How a chemical process works: reactions, separations, catalysts and flowsheets: The anatomy of a chemical plant

1.1 A plant is a reactor wrapped in separations and recycle

A useful first picture of a chemical plant is a small reactor surrounded by a much larger amount of equipment whose job is to prepare what goes in and to sort out what comes out. Shreve describes ammonia manufacture as six phases, and the list works for almost any continuous plant: make the reactant gases, purify them, compress them, react them over a catalyst, recover the product, and recirculate what did not react.

Feed preparation means turning a raw material into something the reactor can use. Natural gas for ammonia must first have its sulphur removed, because sulphur poisons the nickel and iron catalysts further down the line. Kent describes a bed of zinc oxide at about 400 to 500 C that captures hydrogen sulphide before the gas goes anywhere near a catalyst.

The reactor is where the chemical change happens. It rarely converts everything in one pass, so the next block is separation: splitting the reactor outlet into product, unreacted feed, by-products and inert gases. Unreacted feed goes back to the reactor inlet. This recycle loop is why a plant with poor conversion per pass can still turn almost all of its raw material into product.

Recycle has a catch. Anything that enters and never reacts, such as methane or argon in an ammonia loop, builds up. Plants therefore bleed off a small purge stream. Kent notes that the purge is sized to hold inerts in an ammonia loop at about 10 to 15%.

After separation comes purification to the purity buyers will pay for, then product handling: cooling, storage, prilling or flaking solids, loading into tankers or pipelines. Each block has its own equipment, energy use and failure modes, and the reactor is often the smallest of them.

• Shreve (1984), Ch 18 Nitrogen Industries · pdf pp. 324-325
• Kent (2012), Ch 6 Industrial Catalysis · pdf pp. 237-238
• Kent (2012), Ch 22 Synthetic Nitrogen Products · pdf pp. 902-903

1.2 Utilities are a second plant that runs alongside the first

Utilities are the services a process consumes without them ending up in the product. Brennan lists the usual ones: fuel, steam, electricity, cooling water, raw water, compressed air, nitrogen and refrigerated coolants. On a large site they often come from a central steam and power station shared by several plants.

Steam is the plant's main way of delivering heat. It is raised in boilers, or recovered from hot process streams, at several pressures. High-pressure steam is hotter and more valuable, and can drive turbines before being let down to lower pressures for heating. Cooling water does the opposite job: it carries heat away from condensers and coolers to a cooling tower, where some of it evaporates. Electricity runs pumps, compressors and, in electrochemical plants such as chlor-alkali, the reaction itself. Brennan notes that ammonia is also power-hungry because synthesis gas must be compressed. Fuel is needed where temperatures exceed what steam can reach, as in reformer furnaces and kilns.

Nitrogen is the odd one out. It is used as an inert blanket: filling the space above flammable liquids and purging equipment so that no air and fuel mixture can form. It is a safety utility as much as a process one.

Utilities matter for two reasons. First, they cost money every hour. Brennan advises adding about 20% to theoretical steam and cooling water figures for losses and contingencies. Second, losing them is dangerous. Kent lists loss of electricity, nitrogen, water, refrigeration, air or steam among the common initiating events of process accidents, because a reactor that suddenly loses cooling or power can run out of control. A plant is designed so that it can shut down safely when any one utility fails.

• Brennan (2020), Ch 4 Operating cost estimation · pdf pp. 107-110
• Kent (2012), Ch 2 Safety Considerations · pdf p. 65

1.3 Raw materials usually dominate cost, so yield is the first lever

When engineers estimate what a product will cost to make, they fill in a cost sheet with the same lines every time: raw materials, utilities, labour, maintenance, overheads and capital charges. Shreve's rule of thumb, written in 1984, is that the single largest cost is usually raw material, and that materials remain more expensive than energy, so a small gain in chemical yield improves profit more than a similar saving in power.

Brennan's illustrative cost sheet for acetic acid made by methanol carbonylation shows how the lines stack up. The plant makes 40,000 tonnes a year and costs $100 million. Methanol at 0.54 tonnes per tonne and carbon monoxide at 0.52 tonnes per tonne together cost $318 per tonne of acid. Utilities (electricity, 2.5 tonnes of steam and 80 cubic metres of cooling water) add $66. But depreciation alone adds $250 a tonne and maintenance another $76, because this is a small plant for a commodity. Total operating cost comes to about $862 per tonne.

Doing the arithmetic, raw materials are 37% of the total, utilities 8%, and fixed costs, most of them tied to capital, are close to half (418 / 862.5 = 48%). For a world-scale plant the capital lines shrink per tonne and raw materials would dominate even more. For a small or new plant, capital dominates.

This gives three questions to ask of any process. How close does it get to the theoretical raw material use? How much energy does it buy, and how much does it recover? And how big is it relative to the best plants, since capital and labour per tonne fall with scale while raw material use per tonne does not, as Brennan points out.

• Shreve (1984), Ch 1-2 · pdf pp. 20-21, 29-30
• Brennan (2020), Ch 4 Operating cost estimation, Table 4.7 · pdf pp. 126-127
• Brennan (2020), Ch 7 Economies of scale · pdf p. 188

See it in the table

Check your understanding

Why does an ammonia synthesis loop need a purge stream?

In Brennan's acetic acid cost sheet, which line is larger than the whole utilities bill?

Why is nitrogen classed as a utility?