Carbon-Reduction Limits and the Switch to Electrolytic Extraction

On the Ellingham diagram, the line for 2C + O2 → 2CO has a steep negative slope (free energy of formation gets more favourable as temperature rises, because the reaction increases the moles of gas), while the Fe/FeO line has a much shallower, less temperature-sensitive slope. These two lines cross at a characteristic temperature (around 973 K / 700 degrees C in the standard diagram). BELOW that crossing temperature, the CO-formation line lies below the Fe/FeO line, meaning CO is the more effective reducing agent at low temperature — so CO can reduce FeO below roughly 1000 K (or more precisely below the crossover, often quoted around 600 K in simplified diagrams for this comparison), while carbon itself is not yet the better reductant there. ABOVE the crossing temperature, the C-to-CO line drops below the Fe/FeO line and keeps falling steeply, meaning carbon becomes the more effective reducing agent at high temperature — so C can reduce FeO above roughly 1000 K. This temperature-dependent switch (CO better at low T, C better at high T) is exactly why a blast furnace uses both: CO does much of the reducing work in the cooler upper regions, while carbon itself takes over the reducing role in the hottest lower regions near the tuyeres.

The same crossing-point logic applies to any two metals M and M plotted on an Ellingham-type diagram. At any given temperature, the metal whose oxide-formation line sits LOWER (more negative free energy of formation) has the more thermodynamically stable oxide at that temperature, and is therefore the BETTER reducing agent for the other metal oxide there (since a metal with a lower/more negative free-energy line will spontaneously take oxygen away from a metal whose line sits above it). So if the MxOy line sits below the M'xOy line at temperatures below a crossing point T3, then M is the MORE reducing metal below T3 (not less reducing) — a statement claiming M is LESS reducing than M' below T3 gets this backwards and is the incorrect one to flag. Also remember that above any oxide own decomposition/instability threshold (ΔG > 0), that oxide is no longer thermodynamically stable and will decompose on its own.

Very electropositive metals — aluminium, and especially the alkali and alkaline earth metals like potassium and calcium — form oxides so thermodynamically stable (very low, very negative free energy of formation) that their Ellingham lines sit BELOW the carbon lines across the entire practically achievable temperature range. This means carbon (or CO) can never reduce Al2O3, CaO or K2O under any realistic furnace conditions — there is no crossing point within reach. These metals must instead be extracted by ELECTROLYSIS of a molten compound (for aluminium, electrolysis of alumina dissolved in molten cryolite; for reactive alkali/alkaline earth metals, electrolysis of a molten halide), since electrolysis supplies electrical energy directly rather than relying on a chemical reducing agent competing thermodynamically. Iron, by contrast, is reduced by carbon/CO in the blast furnace precisely because its oxide is comparatively less stable and does have an accessible crossing point with the carbon lines — iron is the standard example of a metal NOT extracted by electrolysis industrially, unlike Na, Mg and Al.