Secondary Steel Metallurgy Slag Design and MgO-C Refractory Implications
Theoretical and Practical Considerations
Simon Hellgren
Sustainable Process Engineering, master's level 2019
Luleå University of Technology
Department of Civil, Environmental and Natural Resources Engineering
A CKNOWLEDGEMENTS
I would like to dedicate this thesis work to my family, which have been very supportive throughout my years as a M.Sc. student. Their love and support have been very meaningful to me.
Firstly, I would like to thank my supervisor and course examiner Fredrik Engström for his involvement and interest in this thesis work. His advice has been of great importance and aid for me during the work, and it is well appreciated. Many thanks to all the colleagues at the department of Minerals and Metallurgical Engineering at LTU who has been able to answer all my questions and for their involvement in my thesis work.
I would further like to thank Höganäs AB for the opportunity to be a part in the project, specifically my external supervisors Liviu Brabie and Tommie Edeblom. The subject of the thesis work has been very interesting to be involved in.
Lastly, I would like to express my gratitude towards my friends who has kept these years as a student exciting.
Simon Hellgren Luleå, October 2019
A BSTRACT
MgO-C based refractory materials, often used in secondary steel making, are exposed to various wear mechanisms in its application. The wear could be divided into oxidative, chemical and abrasive categories, all behaving differently and all being influenced by different factors. Due to the importance of minimizing material loss and to the environmental challenges to run a sustainable industry, it is of interest to gain more knowledge of the behavior of the refractory material in use. The present thesis work specifically investigated slag designed of the CaO- SiO2-Al2O3-MgO (CSAM) system as well as the chemical and oxidative wear mechanisms of three different MgO-C based refractory materials from Höganäs AB, Halmstadverken, which contained 5, 10 and 12 wt% carbon (labeled T05, T10 and T12). Different CSAM slags were designed to meet certain process criteria such as MgO and CaO saturations and were investigated through thermodynamic calculations using the FactSage software and through laboratory scaled slag smelting experiments. The oxidation effect on the refractory material was also studied through pre-heating simulations in chamber furnaces, similar to the pre-heating of a re-built ladle furnace.
The thermodynamic calculations made in FactSage 7.0, using the FactPS and FToxid data bases, resulted in a few different slag designs with different properties. A few different slags fulfilled the CaO and MgO saturation limits proposed by Höganäs AB and could be considered to test experimentally for further evaluation. The simulations also showed trends on how the liquid viscosity behaved with different slag compositions and how the solids content changed with temperature.
The oxidation experiments were performed on the different MgO-C refractory types, where the bricks with 10% carbon also contained Al2O3 antioxidants. The experiments showed that the mass loss during the pre-heating is greater for refractory with higher carbon content, with exception to T10, where the mass losses were measured to 3.76 – 4.01%, 1.06 – 1.28% and 6.28 – 6.33% for T05, T10 and T12 respectively. Further, the oxidation depth of each material was measured to 9-10 mm, 2-3 mm and 2-5 mm for the T05, T10 and T12. The experiments also showed that T12 refractory in particular was very susceptible to abrasive wear after being oxidized.
The slag smelting experiments were carried out through two different methods, by melting slag in MgO-C crucibles and by melting pressed slag briquettes on top of refractory bricks. The former tests mainly showed that the methodology was not suitable for this type of refractory material due to the crucibles cracking during the experiments. The latter experiments showed some general behaviour of the different components in the slag, where Ca, Al and Fe stayed near the surface, and Si and Mg penetrated deeper. The spinel formation at the refractory surface was then concluded to be the reason for Al not penetrating deeper. Further it was concluded that no significant difference in refractory dissolution was seen between slags with- and without
L IST OF ABBREVIATIONS
MgO-C: Refractory material made out of Magnesia and Carbon.
CSAM: The CaO-SiO2-Al2O3-MgO slag system.
T05: MgO-C Refractory material containing 5 weight% carbon
T10A: MgO-C Refractory material containing 10 weight% carbon and Alumina anti-oxidants T12: MgO-C Refractory material containing 12 weight% carbon.
Slags M1 – M10: Describes the different slags used in the laboratory experiments, where M1, M3, M5, M7 and M9 contained no MgO.
M2, M4, M6, M8 and M10 contained MgO.
Basicities B2, B3, B4 and B, describing the ratio between Basic and Acidic oxides in a slag:
𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 𝐵𝐵2 =𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 𝐴𝐴𝐵𝐵𝐵𝐵𝐴𝐴𝐵𝐵 =
% 𝐶𝐶𝐵𝐵𝐶𝐶
% 𝑆𝑆𝐵𝐵𝐶𝐶2
𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 𝐵𝐵3 =% 𝐶𝐶𝐵𝐵𝐶𝐶 + % 𝑀𝑀𝑀𝑀𝐶𝐶
% 𝑆𝑆𝐵𝐵𝐶𝐶2 𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 𝐵𝐵4 = % 𝐶𝐶𝐵𝐵𝐶𝐶 + % 𝑀𝑀𝑀𝑀𝐶𝐶
% 𝑆𝑆𝐵𝐵𝐶𝐶2+ % 𝐴𝐴𝑙𝑙2𝐶𝐶3 𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 𝐵𝐵 = % 𝐶𝐶𝐵𝐵𝐶𝐶 + 0.67% 𝑀𝑀𝑀𝑀𝐶𝐶
% 𝑆𝑆𝐵𝐵𝐶𝐶2+ % 𝐴𝐴𝑙𝑙2𝐶𝐶3
Table of Contents
1 INTRODUCTION ... 1
1.1 BACKGROUND ... 1
1.2 PROJECT AIM & LIMITATIONS ... 4
2 LITERATURE REVIEW ... 6
2.1 PREVIOUS METHODOLOGY ... 6
2.2 CHEMICAL- AND ABRASIVE WEAR, DISSOLUTION OF REFRACTORY MATERIAL ... 6
2.3 OXIDATION EFFECTS ON REFRACTORY MATERIAL ... 8
2.4 CONSIDERATIONS DURING SLAG DESIGN ... 8
3 MATERIALS & METHODOLOGY ... 10
3.1 THERMODYNAMIC SIMULATIONS ... 10
3.2 PRE-HEATING SIMULATION OF LADLE-REFRACTORY ... 13
3.3 SLAG SMELTING TRIALS ... 14
4 RESULTS ... 19
4.1 THERMODYNAMIC CALCULATIONS ... 19
4.2 PRE-HEATING SIMULATIONS ... 23
4.3 SLAG SMELTING IN REFRACTORY CRUCIBLES ... 25
4.4 SLAG-BRIQUETTE SMELTING EXPERIMENTS ... 26
5 DISCUSSION ... 31
5.1 THERMODYNAMIC SIMULATIONS ... 31
5.2 PRE-HEATING SIMULATIONS ... 31
5.3 SLAG SMELTING TRIALS ... 31
6 CONCLUSIONS ... 33
7 FURTHER STUDIES ... 34
8 WORKS CITED ... 35
1 I NTRODUCTION
1.1 BACKGROUND HÖGANÄS ABHöganäs AB is a big part of the international metal industry and is today leading on the world market for iron and metal powders, with their yearly production of 500 000 tonnes. Höganäs’
vision as a company is to inspire industry to produce more with less, which is founded in the reduced energy and material costs while handling metal powders in specific. Höganäs lays high priority in helping to develop a sustainable future metal industry through better available techniques that promotes less expensive and more efficient products.
Due to today’s importance of a sustainable industry, Höganäs always aims to surpass environmental regulations. Even though the focus for many industries often lies on the final product, it is just as important to consider material costs and energy efficiency in the different processes implemented to produce that product. Factors such as process optimization connected to a decrease in equipment wear are taken into consideration more thoroughly in today’s industry, partly due to the economic benefits but also due to environmental considerations1 . HALMSTADVERKEN
Halmstadverken, located in Halmstad in Sweden, produces iron powder through a water- atomization process. The powder product is sent to Höganäs for refining, where different customer specific qualities are produced. A basic flow sheet of Halmstadverken is presented in Figure 1.
The electric arc furnace (EAF) is charged with selected scrap and slag formers. After smelting the scrap, liquid steel is tapped into a ladle furnace for further treatment through different additives, such as new slag formers. The treated steel is tapped from the ladle furnace into a tundish, which in turn provides a continuous stream of molten steel to the water atomizer. The final iron powder product is dried, screened and equalized before it is sent to Höganäs for further refining.
Figure 1 – Iron powder production flow sheet at Halmstadverken, Höganäs AB.
The type of ladle furnaces used at Halmstadverken is pictured in Figure 2, where the different ladle zones and refractory types are presented. The inside of the ladle furnace is covered with Magnesia-Carbon bricks (MgO-C), to prevent heat loss.
Figure 2 – A cut-view of the ladle furnaces used at Halmstadverken, showing a more detailed view.
The three different refractory types that are used in this type of ladle furnace are:
• S3T10A, at the upper refractory wall with 10% carbon which also contains alumina anti-oxidants, represented by the top grey layer in the figure below. These refractory pieces are not in contact with either slag or steel, with the exception to the discharge of slag after metal tapping.
• REF6T12, the slag line refractory which contains 12% carbon. This is represented by the top blue layer. The carbon content is the highest in this refractory type due to its constant contact with slag (metal oxides) during operation.
• RES3T05, the lower refractory wall with 5% carbon, which is represented by the yellow layer below the slag line. Only the liquid steel is in contact with this refractory type.
The ladle furnace act as an insulator to the metal bath while transporting, at the same time as providing a possibility to add alloying elements during agitation to reach certain quality standards. While the refractory is required to prevent damaging the ladle furnace, it is constantly worn down at different rates depending on the different refractory types.
MAGNESIA-CARBON REFRACTORY
Magnesia-Carbon (MgO-C) bricks are a commonly used type of refractory material in the secondary steel treatment processes. This refractory material is built up with a fine MgO-grain based matrix, with finer MgO particles filling up the pores together with carbon to reduce the specific surface area. By reducing the specific surface area the slag- and iron melt penetration can be decreased. This type of brick is relatively resistant to wear when in contact with the most frequently slag used in secondary steel treatment, the CaO-SiO2-Al2O3-MgO (CSAM) system2. This type of slag has been modified to fit the secondary steel treatment processes to reduce the overall wear on the expensive refractory bricks. An increased wear on the refractory will not only result in more heat losses from the molten steel reactor, but it also has the potential of damaging the reactor itself, due to the high temperatures of the slag and melt. Refractory wear is unpreventable and depending on process conditions different refractory types last longer than other. Therefore, it is required to rebuild the furnace after some time to prevent unnecessary equipment damage. However, due to both economic and environmental reasons the rebuilding of a ladle furnace is undesired, and the lifetime of the refractory is always sought to be prolonged as long as possible, which is not an easy task.
The refractory in a ladle furnace are exposed to wear through different mechanisms. The bricks are affected differently by different conditions such as oxidation, chemical wear from dissolution, abrasion and spalling caused by stirring or other mechanical forces. Due to the high carbon content in the refractory bricks they become very susceptible to oxidative damage at higher temperatures, both at atmospheric conditions, e.g. during pre-heating of a newly rebuilt ladle furnace, but also in process conditions due to the presence of metal oxides in the slag.
The chemical wear is mostly related to the rate of dissolution of magnesia from the refractory in to the slag. This is determined by the saturation concentration of MgO for a specific slag in use compared to the actual concentration dissolved in the slag. The saturation concentration in a slag is dependent on different variables such as temperature and composition and is therefore difficult to control3,4, but can be estimated through thermodynamic calculations. The temperature will alter the saturation concentrations of the slag system, hence altering the dissolution rate of the refractory material5.
The oxidative wear on MgO-C-based refractory is controlled by the oxidation potential of the carbon in the bricks, which is promoted by the presence of metal oxides. Components such as iron oxides being carried over from the smelting process as well as the metal oxide-based slag will have negative effects on the wear of the MgO-C-bricks. The oxidation of carbon to carbon monoxide leaves pores in the bricks, which in turn promotes slag penetration for a significant increase in chemical and mechanical wear6 . Oxidative wear is also considered during the pre- heating of a newly re-built ladle furnace, where fresh refractory bricks are installed and pre- heated to remove moisture and reduce the thermal shock when liquid steel is tapped in the ladle for the first time.
Abrasive wear is difficult to prevent in ladle metallurgy. The efficiency of the refining of the steel is dependent on a well stirred metal bath, to give a good slag-metal interface. Ladles stirred with inert gases are however susceptible to higher refractory wear closer to the gas injection and will experience higher wear rates compared to the opposite walls. Due to this phenomenon the lifetime of the ladle refractory is significantly shortened, as compared to a process without agitation.
Hence, the different parameters responsible of the ladle refractory wear has to be thoroughly investigated to get a good understanding of the overall lifetime of the refractory. An expanded refractory lifetime could result in significant economic gain, and a better resource efficiency.
When the refractory is worn out, it is taken to a station to be re-built with new refractory bricks.
As the re-built furnace is ready to be taken back into the steel production, it is first pre-heated for 16h at 1000°C, with a slow temperature rise of about 10°C/min. The pre-heating minimizes the thermal shock the refractory material experiences when the hot metal is tapped in to the ladle furnace for the first time, but it also vaporizes any water or humidity present in the refractory bricks. When the pre-heating is completed the furnace is ready to be brought back into the process.
1.2 PROJECT AIM & LIMITATIONS
The purpose of the present thesis work was to specifically investigate the chemical wear and slag penetration in which the slag line refractory in a ladle furnace is exposed to through different slag-refractory interactions. It was of interest at Halmstadverken to design a ladle furnace slag which had a lower viscosity than their current one, where an increased amount of
dissolution respectively. Further, by simulating the oxidizing conditions present during the pre- heating of a re-built ladle furnace, the significant effect the oxidizing conditions have on the fresh refractory material was to be investigated.
The thesis work was limited to investigate and gather basic knowledge of different slag types within the CaO-SiO2-Al2O3-MgO (and FeO) system throughthermodynamic calculations in the FactSage software and through slag smelting experiments in laboratory scale. The pre-heating simulations were limited to heating refractory pieces in chamber furnaces at determined temperature profiles and atmospheric conditions, also carried out in laboratory scale.
2 L ITERATURE REVIEW
2.1 PREVIOUS METHODOLOGYPrevious investigations of similar slag-refractory interactions have applied different methods.
The method of choice affects the results significantly and depending on what interactions are of interest, different methods may be applied. Examples of previous studies’ methodology are:
• Rotating cylindrical refractory rods in a slag or smelting slag in crucibles with inert gas stirring, to account for both chemical and abrasive wear.4,7
• Dissolution of refractory pieces in a designed slag in crucibles. Through this method the chemical wear is mainly studied.8,9
• In some cases, crucibles were made out of refractory material to limit the study to the chemical wear but to also be able to investigate the effect of slag penetration10, 11. 2.2 CHEMICAL- AND ABRASIVE WEAR, DISSOLUTION OF REFRACTORY MATERIAL The dynamics of the chemical wear of MgO-C based refractory are mainly determined by the slag’s ability to dissolve the refractory into itself. Chemical wear could be described as a series of different phenomena, i.e. corrosion mechanisms such as slag penetration, dissolution of refractory compounds, spalling due to corrosion and oxidation/reduction processes4. It has been shown that a slag with magnesia-based additions will respond with less refractory dissolution, due to the decrease in difference between magnesia concentration in the slag and the saturation concentration for that given system. According to Suvorov & Kozlov3, the rate at which the refractory is dissolved and transferred into the slag could be described as equation 1:
𝑄𝑄 =𝐷𝐷
𝐴𝐴 𝑆𝑆�𝐶𝐶𝑀𝑀𝑀𝑀𝑀𝑀∞ − 𝐶𝐶𝑀𝑀𝑀𝑀𝑀𝑀� (1) Where Q describes the rate of transfer of MgO from the surface of the phase boundary; D is the diffusion coefficient which is determined by the specific component being dissolved, the viscosity of the slag as well as the temperature; d is the diffusion layer thickness at the refractory-slag interface, which is dependent on the slag viscosity and the relative velocity at the surface of the refractory; S is a coefficient that accounts for the increase in surface area due to the increase in pores during the continuous dissolution; C∞MgO is the saturation concentration of MgO and CMgO is the actual concentration of MgO in the slag. A larger difference in saturation concentration and actual slag concentration will have a large impact on the dissolution rate and is therefore important to keep as small as possible. The saturation concentration has been determined to be different for each slag system, since it is dependent on temperature and slag composition. Process parameters that could have an effect on the dissolution rate directly are the stirring speeds, which will reduce the diffusion layer thickness;
the viscosity of the slag, which could affect the diffusion coefficient; and difference between
Huang et al5 investigated the dissolution behavior of different refractory cubes (Sintered MgO, MgO-C and MgO-Cr2O3) in a CaO-SiO2-Al2O3 slag, with inert gas (pure argon) injections at different stirring intensity and 1500°C. Further, the effect of oxidizing conditions on the different refractory material’s dissolution in the slag was investigated by adding FeO to the slag. By placing a small refractory cube in a slag bath and injecting inert gas, the slag composition before and after the dissolution of the MgO based refractory piece could be compared to determine the effect of the stirring intensity. The experiments showed that the dissolution rate of the refractory pieces increased significantly with an increasing stirring intensity, as seen in Figure 3. Furthermore, the addition of FeO to the slag showed an increase in dissolution rate in the MgO-C based refractory pieces, due to the oxidation of the carbon.
Figure 3 – MgO-C refractory dissolution with increasing inert gas flow 5.
Wang et al8 studied the dissolution of four different types of MgO based refractory in a synthetic CSAM-slag (53-54% CaO, 6-7% SiO2, 31-33% Al2O3 and 8-9% MgO). The different refractory pieces were exposed to varying conditions such as increasing stirring speeds and stirring times while investigating how much of the refractory piece dissolved into the slag, or how much of its cubic shape remained. By letting the refractory cubes dissolve in melted slag at 1600°C, it could be seen that a stagnant or slowly stirred slag resulted in significantly less MgO dissolution compared to more intense stirring. The authors presented that the main reason for the decrease in dissolution of the MgO cubes was less internal mass transfer within the pores in the refractory matrix, due to a thicker layer of slag penetration which in turn hindered further penetration. An increase in stirring speed would significantly decrease the slag penetration layer, which would promote an increase in internal mass transfer and therefore promote the MgO dissolution.
Kasimagwa et al4 performed a study where rods of commercial refractory material (MgO-C) were rotated in a synthetic slag at a temperature range of 1500 – 1650°C. The slags were saturated with different amounts of MgO, 6% and 12%, to investigate the effect of slag saturation on refractory wear. The experiments showed a significant decrease in wear for the slag containing 12% MgO compared to 6%, hence suggesting that the saturation concentration has an effect on refractory wear. Further, it was observed that the temperature had an impact on
the dissolution rate, where an increased temperature enhanced the dissolution process by affecting the saturation concentration of the slag system.
Mechanical wear is often a consequence of stirring in the metal bath. The agitation is necessary to achieve a good metal-slag interaction but has a negative and significant increase on the refractory wear. Jansson et al7 investigated the wear of MgO-C based refractory rods rotated in a typical CSAM slag, at different stirring rates over time. The refractory wear was compared by exposing MgO-C rods to slags with different MgO content at different rotation speed, where the change in diameter of the rods was used to measure the wear. From the different tests it was clearly noted that rotation speed was a significant parameter regarding the wear rate of the rods, as an increase in rotation speed by a factor of 4 would almost double the wear.
2.3 OXIDATION EFFECTS ON REFRACTORY MATERIAL
The addition of graphite to MgO-refractory bricks introduces two beneficial properties: high thermal shock resistance due to the high thermal conductivity in graphite, as well as a reduction in slag penetration in the MgO matrix due to the low slag-wettability of graphite9,6,12. As Liu et al13 showed in their investigation of dense MgO compared to MgO-C refractory, the addition of carbon in the MgO-C refractory significantly decreased wetting behavior of the slag. The graphite is however very susceptible to oxidation at process conditions, which will result in an increase in pore volume and a promotion of slag penetration. The decrease of graphite could be classified as direct and indirect oxidation. Direct oxidation is the phenomena where carbon is consumed by gaseous oxygen as seen in Equation 2, whereas indirect oxidation is dependent on solid oxides in the slag phase at greater temperatures, see Equation 3.6
2𝐶𝐶(𝐵𝐵) + 𝐶𝐶2(𝑀𝑀) → 2𝐶𝐶𝐶𝐶(𝑀𝑀) (2)
𝐶𝐶(𝐵𝐵) + 𝑀𝑀𝐵𝐵𝐶𝐶𝑋𝑋(𝐵𝐵) → 𝑋𝑋𝐶𝐶𝐶𝐶 + 𝑀𝑀𝐵𝐵(𝑀𝑀) (3) 2.4 CONSIDERATIONS DURING SLAG DESIGN
When designing a steelmaking slag specifically for ladle metallurgy, a few general factors should be considered 2, 10, 14:
• The slag should fulfil certain metallurgical requirements, such as effective desulphurization capacity and absorption of inclusion and oxidized impurities from the metal bath.
• The slag layer should act like a good insulator to the atmosphere, both to reduce heat losses but also to prevent atmospheric elements such as nitrogen and oxygen to dissolve into the metal melt.
• The slag should protect the refractory, e.g. not promote refractory dissolution.
𝐶𝐶𝐵𝐵𝐶𝐶 + [𝑆𝑆] → 𝐶𝐶𝐵𝐵𝑆𝑆 + [𝐶𝐶] (4) Where the components within closed brackets are dissolved in the steel. A saturation of CaO in a slag is therefore promoting desulphurization by shifting the equilibrium of equation 4 towards the right14, 15 . Due to both the benefits of a limiting refractory wear rate and the enhanced removal of Sulphur, it is in general advantageous to have an MgO- and CaO saturation in steel making slags.
J. Liao et al16 investigated the effect of a changing Al2O3/SiO2-ratio in a CSAM type slag.
Previous studies had shown that a (Al2O3)/(Basic oxides) ratio > 1 in a CSAM system would decrease the viscosity, as long as the SiO2 content was kept constant. The work carried out by Liao et al concluded, among other things, that the viscosity of a CSAM type slag would decrease slightly with an increasing Al2O3/SiO2-ratio.
With these different factors in mind the slag design has to be adjusted according to each plant’s specific requirements. The requirements of a slag will vary between different plants due to different process conditions and quality demands on the final product, which will affect the slag design. The often more difficult to control parameters of a slag design would be viscosity; the liquid temperature, to ensure a liquid slag; and slag composition, which in turn affects saturation concentrations and the possibilities of later reuse.
The behavior of a given slag system is at times difficult to confirm, since factors such as slag composition can vary based on the feed material. Basicity measurements B2, B3, B4 and B are four different correlations between acidic and basic slag components that are commonly used to get an indication of the slag behaviour. The different measurements take into account different slag components, where B3, B4 and B are different extensions of the basic B2 measurements17 :
𝐵𝐵2 =𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵𝐵 𝐴𝐴𝐵𝐵𝐵𝐵𝐴𝐴𝐵𝐵 =
% 𝐶𝐶𝐵𝐵𝐶𝐶
% 𝑆𝑆𝐵𝐵𝐶𝐶2 (5)
𝐵𝐵3 = % 𝐶𝐶𝐵𝐵𝐶𝐶 + % 𝑀𝑀𝑀𝑀𝐶𝐶
% 𝑆𝑆𝐵𝐵𝐶𝐶2 (6)
𝐵𝐵4 = % 𝐶𝐶𝐵𝐵𝐶𝐶 + % 𝑀𝑀𝑀𝑀𝐶𝐶
% 𝑆𝑆𝐵𝐵𝐶𝐶2+ % 𝐴𝐴𝑙𝑙2𝐶𝐶3 (7)
𝐵𝐵 = % 𝐶𝐶𝐵𝐵𝐶𝐶 + 0.67% 𝑀𝑀𝑀𝑀𝐶𝐶
% 𝑆𝑆𝐵𝐵𝐶𝐶2+ % 𝐴𝐴𝑙𝑙2𝐶𝐶3 (8)
It is also of importance to produce a slag that is stable after cooling in atmospheric conditions.
By ensuring this, it is possible to reuse the slag e.g. as foundation in infrastructure, as an alternative to CaCO3 in the cement industry, in water cleaning processes to remove metallic impurities as well as in many more different areas.18 By treating the slag as a product rather than a by-product, it is easier to maintain a more sustainable industry.
3 M ATERIALS & METHODOLOGY
Due to both the theoretical and practical part of the master thesis work, a brief project plan describing the different parts was created. The theoretical parts included a fundamental study of the CSAM system using different phase diagrams, as well as designing slags through thermodynamic simulations using the FactSage 7.0 software. The practical parts included pre- heating simulations of MgO-C refractory material of different types, slag melting in MgO-C crucibles as well as a heating microscopy-type investigation using pressed slag briquettes melted on flat MgO-C pieces. The different practical experiments were assumed to, together with the theoretical parts, give a good overview of the slag-refractory interactions and the refractory’s susceptibility to oxidation.
3.1 THERMODYNAMIC SIMULATIONS
FUNDAMENTAL INVESTIGATION OF THE CSAM-SYSTEM
To get an understanding of the CSAM system, different phase diagrams in Slag Atlas19 and Phase Diagrams for Ceramists20 was compared. Since part of the project aim was to decrease the viscosity of the slag by increasing the alumina (Al2O3) content, the ternary CSM-system was briefly evaluated at different Al2O3-fractions as follows:
%𝐶𝐶𝐵𝐵𝐶𝐶 + %𝑆𝑆𝐵𝐵𝐶𝐶2 + %𝐴𝐴𝑙𝑙2𝐶𝐶3+ %𝑀𝑀𝑀𝑀𝐶𝐶 = 100% (𝑏𝑏𝐵𝐵 𝑤𝑤𝐵𝐵𝐵𝐵𝑀𝑀ℎ𝐵𝐵) (9) Where the Al2O3 content was varied between 5-30 weight% with a 5 weight% increase for each diagram. Simple isoplethal studies were performed on each of the six diagrams with different Al2O3 content, with the primary crystallization field set in the Periclase (MgO) area. The primary crystallization field was chosen based on the general decrease in refractory wear by the presence of a MgO saturation, as mentioned in section 2.2.
Figure 4 shows the CSAM system with 30 weight% Al2O3, with points A1 and B1 in the primary crystallisation field of Periclase. The blue and green solid lines in the diagram shows the cooling paths for A1 and B1 respectively, as well as their final points A2 and B2, describing the final crystal composition.
Figure 4 - CSAM ternary system with 30% Al2O37. The red dotted lines represent the alkemade lines for different minerals.
Similar isoplethal studies were performed on the five other diagrams, all with the primary crystallisation field in the Periclase region to get an overview of the CSAM system. All melt compositions for each starting point in the six CSAM phase diagrams were compiled and was later evaluated using the FactSage 7.0 software. The different compositions and in which phase triangle the starting points lies were presented in Table 1.
Table 1 – Basic isoplethal studies on each of the six ternary CSM-phase diagrams with varying alumina content.
Al2O3 content (wt%) Point, phase triangle Melt composition
CaO (wt%) SiO2 (wt%) MgO (wt%)
5 % A1, MgO-Ca2SiO4-Ca3SiO5 53.1% 14.8% 27.1%
5 % B1, MgO-CaO- Ca3SiO5 59.1% 15.5% 20.4%
5 % C1, MgO-Ca2SiO4-Ca3Mg(SiO4)2 38.6% 21.9% 34.5%
10 % A1, MgO-Ca2SiO4-Ca3SiO5 47.9% 14.3% 27.8%
10 % B1, MgO-CaO- Ca3SiO5 57.4% 12.7% 19.9%
10 % C1, MgO-Ca2SiO4-Ca3Mg(SiO4)2 39.4% 19.9% 30.7%
15% A1, MgO-Ca2SiO4-Ca3SiO5 52.0% 12.7% 20.3%
15% B1, MgO-CaO- Ca3SiO5 55.3% 12.1% 17.5%
20% A1, MgO-Ca2SiO4-Ca3SiO5 48.9% 12.9% 18.2%
20% B1, MgO-CaO- Ca3SiO5 57.6% 9.9% 12.5%
25% A1, MgO-Ca2SiO4-Ca3SiO5 44.4% 13.8% 16.8%
25% B1, MgO-CaO- Ca3SiO5 55.4% 10.5% 9.1%
30% A1, MgO-Ca2SiO4-Ca3SiO5 39.4% 15.8% 14.8%
30% B1, MgO-CaO- Ca3SiO5 53.4% 10.4% 6.2%
To be able to determine which slag compositions within the CSAM-system were reasonable to investigate further, thermodynamic calculations were made using FactSage 7.0 software, in the Equilib-module with the selected FactPS and FToxid databases. The systems were investigated briefly where the amount of the different phases, i.e slag and Ca- and Mg based metal oxides, were plotted against the temperature.
EXPERIMENTAL CaO-SiO2-Al2O3-MgO SLAG DESIGN
Due to the expense of synthetic slag components, the slag design was based on the average slag composition used in production at Höganäs Halmstadverken and Table 1. By expanding the high and low compositional ends of the commonly used slag compositions, 25 slags were designed all with increasing Al2O3 content, and with a pre-set amount of iron oxide content.
The slag from Höganäs Halmstadverken, as well as 25 slag designs were compiled in Table 2.
Table 2 –Typical slag composition used at Höganäs Halmstadverken and 25 different slag designs.
Slag CaO % SiO2 % Al2O3 % MgO % FeO % B2 B3 B4 B
Halmstadverken,
typical slag 50-63 4-10 12-15 7-10 6-11 5-13 5-15 2-4 2-4
1 65 5 10 10 10 13.00 15.00 5.00 4.78
2 60 5 15 10 10 12.00 14.00 3.50 3.34
3 55 5 20 10 10 11.00 13.00 2.60 2.47
4 50 5 25 10 10 10.00 12.00 2.00 1.89
5 50 4 30 6 10 12.50 14.00 1.65 1.59
6 65 10 10 10 5 6.50 7.50 3.75 3.59
7 60 10 15 10 5 6.00 7.00 2.80 2.67
8 55 10 20 10 5 5.50 6.50 2.17 2.06
9 50 10 25 10 5 5.00 6.00 1.71 1.62
10 50 9 30 6 5 5.56 6.22 1.44 1.39
11 65 15 10 10 0 4.33 5.00 3.00 2.87
12 60 15 15 10 0 4.00 4.67 2.33 2.22
13 55 15 20 10 0 3.67 4.33 1.86 1.76
14 50 15 25 10 0 3.33 4.00 1.50 1.42
15 50 14 30 6 0 3.57 4.00 1.27 1.23
16 50 20 10 10 10 2.50 3.00 2.00 1.89
17 50 15 15 10 10 3.33 4.00 2.00 1.89
18 50 10 20 10 10 5.00 6.00 2.00 1.89
19 50 5 25 10 10 10.00 12.00 2.00 1.89
20 50 0 30 10 10 N/A N/A 2.00 1.89
21 40 30 10 10 10 1.33 1.67 1.25 1.17
22 40 25 15 10 10 1.60 2.00 1.25 1.17
23 40 20 20 10 10 2.00 2.50 1.25 1.17
24 40 15 25 10 10 2.67 3.33 1.25 1.17
25 40 10 30 10 10 4.00 5.00 1.25 1.17
The slag designs were based on the different basicity-measurements (B2, B3, B4 and B) where the compositions were balanced to either achieve an increasing or decreasing B2 basicity in a systematic manner. Due to the constant increase in alumina content, and the constant FeO content, it was at times difficult to balance realistic slags. Therefore, some slags did not fit a realistic design, due to either too low silica content or extreme basicity values. Even if these slags were extremes in regard to realistic slag design, they were determined to be useful to have as a foundation to create a better overview over the CSAM system.
All slag compositions were inserted in FactSage 7.0, using the equilib module with the two selected FactPS and FToxid data bases, and the data was later compiled in excel sheets. Since the viscosity of the slags was considered a very important parameter in this project, the liquid viscosities of each of the 25 slags were calculated using the viscosity module in FactSage 7.0.
Out of these 25 different slag designs, two of the best considered slags were chosen to be tested in smelting trials involving MgO-C refractory. Furthermore, through a discussion with Höganäs AB, three additional slags were planned to be tested. These five slags were adjusted in MgO content according to Table 3, to be able to compare MgO-dissolution from the crucibles. The slags based on the same compositions, i.e. M1 and M2, M3 and M4 etc., were designed to have different MgO content but to be able to make some sort of comparison between them, the CaO/SiO2-ratio (B2) was kept relatively constant. This way, it was assumed to later be possible to draw some conclusions on MgO-dissolution depending on the initial MgO content in the slags. The melting points of the 10 designed slags was calculated using the equilib-module in FactSage 7.0, to get an understanding of the behaviour of the different slags.
Table 3 – Experimental slag design.
Experimental slag designs CaO
(%) SiO2
(%) Al2O3
(%) MgO
(%) FeO
(%) B2
Amount of liquid slag at T (°C) 80% 99%
M1 55 15 20 0 10 3.7 1275 1450
M2 47 13 20 10 10 3.6 < 1300 1750
M3 55 10 30 0 5 5.5 < 1300 < 1375
M4 50 9 30 6 5 5.6 < 1300 1475
M5 55 15 30 0 0 3.7 1475 1550
M6 47 13 30 10 0 3.6 < 1425 1625
M7 63 17 20 0 0 3.7 < 1550 1650
M8 55 15 20 10 0 3.7 < 1475 1800
M9 55 5 30 0 10 11.0 < 1350 1400
E10 52 5 25 8 10 10.4 < 1350 1725
3.2 PRE-HEATING SIMULATION OF LADLE-REFRACTORY
To be able to simulate the conditions the refractory is exposed to during a pre-heating process, small refractory pieces from Halmstadverken (roughly 10 x 10 x 3 cm) were heated in a muffle furnace in an air atmosphere. The furnace was heated at a rate of 1°C/minute up to 1000°C,
held at 1000°C for 4 hours and then slowly cooled < 1°C/minute to room temperature. No protective atmosphere (such as Argon or nitrogen) was used to ensure the simulation of the pre- heating process would be as realistic as possible. The weight difference of the refractory pieces before and after the heating was measured to determine the amount of mass loss during a pre- heating process. Two pieces of each of the three different types of refractory were used during this simulation:
• Lower wall refractory, RES3T05, containing 5 weight% carbon. (Further referenced as T05).
• Slag line refractory, REF6T12, containing 12 weight% carbon. (T12).
• Upper wall refractory, S3T10A, containing 10 weight% carbon and antioxidants. (T10).
The pieces were heated one at a time, cooled down, weighed separately and was later packaged in plastic bags for storage. As a final measurement, one piece of each refractory type was cut in half and the oxidation depth was measured with a ruler.
3.3 SLAG SMELTING TRIALS PREPARATION OF MGO-C CRUCIBLES
The slag-refractory interactions were determined to be investigated by melting slag in crucibles made by MgO-C material. Ten refractory cubes, 10 x 10 x 10 cm in size, were cut from refractory bricks gathered at Halmstadverken and a 5.5 x 7 cm cylindrical core was drilled out to create a usable crucible as seen in Figure 5. All crucibles were made of the slag line type refractory with 12% carbon (T12).
To ensure that the results from the slag smelting trials would be as realistic as possible, all refractory crucibles were pre-heated in a muffle furnace at a rate of 8-9°C/minute to 800°C, held at 800°C for 4 hours and then cooled to room temperature overnight. This was done to simulate a quicker pre-heating process, to oxidize the carbon surface layer and to be able to simulate process conditions to an extent. The weight loss of each crucible was measured to later be able to compare with the previous pre-heating simulations. A preheated crucible can be seen to the right in Figure 6.
Figure 6 – A pre-heated crucible next to a non-treated one. The inside is not as fully oxidized as the rest of the surface is.
SYNTHETIC SLAG PREPARATION
Ten synthetic slags with different compositions were determined to be melted in MgO-C crucibles to investigate the slag-refractory interactions. The slags contained different amounts of the five major slag components (CSAM+FeO), which was assumed to give a good overview of the different components’ interactions with the refractory. To be able to make sure that the smelting trials would be relatively representative of each slag and to ensure homogenous samples, all slags had to be pre-melted, ground and split to fit the crucibles. Due to the MgO-C crucibles containing 12 weight% carbon it was not suitable to prepare the slags in them, since the slag components would oxidize the carbon present in the MgO-C matrix. Therefore, the base-components of the slag (CSAM) was mixed in the correct ratios and pre-melted in graphite crucibles. FeO was later mixed in with the prepared base slags before they were ground and put in the MgO-C-crucibles. The base slags compositions, their compositional errors due to low quality chemicals as well as the added amount of FeO was compiled in Table 4.
Table 4 – Composition of the slags used for the smelting experiments, as well as the errors in the slag compositions within the parentheses.
Slag recipes and the compositional errors
CaO (wt %) SiO2 (wt %) Al2O3 (wt %) MgO (wt %) FeO (wt %)
M1 54.4 (-0.6%) 14.91 (-0.09%) 19.82 (-0.18%) 0.87 (+0.87%) 10 M2 47.08 (+0.08%) 12.87 (-0.13%) 20.03 (+0.03%) 10.02 (+0.02%) 10
M3 54.44 (-0.56%) 9.89 (-0.11%) 29.8 (-0.2%) 0.87 (+0.87%) 5
M4 50.52 (+0.52%) 8.1 (-0.9%) 30.32 (+0.32%) 6.05 (+0.05%) 5
M5 54.53 (-0.47%) 14.87 (-0.13%) 29.74 (-0.26%) 0.87 (+0.87%) 0 M6 47.77 (+0.77%) 11.54 (-1.46%) 30.52 (+0.52%) 10.17 (+0.17%) 0 M7 62.15 (-0.85%) 16.79 (-0.21%) 20.07 (+0.07%) 0.99 (+0.99%) 0 M8 54.16 (-0.84%) 14.79 (-0.21%) 19.79 (-0.21%) 11.25 (+1.25%) 0 M9 54.94 (-0.06%) 4.68 (-0.32%) 29.51 (-0.49%) 0.87 (+0.87%) 10 M10 51.49 (-0.51%) 5.09 (+0.09%) 24.73 (-0.27%) 8.68 (+0.68%) 10
The different chemicals used to mix the slags were presented in Table 5. The CaO, Al2O3 and MgO chemicals were sent for chemical analysis to Degerfors Laboratorium AB and then analyzed with x-ray diffraction, since their chemical compositions and mineralogy were unknown. The provided FeO chemical was not analyzed due to time limitations and was assumed to be 100 % pure FeO.
The chemicals that were analyzed by x-ray diffraction were performed using a PANalytical Epyrean XRD unit. The chemicals were all measured with the following measurement conditions: Cu Kα radiation, 40mA electron emission current, 45kV accelerating voltage and measuring in the 2θ range 10-90° with a 0.0260 step size.
Not enough CaO was provided for all slags, therefore CaCO3 with a known chemical composition was added to all slags to reach the desired Ca amounts. The slags were not analyzed after they were prepared, they were assumed to be homogenous and to maintain the same composition as prior to melting.
Table 5 – The chemical analysis for each chemical used to prepare the base slags. The chemicals are listed horizontally, and their respective analysis listed vertically.
Chemicals used to prepare base slags
CaO CaCO3 SiO2 Al2O3 MgO FeO CaO (wt %) 93 56.03 0 0.1 2.12 0
SiO2 (wt %) 2.20 0 99 0.13 5.33 0
Al2O3 (wt %) 1.30 0 0 99.2 0.1 0
MgO (wt %) 2.00 0 0 0 92 0
SLAG SMETING
When all the base slags had been pre-melted, the respective amount of FeO was added to each slag, and ground in a ring and puck mill to mix it as homogenous as possible. The complete slags were packed in the 10 pre heated MgO-C crucibles and melted two at a time in a muffle furnace at 1650°C for 2 hours. A temperature profile with a temperature increase of 5°C/minute and slow cooling at < 1°C/minute was chosen to try to ensure the refractory wasn’t damaged, e.g. cracks forming due to thermal expansion and contraction while in contact with the slags.
When the furnace had been cooled down, the crucibles were taken out and stored in steel bowls for later analysis. The crucibles were determined not suitable for microscopic analysis and was therefore only photographed for later visual comparison and evaluation.
SLAG-BRIQUETTE SMELTING EXPERIMENTS
In addition to the slag smelting in the crucibles, ~4 x 4 x 1.5 cm T12 refractory pieces from Halmstadverken were cut out and pre-heated in the same way as the refractory crucibles (see Preparation of MgO-C crucibles) to prepare a heating microscopy-type experiment. The 10 slags prepared for the slag smelting experiments in crucibles were pressed into 5 x 5 x 5 mm briquettes and placed on top of separate pre-heated refractory pieces, see Figure 7.
Figure 7 – Two different pressed slag briquettes placed on top of pre-heated MgO-C T12 refractory pieces.
The chamber furnace was programmed to run as the previous slag smelting experiments, i.e.
ramping up at 5°C/min to 1650°C, holding 1650°C for 2 hours and then slowly cool at a rate
< 1°C/min. After cooling down, the refractory pieces with the melted slag were molded in epoxy resin in pairs (M1 with M2; M3 with M4 etc.), cut in half and polished for analysis using Scanning Electron microscopy with Energy dispersive X-ray spectroscopy (SEM-EDS), see Figure 8.
Figure 8 – Sample M9 (at the top) and M10 (at the bottom) molded in epoxy resin, cut in half and polished for SEM-EDS analysis.
The SEM-EDS analysis was performed using a Gemini Zeiss Merlin microscope with an Oxford EDS detector, at an acceleration voltage of 20kV and a probe current of 1 nA. During the analysis, five pictures per sample with roughly equal distance between each picture was taken and the chemical compositions of the six major components: Ca, Si, Al, Mg, O and Fe were analyzed using the EDS instrument. Notes of visual observations were also taken and compiled, to be able to compare with the SEM-EDS results. The results from the EDS analysis was compiled in an excel-sheet and imported into SIMCA 15.0.2, to evaluate any possible correlations. The different chemical analyses from the EDS instrument were labeled as e.g.
A.1.5 to keep track of the data, where;
A – Denotes the level in the sample from A (at the bottom) to E (at the very top where the pressed briquette melted down).
1 – Denotes which sample the chemical analysis is from, from 1 (M1) to 10 (M10).
5 – Denotes which point (spectrum) at the sample the chemical analysis is from. See example in Figure 17.
For each spectrum point from the EDS equipment, two analysis results were calculated based on atomic weight. Since most slags contained iron, the calculations were balanced according to both Fe (II) and Fe (III), since the oxidation state of iron could vary within the system. All SEM pictures as well as their respective EDS analyses are found in Appendix B.
SIMCA, which is a Multivariate data analysis software tool used to analyze and evaluate possible process correlations, trends and patterns for different large data sets, was used to evaluate the results from the slag-briquette smelting experiments. In SIMCA the chemical analyses were set as primary observation IDs and the different chemical components (Ca, Si,
Model 1) All data points in one data set.
Model 2) The data points from slags M1, M3, M5, M7 and M9 (the slags without any MgO).
Model 3) The data points from slags M2, M4, M6, M8 and M10 (the slags that contained MgO).
A brief evaluation using SIMCA was made, where only the most apparent correlations were noted, by studying Score and Loading plots. It was not considered significant to evaluate any further than by simply using the basic tools in SIMCA, since only the most noticeable correlations was of interest.
4 R ESULTS
4.1 THERMODYNAMIC CALCULATIONS
The results acquired from the thermodynamic simulations for the 25 different slags were compiled in different graphs, see Figure 9 – Figure 13. In the graphs the slag amounts, given in mass percent, are plotted on the primary y axis; the liquid viscosity of each slag, given in Poise, is plotted on the secondary y axis; and a temperature interval from 1500 – 2000 °C on the x axis. The viscosities are only accurate for fully liquid systems, which these slags are not for the most part. It is therefore not an accurate representation of the viscosity of each slag, but it gives a good enough estimate that can be compared between slags with similar solids content.
The two graphs below each liquid slag/viscosity graph displays the CaO and MgO solids content in each slag respectively. The saturation limit lines present in the CaO and MgO solids graphs are limits given by Höganäs AB which represents desired amounts of CaO and MgO saturation in the slags.
In slags 1-15, see Figure 9, Figure 10 & Figure 11, it was observed that the slags with the lower CaO content, higher alumina content and lower basicities in general results in a more liquidous slag at lower temperatures. It could also be assumed that the slag viscosity would decrease as a result since less solid particles would be present in a more liquidous slags, even though the graphs oppose an increasing viscosity. As seen in Figure 9, the top graph shows that slag 3 consists of 80 mass% liquid slag at 1500°C, where the remaining 20 mass% are solid and undissolved slag components. The solids content gradually decreases until the liquid slag amount reaches 100% at 1950°C. Further, the bottom left graph in Figure 9 shows a CaO- saturation of of 14 mass% at 1500°C for slag 3, which decreases and falls under the required saturation concentration of 5 mass% at 1875°C. The bottom right graph shows that slag 3 is saturated with 7 mass% MgO at 1500°C which decreases and falls under the required saturation concentration of 2 mass% at 1800°C. Slag 3 could therefore be considered valid since it met both saturation concentrations at the given process conditions (~1650°C). About half of the designed slags (slags 4, 5, 9, 10, 13, 14 & 15) did not fulfill the desired saturation amounts.
Comparably, slags 1, 2, 3, 6, 7, 8, 11 and 12 all had high MgO- and CaO saturation, but as a result, would probably also have a significantly higher viscosity due to the increased solids content. Out of slags 16-25, slag 20 was the only one to fulfill the saturation limits of both CaO and MgO at the given ladle metallurgy process conditions.
The graphs also show that by increasing the Al2O3 content in the slag while simultaneously decreasing the CaO content, the slags became more liquid. CaO content as low as 50 weight%
would fulfill the saturation limit for slag 20, while at least 55% was required for the other slags mentioned above.
Figure 9 – Thermodynamic calculations of slags 1-5.
0 2 4 6 8 10
1500 1600 1700 1800 1900 2000
MgO-saturation (mass %)
Temperature (°C) Mg-based solids in slags 1-5 (mass%)
Slag 1 Slag 2
Slag 3 Slag 4
Slag 5 MgO-saturation limit
0 0.2 0.4 0.6 0.8 1
50 60 70 80 90 100
1500 1550 1600 1650 1700 1750 1800 1850 1900 1950 2000 Liquid viscosity (Poise)
Slag amount (mass %)
Temperature (°C)
Liquid slag amount and liquid viscosities of slags 1-5
Slag 1 Slag 2 Slag 3 Slag 4 Slag 5
Viscosity Slag 1 Viscosity Slag 2 Viscosity Slag 3 Viscosity Slag 4 Viscosity Slag 5
0 10 20 30 40 50
1500 1600 1700 1800 1900 2000
CaO-saturation (mass %)
Temperature (°C) Ca-based solids in slags 1-5 (mass %)
Slag 1 Slag 2
Slag 3 Slag 4
Slag 5 CaO-saturation limit
Figure 10 – Thermodynamic calculations for slags 6-10
Figure 11 – Thermodynamic calculations for slags 11-15.
0 1 2 3 4 5 6 7 8
1500 1600 1700 1800 1900 2000
MgO-saturation (mass%)
Title
Mg-based solids in slags 6-10 (mass %)
Slag 6 Slag 7
Slag 8 Slag 9
Slag 10 MgO-saturation limit
0 5 10 15 20 25 30 35
1500 1600 1700 1800 1900 2000
CaO-saturation (mass %)
Title
Ca-based solids in slags 6-10 (mass %)
Slag 6 Slag 7
Slag 8 Slag 9
Slag 10 CaO-saturation limit
0 0.2 0.4 0.6 0.8 1
50 60 70 80 90 100
1500 1550 1600 1650 1700 1750 1800 1850 1900 1950 2000 Liquid viscosity (Poise)
Slag amount (mass %)
Temperature (°C)
Liquid slag amount and liquid viscosities of slags 6-10
Slag 6 Slag 7 Slag 8 Slag 9 Slag 10
Viscosity Slag 6 Viscosity Slag 7 Viscosity Slag 8 Viscosity Slag 9 Viscosity Slag 10
0 1 2 3 4 5 6 7 8
1500 1600 1700 1800 1900 2000
MgO-saturation (mass %)
Temperature (°C) Mg-based solids in slags 11-15
Slag 11 Slag 12
Slag 13 Slag 14
Slag 15 MgO-saturation limit
0 5 10 15 20 25 30
1500 1600 1700 1800 1900 2000
CaO-saturation (mass %)
Temperature (°C) Ca-based solids in slags 11-15
Slag 11 Slag 12
Slag 13 Slag 14
Slag 15 CaO-saturation limit
0 0.2 0.4 0.6 0.8 1
50 60 70 80 90 100
1500 1550 1600 1650 1700 1750 1800 1850 1900 1950 2000 Liquid viscosity (Poise)
Slag amount (mass %)
Temperature (°C)
Liquid slag amounts and liquid viscosities of slags 11-15
Slag 11 Slag 12 Slag 13 Slag 14 Slag 15
Viscosity Slag 11 Viscosity Slag 12 Viscosity Slag 13 Viscosity Slag 14 Viscosity Slag 15
