News

Desulphurization of Hot Metal

• satyendra
• October 16, 2013
• 15 Comments
• calcium carbide, Co-injection process, desulphurization, Desulphurizing reagents, Hot metal, Kanbara reactor, KR process, lime, magnesium, MMI process,

Desulphurization of Hot Metal

Sulphur is a desirable element in steel when good machinability is needed from the steel product. However, it is an unwanted element in most of the applications of steel since it affects the internal and surface quality of the steel products, causes hot steel brittleness, behaves as a stress raiser, and deteriorates the mechanical properties of steel. Further, it forms undesirable sulphides which promote granular weakness and cracks in steel during solidification, and lowers the melting point, inter-granular strength, and cohesion of steel. The demands on steel grades with low sulphur content is increasing at a fast rate since these grades improve material properties, lower sulphur inclusions, and have lesser cracking tendencies.

Low sulphur content in hot metal is a prerequisite for producing a steel grade with low sulphur content. Desulphurization of hot metal in the basic oxygen furnace (BOF) is not an attractive alternative because of the highly oxidizing atmosphere existing in the furnace. Further, there is additional sulphur input during steelmaking in the BOF. Sulphur removal in steelmaking becomes less efficient when it is done further down the process chain. Hence, it is important, from a process and economical point of view, to remove most of the sulphur from hot metal before it enters the BOF.

The desulphurization of hot metal provided by the blast furnace (BF) is not sufficient. The pre-treatment of hot metal by desulphurization process before the BOF is an important step to achieve low sulphur content in the hot metal. The high temperature and low oxygen potential in hot metal after tapping from the BF provides a favourable condition for hot metal desulphurization. Removal of sulphur from the hot metal is called desulphurization of hot metal.

Theoretical aspects of the process

Irrespective of where the removal of the sulphur takes place, it is based on the same chemical equations. The conditions of the individual processes only have an impact on the importance of the chemical equations. The only principle for the removal of sulphur is based on moving the dissolved sulphur from the iron to the slag, after which the slag layer is separated from the hot metal. This leads to a reaction for the sulphur transfer between the metal and slag represented by the equation [S]Fe + (O2−)slag = (S2−)slag + [O]Fe. The equilibrium constant of this equation (K1) can be written as K1 = {a[O] · aS2−} / {a[S] · aO2−} where ‘ax’ stands for the activity in steel or slag. This equation shows that for maximal sulphur removal the oxygen activity in the metal phase and the sulphur activity in the slag phase are to be as low as possible.

It is also known that an increased basicity leads to a higher sulphur capacity of the slag, which is good for desulphurization of the hot metal. The basicity is normally calculated based on the weight ratio of basic oxides (like CaO and MgO) to acid oxides (like SiO2, Al2O3 and P2O5). The basicity calculations can differ from plant to plant. It is since there is no general rule on which oxides are included. This also depends on which oxides can be detected. The slag basicity has typical values of 1 to 1.3 (BF) and 3 to 4 (BOF).

Desulphurizing reagents

Desulphurization of the hot metal can be controlled by adding reagents (through injection or mixing), such as sodium carbonate, lime, calcium carbide, and magnesium.

Sodium carbonate – Sodium carbonate, also known as soda ash was a popular agent for the desulphurization of hot metal during the 1960s. Sodium carbonate was popular because of its easy application method. The sodium carbonate was delivered in paper bags and, according to the requirements; number of bags were placed in the empty ladle bottom before pouring the hot metal. This method of desulphurization was easy to use.

However, there were considerable problems associated with the use of the sodium carbonate. Sodium carbonate has a melting temperature of around 851 deg C. It is stable until 1,200 deg C, which is lower than the temperature of the hot metal. So once sodium carbonate is in contact with the hot metal, it decomposes to Na2O and CO2 as per the equation Na2CO3 = Na2O + CO2. Na2O can be reduced by dissolved carbon into sodium (Na) gas. Both Na2O and sodium gas can react with sulphur intensively according to the reactions Na2O (l) + S = Na2S (l) + O (g), and Na (g) + S = Na2S (l). Both these reactions are reversible.

Sodium carbonate can promote separation of hot metal from slag and iron losses in slag are lower. However, presently sodium carbonate is no longer being used as desulphurization reagent. This is because it gives a violent gas evolution. The decomposition of the carbonate in contact with the hot metal produces pungent sodium oxide fumes which are harmful to the human health. Further, sodium carbonate slag is soluble in water. This results into leaching and contamination of land.

Lime – It is the cheapest hot metal desulphurization reagent and is the most applied reagent. However, it has a lower effectiveness. It also produces large quantity of slag and, since iron losses are directly proportional to the generation of slag, this adversely affects the cost. These slags have a lower viscosity than those generated with calcium carbide injection and they decompose more readily, thus they can be more easily treated without having to add fluxing agents. A further advantage is the fact that lime is not subject to dangerous goods regulations and there are no transport and handling problems.

Lime can be used in every desulphurization process from the BF to steelmaking. Lime reacts with dissolved sulphur through the reaction given by the equation CaO(s) + [S]Fe = CaS(s) + [O]Fe. The thermodynamics of this reaction, expressed by the Gibbs free energy [J mol−1]) and the equilibrium constant (log (K)), which are valid for the hot metal temperature range of 1,250 deg C to 1,450 deg C. Both the Gibbs free energy equation and the chemical equilibrium equation show that the reaction between CaO and [S] is favoured at higher temperatures. This is in accordance with practical experience.

Calcium carbide – In the 1970s, calcium carbide was introduced as a reagent for hot metal desulphurization. Initially calcium carbide mixture having around 50 % calcium carbide with 10 % calcium carbonate as a gas separator (the balance was mainly lime), was used for desulphurization of hot metal in the torpedo ladle. With the introduction of transfer ladle treatment, carbide mixture having around 70 % CaC2 with 5 % gas separators is being used since higher gas flow rates can be obtained. This has resulted into greater efficiency of calcium carbide.

The reaction of sulphur with calcium carbide takes place as per reaction CaC2(s) + [S]Fe = CaS(s) + 2C(s). It is assumed that the formed carbon does not dissolve in already carbon-saturated hot metal. The reaction of desulphurization by calcium carbide is around 7 times faster than the reaction of desulphurization by lime.  The thermodynamics of this reaction shows that the reaction is favoured at lower temperatures. This, however, is contradictory to the practical experience, where CaC2 desulphurization efficiency increases at higher temperatures.

As with the reaction with lime, this reaction is also controlled by kinetics rather than thermodynamics. Further, it is to be noted that CaC2 in practice is only 50 % to 70 % pure (the rest is mainly lime (20 % to 30 %) and carbon. These impurities have their influence on the process and can partly explain a gap between theoretical behaviour and practical experience.

Magnesium – Magnesium in powder or granular form is used to remove sulphur through a slag-liquid metal reaction done under reducing conditions, where sulphur is transferred out of the hot metal through the slag-metal interface into the slag. Slag is an ionic solution made up of oxides and fluxes.

Magnesium has a boiling point of 1,105 deg C, and hence in contact with hot metal (1,250 deg C to 1,450 deg C) it vapourizes. Magnesium gas dissolves into liquid iron, after which it reacts with the dissolved sulphur as per reaction [Mg]Fe + [S]Fe = MgS(s). The magnesium gas can also react directly with the dissolved sulphur at the bubble / metal interface, but this has only a small contribution. The reaction of desulphurization by magnesium is around 20 times faster than the reaction of desulphurization by lime and around 3 times faster than the reaction of desulphurization by calcium carbide.

From practical experience, it is seen that the magnesium desulphurization reaction proceeds better at lower temperatures. The stoichiometric consumption of magnesium which forms MgS is 0.76 kg of magnesium per kg of sulphur. The practical data show that, at lower temperatures of the hot metal, lesser magnesium is needed to remove 1 kg of dissolved sulphur. The thermodynamics support the observation that lower temperatures have a positive effect on the desulphurization efficiency of the magnesium.

Resulphurization – A disadvantage of desulphurization with magnesium is the so called resulphurization, the net sulphur transfer from the slag back to the metal. The MgS in the slag reacts with oxygen from the air, or from other sources, forming MgO and unbounded sulphur a per reaction MgS(s) + 1/2 O2(g) = MgO(s) + [S]Fe. To avoid the resulphurization, the sulphur is to be captured in a more stable compound. CaS is more stable than MgS, so that by adding calcium (typically in the form of lime) the resulphurization can be prevented. The reaction which describes the effect of adding lime is MgS(s) + CaO(s) = MgO(s) + (CaS)slag. This reaction still takes place even at higher hot metal temperatures of 1,400 deg C. However, higher temperatures do not only have a negative effect on the desulphurization by magnesium, but also on the stabilization reaction. For magnesium desulphurization, lower temperatures are favourable.

The CaS formed in reactions during the desulphurization with lime and calcium carbide remain attached to the reagent particle, which rises to the slag layer due to the upward pressure within a minute. Reaction during the desulphurization with magnesium is a homogeneous reaction, which means that the magnesium first needs to dissolve in the hot metal before it reacts with the sulphur. The formed MgS hence starts as a single molecule and takes much longer to cluster and rise to the slag (around 5 minutes tp 8 minutes). In practice this means that for effective desulphurization, the skimming cannot be stopped earlier than 8 minutes after the last magnesium particles are injected.

Kinetics of desulphurization

Desulphurization by CaO or CaC2 is in reality controlled by kinetics rather than thermodynamics. When CaO reacts with sulphur, CaS is formed. This CaS forms a layer around the CaO particle, through which other dissolved sulphur atoms need to permeate before they can react with CaO. Since also oxygen is formed in this reaction, the oxygen activity increases around the CaO particle. This oxygen reacts with either carbon (forming CO) or silicon, which leads to the formation of 2CaO.SiO2 as per reaction 2CaO + 2[O]Fe + [Si]Fe = 2CaO · SiO2(s). This 2CaO.SiO2 contributes to the non-reactive shell around the CaO, decreasing its desulphurization efficiency. However, with small CaO particles (less than 50 micrometers) not enough oxygen is created through desulphurization reactions to initiate the reaction for the formation of 2CaO.SiO2. For the reaction between CaC2 and sulphur, the non-reactive shell not only consists of CaS, but also of a graphite layer. This retards the desulphurization even further.

The kinetics of magnesium desulphurization causes some discussion among the experts in the field. Irons and Guthrie claim that 90 % of the magnesium first dissolves before it reacts with [S] and less than 10 % of the magnesium-desulphurization is heterogeneous (magnesium gas at the bubble / metal interface). The formed MgS precipitates on CaO particles. Lindstrom and Yang et al. conclude from their experiments (on lab scale) that more than 90 % of the magnesium desulphurization is heterogeneous and that only a little magnesium first dissolves before it reacts. Visser discussed both views and concluded, also based on plant data, that the kinetic model of Irons and Guthrie predicts the reality on plant scale best, and hence that the route through the dissolved magnesium is dominant.

Desulphurization process

Hot metal which leaves the BF typically contains 0.03 % sulphur, but the demand for many of the steel grades can be as low as 0.001 % sulphur. Hence, the hot metal desulphurization is being used by the steel plants worldwide since it is more efficient and cost effective process to desulphurize before the BOF steelmaking.

A number of processes have been developed for the external desulphurization of hot metal but all of them have the basic requirement of a reagent and a method of mixing. The difference between the processes is the properties of the reagent used for the process, its effectiveness to remove sulphur, and the effectiveness of the mixing method to get the reagent into solution. Also the effectiveness of hot metal desulphurization is inversely proportional to the desulphurization reagent injection rate.

Dip lance or impeller is used for effective mixing and reliable desulphurization of the hot metal. The reagent is fed into the hot metal ladle either by pneumatic means or by gravity with high dosing precision through a dispensing vessel or a silo. For each reagent, one separate dispensing vessel is used. All the vessels are identical. Nitrogen gas is normally used as a carrier gas for the desulphurization reagent.  In case of pneumatic conveying, the reagent transfer in the injection line is under dense flow conditions.  The dense flow conditions maximize reagent delivery as well as reduce abrasion wear of injection lines.

The intimate mixing of the desulphurization reagent with the hot metal removes the sulphur in the hot metal by chemical reaction and converts it into the slag.  Sulphur rich slag generated during the process is removed immediately after completion of the reagent reaction.  The most common method is to tilt the ladle and rake the slag off with the help of a slag skimmer.

For the control of the desulphurization process, mathematical process control and flexible control of the desulphurization plant is adopted. This combination provides a range of possible process variations. The various processes of desulphurization are described below.

Torpedo desulphurization – Initially (in the 1960s and 1970s) hot metal desulphurization took place in the torpedo cars which transported the hot metal from the BF to the steel melting shop. Typical reagents used were calcium carbide, soda ash, and blends of magnesium and lime. During torpedo desulphurization, the reagent is injected into the hot metal through a lance. Nitrogen is used as a carrier gas. The reagent reacts with the sulphur in the hot metal and the sulphides CaS or Na2S ascend to the slag layer. This slag is then raked off with a skimmer. Fig 1 shows schematic view hot metal desulphurization in torpedo ladle.

Fig 1 Schematic view hot metal desulphurization in torpedo ladle

The shape of the torpedo is designed for temperature preservation and not as a metallurgical reactor vessel. The hot metal bath is not very deep (1 m to 2 m), so the reagent particles (which have a lower density than the hot metal) quickly rise to the top. Hence, the reagents only have a short contact time with the hot metal. Reagent mixing is poor, which means that both far ends of the torpedo ladle are not reached by the reagents. Finally a torpedo has only a small opening at the top, which makes it difficult to rake off the slag. This leads to resulphurization through the remaining slag and high iron losses. Because of these drawbacks, desulphurization in the torpedo ladle was replaced by ladle desulphurization.

Kanbara reactor process – The Kanbara reactor (KR) process was developed in 1965 in Hirohata in Japan by Nippon Steel. The low availability of magnesium in Japan was the reason to look for alternatives. The KR process uses relatively cheap coarse lime. Sometimes CaF2 (around 5 % to 10 % of the flow) and / or Al2O3 / calcium carbide is added on top of the hot metal ladle during the first few minutes of the process.  The reagent is injected into the hot metal through a rotating lance together with a carrier gas (normally nitrogen) or, alternatively, the reagent is added from the top.

The stirring lance is equipped with four massive rotor blades, which create turbulence in the hot metal. Due to the turbulence, the bubble size of the transport gas is smaller and the residence time of the lime in the hot metal is longer than during static injection. The increased residence time is of major importance to the process, since lime is a relatively slow reagent.

Typically reagent in the quantity of 5 kg/tHM (kilograms per ton hot metal) to 15 kg/tHM is added. An immersed impellor (at one-third of the bath depth) is used to mix the reagent with the hot metal. The mixing is needed since the reaction between lime and sulphur is relatively slow, so that the contact time needs to be increased. The impellor has a typical rotational speed of 60 rpm (revolutions per minute) to 120 rpm. The average life of the impellor is around 150 heats.

The stirring takes around 5 minutes to 15 minutes after which the impellor is lifted again and the bath is allowed to rest for another 5 minutes to 10 minutes. This is necessary since the slag and the formed CaS need time to ascend to the top. After this the slag layer is skimmed off, which takes around 10 minutes to 15 minutes.

The lime in the KR process is used more efficiently, which means less lime is needed and lime of a lower (thus cheaper) quality can be used. The stirring however also means that the hot metal needs to be skimmed prior to desulphurization as well, in order to remove high SiO2 containing slag from the BF, which decreases the efficiency of the lime. Also the impellor and refractory of the ladle suffer from increased wear. Finally the created turbulence needs a larger free board (typically 1 m more than the co-injection process in the hot metal ladle. Fig 2 gives the schematic view of the Kanbara reactor process for desulphurization.

Fig 2 Schematic view of Kanbara reactor process for desulphurization

Around 1970 a similar process called Rheinstahl-Ruhrer was developed in Germany. It was soon abandoned due to the large slag volumes created. The KR process is widely used. With the KR process, sulphur concentrations below 0.001 % (10 ppm) have been reported.

Magnesium mono-injection process – Magnesium mono-injection (MMI) process, also referred to as the Ukraina-Desmag process is a hot metal desulphurization process which uses only magnesium as a reagent. The process was developed between 1969 and 1971 by the Ukrainian Academy of Sciences. The process is mainly used in Russia and Ukraine, as well as in some plants in China and Taiwan. Tests with this method in North America failed as a result of the violence of the process. In Finland the MMI process was abandoned after some years due to unreliability of the process.

In MMI process, the magnesium (normally salt coated) is injected into the hot metal under pressure through a submerged bell shaped refractory coated lance. The bell at the end of the lance is used as an evaporation chamber for the magnesium to stabilize the process. Nitrogen is most frequently used as a carrier gas. However, there are also plants, with larger ladle sizes, where a straight lance without an evaporation chamber is used. In both cases the evaporation of the magnesium causes enough turbulence to ensure a good reagent distribution in the hot metal.

Turbulence is created by evaporation of the magnesium powder. At higher injection rates the turbulence can become a problem, increasing the iron loss by splashing. Hence, the evaporation chamber at the end of the lance is used to allow the magnesium to evaporate earlier, thus reducing the turbulence. Since magnesium reacts with sulphur much faster than lime and calcium carbide, MMI process is a very fast desulphurization process, in which very little slag is created.

Supporters of the MMI process frequently state that lime does not add significantly to the desulphurization efficiency of magnesium. This is correct, since magnesium is around 20 time faster reagent than lime. An equal amount of lime contributes for less than 5 % to the desulphurization. In contrary it is claimed that lime actually decreases the efficiency of magnesium especially in cases where the lime is not very well burnt. This leads to the reactions CaCO3 = CaO + O(Fe) + CO, and O(Fe) + Mg = MgO.

When only magnesium is used as a reagent, resulphurization is a major problem. It is the major disadvantage of MMI process. Another problem is the severe resulphurization reaction. When no lime is used to prevent this, the sulphur concentration of the hot metal increases considerably before charging in the BOF. One more issue with the MMI process is the thin slag layer compared with KR process and co-injection process. This leads to an increased iron entrainment loss during skimming. In order to stabilize the slag and retard the resulphurization, in most steel plants lime, flux and / or coagulant is added on top of the slag. Fig 3 gives a schematic view of the magnesium mono-injection process.

Fig 3 Schematic view of magnesium mono-injection process

Co-injection process – Co-injection is a hot metal desulphurization process in which both magnesium and fluidized lime is injected into the hot metal. The co-injection of magnesium and lime is a method which combines the advantages of both reagents. Magnesium enables fast desulphurization while lime allows for low final sulphur concentrations. In the past, the lime was sometimes replaced by calcium carbide, which is more efficient, but due to safety issues this option is hardly used in steel plants anymore. Co-injection desulphurization is used worldwide and the process is considered as a standard practice. Multi-injection, which uses all three reagents, is a variation of his process.

In the co-injection process, through a submerged refractory coated lance the reagents are injected at the bottom of the hot metal ladle. An inert carrier gas (normally nitrogen) transports the reagents through the injection line and creates enough turbulence in the ladle for proper mixing. The mixing of the reagents takes place in the injection line, which makes it possible to change the ratio of the reagents during the process. When the reagents react with sulphur, the products (MgS and CaS) ascend to the slag layer, where it is removed with a skimmer.

In the co-injection process, the reagents are stored in different dispensers and are only mixed inside the injection line. The reagents are injected in the hot metal through a straight lance with one opening at the bottom or two or four openings at the side. An inert transport gas (normally nitrogen) is used to ensure a smooth injection. The turbulence in the hot metal is created by the carrier gas and the magnesium which evaporates. This turbulence ensures sufficient distribution of reagent throughout the ladle. An advantage of co-injection process is that the ratio between magnesium and lime can be modified if the situation needs or allows it. For example, if more time is available, more lime and less magnesium can be injected, which makes the process more flexible and more cost effective.

Co-injection process combines the advantages of magnesium (faster process) and lime / calcium carbide (deep desulphurisation). Most sulphur initially reacts with magnesium to form MgS. The lime mostly prevents the resulphurization reaction. With magnesium / lime co-injection, sulphur concentrations below 0.001 % (10 ppm) have been reported.  Fig 4 gives a schematic view of the co-injection process.

Fig 4 Schematic view of the co-injection process of desulphurization 

Technical comparison of different desulphurization processes

The aforementioned desulphurization processes have both the strong and the weak points. It depends on the specific situations and requirements of a steel melting shop, which points put more weight on the balance. However, the three processes can be compared for a few important technical and metallurgical issues of hot metal desulphurization as given below.

Process time – The process time depends on how fast the reagents can react with sulphur. Since magnesium is a much faster reagent than lime, the MMI process and co-injection process are faster processes than the KR process. The KR process has an extra time delay, since skimming prior to injection is frequently needed, in order to prevent the return of impurities from the BF slag into the hot metal during the mixing. Normally, the KR process takes on an average 10 % to 20 % more time than the co-injection process..

The MMI process normally has a shorter injection time than co-injection process (10 % to 20 % faster). However, the gained time is limited for the MMI process, since in both processes skimming can only be ended when all MgS particles reach the slag layer (which can take 8 minutes). Still in general the MMI process is faster than co-injection with magnesium and lime (around 5 %, co-injection with magnesium and CaC2 is in general even faster than MMI). The KR process is the most time consuming process.

Iron loss during skimming – Iron loss during skimming is a major problem. Iron is lost in two different ways. During the slag forming, iron droplets get trapped in the slag, hence forming an emulsion with the slag. When the slag is skimmed, the trapped iron is lost with it. It is called emulsion loss. In general, around 50 % of the slag is iron in emulsion. This means that emulsion loss can be minimized by reducing the total amount of slag. The other major contribution to iron loss is the entrainment loss. When slag is raked off, some iron can come with it. The entrainment loss can be reduced by more careful skimming or by a thicker more viscous slag, which is easier to skim.

Due to the high amount of slag created in the KR process and the required extra skimming prior to the desulphurization, the total iron loss is normally 2 times to 3 times higher than for the co-injection process. The MMI process has the lowest iron emulsion loss, since only little slag is created (around 7 times less than co-injection process). However, due to the lower basicity, MMI process slag contains more iron in emulsion than slag which contains calcium. The entrainment loss of iron for the MMI process is higher than for co-injection process or the KR process, since skimming is more difficult due to the small slag layer and skimming needs to be done more thoroughly due to the higher sulphur concentration of the slag and the high risk of resulphurization later in the process. Normally, the iron loss of the MMI process is similar to co-injection process and it is around 1 % total iron loss. For the KR process the total iron loss is in the range of 2 % to 3 %.

Refractory and lance wear – Wear of the refractory and the lance is mainly caused by the high temperatures and corrosive composition of the hot metal and the slag. For the KR process, the turbulence created by the rotating impellor is a major contributor to the wear. Also the impellor itself is more vulnerable to wear, since the blades can even break off. Decrease of the blades of the impellors then leads to less turbulence and thus lesser efficiency. Because of the wearing problems, a number of have been done on the refractory of especially KR process. The MMI process has less wearing problems than the KR process, due to less turbulence. However, since magnesium is used instead of lime, the basicity of the slag is lower, leading to increased corrosion wear. The co-injection process has less turbulence than the MMI process and a higher basicity in the slag, which explains why the refractory and lance suffer the least from wearing in this process. However, wearing remains also a concern for the co-injection process.

Temperature loss – During the desulphurization process, the hot metal loses temperature. The temperature of the hot metal when it is charged to the BOF has an influence on the amount of coolant (scrap) which can be charged or on the blowing time of the BOF. The colder the hot metal, the lesser amount of the scrap can be added or the longer is the blowing time in the BOF. When the hot metal temperature is already too low before desulphurization, the desulphurization process has to be by-passed completely. This happens more frequently for the KR process. It depends on the situation if temperature loss in the desulphurization process is a problem or not. In countries where hot metal is more expensive than scrap, an increased scrap to hot metal ratio is very beneficial.

Higher temperature losses are caused by longer process times, higher turbulence, lesser slag (slag acts as an isolation material), and the use of reagents which leads to lesser heat. Magnesium causes an exothermic reaction in the hot metal and lime does not. The KR process involves longer process times, high turbulence, and no major exothermic reactions, which lead to a temperature loss which is on an average three times higher than for co-injection process or MMI process. As stated above, injection during co-injection process takes in general longer time than injection for MMI process. On the other hand co-injection process is a less turbulent process and has a thicker isolating slag layer. Hence, the temperature losses for co-injection process and MMI process are in general comparable.

Low sulphur – In the present day scenario, BOF can need hot metal with a sulphur concentration of only 10 ppm to 20 ppm. Due to magnesium-sulphur equilibrium and the resulting resulphurization, only magnesium as a reagent is not sufficient to reach these low sulphur concentrations. There are claims of low sulphur concentrations with the use of only magnesium but only when measurements are taken directly after injection (so before resulphurization shows this effect). In practice hot metal which is desulphurized by MMI process never has a sulphur concentration below 0.006 % when it is charged to the BOF. This can be compensated a little by adding fluxes from the top.

Co-injection process is capable of reaching stable low sulphur concentrations in hot metal. However, since magnesium is not efficient anymore at low sulphur concentrations, only the injected lime has a contribution to the desulphurization as soon as the low sulphur concentrations are reached. Due to the lower turbulence during the co-injection process, reaching the desired low sulphur concentration takes longer time and costs more reagent than for the KR process. In case consistent low sulphur concentrations are needed, the KR process is most suitable.

Flexibility – A desulphurization plant which can respond to the changing situations, like inadequate reagents or lack of time, is beneficial to the overall flexibility of the BOF plant. The KR process is not flexible with regards to the process time, since the optimal lime flow and the stirring speed are already applied. The KR process can only reduce the process time by releasing the initial sulphur aim. Availability of reagents is not a problem under normal conditions for the KR process. Magnesium for the MMI process can become scarce though, leading to a sudden increase in the operational costs of the system or even a production stop. The co-injection system has a high flexibility for both process times and reagent scarceness, since both rate and ratio can be adjusted. Even CaC2 can be injected as an alternative reagent.

Safety – Magnesium is a hazardous flammable compound. Spilled magnesium can catch fire and is not easy to extinguish. In contact with water, magnesium can form hydrogen which is an explosive gas. Magnesium for desulphurization is hence coated, in order to retard its hazards. However, coated magnesium remains a more hazardous reagent than (burnt) lime. In the MMI process (and sometimes in the KR process as well) also frequently CaF2 is added to stabilize the process. When CaF2 reacts, the highly toxic gas fluorine is generated. This, together with the violence during injection (due to the vapourizing and oxidizing of the magnesium), makes the MMI process a relatively unsafe process for human health and the environment. This has also been one of the reasons why the MMI process has been abandoned in some of the countries. Co-injection process is also considered as less safe than the KR process, provided no CaF2 is used in the KR process, due to the use of magnesium. Because of the safety reasons, CaC2 (which can form the explosive gas acetylene when in contact with water) is only very seldom used for new co-injection process installations. When CaF2 is used in the KR process, co-injection (using lime) can even be considered as a safer option.

Economical comparison of different desulphurization processes

The operational cost of the desulfurization plant is considered as the most important factor. The most significant factors which contribute to the operational cost are (i) hot metal loss, (ii) cost of reagent, (iii) refractory and lance wear, and (iv) temperature loss. The comparison of different desulphurization processes with respect to the capital cost and the major operational costs is given below.

Capital cost – The capital cost of the KR process is higher than the co-injection process and the MMI process due to the large structure and support for the lance and motor system. The capital cost of the MMI process is probably a bit lower than the co-injection process since only one dispenser is needed.

Hot metal loss – Iron loss is the most important contributor to the costs of a desulphurization process. The hot metal loss for the MMI process and the co-injection process is around 1 %. For the KR process the hot metal loss is in the range of 2 % to 3 %. In general, the costs for hot metal loss can be reduced in case of the recycling treatment of the slag.

Cost of reagent – Magnesium is a costly reagent than the lime. Further, the KR process uses lime of lower quality while the lime needed for the co-injection process is of higher quality. It is estimated that for the MMI process and the co-injection process an amount of 0.5 kg/tHM (kilogram per ton of hot metal) magnesium is injected. With average magnesium to lime ratio of 1:4, the co-injection process also needs 2 kg/tHM of lime. In general the KR process injects 10 kg/tHM lime. For KR process and MMI process, frequently flux and / or coagulant is added as well. The quantity is around 500 kg/heat.

Refractory and lance wear – The most important equipment wear for the comparison is that of the lances and the ladle refractory. Difference in the maintenance cost for the remaining equipment is normally negligible. The complete lance of the KR process treats on an average 30,000 tons of hot metal (150 heats of 200 t). The lance of the MMI process treats on average 10,000 tons of hot metal (50 heats of 200 t). The lance of the co-injection process also treats on an average 10,000 tons of hot metal (50 heats of 200 t). The average lifetimes of the lances also include the fact that some lances break or block during their first heat.

MMI process and KR process need higher freeboard (at least 500 mm), hence the quantity of the needed refractories is around 10 % higher. For the KR process, the ladle needs replacement of the refractories after handling of an average 18,000 tons of hot metal (90 heats of 200 t). The refractory of a ladle in a MMI process needs a replacement after handling of on an average of 24,000 t tons of hot metal (120 heats of 200 t). For the co-injection process, the refractory of a ladle needs to be replaced after handling of an average 36,000 tons of hot metal (180 heats of 200 t).

Temperature loss – Temperature loss for the hot metal means that less scrap or more hot metal can be added to the BOF. Since there is a cost differential which exists between the cost of the hot metal and the cost of the scrap, it affects the cost of producing steel. The temperature loss for co-injection process and MMI process has an average temperature loss of 10 deg C per heat. KR process has a temperature loss which is on average three times higher at 30 deg C per heat.

There are certain other costs which are made such as the use of electric power and the inert (nitrogen) gas. The KR process needs much more electric power than the other processes, while MMI process needs 5 to 6 times more nitrogen than co-injection process. The influence of spare parts costs on various processes is negligible.

Considering the performance and the operational costs, the KR process has an overall advantage over other processes for the hot metal desulphurization when the main target is to produce low sulphur steel and when process times, temperature loss, and the hot metal loss are not the issues. When no steel grades with very low sulphur concentrations need to be made, resulphurization is not considered a problem and short processing times are needed, the MMI process is the most effective process. Co-injection with magnesium and lime is the most flexible and reliable option for the desulphurization. For a wide range of steel grades (including both low sulphur grades and normal grades), co-injection process is the most effective and economically most attractive process. It depends on the situations, focus, and targets of the steel melting shop which decides the optimal hot metal desulphurization process.