How a chemical process works: reactions, separations, catalysts and flowsheets: Catalysts: what they change and what they cannot

3.1 A catalyst opens a lower road but does not move the destination

Kent explains catalysis with a road analogy. Reactants must climb an energy barrier, the activation energy, to become products. A catalyst offers a different route with a lower barrier, like a bypass around a slow lorry on a one-lane road. The reactants stick briefly to active sites on the catalyst surface (chemisorption), bonds break and form there, and the product leaves. The catalyst returns to its starting state, so a small amount can process an enormous number of molecules.

What a catalyst cannot do is equally important. Kent stresses that it cannot change the equilibrium concentrations and cannot make a thermodynamically unfavourable reaction happen. It only helps the system reach equilibrium faster. Shreve makes the same point: better catalysts shorten reaction time, but equilibrium is unaffected. Ammonia shows the value of speed. Shreve reports that, at a given gas flow, pure iron gives 3 to 5% conversion, singly promoted iron 8 to 9% and doubly promoted iron 13 to 14%.

The bigger prize is selectivity. Most feeds can react several ways, and a catalyst speeds one route far more than the others. Ammonia burned with air at 900 C should, on thermodynamics alone, give mostly nitrogen gas: Kent gives the equilibrium constant for the nitrogen route as 100,000 times larger than for nitric oxide. Yet a platinum-rhodium gauze delivers nitric oxide at 98% selectivity, the first step to nitric acid. A palladium catalyst would give nitrogen instead.

Wittcoff adds that useful reactions often stop at a local energy minimum, not at true equilibrium. Fully equilibrated ethylene and oxygen would be carbon dioxide and water. Silver lets the reaction stop at ethylene oxide. Catalysts give access to these non-equilibrium products, which is much of what modern petrochemistry sells.

• Kent (2012), Ch 6 How does a catalyst work; Selectivity · pdf pp. 213-218
• Wittcoff (2004), Ch 16 Industrial Catalysis · pdf pp. 553-559, 562
• Shreve (1984), Ch 2 and Ch 18 · pdf pp. 25, 323

3.2 Homogeneous or heterogeneous: the choice is really about separation

A homogeneous catalyst is dissolved in the same phase as the reactants, usually a metal atom wrapped in organic molecules called ligands. A heterogeneous catalyst is a separate solid, typically tiny metal or metal oxide particles spread over a porous support such as alumina. Kent notes that 20 grams of a support with 200 to 300 square metres per gram has an internal surface roughly the size of a football field.

Wittcoff sums up the trade-off. Homogeneous catalysts are reproducible and highly selective, which is why they are favoured where precise chemistry matters, as in pharmaceuticals. But they are hard to get out of the product, and an expensive metal such as rhodium must be recovered, by distillation or ion exchange, for the process to pay. Heterogeneous catalysts are easy to separate, can work at high temperatures and suit continuous fixed or fluidised beds, so they dominate industry. Their weaknesses are diffusion, because molecules must travel into pores to reach the active sites, and heat transfer, because hot spots form inside pellets.

Two cases show why a homogeneous catalyst can win despite the separation cost. Making butyraldehyde from propylene, carbon monoxide and hydrogen (hydroformylation) with a cobalt catalyst needs 3,000 to 4,500 psig and 150 C and gives a 4:1 ratio of the useful straight-chain product to the branched one. A rhodium complex works at 225 psig and 100 C and is more selective for the straight chain. Kent says the milder conditions more than pay for rhodium costing 1,000 times as much as cobalt.

Acetic acid is the other. Wittcoff reports that the older BASF cobalt process needed 250 C and 60 bar with many by-products; Monsanto's rhodium process gives over 99% selectivity on methanol. Methanol carbonylation now makes about 88% of US acetic acid.

• Kent (2012), Ch 6 Structure of heterogeneous catalysts; Homogeneous catalytic reactions · pdf pp. 215-216, 221, 226, 240
• Wittcoff (2004), Ch 16.2 Homogeneous and heterogeneous catalysis · pdf pp. 559-561
• Wittcoff (2004), Ch 10 Acetic acid · pdf pp. 391-392

3.3 Every catalyst dies, and plants are built around that fact

Textbooks say catalysts are not consumed. Wittcoff points out that this is only loosely true: practically every industrial solid catalyst must be taken out at intervals of a few months to a year for regeneration or replacement. Both he and Kent group the causes into three.

Sintering is heat damage. The tiny metal particles slowly migrate and merge into bigger ones, losing surface area. The support can also collapse: Kent notes that high-area gamma alumina, around 200 square metres per gram, turns into alpha alumina of 1 to 5 square metres per gram above about 800 C. Wittcoff describes a nickel catalyst in which this change began at only 400 C in plant service. Sintering is irreversible, so its rate sets catalyst life.

Poisoning is chemical. Something in the feed binds to the active sites and blocks them. Sulphur is the most common poison for nickel, copper and cobalt, and is usually irreversible, which is why ammonia plants desulphurise their gas first and add guard beds. Kent lists carbon monoxide, carbon dioxide and water as poisons for the ammonia catalyst. Not all poisoning is bad: refiners deliberately pre-sulphide hydrotreating catalysts to curb unwanted side reactions.

Fouling is physical. Dust, rust or carbon (coke) coats the surface. Fluid catalytic cracking turns this into a design feature: coked catalyst flows continuously to a second vessel where the carbon is burned off, and the heat preheats the feed. Wittcoff says only this makes the process feasible.

Some catalysts are also hazardous to handle. Kent warns that reduced nickel, copper and iron catalysts react violently with air and must be carefully passivated with small doses of air before they are unloaded.

• Wittcoff (2004), Ch 16.1.3 Catalyst deactivation · pdf p. 558
• Kent (2012), Ch 6 Catalyst deactivation; Commercial applications · pdf pp. 223-226, 237-239
• Kent (2012), Ch 22 Nitric acid · pdf pp. 913, 915

See it in the table

Check your understanding

Which of these can a catalyst NOT do?

Why can a rhodium hydroformylation catalyst beat cobalt even though rhodium costs about 1,000 times more?

Fluid catalytic cracking deals with coking by: