Matter at the atomic scale: bonds, reactions and process engineering: Cement, steel and polymer plants, and process safety
All courses / Understanding Engineering / Module 10: Matter at the atomic scale: bonds, reactions and process engineering Part 5 of 5Cement, steel and polymer plants, and process safety
Three industries show the earlier ideas working at full scale. Cement and steel turn rock into the shell of a city, and each releases carbon dioxide that the chemistry itself demands. A polymer plant is held back by heat. The part closes with the principle behind process safety, since every plant stores energy and hazardous material.
5.1 Cement shows the limit of chemistry: most of the carbon dioxide from a kiln comes from the limestone, not from the fuel
Portland cement starts with limestone, which is calcium carbonate. Heated to about 900 degrees C in a preheater and the kiln, it decomposes: CaCO3 → CaO + CO2. The lime then combines with clay minerals at about 1,450 degrees C into hard nodules called clinker, which is ground with gypsum into cement (Portland cement; Hydraulic setting of lime and cement).
By weight, clinker is about 65 per cent calcium oxide. Per tonne of clinker that is 0.65 t of CaO, and since 56 g of CaO comes with 44 g of CO2, the limestone releases 0.65 x 44/56 = 0.51 t of carbon dioxide. No fuel or kiln design can reduce that while the product is clinker made from carbonate. Fuel adds more: Indian kilns averaged 731 kcal of heat per kg of clinker in 2020, about 3.1 GJ per tonne (GCCA India), and coal gives about 95 kg of CO2 per GJ, so 0.29 t. The total is about 0.80 t per tonne of clinker, of which 64 per cent is the chemistry.
To make cement, clinker is ground with gypsum and often fly ash, slag or limestone. India's average clinker factor in 2020 was 0.75, so a tonne of cement carries 0.60 t from the kiln, plus grinding and power, and that fits the 680 kg per tonne reported for India as a whole. Switching fuels trims the 0.29 t. Lowering the clinker factor trims both parts: a fall from 0.75 to 0.65 saves 0.08 t per tonne of cement, about 27 million tonnes a year at the 334 Mt that India made in 2019-20. The rest is a job for capture, and kiln gas, rich in carbon dioxide, is a good candidate (Carbon capture and storage; commodity Cement).
| Step | Working | Result |
|---|---|---|
| Lime in clinker | 65 per cent of 1 t | 0.65 t CaO |
| Limestone needed | 0.65 t x 100.1/56.1 | 1.16 t CaCO3 (about 1.5 t of raw meal with the clay) |
| CO2 from the limestone | 0.65 t x 44.0/56.1 | 0.51 t |
| Fuel heat (Indian average, 2020) | 731 kcal per kg x 4.187 kJ per kcal | 3.1 GJ per t of clinker |
| CO2 from coal | 3.1 GJ x 95 kg per GJ | 0.29 t |
| Total per tonne of clinker | 0.51 + 0.29 | 0.80 t |
| Share from the chemistry | 0.51 / 0.80 | 64 per cent |
| Per tonne of cement at clinker factor 0.75 | 0.80 x 0.75 | 0.60 t, before grinding and power |
| Check | reported Indian average 680 kg per tonne of cement (2020, including power) | consistent |
| Saving if the clinker factor falls to 0.65 | 0.80 x 0.10 | 0.08 t per t of cement, about 27 Mt on 334 Mt |
The 65 per cent CaO share and the coal factor are typical values, assumed round; the kiln heat is the GCCA India 2020 average. Values in the right-hand column are calculated from the inputs shown.
Source: Global Cement and Concrete Association India, Roadmap executive summary (2025, read November 2026), India emissions 680 kg CO2 per tonne of cement in 2020 (Scope 1 and 2), clinker factor 0.75, 334 Mt cement in 2019-20, kiln heat 731 kcal per kg of clinker (2020); https://gccassociation.org/wp-content/uploads/2025/03/Excutive-summary-Roadmap.pdf- Kosky et al., Exploring Engineering (Academic Press, 2013) · ch. 11, 'Materials Engineering' (idea source only; all wording, numbers and examples are the author's own)
- Global Cement and Concrete Association India, Roadmap executive summary (2025, read November 2026) · web
5.2 Iron ore is an oxide, so steelmaking is the removal of oxygen, and carbon is both the cheapest remover and the main source of carbon dioxide
Iron ore is rusted iron, Fe2O3 mixed with rock. To get the metal, take away the oxygen; the practical agents are carbon and hydrogen. In the Blast furnace and cast iron, coke and hot air make carbon monoxide, which strips oxygen from the ore on its way up the stack: Fe2O3 + 3 CO → 2 Fe + 3 CO2. At the bottom, near 1,500 degrees C, liquid iron collects. The oxygen steelmaking vessel (Basic oxygen steelmaking) then blows oxygen through the melt to burn its 4 per cent carbon down to a fraction of a per cent. Scrap and sponge iron can be melted in an Electric arc furnace instead; which agent can reduce which metal is plotted on the Ellingham diagram (Thermodynamics of ore reduction (Ellingham diagram)).
The table works out the chemistry for one tonne of iron. The oxygen to be removed is 0.43 t, or 26.9 kmol of atoms. If carbon takes it away as CO2, the least carbon is 13.4 kmol, 161 kg, giving 0.59 t of CO2. Real blast furnaces burn about half a tonne of coke and coal per tonne of hot metal, and the route emits 2.32 t of CO2 per tonne of steel on the 2023 worldsteel average, four times the minimum. The gap exists because carbon also supplies the heat that melts iron and slag, the exit gas is never fully used (reduction is an equilibrium, so the gas must stay rich in CO), and flux and the oxygen vessel add more.
The alternatives follow the same table. Scrap in an arc furnace averages 0.70 t, since the iron is already reduced; sponge iron followed by an arc furnace, 1.43 t. With hydrogen instead of carbon, the oxygen leaves as water, needing 54 kg of hydrogen per tonne of iron, about 2.8 MWh of electrolysis at 52 kWh per kg (Hydrogen direct reduction (green steel); Direct reduction of iron). India made 149 Mt of crude steel in 2024, 59 per cent by electric routes (including induction furnaces). Scrap is limited by supply, hydrogen by cost.
| Step | Working | Result |
|---|---|---|
| Iron atoms | 1,000 kg / 55.85 | 17.9 kmol |
| Ore (pure Fe2O3) | 17.9 kmol / 2 x 159.7 kg | 1.43 t; about 1.6 t of ore at 62 per cent Fe |
| Oxygen to remove | 1.43 t - 1 t = 0.43 t; 1.5 x 17.9 | 26.9 kmol of O atoms |
| Minimum carbon (as CO2) | 26.9 / 2 kmol x 12 kg | 161 kg |
| Minimum CO2 | 13.4 kmol x 44 kg | 0.59 t |
| Real blast furnace route, worldsteel 2023 | BF-BOF per tonne of crude steel | 2.32 t CO2 |
| Sponge iron + arc furnace; scrap + arc furnace | worldsteel 2023 | 1.43 t; 0.70 t CO2 |
| Hydrogen needed instead of carbon | 26.9 kmol x 2.016 kg | 54 kg |
| Electricity to make it (52 kWh per kg, close to the IEA 2024 average) | 54 x 52 | 2.8 MWh |
The first five rows are stoichiometry; the others are reported or assumed values. The minimum CO2 is not the real emission, since carbon also supplies heat and the gas is only partly used. Values in the right-hand column are calculated from the inputs shown.
Source: World Steel Association, World Steel in Figures 2025 (read November 2026), 2023 CO2 intensity 1.92 t per t crude steel; BF-BOF 2.32, DRI-EAF 1.43, scrap EAF 0.70; India 2024 crude steel 149.4 Mt, 58.8 per cent electric routes; https://worldsteel.org/wp-content/uploads/World-Steel-in-Figures-2025.pdf- Kosky et al., Exploring Engineering (Academic Press, 2013) · ch. 11, 'Materials Engineering' (idea source only; all wording, numbers and examples are the author's own)
- World Steel Association, World Steel in Figures 2025 (read November 2026) · web
5.3 A polymer plant is limited by heat removal, not by equilibrium, because making polyethylene releases 3.3 GJ for each tonne
Making polyethylene opens the double bond of each ethylene molecule and links it to the next. The reaction gives out about 93 kJ per mole of ethylene. A tonne of polyethylene is 35.7 kmol of ethylene units, so the heat released is 35.7 x 93, which is 3.3 GJ, as much heat as burning about 0.13 t of coal. With a specific heat of about 2.3 kJ per kg per K, full conversion in a vessel that lost no heat would raise the temperature by 3,300 / 2.3 = 1,400 K, far past where the polymer decomposes.
This is why the plant is built around cooling. In the high-pressure process (High-pressure polyethylene process; Polyethylene (LDPE, LLDPE)) ethylene at 1,500 to 3,000 bar and 150 to 300 degrees C is started with a trace of free-radical initiator in long tubes with cooling jackets. Only 20 to 35 per cent (typical) is converted per pass, so most ethylene is recycled, as in the ammonia loop, but the cause is heat, not equilibrium. In the lower-pressure plants that make High-density polyethylene, a Ziegler-Natta or similar catalyst starts the chains; the reaction runs in a stirred slurry or a gas-phase fluidised bed, and gas circulation carries away the heat. A catalyst, not pressure, then governs chain length and branching, and so stiffness and melting point (Ziegler-Natta polymerisation).
As Part 1 suggests, branched chains pack loosely and give soft film (LDPE); straight chains pack closer into a stiffer grade (HDPE). Molten polymer is thick, so heat leaves it slowly. A reaction can only go as fast as its heat can be removed.
- Kosky et al., Exploring Engineering (Academic Press, 2013) · ch. 11, 'Materials Engineering' (idea source only; all wording, numbers and examples are the author's own)
- Kosky et al., Exploring Engineering (Academic Press, 2013) · ch. 6, 'Chemical Engineering' (idea source only; all wording, numbers and examples are the author's own)
5.4 Process safety starts from the fact that a plant stores energy and hazardous material, and the first defence is to hold less of both
A chemical plant holds fuel, hot fluid, pressure and sometimes poison. Four kinds of event cover most of the list: fire and explosion of a flammable vapour or dust (sugar and flour dust can explode), release of a toxic substance, loss of containment of hot or pressurised fluid, and a runaway reaction.
The runaway shows the principle. Part 4's scale-up table showed the cooling limit. Heat made by a reaction grows exponentially with temperature, since the rate does (Part 2), while heat removed by a jacket grows only in proportion to the temperature gap. Plotted, the first is a rising curve and the second a straight line. Where the line stays above the curve the reactor settles; beyond the last crossing a small warm-up raises the heat made faster than cooling can follow, and the temperature climbs until a relief valve opens or a vessel bursts. An explosion, in the plain sense, is just a reaction so quick that its heat and gas cannot escape. To stay safe, limit what can go wrong: smaller batches, slow feed, weaker concentrations.
Engineers think in layers. First comes inherent safety: a less hazardous substance, less inventory, milder conditions, a simple design. Then passive devices that need no power, such as relief valves, bursting discs and bunds. Then active controls, alarms and trips (the next module, control and systems, covers these), and last procedures and emergency response. Each layer is imperfect, so several are stacked.
Failed layers show in history. In 1974 at Flixborough, England, a temporary pipe in a cyclohexane plant failed and a vapour cloud killed 28 people. In 1984 at Bhopal, water entered a tank holding tonnes of toxic methyl isocyanate, a runaway reaction followed, and the gas killed thousands. Both led to legal change, including India's 1989 hazardous chemical rules. Questions for a fund: how much hazardous material is held, what happens on loss of cooling, and who reviews changes?
- Kosky et al., Exploring Engineering (Academic Press, 2013) · ch. 6, 'Chemical Engineering' (idea source only; all wording, numbers and examples are the author's own)
See it in the table
H₂O WaterFe Iron oreSt SteelCem CementGyp GypsumCaO LimeSu SugarH2 HydrogenO2 OxygenCO2 Carbon dioxideNH3 AmmoniaC2H4 EthyleneLLDPE Polyethylene (LDPE, LLDPE)HDPE High-density polyethyleneCheck your understanding
Which statement about a tonne of clinker is correct?
The reaction CaCO3 -> CaO + CO2 releases 44 g of CO2 for each 56 g of lime, so 0.65 t of CaO means 0.51 t of CO2. Fuel adds a further 0.29 t.The minimum carbon dioxide from reducing one tonne of iron ore with carbon is 0.59 t, yet blast furnace steel emits about 2.3 t. The main reason for the gap is:
Stoichiometry gives only the chemical minimum. Furnace heat, an equilibrium that wastes some of the gas, flux and the oxygen vessel make up the rest.A high-pressure polyethylene reactor converts only 20 to 35 per cent of the ethylene per pass. The limit is mainly set by:
Polymerisation releases 3.3 GJ a tonne, enough to raise the temperature by well over 1,000 K if undiluted. Conversion is held low to keep cooling workable, and unreacted ethylene is recycled.