Co-hydrogenation: vegetable oil co-processed in a gasoil hydrotreater

Rapeseed-type vegetable oil (modelled as triolein) co-processed at 7.3 wt% in a sulfur-bearing gasoil (n-hexadecane carrier with a dibenzothiophene slice) over a sulfided hydrotreating catalyst at 70 bar and 613-633 K, the way refineries put renewable carbon into diesel without a separate HVO unit. The triglyceride leaves by three routes, each written as its own conversion reactor with an exact element balance: hydrodeoxygenation (triolein + 15 H2 -> 3 n-C18 + propane + 6 H2O), decarboxylation (triolein + 6 H2 -> 3 n-C17 + propane + 3 CO2) and decarbonylation (triolein + 9 H2 -> 3 n-C17 + propane + 3 CO + 3 H2O); the glycerol backbone becomes propane in all three, and the double bonds are saturated in the same step. The gasoil is desulfurized in parallel (dibenzothiophene + 2 H2 -> biphenyl + H2S). The effluent goes through a hot and a cold high-pressure separator; an amine wash (a fixed-recovery separator) takes H2S and CO2 out of the recycle gas, a 5% purge bleeds the CO and propane the wash cannot, and the recycle compressor closes the hydrogen loop. The separator liquids lose their water in a knock-out, are let down to 3 bar, flashed, and steam-stripped into the co-processed diesel.

What the split costs in hydrogen is the point of the example: at 60% HDO, 30% decarboxylation and 10% decarbonylation the oil consumes 11.7 mol H2 per mol triglyceride, against 15 for pure HDO and 6 for pure decarboxylation; with the 2.97 mol/s the desulfurization takes, chemical consumption is 14.67 mol/s, and the hydrogen leaving in the purge, the flash gas and the stripper overhead is set against the 30 mol/s makeup. The split leaves its fingerprint in the product: the diesel carries n-C17 and n-C18 in a 0.67 mole ratio, which is the measurement a refinery uses to infer the route.

Honesty notes: (1) The pathway split is a stated screening INPUT, not a prediction. The co-processing literature reports both routes operating on sulfided NiMo/CoMo, with their balance set by catalyst and conditions (Huber, O'Connor & Corma, Appl. Catal. A 329 (2007) 120; Donnis et al., Top. Catal. 52 (2009) 229; Kubicka & Kaluza, Appl. Catal. A 372 (2010) 199); the 60/30/10 numbers are illustrative within that picture, and the n-C17/n-C18 ratio of a real product is what should replace them. (2) The triglyceride record carries no ideal-gas heat-capacity correlation and no heat of formation, so no heat of reaction is booked: every reactor is temperature-specified and its duty is sensible heat only, not the exotherm a real bed removes with quench gas. (3) Water-gas shift and CO methanation, which move CO and CO2 around in a real loop, are not modelled; CO accumulates until the purge carries it out. (4) The oil is one triglyceride and the gasoil one paraffin plus one sulfur species; there is no aromatic saturation, no cracking and no hydroisomerization. The atom balances are exact; the hydrogen bookkeeping follows from them.