DEFINING THE OPERATING REGIME AND METHODOLOGY FOR THE FURNACE METHOD FOR THE PRODUCTION OF LOW CARBON
FERROCHROME
Heine Weitz1,2, Andrie Garbers-Craig1
1University of Pretoria, Lynnwood Road, Hatfield, South Africa
2OptiProc, 12 OโHare Street, Highveld, Centurion, South Africa Keywords: Low carbon ferrochrome, Furnace Method, Silicothermic reduction
Abstract
The Furnace Method for the production of low carbon ferrochrome has been found to offer significant savings in energy and raw material consumptions, compared to the Mixing Method processes. The operating conditions for the Furnace Method differ significantly from those of the Mixing Method processes. This paper describes the characterisation of the slag operating regime for the Furnace Method in order to minimise risks during implementation.
The slag liquidus temperature, as well as its chemical and thermal compatibility with different refractory systems was evaluated to identify the preferred fluxing agent, basicity ((CaO+MgO)/SiO2) range and refractory system. A lime fluxed slag was found to provide the best combination of liquidus temperature and alloy product quality, while being compatible with a magnesia refractory at an operating temperature of 1750ยฐC.
Slag chemistry variations throughout batches pose a risk to the integrity of the refractory lining.
The impact that these variations have on the process chemistry was quantified in order to arrive at a feeding methodology for the process.
Introduction
Ferrochrome is one of the main raw materials used in the production of stainless steel (typically 10 and 20% Cr). An important benefit thereof is the increase in the steel corrosion and oxidation resistance [1]. Low carbon ferrochrome, required for the production of many steel types, cannot be produced directly by carbothermic reduction of chromite ore due to the amount of carbon that can go into solution with the metallic chromium and iron [2]. Most processes for producing low carbon ferrochrome from high carbon ferrochrome or charge chrome require very high process temperatures, result in high chromium losses or require uneconomically long reaction times [3].
Metallothermic production of low carbon ferrochrome is therefore preferred [4].
A number of metals that are produced in bulk can be used to reduce chromium from chromite ore. The stoichiometric consumption for the most notable ones is shown in Table 1, along with their cost per tonne of chromium metal produced. In practice, the consumptions may differ slightly, depending on the targeted chromium recovery.
From this data, it is evident that aluminium and silicon are economically the most viable options for use as metallothermic reductants. Of the two, only silicon (in the form of ferrochrome
Advances in Molten Slags, Fluxes, and Salts: Proceedings of The 10th International Advances in Molten Slags, Fluxes, and Salts: Proceedings of The 10th International
Conference on Molten Slags, Fluxes and Salts (MOLTEN16) Edited by: Ramana G. Reddy, Pinakin Chaubal, P. Chris Pistorius, and Uday Pal TMS (The Minerals, Metals & Materials Society), 2016
silicide) can readily be produced on a smaller scale on site. This is likely the reason for the prevalence in its use.
Table 1: Metal consumption for Cr production
Metal Consumption
(t/t Cr)
Cost per tonne Cr*
(US$)
Aluminium 0.519 891
Magnesium 0.701 1 578
Manganese 1.585 2 877
Silicon** 0.405 704 - 1 069
* Prices from www.metalbulletin.com (15 August 2015).
** Calculated using 75% FeSi for the lower value and Si metal for the higher value.
Process Routes
Only two process routes (not considering the variations thereof) are considered to be commercially feasible for the production of LC FeCr, namely the Mixing Method and the Furnace Method. These have been described extensively in earlier works [1]. Of the two methods, the most commonly used process is the Mixing Method [4]. Both methods are illustrated in Figure 1. The process steps and furnace components indicated in blue are specific to the Mixing Method, while those in orange are used in the Furnace Method only. The remainder is common between the processes. Solid FeSiCr feed variations for both Methods are also possible (indicated with dotted line).
Figure 1: Mixing and Furnace Method processes
The Mixing Method
Processes that involve mixing of materials (cocktailing) in ladles will be referred to as the Mixing Method, in agreement with the work by Gasik [1]. The Mixing Method is a three-step process. FeSiCr is produced in a semi-open submerged arc furnace. Simultaneously, an ore-lime melt is produced in an open arc furnace. Lime is added to obtain a Cr2O3 content of
C h r o m i t e f i n e s
B u r n t l i m e
C h r o m e o r e
Q u a r t z
C o k e
R i c h a l l o y
> 4 0 % S i
F i n a l a l l o y
6 5 - 7 0 % C r
S l a g ( w a s t e )
S l a g
S l a g ( w a s t e )
< 3 % C r 2 O 3 I n t e r m e d i a t e a l l o y
F U R N A C E M E T H O D
M I X I N G M E T H O D
The products from both of these steps are then mixed (cocktailed) in ladles to produce LC FeCr and a low Cr2O3 slag. Some Cr2O3 is retained to ensure that the Si content in the LC FeCr product is below 1.5%. The simplified reduction reactions are show in Equations 1 and 2. The processes are exothermic, so the temperature in the ladle could be somewhat higher.
๐ถ๐2๐3(๐) + 1.5๐๐(๐) = 2๐ถ๐(๐) + 1.5๐๐๐2(๐) ๏H1900ยบC = -348.194 MJ โฆEq. (1) 2๐น๐๐(๐) + ๐๐(๐) = 2๐น๐(๐) + ๐๐๐2(๐) ๏H1900ยบC = -421.379 MJ โฆEq. (2)
The Mixing Method has some distinct disadvantages, which include high energy losses due to the exothermic reactions during the ladle cocktailing step, high process temperatures that result in high refractory wear, as well as material losses due to spillage. In addition, the oxidising conditions in the ore-lime melt furnace, together with the high process temperature are conducive to the formation of Cr(VI) [6] and could result in the formation of up to 1% CrO3 [1].
The Furnace Method
The Furnace Method is an alternative to the Mixing Method. In the Furnace Method, the FeSiCr is added directly to the ore-lime melt. The Liquid Feed Furnace Method (LFFM) was developed and patented by Mintek in South Africa [7], although not named as such. The reaction equations are similar to those of the Mixing Method. However, a lower operating temperature can be used, as discussed later.
High level mass and energy balances were developed for the Mixing and Furnace Methods, as well as their solid FeSiCr feed variations. In all cases, ore with a Cr2O3 content of 48.7% and FeSiCr with 40% Cr and 42% Si was used. The energy requirements are shown in Table 2. The Furnace Method, particularly the LFFM, offers a significant saving in the electrical energy requirement. This is due to the fact that the silicothermic reduction takes places inside the low carbon ferrochrome furnace, instead of in the ladles. The exothermic energy is therefore available for heating and melting of the ore-lime mixture.
Table 2: Smelting energy consumption for four process options
Process Consumption *
(kWh/t LC FeCr)
Ore-lime and LC FeCr furnace process temperature (ยบC)
Mixing Method (Perrin) 2254 1900
Mixing Method (Duplex, solid FeSiCr feed) 2326 1900
Solid Feed Furnace Method (SFFM) 1589 1750
Liquid Feed Furnace Method (LFFM) 1202 1750
* The energy consumption excludes that for the production of FeSiCr, which is common for all options.
Thermochemical Modelling Setup
At the time of writing there was no commercial installation producing low carbon ferrochrome using the LFFM. No information was therefore available relating to the operational parameters required to promote the product quality and recovery, nor for the refractory and slag systems that are the most appropriate for the process.
Thermochemical modelling was subsequently performed to evaluate different slag compositions and refractory materials to ascertain whether or not LC FeCr can be produced that is within specification, whether or not the process would be operable in terms of the slag liquidus temperature (and superheat), and to determine if the slag and refractory materials would be compatible at the selected process temperature. Modelling was done using FactSage 6.4.
The liquidus temperature of a 70% Cr LCFeCr product is approximately 1700ยบC. A slag superheat of between 50 and 200 degrees Celsius is generally required to achieve a balance between maintaining slag fluidity and refractory protection [8]. The operating temperature should therefore be in the region of 1750ยบC and the slag liquidus between 1550 and 1700ยบC.
Refractory systems were limited to magnesia and doloma (burnt dolomite, CaMgO2), as they are readily available and low cost. Fluxes were limited to lime and doloma, again due to their availability and cost. The systems that were modelled are listed in Table 3. For the purposes of modelling, magnesia and lime were considered to be pure MgO and CaO respectively.
Table 3: Refractory and slag flux systems used in models.
System number Refractory Flux
1 Magnesia Lime
2 Magnesia Doloma
3 Doloma Lime
4 Doloma Doloma
The final slag basicity ((CaO+MgO)/SiO2) for the Mixing Method processes ranges between 2.0 and 2.5. The natural basicity (no fluxes added) of the slag is 0.45. The basicity range used in the evaluation was between these two limits (0.45 to 2.5).
Thermochemical Modelling Results Reaction Products
The alloy silicon content decreases with an increase in the slag basicity. The reason for this is that, as more flux is added, the activity of SiO2 in the slag decreases. This promotes further reduction of the chromite with Si (to form SiO2). Although a Si content of 1.5% is acceptable, 1.0% Si was chosen as the target at equilibrium, in order to compensate for process variability.
Figure 2: Alloy Si content and solid oxide content as a function of slag basicity (T = 1750ยบC)
For a lime fluxed slag, the targeted 1% Si is reached above a basicity of 1.5. Above a basicity of 2.0 a solid metal oxide (MO), consisting of mostly MgO (~95%) and Cr2O3 (~4%), starts forming. For a doloma fluxed system, the 1% Si is again reached above 1.5, but the solid MO already starts forming slightly above a slag basicity of 1.0. An operating basicity between 1.5 and 2.0 would therefore be preferable for a lime fluxed slag, considering the reaction products only. For a doloma fluxed slag, operating above a slag basicity of 1.5 may result in a high viscosity slag due to the presence of the solid MO phase. These results are shown in Figure 2.
Slag Liquidus Temperature
A plot of the slag liquidus temperature (Figure 3) shows both the lime and doloma fluxed slags to have liquidus temperatures within the required 1550 to 1700ยบC range between slag basicities of 1.68 to 1.90 and 1.07 to 1.23 respectively. Using a slag within these ranges would therefore ensure that the slag is fluid enough to be handled with ease, without having an excessively high superheat.
Figure 3: Slag liquidus temperatures for lime and doloma fluxed slags
0%
5%
10%
15%
20%
25%
30%
35%
0%
1%
2%
3%
4%
5%
6%
7%
8%
9%
10%
11%
12%
0.0 0.5 1.0 1.5 2.0 2.5 Solid oxide content (percentage of total oxides)
Alloy Si content
Slag basicity
Alloy Si (Lime flux) Alloy Si (Doloma flux) Solid oxide (doloma flux) Solid oxide (Lime flux)
1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 2100 2200 2300
0.7 0.8 0.9 1 1.1 1.2 1.3 1.4 1.5 1.6 1.7 1.8 1.9 2 2.1 2.2 2.3 2.4 2.5
Liquidus Temperature (ยบC)
Slag Basicity (CaO+MgO)/SiO2 Doloma slag
Lime slag
Refractory-Slag Interaction
The commercial furnace will operate with a slag freeze lining on the refractory hot face. This is achieved by water cooling of the furnace shell. However, at times the freeze lining may be lost when it breaks off from the refractories due to thermal shock, or it may be worn away during periods of poor control. While the freeze lining is being re-established, the turbulence in the slag bath may cause the refractories to be exposed to slag with a temperature near the process temperature.
Modelling the interface between the slag and refractory is complex, as it is effectively a layer with a thickness of one molecule. This also does not take into consideration the effect of stirring, which continually replaces the slag on the interface with bulk slag. A more sensible way of modelling the potential refractory wear is by considering the maximum solubility of the refractory in the slag. The equilibrium slag products from the modelling were added (in new models) to 100 g of magnesia refractory. The liquid slag addition limit was chosen to be 100 g, as the mass of slag in contact with the refractory is not expected to exceed the mass of solid refractory at the refractory-slag interface.
The equilibrium amount of solid oxide material should be equal to or more than the 100 g of refractory that was initially added to the system. This indicates that the refractory is not absorbed into the slag and remains solid. For the systems with magnesia refractories (Systems 1 and 2), the solid MgO was also plotted. This is an important measure, as the total solid oxides can be high due to the precipitation of other oxides, even though the MgO from the refractory is being eroded by the slag. The results of the evaluation are shown in Figure 4.
Figure 4: Solid oxides and solid MgO as a function of slag basicity for refractory slag systems at 1750ยบC (100 g initial refractory and 100 g slag added).
0 20 40 60 80 100 120
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Equilibrium phase mass (g)
Slag Basicity 1. MgO refractory, Lime flux
Solid oxides Solid MgO
0 20 40 60 80 100 120
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Equilibrium phase mass (g)
Slag Basicity 2. MgO refractory, Doloma flux
Solid oxides Solid MgO
0 20 40 60 80 100 120
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Equilibrium phase mass (g)
Slag Basicity 3. Doloma refractory, Lime flux
Solid oxides
0 20 40 60 80 100 120
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Equilibrium phase mass (g)
Slag Basicity 4. Doloma refractory, Doloma flux
Solid oxides
The curves in Figures 4.3 and 4.4 show that the equilibrium amount of solid oxides in the system never exceeds the 100 g of refractory that was added for slag in contact with a doloma refractory.
In fact, it was found that approximately 14% of the refractory would have been taken into solution at the process temperature. Doloma refractory is therefore not suitable for use in this basicity range and temperature.
The results from Figures 4.1 and 4.2 were much more encouraging. For the lime fluxed slag, the total solid oxide and solid MgO contents exceed 100 g at basicities of 1.65 and 1.90 respectively.
For the doloma fluxed slag these values are 1.35 and 1.48 respectively. Comparing these results to the liquidus temperatures plotted in Figure 3, one can see that the ranges overlap for the lime fluxed slag between 1.68 and 1.90, while there is no overlap for the doloma fluxed slag. Of the systems evaluated, the only suitable alternative is therefore a lime fluxed slag in contact with a magnesia refractory.
Operation
Although the furnace in the LFFM is operated on a continuous basis, the slag and alloy undergo distinct compositional changes with time, and is therefore classified as a semi-batch processes.
As exothermic reactions take place in the LFFM LC FeCr furnace, control is somewhat more complex than for the Mixing Method processes. The control of the energy balance is therefore crucial for maintaining the integrity of the refractory lining. The feed ratio and power input are adjusted throughout the heat to maintain a constant energy balance and process temperature.
Deciding on when to feed the different raw materials in the LFFM process is probably the greatest challenge of the entire operation. Feeding all of the ore-lime mixture at the start of the batch will result in a slag with a high liquidus temperature, which will likely solidify in the furnace. However, feeding the FeSiCr at the start of the batch poses a severe risk to the refractory integrity.
FactSage models, with slag and alloy at equilibrium, show the magnesium metal content to be extremely high when gradually adding the ore-lime mixture to liquid FeSiCr (Figure 5).
Although some of the Mg may originate from the slag, it is equally likely to come from the refractory. A balance therefore needs to be found between maintaining the integrity of the furnace refractory, while having a liquid slag.
Figure 5: Metallic magnesium and silicon contents for basicities of 1.0 to 2.5 when adding ore-lime mixture to FeSiCr.
In practice, this requires a calculation of how much of the ore-lime mixture can be heated and melted by the exothermic energy from silicothermic reduction. This fraction of the ore-lime mixture should be fed at the end of the batch (after the FeSiCr), while the balance of the ore-lime mixture should be fed at the start of the batch. There should therefore be as much as possible of the ore-lime mixture in the furnace before feeding the FeSiCr. The Si would then preferentially react with the metal oxides from the ore, instead of the MgO in the refractory. Feeding the FeSiCr close to the centre of the furnace would be preferred to side feeding for two reasons.
Firstly, contact between molten FeSiCr and the sidewall is limited and secondly, feeding close to the arc will improve reaction kinetics. Where the Mixing Method relies on the momentum imparted by the cocktailing process to ensure good contact between the reactants, the LFFM relies on the momentum from the electrical arc, as well as the large slag-metal interface in the furnace (relative to that of a ladle).
Conclusions
The Furnace Method, in particular the Liquid Feed Furnace Method, provides a saving in the electrical energy requirement when compared to the Mixing Method processes.
For both doloma and lime fluxed slags, the required alloy Si content (1%) is achieved above a slag basicity of 1.5. However, the doloma slag only has a liquidus temperature within the 1550 to 1700ยบC range between slag basicities of 1.07 and 1.23. Doloma fluxed slag is therefore not suitable for producing low carbon ferrochrome. A lime fluxed slag has a liquidus temperature within the required 1550 to 1700ยบC range for a slag basicity range of 1.68 and 1.90. The slag is also compatible with a magnesia lining between a slag basicity of 1.65 and 1.90. This system therefore has a suitable basicity range within which to operate.
Doloma refractories can also not be used for the production of low carbon ferrochrome at the required operating temperature, as approximately 14% of the refractory was found to be taken into solution at a process temperature of 1750ยบC.
0.0%
0.5%
1.0%
1.5%
2.0%
2.5%
3.0%
3.5%
4.0%
4.5%
5.0%
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
40.0%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
Metal Mg content (including Mg(g))
Metal Si content
Ore-lime Added
Si (B=1.0) Si (B=1.5) Si (B=2.0) Si (B=2.5) Mg (B=1.0) Mg (B=1.5) Mg (B=2.0) Mg (B=2.5)
Acknowledgements
The authors would like to thank the Department of Materials Science and Metallurgical Engineeringโs Centre for Pyrometallurgy at the University of Pretoria for providing the facilities and funding for this project.
References
[1] M. Gasik, "Technology of chromium and its ferroalloys," in Handbook of Ferroalloys - Theory and Technology, M. Gasik, Ed. Oxford: Butterworth-Heinemann, 2013, pp. 267-316.
[2] A.G.E. Robiette, Electric smelting processes. London: Charles Griffin & Company Limited, 1973.
[3] P.J. Bhonde, A. M. Ghodgaonkar, and R. D. Angal, "Various techniques to produce low carbon ferrochrome," in Proceedings of the Eleventh International Ferroalloy Conference, New Delhi, 2007, pp. 85-90.
[4] European IPPC Bureau, "Integrated pollution prevention and control (IPPC) reference document on best available technique in the non ferrous metals industry," Seville, 2001.
[5] S. Ghose, J. K. Nanda, and B. B. Patel, "Duplex process for production of low carbon ferrochrome," in Proceedings of the seminar on problems and prospects of ferro-alloy industry in India, Jamsedpur, 1983, pp. 215-217.
[6] J.P. Beukes, N.F. Dawson, and P.G Van Zyl, "Theoretical and practical aspects of Cr(VI) in the South African ferrochrome industry," The Journal of the South African Institute of Mining and Metallurgy, vol. 110, pp. 743-750, 2010.
[7] H.L. Smith, G. M. Denton, and N. A. Barcza, "Ferrochromium production," 96/0877, 1996.
[8] M.W. Kennedy, "Electric slag furnace dimensioning," , Orlando, 2012.