The Oxygen Steelmaking Process: Part One


The oxygen steelmaking process is a generic name given to those processes in which gaseous oxygen is used as the primary agent for autothermic generation of heat as a result of the oxidation of dissolved impurities like carbon, silicon, manganese and phosphorus and to a limited extent the oxidation of iron itself. Several types of oxygen steelmaking processes, like top blowing, bottom blowing and combined blowing have been invented.

The oxygen steelmaking process is a generic name given to those processes in which gaseous oxygen is used as the primary agent for autothermic generation of heat as a result of the oxidation of dissolved impurities like carbon, silicon, manganese and phosphorus and to a limited extent the oxidation of iron itself. Several types of oxygen steelmaking processes, like top blowing, bottom blowing and combined blowing have been invented.

The essential features of conventional steelmaking are the partial oxidation of the carbon, silicon, phosphorus and manganese present in pig iron and the accompanying reduction in the sulfur level. Blast furnace hot metal for LD (Basic Oxygen Furnace) steelmaking process ideally contains about C=4.2%, Si=max 0.8%, Mn=max 0.8%, S=max 0.05%, P=max 0.15% but these solute elements are diluted by the addition scrap which forms some 20-30% of the metallic charge.

Refining Reactions

In LD basic oxygen steelmaking process, the oxygen required for the refining reactions is supplied as a gas and both metal and slag are initially oxidized

½O2(g) ↔ [O] .....(1)

Fe + [O] ↔ (FeO) .....(2)

2(FeO) + ½O2(g) ↔ (Fe2O3) .....(3)


The actual distribution of oxygen between slag and metal is not easily determined since it is a function of a number of variables including lance height and oxygen flow rate. The principal refining reactions is of course the removal of carbon:

[C] + [O] ↔ CO2 .....(4)

[C] + (FeO) ↔ CO2 + Fe .....(5)

The Figure 1 represents an idealized diagram, showing the changes in concentrations of the elements in LD metal bath during oxygen blowing. The basic thermodynamic data for these reactions are well established and the equilibrium carbon and oxygen contents may be readily calculated for all the temperatures and pressures encountered in steelmaking.

Figure 1: The changes of bath composition during the blow in a basic oxygen steelmaking converter (idealized)

Oxidation of carbon during the oxygen converter process is most important, since the reaction increases the temperature and evolves a large amount of gases CO and CO2 that cause agitation of metal and slag and remove hydrogen, nitrogen and part of non-metallic inclusions from the metal. Owing to the pressure of the oxygen supplied and the evolution of large quantities of gases, the liquid bath becomes an intimate mixture of slag, metal and gas bubbles, with an enormous contact surface. Because of this, the reaction of carbon oxidation is self-accelerated and attains a very high rate.


In accordance with thermodynamic predictions, the removal silicon is usually completed relatively early in the blow. The reaction may be represented by equations (6) and (7).

[Si] + 2[O] ↔ (SiO2) .....(6)

[Si] + 2(FeO) ↔ (SiO2) + 2Fe .....(7)


Similar equations can be applied to manganese removal

[Mn] + [O] ↔ (MnO) .....(8)

[Mn] + (FeO) ↔ (MnO) + Fe .....(9)

Initially, the bath manganese level falls as a result of oxidation, but later, a slightly reversion, followed by a second fall occurs. These changes in the manganese content of the bath are attributed to the combined effects of rising temperature and variable slag composition, on the activities of manganese and ferrous oxides, suggesting that the reaction is close equilibrium. This view is supported by the observation that at the end of blowing, the manganese content was found to be 82% of the equilibrium value when lump lime was used and 85% of the equilibrium value when powdered lime is injected.

In the middle part of blow, the (FeO) level in the slag falls as a consequence of the decarburization process and the dilution that accompanies lime fluxing. However, towards the end of the blow, the (FeO) increases again, as carbon removal becomes less intense and dilution begins to affect the activity of manganese oxide with the result that manganese transfers from bath to slag. To some extent the manganese loss may be minimized by rising the temperature.


The partitioning of phosphorus between the slag and metal is known to be very sensitive to process conditions and so far it has been possible to build a kinetic model based on simple assumptions.

The distribution of phosphorus between slag and metal has been reviewed by Healy, who concluded that the thermodynamic behavior of phosphorus is best explained by a modified version of the ionic theory first proposed by Flood and Grjotheim. The slag-metal reactions is written in ionic form in equation (10)

2[P] + 5[O] + 3 (O2-) ↔ 2(PO3-)4 .....(10)

Healy has expressed the equilibrium distribution of phosphorus by equations that apply to specific concentration ranges in the CaO-SiO2-FeO system, i.e.:

log (%P)/[P] = 22 350/T + 7 log%CaO + 2.5 log Fet – 24.0 .....(11)

log (%P)/[P] = 22 350/T + 0.08 log%CaO + 2.5 log Fet – 16.0 .....(12)

Equation (11) is applicable to slag containing over 24%CaO while equation (12) is valid from zero %CaO to saturation.

Unfortunately, in practice the phosphorus partition ratios are far from the values calculated for equilibrium with carbon-free iron, because the oxygen potential of the slag-metal system is influenced by decarburization. A limited correlation with the carbon content in the bath has been reported, although other workers have suggested that extensive dephosphorization should be possible at high carbon levels, provided that the slag is sufficiently basic.

On the other hand, the dependence of phosphorus distribution on the FeO content of slag in LD and Q-BOP is shown in Figure 2. A parameter kPS is defined as

kPS = (%P2O5)/[%P]•(1 + (%SiO2)) = φ ((%FeO),B) .....(13)

where B is basicity; for B>2.5, kPS is found to be independent of B.

The distribution of phosphorus is also found to be related to the content of carbon in steel at the time of tapping; owing to lower carbon levels achieved in bottom-blown process, the phosphorus distribution is expected to be better than in LD. In general, high basicity and the low temperature of slag (irrespective of the FeO content) favor dephosphorization.

Figure 2: Effect of FeO content of slag on phosphorus distribution and log kPS value

Sulfur removal

Sulfur transfer takes place through the following reactions:

[S] + (O2)g = (SO2)g .....(14)

It is found that approximately 15-25% of dissolved sulfur is directly oxidized into the gaseous phase due to the turbulent and oxidizing conditions existing in the jet impact zone.

In Basic Oxygen Furnace, the metal desulphurization proceeds slowly because it is a diffusion process. It may be speeded up by improving the bath mixing and increasing the temperature, fluidity and basicity of the slag, and the activity of sulfur. At the initial stage of the heat, when the metal is rich in carbon and silicon, the activity of sulfur is high. Besides, part of sulfur is removed at the initial stages of the process when the temperature of melt is still relatively low through its reaction with manganese:

[Mn] + [S] = (MnS) .....(15)

A rise in the concentration of iron oxides in the slag promotes dissolution of lime, and therefore favors desulphurization. But the secondary and most intensive desulphurization occurs at the end and of the heat when the lime dissolves in the slag with a maximum rate and the slag basicity reaches B=2.8 and more. Thus the total desulphurization of the metal is mainly decided by the basicity of the homogeneous final slag which is formed in the oxygen converter process during the last minutes of metal blowing.

With an increase of slag basicity, the residual concentration of sulfur in metal bath becomes lower, so that the coefficient of sulfur distribution between slag and metal can be raised up to 10. The greater the bulk of slag is the largest part of the sulfur will pass into slag at the same sulfur distribution coefficient. But it is not beneficial to form a very large bulk of slag, since this increases the iron loss due to burning, causes splashings and rapid wear of the lining.

In a steel plant, regression equations based on operational data are employed to predict the end point sulfur within acceptable limits. One such equation, for example is

(%S)/[S] = 1.42B – 0.13(%FeO) + 0.89 .....(16)

The equation 16 shows the beneficial influence of slag basicity and the retarding influence of FeO on sulfur distribution. A large number of such correlations are reported in the literature, but they are suitable and applicable to a local situation only.

Search Knowledge Base

Enter a phrase to search for:

Search by

Full text


The Total Materia database contains many thousands of structural steel materials across a large range of countries and standards.

Where available, full property information can be viewed for materials including chemical composition, mechanical properties, physical properties, advanced property data and much more.

Using the Advanced Search page, it is possible to search for materials by general application by using the Full Text Search function of Advanced Search.

It maybe that you need to further narrow the search criteria by using the other fields in the Advanced Search page e.g. Country/Standard.

Then click Submit.

solution img

Total Materia A list of materials will then be generated for you to choose from.

solution img

After clicking a material from the resulting list, a list of subgroups derived from standard specifications appears.

solution img

From here it is possible to view specific property data for the selected material and also to view similar and equivalent materials in our powerful cross reference tables.

For example, by clicking on the chemical composition link on the subgroup page it is possible to view chemical composition data for the material.

solution img

Other critical property data for structural steels can also be viewed including mechanical property data covering yield stress, tensile stress, elongation and impact data.

solution img

For you’re a chance to take a test drive of the Total Materia database, we invite you to join a community of over 150,000 registered users through the Total Materia Free Demo.