Still the Iron Age

by Vaclav Smil

  McLouth Steel in Trenton, MI, a small company with less than 1% of the country’s ingot capacity, became the technique’s US pioneer in late 1954, capable of producing about 490,000 t a year from three 60-t furnaces with 45-min heats (ASME, 1985). The next installations came only in 1958 (two 100-t furnaces at McLouth and three 100-t furnaces at Kaiser Steel in Fontana, CA), and the two leading US producers, US Steel and Bethlehem Steel, began to operate their first BOFs only in 1963 and 1964 (Hogan, 1971). Adams and Dirlam (1966) used this much belated adoption as the principal example to argue against the universal validity of the Schumpetrian hypothesis, according to which large firms with great market power have greater incentives and more abundant resources to be the innovation leaders (Schumpeter, 1942).

  The adoption pace finally quickened during the 1960s, and by 1970 oxygen furnaces produced 40% of American steel. But Japanese steelmakers, rebuilding the plants destroyed during WW II, embraced the technique so rapidly that by 1970 80% of the country’s steel production originated in BOFs. Subsequent steady progress brought the global share to nearly 60% by the century’s end, when the US and Japanese shares were, respectively, a bit above 50% and more than 70%, and Germany produced 70% of its steel in BOF and Austria 90%. By 2013, the global share reached 70% and many national shares changed only a bit, but the US share declined to about 40%.

  Modern BOFs are massive, open-mouthed, pear-shaped vessels that require 50–60 m3 of oxygen to produce 1 t of crude steel; the gas is blown from both the top and the bottom, as a supersonic jet from a water-cooled vertical lances onto molten pig iron and through the tuyères at the furnace bottom, which are also used to blow in argon (an inert gas) to stir the charge and reduce phosphorus levels in the metal (Miller et al., 1998; Fig. 5.3). The resulting oxidation is much more rapid than in open hearth furnaces: a large BOF decarburizes 300 t of iron from 4.3 to 0.04% C in just 35 min, while the most advanced open hearth would need at least 9 h. The entire process is slightly exothermic (hot gas leaving the furnace, typically 90% CO and 10% CO2, is often used to melt the charged scrap), yielding about 200 MJ/t of crude steel. The largest furnaces are 10 m high and up to 8 m in diameter, and their capacities range mostly between 150 and 300 t/heat.

  Figure 5.3 Molten pig iron is poured into a BOF to make steel at JFE Steel’s Kimitsu Works in Chiba, east of Tōkyō. Reproduced by permission from JFE Steel.

  Productivity advances have been related primarily to the use of oxygen: the early norm was less than 9 m3 of oxygen per tonne of steel; by the late 1990s the typical rate rose to 30 m3/t. In the year 2000, the world’s largest BOF, with capacity of 375 t and tap weights of 325 t in 94 min, was at Northwestern Steel & Wire Company in Sterling, IL.

  In the latest furnaces, oxygen use is up to 50 m3/t, typical inputs per tonne of steel include 1033 t of hot metal, about 40 kg of scrap and ferroalloy, and about 30 kg of flux (lime and dolomite), and by-products include about 100 m3/t of hot gas (about 70% CO, 10–15% CO2) and 50 kg of slag. Charging a furnace with pig iron (delivered by torpedo cars, poured into a ladle, and delivered by an overhead crane) requires tilting of the furnace on its pinions. After it is charged the furnace is returned to the upright position, liquid-cooled lances are inserted, blowing with a high-purity oxygen begins, and slag-forming material is added. About two-thirds of the required oxygen converts C to CO, and 9% goes for oxidizing Si to SiO2 and another 9% for turning Fe to FeO in slag. After a heat is finished the molten steel is transferred to a ladle and either delivered to a tundish to begin continuous casting (see the following section) or cast into ingots.

  The economic impact of BOFs was truly revolutionary. Capital expenditures were cut by an order of magnitude as a single BOF can replace 10–12 open hearth furnaces, and labor productivity rose by 3 orders of magnitude, from 3 man-hours/t of hot metal produced by open hearths in 1920 to 10 s/t by BOFs in 1999 (Berry, Ritt, & Greissel, 1999; Shepard, 2004). Reduced flexibility of charges has been the only drawback: scrap could make up as much as 80% of the open hearth charge but it is limited to no more than 30% in oxygen furnaces. But the unchallenged dominance of these furnaces was much shorter than the reliance on open hearths: by the end of the twentieth century their share of steel output was declining in every major steel-producing nation because the share of the metal produced in EAFs was rapidly increasing (see Fig. 9.5).

  Electric Arc Furnaces

  Origin of electric arc furnaces goes back to the experiments conducted in 1878 and 1879 by William Siemens, one of the eminent inventors and innovators of the nineteenth century, whose other notable metallurgical innovation, the regenerative furnace, was described in the previous chapter. Siemens built two furnaces, one with electrodes at the top and at the bottom (charged material covered the lower electrode), and the other with horizontally opposing electrodes melting the charge beneath them by radiation. But during the late 1870s, there was no way to generate electricity on large scale and at an affordable cost (the first commercial coal-fired power plants began to operate in 1882), and the first commercial use of relatively small EAFs came in 1888 for the newly invented process of smelting aluminum invented by Charles Hall (1863–1914) and Paul Héroult (Smil, 2005; Toulouevski & Zinurov, 2010).

  Before WW I, arc furnaces were used (in small numbers) in places with cheap hydroelectricity as smelters (Noble Electric Steel in California, Stassano furnace design in Torino), and they were charged with pulverized iron ore and charcoal, but most of them were installed as replacements of Bessemer converters or open hearths. By 1913, Héroult’s furnaces produced about 750,000 t of steel, with Germany accounting for 50% and the United States for nearly 20% (Scientific American, 1913). Furnaces were used to make small batches of specialty steel, and their capacities gradually increased from less than 5 t before 1910 to about 25 t by the late 1920s. American electric arc steel production grew by nearly 90% during the 1920s, then dropped and stagnated during the 1930s, surpassed 1 Mt in 1940, and tripled by the end of WW II as the largest furnace capacities reached 100 t/heat.

  EAFs gained a large share of steel production only during the great post-WW II industrial expansion when electricity prices declined and when the previous decades of iron and steel production resulted in accumulation of relatively large amounts of scrap metal. Another key factor in their rapid adoption was the lower capital cost of these furnaces, no more than 15–20% of the total cost of operation combining blast furnace and BOF (Jones, Bowman, & Lefrank, 1998). Key technical advances since the middle of the twentieth century included larger capacities, shorter tap-to-tap times (hence higher productivities), and reduced energy and electrode-graphite consumption.

  By the end of the twentieth century, the best melt shop, at Badische Stahlwerke in Kehl, set a new world record with its best furnace producing 46 heats of 87 t, with tap times down to 36 min (power-on time less than 28 min), average electricity consumption of 302 kWh/t, and oxygen use of 37.5 m3/t (Greissel, 2000). Typical performances have improved as follows: tap-to-tap times declined from 3 h in 1970 to just 30–40 min three decades later, and specific electricity requirements fell from more than 700 kWh/t in 1950 to 475 kWh/t by 1980 and then to less than 350 kWh/t by 2010, while graphite consumption was reduced from about 7 kg/t in the early 1960s to as little as 1.1 kg/t by 2010 (Iron Age New Steel, 1999; Madias, 2014; Stubbles, 2000; Toulouevski & Zinurov, 2010). Electrode diameters increased from the previously common 6.1 cm to 7.1–7.5 cm, and the highest voltages rose from 1000 V to as much as 1600 V. Moreover, all processing required to obtain specific steel qualities was moved out of EAFs to secondary ladle treatment, a shift helping to increase the overall efficiency.

  DC furnaces came into operation with the more common availability of reliable sources of that current; the worldwide shift from AC to DC EAFs began in the mid-1980s (Jones, 2003; Takahashi, Hongu, & Honda, 1994). The main advantages of DC furnaces are a lower specific demand for electricity, electrodes and refractories, and reduced operating noise. But these advantages are offset by higher fur
nace costs and higher specific cost of large-diameter electrodes and have been reduced thanks to the advances in AC operation (Madias, 2014). While the specific power of furnaces with capacities of 50–100 t was between 200 and 250 kVA/t, ultra high-power furnaces (first used in the United States in 1963) have capacities well in excess of 100 t, 70–80 MVA transformers, and specific power input up to 1 MVA/t, with Simetal’s Ultimate design reaching 1.5 MVA/t (Siemens VAI, 2012).

  Refractory linings are now mostly made of water-cooled panels for longer durability. Use of oxygen has increased productivity and reduced electricity demand. During the 1990s, typical injections were 10–15 m3/t of steel, and higher rates, now commonly 40–50 m3 and up to 70 m3 (when the gas is also used for postcombustion of CO), have been combined with the concurrent injection of carbon (up to 15–17 kg/t). This was necessary to prevent the formation of larger amounts of oxidized iron and the resulting decline in yields. In addition, injected carbon produces foaming slag and arc immersion in the slag improves the efficiency of energy use.

  EAFs have remained competitive and gained higher market shares thanks to a continuing quest for higher productivity by cutting operating and maintenance costs. This effort has been particularly important at smaller mini-mills with a single furnace. Relative prices of scrap and hot metal determine the ratios of charged materials and the degree of profitability: lower prices led to increased scrap charging. Common advances have included scrap preheating using oxy-fuel burners, introduction of coal and other carbon additives, and postcombustion of CO produced during the conversion. For furnaces relying on scrap, the raw material amounts to two-thirds of operating cost.

  EAFs are now used to produce all kinds of steel, from low-carbon alloy to stainless grades, and they have flexible charging, able to use not only pig iron but also HBI and DRI (Siemens VAI, 2012). They are installed on an upper level above the shop floor (the earlier models were commonly at grade level and required dug-out pits for discharging). The furnace’s cylindrical shell sits on a spherically shaped bottom and is covered by a flattened curving roof, its sidewalls above the slag line are covered with water-cooled panels, and the entire interior is lined with refractory material. Electrodes are lowered and raised through the central portion of the roof, and the furnace can be tilted for easy tapping.

  Gantry design has a self-supporting roof, oxygen consumption is up to 45 m3/t, and carbon is both charged (about 10 kg/t) and injected (about 7 kg/t). They have high-power inputs (as much as 1.5 MVA/t) and arc voltage up to 1500 V, melting power up to 130 MW, and electricity consumption less than 350 kWh/t, and those with efficient energy recovery and 100% scrap preheating (to at least 600 °C and up to 800 °C) require less than 280 kWh of electricity per tonne of hot metal. The tap weight of the largest furnaces is in excess of 300 t: Simetal’s EAF Ultimate in Gebze mill (Çolakoĝlu Metalurji, Turkey) has a capacity of 315 t, the world’s largest furnace transformer (240 MVA), melting power of 205 MW, tap-to-tap time of 55 min, and electricity requirements of 350 kWh/t (Siemens VAI, 2012; Toulouevski & Zinurov, 2010)—but microdesigns (with tap weights less than 35 t and annual output of 50,000–200,000 t) for small foundries and mills are also available. The other important developments have been the introduction of oxygen blowing (during the 1970s), hot heel, and foaming slag.

  Unlike the EAF’s complex and expensive equipment (including the electrodes), oxy-fuel burners delivering an equivalent amount of heat are simple and cheap. Oxygen was first introduced through consumable hand-held pipes, then by using supersonic lances, and finally through fixed wall injectors obviating any door opening (Madias, 2014). The hot heel practice leaves behind up to 15–20% of hot metal and some slag at the furnace bottom after each tapping; this eliminates the possibility of damaging bottom refractories by electric arcs and allows operating with higher electric power and commencing auxiliary oxygen blowing right after scrap charging. A recent trend has been toward more massive hot heels, even up to 50% of the heat weight (Madias, 2014). Foaming slag, introduced during the 1980s, caused by the ascent of small CO bubbles, is generated by concurrent blowing of oxygen and carbon into the bath; it results in a complete immersion of electrodes, prolongs their life, shields the refractories from the arcs, protects the furnace’s roof and walls from excessive heating, speeds up smelting by increasing heat transfer to steel, and reduces electricity requirements.

  In the United States, EAF output began to rise slowly, from 5.4 Mt of steel in 1950 to 7.6 Mt in 1960, and then it jumped to nearly 18 Mt in 1970 (Hogan, 1971). EAFs made particular large inroads in the United States with the rise of mini-mills. The shift from BOF to EAF severed the link between primary ironmaking (including coking and ore beneficiation) and steelmaking, and it favored setting up smaller mills that could be located without any regard to the sources of coal, ore, or fluxing materials. The rise and the expansion of mini-mills (plants with annual capacities as small as 50,000 t and as large as 600,000 t of metal) became the most dynamic and the most competitive component of US steelmaking. Mini-mill output was initially geared toward low-grade products (bars, structural shapes, wire rods), but later many enterprises began to specialize in higher grades of steel (Szekely, 1987). Mini-mill expansion was also helped by their low capital cost (no more than 15–20% of the total needed for integrated steelmaking) and low operating costs when EAFs were combined with continuous casting (Jones et al., 1998).

  By 1975, there were more than 200 mini-mills worldwide, with nearly half of them in Western Europe and a quarter in the United States; by 1990 the output of US mini-mills surpassed 33 Mt of steel, and by the year 2000, when one-third of the world’s steel originated in EAF, the US steel:pig iron ratio stood at 2.1 as the country’s steel output was split between EAFs and BOFs. By 2013, the United States became an even greater consumer of scrap as the steel:pig iron ratio rose to 2.87 and as American EAFs produced just over 60% (52.6 Mt) of all steel, compared to 40% in the EU, 23% in Japan, and less than 9% in China (WSA, 2014). The share of EAFs in global steel production was rising steadily during the last two decades of the twentieth century, and it was expected to surpass 50% by 2010 (Manning & Fruehan, 2001). But the share was still less than 30% by 2013, mainly due to the combination of China’s dominance of global steelmaking and the country’s very low share of EAF production.

  Secondary refining of steel produced by BOFs and EAFs reduces the presence of all undesirable elements below specified levels. Modern mass-produced high-quality steels should contain less than 50 ppm of phosphorus, less than 5 ppm of sulfur, 10 ppm of carbon, and 1.5 ppm of hydrogen (Iwasaki & Matsuo, 2012). Secondary refining (desulfurization, dephosphorization, and decarbonization) is done in converters or in ladle furnaces (Kumakura, 2013), while concentrations of dissolved H2, N2, and O2 are reduced in degassers.

  Continuous Casting

  Post-WW II changes in the treatment of steel after its production have been no less important than was the shift to BOFs and EAFs. Long-standing practice was inherently energy intensive as the hot metal was first cast into steel ingots, oblong pieces that could weigh between 50 t (for specialty steels) and 500 t (for steel destined to be forged into large pieces) and had to be reheated before they could go through a primary rolling mill, where they were formed into one of the three basic kinds of semifinished shapes: wide (some more than 3 m) and thick (up to 25 cm) slabs; square-profile (up to 25 cm) billets; and rectangular-profile blooms (commonly 40 ×60 cm). Final processing (hot- or cold-rolling) turned standard- and medium-thickness slabs into thin slabs and plates, billets into bars and rods, and blooms into I and H beams. The complete casting and rolling sequence could require no less energy than the steelmaking itself, but it persisted for generations, being the final component of the traditional combination that began with blast furnace pig iron and continued with open hearth steelmaking.

  Steps eliminated by continuous casting include the pouring of hot metal into ingot molds, removing the molds from the ingots, putting the ingots into soaking pits in order to
equalize their temperature, and then rolling them into semifinished products (slabs, billets, or booms). Much like the idea of the oxygen furnace, the concept of continuous casting had also originated with Henry Bessemer: his first patent for a twin-roll caster (using water-cooled rolls with edges sealed by a flange or a dam in groove) was granted in 1865, and the original design was improved in 1891 (Bessemer, 1891), a couple years after R.M. Daelen received a German patent for a vertical casting machine. But Bessemer’s peers were unimpressed:

  One need not, therefore, be greatly surprised that the production of continuous sheets direct from fluid iron did not excite a great amount of enthusiasm in the minds of tin plate manufacturers of that day; in fact, the whole scheme was simply pooh-poohed and laid aside, without any serious consideration of its merits

  (Bessemer, 1891, p. 27).

  Problems with metal sticking and uneven cooling prevented the conversion of these early designs into functioning machinery, and the first successful commercial applications of continuous casting, during the 1930s, were not for steel but for the castings of color metals with significantly lower melting points. Continuous casting of steel owes it eventual success to the combination of the metallurgical expertise of Siegfried Junghans (1887–1954) and entrepreneurial effort of Irving Rossi (1889–1991). A key to this advance was the invention of a vertically oscillating (reciprocating) mold by Siegfried Junghans: it eliminated the possibility of the cooling metal sticking to the mold.

  The first working prototype was ready in 1927; Junghans filed a patent claim in 1933 (inexplicably, it was not granted until 1944) and afterwards devoted himself to adapting the process for steel casting. When, in 1936, he demonstrated his technique with brass casting to Irving Rossi, an American engineer doing business in pre-WW II Germany, it had actually ended in failure as the brass billet skin tore open and hot metal spewed out. But Rossi had correctly distinguished between the challenges with a prototype and the far-reaching commercial potential of the demonstrated technique and immediately secured exclusive rights to Junghans’ patent and to its follow-ups for the United States and England and nonexclusive rights for all countries outside of Germany, and in 1938 he added an agreement for sharing information required to build new continuous casting plants in return for financing such developments outside of Germany (Tanner, 1998).