Basics of slag production
Written by Written by Lewis Juckes
Thursday 27 October 2011
Slags from the Iron and Steel Industry
With world steel production now well over a billion tonnes per year, the slag that arises from some of the processes involved is a major resource. Traditionally it has been used mainly as an aggregate but for some types there are other applications, such as a raw material for cement or as a fertiliser.
Slag, as the term will be used here, is any siliceous melt that arises in significant quantity from the various processes used in the production of iron and steel, and more particularly the solid materials that forms when such melts cool. Slag from the production of ferrochrome is also included here; this material is produced in substantial tonnages and the main use of ferrochrome is in the steel industry.
Slags also arise from other processes, particularly the smelting of non-ferrous metals, but these materials can be very different and each needs to be studied individually. Moreover, in colloquial English an even wider range of materials such as clinker, ash and even colliery waste is sometimes referred to as “slag”. Such materials are not covered here.
In the earliest processes the slag consisted mainly of the impurities (known as the gangue) that were present in the iron ore; in more modern times there is also a large contribution from materials added to aid the underlying process or to remove some specific deleterious element. An early example of this was the lime added to the feed for blast furnaces to give a more fluid slag, to reduce loss of iron as oxide in the slag, and to remove much of the sulphur introduced by the coke.
In tonnage terms, blast furnace slag is at present the main type followed by slags from the Basic Oxygen Steelmaking (BOS) and Electric Arc Furnace (EAF) processes. Various other processes such as external desulphurisation of steel may generate significant amounts of slag at some plants but overall they are only a small proportion of the total. In a few parts of the world, older or even ancient slags may be so abundant as to constitute an important industrial resource.
Disposing of this material by dumping would not only be expensive but would also represent a waste of valuable natural resources. Thus much of this slag is utilised, and to achieve this it is normally necessary for certain properties to be kept with specified limits. For some outlets the chemical properties may be most relevant, while for others it can be the physical properties and also stability, such as susceptibility to breakdown or weathering.
Recycling slags by feeding them back into the process has the attraction of efficiency but there are usually strict limits to the amount that can be recirculated in this way. Normally one of the functions of the slag is to remove some unwanted element, in which case adding it back to the stream may simply increase the removal task at some later stage (the re-introduction of phosphorus into steel is a particular danger of this type). This can sometimes be countered by recycling to some other branch of the process stream or by separating out certain fractions and recycling those. An example of the latter is the magnetic extraction of metal globules from BOS slag and their addition to the blast furnace feed, although the amount of phosphorus in the adhering silicate slag has to be monitored closely.
Blast Furnace Slag
In modern ironmaking, using rich ores, the ore may contain several percent of silica and this becomes a major component in the slag. Alumina is usually present, though the levels can vary quite widely between sources. In most commercially traded ores the other elements tend to be present only at much lower levels although there are some sources which can contain significant amounts of other components. These are sometimes used for specific metallurgical purposes (for instance, to raise the manganese or titanium content of the burden) or for reasons such as local availability and cost.
The ash from the coke is another contributor to the slag, and while this is predominantly siliceous there are differences which can be linked to geography – although there are many exceptions, southern hemisphere coals (including India) tend to have lower alumina contents than those from the northern hemisphere.
Other components may be included as additives of their own, to improve some part of the metallurgical process. Lime is the most important of these, and there are several points of the process before the blast furnace where it may be added; another is titanium, sometimes used to counteract wear in the furnace lining.
Air-cooled Blast Furnace SlagThe simplest treatment of the liquid slag is to lead it into a slag pit where layers build up until the pit is full. When it has fully solidified and then cooled somewhat, the slag (commonly known as air-cooled blast furnace slag or ACBFS) is excavated and transported away for further treatment (such as magnetic removal of any iron, or crushing and screening). Slag pits beside the furnace are most convenient but where space is restricted the liquid slag can be run into slag ladles and then conveyed to pits elsewhere. While one pit is being excavated, another is being filled, and water sprays can accelerate cooling so that excavation can begin promptly.
To maintain slag quality, good housekeeping is essential at all stages. General rubbish must be kept from entering the slag pits, and the bunds or walls enclosing the pits, which may be simply a ridge of aggregate, should be made of the same material in case some is recovered with the cooling slag during excavation. As the slag is removed and processed, the different slags (and other materials) must be kept in well-separated streams.
Such air-cooled slag is usually wholly or largely crystalline. Rapid cooling gives a glassy product, and several techniques are available whereby this can be achieved with water-cooling to give a product that is in most cases fine particles or granules. In the simplest versions the molten slag stream is poured into a powerful jet of water to give granulated slag but there are other variants such as pelletising, where the slag is poured on to a spinning cylinder and flung as droplets that cool in the air. Water-cooling was also important in the foaming process, where the liquid slag was poured over a group of upward-facing sprays to give a lightweight, pumice-like product. Environmental safeguards are needed with such processes, partly for the dust or slag fibres that may be generated but also because the sulphur in the slag can react to release compounds into the air and water.
Air-cooled blast furnace slag is a rock-like material, but often more porous than most natural rocks. This can be an advantage in some applications, such as aggregate in sewage treatment beds.
Depending on the application, there may be desirable or statutory constraints on the chemical composition. Ground granulated blast furnace slag (GGBS) can be blended with Portland cement to give a concrete that may have technical or price advantages, but in some countries there are limits on the alumina content (which may in turn restrict the use of iron ores from parts of the world where the gangue is more aluminous).
If ACBFS is used for aggregate it will be the physical properties that are most important, and the European Standards can measure various aspects of the strength. In the earlier years of the twentieth century there was concern over the presence of the β form of dicalcium silicate, a mineral rare in natural rocks but which can be present to a level of several percent in ACBFS from some sources. Under certain conditions this can invert to the γ form while cooling, with an expansion that shatters the surrounding material. This process, known as “falling”, was common at some plants in the early years of the twentieth century and was found to make the slag unusable as an aggregate.
Above: A video of a 'falling' stainless steel slag
In the second half of that century a belief arose in Britain, but nowhere else, that this inversion might occur spontaneously at some substantially later time and that it might, for instance, disrupt concrete in which the slag had been used. Strict chemical limits were developed and incorporated into British Standards to guard against this postulated “late falling”.
When British Standards were being replaced by European Standards in the late 1990s, a review showed that the idea of late falling had been based on a misinterpretation of earlier reports and that there were no known descriptions of the phenomenon ever being observed, let alone causing any disruption to an aggregate. Thus the European Standards only specified a test to confirm whether any falling had occurred during cooling. As it is, the slag compositions used in modern ironmaking mean that falling slag in the slag pits is now a rare occurrence.
In the latter half of the twentieth century, steelmaking came to be dominated by two processes. These are BOS, where oxygen is blown into the molten metal, and Electric Arc, where energy is provided by large electrodes. Earlier processes included the Open Hearth, where heated air and fuel were blown across the surface of molten steel held in a vessel, and Bessemer, where air was blown through the molten steel via holes in the base of the steelmaking vessel. These processes were superseded in economies with modern industrial methods although examples may survive in areas with special considerations.
Fluxing materials may be added to accelerate the slag formation and the assimilation of additives such as lime. Among the common historical fluxes are calcium fluoride and certain borates, although a wide range of other materials has been tried at some locations such as red mud from the processing of alumina. The possible presence of such elements must be borne in mind during the environmental assessment of any such slag.
Slags from these processes are usually strong and rock-like, with the potential for use as aggregates in a variety of applications. The density is higher than that of blast furnace slag or many commonly used natural rocks, and this can influence marketing; for some applications, such as use in waterways, the density is an asset. The high total lime content, often over 40%, and the fact that much of this lime content is held by minerals that are reactive under soil conditions, allows some slags to be used as agricultural liming agents.
When molten iron (referred to as “hot metal”) from the blast furnace is to be converted into steel, the main tasks are the removal of carbon, silicon and phosphorus. The LD (Linz-Donawitz) process is by far the most widely used, and takes its name from the towns where it was developed. Other terms such as BOS (Basic Oxygen Steelmaking) or BOF (Basic Oxygen Furnace), which have a broader meaning, are also commonly used for this process. The word “Basic” is used to indicate the type of refractory lining needed, normally magnesia, to resist the lime-rich slag.
This is a batch process (with each batch known as a “heat”) where oxygen is blown from above into the steel, and the different elements present oxidise in a predictable order. The carbon escapes as gaseous carbon monoxide and the silicon becomes silica which enters the slag. Phosphorus is normally only present in very small amounts but its profound effect on steel properties means that it receives close scrutiny. (In modern practice there could be around 0.1-0.2% P in the hot metal, which will be reduced to about a tenth of that value, giving a content of about 1% P2O5 in the slag. The exact values will of course vary considerably depending on raw materials, plant policies, and the target steel composition at any given time.)
The key to phosphorus removal is a lime-rich slag, achieved by addition of a calculated weight of lime to each heat in the furnace. This is an area of technical trade-offs: coarse lime takes longer to dissolve into the slag and begin the removal of phosphorus but with finer lime there would be excessive dust losses. Slag with a higher lime/silica ratio is not sufficiently fluid but a lower ratio gives reduced phosphorus removal. A flux, such a fluorspar (calcium fluoride) can be added to accelerate slag formation but increases the cost of the process.
The final slag can have a lime/silica ratio of about 3 or 4, although this can vary substantially from one heat to the next. Depending on the grade of steel being produced, the details of the process for each heat may be governed largely by the target level of phosphorus. For the purposes of slag utilisation, the most important point is that several percent of free lime can remain unreacted. This can later react expansively with water and is capable of disrupting an aggregate or the structure in which it has been used.
Much research has gone into ways to counteract or modify these expansive reactions. Although there is theoretically scope for changes in various aspects of the steelmaking process, it is usually accepted that the optimisation of this process has the highest priority and so efforts to modify the slag properties begin at the next stage, when the slag is poured from the steelmaking vessel. Introducing siliceous additives could accelerate the reaction of the melt with any remaining free lime but it is necessary to ensure that such particles are not immediately surrounded by a chilled coat of solid slag as that would prevent any further reaction. Processes have been tested using injection of sand, and addition of cullet (scrap glass).
Weathering of the slag before utilisation can allow some hydration of the free lime to take place so that the potential for expansion is reduced before the slag is used as an aggregate. To assess the success of such weathering, a suitable test is needed. Many expansion tests have been devised although their suitability as indicators of behaviour under, say, conditions within a road is rarely tested. The European Standard BS EN 1744 specifies a test where the results have been correlated with a series of road trials using BOS slag. Chemical analysis for free lime can also be a useful indicator of factors such as the progress of the weathering process. The weathering can be done under atmospheric conditions, where it can typically take about a year although details may depend on local conditions. (In arid areas, for instance, little weathering will occur unless the slag piles are regularly watered.) In addition there are processes, some already in commercial operation, for accelerated weathering by use of hot water (from plant cooling circuits) or by use of steam. These can reportedly reduce the time required per batch to days or even hours.
Particles of free magnesia can also be present in steelmaking slags, usually as fragments broken from the magnesia refractories. These can hydrate expansively in the same way as free lime, although the reactions are normally slower and it is less common to find contents high enough to cause concern. In any case, any expansion from this reaction would normally be included in that measured by the test specified in BS EN 1744.
The BOS process needs to be fed with molten iron, whereas the Electric Arc Furnace (EAF) can operate with a feed consisting partly or entirely of solid metal such as scrap. This gives rise to two significant differences in the slag, which is otherwise fairly similar in chemical composition. First, because much of the phosphorus has already been removed the slag does not need to be so basic. Thus any free lime present is normally at a lower level, and less likely to cause any significant expansion of the aggregate. Secondly, depending on the nature of the scrap used, the feed may contain significant proportions of other components (such as copper from motor vehicle scrap), and some of these may enter the slag.
In countries such as Britain, in earlier years small steel plants existed in many areas that are not now thought of as associated with that industry. Over the years large tonnages of slag may have accumulated, and these may now be encountered during operations such as the development of brownfield sites. In that case, it becomes necessary to know their nature and properties in order to establish whether they are a valuable resource or a liability. In particular, their purity needs close examination. If the slags were regarded as materials to be dumped, at the time of production, then it is quite possible that other ironworks materials may also have been dumped with them.
While each case needs individual study, some guidance is found in the nature of the processes of those times. Blast furnace slags, for instance, may have higher sulphur contents than their modern equivalents, and the longer time means that alteration reactions (such as with ground water) that are normally insignificant might be important.
Open Hearth slags were used as road aggregate in some countries when that was a major process in the industry. Although failures were not common, those that did occur were often widely reported in publications. This not only caused a loss of confidence in the material but such cases have at times been confused with BOS slag. In fact, the descriptions of the failures have usually revealed that the slag was used with no form of quality control, and in some cases the users were not even aware that of what type of slag was being used. It was such events that led to the more thorough attention that is now given to the handling, treatment and properties of modern slags.
The nature of ancient slags, in those areas where they occur, will depend on local materials and ironmaking traditions. However, slags from the earliest processes often have a high FeO content, giving a dense rocky material rich in the minerals wustite and fayalite.
Chromite is the only ore of chromium in commercial use, and as deposits of this mineral are sparsely scattered across the globe there are far fewer production sites for ferrochrome than for iron and steel. Production is usually in an electric arc furnace, and the slag arises from the other components present within the chromite. These components are mainly oxides of Fe, Mg and Al, in proportions that vary quite widely between different ore deposits. Such combinations alone will not give a good fluid slag and so silica is usually added to promote melt formation.
This gives a slag of different composition from either blast furnace or steelmaking slag, and which may be glassy or crystalline depending on the cooling routine. The result can be a strong and stable material, suitable as an aggregate for roads and similar applications. Typically the density is relatively low, closer to that of blast furnace slag than steelmaking slag, which can influence its suitability for certain markets.
About the author
Lewis Juckes graduated from the University of Natal in 1963 and worked as a geologist during the 1960s (including two years in the Antarctic). After obtaining his PhD from Birmingham University in 1969 he joined British Steel and remained with them and their successors until his retirement in 2000. His work there covered a wide range of non-metallic materials such as ores, slags and dusts. When the British Standards were being replaced by European Standards in the 1990s, he was involved in the drafting of those affecting the utilisation of slags. Since his retirement he has retained an interest in such matters, publishing two papers concerning slag utilisation.