), 2602. 19) The Japanese Society for the Promotion of Science: Steelmaking comparable P contents in the final tapped steel. Hot metal Data Sourcebook Rev. Ed., Gordon and Breach Science Pub., New utilization for the basicity and oxygen partial pressure of the York, NY, (1988), 59. 20) S. Basu, A. K. Lahiri and S. Seetharaman: Metall. Mater. Trans. B, present slag system did not particularly promote sulfur 39B (2008), 447. removal compared to the 100% scrap heat. The addition of 21) H. Suito and R. Inoue: ISIJ Int., 35 (1995), 258. 22) X. Guo, H. Matsuura, M. Miyata and F. Tsukihashi: ISIJ Int., 53 carbon saturated iron units up to 22% into the furnace did (2013), 1381. not significantly lower the nitrogen content, but as the 23) N. Sano, W.-K. Lu, P. V. Riboud and M. Maeda: Advanced Physical amount exceeded 22% of the charge, the overall trend seems Chemistry for Process Metallurgy, Academic Press, San Diego, CA, (1997), 51. to indicate the nitrogen content in the tapped steel to 24) K. Balajiva and P. Vajragupta: J. Iron Steel Inst., 155 (1947), 563. decrease. In particular, hot metal additions slightly decreas- 25) H. Kishitaka: The 43rd Nishiyama Memorial Seminar, ISIJ, Tokyo, (1977), 50. es the N content compared to the cold pig iron unit additions 26) Y. Nakai, N. Kikuchi, Y. Miki, Y. Kishimoto, T. Isawa and T. due to the decrease in the power on time of the EAF. With Kawashima: ISIJ Int., 53 (2013), 1020. hot metal additions of 36%, the tramp elements of Cu and 27) P. Yan, X. Guo, S. Huang, J. Van Dyck, M. Guo and B. Blanpain: ISIJ Int., 53 (2013), 459. Sn is correspondingly diluted and ensured the critical defect 28) J. Lehmann and M. Nadif: Rev. Mineral. Geochem., 73 (2011), 493. index to be less than 10% resulting in 50% reduction of the 29) B. Bowman and K. Kruger: Arc Furnace Physics, StahlEisen Com- munications, Düsseldorf, (2009), 5. quality defect index. The existing inclusion chemistry of the 30) C. F. Pilliod: 46th Electric Furnace Conf. Proc., The Iron and Steel heats was identified to be closer to the C3A (3CaO·Al2O3), Society, Pittsburgh, (1988), 107. but was comparable for both the hot metal free and 36% hot 31) A. Takemura, Y. Ugawa, K. Kunishige, Y. Tanaka, S. Hashimoto and K. Ootsuka: Mater. Trans., 52 (2011), 1905. metal added heats. 32) N. Imai, N. Komatsubara and K. Kunishige: ISIJ Int., 37 (1997), 224. 33) ASM International: ASM Metals Hand Book, 10th ed., Properties and Selection: Iron; Steels and High Performance Alloys, Vol. 1, Materials Acknowledgements Park, OH, (1995), 389. This study was partially supported by the Brain Korea 21 34) K. Shibata, S.-J. Seo, M. Kaga, H. Uchino, A. Sasanuma, K. Asakura and C. Nagasaki: Mater. Trans., 43 (2002), 292. PLUS (BK21 PLUS) Project at the Division of the Eco- 35) E. T. Turkdogan: Physical Chemistry of High Temperature Technol- Humantronics Information Materials and the Ministry of ogy, Academic Press, New York, (1980), 5. Trade, Industry, and Energy project No. 2013-11-2008.