Still the Iron Age

by Vaclav Smil

0.4 mm) of polyethylene and reduce the weight by nearly 40% compared to a standard steel roof (Hoffmann, 2012). Among the promising innovations is a new alloy of iron, aluminum, carbon, and nickel (which forms a shearing-resistant intermetallic compound with Al) that has the strength and lightness of titanium alloys but costs only a tenth as much (Kim et al., 2015).

  The benefits of substitution cannot always be judged simply by the cost or energy requirements of more expensive and more energy-intensive material, and only a life cycle analysis allows us to make the right conclusions. An excellent example is PE International’s (2012) life cycle assessment (LCA) for complete sets of forged aluminum and steel truck wheels (18 in the United States, 12 in Europe, total distances of 1 and 1.5 million km). The study found that aluminum’s higher specific burden of production (energy cost of smelting aluminum is approximately 10 times as much as producing steel) is more than compensated for during the use phase as lightweight aluminum wheels make it possible to carry additional cargo or deliver improved fuel efficiency. In a world where CO2 emissions would be a key parameter determining material preferences, aluminum would be an indisputably better choice.

  But the process can go the other way: the majority of recent replacements have used lighter materials to displace steel, but, even after 150 years of advancing steel uses, there are still applications where steel could become the dominant material. A prominent example is the replacement of wooden utility poles by steel structures. North America is a continent of wooden electricity distribution poles, with a total of 185 million units, requiring about 2.5 million annual replacements. So far, about 1 million steel distribution poles have been installed by the US utilities, and an obvious question to ask is what would be the consequences of replacing a substantial share of them by galvanized steel poles?

  A revealing illustration of the complexities of this material substitution has been provided by a detailed LCA of the two options (SCS Global Services, 2013). The comparative LCA looked at six impact categories and more than 40 indicators and found that in 21 cases steel poles have a significantly lower impact (a difference of 25% or more) than wooden poles, and that for 12 indicators the difference was 100% or greater. These advantages included above all reduced depletion of energy resources (by about half) and lower (about 60%) greenhouse gas emissions, absence of exposures (of humans and ecosystems) to hazardous materials (containing arsenic and chromium) used to treat wood, absence of exposure to toxic herbicides used in forest management, and elimination of habitat disturbance and biodiversity loss attributable to large-scale tree plantations in the US Southeast. In contrast, wooden poles result in lower regional acidification (33% less), lower (40% less) ground-level ozone exposure, and, of course, vastly lower depletion of lead and zinc.

  Substitution opportunities vary greatly by product category, and they can go only so far: applications ranging from the hulls of large vessels to reinforcing bars in concrete, and from high-pressure boilers to durable cutlery have no prospect to see steel displaced by other materials. In contrast, steel’s share of the mass of common consumer goods will continue to decline as aluminium alloys, new plastics, and new composite materials will be making greater contributions in production of electronic gadgets and tools as well as in transportation (bogies of rapid trains are steel, but their bodies and interiors are already nothing but aluminium alloys and plastics).

  Substituting Coke with Charcoal

  Replacing coke with metallurgical charcoal in modern BFs would be an extraordinary challenge. Coke consumption for iron ore reduction in BFs was about 600 Mt in 2013 (additional coke is used for sintering and pelletizing ores, in cupola furnaces and in nonferrous metallurgy), and even when assuming a very conservative growth rate of pig iron production it would be about 700 Mt by 2030. As already explained, coke and charcoal have very similar energy densities of about 30 GJ/t. But their specific densities differ: preferred bulk density of metallurgical charcoal is at least 0.4 g/cm3, and while a widely used Brazilian charcoal from eucalyptus surpasses that with densities of 0.53–0.59 g/cm3 (Pereira et al., 2012), many trees yield charcoal whose density is well below the desirable level.

  Brazil entered the twenty-first century as by far the world’s largest charcoal producer and as the only major consumer of metallurgical charcoal. The fuel is converted both from illegally harvested natural forests and from plantations of fast-growing eucalyptus located mostly in Minas Gerais (Peláez-Samaniegoa, 2008). No by-products are recovered, and the conversion efficiency remains no better than 25%. About 75% of all produced charcoal is destined for BFs, with typical specific requirements of about 2.9 m3 (or 725 kg) of charcoal to smelt a tonne of hot metal (Ferreira, 2000). Brazilian charcoal is used in BFs whose internal volumes are an order of magnitude smaller than those of modern coke-fueled units: the largest Brazilian furnace has just 568 m3 (Pfeifer, Sousa, & Silva, 2012).

  The size limit is explained by the different compressive strength of the two fuels, a key quality required to support heavy charges in large furnaces. Charcoal is too fragile to support heavy burdens in BFs whose shafts are commonly taller than 10 m and as high as 30 m and whose internal volumes are in excess of 5000 m3. And while charcoal furnaces have typical volumes of just 350 m3, coke-fueled furnaces are now commonly larger than 3000 m3. These qualitative differences preclude any simple mass-for-mass substitution of coke with charcoal in today’s large BFs using coal-based fuel.

  Consequently, even if wood supply posed no constraints, charcoal’s inferior compressive strength would require a massive restructuring of the industry in order to produce pig iron in smaller furnaces where charcoal could support lighter burdens. The cost of such transition would include the construction of many smaller furnaces and of the establishment of extensive tree plantations and of mass harvesting operations needed to produce the requisite wood, a change inevitably resulting not only in lower smelting productivities, increased energy costs, and higher metal prices but also in significant negative environmental impact.

  Brazil offers the only instance of relatively large-scale modern reliance on charcoal in ferrous metallurgy, but using its experience as the basis for further expansion of wood-based iron ore smelting would be quite problematic. During the first decade of the twenty-first century about a third of the country’s pig iron output originated in small charcoal-fueled BFs located mostly in the states of Pará, Minas Gerais, and Mato Grosso do Sul (Uhlig, 2011). Up to a third of the required wood came from illegal cutting of natural (primary or secondary) forests (Monteiro, 2006), and according to Uhlig (2011) about 15% of the Amazon’s deforestation has been due to the harvests of wood for charcoaling. The rest of the charcoal was made from eucalyptus trees grown in expanding plantations reaching nearly 5 Mha in 2011 (Pereira et al., 2012).

  Expanding the prevailing Brazilian charcoal-making method to meet much larger global needs would be disastrous. About four-fifths of the fuel comes from small, inefficient, and polluting semicircular brick-and-mud kilns known as hot tail, rabo quente. Much like their coal-coking predecessors in the early industrial United States, these 2.5-m-high brick beehives are built in massed rows, they are charged with air-dried wood, lit and let to smoulder for up to a week, and after a 3-day cooling period the fuel is unloaded and transported to BFs. Greenpeace (2013) described the typical working conditions at these charcoaling operations as a kind of hazardous slave labor. Studies have shown how underpaid workers are exposed to high temperatures, dust, smoke, and uncontrolled emissions of nitrogen and sulfur oxides, benzene, methanol, phenols, naphthalene, and polycyclic aromatic hydrocarbons (Kato et al., 2005).

  Enforcement of appropriate labor regulation could solve that, but these inefficient furnaces could not be used to supply vastly increased charcoal needs. Brazilian experience shows that the charge of 450 kg of carbon per kg of pig iron can be supplied by 630 kg of charcoal (Sampaio, 2005). Global output of about 1.2 Gt of pig iron and average charge of 630 kg of charcoal per tonne of pig iron would require about 750 Mt of c
harcoal, about 15 times as much as the fuel’s recent worldwide production (FAO, 2014). With average charcoaling efficiency of 25% (Bailis et al., 2013), this charcoal-based smelting would require about 3 Gt of wood. The actually produced total would have to be at least 5% higher because long-distance transportation (inevitable given the much higher overall global demand) and handling of such a friable fuel would result in unavoidable diminution losses, especially for charcoals made from less dense wood.

  Of course, these losses could be avoided by shipping roundwood to charcoaling facilities located close to concentrations of BFs, but again, this option would result in unprecedented levels of wood trade. With a 5% markup, the charcoal-based industry would need about 3.2 Gt of wood. In comparison, recent global wood harvest has been on the order of 3.5 Gm3 or roughly 2.3 Gt (FAO, 2014)—and wood for making charcoal would thus claim at least 40% more wood than the recent worldwide harvest used for lumber, pulp, and fuel. In order to cover all those requirements, the global wood harvest would have to rise to 5.5 Gt/year, roughly a 2.4-fold increase. And while in 2012 all wood-in-rough traded amounted globally to about 70 Mt (FAO, 2014), exporting just half of the wood needed to make charcoal (about 1.6 Gt at the 2015 smelting rate) would require a nearly 23-fold expansion of such sales.

  The extent of the area that would be used to harvest this wood from fast-growing tree plantations would depend on prevailing yields. With a rather high average of 15 t/ha, it would take about 210 Mha, slightly more than half of the area of the entire Amazon basin (410 Mha). But all of these calculations could be seen as too pessimistic because the necessity to replace coke and the new, mass-scale demand for charcoal would engender many innovations that could, on the one hand, lower the specific requirements, and could, on the other hand, increase the yield of cultivated tress and the efficiency of their conversion to charcoal.

  If the harvest were to come only from high-yielding (25 t/ha) clones in eucalyptus plantations (Pfeiffer, Sousa, & Silva, 2012), and if all charcoaling were to be done in modern continuous retorts that can convert 35–40% of wood to nearly pure carbon (Rousset et al., 2011), then the needed area would be reduced by nearly 60% to about 125 Mha. But, in turn, that combined assumption of high yields and high conversion efficiencies may be too unrealistic as it would be difficult to sustain 25 t/ha yields everywhere in the tropics and as such large-scale cultivation would require a great deal of nontropical charcoaling with inevitably lower harvests. Although short-term experiments on small plots indicate some impressively high harvests of temperate plantation trees, it would be unrealistic to expect yields higher than 10–15 t/ha for fast-growing hybrid poplars, pines, or willows (Smil, 2015).

  For the sake of completeness, I should note that there is one exceptional source of charcoal, whose quality is easily comparable to that of metallurgical coke, but whose natural availability is very limited. Babassu palm (Orbignya martiana) grows in northern, northeastern, and central regions of Brazil, forming extensive monospecific forests from the inland state of Goiás to the coastal Maranhão, and it produces ellipsoidal nuts about 10 cm long and 6 cm in diameter, whose very hard endocarp has exceptionally high density (Protásio et al., 2014). Charcoal produced from the babassu nut endocarp has apparent density of 1 g/cm3 (similar to coke) and compressive stress higher than 40 MPa, an order of magnitude higher than typical tree-wood charcoal and almost three times as high as metallurgical coke used in Brazilian BFs (Emmerich & Luengo, 1996).

  Another advantage is babassu charcoal’s low sulfur content. This charcoal could be thus used as a direct substitute for metallurgical charcoal in BFs—but, obviously, current supply could cover only a tiny share of potential demand. Recent Brazilian production of babassu shells is only about 1.5 Mt/year (Protásio et al., 2014), and only large-scale tropical plantations could increase its availability. Regardless of the phytomass origins, worldwide smelting of more than 1 Gt of pig iron with charcoal would have enormous environmental impacts. Relying solely on natural forests would be impossible; converting large shares of tropical forests to tree plantations would obviously be a recipe for intolerably massive deforestation and soil erosion as well as a further assault on tropical biodiversity, while maintaining large areas of high-yielding temperate trees would require repeated fertilization, applications of insecticides, and often also supplementary irrigation.

  How renewable would that be? Obviously, coke-base smelting taps a finite, nonrenewable source of fossil energy, but given the magnitude of the requirement and the size of existing coal resources, we could rely on this option for generations to come even as we find other ways to reduce our mobilization of fossil carbon, above all by increasing the shares of renewable generated electricity and by replacing coal combustion by natural gas in many industrial uses.

  In any case, returning to reality (recall that all of this would be possible only with creating new capacities in small BFs!), there are only two direct substitutions of coal by charcoal that make sense from both the technical and environmental points of view. The first one is to substitute charcoal for a small share of coal in coke production, and Mašlejová (2013) showed by laboratory experiments using 1–5% of wooden biomass instead of volatile coal that this substitution works. She also experimented with an indirect replacement, by using charcoal instead of fine coke in sintering furnaces. Clearly, the most appealing substitution is to use charcoal instead of injected coal: this does not create any problems with burden support and yet it could replace up to 200 kg/t of the fossil reductant. Babich, Senk, and Fernandez (2010) experimented with and modeled this substitution and found that the conversion efficiency of all the tested charcoals was either better than or comparable with the use of coal dust.

  Again, just in order to indicate the magnitude of such displacement, if applied to the entire global output of 1.2 Gt of pig iron, this substitution would require about 240 Mt of charcoal, a much more manageable total that could be grown on as little as 10 Mha and no more than 20 Mha of tree plantations. But all of these calculations were done just in order to reveal the magnitudes of new material flows and ignored the economics of such substitutions. Suopajärvi and Fabritius (2013) showed that while forest-rich Finland has enough wood for current users as well as for possible ironmaking, the economics of the switch would be unfavorable, and Norgate and Landberg (2009) concluded that eventual use of charcoal in iron- and steelmaking will not be competitive with fossil fuel carbon on price alone.

  Partial relief could come by using other phytomass to produce what is now called biocoke, high-density and high-quality fuel that could be made from waste phytomass, including not only woody matter from plantation thinnings but also such cellulosic wastes as bagasse (sugar cane stalks after sugar extraction) and other crop residues.

  Dematerialization

  Dematerialization has been commonly defined as the reduction of material used, be it per finished product (kg/kg) or per unit of economic output (kg/$), unit of power (kg/W), performance, or service delivered (e.g., mass of a computer per instructions per second), and during the past two decades many analyses have demonstrated the ubiquity of such gains (Ausubel & Wagonner, 2008; Smil, 2013). Examples of relative dematerialization abound because the process has been one of the unmistakable defining trends in modern extractive industries, in manufacturing, energy supply, transportation, and service delivery: in all of these cases we have been using a progressively smaller mass of materials to deliver the same or even better products (lighter beverage cans, scratch-resistant yet thinner glass) and higher efficiency (plastics and composite materials in cars and airplanes) at a lower cost.

  I have shown throughout this book that this relative, and persistent, dematerialization has been one of the most important accomplishments of the modern iron and steel industry. Declining quantities of ore, coal, fluxing materials, and total energy have been used to produce a tonne of hot metal in BFs, and the combination of rising conversion efficiencies in BOFs and EAFs and the universal adoption of continuous casting have redu
ced the material and energy demands of steelmaking. This relative dematerialization can also be illustrated by tracing the steel intensity of mature economies. When using constant GDP values adjusted for inflation and expressed in constant 2009$ (BEA, 2015) and crude steel consumption totals from Kelly and Matos (2014), the steel intensity of the US economy fell from 37 kg/$ in 1929 to 33 kg/$ in 1950. In 1973 when the country’s steel output peaked, it was 20 kg/$, by 1990, following the industry’s retreat (see Chapter 4), it was just 9.7 kg/$, it changed little by the year 2000 (9.5 kg/$0), and by 2013 it was down to 6.7 kg/$.

  And steel has been no exception as impressive examples of relative dematerialization can be provided by analyzing secular changes in consumption of other key materials (aluminum, copper, wood, cement) per unit of national GDP or per unit weight of final products, be they airplanes or buildings. Over longer periods of time (including the 1929–2013 rates cited in the previous paragraph) this decline reflects not only gradual technical advances but it is also heavily affected by different stages of economic development: the ratio rises during the period of intensive investment in basic infrastructures; it is stagnant or falling in mature economies with slowly growing and aging populations whose GDP comes largely from less material-intensive services.

  Absolute Dematerialization

  Dematerialization in absolute terms has been, so far, a much rarer occurrence, but steel consumption in some mature, affluent economies is actually among the infrequent examples of this kind. Because economies continue to expand, because very few countries have declining populations, and because stationary populations may be still increasing their average per capita consumption, absolute dematerialization—declining aggregate use or even the complete elimination of a particular material on a national or even global basis—has been very rare. Perhaps its two most notable examples, elimination of chlorofluorocarbons and the use of lead in household paints—have been due to legislative actions caused by environmental and health concerns.