Impacts of Modification of Alloying Method on Inclusion Evolution In

materials

Article Impacts of Modiﬁcation of Alloying Method on Inclusion Evolution in RH Reﬁning of Silicon Steel

Fangjie Li 1 ID , Huigai Li 1,*, Shaobo Zheng 1, Jinglin You 1, Ke Han 2 and Qijie Zhai 1 1 State Key Laboratory of Advanced Special Steel, Shanghai Key Laboratory of Advanced Ferrometallurgy, School of Materials Science and Engineering, Shanghai University, Shanghai 200072, China; [email protected] (F.L.); [email protected] (S.Z.); [email protected] (J.Y.); [email protected] (Q.Z.) 2 National High Magnetic Field Laboratory, Florida State University, Tallahassee, FL 32310, USA; [email protected] * Correspondence: [email protected]; Tel.: +86-021-5633-4045

Received: 28 September 2017; Accepted: 16 October 2017; Published: 19 October 2017

Abstract: This study explores the effect of introducing additional alloy elements not only in a different order but also at different stages of the Ruhrstahl-Heraeus (RH) process of low-carbon silicon steel production. A more economical method, described as “pre-alloying”, has been introduced. The evolution of MnO-FeO inclusions produced by pre-alloying was investigated. Results show that spherical 3FeO·MnO inclusions form ﬁrst, then shelled FeO·zMnO (z = 0.7–4) inclusions nucleate on the surface of pre-existing 3FeO·MnO. Spherical FeO·zMnO (z = 3–5) is further evolved from shelled 3FeO·MnO by diffusion. Because these MnO-FeO inclusions ﬂoat up into the slag before degassing, the pre-alloying process does not affect the quality of the melt in the end. Both carbon content and inclusion size conform to industry standards.

Keywords: silicon steel; RH reﬁning; inclusions; evolution

1. Introduction Silicon steel is a soft magnetic material that is widely used in the manufacture of electrical machinery, motors, and transformers [1]. Production of this low-carbon steel requires careful control of the oxide inclusions that occur during the steelmaking process [2–4]. One commonly used technique for refining low-carbon steels is the Ruhrstahl-Heraeus (RH) process [5–7]. The RH refining process has three stages: (1) decarburization, (2) deoxidization and alloying, and (3) degassing. During the first stage, oxygen is blasted into the melt to keep the carbon level low. The second stage has two aims: (1) to bring the oxygen level down before the processing continues, and (2) to ensure the proper combination of the additional elements introduced to form a new alloy. The third stage also has two aims: (1) to remove the hydrogen and nitrogen that had been naturally absorbed from the air during treatment prior to the RH process, and (2) to allow inclusions to float up into the slag. During deoxidization and alloying, the second stage, aluminum (Al) and ferrosilicon (FeSi) are frequently used as deoxidizers. When these alloys are added separately, they produce either Al2O3 or SiO2 inclusions. Zhen et al. pointed out that slabs produced by Al deoxidization are inferior in quality to those produced by Si deoxidization [8]. This may be in part because SiO2 inclusions (produced by Si deoxidization) are less likely to cluster than Al2O3 inclusions (produced by Al deoxidization) [9]. When FeSi is added first, the resulting SiO2 inclusions combine with any Al that is subsequently added, but the reverse does not occur. The behavior of inclusions in melts varies, depending on the sequence of alloy addition [10]. After deoxidization, conventional alloying occurs, not only with Si and Al but also with Mn. As an alloying agent, low-carbon ferromanganese (FeMn) is as effective as pure Mn but much more economical, so it is often used in this stage. In the current study, medium-carbon FeMn, which is even

Materials 2017, 10, 1206; doi:10.3390/ma10101206 www.mdpi.com/journal/materials Materials 2017, 10, 1206 2 of 8 less expensive than low-carbon FeMn, was introduced into the melt in the first stage, that is, before the beginningMaterials of2016 decarburization., 9, 1206 This fabrication method is referred to in this study as “pre-alloying”.2 of 8 The disadvantage of pre-alloying is that Mn oxide inclusions begin immediately to form and build up in theless melt. expensive The countervailing than low-carbon advantage, FeMn, was however, introduced is that into the the Mn melt that in the remains first stage, after that oxides is, before are formed is availablethe beginning during of the decarburization. second stage to This reduce fabrication the consumption method is ofreferred costlier to low-carbon in this study FeMn. as “pre- Previous alloying”. The disadvantage of pre-alloying is that Mn oxide inclusions begin immediately to form studies have analyzed the manganese balance between slag and steel in the pre-alloying process and build up in the melt. The countervailing advantage, however, is that the Mn that remains after from both thermodynamics and dynamics [11–13]. The yielding rate of FeMn and its influencing oxides are formed is available during the second stage to reduce the consumption of costlier low- factorscarbon has alsoFeMn. been Previous investigated studies have [14]. analyzed Nevertheless, the manganese the evolution balance of between inclusions slag inand silicon steel in steel the after introducingpre-alloying the pre-alloyingprocess from procedureboth thermodynamics has rarely beenand dynamics studied. One[11–13]. of the The objectives yielding rate of this of FeMn study was to investigateand its influencing the growth factors and has behavior also been of theinvestigat Mn oxideed [14]. inclusions Nevertheless, produced the evolution by the early of inclusions introduction of medium-carbonin silicon steel after FeMn. introducing the pre-alloying procedure has rarely been studied. One of the objectivesThe third of stage this ofstudy the RHwas processto investigate is a procedure the growth named and behavior “degassing”. of the Mn Degassing oxide inclusions occurs when argonproduced is blasted by intothe early the melt, introduction causing of inclusions medium-carbon to float FeMn. up into the slag. This process is important for removingThe the third silicon stage oxide of the and RH aluminum process is a oxide procedure inclusions named that“degassing”. form during Degassing the deoxidization occurs when and argon is blasted into the melt, causing inclusions to float up into the slag. This process is important alloying stage [15]. Zhang et al. claimed that the smaller inclusions collide with one another to form for removing the silicon oxide and aluminum oxide inclusions that form during the deoxidization larger inclusions, which then float up to the slag where they are absorbed and thus removed from the and alloying stage [15]. Zhang et al. claimed that the smaller inclusions collide with one another to meltform [16]. larger inclusions, which then float up to the slag where they are absorbed and thus removed fromThe aimthe melt of this [16]. work is to investigate the impacts of modification of the alloying method on the inclusionThe evolution aim of this in RH work refining is to investigate of silicon the steel, impact withs of an modification emphasis on of the the compositionalloying method and on size the of the inclusions.inclusion The evolution carbon in content RH refining induced of silicon by medium-carbon steel, with an emphasis FeMn additionon the composition in pre-alloying and size was of also probedthe toinclusions. make sure The it carbon met the content industry induced expectation. by medium-carbon FeMn addition in pre-alloying was also probed to make sure it met the industry expectation. 2. Experimental Procedure 2. Experimental Procedure RH refining was processed in a 160-t ladle with MgO-C-type lining refractory. The molten steel RH refining was processed in a 160-t ladle with MgO-C-type lining refractory. The molten steel in the ladle undergoes the processes of decarburization, deoxidization alloying, and degassing to in the ladle undergoes the processes of decarburization, deoxidization alloying, and degassing to obtain a semi-finished product. Samples were taken by drawing 3 kg of steel melt from the industrial obtain a semi-finished product. Samples were taken by drawing 3 kg of steel melt from the industrial ladleladle at four at four key key states states of of RH RH refining. refining. The The firstfirst drawing occurred occurred before before the the start start of decarburization of decarburization but afterbut after medium-carbon medium-carbon FeMn FeMn alloy alloy hadhad beenbeen added; the the second second after after decarburization decarburization had been had been completed;completed; the thirdthe third after after FeSi FeSi and and Al Al had had been been introduced introduced for for deoxidization/alloying, deoxidization/alloying, followed followed by by the additionthe addition of low-carbon of low-carbon FeMn for FeMn further for further alloying; alloying; and the and fourth the afterfourth degassing after degassing had been had completedbeen (See Figurecompleted1 ). The(See fourFigure samples 1). The four were samples investigated were in tovestigated clarify the to formationclarify the formation and evolution and evolution of inclusions, not onlyof inclusions, in the pre-alloying not only in the process pre-alloying but also process in the but deoxidization also in the deoxidization and alloying and process. alloying process.

FigureFigure 1. Illustration 1. Illustration of of sampling sampling duringduring the the RH RH (Ruhrstahl-Heraeus) (Ruhrstahl-Heraeus) refining reﬁning process. process.

Each drawing was cooled down to room temperature and allowed to solidify into a cylindrical Each drawing was cooled down to room temperature and allowed to solidify into a cylindrical ingot. From each ingot, a sample was cut in the form of a thin sheet (50 × 40 × 2 mm3), taken from the ingot. From each ingot, a sample was cut in the form of a thin sheet (50 × 40 × 2 mm3), taken from

Materials 2017, 10, 1206 3 of 8 Materials 2016, 9, 1206 3 of 8 thelongitudinal longitudinal plane plane at the at thehalfway halfway point point of the of radius the radius (see Figure (see Figure 2a). Each2a). of Each these of four these sample four sheets sample sheets(S1, (S1,S2, S3, S2, S4) S3, S4)was was then then electrolyzed electrolyzed in a in non- a non-aqueousaqueous solution solution composed composed of ofmethanol, methanol, maleic maleic anhydride,anhydride, and and tetramethyl tetramethyl ammonium ammonium chloridechloride [[17].17]. After After 6 6 hours hours of of electrolysis, electrolysis, a good a good portion portion of of thethe steel steel matrix matrix had had dissolved dissolved into into the the non-aqueousnon-aqueous solution. When When this this solu solutiontion was was passed passed through through a polycarbonatea polycarbonate membrane membrane filmfilm filterfilter (open(open pore size 3 3 μµm),m), the the oxide oxide inclusions inclusions from from the the sample sample sheetsheet were were deposited deposited on on the the filter filter (see (see FigureFigure2 2c)c) [[1188].]. This This filter filter was was coated coated in in platinum platinum to toimprove improve conductivity,conductivity, and and inclusions inclusions were were then then examined examined in in a scanninga scanning electron electron microscope microscope (SEM, (SEM, Hitachi Hitachi SU 1510,SU Hitachi,1510, Hitachi, Tokyo, Tokyo, Japan) Japan) with anwith energy-dispersive an energy-dispersive X-ray X-ray spectrometer spectrometer (EDS, (EDS, Hitachi, Hitachi, Tokyo, Tokyo, Japan). FiftyJapan). inclusions Fifty inclusions from each from filter each were filter randomly were randomly selected selected for statistical for statistical analysis. analysis.

Figure 2. (a) Sheet cut from the ingot at the halfway point of the radius (“Height direction” and “RadialFigure direction” 2. (a) Sheet are cut the verticalfrom the and ingot horizontal at the halfway directions point of theof RHthe ladle,radius respectively). (“Height direction” (b) Schematic and of“Radial non-aqueous direction” solution are electrolysisthe vertical of and a steel horizontal sample. Thedirections anode of is athe sheet RH cutting ladle, fromrespectively). the sample and(b the) Schematic cathode of is non-aqueous a platinum plate. solution (c) electrolysis Collection ofof inclusionsa steel sample. separated The anode by vacuumis a sheet filtration. cutting from the sample and the cathode is a platinum plate. (c) Collection of inclusions separated by vacuum filtration. 3. Results and Discussion 3. Results and Discussion 3.1. Analysis of Oxide Inclusions in RH Refining 3.1.Inclusions Analysis of in Oxide sample Inclusions S1 were in analyzedRH Refining using EDS and found to be composed of a MnO-FeO oxide · rich inInclusions Fe, which in was sample called S1 3FeO wereMnO analyzed according using EDS to the and average found atomicto be composed ratio calculated of a MnO-FeO from the EDSoxide results rich (seein Fe, Figure which3a,e). was Thesecalled 3FeO·MnO inclusions wereaccord photographeding to the average using atomic SEM ratio and werecalculated found from to be sphericalthe EDS in results all cases. (see Figure 3a,e). These inclusions were photographed using SEM and were found to be Thespherical images in all of cases. inclusions in sample S2 were more complex. These inclusions appeared in two separateThe spherical images of forms, inclusions one with in sample a shell S2 and were the more other complex. without (seeThese Figure inclusions3b,c). appeared EDS showed in two that theseseparate two forms spherical were forms, distinct one from with one a shell another. and the The other unshelled without form (see was Figure composed 3b,c). EDS entirely showed of Mn-rich that oxide,these which two forms wasidentified were distinct as FeO from·zMnO. one another. The average The unshelled value of form z calculated was composed from EDS entirely data isof closeMn- to 4 (Figurerich oxide,3e). which In the shelledwas identified form, theas FeO·zMnO. outer shell The was average also composed value of ofz calculated FeO ·zMnO, from but EDS the data average is valueclose of to z is4 (Figure close to 3e). 2. TheIn the inner shelled core, form, however, the ou waster shell composed was also of 3FeOcomposed·MnO, ofthe FeO·zMnO, same composition but the asaverage in the whole value sphere of z is ofclose the S1to 2. inclusions The inner (see core Figure, however,3d,e). was composed of 3FeO·MnO, the same composition as in the whole sphere of the S1 inclusions (see Figure 3d,e).

Materials 2017, 10, 1206 4 of 8 Materials 2016, 9, 1206 4 of 8

Figure 3. (a) Morphology (black arrow) and EDS spectrum of the spherical inclusions typical of S1. Figure 3. (a) Morphology (black arrow) and EDS spectrum of the spherical inclusions typical of S1. The Pt peak is derived from the platinum plating that is applied to the inclusions before analysis. On The Pt peak is derived from the platinum plating that is applied to the inclusions before analysis. average, S1 inclusions are smaller than S2 inclusions. (b) Morphology and EDS spectrum of the inner On average, S1 inclusions are smaller than S2 inclusions. (b) Morphology and EDS spectrum of the core (black arrows) and outer shell (red arrows) of the shelled inclusions that occur in S2. inner(c) coreMorphology (black arrowsand EDS) and spectrum outer shellof the (redunshelled arrows) spherical of the inclusions shelled inclusionsthat also occur that occurin S2. in S2. (c) Morphology(d) Mapping analysis and EDS of the spectrum distribution of of the Mn, unshelled O, and Fe sphericalelements in inclusions shelled inclusions. that also Note occur that in S2. (d) Mappingthe intensity analysis of Mn and of theO elements distribution is high of in Mn, the ou O,ter and area Fe (shell) elements and weak in shelled in the inner inclusions. area (core), Note that thebut intensity that of ofFe Mnis high and in O the elements core and is weak high in in the the shell. outer (e) area Composition (shell) and analysis weak of in MnO-FeO the inner oxide area (core), butinclusions that of Fe in is S1 high and inS2. theThe corefirst two and columns weak in on the the shell. left show (e) that Composition the Mn/Fe ratio analysis for all of S1 MnO-FeO inclusions oxide inclusions(calculated in S1 to andbe 0.36 S2. ± The 0.09) first is at two the columnssame level on as thefor leftthe core show of thatshelled the inclusions Mn/Fe ratio in S2 for (0.35 all ± S1 0.05). inclusions (calculatedThe last two to be columns 0.36 ± on0.09) the is right at the show same that level the Mn/Fe as for ratio the core for S2 of spherical shelled inclusionsinclusions (average in S2 (0.35 of 4± 0.05). Thewith last twomaximum columns 5 and on minimum the right show3) is at that the the same Mn/Fe level ratioas for for the S2 shells spherical alone inclusions in the S2 shelled (average of 4 inclusions (average of 2 with maximum 4 and minimum 0.7), and the shell has a wider ratio range with maximum 5 and minimum 3) is at the same level as for the shells alone in the S2 shelled inclusions from 0.7 to 4. (average of 2 with maximum 4 and minimum 0.7), and the shell has a wider ratio range from 0.7 to 4. Above results indicate that the spherical 3FeO·MnO forms a molten-glass-like product before theAbove melt reaches results the indicate solidification that the temperature spherical 3FeOand later·MnO acts forms as a heterogeneous a molten-glass-like nucleation product site for before theshelled melt reaches inclusions. the Dissolved solidification Mn in temperature the melt diffuses and towards later acts 3FeO·MnO as a heterogeneous to nucleate FeO·zMnO nucleation on site for shelled3FeO·MnO. inclusions. At this Dissolved stage, such Mnoxides in theform melt molten-glass-like diffuses towards products 3FeO made·MnO of FeO·zMnO to nucleate shell FeO and·zMnO 3FeO·MnO core. The solidification temperature of FeO-MnO increases with the increase of MnO on 3FeO·MnO. At this stage, such oxides form molten-glass-like products made of FeO·zMnO shell content [19]. Because FeO·zMnO has a higher MnO content and hence a higher solidification and 3FeO·MnO core. The solidification temperature of FeO-MnO increases with the increase of temperature than 3FeO·MnO, FeO·zMnO solidifies first, forming the shell. Subsequently, 3FeO·MnO, MnOwith content its lower [19 solidification]. Because FeO temperature,·zMnO has forms a higher the solid MnO core, content leaving anda distinct hence shell-core a higher structure solidification temperaturein a solid state. than 3FeO·MnO, FeO·zMnO solidifies first, forming the shell. Subsequently, 3FeO·MnO, with its lower solidification temperature, forms the solid core, leaving a distinct shell-core structure in a solid state. Materials 20162017, 910, 1206, 1206 5 of 8

In sample S2, the ratio (z) of Mn/Fe in FeO·zMnO varies from 0.7 to 4 in shelled inclusions and fromIn 3 to sample 5 in spherical S2, the ratio FeO·zMnO. (z) of Mn/Fe When in z FeO= 0.7,·zMnO the composition varies from is 0.7close to 4to in 2FeO·MnO. shelled inclusions When z and= 4, thefrom composition 3Materials to 5 in 2016 spherical, is9, 1206 FeO·4MnO. FeO·zMnO. This Whenindicates z = that 0.7, theinter-diffusion composition occurred is close towithin 2FeO the·MnO. inclusions. When5 of 8 z The = 4, higherthe composition chemical potential is FeO·4MnO. of Mn in This the indicatesoxide shell that of S2 inter-diffusion inclusions may occurred have resulted within in the the inclusions. diffusion In sample S2, the ratio (z) of Mn/Fe in FeO·zMnO varies from 0.7 to 4 in shelled inclusions and Theof Mn higher from chemicalthe shell into potential the inner of Mn core. in Conversely the oxide shell, the ofhigher S2 inclusions Fe content may in the have core resulted diffuses in into the diffusionfrom of 3 Mnto 5 fromin spherical the shell FeO·zMnO. into the When inner z core. = 0.7, Conversely, the composition the higheris close to Fe 2FeO·MnO. content in When the core z = diffuses4, the shell.the compositionWhen diffusion is FeO·4MnO. leads to This z > indicates 3, a distinct that inter-diffusion interface between occurred shell within and the core inclusions. disappears The and into the shell. When diffusion leads to z > 3, a distinct interface between shell and core disappears and sphericalhigher FeO·zMnO chemical potential forms (see of Mn Figure in the 4). oxide shell of S2 inclusions may have resulted in the diffusion · sphericalof Mn FeO fromzMnO the shell forms into (see the Figureinner core.4). Conversely, the higher Fe content in the core diffuses into the shell. When diffusion leads to z > 3, a distinct interface between shell and core disappears and spherical FeO·zMnO forms (see Figure 4).

Figure 4. Schematic illustration of the evolution of MnO-FeO inclusions. Spherical 3FeO·MnO Figure 4. Schematic illustration of the evolution of MnO-FeO inclusions. Spherical 3FeO·MnO (left- (left-hand) forms a molten-glass-like product before the melt reaches the solidification temperature. hand) forms a molten-glass-like product before the melt reaches the solidification temperature. The The centerFigure sphere 4. Schematic is a shell-core illustration inclusion of the evolution formed of by MnO-FeO dissolved inclusions. Mn diffusing Spherical towards 3FeO·MnO 3FeO (left-·MnO to center sphere is a shell-core inclusion formed by dissolved Mn diffusing towards 3FeO·MnO to nucleatehand) FeO forms·zMnO a molten-glass-like shell on 3FeO· MnO.product 3FeO before·MnO the me formslt reaches the solid the solidification core after solidification, temperature. The leaving a nucleatecenter FeO·zMnO sphere is shella shell-core on 3FeO·MnO. inclusion 3FeO·MnOformed by dissolved forms the Mn solid diffusing core aftertowards solidification, 3FeO·MnO toleaving distinct shell-core structure. a distinctnucleate shell-core FeO·zMnO structure. shell on 3FeO·MnO. 3FeO·MnO forms the solid core after solidification, leaving a distinct shell-core structure. When samples S3 and S4 were examined, three di differentfferent types of Mn-free oxide inclusions had When samples S3 and S4 were examined, three different types of Mn-free oxide inclusions had emerged: SiO 22,, Al-Si Al-Si oxide, oxide, and and Al Al2OO33 (see(see Figure Figure 55).). No MnO-FeO oxides werewere foundfound eithereither atat thethe beginningemerged: or the SiO end2, Al-Si of oxide, the third and andAl2O3 final (see Figure (degassing) 5). No stage,MnO-FeO despite oxides the were fact found that either FeMn at hadthe been beginningbeginning or the or endthe endof theof the third third and and final final (degassi (degassing)ng) stage,stage, despite despite the the fact fact that thatFeMn FeMn had been had been added not only in the second stage (traditionally reserved for deoxidization and alloying) but also addedadded not only not onlyin the in secondthe second stage stage (traditionally (traditionally reserved for for deoxidization deoxidization and and alloying) alloying) but also but in also in thein the firstthe first stagefirst stage stage (traditionally (traditionally (traditionally reserved reservedreserved only only forfor for decarburization).decarburization). decarburization). This This This means means means that that thatthe pre-alloyingthe the pre-alloying processprocess had not had interfered not interfered with with the the levellevel level ofof of inclusionsinclusions inclusions in in the thethe final finalfinal product. product.product.

Figure 5. The type and three-dimensional images of Mn-Free oxides in S3 and S4 (a) SiO2 in S3, Figure 5. The type and three-dimensional images of Mn-Free oxides in S3 and S4 (a) SiO2 in S3, (b) Al-Si (b) Al-Si oxide in S3, and (c) Al2O3 in S3; (d) SiO2 in S4, (e) Al-Si oxide in S4, and (f) Al2O3 in S4. oxide in S3, and (c) Al2O3 in S3; (d) SiO2 in S4, (e) Al-Si oxide in S4, and (f) Al2O3 in S4.

Figure 5. The type and three-dimensional images of Mn-Free oxides in S3 and S4 (a) SiO2 in S3,

(b) Al-Si oxide in S3, and (c) Al2O3 in S3; (d) SiO2 in S4, (e) Al-Si oxide in S4, and (f) Al2O3 in S4.

Materials 2017, 10, 1206 6 of 8 Materials 2016, 9, 1206 6 of 8

A decrease in size occurred between the second and third stages in all types of oxide inclusions (see Figure6 6),), becausebecause thethe largerlarger inclusions inclusions had had already already ﬂoated floated into into the the slag. slag.

Figure 6. The average sizes of oxide inclusio inclusionsns in the S3 and S4 samples.

3.2. Quality and Cost Analysis of the Pre-Alloying Innovation Carbon content is an important factor in the productionproduction of low-carbon silicon steel. The carbon −2 content in the the first first stage stage of of the the conventional conventional al alloyingloying process process is is in in the the range range of of 3.0–4.0 3.0–4.0 × ×10−102 wt. wt.% for % high-qualityfor high-quality silicon silicon steel steel [20,21]. [20, 21Be].cause Because of the of theearly early addition addition of medium-carbon of medium-carbon FeMn FeMn alloy alloy in the in pre-alloyingthe pre-alloying process, process, the thecarbon carbon content content at atthe the en endd of offirst first stage stage was was considerably considerably higher higher than than the level requiredrequired forfor highhigh quality quality steel. steel. The The carbon carbon content content dropped dropped sharply sharply during during the the second second stage stage and −3 andreached reached 1.0 × 1.010 × 10wt.−3 wt. % at% theat the end end of RHof RH refinement, refinement, a level a level even even lower lower than than that that produced produced in the in theconventional conventional alloying alloying process process (see Table (see1 ).Table Thus, 1) the. Thus, carbon the content carbon can content be expected can be to consistentlyexpected to consistentlymeet or exceed meet the or industry exceed the standard industry for standard high quality for high silicon quality steel. silicon steel.

Table 1. Chemical compositions of steel samples (wt. %) in different different stages of the the pre-alloying pre-alloying process and in the final ﬁnal stage of the conventionalconventional alloying process.

ProcessProcess PointPoint C C Si Si Mn Mn Al Als s S1 S1 4.4 × 10 4.4− 2× 10−2 2.1 × 2.110 ×− 310−3 6.1 6.1× ×10 10−2−2 2.02.0 ×× 1010−−3 3 −3 −3 −2 −3 S2 1.0 × 10 −32.3 × 10 −3 4.9 × 10 −2 1.9 × 10−3 Pre-alloying process S2 1.0 × 10 2.3 × 10 4.9 × 10 1.9 × 10 Pre-alloying process S3 5.0 × 10−3 2.6 × 10−1 3.0 × 10−1 3.3 × 10−1 −3 −1 −1 −1 S4 S3 1.0 × 10 5.0− 3× 10 2.6 × 2.610 ×− 110 3.6 3.0× ×10 10−1 3.32.8 ×× 1010− 1 S4 1.0 × 10−3 2.6 × 10−1 3.6 × 10−1 2.8 × 10−1 Conventional alloying process End of RH <5.0 × 10−3 2.5 × 10−1 3.5 × 10−1 2.8 × 10−1 Conventional alloying process End of RH <5.0 × 10−3 2.5 × 10−1 3.5 × 10−1 2.8 × 10−1

Because MnO-FeO oxides form immediately after the addition of low-cost, medium-carbon FeMn Because MnO-FeO oxides form immediately after the addition of low-cost, medium-carbon in the pre-alloying process, part of the added Mn element (about 21%) goes into inclusions during the FeMn in the pre-alloying process, part of the added Mn element (about 21%) goes into inclusions first stage. This loss of Mn is compensated by the later addition of high-cost, low-carbon FeMn during during the first stage. This loss of Mn is compensated by the later addition of high-cost, low-carbon the second stage. The amount of high-cost FeMn alloy needed in the second stage is thus reduced FeMn during the second stage. The amount of high-cost FeMn alloy needed in the second stage is about 20%, by the previous addition of low-cost, medium-carbon FeMn in the first stage. Based on thus reduced about 20%, by the previous addition of low-cost, medium-carbon FeMn in the first the price of the two kinds of FeMn alloys in 2017 [22,23], the overall cost reduction introduced by stage. Based on the price of the two kinds of FeMn alloys in 2017 [22,23], the overall cost reduction pre-alloying amounts to 3%. introduced by pre-alloying amounts to 3%. 4. Conclusions 4. Conclusions A new fabrication method of pre-alloying was introduced in the RH refining process of low-carbon A new fabrication method of pre-alloying was introduced in the RH refining process of low- silicon steel. The evolution of inclusions produced by pre-alloying was investigated. The quality of the carbon silicon steel. The evolution of inclusions produced by pre-alloying was investigated. The final product of RH refining and the cost of the pre-alloying RH process were analyzed. Three main quality of the final product of RH refining and the cost of the pre-alloying RH process were analyzed. conclusions were obtained: Three main conclusions were obtained:

Materials 2017, 10, 1206 7 of 8

(1) MnO-FeO inclusions formed as a by-product of pre-alloying. The evolution of the oxide is as follows: spherical Fe-rich 3FeO·MnO becomes shelled Fe-rich, low-melting point 3FeO·MnO in the core and Mn-rich, high-melting point FeO·zMnO on the surface of the shell. Some of shelled inclusions become spherical Mn-rich FeO·zMnO. These inclusions are eliminated at the beginning of the degassing stage. (2) Inclusion evolution in pre-alloying RH refining is different from that in conventional RH refining, but the ingots produced by pre-alloying are of a quality that meets or exceeds industry-wide specifications with respect to quality, carbon content, and inclusion size. (3) When pre-alloying is used, the amount of high-cost, low-carbon FeMn alloy required during the deoxidization/alloying stage is reduced by 20%, thus reducing the overall cost of the process by 3%.

Acknowledgments: This work was supported by the National Natural Science Foundation of China (No. U1460103). Support was also provided by the Instrumental Analysis & Research Center in Shanghai University. The manuscript was finalized when Huigai Li, as a visiting scholar, visited the National High Magnetic Field Laboratory, which is supported by NSF DMR-1157490, the State of Florida, and department of energy. Thanks also to Mary Tyler for preliminary editing. Author Contributions: Shaobo Zheng, Huigai Li, Jinglin You, and Qijie Zhai conceived and designed the method and experiments; Shaobo Zheng and Fangjie Li performed the experiments; Huigai Li, Fangjie Li, and Ke Han analyzed the data and wrote the manuscript. Conflicts of Interest: The authors declare no conflict of interest.

References

1. Liang, Y.F.; Ye, F.; Lin, J.P.; Wang, Y.L.; Chen, G.L. Effect of annealing temperature on magnetic properties of cold rolled high silicon steel thin sheet. J. Alloys Compd. 2010, 491, 268–270. [CrossRef] 2. Bellot, J.P.; Descotes, V.; Jardy, A. Numerical modeling of inclusion behavior in liquid metal processing. JOM 2013, 65, 1164–1172. [CrossRef]

3. Park, J.H.; Todoroki, H. Control of MgO·Al2O3 spinel inclusions in stainless steels. ISIJ Int. 2010, 50, 1333–1346. [CrossRef] 4. Katsuhiro, S.; Mizukami, Y. Reaction mechanism between alumina graphite immersion nozzle and low carbon steel. ISIJ Int. 1994, 34, 802–809. [CrossRef] 5. Geng, D.Q.; Lei, H.; He, J.C. Effect of traveling magnetic field on flow, mixing, decarburization and inclusion removal during RH refining process. ISIJ Int. 2012, 52, 1036–1044. [CrossRef] 6. Fernandes, M.; Pires, J.C.; Cheung, N.; Garcia, A. Influence of refining time on nonmetallic inclusions in a low-carbon, silicon-killed steel. Mater. Charact. 2003, 51, 301–308. [CrossRef] 7. Van Ende, M.A.; Kim, Y.M.; Cho, M.K.; Choi, J.; Jung, I.H. A kinetic model for the Ruhrstahl Heraeus (RH) degassing process. Metall. Mater. Trans. B 2011, 42, 477–489. [CrossRef] 8. Zhen, L.I.; Liu, C.J.; Sun, Q.; Jiang, M.F. Effect of deoxidation process on distribution characteristics of inclusions in silicon steel slabs. J. Iron Steel Res. Int. 2015, 22, 104–110. [CrossRef] 9. Nurmi, S.; Louhenkilpi, S.; Holappa, L. Optimization of intensified silicon deoxidation. Steel Res. Int. 2013, 84, 323–327. [CrossRef] 10. Kwon, Y.; Choi, J.; Sridhar, S. The morphology and chemistry evolution of inclusions in Fe-Si-Al-O melts. Metall. Mater. Trans. B 2011, 42, 814–824. [CrossRef] 11. Shinozaki, N.; Ishido, K.; Mori, K.; Kawai, Y. Rate of reduction of MnO in slag by liquid iron containing carbon. Tetsu-to-Hagané 1984, 70, 73–80. [CrossRef] 12. Suito, H.; Inoue, R. Assessment of maganese distribution in hot metal and steel. ISIJ Int. 1995, 35, 266–271. [CrossRef] 13. Hils, G.; Newirkowez, A.; Kroker, M.; Grethe, U.; Jürgensen, R.S.; Kroos, J.; Spitzer, K.H. Conventional and tailored Mn-bearing alloying agents for the production of high manganese steels. Steel Res. Int. 2015, 86, 411–421. [CrossRef] 14. Jung, S.M.; Rhee, C.H.; Min, D.J. Thermodynamic properties of manganese oxide in BOF slags. ISIJ Int. 2002, 42, 63–70. [CrossRef] Materials 2017, 10, 1206 8 of 8

15. Wu, P.; Eriksson, G.; Pelton, A.D. Critical evaluation and optimization of the thermodynamic properties and phase diagrams of the CaO-FeO, CaO-MgO, CaO-MnO, FeO-MgO, FeO-MnO, and MgO-MnO systems. J. Am. Ceram. Soc. 1993, 76, 2065–2075. [CrossRef]

16. Van Ende, M.A.; Guo, M.; Proost, J.; Blanpain, B.; Wollants, P. Formation and morphology of Al2O3 inclusions at the onset of liquid Fe deoxidation by Al addition. ISIJ Int. 2011, 51, 27–34. [CrossRef] 17. Zhang, L.; Taniguchi, S.; Cai, K. Fluid ﬂow and inclusion removal in continuous casting tundish. Metall. Mater. Trans. B 2000, 31, 253–266. [CrossRef] 18. Doostmohammadi, H.; Karasev, A.; Jönsson, P.G. A comparison of a two-dimensional and a three-dimensional method for inclusion determinations in tool steel. Steel Res. Int. 2010, 81, 398–406. [CrossRef] 19. Dekkers, R.; Blanpain, B.; Wollants, P. Crystal growth in liquid steel during secondary metallurgy. Metall. Mater. Trans. B 2003, 34, 161–171. [CrossRef] 20. Sumida, N. Production of ultra-low carbon steel by combined process of bottom-blown converter and RH degasser. Kawasaki Steel Tech. Rep. 1983, 8, 69–76. 21. Hirata, R. Low-carbon steels manufactured by circulating ﬂow vacuum degassing process. JOM 1966, 18, 617–622. [CrossRef] 22. You, B.D.; Lee, B.W.; Pak, J.J. Manganese loss during the oxygen refining of high-carbon ferromanganese melts. Met. Mater. Int. 1999, 5, 497–502. [CrossRef] 23. Asian Metal. Available online: http://www.asianmetal.com/FerromanganesePrice/Ferromanganese.html. (accessed on 25 August 2017).

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).