A literature study on measurement methods for calculating the FeO activity in slag

INOM

EXAMENSARBETE TEKNIK, GRUNDNIVÅ, 15 HP

STOCKHOLM SVERIGE 2020,

A literature study on measurement methods for calculating the FeO activity in slag

REBECCA ROSÉN

KTH

SKOLAN FÖR INDUSTRIELL TEKNIK OCH MANAGEMENT

Abstract

During the time we live in, climate change is more important than ever. HYBRIT, formed by SSAB, LKAB, and Vattenfall, is an initiative set out to make the steel production fossil free, by developing a novel process that produce direct reduced iron (DRI) using hydrogen as reagent. Electric arc furnaces (EAF) will then be employed to melt the DRI. For SSAB, it is important to obtain a satisfied dephosphorisation in the EAFs for the production of high strength steel. The composition of the DRI, has a big impact on the dephosphorisation and it is important that the slag has an optimal basicity and FeO content. To better understand the impact that FeO has on dephosphorisation, experimental data has to be collected and the FeO activity calculated in different slag systems. Activity measurements of FeO is rather complex. In this thesis, three methods were looked at: experimental determination, electrochemical methods and equilibrating the slag. Equilibrating the slag turned out to be the best suitable method for calculating the FeO activity and is recommended to be evaluated further.

Sammanfattning

I den tid vi lever i nu är frågan om klimatförändringarna viktigare än någonsin. HYBRIT, bildat av SSAB, LKAB och Vattenfall, är ett initiativ som syftar till att göra stålproduktionen fossilfri genom att utveckla en ny process som producerar direkt reducerat järn (DRI) med väte som reagens. Elektriska bågugnar (EAF) kommer sedan att användas för att smälta DRI. För SSAB är det viktigt att få en bra fosforrening i EAF:erna för produktion av höghållfast stål. DRI:s sammansättning har stor inverkan på fosforreningen och det är viktigt att slaggen har en optimal basicitet och FeO-innehåll. För att bättre förstå vilken inverkan FeO har på fosforreningen måste experimentella data samlas in och FeO-aktiviteten beräknas i olika slaggsystem.

Aktivitetsmätningar av FeO är ganska komplicerade. I denna avhandling undersöktes tre metoder: experimentell bestämning, elektrokemiska metoder och jämvikt av slaggen. Jämvikt av slaggen visade sig vara den mest lämpliga metoden för beräkning av FeO-aktiviteten och rekommenderas att utvärderas ytterligare.

Table of contents

1. Introduction ...1

1.1 Problem ...1

1.2 Aim ...1

1.3 Limitations ...1

2. Steel production ...2

2.1 Blast furnace ...3

2.2 DRI – Direct Reduced Iron ...4

2.3 Electric arc furnace – EAF ...5

2.4 Dephosphorisation...5

2.5 The HYBRIT-process...7

3. Method ...8

4. Result...9

4.1 Experimental determination of activities ...9

4.2 Electrochemical methods ...9

4.3 Equilibrating the slag... 10

5. Discussion... 12

5.1 Dephosphorisation with DRI ... 12

5.2 Environmental aspects ... 12

5.3 Measuring methods of FeO activity ... 13

6. Conclusions ... 14

7. References ... 15

1

1. Introduction

1.1 Problem

Climate change is on everyone’s minds these days. Companies and countries are looking for different ways to reduce their effect on the environment and especially their carbon dioxide, CO2, emissions. The reduction of the CO2 emissions is an important step on the way to reach the UN’s 1.5-degree target. The biggest source of CO2 emission in Sweden is the steel industry, with 10 % of Sweden’s overall CO2 emissions. To make sure Sweden meets its climate goals of zero CO2 emission by the year 2045, three companies; SSAB, LKAB, and Vattenfall, joined forces and started HYBRIT. The goal with HYBRIT is to direct reduce iron with hydrogen, instead of charging iron oxide together with coke, using fossil free pellet and electricity, making the CO2 emission zero. Fossil free sponge iron, or DRI, is a completely new product and that is why it is desirable to know the most efficient way to melt the sponge iron, but also to manufacture a new pellet suitable for the process to minimise additives in the process. The impurities, mainly phosphorus, of the pellet will also be of great importance. Phosphorus is unwanted due to causing embrittlement in steel, and since DRI makes dephosphorisation harder, due to composition, it is desirable to make the dephosphorisation as efficient as possible. To do so more knowledge of how FeO and the FeO activity affects the dephosphorisation is needed.

At low basicity, data for the activity of FeO is missing.

1.2 Aim

The aim of the project is to look at the dephosphorisation of iron and the effect the FeO activity has on it, ultimately finding different measuring methods for calculating the activity of FeO and comparing them and give suggestions on which method that is most appropriate for future experiments.

1.3 Limitations

Because developments and recesses are limited by time, this project will only look at three methods for measuring the activity of FeO.

1. Experimental determination of activities 2. Electrochemical methods

3. Equilibrating the slag

2

2. Steel production

Back in 2017, Sweden established a climate law, a climate goal, and a climate policy council.

The long-term goal is to have net zero greenhouse gas emissions by 2045 [1]. This, in itself, is a part of EU’s climate goal for 2050, and thus making Europe the world’s first climate neutral continent by the same year [2]. Another part is to keep to the Paris Agreement, made by the UN, and keep global warming to a maximum of 2 degrees, but preferably below 1,5 degrees [3]. To meet these goals, the world needs to revolutionise its way to deal with the greenhouse gasses. Some scientists even say that to meet these goals we can not only minimise the emissions but we also need to store the greenhouse gasses under ground, from where they once came, ironically [4].

This is where HYBRIT comes into the mix. The steel industry stands for approximately 7 % of the global CO2 emissions, or two tonnes of CO2 emissions created for every tonne steel produced [5]. That means, if HYBRIT is successful, a big part of the emissions will be gone.

The alternative to the blast furnace today is scrap based steelmaking in an electric arc furnace, EAF. What has to be taken into consideration is the world’s consumption of steel. According to a forecast by HYBRIT, the demand of steel will increase, both recycled and ore based, as seen in figure 1, at least to the year of 2050 [6].

Figure 1: The demand of steel in 2016 and prediction for 2050 [6].

To be able to keep up with this demand while still maintaining the goal of fossil free steel there has to be a revolutionary change in how steel is made, and HYBRIT is just that. There are two

3

main processes for ore-based steelmaking: in blast furnace and direct reduced iron, DRI, in an electric arc furnace. A schematic overview can be seen in figure 2:

Figure 2: A schematic overview of how the steelmaking process works today [6].

2.1 Blast furnace

The blast furnace is where iron oxide is reduced with carbon in several steps at elevated temperatures to form pig iron [5]. It is a counter-current gas/solids reactor where charged material, in this case coke, iron ore and fluxes, reacts with hot gases that is rising up in the furnace [7]. The reduction usually starts with Fe2O3, followed by the reduction of Fe3O4 and then FeO, which is then reduced to Fe. Since both the iron oxide and coke are in solid phase, it needs CO or CO2 gas to react with for the reduction to happen [8]. It all starts with that CO reacts with O, as in equation 1:

𝐶𝑂 +¹

2𝑂₂ = 𝐶𝑂₂ 𝑒𝑞. 1

4

∆𝐻₂₉₈= −283.5 [𝑘𝐽/𝑚𝑜𝑙]

After that, Boudouard’s reaction happens, as seen in equation 2 [8]:

𝐶 + 𝐶𝑂₂ → 2𝐶𝑂 𝑒𝑞. 2

∆𝐻₂₉₈ = 172.5 [𝑘𝐽/𝑚𝑜𝑙]

Then, the reduction of Fe2O3 and Fe3O4 takes place, equation 3, 4 and 5 [5] [9] [10]:

3𝐹𝑒₂𝑂₃(𝑠) + 𝐶𝑂(𝑔) ↔ 2𝐹𝑒₃𝑂₄(𝑠) + 𝐶𝑂₂(𝑔) 𝑒𝑞. 3

∆𝐻₂₉₈^° = −7.8 𝑘𝐽/𝑚𝑜𝑙𝐹𝑒

2𝐹𝑒₃𝑂₄(𝑠) + 𝐶𝑂 ↔ 6𝐹𝑒𝑂(𝑠) + 𝐶𝑂₂(𝑔) 𝑒𝑞. 4

∆𝐻₂₉₈^° = 11.2 𝑘𝐽/𝑚𝑜𝑙𝐹𝑒

6𝐹𝑒𝑂(𝑠) + 6𝐶𝑂(𝑔) ↔ 6𝐹𝑒(𝑙) + 6𝐶𝑂₂(𝑔) 𝑒𝑞. 5

∆𝐻₂₉₈^° = −15.7 𝑘𝐽/𝑚𝑜𝑙𝐹𝑒𝑙

The hydrogen from the moister reacts with itself and becomes hydrogen gas, but the reduction of the iron oxides can also be made with hydrogen. Instead of CO2 gas, H2O is produced [8], as seen in equation 6 [6]:

𝐹𝑒₂𝑂₃+ 3𝐻₂ = 2𝐹𝑒 + 3𝐻₂𝑂 𝑒𝑞. 6

2.2 DRI – Direct Reduced Iron

Direct reduced iron is a product of a reduction process where iron ore is reduced by natural gas in a shaft oven. It too uses the counter-current principal where iron pellet is charged at the top of the shaft and reduced by the reduction gas [6]. The reduction gas used in this case is natural gas that has been cracked to primarily hydrogen, H2, and carbon monoxide, CO, [5]. The reduction also uses equation 1 – 5 to reduce the iron and the temperature is well below the melting point of iron [6]. One step in making the DRI fossil free is to use hydrogen instead of the natural gas, the reactions then are:

3𝐹𝑒₂𝑂₃(𝑠) + 𝐻₂(𝑔) ↔ 2𝐹𝑒₃𝑂₄(𝑠) + 𝐻₂𝑂(𝑔) 𝑒𝑞. 7

∆𝐻_{843 𝐾}⁰ = 6.440 [𝑘𝐽]

𝐹𝑒₃𝑂₄(𝑠)+ 4𝐻₂(𝑔)↔ 3𝐹𝑒(𝑠)+ 4𝐻₂𝑂(𝑔) 𝑒𝑞. 8

∆𝐻_{843 𝐾}⁰ = 101.482 [𝑘𝐽]

5

To look further at the fossil free route, a pellet developed especially for direct reduction needs to be optimized to match slag requirements whilst still maintaining a low concentration of gangue [5].

2.3 Electric arc furnace – EAF

The electric arc furnace is basically a big melt oven. Today it is mostly used for scrap melting.

Most of the refining is done after in the ladle metallurgy [6]. In the furnace there are three big graphite electrodes and it is the electric arcs between the electrodes and the scrap in the furnace that generates the heat that melts the scrap [6] [8]. The furnace is charged with big buckets and it could be either scrap and/or pig iron or DRI. When charging with DRI, it could be a continues flow. If the charge is mostly or 100 % DRI, the buckets could be completely replaced with a continues flow and high amounts of DRI (>35 %) reduces the risk of unmelted areas in the furnace.

The slag plays an important part in the electric arc furnace. It is a foaming slag which reduces the heat loss to the surroundings, protects the electrodes, reduces the radiation of the arcs onto the refractory walls and transports impurities, like phosphorus, from the melt to the slag [5] [6]

[8]. The foaming slag is made when oxygen is injected into the melt, oxidising Fe to FeOx, which is transported to the slag, where it is reduced by injected coal making CO(g). When EAF is combined with DRI, it should not only be seen as a way to melt, it could be used for steel refinement also. The electric arc furnace uses a lot of energy, with 500-700 kWh/tonne steel. It takes 2.5 times more energy to make a steel product from iron ore than from scraps. That shows how important it is to recycle steel [5] [6] [8].

2.4 Dephosphorisation

After the blast furnace or EAF, the iron goes through a refining step. The oxidising process of removing phosphorus by oxidation, dephosphorisation, occurs during the refining process, that is essentially the production of steel from pig iron [8] [9]. The removal of carbon, silicon, phosphorus, and manganese is made with a basic slag rich in oxygen and lime. Important slag components used for dephosphorisation of steel are CaO, MgO, SiO2 and FeO. Oxygen is added to the slag as iron oxide, FeO. What favours dephosphorisation is a high content of FeO in the slag, which corresponds to highly oxidized metal. But CaO has a higher dephosphorising power than FeO, and as a result of raising the FeO in the slag, there is an increase in the oxidation level of the metal which corresponds to a dilution of CaO, and therefore to a decrease in the dephosphorising power of the slag. In the dephosphorisation, iron oxide plays a double role.

6

The oxidising effect is dominant up to 16-20% FeO and the partition coefficient, tending to transfer phosphorus to the slag, increases with the FeO. Above 20% FeO, the partition coefficient decreases [9]. Phosphorus reacts with oxygen to make P2O5, as in equation 9 [5]:

𝑃₂(𝑔) +⁵

2𝑂₂(𝑔) ↔ 𝑃₂𝑂₅(𝑙) 𝑒𝑞. 9

The phosphorus in liquid metal comes from the iron ore, so in a process with hydrogen reduced DRI it is really important to have a good dephosphorisation [11]. Phosphorus is reduced from compound and dissolved by the iron and 1323℃ is the inversion temperature for the equation 10 [9]:

𝑃₂𝑂₅+ 5𝐹𝑒 ↔ 5𝐹𝑒𝑂 + 2𝑃 𝑒𝑞. 10

The basicity of the slag plays an important role in dephosphorisation, especially when DRI is used in EAF. The basicity is defined as:

𝐵 = ^{%𝐶𝑎𝑂}

%𝑆𝑖𝑂2 𝑒𝑞. 11 And the slag basicity is the driving force of dephosphorisation reaction based on equation 12:

1

2𝑃₂(𝑔) +³

2(𝑂²⁻) +⁵

4𝑂₂(𝑔) = (𝑃𝑂₄³⁻) 𝑒𝑞. 12 Equation 13 points to that dephosphorisation is favoured by a high oxygen potential and a high activity of O^2- anion, or slag basicity [11]:

𝐾₂ = ^{(𝑎𝑃𝑂43−)}

𝑎_[𝑃]𝑝_𝑂2^5⁄4(𝑎_𝑂2−)^3⁄2

𝑒𝑞. 13

The basicity decreases with increasing DRI content. Due to that, SiO2 concentration in the slag is proportional to DRI addition. An increase in use of DRI in EAF to make high quality steel would also increase the risk of more phosphorus in the steel. The dephosphorisation ability strongly depends on the CaO activity and FeO content in a CaO-based slag. The degree of dephosphorisation is strongly dependent of the basicity, and therefore on the DRI mixing ratio.

For dephosphorisation with DRI, slagmaking due to gangue oxides is important because if the DRI contains SiO2, which is an acidic component, the thermodynamic driving force to remove phosphorus is decreased, resulting in suppression of dephosphorisation efficiency. That is why it is important to increase basicity when using DRI in EAF to improve the dephosphorisation efficiency [12].

7

2.5 The HYBRIT-process

The concept behind HYBRIT is to replace coal in the steelmaking process with hydrogen and fossil free electricity, making the whole steel production fossil free. Sweden has good conditions for making this possible due to the fact that Sweden has great access to fossil free electricity, Europe’s biggest iron ore mine and multiple places where steel is produced in big scale.

For this to happen there needs to be a radical change in the industry [6]. That is why SSAB, LKAB, and Vattenfall announced that they together will have the primary goal of replacing carbon as the main reducing agent [5]. The desired production process can be seen in figure 3.

Figure 3: HYBRIT-process from iron ore to finished steel [6].

The pellet will be produced with fossil free heat sources and the characteristics of the pellet will be carefully adjusted to fit the process of hydrogen reduction. The biggest challenge is to find the best renewable fuels. In the ironmaking, a shaft oven will be used. The pellet will be charged at the top and on its way down it will meet hydrogen and water vapor will be produced, like in equations 6, 7, and 8. Just like today’s production, the reactions will be in solid phase. The steel production will have sponge iron, or DRI, as raw material which will be melt in an EAF together with scrap. Refining and casting will take place as usually. A common theme throughout the HYBRIT-process is the electricity. It will take a lot of electricity, primarily to make the hydrogen and in the EAF. Since the electricity comes from renewable sources, a storage for hydrogen and/or DRI can be made [6].

8

3. Method

The focus in this report is the dephosphorisation in EAF, the impact of FeO activity, and how to determine the FeO activity. Therefor the literature study was divided into four steps. First, HYBRIT + EAF versus blast furnace route was studied, to be followed by dephosphorisation in EAF. Thereafter the focus was on FeO and dephosphorisation and what effect FeO has on the dephosphorisation. Lastly, methods on measuring the activity of FeO was studied.

Measuring of FeO is very advanced and still a somewhat unexplored area. So to limit the project, only three methods are examined.

9

4. Result

The following methods of measuring the activity of FeO were studied.

4.1 Experimental determination of activities

A common method for calculating the activities of FeO in liquid form is equilibrating pure metallic iron with, in the examined study and many more, a silicate melt at 1 atm with fO2

slightly below the iron-wustite (IW) buffer. That corresponds to only Fe²⁺ in the melt and equations 14 and 15:

𝐹𝑒 +¹

2𝑂₂ ↔ 𝐹𝑒𝑂 𝑒𝑞. 14

𝑎_𝐹𝑒𝑂𝑙𝑖𝑞 = 𝑒^∆𝐺(1)

°

𝑅𝑇 ∙ 𝑎_𝐹𝑒∙ √𝑓𝑂₂ 𝑒𝑞. 15

Where aFeOliq is the activity of FeO in the melt, aFe is the activity of iron in metal, ∆G^°(1) is the Gibbs free energy change associated with equilibrium in equation 14. However, this method is limited to a temperature – fO2 domain restricted below the IW buffer. Instead, one can measure aFeOliq and aFe2O3liq under relatively oxidizing fO2 conditions by equilibrating a Fe-bearing alloy with silicate melt. The melt components chosen were stoichiometric FeO and Fe2O3 liquids.

From a thermodynamic analysis of the pseudobinary system SiO2-FeO-±Fe2O3 at 1 atm, the properties of liquid FeO can be estimated. With equation 14 and 15 and known P, T, fO2 and aFe (from XFemetal), the activity of liquid FeO is calculated, as seen in equation 16 [13]:

𝑎_𝐹𝑒𝑂𝑙𝑖𝑞 = 𝑎_𝐹𝑒√𝑓𝑂₂𝑒

−226244+42.49𝑇+𝑃(0.483+8.633∙10−5𝑇−3.784∙10−6𝑃)

𝑅𝑇 𝑒𝑞. 16.

4.2 Electrochemical methods

Another method is using electrochemical methods similar to square wave voltammetry or differential pulse voltammetry. In these methods electrodes are lowered into the melt and voltage is controlled while the current is measured. The diffusion rates are used when calculating the activities. With the measured differential current peak height, the diffusion rates can be calculated with equation 17:

𝐼_𝑚𝑎𝑥 = [𝑛𝐹𝐴√𝐷𝐶]/[√𝜋√𝜏 = 𝜏^′][(1 − 𝜎)/(1 + 𝜎)] 𝑒𝑞. 17 where Imax is the peak current divided by the electrode surface area, n is 2 (the number of electrons in the reduction), F is Faraday’s constant, A is electrode surface area, D is diffusion

10

rate, C is initial concentration in the melt, τ and τ’ are the sampling times, σ=exp[nFΔE/2RT], ΔE is pulse height, R is the gas constant, and T is the temperature.

What the electrochemical experiments are measuring is the voltage it takes to either reduce FeO to metal and oxygen or to oxidise the reduced Fe. The peak from the voltammetry, which determents reduction potentials, does not measure activities (which depends on the concentration), but the activity coefficients, which reflects the point when the molar concentration of oxidised and reduced species are equal. Standard reduction potentials can be used to calculate activity coefficients. They are a measure of the strength of the driving force of reactions under standard state conditions. Equation 18 shows how to calculate the standard reduction potentials.

𝐸^∗ = 𝐸𝑝𝑘 −^𝑅𝑇

𝑛𝐹ln (^𝐷𝑟

𝐷_𝑜)

1 2−^∆𝐸

2 + ^𝑅𝑇

4𝐹 ln 𝑓𝑂₂ 𝑒𝑞. 18 Where E* is the standard reduction potential in volts, Epk is the potential in volts at which current reaches a maximum, Dr is the diffusion rate in m2/s of the reduced species in Pt, Do is the diffusion rate in m²/s of the oxidized species in silicate melt, ΔE is the pulse height in volts, fO2 is the activity of oxygen in atmospheres, R is the gas constant, T is the temperature, F is Faraday’s constant, and n is the number of electrons in the reduction. Thermodynamically in a melt with Fe²⁺ and O^2-, you can write [14]:

𝐹𝑒²⁺+ 𝑂²⁻ → 𝐹𝑒𝑂 𝑒𝑞. 19

Then the activity of FeO can be written like:

𝑎_𝐹𝑒𝑂 = 𝑎_𝐹𝑒²⁺∙ 𝑎_𝑂²⁻ 𝑒𝑞. 20 Where

𝑎_𝐹𝑒²⁺ = 𝛾_𝐹𝑒²⁺_∙𝑋_𝐹𝑒²⁺ 𝑒𝑞. 21 and

𝛾𝐹𝑒𝑂 = 𝑎_𝑂²⁻ ∙ 𝛾_𝐹𝑒²⁺ 𝑒𝑞. 22

4.3 Equilibrating the slag

When developing new metallurgical processes, the thermodynamic information of activity and activity coefficient are important. With wet chemical analysis the mass pct. Fe³⁺ /mass pct. Fe²⁺

11

ratio can be obtained but you can also use spectrophotometry. In a slag system FeO-Fe2O3- CaO-SiO2, the activity calculation is based on equation 23:

𝐹𝑒𝑂_𝑥(𝑖𝑛 𝑠𝑙𝑎𝑔) = 𝐹𝑒 (𝑖𝑛 𝑃𝑡) +^𝑥

2𝑂₂ 𝑒𝑞. 23

And the activity for iron oxide in the liquid comes from equation 24 and equation 25:

𝐹𝑒_(𝑠)+¹

2𝑂_{2 (𝑔)} = 𝐹𝑒𝑂 (𝑠, 𝑙) 𝑒𝑞. 24

𝐹𝑒_(𝑠)+²

3𝑂₂ = 𝐹𝑒𝑂_{1.33 (𝑠)} 𝑒𝑞. 25

By using the established polynomial relationship, the activity coefficient of iron can be calculated and using the obtained activities together with mole fraction composition from the slag system, the activity coefficient for iron oxide can be calculated. The mole fraction is calculated by combining the Fe³⁺/Fe²⁺ ratio and the composition obtained by electron prob microanalysis (EPMA). The activity coefficient of iron can be calculated from equation 26:

𝛾𝐹𝑒_(𝑠)= ^𝑎𝐹𝑒

𝑁_𝐹𝑒(𝑖𝑛 𝑃𝑡) 𝑒𝑞. 26

where γFe(s) indicates the activity coefficient of gamma-iron in the Pt-Fe solid solution, aFe

corresponds to the activity of iron in the Fe-O system, and NFe (in Pt) to the mole fraction of iron in the platinum crucible equilibrated with the liquid slag. At 1573 K and the oxygen partial pressure of 10^-8 and 10^-9 atm, the activity of FeO is calculated with equation 27:

𝐾 = ^{𝑎𝐹𝑒𝑂(𝑙)}

𝛾𝐹𝑒(𝑠)𝑁𝐹𝑒𝑝𝑂₂^1/2 𝑒𝑞. 27

where aFeO (l) indicates the activity of FeO(l) and pO2 indicates the controlled oxygen partial pressure [15].

12

5. Discussion

5.1 Dephosphorisation with DRI

The focus of this work is the dephosphorisation and what effect the FeO has on it. The existing dephosphorisation works well but has to proceed a bit different when the charge is mainly or all DRI. During melting of the DRI, the gangue in the pellet becomes the slag i.e. the raw material of the pellet decides the slag components. Converting to hydrogen reduced DRI does impact the impurities of the steel too. The sulphur content of steel is largely due to the fossil compounds, so in a fossil free steel the content of sulphur would be minimum to none. That in turn decreases the number of steps in the steelmaking process. But the phosphorus content of the pellet needs to be kept to a minimum since phosphorus comes from the iron ore and with DRI, the natural thermodynamic driving force for removal of phosphorus decreases, due to the decrease in the basicity, so it is important to keep that in mind when design the pellet.

It is mentioned that the FeO plays a double role in dephosphorisation. This has to do with the slag composition. Depending on the oxygen potential FeO could behave either acidic or basic.

If the content is too high, that means that the activity of CaO decreases and that FeO behaves acidic. That leads to a lower thermodynamic driving force of the removal of phosphorus, thus leading to a risk of higher phosphorus content in the steel. To prevent this from happening, extra CaO should be added to keep the slag basic.

5.2 Environmental aspects

It will take time before the steel industry is totally fossil free. With that in mind, HYBRIT is a big step in the right direction and it is plausible. Some of the process steps that HYBRIT describes is ongoing right now. It is just one element that separates the not-so-sustainable carbon-based production and the HYBRIT hydrogen-based production. All though, just replacing carbon with hydrogen will not work. The biggest problem for years has been the producing of hydrogen. To cleave water takes a lot of energy and for it to be sustainable the energy must come from a renewable resource. Today, however, that sort of process is possible and the possibility of storing hydrogen is also being looked at. That means that hydrogen can be used for other applications too.

Another difference that has to be taken under consideration is the differences in the enthalpy in the equations. The carbon reducing equations are exothermic while the equations with hydrogen are endothermic. That means that energy has to be added for the reaction to happen. This can be fixed with preheating the hydrogen before entering the shaft.

13

5.3 Measuring methods of FeO activity

The three measuring methods of the activity of FeO looked at in this study is experimental determination, electrochemical methods, and equilibrating the slag. Each method differs a lot from the others. The experimental method is pretty straight forward. The experiment is done on a silicate melt but still has similarities with an iron melt. Stoichiometric FeO and Fe2O3 liquids gives the calculations validity but it must still be tested on an iron melt before any real conclusions can be made. On the other hand, when they substituted Na with Ca in their hydrous melt, they saw an increase in the Fe³⁺ activity coefficient and therefore a decrease in the Fe³⁺/Fe²⁺ which could be useful information. In the second method a voltammetry was used to determine the activity coefficients. To do so, they very early divide the activity of FeO, aFeO, in to aFe2+ and aO2-, so that:

𝑎_𝐹𝑒𝑂 = 𝐾 ∙ (𝑎_𝐹𝑒²⁺)(𝑎_𝑂²⁻) 𝑒𝑞. 28 Which is thermodynamically correct but in this case the activity of the oxide is not of interest in this study and the aFe2+ was not studied enough. This cannot be directly applied to steel plants.

And finally, the third method looked at a slag containing almost the same elements as a slag used for dephosphorisation and looks like the most appropriate for further experiments. With wet chemical analysis they found the mass pct. Fe³⁺ /mass pct. Fe²⁺ ratio but it could also be done with spectrophotometry. With the mass pct. of the ions and mole fraction composition of the slag the activity of FeO can be calculated, making this method the most possible and one to do experiments with. Ongoing, to be able to get data on the activity of FeO at lower basicity, there needs to be more experiments done and more experiments with wet chemical analysis and spectrophotometry to see which of these methods that is the best at calculating the amount of Fe²⁺ and Fe³⁺.

14

6. Conclusions

This study sought out to find a method to calculate the FeO activity in steel. The effect that the FeO activity has on the dephosphorisation of iron was also looked at. Among the three methods that were evaluated in this report, one stood out: equilibrating the slag., where wet chemical analysis and spectrophotometry was used to calculate the activity of FeO, with pleasant results.

It is recommended that this method is further evaluated to see if one of these techniques could work within HYBRIT and other aspects of fossil free steel production.

15

7. References

[1] J. Bengtsson, “Naturvårdsverket,” 12 December 2019. [Online]. Available:

https://www.naturvardsverket.se/Miljoarbete-i-samhallet/Miljoarbete-i-Sverige/Uppdelat- efter-omrade/Klimat/Sveriges-klimatlag-och-klimatpolitiska-ramverk/#. [Accessed 3 Maj 2020].

[2] “European Commission,” 4 Mars 2020. [Online]. Available:

https://ec.europa.eu/commission/presscorner/detail/en/ip_20_335. [Accessed 4 Maj 2020].

[3] M. Smith, “Naturvårdsverket Parisavtalet,” 31 Maj 2019. [Online]. Available:

https://www.naturvardsverket.se/Miljoarbete-i-samhallet/EU-och- internationellt/Internationellt-

miljoarbete/miljokonventioner/Klimatkonventionen/Parisavtalet/. [Accessed 4 Maj 2020].

[4] “Ny Teknik,” 3 December 2018. [Online]. Available: https://www.nyteknik.se/miljo/fn-s- klimatmal-nara-omojligt-att-na-6941116. [Accessed 4 Maj 2020].

[5] J. Huss, “Optimizing the slag system for phosphorus removal in a DRI-based EAF-process using the dictionary attack method,” KTH, Materialvetenskap, Stockholm, 2018.

[6] O. Wijk, G. Hyllander, J. Sandberg, M. Tottie, B. Sundelin, M. Thorén, M. Nordlander, N.

Simonsson and R. Vikman, “Slutrapport HYBRIT - Hydrogen Breakthrough Ironmaking Technology,” Stockholm, 2018.

[7] “International Iron Metallics Association,” 2017. [Online]. Available:

https://www.metallics.org/pig-iron-bf.html. [Accessed 8 Maj 2020].

[8] M. Andersson and T. Sjökvist, Processmetallurgins grunder, Stockholm: KTH, 2004.

[9] A. Andersson and B. Bertilsson, Tretise on Process Metalurgy, Amsterdam: Elsevier, 2014.

[10] R. J. Longbottom and L. Kolbeinsen, “Iron Ore Reduction with CO and H2 Gas Mixtures - Thermodynamic and Kinetic Modelling,” University of Wollongong, 2008.

[11] J. Huss, M. Berg, N. Kojola, S. Du and P. Jönsson, “Experimental study on phosphorus partitions between liquid iron and liquid slags based on DRI,” Internal report Hybrit RP1, Stockholm, 2019.

[12] J. P. J. Heo, “Effect of Direct Reduced Iron (DRI) on Dephosphorization of Molten Steel by Electric Arc Furnace Slag,” Metall and Materi Trans B, vol. 49, pp. 3381-3389, 2018.

[13] B. Scaillet, M. Pichavant and F. Gaillard, “Experimental determination of activities of FeO and Fe2O3 components in hydrous silicic melts under oxidizing conditions,” Geochimica et Cosmochimica Acta, vol. 67, no. 22, pp. 4389-4409, 2003.

[14] R. O. Colson, A. Floden, T. Haugen, K. Malum, M. Sawarynski, M. Nermoe, K. Jacobs and D.

Holder, “Activities of NiO, FeO, and O2- in silicate melts,” Geochimica Cosmochimica Acta, vol.

69, no. 12, pp. 3061-3073, 2005.

16

[15] H. M. Henao and K. Itagaki, “Activity and Activity Coefficient of Iron Oxides in the Liquid FeO- Fe2O3-CaO-SiO2 Slag Systems at Intermediate Oxygen Partial Pressures,” Metallurgical and Materials Transactions B, vol. 38, no. Oct 2007, pp. 769-780, 2007.

[16] W. Urban, M. Weinberg and J. Cappel, “De-Phodpgorization Strategies and Modelling in Oxygen Steelmaking,” Iron Steel Technol, vol. 134, pp. 27-39, 2014.

TRITA ITM-EX 2020:310

www.kth.se