Concept
Conversion, yield, and selectivity are the three metrics that quantify how well a reaction step turns feed into the desired product. Conversion is the fraction of a reactant consumed; selectivity is the fraction of that consumed reactant that goes to the desired product rather than to byproducts; and yield is the fraction of fed reactant that ends up as desired product — the product of the other two.
The three definitions (all on a consistent, stoichiometry-adjusted molar basis):
X = reactant consumed / reactant fed. It is a statement about a reactant — how much of it disappeared.S = desired product formed / reactant consumed (in stoichiometric equivalents). It is a statement about where the consumed reactant went — to product or to byproducts.Y = desired product formed / reactant fed. It is the feed-to-product statement, and it is exactly the product of the other two:Y = X · S
The three are not independent: fix any two and the third follows. Specifying conversion alone leaves selectivity — and therefore yield — undetermined.
Why all three are needed. Conversion and yield coincide only when selectivity is 100%. High conversion with poor selectivity burns feed into byproducts that must be separated and disposed of; high selectivity at low conversion wastes nothing but forces a large recycle. A single headline conversion number hides selectivity losses, which is why the reaction performance of a unit operation is parameterized by conversion and selectivity, not conversion alone.
What they set in the model. These metrics fix the generation and consumption terms of the reactor’s mass balance: conversion sets how much reactant is consumed, selectivity sets how that consumption splits between product and byproduct. They also drive the feed requirement and therefore feedstock cost — a tonne of product made at low yield needs more feed.
Per pass or overall. Each metric can be quoted for one trip through the reactor or for the whole process with recycle. The two are very different numbers and conflating them is a frequent error — large enough that the distinction has its own page.
Ammonia synthesis is the clean case because its selectivity is essentially 100% — the only product of N₂ + 3H₂ → 2NH₃ is ammonia, with no meaningful byproduct — so yield collapses onto conversion. At the running-example per-pass conversion of ~20% (the feed at a stoichiometric 3:1 H₂:N₂ ratio, so H₂ and N₂ convert together), the per-pass yield is also ~20%. This is why the ammonia loop is usually discussed in conversion terms alone: with S ≈ 1, the three metrics reduce to one.
Separately, to show where selectivity bites: in a route where that same 20% conversion came with 80% selectivity, the per-pass yield would be only 0.20 × 0.80 = 16%, and the reactant lost to byproducts would leave as a waste or co-product stream that the success case — ammonia’s clean reaction — simply does not have.