BEHAVIOR OF SILICON, CARBON AND CHROMIUM ...

BEHAVIOR OF SILICON, CARBON AND CHROMIUM IN THE FERROCHROME CONVERTER - A COMPARISON BETWEEN CTD AND PROCESS SAMPLES Eetu-Pekka HEIKKINEN, University of Oulu, Finland Topi IKÄHEIMONEN, Outokumpu Stainless, Finland Olli MATTILA, Ruukki Metals Oy, Finland Timo FABRITIUS and Ville-Valtteri VISURI, University of Oulu, Finland

Abstract The purpose of a ferro-chrome converter (hereafter abbreviated as CRK) is to remove silicon as well as part of carbon from the molten ferro-chrome by blowing oxygen into the melt. The product of the CRK is a raw material for the AOD converter used in stainless steel production, whereas the raw materials of the CRK are molten ferro-chrome with high silicon and carbon contents (from the ferro-chrome plant) and stainless steel scrap that is used to decrease the melt’s temperature that is increased due to oxidation reactions of silicon and carbon. The CRK process could therefore be considered as a link between the ferro-chrome production and the production of stainless steels. Oxidation reactions of silicon, carbon and chromium in the CRK process have been considered by comparing the results of computational thermodynamics with process samples from the Outokumpu Stainless Tornio steelworks. The results of this study indicated that the oxidation reactions are not in mutual equilibrium at the end of the CRK processing. Due to limitations of this preliminary study, namely isothermal computations in which divalent chromium was not taken into account and process samples taken only from the end of the CRK process, it was necessary to verify the results with an improved thermodynamic modelling as well as samples from the different stages of the CRK process.

Introduction Computational thermodynamics provides a suitable tool for the modelling, understanding and control of pyrometallurgical processes. On the other hand, the measurements from the metallurgical processes provide process data which could be used in the validation of new thermodynamic models and thermochemical data. However, in comparison to some widely used metallurgical unit operations (such as BOF), there is much less information available concerning the thermodynamic state of less common metallurgical processes such as the CRK process, in which the silicon and carbon contents of ferrochromium are lowered before the stainless steel production. In order to estimate whether the oxidation reactions of carbon, silicon and chromium taking place in the CRK process are in mutual equilibrium or not, a comparison between computational values representing thermodynamic equilibria and measured process data was made1). The results of this study indicated that the oxidation reactions are not in mutual equilibrium at the end of the blowing and one should therefore be careful when veryfying thermodynamic modelling with the process measurements1). However, there were some significant limitations in calculations (i.e. isothermal approach and neglect of divalent chromium) as well as in process sampling (i.e. samples were taken only from the final state of the CRK process)1). Due to these

The 6th European Oxygen Steelmaking Conference - Stockholm 2011

Programme No. 3-04

limitations it was felt necessary to expand the evaluation with a new comparison, for which the process samples were obtained from different stages of the CRK process and thermodynamic equilibria were calculated to correspond the actual temperatures measured from the CRK process. Additionally, the oxidation of chromium to both divalent and trivalent states (CrO and Cr2O3, respectively) was taken into account.

CRK process The CRK process is a bridge between the production of ferrochromium and stainless steel. Its task is to serve as a buffer for molten ferrochromium (between the ferrochromium plant and the stainless steel melting shop) as well as to enable the utilisation of ferrochromium’s chemical energy for scrap melting. With the use of CRK process, the concumption of electric energy in the electric arc furnace (EAF) is decreased not only due to utilisation of oxidation reaction’s chemical energy, but also due to logistical benefits, i.e. possibility to transport ferrochromium from ferrochromium plant to stainless steel melting shop in a molten state.1) The location of the CRK process in the stainless steel production chain (used in the Outokumpu Stainless Tornio steelworks) is presented in Figure 12).

Figure 1. Stainless steel production in the line 1 of Outokumpu Stainless Tornio steelworks2). Liquid ferrochrome contains approximately 4 w-% silicon and 7 w-% carbon. All silicon and some of the carbon are oxidized in the CRK process by blowing pure oxygen or oxygen/compressed air mixture into the bath. Thermal energy generated in the oxidation reactions is utilized for scrap and alloy melting. During the process, 550 kg of scrap and 150 kg of slag formers per ton of produced ferrochromium are melted. Simultaneously the charge temperature is risen up to 1650 - 1700 °C. The oxidation of silicon, carbon and chromium is illustrated in Figure 2 in which the Cr2O3-content of the slag as well as the Si- and C-contents of the metal are shown as a function of the amount of blown oxygen.1,3,4)


Programme No. 3-04

Figure 2. Changes of slag and metal phase compositions during blowing in CRK process4). Nominal capacity of the CRK process in Outokumpu Stainless Tornio Works is 95 tons. The geometry of the vessel is similar to AOD converter at the same plant and it is lined with MgO-C bricks. The charge weight varies normally from 30 to 100 tons. The CRK process is equipped with five side-mounted tuyeres located at 40 centimeters height from the vessel bottom. Top lance is of Laval type with three holes (3HL). In the beginning of the blowing approximately 60 m3/min (STP) of oxygen is blown through the tuyeres (cf. stage 6 in Figure 2). In stage 22 (cf. Figure 2) approximately 50 m3/min (STP) of air is blown through the tuyeres in addition to approximately 130 m3/min (STP) of oxygen that is blown through the lance. The final stage is similar to the first stage with the oxygen blowing of approximately 60 m3/min (STP) through the tuyeres.1,3) Average compositions of the metal and slag at the different stages of the CRK process are presented in Tables 1 and 2, respectively. For explanations concerning the markings ‘1’, ‘2A’ and ‘2B’ in Tables 1 and 2, cf. chapter about ‘Sampling and analytical methods’. Table 1. Average melt compositions and temperatures at the different stages of the CRK process. Cr (w-%) C (w-%) Si (w-%) T (qC)

Start 54.0 7.1 4.7 1520

Intermediate ‘1’ 53.8 4.9 1.0 1655

Final ‘2A’ 54.6 0.45 4.5 1728

Final ‘2B’ 36.1 2.9 0.4 1626

Table 2. Average CRK slag compositions at the different stages of the CRK process. 1 2A 2B

CaO (w-%) 45.1 48.0 43.5

SiO2 (w-%) 27.0 35.7 29.7

MgO (w-%) 18.2 11.0 18.3

Fe2O3 (w-%) 1.0 0.3 1.0

Cr2O3 (w-%) 2.7 1.9 2.5

Al2O3 (w-%) 4.1 2.1 2.9


TiO2 (w-%) 1.2 0.4 1.5

MnO (w-%) 0.2 0.1 N/A

Basicity 2.1 1.6 1.9

Programme No. 3-04

Sampling and analytical methods Two metal and slag samples were taken from each of the 11 charges considered in this study. Both samples were taken manually from the tilted CRK. Metal samples were taken through slag with a lollipop type sampling probe, whereas slag samples were taken with an iron rod. Differing from our previous study in which all the samples were taken at the end of the blowing1), samples were now taken on two different points of time (intermediate samples refered as samples ‘1’ as well as final samples refered as samples ‘2’). Additionally, the final samples are in this paper divided in two separate categories, ‘2A’ and ‘2B’, the first of which represents the samples from the charges in which low-carbon ferrochromium was produced and additional silicon was added to metal in order to enhance its crushability, since it could not be directly charged into the AOD process. The latter represents the samples from the charges in which no additional silicon was added (i.e. direct charging to the AOD process). Carbon contents of the metal samples were analyzed with optical emission spectrometer (ARL4460) and other elements with XRF (Panalytical Axios Advanced). Grinded slag briquettes were analyzed with a similar XRF as steel samples.

Equilibria of the oxidation reactions The mutual equilibrium of the oxidation reactions of silicon, chromium and carbon was estimated by calculating the partial pressures of oxygen describing the equilibria of the oxidation reactions presented in equations (1) to (4). Individual reactions (1) and (2) were chosen to describe the oxidation of silicon and carbon, respectively, whereas the oxidation of chromium was described with two separate reactions (3) and (4) the previous describing the formation of chromium(III)oxide (Cr2O3) and and the latter the formation of chromium(II)oxide (CrO). The equilibrium constants for the reactions (1) to (4) are presented in equations (5) to (8), respectively. Si Fe O2 g SiO2 2 C Fe O 2 g 2 CO g

(1) (2)

4 2 Cr2 O3 Cr Fe O2 g 3 3 2 Cr Fe O2 g 2 CrO

K 1

K 2

K 3

K 4

a SiO2

a SiO2

a Si p O2

f Si X Si p O2

2 p CO a C2 p O2

2 p CO f C2 X C2 p O2

23 a Cr 2 O3

23 a Cr 2 O3

43 a Cr p O2

43 f Cr4 3 X Cr p O2

2 a CrO 2 a Cr p O2

f Cr2

2 a CrO 2 X Cr p O2

(3)

(4)

e

e

§ 'G01 'G R o H ¨ ¨ RT ©

· ¸ ¸ ¹

§ 'G02 'G R o H ¨ ¨ RT ©

· ¸ ¸ ¹

e

e

(5)

(6)

§ 'G03 'G R o H ¨ ¨ R T ©

§ 'G04 'G R o H ¨ ¨ R T ©

· ¸ ¸ ¹

· ¸ ¸ ¹


(7)

(8)

Programme No. 3-04

for which ai represents the activity of component i, pi the partial pressure of component i, Xi the mole fraction of component i and fi the activity coefficient of component i. 'G0(n) equals the standard Gibbs energy for the reaction presented in Equation (n), 'GRoH equals the Gibbs energy for the change between the Raoultian and Henrian standard states, R equals the gas constant (1,987 calmol-1K-1 = 8,314 Jmol-1K-1) and T equals temperature. Partial pressures of oxygen corresponding the equilibrium between silicon and SiO2, carbon and CO, chromium and Cr2O3 as well as chromium and CrO were calculated using Equations (9) to (12), that were obtained from the Equations (5) to (8), respectively.

p O2 pO2 p O2 p O2

a SiO2 f Si X Si

p e 2 f C X C2 23 a Cr 2 O3 43 f Cr4 3 X Cr

a

2 f Cr2 X Cr

· ¸ ¸ ¹

§ 'G02 'G R o H ¨ ¨ RT ©

· ¸ ¸ ¹

e

2 CO

2 CrO

§ 'G01 'G R o H ¨ ¨ RT ©

e

e

(9)

(10)

§ 'G03 'G R o H ¨ ¨ RT ©

§ 'G04 'G R o H ¨ ¨ R T ©

· ¸ ¸ ¹

· ¸ ¸ ¹

(11)

(12)

The values needed for the calculations of the partial pressures as well as the assumptions behind the values are presented in Table 3.


Programme No. 3-04

Table 3. The values used in the calculations as well as the list of used models and assumptions. Parameter in Equation (10) (9) (11)

(12)

T

T

T

T

aSiO2

Used value(s) 1841-2038 (K) 0.0066-0.098

2.85-12.20

fSi

0.000160.090 47.2T(qC) - 213000 (cal/mol) 24300-26900 (cal/mol) 2.15 (bar)

XSi 'G0(1) 'GRoH pCO

0.174-2.52

fC

XC

0.014-0.213 -40.2T - 67400 (cal/mol) 4110-4550 (cal/mol)

'G0(2) 'GRoH aCr2O3

0.061-0.190

fCr

fCr

0.397-0.926

XCr

XCr

0.33-0.56 40.16T(qC) - 168315 (cal/mol) -708 - -639 (cal/mol)

'G0(3) 'GRoH aCrO 'G0(4) 'GRoH

0.007180.258 16.0T(qC) 128000 (cal/mol) -1061 - -959 (cal/mol)

Models and assumptions that were used Measured temperatures from the CRK process while taking samples. Calculated SiO2 activities (using quadratic formalism by Ban-Ya5) and Xiao et al.6) and measured contents of the CRK-slag at the different stages of blowing). Calculated activity coefficients of Si (using the unified interaction parameter formalism by Pelton and Bale7) and Henrian standard states) for different metal compositions (obtained from the metal sample analyses). Measured Si-contents of the metal at the different stages of blowing. Standard free energy change for the reaction presented in Equation (1) achieved from the database of the HSC Chemistry for Windows, V6.18). -RTlnf0Si (value of f0Si achieved from reference 9)). Estimated ferrostatic pressure in the CRK process. Calculated activity coefficients of C (using the unified interaction parameter formalism by Pelton and Bale7) and Henrian standard states) for different metal compositions (obtained from the metal sample analyses). Measured C-contents of the metal at the different stages of blowing. Standard free energy change for the reaction presented in Equation (2) achieved from the database of the HSC Chemistry for Windows, V6.18). -2RTlnf0C (value of f0C achieved from reference 9)). Calculated Cr2O3 activities (using quadratic formalism by Ban-Ya5) and Xiao et al.6) and measured contents of the CRK-slag at the different stages of blowing). Calculated activity coefficients of Cr (using the unified interaction parameter formalism by Pelton and Bale7) and Henrian standard states) for different metal compositions (obtained from the metal sample analyses). Measured Cr-contents of the metal at the different stages of blowing. Standard free energy change for the reaction presented in Equation (3) achieved from the database of the HSC Chemistry for Windows, V6.18). -(4/3)RTlnf0Cr (value of f0Cr achieved from reference 9) ). Calculated CrO activities (using quadratic formalism by Ban-Ya5) and Xiao et al.6) and measured contents of the CRK-slag at the different stages of blowing). Standard free energy change for the reaction presented in Equation (4) achieved from the database of the FactSage, V5.510). -2RTlnf0Cr (value of f0Cr achieved from reference 9)).


Programme No. 3-04

Calculated partial pressures of oxygen corresponding to the equilibria between measured Si- and SiO2-contents, measured C-contents and estimated partial pressure of CO as well as measured Cr- and Cr2O3/CrO-contents are presented in Figure 3 for different samples ‘1’ representing the intermediate samples, ‘2A’ representing the final samples with silicon addition and ‘2B’ representing the final samples without silicon addition. 1.E-07 1.E-08 1.E-09

p(O2)

1.E-10 1.E-11 1.E-12 1.E-13 1.E-14 1.E-15 1.E-16 1

1

1

1

1

1

1

1

1

1

1 2A 2A 2A 2A 2A 2B 2B 2B 2B 2B 2B

Sample Si + O2 = SiO2 2Cr + 1,5O2 = Cr2O3

2C + O2 = 2CO Cr + 0,5O2 = CrO

Figure 3. Partial pressures of oxygen calculated from the measured process data concerning the slag and metal compositions at the different stages of the blowing. In order to be in mutual equilibrium with each other, the partial pressures of oxygen calculated for the oxidation reactions of silicon, carbon and chromium should have values that are close to one another. It is seen from Figure 3 that for intermediate samples (marked with ‘1’ in Figure 3) as well as for final samples with additional silicon addition (marked with ‘2A’ in Figure 3) the difference between the partial pressures of oxygen describing the equilibria of different oxidation reactions is significant (up to three or four decades), although there are few samples in which the difference between the values of pO2 is considerably lower (especially if one assumes that chromium is oxidized into CrO and not into Cr2O3). This indicates that the oxidation reactions of silicon, carbon and chromium are generally not in mutual equilibrium, although it is close in some individual charges. The mutual equilibrium of the oxidation reactions is most likely achieved in charges in which no additional silicon was added (samples marked with ‘2B’ in Figure 3), although even these samples show some variation in the values of pO2. The variation is largest (i.e. the oxidation reactions are least likely to be in mutual equilibrium with each other) in samples ‘2A’ which represent the charges in which additional silicon was added to the melt after the blowing. However, rather than the silicon addition itself, the reason for the difference is more likely to be the lower carbon content in the metal in these samples, since higher partial pressures of oxygen (i.e. more oxidizing conditions) are required to oxidize carbon when its content decreases. These observations concerning the lack of mutual equilibrium are in accordance with our previous study in which only the final state of the CRK process was considered1).


Programme No. 3-04

In addition to mutual equilibrium, it is possible to estimate the order of oxidation for silicon, carbon and chromium by comparing the partial pressures of oxygen that are needed to oxidize these elements. According to Figure 3 it seems that for intermediate samples (once again marked with ‘1’) the order of oxidation is either Si o C o Cr (to CrO) o Cr (to Cr2O3) or Cr (to CrO) o C o Si o Cr (to Cr2O3) depending the amounts of silicon and carbon in the metal. This order indicates that chromium does not necessarily oxidize to Cr3+ (Cr2O3), if the oxidizing conditions are assumed to be determined by the oxidation reactions of silicon and/or carbon. Due to additional silicon addition, the order of oxidation is more difficult to define for the final samples. For the samples ‘2B’ (i.e. no silicon addition) the order of oxidation seems to be Cr (to CrO) o Si/C o Cr (to Cr2O3), although the difference between the values of pO2 describing the formation of Cr2O3, SiO2 and CO is not very clear. This indicates that the chromium is more likely to have a higher valence (Cr3+) in final samples than in the intermediate ones. The order of oxidation for the samples ‘2B’ (in which silicon was added to the charge) is Cr (to CrO) o Si o Cr (to Cr2O3) o C, which reflects the difficulties to achieve low carbon contents without oxidizing the chromium. In addition to partial pressures of oxygen, the mutual equilibrium of the oxidation reactions was considered using the CO/CO2 -ratios corresponding the equilibria (cf. Figure 4). These CO/CO2 ratios were calculated using equation (14) that shows the relation between the partial pressures of oxygen and the corresponding CO/CO2 -ratios based on the chemical reaction presented in equation (13). 2 CO g O2 g 2 CO2 g

pCO pCO2

§ 'G13 ¨ e RT ¨ ¨ p O2 ©

· ¸ ¸ ¸ ¹

(13)

12

(14)

for which pi represents the partial pressure of component i, 'G0(13) equals the standard Gibbs energy for the reaction presented in equation (13), R equals the gas constant (1,987 calmol-1K-1 = 8,314 Jmol-1K-1) and T equals temperature.


Programme No. 3-04

100000

10000

CO/CO2

1000

100

10

1 1

1

1

1

1

1

1

1

1

1

1 2A 2A 2A 2A 2A 2B 2B 2B 2B 2B 2B

Sample Si + O2 = SiO2 2Cr + 1,5O2 = Cr2O3

2C + O2 = 2CO Cr + 0,5O2 = CrO

Figure 4. CO/CO2-ratios calculated from the measured process data concerning the slag and metal compositions at the different stages of the blowing. In addition to the conclusions that could already be drawn from Figure 3, it is seen from Figure 4 that even though the difference between the CO/CO2 -ratios (or partial pressures of oxygen as seen from Figure 3) can be relatively large (up to several decades), it does not necessarily require big amounts of CO2 to change the CO/CO2 -ratio quite drastically, since the equilibrium is heavily on the CO’s side. The effect of metal composition on the state of equilibrium of the oxidation reactions is presented in Figures 5 to 7, in which the partial pressures of oxygen representing the equilibrium of the oxidation reactions of silicon, carbon and chromium (i.e. reaction presented in equations (1) to (4)) are plotted as a function of silicon, carbon and chromium contents in the metal. In Figure 5 it is shown that the partial pressure of oxygen in equilibrium with Si and SiO2 depends strongly on the metal’s silicon content, since the amount of oxygen needed to oxidize silicon increases when the silicon content in metal decreases. The higher values of pO2 in the silicon content range of 6-9 mol-% (samples ‘2A’) are due to silicon addition at the end of the blowing and therefore these points are not comparable with the others. A behavior similar to silicon can be observed from Figure 6 for the equilibrium between carbon and carbon monoxide. With higher carbon contents (i.e. the intermediate samples ‘1’) the partial pressure of oxygen needed to oxidize the carbon is low, but it increases as metal’s carbon content decreases during the blowing. In the final samples (i.e. samples ‘2’) the values of pO2 describing the C-CO-equilibrium are higher the further the decarburization proceeds. Since no extra carbon is added to the metal after the blowing, all the final samples are comparable unlike in Figure 5. Figure 7 is not as informative as Figures 5 and 6. Due to high chromium content in the molten metal during all the stages of blowing, the partial pressure of oxygen needed to oxidize the chromium into either CrO or Cr2O3 did not vary much as a function of blowing time nor as a function of chromium content. Hence the behavior typical to oxidation of silicon and carbon as seen in Figures 5 and 6 is not observable for the chromium (this would require samples with a The 6th European Oxygen Steelmaking Conference - Stockholm 2011

Programme No. 3-04

much lower chromium content in the metal). Figure 7 also shows that the formation of CrO requires less oxidizing conditions in comparison to the formation of Cr2O3, although this is rather obvious since chromium has a lower valence in CrO and higher in Cr2O3. 1.E-07 1.E-08 1.E-09

p(O2)

1.E-10 1.E-11 1.E-12 1.E-13 1.E-14 1.E-15 1.E-16 0

0.02

0.04

0.06

0.08

0.1

X(Si) in metal Si + O2 = SiO2 (Sample 1)

Si + O2 = SiO2 (Sample 2)

Figure 5. Partial pressures of oxygen representing the equilibrium between silicon (in metal) and SiO2 (in slag) as a function of silicon content in metal calculated from the measured process data concerning the slag and metal compositions at the different stages of the blowing. 1.E-07 1.E-08 1.E-09

p(O2)

1.E-10 1.E-11 1.E-12 1.E-13 1.E-14 1.E-15 1.E-16 0

0.05

0.1

0.15

0.2

0.25

X(C) in metal 2 C + O2 = 2 CO (Sample 1)

2 C + O2 = 2 CO (Sample 2)

Figure 6. Partial pressures of oxygen representing the equilibrium between carbon (in metal) and carbon monoxide (gas) as a function of carbon content in metal calculated from the measured process data concerning the metal compositions at the different stages of the blowing.


Programme No. 3-04

1.E-07 1.E-08 1.E-09

p(O2)

1.E-10 1.E-11 1.E-12 1.E-13 1.E-14 1.E-15 1.E-16 0

0.1

0.2

0.3

0.4

0.5

0.6

X(Cr) in metal 2 Cr + 1,5 O2 = Cr2O3 (Sample 1) 2 Cr + 1,5 O2 = Cr2O3 (Sample 2)

Cr + 0,5 O2 = CrO (Sample 1) Cr + 0,5 O2 = CrO (Sample 2)

Figure 7. Partial pressures of oxygen representing the equilibrium between chromium (in metal) and either Cr2O3 or CrO (both in slag) as a function of chromium content in metal calculated from the measured process data concerning the slag and metal compositions at the different stages of the blowing. As it was already considered in the context of Figure 3, the calculated partial pressures of oxygen presented in Figures 5 to 7 can also be used to estimate the order of oxidation for silicon, carbon and chromium at the different stages of blowing. It is seen from Figures 5 to 7 that with higher silicon contents the driving force for the oxidation of silicon is greater than the one of carbon and chromium (i.e. the partial pressures of oxygen describing the equilibrium between carbon and CO as well as chromium and CrO or Cr2O3 in Figures 6 and 7 are higher than the ones describing the equilibrium between silicon and SiO2 in Figure 5). However, as the silicon content in metal decreases towards the end of blowing, the driving force for the oxidation of carbon and chromium becomes greater. And as the decarburization proceeds, the partial pressure of oxygen needed to oxidize the carbon into CO increases, too, thus making the driving force for the oxidation of chromium greater than the ones for either silicon or carbon. According to Figures 6 to 7, the partial pressures of oxygen required to oxidize chromium and carbon are approximately 10-13 for chromium into CrO, approximately 10-12 for chromium into Cr2O3 and approximately 10-13 for carbon into CO with higher carbon contents in metal. On the other hand, Figure 5 shows that the values of pO2 describing the equilibrium between silicon and SiO2 exceed 10-13-10-12 when the mole fraction of silicon in metal decreases below 0.004-0.010. Hence, the silicon content below which it becomes more favourable for chromium and carbon to be oxidized in comparison to silicon, is approximately 0.4-1.0 mol-% or 0.29-0.62 w-%. This limit was estimated to be approximately 0.3 w-% in our previous study, whereas the observations from the CRK process have indicated that the oxidation of chromium becomes significant when the silicon content of the metal decreases below 0.1 w-%1). Figures 6 and 7 also illustrate, that since the oxidation of chromium increases below the carbon contents of approximately 10 mol-% or 2.5 w-%, the reduction of chromium is required when producing ferrochromium with carbon contents lower than this.


Programme No. 3-04

The effect of temperature on the state of equilibrium of the oxidation reactions was illustrated by using Ellingham-type diagrams (presented in Figures 8 to 11), in which the values of RTln(pO2) as a function of temperature are used to describe the equilibrium of each oxidation reaction and in which the driving force for the oxidation is larger for the reactions that are located lower in the diagrams. In this study the values of RTln(pO2) were calculated separately for each sample (i.e. for each charge) using the values of pO2 obtained from equations (9) to (12) in order to describe the dispersion of the values. Additionally, equations (15) to (18) were used to calculate the boundaries between which the process is operated.

>RT ln p >RT ln p >RT ln p >RT ln p

@ @ @ @

O2 Si SiO 2

O2 C CO

O2 Cr Cr O 2 3 O2 Cr CrO

'G01 RT ln a SiO2 RT ln a Si

(15)

'G02 2 RT ln pCO 2 RT ln aC

(16)

2 4 RT ln aCr2O3 RT ln aCr 3 3 0 'G4 2 RT ln aCrO 2 RT ln aCr 'G03

(17) (18)

for which ai represents the activity of component i and pi the partial pressure of component i. 'G0(n) equals the standard Gibbs energy for the reaction presented in Equation (n), R equals the gas constant (1,987 calmol-1K-1 = 8,314 Jmol-1K-1) and T equals temperature.

-60000

-60000

-70000

-70000

-80000

-80000

-90000

-90000 RTln(pO2) (cal/mol)

RTln(pO2) (cal/mol)

The boundaries were obtained by defining the lines representing the maximum and minimum values of RTln(pO2). The maximum values were defined by using the highest values of oxide activities (i.e. aSiO2, pCO, aCr2O3 and aCrO) and the lowest values of element activities (i.e. aSi, aC and aCr) in equations (15) to (18), whereas the minimum values were achieved by using the highest values of element activities and the lowest values of oxide activities. In addition to maximum and minimum, a line describing the oxidation of pure elements into pure oxides (i.e. aSiO2 = pCO = aCr2O3 = aCrO = aSi = aC = aCr = 1) was also drawn to Figures 8 to 11 for comparison.

-100000 -110000 -120000

-100000 -110000 -120000

-130000

-130000

-140000

-140000

-150000

-150000

-160000 1400

-160000 1500

1600

1700

1800

1400

Temperature (Celsius)

Figure 8. Ellingham-type diagram for the Si/SiO2 -equilibrium.

1500

1600

1700

1800


Figure 9. Ellingham-type diagram for the C/CO -equilibrium.


Programme No. 3-04

-60000

-70000

-70000

-80000

-80000

-90000

-90000 RTln(pO2) (cal/mol)

RTln(pO2) (cal/mol)

-60000

-100000 -110000 -120000

-100000 -110000 -120000

-130000

-130000

-140000

-140000

-150000

-150000

-160000 1400

-160000 1500

1600

1700

1800

1400


Figure 10. Ellingham-type diagram for the Cr/Cr2O3 -equilibrium.

1500

1600

1700

1800


Figure 11. Ellingham-type diagram for the Cr/CrO -equilibrium.

The boundary values for RTln(pO2) are illustrated in Figures 8 to 11 as thick lines with diamonds, whereas the line describing the oxidation of pure elements into pure oxides is illustrated with a thick line with squares. Open circles as well as open triangles represent the values calculated for individual intermediate samples (i.e. samples ‘1’), whereas closed circles represent individual final samples with a silicon addition (i.e. samples ‘2A’) and closed triangles individual final samples without a silicon addition (i.e. samples ‘2B’). Thin lines are drawn to represent the values of pO2 at intervals of two decades the uppermost corresponding the value of 10-8 and the lowest corresponding the value of 10-18. Figures 8 to 11 show that although there are considerable variations in the temperatures of different samples, it is difficult to separate out the actual effect of temperature on the reaction equilibria, since it is not the only changing variable in the samples. This is clearly seen in Figure 9 which shows that the equilibrium between carbon and carbon monoxide is totally different in samples ‘2A’ (closed circles) and in samples ‘1’ (open circles) even though they appear in the same temperature range. This indicates that the equilibria are more dependent on the metal and slag compositions than on temperature. Nevertheless, Figure 11 shows some indication on the temperature dependence of the Cr/CrO -equilibrium.

Conclusions CRK is a process between the production of ferrochromium and the production of stainless steel, in which the contents of silicon and carbon in molten metal are decreased by blowing oxygen into the melt. The oxidation reactions of silicon, carbon and chromium in the CRK process were considered based on computational thermodynamics as well as metal and slag compositions obtained from the process samples taken from the Outokumpu Stainless Tornio steelworks.


Programme No. 3-04

The results indicate that the oxidation reactions of silicon, carbon and chromium are not necessarily in mutual equilibrium during the process, although the lack of mutual equilibrium is not quite clear in all the samples. The differences in the values of partial pressure of oxygen required to oxidize silicon, carbon and chromium indicate the order of oxidation during the process. Based on the results concerning the intermediate samples the order of oxidation is Si o C o Cr as long as the silicon content in the metal is higher than approximately 0.29-0.62 w-%. As the amount of silicon decreases, the driving forces for the oxidation of carbon and chromium become larger than the one for silicon. Based on the intermediate samples it is not clear whether chromium is oxidized into CrO or Cr2O3. At the end of the blowing the order of oxidation depends heavily on the carbon content that is aimed for. With higher final carbon contents, the driving forces for the oxidation of silicon and carbon are close to one another. However, with lower final carbon contents the further oxidation of carbon requires considerably higher partial pressures of oxygen than the oxidation of both chromium and silicon. Hence, the reduction of chromium is required when producing ferrochromium with carbon contents lower than 2.5 w-%.

Acknowledgements The authors would like to thank the operators of the CRK process as well as the personnel of the analytical laboratory of the Outokumpu Stainless Tornio works for their contribution. This research was partly funded by Finnish Funding Agency for Technology and Innovation (TEKES) and it is a part of the Energy Efficiency & Lifecycle Efficient Metal Processes (ELEMET) research program coordinated by Metals and Engineering Competence Cluster (FIMECC).

References 1) E-P. Heikkinen, T. Ikäheimonen, O. Mattila and T. Fabritius, INFACON XII - The Twelfth International Ferroalloys Congress, Helsinki, Finland, 6-9th June, 2010. 229237. 2) Teräskirja (‘Steelbook’), 8th ed. Metallinjalostajat ry, Tampere, 2009. 104 p. [In Finnish] 3) T. Fabritius, PhD thesis, University of Oulu, Finland, 2004, 77 p. 4) V. Juntunen, J. Pitkälä and P. Kupari, Material for supplementary education, POHTO, Oulu, Finland. [In Finnish] 5) S. Ban-Ya, ISIJ Int, Vol. 33 (1993), No. 1, 2-11. 6) Y. Xiao, L. Holappa and M.A. Reuter, Met and mat trans, Vol. 33B (2002), 595-603. 7) A. Pelton and C. Bale, Met and mat trans, Vol. 17A (1986), 1211-1215. 8) A. Roine, HSC Chemistry for Windows, version 6.1 and its database, Outotec research oy, 2006. 9) G.K. Sigworth and J.F. Elliott, Metal science, Vol. 8 (1974), No. 9, 298-310. 10) C.W. Bale, A.D. Pelton, W.T. Thompson, G. Eriksson, K. Hack, P. Chartrand, S. Decterov, J. Melançon & S. Petersen, FactSage Version 5.5 Software and its databases, Thermfact and GTT-Technologies, 2007.


Programme No. 3-04