Aluminium Remelt Loss Is a Geometry Problem: A Surface-Area Bound, the Briquetting Gain for Chips, and Why Direct Foil Remelting Does Not Pay

When aluminium scrap is remelted, part of it is lost as oxide and dross. This paper shows that, to first order, the loss is set by geometry: by the surface-area-to-volume ratio of the scrap before it enters the furnace. A model in which an equivalent oxide skin of 2 micrometres forms over the exposed surface, multiplied by an entrainment factor for turbulent thin scrap and with a contamination term for oily turnings, gives melt losses of 1.02 per cent for ingot, 1.35 per cent for 2 mm sheet, 9.29 per cent for turnings, 13.87 per cent for used beverage cans, and 45.0 per cent (the model cap) for 0.02 mm foil. Compacting turnings into 30 mm briquettes cuts their loss from 9.29 to 1.02 per cent, recovering 82.7 kg of metal per tonne charged, worth $198.48 at a metal price of $2,400 per tonne against a briquetting cost of $55 per tonne — a net gain of $143.48 per tonne. Sorting scrap by geometry class pays whenever the price spread between classes exceeds the sorting cost of $85 per tonne; the computed spread between ingot and can scrap is $308.40 per tonne, so sorting pays by a wide margin. At a scrap purchase price of $1,900 per tonne charged, direct foil remelting is a net loss at both energy prices tested ($40 and $120 per MWh): foil would require a metal price of $3,618.18 per tonne to break even. Direct foil remelting is therefore printed here as a clean kill. SIMULATED, NOT MEASURED.