2015 Volume 55 Issue 8 Pages 1661-1668
The alloy/liquid Fe interaction behaviour at early stages of deoxidation/alloying have considerable influence on the mechanical properties of steel products. The interfacial reaction between FeMnSi and liquid Fe, as well as the inclusion formation, were experimentally studied at the addition early stages based on quenched FeMnSi–Fe diffusion couples. The microstructure and composition of the quenched diffusion couple were evaluated and the inclusions in the reaction zone were characterised. Five regions were distinguished in the quenched reaction zone, i.e. (1) a solidified Fe shell, (2) a α1 phase consisting of a solid α1 phase and a Fe–Mn–Si liquid melt at the experimental temperature, (3) eutectic α1+Fe5Si3 and (4) eutectic Mn5Si3+MnSi phases, which are liquid at the experimental temperature and (5) the original FeMnSi alloy which did not melt during the interaction. Plenty of oxysulphide inclusions are observed in the reaction zone: Fe(O,S) inclusions formed in the bulk iron and an inclusion poor zone is observed in the inner part of the Fe shell. Large amounts of tiny (Mn,Si)O inclusions formed in the α1 phase near the Fe/α1 interface and large (Mn,Si)(O,S) inclusions precipitated in the liquid phase after solidification.
The addition of Mn and Si to molten steel during ladle refining serves several purposes: (1) adjustment of the steel composition as alloy element. Mn addition to steel products improves their hardness and wear resistance, while the use of Si improves the strength and elasticity properties;1,2) (2) removal of dissolved oxygen as deoxidation agents.3,4) Mn–Si combined deoxidation is used for several special steel grades, e.g. coarse-grained steel, tire-cord steel and stainless steel.5) Due to their low melting temperature and deformability, the deoxidation products (MnO–SiO2–(Al2O3) glassy inclusions), have a less detrimental effect on the mechanical properties of steel products than hard Al2O3 inclusions, i.e. the Al deoxidation products.5,6) Another advantage of Mn–Si combined deoxidation compared to Al deoxidation is that the remaining MnO–SiO2 based inclusions are more uniformly dispersed in solidified steel.7) The uniformly dispersed inclusion could be used to improve the steel performance. Al2O3 inclusions tend to be pushed to the region of final solidification, while both solid and liquid MnO–SiO2 inclusions are easily engulfed in solid steel during solidification.8) More importantly, the liquid MnO–SiO2 inclusions exhibit a narrow size distribution compared to Al2O3 inclusions, allowing an effective control in terms of inclusion size.9,10,11) Al2O3 inclusions are easily attracted by each other to form large clusters in liquid steel, while liquid MnO–SiO2 inclusions do not agglomerate, even when they are fairly close to each other.9,10,11) In addition, the formation of liquid MnO–SiO2 inclusions can also effectively prevent nozzle clogging and consequently improve steel castability.7)
Mn and Si can be added to the molten steel in the form of a FeMnSi alloy. To maximise the beneficial effects of Mn and Si, their content and inclusion formation has to be precisely controlled. The former requires thorough knowledge on the dissolution of FeMnSi alloys into the molten steel and on the interfacial reactions between the added alloys and the steel. The latter relies on the better understandings of formation mechanism of the Mn-containing inclusions at the early stages of deoxidation. These Mn-containing inclusions determine the inclusion characteristics in final steel product and consequently influence the mechanical properties of steel product. In our previous research, the interfacial reactions between alloying agents (i.e. Al, Ti and Mn) and molten steel have been investigated.12,13,14,15,16,17) The liquid Fe is found to solidify immediately after contacting the cold alloy, while the alloying agent (Al/Mn) heats up and melts due to its low melting temperature.12,13,14) Also for additions with a melting point higher than Fe, like Ti, internal dissolution can take place.15,16,17) Thereafter, Fe/alloy interdiffusion results in the formation of a reaction zone which can be partially solid and liquid depending on the local concentration and temperature. Regarding the inclusion formation, an inclusion free zone is observed in the inner part of the Fe shell.12,14) Its formation is believed to be due to a local depletion of O and/or S, caused by the diffusion of the alloy element, O and/or S and deoxidation/desulphurisation reactions at the diffusion front. Besides, plenty of deoxidation products are also formed in the reaction zone. In case of Al additions, large angular Al2O3 inclusions are observed at the initial Fe/Al interface, while aggregates of fine Al2O3 inclusions are found further away from the diffusion front. The large angular Al2O3 inclusions result from the heterogeneous nucleation of Al2O3 on FeOx and SiO2 inclusions and from reduction of these oxides, whereas the fine Al2O3 inclusions arose from homogeneous nucleation at the reaction front.12) Since Ti is added after Al, pure TiOx inclusions are not formed due to the low dissolved oxygen in liquid steel (~4–6 ppm). The modification of original Al2O3 inclusions, however, is observed in regions of high Ti content.15) Only small spherical MnO/Mn(O,S) inclusions are observed in case of Mn addition, suggesting that these inclusions are formed during solidification of the reaction zone.14) For combined alloy addition and deoxidation, e.g. by FeMnSi, the precise control of the alloy content remains a challenge since the dissolution and interfacial reactions are more complicated. Moreover, the (Mn,Si)(O,S) inclusion formation mechanism and their characteristics at the early stages of deoxidation are still not clear.
In this study, the interactions between Fe and a FeMnSi alloy shortly after alloy addition were investigated based on quenched diffusion couple. A small volume of liquid Fe containing various contents of dissolved oxygen and sulphur was brought into contact with a cold FeMnSi alloy. The dissolution behaviour of the FeMnSi and the interfacial reaction between liquid Fe and FeMnSi were studied. The inclusion formation mechanism in the reaction zone was discussed. The influences of the oxygen and sulphur contents and the relative position to the diffusion front on the inclusion characteristics were studied. This work contributes to improve the Mn–Si combined alloy addition/deoxidation during ladle refining.
The detailed experimental method has been described in our previous paper.12,13,14,15) 100 g electrolytic Fe was mixed with Fe2O3 and FeS (reagent grade) in a magnesia crucible. The mixture was then melted within a vertical tube furnace (GERO HTRV 100-250/18, MoSi2 heating elements) at 1600oC under Ar atmosphere. The Ar was purified by passing over Mg chips at 550°C. The oxygen content in the off-gas was measured with a solid state ceramic oxygen sensor (Rapidox 2100), yielding a typical value of about 10−18 ppm. After the Fe was melted, the melt was stabilized at 1600°C for 60 min. A small piece of FeMnSi (5×5×5 mm3 cubic shape) was placed in a quartz tube. The end of the quartz glass tube was narrowed in order to maintain the FeMnSi alloy piece inside. The quartz glass tube with the alloy was then quickly introduced into the furnace and lowered into the melt (Fig. 1(a)). A small volume of liquid Fe was sucked and brought into contact with FeMnSi for a designated interaction time (Fig. 1(b)). Thereafter, the tube was rapidly withdrawn from the furnace and quenched in water. The detailed conditions of each test are listed in Table 1. The FeMnSi alloy (Table 2) was collected from a steel plant and its composition was measured with inductively coupled plasma atomic emission spectroscopy (ICP-AES). The lower part of the obtained diffusion couple, i.e. metallic Fe (Fig. 1(c)), was cut into small pieces (0.5–1.0 g) and cleaned with acetone in an ultrasonic bath. At least two pieces of Fe were analysed with LECO combustion analysis (TC-400 and CS-230) for total oxygen (T.O) and total sulphur (T.S) measurement. The measured values (average of two samples) are listed in Table 1. Since the impurity level of the electrolytic Fe was negligible, the measured T.O and T.S was considered to be, respectively, the dissolved [O] and [S] at experimental temperature. The vertical central cross section of upper part of the diffusion couple (reaction zone in Fig. 1(c)) was mirror polished and subjected to microstructural characterization using electron probe microanalysis (FE-EPMA, JXA-8530F). The composition of reaction zone is analysed with EPMA-WDS, i.e. wavelength dispersive spectroscopy.
Experimental set-up and procedure: (a) introducing the quartz tube into the melt, (b) bringing liquid Fe into contact with the FeMnSi alloy and (3) overview of the quenched diffusion couple.
| Test No. | Contact time (s) | Composition (ppm in mass) | |
|---|---|---|---|
| T.O | T.S | ||
| 1-a | 2 | 200 | 0 |
| 1-b | 5 | ||
| 1-c | 10 | ||
| 2-a | 5 | 300 | 140 |
| 2-b | 10 | ||
| 3-a | 5 | 600 | 310 |
| 3-b | 10 | ||
| 4 | 10 | 900 | 450 |
* O and S are introduced into melts through Fe2O3 and FeS addition
| Fe | Si | Mn | |
|---|---|---|---|
| FeMnSi | 11.6 | 29.4 | 52.9 |
Figure 2 gives the Fe–Si and Mn–Si binary phase diagram (up to 50 wt% Si) calculated with FactSage, using the FSstel database. In the Fe–Si phase diagram (Fig. 2(a)) there are two eutectic points: (1) at 1195°C and 19 wt% Si and (2) at 1202°C and 21.5 wt% Si; there are also two eutectic points for the Mn–Si phase diagram (up to 50 wt% Si Fig. 2(b)): (1) at 1040°C and 11 wt% Si and (2) at 1241°C and 28.5% Si. As for the intermetallic phases, it can be seen from the table in Fig. 2 that α1 (Fe3Si) and Mn3Si; Fe5Si3 and Mn5Si3; and FeSi and MnSi belong to same space group and have a similar Si content respectively. Narasimha et al. found that Fe can be partially substituted by Mn in the Fe5Si3 phase18)and Niculescu et al. found that Mn can partially occupy Fe sites in α1.19)
Phase diagram (a) Fe–Si and (b) Mn–Si, and the Si content of corresponding phases.
After bringing the liquid iron in contact with the cold solid FeMnSi for the targeted interaction time, the diffusion couple was withdrawn from the furnace and quenched in water. A typical macroscopic view of the obtained diffusion couple is shown in Fig. 1(c) (Test 1-a as an example). The bottom part of the addition deformed from its original cubic shape into a cylindrical shape, suggesting that the alloy partially melted in this short interaction time. Fe and FeMnSi make good contact and macroscopically a reaction zone can be clearly observed, marked as R-Zone in Fig. 1(c). Figure 3(a) shows the typical microstructure of the Fe–FeMnSi diffusion couple (BSE image of Test 1-a), in which five phases can be distinguished, i.e. α-Fe, α1, Fe5Si3, Mn5Si3 and MnSi, as indicated as dashed lines. Due to the fine structures (around 1 μm) and low contrast, the α1/α1+Fe5Si3 interface cannot be clearly distinguished in Fig. 3(a). Figures 3(b) and 3(c) show the enlarged interface of α1/α1+Fe5Si3 and that of Mn5Si3/Mn5Si3+MnSi. Figure 4 gives the measured concentration profile as a function of the perpendicular distance from the diffusion front, i.e. the α-Fe/α1 interface in Fig. 3(a). It should be noted that the concentration profile is measured on the position of or at least close to the cylindrical centreline.
The microstructure of reaction zone: (a) overview of Test 1-a, and enlarged (b) α1/α1+F5S3 interface and (c) M5S3/M5S3+MS interface, in which F5S3 represents the Fe5Si3 phase; MS the MnSi phase and M5S3 the Mn5Si3 phase.
Concentration profile and distinguished phases in Test 1-a (Fig. 3(a)), L and PL represent the liquid phase and partially liquid/solid phase respectively.
The total diffusion distance marked with Fe content in Fig. 4 is around 1250 μm, which approximately equals the thickness of the melted FeMnSi alloy during the interaction. The concentration profile in Fig. 4 can be divided into two domains, i.e. Fe and Mn dominant regions based on the relative Fe/Mn contents. The diffusion couple can be further divided into five regions depending on the Si content (also shows on Fig. 3(a)). Si is not detected in region I, suggesting it is the solidified Fe shell, i.e. the Fe (1600°C) which solidified on the cold FeMnSi (25°C or slightly preheated before addition). Region II contains α1 with up to 15 wt% Si, representing a liquid (partially liquid) phase at high temperature. As can be seen from Fig. 4, the Si content in this α1 phase increases from 13 wt% to 15 wt%, i.e. near-stoichiometric α1 phase (Fe3–xMnxSi).19,20,21) It should be noted that the aggregation state (i.e. liquid or solid) of region II at the experimental temperature depends on the local concentration and temperature. During the solidification of this region, part of the liquid phase first transformed to α2 (primitive cubic) or α-Fe at high temperature and then to α1 (Face-centred cubic) according to the Fe–Si phase diagram (Fig. 2(a)). The phase transformation from α-Fe and/or α2 to nearly stoichiometric α1 is accompanied by a 4% volume decrease and may result in the cracks shown in Fig. 3(a).20) Regions III (Fig. 3(b)) and IV (Fig. 3(c)) exhibit a typical eutectic structure, indicating they were liquid during the experiment. The bright and grey phases in region III (Fig. 3(b)) contain around 15 wt% and 23 wt% Si, suggesting they are α1 and Fe5Si3, respectively, in which Fe is partially substituted by Mn atoms, stabilizing the ‘high temperature phase’ Fe5Si3. The grey and dark phases in region IV (Fig. 3(c)) contain around 23 wt% and 32 wt% Si, suggesting they are Mn5Si3 and MnSi, respectively, in which Mn could be partially substituted by Fe. These phases and eutectic structure found in regions III and IV agree well with the phase diagram shown in Fig. 2. Region V also exhibits a typical eutectic structure, which has the same microstructure and composition as the original FeMnSi alloy, indicating region V is the original FeMnSi alloy and did not melt during the experiment.
3.2. The Reaction Zone Development and the Interfacial Reaction 3.2.1. Influence of Interaction TimeFigure 5 shows the microstructure evolution of the reaction zone as a function of interaction time for Test 1 with 200 ppm oxygen. The corresponding concentration profiles are given in Fig. 6. It should be noted that because the position of the initial Fe/FeMnSi interface cannot be deduced from the quenched samples, the x-axis for the different samples lies not necessarily on the same absolute position. The measured average thickness of various phases is listed in Table 3. Due to the fine structures (around 1 μm) and low contrast at the α1/α1+Fe5Si3 interface, the thickness of α1 is estimated on the concentration profiles (Fig. 6). It can be seen that the thickness of regions III and IV, as well as the total diffusion distance, increase with increasing interaction times. As discussed previously, regions III and IV have a typical eutectic structure and therefore are liquid at experimental temperature; the total diffusion distance approximately equals the thickness of the liquid region, i.e. part of region II + regions III and IV. Their thickness increases due to the melting of the FeMnSi alloy and the more extensive interdiffusion at longer interaction times. As discussed in our previous paper,14) the melting of the added alloy is primarily heat controlled: the steep temperature gradient between liquid Fe and cold alloy, the heat of mixing generated at reaction zone and convection in this region. On the other hand, the thickness of region II (α1) decreases with interaction time, i.e. from 500 μm after 2 s interaction to 330 (5 s) and 150 μm (10 s). This is probably linked to the aggregation state of this region at the experimental temperature. The interdiffusion of Si and Fe near the initial Fe/FeMnSi interface leads to the formation of a partially solid and liquid region. The measured Si content is constant in the quenched α1 phase, also indicating part of this phase (the solid α1 at experimental temperature) grows through the solid state diffusion at experimental temperature. Pierre et al. investigated the interaction between mild steel and Mg–Si melts at a temperature around 730°C and also observed the formation of stoichiometric α1 near the initial steel/melt interface.22) They concluded that the growth of this phase is limited by Fe diffusion (diffusion coefficient is around 10−12−10−11 m2/s at 1220°C) in α1.22) In the present case, the newly formed α1 phase is solid at the early stages and its growth is mainly governed by the Fe release from Fe shell/newly formed α1 phase to the liquid reaction zone. Due to the continuous heat supply from the Fe melt, the newly formed α1 phase gradually melts and its thickness decreases in increasing interaction time. Therefore, the total thickness (newly formed α1 phase + part of liquid phase at experimental temperature) decreases in increasing interaction time as well.
Interfacial microstructure of reaction zone with interaction time of (a) 2 s; (b) 5 s and (c) 10 s in Test 1.
The measured concentration profile of (a) 2 s; (b) 5 s and (c) 10 s in Test 1 (Fe–FeMnSi diffusion couple).
| Interaction time/s | α1* | α1+Fe5Si3 | Fine MnSi+Mn5Si3 | Total diffusion* |
|---|---|---|---|---|
| 2 | 500 | 433 | 410 | ~1250 |
| 5 | 330 | 490 | 800 | ~1500 |
| 10 | 150 | 540 | 1600 | >1500 |
In previous study (Mn–Fe[O,S] interaction couple),14) it was found that oxygen and sulphur additions locally raise the temperature of the Fe/Mn interface by releasing chemical heat of deoxidation and desulphurisation reactions,23,24) consequently enhancing internal dissolution. Figure 7 shows the influence of the oxygen and sulphur addition on the current reaction zone development after 10 s interaction. Table 4 lists the average thickness of different phases of Figs. 5(b) and 7. It can be seen that the thickness of region III (α1+Fe5Si3 eutectic) slightly increases for increasing [O] and [S] contents, while that of region II (part liquid phase + solid α1 phase) decreases. Similarly, this is probably due to the deoxidation and desulphurisation reaction at the Fe/FeMnSi interface, which releases chemical heat and accelerates the melting of the newly formed α1 phase at experimental temperature.23,24) On the other hand, the thickness of region IV (MS+M5S3 eutectic) remains constant, being independent of [O] and [S] additions. This region is liquid at experimental temperature and the liquid/solid interface at the FeMnSi side is primarily determined by the melting of the FeMnSi alloy. Since the deoxidation and desulphurisation reactions take place at the Si/Mn and [O]/[S] diffusion front, close to the initial Fe/FeMnSi interface, they can only locally raise the temperature at their reaction front and do not significantly influence the melting of the FeMnSi alloy.
Interfacial microstructure of reaction zone with various oxygen and sulphur addition (a) Test 2-b; (b) Test 3-b and (c) Test 4.
| Test | [O] | [S] | α1* | α1+Fe5Si3 | Fine MnSi+Mn5Si3 |
|---|---|---|---|---|---|
| 1-b | 200 | 0 | 150 | 540 | 1600 |
| 2-b | 300 | 140 | 150 | 600 | 1600 |
| 3-b | 600 | 310 | 130 | 640 | 1600 |
| 4 | 900 | 450 | 120 | 700 | 1500 |
Oxysulphide inclusions (Mn,Si)(O,S) are formed in reaction zone (regions II–IV, Fig. 8) in case of high [O] and [S] content in liquid Fe. The formation of these inclusions is believed to be a consequence of the diffusion of O, S, Mn and Si towards the diffusion front, where deoxidation and desulphurisation reactions take place. Depending on the interaction time, iron composition ([O] and [S] content) and location, the inclusions show different characteristic, which can be categorised into three types based on their size, morphology and composition:
Inclusion free zone formation in (a) Test 2-b and (b) Test 3-b.
(i) Small Fe(O,S) inclusions (Fig. 8) in the bulk Fe with high dissolved [O] (>200 ppm) and [S] content.
(ii) (Mn,Si)O (Fig. 8) in α1 phase near the Fe/α1 interface (region II). Tiny spherical (Mn,Si)O inclusions (Fig. 8(a)) are observed close to Fe/α1 interface in case of low [O] and [S] content in liquid Fe, while both tiny spherical and large non-spherical inclusions are found (Fig. 8(b)) in case of high [O] and [S] content.
(iii) Spherical or nearly spherical (Mn,Si)(O,S) inclusions (Fig. 9) from interface of region II/III to region IV.
(Mn,Si)(O,S) inclusions of Test 2-b: (a) near the interface of region II (α1) and III, (b) in region III and (c) in region IV. Dash line represents α1/α1+F5S3 interface.
Plenty of Fe(O,S) inclusions are observed at the side of bulk Fe (region I, Fig. 8) in tests with high [O] and [S] content. Similar to the Fe–Mn diffusion couple,14) an “inclusion poor” zone is observed between bulk Fe/Mn (or Si) diffusion front. Much less and smaller inclusions are observed in this region comparing to the bulk Fe, suggesting locally lower [O] and [S] contents in this region at high temperature. The formation mechanism of this “inclusion poor” zone has been discussed in detail elsewhere:14) (1) immediately after contact, the liquid Fe is solidified and due to the high [O] and [S] contents in the Fe melt, Fe(O,S) inclusions formed in this solidified Fe shell; (2) the remaining [O] and [S] in solidified Fe are consumed at the Mn/Si diffusion front, leading to a concentration gradient of [O] and [S] from bulk Fe to Mn/Si diffusion front, and consequently their diffusion. This diffusion leads to a decrease of [O] and [S] contents in the solidified Fe, creating a [O] and [S] depleted region; (3) the Fe(O,S) are not stable and tend to dissolved again in this [O] and [S] depleted region. Due to the larger diffusivity of [O] and [S] in solidified Fe than that of Mn, the reaction front of Fe(O,S) dissolution moves faster than that of (Mn,Si)(O,S) formation, creating the “inclusion poor” region as observed in Fig. 8;24,25,26) (4) with increasing interaction time, the diffusion of [O] and [S] from bulk Fe to the reaction zone proceeds, enlarging the area in the solid Fe with low [O] and [S] contents, and consequently resulting in a larger inclusion poor zone.
3.3.2 Deoxidation and Desulphurisation ProductsFigure 10 shows the inclusion size and number density changes as a function of the distance from the Fe/α1 interface (towards the FeMnSi alloy). Large amount of tiny (Mn,Si)O inclusions are observed in α1 phase near the Fe/α1 interface. These tiny inclusions are most likely the deoxidation products formed by homogenous nucleation. With moving away from the Fe/α1 interface (towards FeMnSi alloy), (1) the number density of (Mn,Si)O inclusions decrease rapidly, indicating that the oxygen from the Fe are largely consumed at the Fe/α1 interface through deoxidation reaction; (2) the size of inclusions increases and the composition changes into oxysulphide (the inclusion composition analysed with EPMA-WDS). The latter is probably due to the desulphurisation reactions, which occurs on oxide sites and results in the formation of oxysulphide inclusions. Plenty of large spherical or nearly spherical (Mn,Si)(O,S) inclusions are formed in the area near the α1/α1+F5S3 interface (Fig. 9(a)). As shown in Fig. 4, the reaction zone near the α1/α1+F5S3 interface is liquid at experimental temperature. These (Mn,Si)(O,S) inclusions are then believed to be formed during the solidification of the reaction zone. Unlike the oxygen, sulphur from the solidified Fe only partially reacts with Mn in α1 phase to formed sulphide inclusions (MnS) at experimental temperature. This is because the MnS is not stable and tends to decompose at low sulphur content region at high temperature. The sulphur then transfers through the α1 phase to liquid phase. During the solidification, the solubility of sulphur decreases and sulphide inclusions are precipitated out. With further moving away from the α1/α1+F5S3 interface, the inclusion number density decreases and only few inclusions are found in region IV (Mn5Si3+MnSi).
Inclusion size and number density changes as a function of the distance from the Fe/α1 interface: (a) Test 2 and (b) Test 3.
It can be seen from the Fig. 10 that the position to form large oxysulphide inclusions is approaching to the Fe/α1 interface with increasing interaction time. This is because the melting of new formed α1 phase with interaction time due to the heat supply from the furnace and Fe melt, i.e. the dot line moves to the dash line in Fig. 10. Consequently, the liquid phase approaches the Fe/α1 interface and the precipitated sulphide inclusions are close to the Fe/α1 interface as well. The comparison of Tests 2 (Fig. 10(a)) and 3 (Fig. 10(b)) shows the influence of [O] and [S] content in liquid Fe on the inclusion formation in reaction zone. High [O] content in liquid Fe (Test 3) results in a formation of large amount of (Mn,Si)O inclusions in α1 phase close to the Fe/α1 interface. The average diameter of these (Mn,Si)O inclusions (0.4–1.2 μm) is also larger than that (0.4–0.8 μm) formed in diffusion couple with low [O] in liquid Fe. High [S] content seems lead to a formation of large sulphide inclusions in solidified liquid phase (part of region II and region III). Moreover, the precipitated sulphide inclusions are closer to the Fe/α1 interface in case of high [O] and [S] content in liquid Fe.
Compared to the Al deoxidation,12,13) much smaller deoxidation products, i.e. Mn-containing inclusions are observed in case of Mn14) or Mn–Si deoxidation in the quenched reaction zone. In the present work, only part of these inclusions (near the Fe/α1 interface) is formed at high temperature, i.e. primary deoxidation products. These primary inclusions have small size and would have less detrimental effect on steel properties. On the other hand, large oxysulphide inclusions are observed in case of high sulphur content in liquid Fe, which are believed to be formed during the solidification of the diffusion couple. Therefore, the dissolved sulphur cannot be removed by forming Mn-containing sulphides, i.e. by means of Mn addition. Top slag desulphurisation or Ca/Mg desulphurisation would be suggested to control the sulphur content in practice and to prevent the formation of large oxysulphide inclusions. Two solid phases, i.e. a Fe shell and a solid α1 phase are formed at the reaction zone in case of FeMnSi addition. Depending on the size of the added alloy, the thickness of these two solid layers varies and may exist in a prolonged duration, which would remarkably delay the alloy dissolution. The delayed alloy dissolution results in (1) undissolved alloy particles and also less available time for the dissolved alloy to mix with liquid steel (i.e. the compositional inhomogeneity) and (2) less time for inclusion to float out and even an inclusion enriched region around the added alloy (i.e. the decreased steel cleanliness). On the other hand, the dissolution of FeMn and FeSi were much quicker. For instance, Argyropoulos et al.27) found that after the formation of the Fe shell around cold FeSi alloy, a liquid film of eutectic composition formed at the Fe shell/FeSi interface at around 1200°C and an exothermic reaction was initiated. This exothermic reaction would facilitate the melting of Fe shell, and shorten FeSi alloy dissolution time. Although two solids layers, a Fe shell and a γ-FeMn, were also observed in case of FeMn addition, the thickness of the solid γ-FeMn at experimental temperature is believed to be very thin (few microns) and overall dissolution is limited only by the melting of Fe shell.14) What’s more, the inclusion dispersion was also found to be more homogeneous in case of the combined FeSi and FeMn addition than that of FeMn alone.28) Therefore, to precisely control the Mn and Si content and to obtain an improved steel cleanliness, the addition of either FeMnSi with small size or the combination of FeMn and FeSi alloys should be considered in practice.
The interfacial reaction and inclusion formation at early stage of FeMnSi addition to liquid Fe containing various dissolved oxygen and sulphur are investigated based on quenching FeMnSi–Fe diffusion couples. The development of reaction zone and their aggregation state at experimental temperature are discussed. The inclusion formation and their characteristic are evaluated as well. The main results are summarised as follows:
(1) Five regions are distinguished in the quenched FeMnSi–Fe diffusion couple based on their microstructure and composition, i.e. α-Fe, stoichiometric α1 phase, eutectic α1+Fe5Si3 phase, eutectic Mn5Si3+MnSi and original Mn5Si3+MnSi phase. The stoichiometric α1 phase consists of two contributions: a solid α1 phase at experimental temperature formed by releasing Fe into melted FeMnSi alloy and a Fe–Mn–Si melt which transforms upon quenching into α1 phase. Both the eutectic α1+Fe5Si3 and Mn5Si3+MnSi are liquid at experimental temperature and the original Mn5Si3+MnSi phase is not melted during the interaction.
(2) The thickness of the solid α1 phase decreases in increasing interaction time at experimental temperature due to the continuous heat supply from the furnace and melt. Oxygen and sulphur addition locally raise the temperature by releasing chemical heat of deoxidation and desulphurisation, consequently accelerating the melting of the solid α1 phase.
(3) Three types of inclusion are found in the reaction zone: (1) Fe(O,S) inclusions in solidified Fe. An inclusion poor zone is observed in the inner part of the Fe shell closing to the α-Fe/α1 interface; (2) Large amount of tiny (Mn,Si)O inclusions in α1 phase near the α-Fe/α1 interface, which arose from the homogeneous nucleation of deoxidation products; and (3) larger sulphides inclusions near the α1/α1+Fe5Si3 interface, which are precipitated during the solidification of liquid Fe–Mn–Si–S melt.
(4) Both the amount and size of (Mn,Si)O inclusions increase in increasing [O] and [S] contents in liquid Fe. The location of precipitated (Mn,Si)(O,S) sulphides inclusions approaches to the Fe/α1 interface with interaction time and the addition of [O] and [S].
