The effect of magnesium addition on the formation and evolution of inclusions in Mn–Si killed steel is studied by laboratory experiments and thermodynamic calculation. Samples were taken at different holding time after magnesium addition to study the transient evolution of inclusions. The morphology, composition, number and diameter of inclusions were thoroughly analyzed by SEM-EDS. The results show that four types of inclusions (MgO–SiO₂–MnO liquid inclusion, MgO–SiO₂–MnO complex inclusion, MgO–SiO₂ and MgO solid inclusion) form with different amount of magnesium addition. Liquid inclusion of MnO–SiO₂ is promptly modified to MgO–SiO₂–MnO liquid inclusion with 0.0005% magnesium addition. As the amount of magnesium is increased to 0.0011%, intermediate product MgO forms immediately and then gradually transforms to MgO–SiO₂ solid inclusion with holding time. While MgO solid inclusion keeps stable throughout with 0.0019% magnesium addition.

Thermal conductivities of iron oxide scales have been determined in a temperature range of room temperature to 1176 K by the laser flash method, where the interfacial thermal resistance between FeO and iron has been corrected for each temperature. Two series of samples were prepared from iron plates with 99.99% purity. One was samples with FeO scale only and the other was samples with multi-layered scale of Fe₂O₃/Fe₃O₄/FeO. The laser flash method was applied to measure apparent thermal diffusivities, which were converted to apparent thermal conductivities. Apparent thermal conductivities obtained are contaminated by the interfacial thermal resistance between scale and iron, and are in linear proportion to scale thickness. Using apparent thermal conductivities at room temperature, thermal conductivities of scale have been derived from the slope of the linearity as 2.4 Wm⁻¹K⁻¹ for FeO scale and 1.8 Wm⁻¹K⁻¹ for multi-layered scale, and interfacial thermal resistances have been derived from the intercept as 7.3×10⁻⁶ m²KW⁻¹ for the interface at FeO/iron and 6.4 × 10⁻⁶ m²KW⁻¹ for the interface at multi-layered scale/iron. The interfacial thermal resistance between FeO and iron decreases with increasing temperature. Considering these temperature-dependent interfacial thermal resistances, thermal conductivity values of FeO scale have been determined as a function of temperature. The values are roughly 2.2 Wm⁻¹K⁻¹ up to 1176 K except 6.9 Wm⁻¹K⁻¹ at 674 K where the decomposition of FeO tends to take place.

The influence of temperature, slag basicity and Cr₂O₃ content on the viscosity of stainless steelmaking slags has been experimentally determined. The viscosity is mostly affected by the combination of basicity and Cr₂O₃ content. The results are discussed by considering the Iida model for the fully liquid slag and the Herschel-Bulkley equation for the solid-bearing slag. For the liquid slag, the predicted viscosity values by the original Iida model do not agree very well with the measured ones, because the model does not include the amphoteric behavior of Cr₂O₃ and underestimates the basic character of Al₂O₃. With the presently optimized parameters in the Iida model, the calculation is consistent with the measured viscosity values of the fully liquid slag with an error of 30%, which is fairly acceptable. At high basicity and high Cr₂O₃ contents, the presence of solid particles in slag has a paramount influence on the viscosity. It was also found that the solid-bearing slags exhibit evident Bingham pseudoplastic behavior with a residual torque of 0.12 mNm and 0.99 mNm in the investigated samples.

Carbon composite brick has attracted increased attention because it is widely used on the hot face of blast furnace hearth. In order to reduce the erosion rate of carbon composite brick and investigate effects of coke ash on the erosion of carbon composite brick, the minerals CaO·Al₂O₃, CaO·2Al₂O₃, and CaO·6Al₂O₃ (CA, CA2, and CA6), which were considered to be the main components of coke ash in hearth, were prepared by heating a mixture of CaCO₃ and Al₂O₃. Experiments of the minerals coupled with carbon composite brick were carried out at different temperatures in a high temperature tube furnace. The mass change fraction of the experimental assembly was calculated using mass measurements before and after the experiment. The reaction products at different temperatures were predicted on the basis of ternary phase diagrams, and the thickness of the reaction layer was measured using scanning electron microscopy (SEM). The results show that the reaction rate of carbon composite brick with the minerals increased with an increase of the Ca content in the minerals. The results of reaction interface analysis show that the diffusion of Ca²⁺ was the dominant process during reaction. Cation diffusion was considered to be the rate determining step, and this was in agreement with experimental results. Therefore, reducing the ash content of coke, especially the CaO in ash, is the key to reduce the erosion rate of carbon composite bricks below the centerline of taphole.

Molten oxides such as silicate melts are used in glass manufacturing processes and the chemical structure of the melts affects their physical properties and hence, the efficiency of the process in which they are used and the quality of the manufactured product. Analysis of the chemical structures using Raman or nuclear magnetic resonance spectroscopy is time consuming due to the process of preparing quenched samples and the long relaxation time of atomic nuclei. Hence, a technique for faster structural analysis is desirable. In this study, in order to accumulate basic data for in-situ estimation of the network structures of molten oxides, we systematically investigated the relationships between the alkali oxide composition and measured impedance behavior. Nyquist plots were fitted using an equivalent circuit consisting of solution resistance, charge transfer resistance, and double layer capacitance. In the present samples, the solution resistance and charge transfer resistance decreased, and double layer capacitance increased with increasing K⁺ concentration. These results were attributed to K⁺ behaving as a charge carrier or the double layer becoming thinner due to increasing concentration of K⁺ ions, which increased interfacial polarization. We observed that the solution and charge transfer resistances increased, and double layer capacitance decreased, in the order of Li, Na, and K. Hence, these resistances were dependent on the ionic radius, as well as the macrostructure of the melts.

Electrical conductivity is one of the most important thermophysical properties of oxide melts. In the present study, the electrical conductivities of CaO–SiO₂–MgO–Al₂O₃ type high aluminum blast furnace slags were measured by using four electrode method. It can be concluded that the electrical conductivity decreases as increasing the content of Al₂O₃, because Al³⁺ has a strong preference to form AlO₄⁵⁺ tetrahedron and incorporates into the network of SiO₄⁴⁺ which enhances the degree of polymerization of melts. As adding MgO to the slags, the electrical conductivity also decreases due to the stronger polarization ability and thus weaker diffusion ability of small Mg²⁺ ion relative to Ca²⁺ ion. However, the electrical conductivity increases with gradually increasing the CaO/SiO₂ ratio, owing to the increase of concentration of Ca²⁺ ion which acts as the charge carrier and the decreases of degree of polymerization.

The self-sintering of BOS filter cake has been investigated by studying its oxidation in air between 100 to 1000°C. The aim of the study was to gain an understanding of the self-sintering of the BOS filter cake in stockpiles, in terms of what reactions occurred, and how strength was developed in the filter cake during self-sintering.

Upon heating, the BOS filter cake underwent a sequence of drying, oxidation and calcination events. The self-sintering process was driven by the oxidation of very fine (200–300 nm) particles of metallic iron and wüstite at low temperatures. Reactions in self-sintering were found to be the oxidation of metallic iron and wüstite to hematite and zinc ferrite, beginning at approximately 120°C and largely complete by 500–600°C. Phase analysis, thermodynamic modelling and enthalpy measurements were used to propose a probable reaction path consistent with the observed reaction products. These exothermic oxidation reactions at low temperatures provided the energy required to heat the stockpiles to drive self-sintering.

Bonding within the reacted filter cake was from a network of particle-particle bonds formed between the very fine iron oxide particles in the matrix during oxidation at elevated temperatures. Exposure of the BOS filter cake to temperatures above 600°C under oxidising conditions is likely sufficient to form adequately strong material for transport and recycling in the BOS. Fluxes played little role in the development of the bonding within the filter cake.

To evaluate the effect of carbonization conditions on the bamboo, the relationship between carbonization parameter and physicochemical characteristics was studied. The results indicated that the volatile matter drastically decreased with the increase of carbonization temperature, while the fixed carbon and fuel ratio (fixed carbon/volatile matter) increased. Excellent linearity between the fuel ratio and carbonization temperature was obtained. The energy yield decreased gradually when rising the carbonization temperature, whereas the change of heating value was not obvious. A new calculation model of higher heating value (HHV) was developed, and it could be used to predict HHV of the bamboo char more precisely at temperatures above 300°C. The positive impact of functional groups, specific surface area as well as catalysis of alkali metal may contribute to the combustion of bamboo char. The results showed that there is a feasible operating condition for the transformation of bamboo into char with the carbonization parameter at 400°C for 30 min.

An Electropulsing-assisted Mold Flux Heat Transfer Simulator technique (EPMFHTS) has been developed in this article, through which the pulsed electric current was applied to the molten mold flux pool during the initial solidification process in the continuous casing mold. The aim of this study is to investigate the effect of pulsed electric current on the mold flux crystallization and heat transfer behavior in the casting mold. The results suggested that the general variation trend of the responding temperature and heat flux were nearly same, which were divided into four stages, namely, rapidly rising, sharp declining, gradually declining and relatively stable. The temperatures and heat fluxes showed a faster reduction rate in the declining stage and their steady state values are lower under the Electric Pulse treatment (EPT) condition due to the increase of the thermal resistance resulted from the enhancement of the crystallization of the mold flux. The interface thermal resistance at the mold/slag film interface and the total thermal resistance of the slag film increased due to the improved crystallization through the accelerated particles movement resulted from EPT. The SEM image showed a larger crystalline fraction, thicker crystalline layer and smaller grain size under EPT, indicating that the EPT did promote the molten mold flux crystallization and refine crystals size in the continuous casting mold.

It has been realized that a massive-like transformation, in which δ phase (ferrite) transformed to γ phase (austenite) in the solid state during and after solidification, was selected in Fe–C steels. X-ray radiography confirmed that the massive-like transformation also occurred in Fe-18 mass%Cr-Ni alloys with Ni contents of 8, 11, 14 and 20 mass%Ni. According to the equilibrium phase diagram, δ phase is the primary phase in 8 and 11 mass%Ni alloys while γ phase in 14 and 20 mass%Ni alloys. Solidification was always initiated by nucleation of δ phase and consequently fine γ grains were formed by the massive-like transformation in 8 and 11 mass%Ni. On the other hand, nucleation of δ phase as a metastable phase was preferably selected at lower undercoolings (<50 K) in 14 and 20 mass%Ni and consequently the massive-like transformation occurred even in 14 and 20 mass%Ni alloys. Solidification of γ phase can be triggered by nucleation of δ phase followed by the massive-like transformation in the Fe–Cr–Ni with lower Cr/Ni values (the primary γ alloys). Moreover, the present study demonstrates that the massive-like transformation will be commonly observed in Fe-based alloys, in which δ and γ phases are competitive each other from a thermodynamic perspective.

The present paper is the first part of a three paper series discussing raceway blockages in blast furnaces for ironmaking. During raceway blockages the hot blast flow rate can be reduced significantly and in such cases it is beneficial to shut down additional fuel systems like pulverized coal injection (PCI) on that specific tuyere. Currently this is mainly done by a simple threshold comparison of the hot blast flow rate data. Using a constant threshold level cannot account for drifting sensor levels or local changes in burden permeability around the raceway. Thus, the authors investigate various strategies to find more reliable solutions for blockage detection and the shutdown of PCI branches. Part 1 discusses the different nature of various blockage events on the basis of tuyere camera images and hot blast flow rate data. For the detailed testing of signal processing (part 2) and image processing algorithms (part 3) a universal test-bench is presented which also includes a common quality assessment of the results. Based on the collection of tuyere camera images a simple classification of blockage events is established. The analysis of 1500 hours of blast furnace operation data delivers a proper statistics about the frequency of occurrence of blockages on a small size blast furnace. The detailed analysis of hot blast signals and tuyere camera images presented in this paper series aims to improve the understanding of the raceway regions and their influence on optimal blast furnace operation.

The present paper is part 2 of a three paper series discussing raceway blockages and various approaches for automated detection based on blast furnace (BF) plant data. Blockages of the raceway areas occur on a regular basis and in cases where the hot blast flow rate is reduced significantly, it is beneficial to shut down additional fuel systems like pulverized coal injection (PCI) on that specific tuyere. While part 1 gave an overview on the different appearances of raceway blockages and presented a common test-bench used for signal and image processing, part 2 focuses on the discussion of various signal processing algorithms and their applicability for raceway blockage detection. The algorithms are tested on data from a real BF with supplementary tuyere camera images to validate the results. Beyond this test dataset, a long-term test is performed by processing three month of BF data offline.

The present paper is part 3 of a paper series discussing raceway blockages and various approaches for automated detection based on blast furnace (BF) plant data. While part 1¹⁾ gave an overview on the different appearances of raceway blockages and part 2²⁾ focussed on signal processing of hot blast data, this third part discusses various approaches for image processing of tuyere camera data. The visual impression of raceway blockages strongly varies between different events. This makes automated detection based on digital image processing of tuyere cameras a difficult task. On one hand the image processing algorithm should be robust and easy to tune for different tuyeres or different blast furnaces, on the other hand it should be fast enough, so that all tuyeres of a blast furnace can be processed on-line with a sufficiently high image frame rate. While algorithms optimized for motion detection fail due to the lack of a homogeneous background, adaptive thresholding of the grey-level histograms delivers useful results. Due to the nature of chaotic motion of coke particles inside the raceway also line based processing methods can extract the information from tuyere images in a sufficient manner and are very fast with regards to online implementation in a process control system. However, image processing of tuyere camera data has some disadvantages compared to the signal processing of hot blast data as discussed in part 2 of this paper.

The curling of a sheet in cold asymmetric rolling was investigated experimentally by interrupted rolling. Influences of speed difference as well as entry table height were studied. The upper and lower profiles of the aluminum sheet obtained by interrupted rolling were measured by a laser profilometer. The contact lengths on both upper and lower surfaces L_U and L_L were detected from the profiles. The curvature radius of sheet was calculated from the curling profiles. Under differential-speed rolling, the sheet curled to the slower roll side at low reduction in thickness, while it curled to the faster roll side at higher reduction. Meanwhile, L_U on the faster (upper) roll surface was longer than L_L on the slower (lower) roll surface at low reduction. On the other hand, L_U was shorter than L_L at higher reduction in thickness. It is found that the ratio of contact length L_L/L_U is correlated with the curling direction in the differential-speed rolling. If L_L/L_U < 1, the sheet curls towards the slower roll side, while the sheet curls towards the faster roll side in case of L_L/L_U > 1.

In the hot-rolling process, defects resulting from running instability when a thin strip runs on the Run-Out-Table (ROT) at high speed in the state of non-tension (head or tail of the strip) pose a chronic problem. As a measure to prevent unstable strip running, the maximum line speed has been limited. Consequently, this has also become an aggravating factor that impedes higher productivity. Although various methods to solve and improve this problem have been proposed from an early date, it has not been eradicated completely.

In this paper, we propose a numerical simulation method to simulate the strip running phenomenon on the ROT. It was found that the numerical results correspond to the theoretical and experimental results. Use of the proposed method has reduced both the cost and time necessary to improve the unstable strip running phenomenon.

The phenomenon of uneven contact pressures between cylindrical backup roll and CVC (Continuously Variable Crown) work roll in four-high hot strip mill has always existed, which can lead to seriously uneven wear of cylindrical backup roll and even spalling defect. In order to obtain an available solution plan for above mentioned problem, a new kind of backup roll contour called reverse CVC backup roll contour is proposed in this paper. Reverse CVC backup roll contour is generated by fitting four cubic chamfer contours and one backup roll surface contour designed by CVC work roll contour. And one roll elastic deformation simulating software coded based on Influence Function Method has been used to calculate and study the contact pressure distributions between CVC work roll and reverse CVC contour backup roll or cylindrical contour backup roll. Compared with cylindrical contour backup roll, contact pressures between CVC work roll and reverse CVC contour backup roll are much smoother without steep peak. Simulating results meet well with data collected from the application field, wear of new reverse CVC contour backup roll is more well-distributed than cylindrical contour backup roll.

The effects of solution composition and additives on the formation of electrodeposited invar Fe–Ni alloys with low thermal expansion were investigated. In all solutions, increasing the current density from 10 A·m⁻² significantly decreased the Ni content in the deposits and resulted in anomalous codeposition, in which the electrochemically less noble Fe was preferentially deposited. Further increasing the current density increased the Ni content in the deposits as Fe deposition reached the diffusion limitation of the Fe²⁺ ions. Increasing the concentration of Fe²⁺ ions increased the current density required to initiate an increase in the Ni content in the deposits because of increase in the diffusion-limited current density of Fe. With the increasing concentration of malonic acid, the current density region in which Ni deposition was suppressed was extended, and the potential at which Fe deposition reached the diffusion limitation of the Fe²⁺ ions was shifted to a less noble direction. As a result, the relationship between the Ni content in the deposits and the current density shifted to a higher-current-density region with increasing malonic acid concentration. At malonic acid concentrations above 0.05 mol·dm⁻³, the current efficiency for alloy deposition was greatly reduced owing to the promotion of hydrogen evolution. The Ni content in the deposits was significantly increased with the addition of thiourea. Boric acid somewhat increased the Ni content in the deposits in the lower-current-density region. In solutions containing both boric acid and saccharin, an invar alloy of 36 mass% Ni was obtained in the wider-current-density region.

A novel quantitative characterization method of MnS distribution was proposed based on the data from automatic inclusions analysis software to evaluate their effects on intragranular polygonal ferrite (IPF) formation in medium carbon non-quenched and tempered steels. The results show that agminated MnS inclusions in the steel are less effective to promote the IPF formation and lots of inclusions are actually wasteful because their space distances are too small to act as nucleation sites individually, even their number is larger than that in the steel with plentiful uniformly distributed MnS inclusions. Due to the difference of observation method, three-dimensional dendritic MnS inclusions would be revealed as several agminated small-sized ones in two-dimensional observation. In order to avoid misunderstanding, the characteristics of the nearest space distances between inclusions in widely used two-dimensional observation could be applied to recognize and define these agminated MnS inclusions to more accurately evaluate their effects on IPF formation.

Tensile tests and deep drawing under different drawing ratios were carried out on 2507 duplex stainless steel (DSS). The effects of the microstructure and microtexture on the formability of the as-received sheets were examined by optical microscope and electron backscatter diffraction. Deformation microstructures in different regions of the drawn cup were also analyzed with transmission electron microscopy. The results indicated that 2507 DSS sheet had serious anisotropy and obvious 45° ears. Textures of the as-received sheet showed that the strong α-fiber texture in the ferrite (α) was dominant, while the texture intensity in the austenite (γ) was weak. The {001}<110> and {112}<110> orientations were the main factor that resulted in the low Lankford parameter (r-value) and high planar anisotropy (Δr). The microstructures in different regions of the drawn cup showed that the different crystal structure of α and γ phases led to the inhomogeneous strain distribution. The decrease of the deep drawability was attributed to the incoordinate deformation between the α and γ phase.

Extraordinarily strong recrystallization texture components near rotated cube ({100}<011>) developed in a low carbon steel processed with a severe cold rolling at 99.8% reduction followed by an annealing at 550°C or more. The rotated cube orientation is beneficial for magnetic applications, while it has been recognized as an orientation difficult to recrystallize. Unlike the conventional 90% cold rolling sample, the 99.8% cold rolling sample showed an unusually strong α fiber deformation texture and a fine lamellar structure understood from previous work to be generated via grain subdivision. In addition, texture analysis of a partially recrystallized sample at lower temperature anneal of 450°C revealed that the {100}<0 7 10> oriented grains recrystallized discontinuously from near the domain boundaries in the 99.8% cold rolled sample, while the {100}<011> oriented grains were likely to recrystallize continuously. Kernel average misorientation analysis of EBSD data suggested that the {100}<011> oriented grains underwent a faster recovery than the other α fiber orientations, resulting in a fast nucleation and growing toward the deformed matrix. Since this oriented nucleation of {100}<011> oriented grains was not observed in 99.8% cold rolled pure iron and 99.2% cold rolled interstitial free steel, solute carbons in the steel would play a significant role for the selective and fast recovery of the {100}<011> oriented grains during the annealing process.

Induction heating (IH) is commonly used for heating and heat treatment. An accurate prediction of temperature distribution is required to optimize the heating parameters, such as the heating coil dimensions, coil current, and heating time, without trial and error. In this study, temperature-dependent B-H curves and changes in phase transformation under rapid heating were measured and used for IH analysis to improve the prediction accuracy. IH experiments were also carried out to confirm the accuracy of the analysis. Consideration of the temperature-dependent B-H curves improved the prediction accuracy up to the Curie temperature, and consideration of changes in phase transformation temperature under rapid heating improved the accuracy for temperatures above the Ac₁ temperature. The maximum error between the experimental and calculation results for a columnar test piece was approximately 25 K, when the experimental temperature was 1300 K. This error is 9 times smaller than that obtained with a conventional analysis method. Good agreement was also obtained between the experimental and calculated values for a test piece with a convex projection, for which the temperature distribution and evolution are complex.

The deformation behaviors of metastable retained austenite and martensite in a high carbon quenched-tempered steel were investigated complementarily using electron backscattered diffraction, nanoindentation and transmission electron microscopy (TEM) to determine the effect of the austenite grain size on its mechanical stability. Investigation by the nanoindentation have shown that martensite exhibits a single plastic deformation stage, i.e., a dislocation glide motion, whereas metastable retained austenite exhibits double plastic deformation stages. TEM observations of a cross-sliced foil with an indentation mark has suggested that the first stage deformation in the austenite phase is dominated by stress-induced martensite transformation, while the second stage is governed by the slip deformation of dislocations in the transformed martensite. Furthermore, we have found that the stress required for the phase transformation and for the transition from the phase transformation to the plastic deformation of transformed martensite increase with decreasing in austenite grain size, indicating that fine austenite grains are mechanically more stable than coarse austenite grains.

X-ray diffraction is a powerful tool for dislocation characterization, which includes evaluation of dislocation distribution, nature of dislocation, and dislocation density. In the Williamson–Hall (WH) plots, the full width at half maximum (FWHM) is plotted relative to the diffraction angle for each diffraction peak and the method corresponds to the basic approach for dislocation characterization. However, the elastic anisotropy in each crystal plane makes the analysis of WH plots difficult because elastic anisotropy also affects the FWHM of diffraction peaks. In order to correct the effect of elastic anisotropy, Ungár developed a unique methodology by using the contrast factor C, and this is termed as the modified Williamson–Hall (mWH) method. Conversely, researchers developed a new methodology termed as the “direct-fitting (DF) method,” in which the elastic anisotropy is corrected by directly applying the diffraction Young’s modulus ratio (ω). In the DF method, a linear relation is realized in the corrected WH plots, and reliable values are obtained for the parameter α that contains information on the crystallite size. In the present study, the α-value obtained using the DF method was applied to the mWH equation, and dislocation characterization was performed in a low carbon ferritic steel (Fe-0.0056%C) by cold rolling. The results indicated that increasing the extent of cold rolling decreases the screw component of dislocation, and monotonically increases the parameter φ (which contains the information of dislocation density). Additionally, the parameter A (which depends on the dislocation arrangement) was evaluated at approximately 0.50 for cold worked ferrite.

Growth ledges are commonly observed on interphase boundaries during diffusional phase transformations and are of great importance for understanding inter-sheet spacing of interphase precipitates. A simple model based on Gibbs Energy Balance (GEB) for describing growth kinetics via diffusional growth-ledges of height λ is presented for the case of ferrite growth into austenite. The model is validated against the case of austenite to ferrite transformation involving interphase precipitation in a V, Mn, Si alloyed HSLA steel where, λ is assumed to be equal to the inter-sheet spacing of interphase carbide precipitates. The presented model provides a computationally efficient and versatile method for predicting the ledge height, λ, and the growth kinetics of ferrite from initial nucleation through to final soft impingement considering the evolution of solute drag at growth ledge risers. It is suggested that the intrinsic mobility of growth ledge risers is: , with R the gas constant and T the absolute temperature in K.

Immobilization of chromium in a stable spinel by modification is a powerful way to prevent chromium pollution of stainless steel (SS) slags. The precipitated spinel grain size is usually smaller than 30 µm, however, which limits the effectiveness of the modification. In the literature, very few efforts have been reported on promotion of spinel growth rate by optimizing the dynamic conditions. In this study, the effects of shear force on the spinel grain size evolution under isothermal conditions and during cooling were investigated. The experimental results indicate that the employment of shear force significantly changes the growth behavior of spinel at 1500°C. The growth mechanism of spinel was studied by using crystal size distribution (CSD) theory, showing different regimes of supply-controlled Ostwald ripening, surface-controlled growth with decaying nucleation rate, and constant rate nucleation and growth at shear rates of 0 s⁻¹, 10.83 s⁻¹, and 21.67 s⁻¹, respectively. On the other hand, at a cooling rate of 5°C·min⁻¹, the shear force was found to have little effect on the crystallization behavior of spinel. The results of static leaching tests indicated that hardly any chromium had leached, which makes the modified SS slag more environmentally friendly when used as a raw material.

The emissions of a blast furnace and the separation efficiency of the top-gas scrubber were investigated for the gaseous components HCl, H₂S, SO₂, COS, CS₂, CH₃SH, HCN and NH₃. The clean gas emission of sulphur was dominated by COS which accounted for about 85% of the total sulphur emissions. From the hydrogen halides only HCl was present above the detection limit and the concentrations of NH₃ and HCN were also low. The average separation efficiency for HCl, SO₂ and NH₃ was 96%, 81% and 46%, respectively. COS was not separated in the scrubber system. The concentrations of the weak acids HCN and H₂S were higher in the off-gas after the scrubber compared to the concentration before the scrubber. This phenomenon was explained by the release from the scrubber water into the gas of un-dissociated H₂S and HCN originating from cyanides and sulphides contained in the separated dust. Separation efficiencies similar to those of the scrubber system are expected to be achieved also by dry sorption gas cleaning processes. However, this has to be verified in experiments. In the case of applying a dry gas cleaning system, the concentrations of H₂S and HCN in the clean gas would be even less.