Wet-process phosphoric acid: dihydrate reactor-filter, 1500 t/d P2O5

A dihydrate (DH) phosphoric acid line at the scale of one large Moroccan train: 4,830 t/d of a 33% P2O5 phosphate rock (78 wt% fluorapatite, 13% calcite, 9% silica) attacked at 75 C with 98 wt% sulfuric acid at 2.5% sulfate excess, the gypsum filtered off and washed in three countercurrent stages. The phosphoric_acid_reactor block makes 1,500 t/d of P2O5 as a 28.4 wt% P2O5 filter acid at 94.2% recovery; the three losses close it exactly: 1.0% unattacked rock, 3.0% co-crystallized in the gypsum lattice, 1.9% water-soluble P2O5 left in the washed cake by the countercurrent washing equation. It throws 4.79 t of dry gypsum and uses 2.80 t of sulfuric acid per t P2O5. Gypsum is the stable hydrate here, and that is computed, not assumed: the acid's water activity (0.838, from the H3PO4/water model fitted to measured data) sits above the gypsum/bassanite transition activity (0.565, from LLNL log K(T)). The attack releases 50.0 MW on formation enthalpies (-464 kJ per mol of apatite) plus the acids' heats of solution: the 98 wt% acid enters already partly diluted and takes 1.1 MW back, while the phosphoric acid made dissolves at its own strength and adds 3.5 MW (NBS 1982). The flash cooler removes 25.6 MW by evaporating 11.0 kg/s of water, which leaves through the off-gas with the CO2 and SiF4. At the 6 h digestion residence EPA's 1976 phosphoric acid report cites, the attack system holds about 4,060 t of slurry (48 wt% solids). Fluorine is split 7% to the gas, 43% to the cake as CaF2 and 50% dissolved in the acid, the typical dihydrate distribution in the EFMA BAT booklet on phosphoric acid (most of the dissolved share leaves later, in the concentrator). The tilting-pan gypsum filter is sized by Ruth's constant-pressure law: 230 m2 at 6.5 t P2O5/m2/day, limited by cake formation (the cake forms in 18 s of its 18 s sector; the three-stage wash needs 20 s of its 81 s). The vacuum (14.5 inHg), the 3 min cycle and the 18 s formation time are within the ranges in Pillai et al., US 5,185,135; the filtrate viscosity, 3.95 mPa s, is the middle of the 3.3-4.6 mPa s Anouar et al. (Heliyon 2024) measured on wet-process acid, at a temperature they do not state; the wash liquor's is the pure-acid fit at the first wash filtrate's strength. The cake's specific resistance, 1.44e8 m/kg, is back-calculated to the middle of the 5.5-7.8 t P2O5/m2/day Anouar et al. measured on dihydrate slurries. The cloth resistance (1e9 1/m) and the wash/dry sector split are representative, not sourced; at 1e10 1/m no cake reaches 5.5. Scope, stated rather than implied: (1) one block for the attack tanks, flash cooler and filter, so there is no tank-by-tank residence time or crystal size, and the filter's cake resistance is an input, not a prediction from crystal size; (2) the attack efficiency, co-crystallized fraction, fluorine split and cake liquor are plant inputs, not predictions (the 4% insoluble loss, mostly co-crystallized, follows the EFMA booklet's 4-6% DH losses "most of them co-crystallised"); (3) the heats of solution are 298.15 K values, one strength per liquid stream; (4) the water activity reads the P2O5 strength only, not the sulfate, fluoride or impurities that also lower it; and (5) anhydrite, the truly stable phase above about 42 C, is not considered, as in the real process, where gypsum forms metastably.