Considerations For The Use Of Alternative Iron Materials In The EAF– Current Understanding

Jeremy A.T. Jones
Nupro Corporation


The production of steel via the EAF route continues to grow both in North America and worldwide. The past 5 years has seen additional growth both in the supply and use of alternative iron materials for use in the EAF. Though initially, the demand for alternative iron materials was based on a need for a lower residual product from melting operations, many EAF operations now use alternative iron materials for other reasons as well. As demand for AI materials has grown, more materials have become available on the merchant market and the number of choices facing EAF operators has grown. In fact, in another 5 years, it is expected that several new processes will be in commercial operation, further increasing the number of options available to steelmaking operations. This paper extends the work presented by the author 2 years ago and is meant to offer the EAF operator a means of deciphering the staggering number of options with which they are now faced. It will be demonstrated that the choice of feed materials is highly dependent on local factors and on the operating practices employed at each unique facility. The choice of raw materials is clearly based on a keen understanding of objectives and goals and cannot be applied universally to other facilities.

Do We Need Cleaner Scrap?

The major emphasis in EAF steelmaking has been related to achieving maximum energy efficiency in the EAF. In addition, material feedstocks are also influencing the design of electric arc furnaces and the way in which they are operated. The use of alternate iron sources is the biggest driving force in this area. This includes direct reduced iron (DRI), hot briquetted iron (HBI), iron carbide, and pig iron (either solid or as hot metal). The demand for these “clean steel units” is brought about by the desire to lower the levels of tramp elements (residuals) in the steel. Levels of these residual elements (copper, tin, nickel, chrome and molybdenum) are high in obsolete scrap and can affect casting operations and product quality if they are not diluted. One option is to use only high quality scrap in the EAF but this has major cost implications and with the additional flat rolled EAF capacity coming on line, it is expected that there will be a shortage of high quality scrap. Alternatively by adding clean iron units as part of the scrap mix along with obsolete scrap, the levels of these residuals in the liquid steel can be reduced to acceptable levels.

Many advantages have been attributed to the use of alternative iron materials in the EAF including:

    ∙ Dilute residuals – AI materials contain low levels of residuals
    ∙ Consistent chemistry – AI materials are consistent in their chemistry. Thus utilization of AI materials will tend to offset the wide chemistry fluctuations common in obsolete scrap.
    ∙ Consistent C recovery – recovery of carbon from AI materials typically exceeds 90 %. Charge carbon recovery can vary widely (30 – 80 %) and some charge carbon is also high in ash content (up to 14 %) and sulfur content.
    ∙ Better slag foaming – slag chemistry is more consistent when properly using AI materials. The period of CO evolution is typically extended as compared to conventional EAF operations. These factors lead to more optimum slag foaming conditions.
    ∙ Greater slag volume – greater slag volume can lead to higher Fe losses but may be beneficial for slag foaming especially for DC operations and AC operations utilizing long arc practices in the EAF.
    ∙ Consistent EAF operations
    ∙ Lower N content in steel – AI materials are low in nitrogen content and thus result in low bath melt-in nitrogen content. If the AI material results in elevated bath carbon levels, CO formed in the bath during decarburization will strip nitrogen from the bath.

Alternative iron sources can take on many different forms. Even within a single product type such as DRI, variations in chemistry and metallic content can greatly affect the way in which the material can be used in the EAF. One of the biggest concerns facing EAF operations is to meet product quality requirements at minimum cost. Without a clear understanding of the effect of alternative iron characteristics on EAF operations, it will be difficult for the EAF facility to determine the least cost mix of scrap to melt in the furnace.

Several parameters are used to specify the quality and characteristics of DRI. As other technologies have been developed, these parameters have also been used to grade these products. Thus the following parameters can be applied to any alternative iron product.

Key parameters that have been identified in the past as measures of the quality/value of AI materials include:

    ∙ Metallization
    ∙ Carbon content
    ∙ Gangue content

In addition, several other parameters have been identified in recent years as having an effect on value-in-use of AI materials. These include:

    ∙ Yield
    ∙ Iron, silicon and aluminum content
    ∙ Sizing
    ∙ Density
    ∙ Propensity to generate fines

In fact, it can be shown that all of these parameters can in fact be used to classify any type of metallic feed for the steelmaking process.

Key Quality Parameters For Eaf Iron Bearing Feed Materials


Percent metallization refers to the % of iron in the DRI, which is present as metallic Fe:

% Met = [(% Fe metallic)/(% Fe Total)] * 100

Typically steelmakers prefer a metallization of about 92 % or higher. Some operations now classify the alternative iron source by its’ “useable iron content”. Useable iron content refers to the amount of metallic iron present in the feedstock (% Fe metallic) and is an indication of the amount of the feed, which will directly contribute to the steel bath. Any iron that is not in this form will be present as iron oxide and must be reduced in the bath to recover the iron units. This will require both energy input and a source of reductant, usually carbon.

Other forms of AI such as pig iron have very high degrees of metallization, on the order of 98 – 100 %. Ironically, this measure of quality has seldom, if ever, been applied to conventional steel scrap types.

Obsolete scrap tends to contain more rust. Shredded scrap tends to oxidize easily if exposed to rain. Borings and turnings are especially prone to oxidation and in some cases may even ignite. When overly oxidized scrap is added to the EAF, the result is that the FeO levels in the slag may rise quite rapidly resulting in the inability to foam the slag. It is important to be aware of the fact that rust is actually a combination of iron oxide and water. If we have scrap, which contains 2% rust, the net effect (when energy efficiency is taken into account) can be an increase in energy requirements by 43 kWh per ton of product. At least one study has shown that a large portion of the rust actually reports to the offgas system and ends up in the EAF dust, those iron units being non-recoverable.

Analysis of the wide variation in yield of various scrap types indicates that iron content in some scrap may be as low as 70 % and for most conventional scrap types will lie in the range of 88 – 92 %. This is due to the coatings, oil and grease, dirt and moisture associated with the scrap in addition to oxidation of the scrap iron content. Even #1 grade scrap will typically contain surface coatings and oil and grease, which will result in an iron yield of less than 96 %.

It can be seen that metallization in conventional scrap may be fairly high, though the yield of metallic iron to the finished product is low compared to some AI materials. This is because many melting operations overlook the actual iron content of the scrap. Indeed yield is difficult to determine and in many cases, scrap cost becomes the predominant factor for scrap purchase. In AI materials, the iron content is easily identified and this fact explains why it is used as a quality control parameter for these materials. For most DRI/HBI materials, the feedstock is a high grade iron ore, in which case, the iron content in the finished product will be very high. For the purpose of quality control over all types of EAF feed it is best to consider metallization only when reviewed along with total iron content as well.


Gangue content in DRI is probably the reason cited most often for steelmakers’ reluctance to use DRI in the EAF. Gangue consists of both acidic and basic components. Acid components include silica, alumina and titanium oxide. Alkali oxides such as Na2O and K2O and titanium oxide may also be present in the iron ore though usually in smaller quantities (<0.1 %). Basic components include lime and MgO. Gangue does not contribute to steelmaking in the sense that it basically “comes along for the ride” with the iron in the DRI. Silica and alumina are usually present in the iron ore and also in binders used in pelletizing. CaO and MgO may be present in small quantities in the iron ore but usually result from additions to iron ore pellets. Silica and alumina levels in the DRI are in a large part dependent on the selection of iron ore for the process. Selection of appropriate iron ores is essential for acceptable quality DRI. Sulfur and phosphorus contents in the iron ore are also a concern because they are not removed in the direct reduction process. Usually, about 1.5 tons of iron ore is required to produce a ton of DRI. Thus any undesirable materials in the ore actually become more concentrated in the DRI product. For most gas based DRI processes, silica levels of 1-3 % in the DRI are typical. Alumina levels of 0.5 -1.5 % are also typical. The biggest concern with gangue materials is that they may require additional flux additions and the resulting slag components consume energy within the steelmaking process. Steelmakers want the gangue content to be as low as possible because this material reports to the slag in the EAF and lime must be added to help maintain the desired slag basicity. The effect of gangue content has been studied extensively in other papers and will not be discussed here (3, 13, 14). Gangue is typically not associated with pig iron, though most pig iron contains silicon, which will eventually form silica, which will report to the slag. Silicon content can vary from as low as 0.3 wt % up to 2.5 wt %. Manganese content is typically around 0.4 wt % but can be as high as 1.0 %. Both the silicon and the manganese will be oxidized during oxygen injection operations and will supply energy to the steelmaking process. However, the silica that is generated will require additional basic flux additions and essentially can be considered as gangue content in the pig iron. In the case of “beach iron” or “flat iron”, the material is sometimes seriously contaminated with sand and concrete. This can results in significantly lower metallic yield and will contribute significant quantities of gangue to the EAF. This type of material varies very widely in quality and a visual inspection should always be carried out when a shipment is received. Gangue content associated with this material can vary from 2 – 20 % by weight. Conventional steel scrap typically contains 0.1 % silicon, 0.1-0.3 % carbon and 0.3-0.6 % manganese. Thus scrap can be considered to have a relatively low contained gangue content. That is not to say however that conventional scrap does not contain gangue. Most gangue associated with conventional scrap is in the form of surface contamination or material mixed with the scrap as is common in bundles. In the case of scrap types that contain surface oil or grease, dirt will tend to adhere to the scrap surface thus contributing to gangue associated with the scrap. It has been determined that dirt can contribute between 0.25 and 1.0 weight % gangue to most conventional scrap types including prime #1 grades. When “zeroing out a scrap pile”, a large amount of dirt may be carried over to the scrap bucket. This dirt really serves no purpose in the EAF and actually increases melting costs. If large quantities of dirt are present in the charge, additional CaO and MgO must be added to the furnace to prevent accelerated refractory wear. This will result in additional power requirement, increased electrode consumption and perhaps extended tap-to-tap time. In addition, there will be additional iron yield loss due to the greater slag volume. Thus it can be seen that gangue content is a concern for conventional scrap types as well as alternative iron materials.

Carbon, Si and Al content

Scrap chemistry and residual levels play an important part in the scrap selection process. Obviously, the melt-in chemistry must meet the required limits for residual elements (Cu, Sn, Mo, Ni). Other elements that have an important affect on steelmaking include carbon, manganese, silicon and aluminum. Typically, aluminum levels are less than 0.05% in scrap (unless a large proportion of the scrap is galvalume) and thus are not a major concern. Silicon and manganese will react with oxygen to form oxides, which will report to the slag. Carbon will react with either oxygen or FeO to form CO gas, which leaves the furnace in the offgases. In most cases, the reactions with oxygen are exothermic (the reaction of carbon with FeO is mildly endothermic) and supply energy to the EAF process.

The % carbon in the DRI can be manipulated based on the flow of reducing gas, the residence time of the DRI in the shaft and most easily by adjusting the amount of cooling gas used in the lower zone of the shaft furnace. Usually, steelmakers prefer a carbon content in the DRI above 1 %. In the past, the carbon content in the DRI was engineered to match the metallurgical requirement to recover the Fe tied up in the form of FeO. In recent years, an excess of carbon in DRI has become a popular method for introducing additional carbon into the melt and thus reducing the charge carbon requirement. However, it should be noted that DRI has a tendency to re-oxidize both in transit and in storage. Thus the additional carbon in the DRI may in fact be necessary to recover the iron units in the DRI.

HBI can also have high carbon content. Typically though, merchant HBI products rarely contain more that 2.8 % C. Most HBI is formed from pellet based DRI processes and the carbon content in DRI pellets tends to be concentrated in the outer portion of the pellet. If the pellets contain a high carbon content, they become more difficult to briquette. The integrity of the briquette suffers and much higher handling losses are likely to result.

Pig iron is a material in which the chemical composition can vary quite considerably (8). The silicon content can vary from as low as 0.3 wt % up to 2.5 wt %. Manganese content is typically around 0.4 wt % but can be as high as 1.0 %. Carbon content varies between 3.5 and 4.5 and is usually greater than 4 % in most operations. This contrasts sharply with scrap, which typically contains 0.1 % silicon, 0.1-0.3 % carbon and 0.3-0.6 % manganese. In some cases, the silicon content can be as high as 2.5% in cast iron and 1.8% in some types of borings.

It is important to note that the amount of carbon charged to the furnace can affect operating parameters:

    ∙ The melt-in carbon level determines the initial rate of FeO generation during oxygen injection. For bath carbon level above 0.3 weight%, almost all of the oxygen reacts to form CO. As the carbon level drops between 0.3 and 0.15 weight%, increasing amounts of the oxygen react with iron to form FeO. FeO generation is necessary for slag foaming operations. In order to sustain slag foaming, the FeO concentration in the slag must be maintained within a relatively narrow range. Thus, the total amount of carbon charged to the furnace coupled with oxygen injection practices will determine when the slag attains optimum foaming conditions.
    ∙ Charge carbon must be balanced with total oxygen injection in order to achieve the desired tap carbon level. If the bath is blown “flat” with higher oxygen content than the desired level, it is likely that the yield loss is too high.

When large quantities of pig iron are used, it may not be necessary to add charge carbon at all. Every 1% of pig iron in the charge supplies 0.87 lbs per ton of charge carbon (assumes 4% C in pig iron, and 92% scrap yield). Thus 20% pig iron in the charge would supply the equivalent of nearly 18 pounds per ton of charge carbon. The recovery of carbon contained in metallic feed materials is very high (typically 90 – 100 %). When improved carbon recovery is taken into account, this amount of pig iron might replace 20 to 60 pounds per ton of charge carbon. Note that the use of large amounts of high carbon materials can lead to extended decarburization requirements, which must be planned for.

In electric furnace steelmaking, some carbon will be contained in the scrap feed, in pig iron, DRI, HBI or other alternative iron furnace feeds. The total amount of carbon contained in these EAF feeds will generally be considerably lower than that required to meet the desired tap carbon level and oxygen injection requirements, so some additional carbon is typically charged to the EAF. In general, the amount of charge carbon used will correspond to a carbon/oxygen balance, as the steelmaker will try to maximize the iron yield. Typical charge carbon rates for medium carbon steel production lie in the range of 5 – 25 pounds per ton of liquid steel.

Injection of carbon materials for slag foaming will typically lie in the range of 10 to 30 pounds per ton of steel produced. It is important to note that the fixed carbon in charge materials can vary significantly and ash content can be as high as 12%. The method of placement in the charge bucket can have a major impact on carbon recovery, as can the size of the carbonaceous material. Studies have shown that material sized between ½ inch and ¾ inches gives the best recovery for conventional EAF operations. Coal can have fixed carbon levels ranging from 60 to 80% and recoveries in the EAF also vary from 20 to 80% depending on the coal size and method of addition to the furnace. Most EAF operations aim for a melt-in carbon level of approximately 0.1 to 0.3%.

Due to erratic charge carbon recovery in the EAF, many operations have turned to high carbon feed materials as a way to reduce the variations in steelmaking operations.

Other AL Quality Parameters

Scrap Yield

Yield is dependent on both material physical properties and on method of use in the EAF operation. Scrap yield is an important issue and impacts heavily on the true cost of using various types of scrap. Yield loss can be attributed to:

    ∙ Non-metallic content of the scrap – gangue, paint, oil, grease, coatings, dirt etc.
    ∙ Operating losses – iron units that are oxidized and are not recovered.
    ∙ Elemental content that is refined from the steel – aluminum, silicon, manganese, and phosphorous.
    ∙ Iron that enters the EAF as an oxide and is not recovered.

Any non-metallic elements associated with the scrap typically will not contribute to the tapped weight of the steel. Metallic content that oxidizes or vaporizes in the EAF will likewise not contribute to the finished mass of liquid steel. Some of the factors that should be considered include:

    ∙ Almost all the silicon and most of the manganese will be oxidized from the scrap during the melting and refining processes. Thus, high levels of either of these two elements in the scrap (particularly silicon) will tend to lower yield.
    ∙ Typical carbon content in the steel when tapped is between 0.05 and 0.2%. If the scrap contains levels of carbon higher than this, it will be removed during oxygen blowing operations. This will reduce yield.
    ∙ Oil and grease will burn away from the scrap during meltdown, thus reducing yield.
    ∙ Dirt, concrete and other materials that are typically found in the scrap will report to the slag, thus reducing yield and potentially increasing flux and energy requirements.
    ∙ Paint and zinc coatings will report to the EAF dust and do not contribute to the weight of liquid steel.
    ∙ Any scrap that is oxidized during burner or oxygen lance operations will form FeO. Unless sufficient carbon is added to the slag, this iron will not be recovered.
    ∙ Any rust present on the scrap represents a yield loss, with the exception of any iron units recovered by charged or injected carbon.

Yield losses associated with AI materials are mostly related to gangue materials. Other forms of yield loss listed above are typically not present in AI materials. In the case of pig iron, yield losses are associated with silicon, manganese, iron and carbon content, which are oxidized in the steelmaking operations.

Oil, grease and coatings are fairly common in both prompt and obsolete scrap types. Oil and grease are commonly associated with borings and turnings, a result of the lubricants used during processing of the steel. Some scrap operations allow these types of scrap to sit in a drainage area so that the oil can be separated from the scrap prior to shipping the scrap to the steel mill. Frequently, however, the scrap is shipped directly to the steel mill from the source and levels of 2 to 3% oil may be present in the scrap. This represents a large yield loss when the scrap is melted, and can also have a significant impact on offgas system operations.

Coatings are fairly common on shredded scrap, resulting from automobile processing or from the processing of “white goods” (appliances). These coatings may include paint or zinc in the form of galvanized scrap. In the case of galvanized scrap use, high zinc levels in the furnace freeboard can occur, resulting in arc stability problems. Zinc coatings will typically amount to 0.5 to 2.0% by weight on galvanized scrap. This represents a significant yield loss and also greatly increases the amount of EAF dust generated per ton of steel produced.

Scrap may also contain free water, i.e. surface water that is not chemically bound to the scrap. This water will evaporate from the scrap when it is placed in the furnace. Free water represents both a yield loss and a drain on energy. In many cases, it is difficult to keep scrap dry unless it is stored under covered enclosures. Some operations have found it beneficial to store scrap on sloped concrete pads. This helps to reduce dirt carry-over when the bottom of the pile is charged to the bucket and also allows some water and oil to drain from the scrap to a collection tank. In locations where temperature drops below freezing in the winter, water contained in the scrap can sometimes lead to furnace “explosions”.

Borings, turnings and shredded scrap are usually associated with the greatest amount of free water content. In some cases this is because this scrap is “washed” to remove oil and lubricants. Every 0.5% of free water in the scrap represents an energy drain of 11.2 kWh per ton of scrap charged to the EAF. When energy efficiency and scrap yield are taken into account, this represents an increase of approximately 20 kWh per ton of product.

Scrap Fines Generation

Two major concerns with DRI use that are frequently over-looked are fines generation during handling and re-oxidation of the DRI. Both of these mechanisms lead to increased yield losses. Typically DRI pellets can be expected to generate much more fines than briquetted material. Several users of DRI have measured fines generation during shipping due to handling of the pellets and estimate that the –1 mm fraction typically increases by 3-5 %. This is an issue for the end user as this material re-oxidizes easily and can potentially be a fire hazard. Even if this material makes it into the furnace, it is likely that a large portion of the fine material will report to the EAF offgas system. Some users have collected the fine material and have injected it into the furnace at the bath/slag interface. When used in this manner, this material can be a very effective slag foaming agent.

Some DRI facilities use lump ore as a feed material. This results in lower feed material costs per ton of DRI produced but the lump ore portion of the DRI will generate a greater amount of fines during processing and will produce much greater quantities of fine material during shipping and handling. Thus yield losses will tend to be much higher when using this material in the EAF.

Great care must be taken to keep DRI dry. Usually, operations using large quantities of DRI will store the DRI in silos. These silos are usually passivated with nitrogen. DRI is extremely porous and re-oxidation will occur very easily if it is not properly stored. DRI can absorb 10 to 15 % moisture if it is stored uncovered in wet weather (15). HBI has a much lower surface area and several re-oxidation tests indicate that it re-oxidizes at approximately 30 % of the rate experienced by DRI pellets. HBI tends to absorb only 3 to 5 % moisture when stored uncovered (15). DRI produced from lump ore tends be as reactive or more reactive than DRI pellets.

Tests conducted on DRI pellets stored in unprotected piles indicated a reduction in metallization of 7 %. For every 1 % increase in FeO in the DRI, an additional 10.4 kWh/tonne DRI, (theoretically), is required to recover the iron. EAFs are usually 58 to 60 % efficient and therefore as much as 17 kWh may be needed. If the DRI iron yield is assumed to be 89 %, a total of 19 kWh per tonne of steel would be required. Thus a decrease of 1 % in DRI metallization could increase the power requirement per tonne of steel by almost 25 kWh/ton. This could represent an increase of 6.3 % in tap-to-tap time for a 100 % DRI charge based on FeO content alone. Several analyses of the effect of DRI characteristics on the cost of steelmaking have been conducted (13, 14). These studies indicate that a decrease of 1% in DRI metallization results in an increase in conversion costs of $5/ tonne of liquid steel.

Of greater importance than the additional cost associated with recovering the iron units is the effect that the re-oxidized iron will have on the slag chemistry in the EAF. The DRI will typically only contain sufficient carbon to reduce the un-metallized portion of the pellet ( 1 % C for every 6 % FeO). Thus unless the pellet contains excess carbon, there will be a shortfall of carbon in the EAF. This can result in wide fluctuation in the slag FeO content and may result in accelerated wear of furnace refractories. If sufficient carbon injection capacity is not available, iron yield will suffer, as will slag foaming, which will in turn reduce energy efficiency in the EAF.

When all factors are taken into account, most high tonnage steel producers will prefer to maintain as high a metallization as possible in the DRI. It is however surprising to see how few facilities utilize closed storage bins for DRI. For facilities using less than 20 % DRI/HBI in their operations, the best course of action is to purchase HBI if it is available. Even though the cost to briquette HBI is approximately $3 to $5 per ton, the potential losses in re-oxidation of DRI pellets easily out-weigh this additional cost. For facilities utilizing greater quantities of DRI, an enclosed storage system coupled with a continuous feed system should be considered a necessity.

Scrap Sizing

Sizing of the scrap can have a significant effect on melting operations. The acceptable scrap size is generally a function of the size of the furnace and the power rating of the transformer. Many operations have reduced their use of scrap bundles because they take too long to melt and tend to cave in, resulting in electrode breakage and lost operating hours.

Generally speaking, smaller scrap melts faster than larger pieces due to the relatively larger surface to volume ratio. Similarly, bulk scrap volumes with higher density take longer to melt than similar volumes of less dense scrap. Despite the melting rate advantage, a 100% charge of small scrap pieces is not advisable since it might result in excessive scrap cave-ins and arc instability.

Usually, a maximum of 20 % cold pig iron is used in the EAF because it takes longer to melt in than scrap, especially if it is supplied as large pigs. Small sized pieces are preferable. If higher percentages of pig iron are used, the oxygen blowing profile must be adjusted to ensure that decarburization delays do not result. DRI/HBI charged to the EAF in the scrap bucket can form clumps as it heats up and these can result in cave-ins later in
meltdown. These cave-ins frequently lead to violent reactions in the slag and pre-mature flushing of slag from the furnace. It has been found that proper layering of the AI materials in the charge is very important for yield, energy consumption and to prevent sticking of these materials. Most AI materials are small in size and can be blended in with larger scrap fairly easily. If large quantities of DRI are to be used, roof feeding is usually employed. HBI can also be fed through the roof though care must be taken to ensure that the roof feed system has been designed to handle HBI, which is larger and denser than DRI.

Scrap Density

Scrap bulk density is of major importance in EAF operation. If low-density scrap is used, it may be necessary to increase the number of back-charges to the EAF. Low-density scrap can be difficult to handle when charging the furnace. Low-density scrap tends to sit up in the furnace, making it difficult to close the furnace. In such cases, it is common to see the crane operator using the scrap bucket to “pat down” the scrap in the furnace so that the roof may be swung. Alternatively, it may become necessary to use the crane magnet to remove some of the scrap from the furnace. Almost all AI materials are popular for increasing charge density and reducing the number of back-charges to the EAF.

Sometimes a charge does not fully fill the furnace due to high bulk density, resulting from substantial pig iron or DRI/HBI use. This can cause damage to the furnace shell if the electrical melt-in profile is not adjusted to accommodate the denser charge. If the arc is not buried well into the scrap, there may be excessive radiation to the furnace sidewalls and roof, which may result in panel and delta damage. Therefore, scrap density should be visually checked on a regular basis and electrical power input profiles should match scrap charging profiles in the furnace to ensure high melting efficiency and minimal damage to the furnace shell.

Key Operating Parameters For EAF Iron Bearing Feed Materials

Melting Energy Requirements

In the past, AI materials have been blamed with higher electrical power use related to greater slag volumes and increased flux requirements. Electrical power requirements are obviously tied to the amount of chemical energy use. It has been shown in many operations using AI materials that yield can actually be improved if the materials are used properly. Many EAF operations now utilize slag v-ratios of between 1.8 and 2.2. This has
greatly minimized the impact of gangue content on flux requirements and the resulting slag volume. As a result, yield becomes the biggest factor affecting melting energy requirements. This is not unique to AI materials but is also very important for conventional scrap types. It is anticipated that much greater attention will be paid to scrap yields in the future.

Flux Requirements

As has been mentioned previously, many EAF operations now utilize slag foaming strategies, which employ lower slag V-ratios (1.8 – 2.2) and use more MgO in the flux charge. Therefore the effect of gangue content in the charge is not as significant a factor as it was when V-ratios of 3 and higher were employed. The reduction in slag volume has also resulted in reduced iron losses to the slag. However, few EAF operations track slag chemistry on a regular basis and therefore fluctuations in the composition of feed materials and scrap gangue content can still result in wide variations in slag foaming efficiency. A wide variation in the silicon content in pig iron is one of the biggest offenders. As EAF operations begin to track raw material compositions and slag chemistry better, it is expected that slag foaming operations will improve. Users are beginning to understand gangue content of traditional scrap types to a greater degree and as better processes are put in place to control slag chemistry in the EAF, operations will only improve.

Slag Foaming

Foamy slag practice originated with DRI melting operations where FeO and carbon from the DRI would react in the bath to produce CO which would “foam'” the slag. At the start of meltdown the radiation from the arc to the sidewalls is negligible because the electrodes are buried in the scrap. As melting proceeds the efficiency of heat transfer to the scrap and bath drops off and more heat is radiated from the arc to the furnace sidewalls. By covering the arc in a layer of slag, the arc is shielded and the energy is transferred to the bath. When foamed, the slag cover increases from 4 inches thick to 12 inches or more. HYL estimates that the slag thickness when using 80 % DRI will be approximately double those for a 100 % scrap charge. HYL also indicates that at high carbon removal rates and 80% DRI, the foamy slag thickness can be as high as 40 inches (17).

Suitable slag chemistry is critical for slag foaming. Thus, if the dirt content of the scrap is unexpectedly high or low, slag foaming will be difficult to achieve because the slag V ratio will not be in the optimum range. Sufficient foaming during the flat bath period is required so that high electrical input can be used without sacrificing efficiency and or causing damage to the furnace shell. Slag foaming is possible over a range of V ratios, but optimal slag chemistries ensure the sustainability of the foamy slag.

Though it is recognized that DRI melting contributes significantly to effective slag foaming, very few operations actually attempt to place a value on this. Slag foaming is probably one of the least understood phenomena in EAF steelmaking today and yet it is probably the most critical factor for achieving good energy efficiency in the EAF.

When roof feeding DRI into an EAF, slag foaming results almost immediately as the carbon in the DRI begins to react with FeO in the slag. If the DRI federate is too high, the pellets may freeze and agglomerate and the rate of foaming will decrease substantially. Top feeding of HBI does not result in such vigorous foaming as the briquette must partially melted before carbon is exposed to react with FeO in the slag. More of the carbon in the HBI is likely to react with FeO in the briquette as the briquette melts. This may result in improved recovery of the iron units in the briquette but will also result in a less vigorous boil and therefore less intense slag foaming.

When CO is evolved within the bath, it will also help to strip nitrogen and hydrogen from the steel.Decarburization is also beneficial for the removal of hydrogen. It has been demonstrated that decarburizing at a rate of 1 % per hour can lower hydrogen levels in the steel from 8 ppm down to 2 ppm in 10 minutes. In the case of DC furnaces, it is even more critical to ensure that the arc is buried because, typically, these furnaces tend to operate at much greater arc voltages than conventional AC furnaces. Thus the use of DRI in a DC furnace may be an effective way to ensure good electrical efficiency.

It may prove to be a worthwhile study for some EAF steelmakers to evaluate whether the additional benefits of improved slag foaming from DRI, are sufficient to offset the additional power requirements traditionally associated with the use of these materials.

Effect of scrap layering

Batch charging of large quantities (> 20 %) of DRI/HBI can result in large swings in both bath chemistry and slag chemistry especially if the DRI/HBI segregates out in the charge. These fluctuations make it difficult to maintain consistent operations in the EAF and will likely increase power consumption, electrode consumption and tap-to-tap time. A more flexible flux addition strategy must be used with a corresponding reduction in effectiveness of slag foaming. If DRI/HBI is not well distributed in the charge bucket, it can stick to the furnace sidewalls. When this material eventually caves in, a violent reaction may result as FeO reacts with carbon. This can be a safety hazard, and slag may violently shoot out of the slag door. It is likely that a CO spike in the offgas will occur, temporarily over-loading the offgas system and potentially causing explosions downstream in the system. Slag chemistry may change very rapidly, with the result that slag foaming may be difficult to control and maintain. If the DRI has insufficient carbon content to reduce the FeO contained in the DRI, the slag FeO level may become too high making it aggressive to furnace refractories and making it difficult to regain good slag foaming in the furnace. Thus several potential issues can arise if DRI/HBI is improperly batch charged to the EAF.

More recently, several operators have discovered the benefits of charging DRI/HBI in close proximity to pig iron. Several operations have found that this resulted in lower power consumption and greatly improved yield (in some cases the HBI yield improved more than 5 %). The reasoning behind this phenomenon is that the HBI melts in thus releasing FeO, which is then reduced due to the high carbon content resulting from the pig iron. The net result is that almost all of the iron content in the DRI/HBI is recovered.

In general, it has been found that good blending of the AI materials with the conventional scrap results in improved yield and reduced power consumption while minimizing the negative melting aspects of these materials.

Pig iron should be placed in the bottom half of the first charge bucket so that it will melt into the bath early in the heat. This allows for adequate decarburization time, (thus avoiding the possibility of an extended tap-to-tap time), and allows for maximum recovery of energy from CO post-combustion to cold scrap in the furnace.

Uniformity of melt in

The preceding discussion on the effects of proper scrap layering is closely related to the uniformity of melt in of the scrap. With proper layering, the scrap will melt in uniformly and cave-ins can be avoided. This is especially important when using AI materials as the rapid melt-in of AI materials causes rapid swings in slag chemistry leading to loss of foaming efficiency. This has been discussed in the section on slag foaming.

Maximizing The Utilization Of AI Materials In The EAF

It is hoped that the preceding sections have demonstrated the importance of both physical parameters and operating parameters with respect to optimizing scrap use in EAF steelmaking operations. While it has been intended that the focus of this paper is the use of AI materials, it can easily be seen that these concepts can and should be extended to all scrap types used in EAF steelmaking. In order to maximize the utilization of materials in the EAF, the following procedure is recommended:

    1. Establish operating goals up front.
    2. List any operating constraints (O2 injection rates, burner rates, max power input etc.)
    3. Understand the characteristics of the materials being used
    4. Perform scrap inspection regularly and use the information gained to help optimize EAF operations.
    5. Use process data feedback to gain a better understanding of scrap parameters
    6. Identify scrap quality issues and maximize effective use of materials.


The biggest obstacle that AI materials must overcome to become conventional EAF feed materials is the way in which they are evaluated. In many cases, AI materials are held to a higher standard than many of the conventional forms of scrap that are used in the EAF today. This is not to say that the parameters that are used for evaluating AI materials are inappropriate. Rather, it becomes apparent that we might do a better job of evaluating conventional scrap materials if we applied these same quality parameters. There is a need within the EAF industry to adopt a more universal set of parameters for the evaluation of EAF feed materials.

Initially alternative iron sources were viewed only as diluents for scrap with unacceptable residual levels. A more recent approach has been to look at some of these materials as both a source of clean iron units and as a source of fuel for the process. In some cases this coupling is synergistic but care must be taken to clearly define objectives and constraints prior to using these materials in the EAF. The next step in the optimization process is to control the amount of variability in the EAF. Gaining a better understanding of feed materials and their use in the EAF will go a long way towards satisfying this requirement.

EAF operators will continue to look for more innovative ways to combine and use various forms of iron alternatives in the furnace. The ultimate consideration will be dependent on many factors including furnace design, availability of raw materials, cost of raw materials, availability and cost of energy sources, desired product mix, level of post furnace treatment/refining available, capital cost and availability of a trained work-force. Ultimately the “right” selection will be one that meets the specific requirements of the individual facility at least cost, while meeting product requirements and maintaining process flexibility. A lot of valuable experience has been gained with respect to the use of AI materials in the EAF.

Ranking of various forms of alternative iron materials for use in the EAF is many times subjective in nature. Many factors enter into the evaluation including: feed temperature, iron content, chemical fuel content (Si, C, Mn etc.), gangue content, melting power requirements, oxygen requirements, flux requirements, refractory consumption etc. EAF operators must continue to improve their understanding of the raw materials that they use. There are several good process models available and these can be calibrated to existing operations prior to evaluating the effects of changing the furnace charge make-up.

The use of alternative iron units in the EAF will continue to grow in the developed countries due both to product quality requirements and some of the process advantages that some materials offer. In the developing world, alternative iron will supplement the limited amount of scrap available. As infrastructure grows in these countries, the economy will support the production of higher value-added products and the transition into these markets will be very rapid due to the significant capacity for production of scrap substitutes, which will already be in place.

Operating practices will continue to evolve and will not only seek to optimize energy efficiency in the EAF, but will also seek to discover the overall optimum for the whole steelmaking facility. After all, the most important factor is to optimize operating costs for the whole facility, not necessarily resulting in the optimum costs for individual steps in the process chain. The key factors that will be evaluated will include process flexibility, productivity, product quality and environmental considerations. The selection and use of raw materials will be made to match these operating requirements.