Heavy inorganics: salt, sulphur, nitrogen and phosphorus: From ammonia and rock to fertiliser: nitric acid, urea, ammonium nitrate and phosphates

4.1 Nitric acid: ammonia burns on platinum gauze in a millisecond, then absorption takes the time

The Ostwald process, discovered in 1900, has three steps (Kent). First, 4NH3 + 5O2 → 4NO + 6H2O on a pad of woven platinum-rhodium gauze. In Shreve's example, air with about 10% ammonia passes through about 30 gauze layers. The reaction is over in under a millisecond, and the gas leaves at about 940 C with 94-95% of the ammonia converted to nitric oxide (NO). Second, NO is oxidised to NO₂ as the gas cools, a slow step that speeds up with pressure. Third, NO₂ is absorbed in water in a tall column with cooled trays: 3NO2 + H₂O → 2HNO3 + NO. The product is acid of 50-70%, below the 68.4% azeotrope (the strength at which boiling no longer concentrates it).

The design conflict is that oxidation of ammonia prefers low pressure, and absorption high pressure. Single-pressure plants, at about 10 bar, cost up to 30% less to build and were preferred in the US. Dual-pressure plants burn the ammonia at medium pressure and compress the gas for absorption. They give better ammonia yield and lose less platinum, so they were preferred in Europe, where ammonia was dearer (Kent).

Side reactions turn some ammonia into nitrogen and some into nitrous oxide (N₂O). Kent gives 300-3,500 ppmv of N₂O in absorber tail gas. N₂O is a greenhouse gas, although Kent's chapter discusses only NOx limits, such as 1.5 kg of NOx per tonne in the US and 200 ppm in the EU. These are met by extra absorption and catalytic reduction with ammonia.

Nitric acid plants export steam. About 65% of the world's nitric acid, 53 Mt in 1999, went into ammonium nitrate fertiliser (Kent).

• Kent (2012), Ch 22 Synthetic Nitrogen Products · pdf pp. 910-917
• Shreve (1984), Ch 18 Nitrogen Industries · pdf pp. 332-336
• Dryden (1997), Ch IIE Nitrogen Industries · pdf pp. 138-142

4.2 Urea joins ammonia to the CO2 from the same plant, and ammonium nitrate neutralises ammonia with nitric acid

Urea forms in two steps (Kent). Ammonia and CO₂ combine quickly, giving off heat, to form ammonium carbamate: 2NH3 + CO₂ → NH₂COONH₄. The carbamate then slowly loses water to give urea, NH₂COONH₄ → CO(NH₂)2 + H₂O. This second step absorbs heat and stops well short of completion, at 50-80% conversion of the CO₂ per pass. Shreve describes a reactor at 14 MPa and 170-190 C.

Most of the plant handles the unconverted carbamate, which is very corrosive. Total-recycle plants now dominate. In the Stamicarbon process, the incoming CO₂ strips the reactor effluent at full synthesis pressure, so recovered gases return with only a pump, not a gas compressor, and the heat of carbamate formation raises steam (Shreve). Over 99% of the ammonia and CO₂ become urea. Dryden's figures show why recycle won. A once-through plant used 1.15 t of ammonia and 210 kWh per tonne; total recycle needs about 0.6 t of ammonia, 0.77 t of CO₂ and 145 kWh, close to the theoretical 0.567 t of ammonia and 0.733 t of CO₂. Long heating forms biuret, which harms some crops.

Urea plants sit beside ammonia plants, needing both their ammonia and their CO₂. Kent's rule is that the cash cost of urea is 1.4-1.6 times the price of ammonia. At 46% nitrogen it is the most concentrated solid nitrogen fertiliser.

Ammonium nitrate is simpler. Ammonia gas neutralises 50-70% nitric acid, and the heat of reaction boils off much of the water. An evaporator concentrates the melt to 96-99%, and it is prilled in towers or granulated (Kent). Dryden gives 0.22 t of ammonia and 1.38 t of 60% acid per tonne. At 33-35% nitrogen it acts fast, but it absorbs water from air and becomes explosive when mixed with fuel oil.

• Kent (2012), Ch 22 Synthetic Nitrogen Products · pdf pp. 885, 917-921
• Shreve (1984), Ch 18 Nitrogen Industries · pdf pp. 328-330
• Dryden (1997), Ch IIE Nitrogen Industries · pdf pp. 144-155

4.3 Wet-process phosphoric acid trades sulphuric acid for gypsum, about five tonnes per tonne of P2O5

Phosphate rock is mostly fluorapatite, Ca10(PO₄)6F2. In the wet process the rock reacts with sulphuric acid in a slurry of recycled phosphoric acid and gypsum crystals: Ca3(PO₄)2 + 3H2SO4 + 6H2O → 2H3PO4 + 3CaSO4.2H2O (Kent). The skill is to grow gypsum crystals that filter well without coating the rock. In the dihydrate process the slurry is held at 80-85 C with 2-3% free sulphuric acid. About 10-15 volumes of slurry are recycled for every volume sent to the filter, and reactors hold 2-6 hours of slurry. The acid leaves the filter at 26-32% P₂O₅ and is evaporated to 40%, or to 52-54% merchant-grade acid for trade.

Hemihydrate processes run hotter, at 95-100 C, and stronger, giving 38-42% acid directly. This saves 2.3-3.5 GJ of evaporation steam per tonne (Shreve). Yields through the filter are 95-97% for dihydrate, 91-94% for single-stage hemihydrate and 98-99% for two-stage processes (Kent). Dihydrate still leads: it accepts wet rock slurry, runs reliably and is proven at over 2,000 t/day of P₂O₅.

The waste is enormous. Kent gives about 5 t of dry gypsum per tonne of P₂O₅. It carries P₂O₅, fluorine, low-level radioactivity and traces of heavy metals. Most is stacked; only in Europe and Japan, where natural gypsum is dear, has much gone into building products. Fluorine, about 3.5% of the rock, is partly recovered as fluosilicic acid.

Pure acid once came from the electric furnace, which smelts rock, silica and coke to phosphorus vapour that is burned and hydrated. Shreve's figures are 18,840 MJ of electricity per tonne of 85% acid. By 2001 only one US furnace plant remained. Emission rules, high costs and cheaper purification of wet acid by solvent extraction closed the rest (Kent).

• Kent (2012), Ch 23 Phosphorus and Phosphates · pdf pp. 945, 951-962
• Shreve (1984), Ch 16 Phosphorus Industries · pdf pp. 294-298
• Dryden (1997), Ch IID Fertilizer Industry and IIF Phosphorus Industries · pdf pp. 156-157, 163-165

4.4 DAP, MAP, superphosphates and NPK: the choice is about nutrient per tonne shipped

Fertiliser buyers pay for nutrients, not tonnes, so concentration matters. Single superphosphate (SSP), the oldest product, is made by mixing ground rock with sulphuric acid, about 0.6 kg of acid per kg of rock. It holds only about 20% available P₂O₅, because the gypsum stays in. Kent notes it is cheaper to ship 32% rock to local SSP works than to ship SSP itself. Triple superphosphate (TSP) attacks the rock with phosphoric acid instead, so no gypsum is added, and reaches 45-46% P₂O₅.

Ammonium phosphates took over. In the TVA process, wet-process acid at about 40% P₂O₅ is part-neutralised with ammonia in a preneutraliser to a mole ratio of about 1.4 NH₃ to H₃PO₄. At that ratio the salt is most soluble, so the slurry stays fluid with only 16-20% water. It is sprayed onto a rolling bed of recycled granules in a rotary drum, where more ammonia brings the ratio to 1.85-1.94. Solubility drops, the slurry solidifies on the granules, and the product is dried gently (Kent). The product grade is 18-46-0 (per cent N, P₂O₅ and K₂O). MAP, stopped near a ratio of 1.0, carries 52-54% P₂O₅ and so more phosphate per tonne shipped.

The rise was fast: 0.5 Mt of P₂O₅ as ammonium phosphates in 1960, 16.7 Mt in 1999, and two-thirds of world phosphate fertiliser (Kent). Nitrophosphates attack rock with nitric acid instead of sulphuric, which suits countries short of sulphur; India's Trombay plant used this route (Dryden).

Policy shaped India's mix. Dryden records that no new superphosphate plants were licensed after 1962, to push farmers towards higher-analysis mixed fertilisers. Kent shows India using 28.1 Mt of nutrients in 2010/11, second only to China.

• Kent (2012), Ch 24 Fertilizers and Food Production · pdf pp. 969, 975-981
• Shreve (1984), Ch 16 Phosphorus Industries · pdf pp. 293, 302-303
• Dryden (1997), Ch IID Fertilizer Industry and IIF Phosphorus Industries · pdf pp. 122, 172-176

See it in the table

Check your understanding

Why do European nitric acid plants tend to use dual-pressure designs?

In the TVA DAP process, why is the acid first ammoniated only to a mole ratio of about 1.4?

Roughly how much phosphogypsum does a wet-process plant make per tonne of P2O5, according to Kent?